NX Nastran 5 Release Guide

NX Nastran 5 Release Guide
Proprietary & Restricted Rights Notice
2007 UGS Corp. All Rights Reserved. This software and related documentation are proprietary to UGS Corp. NASTRAN is a registered trademark of the National Aeronautics and Space Administration. NX Nastran is an enhanced proprietary version developed and maintained by UGS Corp. MSC is a registered trademark of MSC.Software Corporation. MSC.Nastran and MSC.Patran are trademarks of MSC.Software Corporation. All other trademarks are the property of their respective owners.
NX Nastran 5.0 Release Guide
Contents
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-1 Dynamics . . . . . . . . . . . . . . . . . . . . . . Numerical Capabilities . . . . . . . . . . . . Multi-body Dynamic Software Interfaces New Optimization Option . . . . . . . . . . Advanced Nonlinear . . . . . . . . . . . . . . Linear Contact Enhancements . . . . . . . Strength Ratio Output for Composites . Miscellaneous Enhancements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111111112 2 3 3 3 4 5 5
Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-1 SPC/SPCD Enforced Motion . . . . . . . . . . . . . . Mode Acceleration Method . . . . . . . . . . . . . . . Random Response Output . . . . . . . . . . . . . . . DDAM Process Automation . . . . . . . . . . . . . . Elemental Energy Output . . . . . . . . . . . . . . . Modal Frequency Response in Rotor Dynamics Modal Contribution Computation . . . . . . . . . . Direct Structural Damping Matrix Input . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2- 2 2-16 2-19 2-25 2-33 2-44 2-50 2-53
Numerical Capabilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-1 Sparse Data Recovery for Modal Solutions . . . . . Distributed Memory Parallel Improvements . . . Sparse Cholesky Technique for Linear Solutions Eigenvalue Analysis Improvement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33332 2 5 6
Multi-body Dynamic Software Interfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-1 RecurDyn Flex Input File Creation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4- 2 ADAMS Stress Recovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-10 Component Mode Reduction of the Residual Structure Procedure for Flexbody Solutions . . 4-23 New Optimization Option . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-1 Input Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5- 2 Advanced Nonlinear . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-1 Contact Improvements . . . . . . . . . . . . . . . . . . Improved Convergence for Contact with Friction Surface-to-Surface Glue Support . . . . . . . . . . . Element Birth and Death . . . . . . . . . . . . . . . . . Bolt Preload Capability . . . . . . . . . . . . . . . . . . Shell Thickness Result Recovery . . . . . . . . . . . Iterative Solution Improvement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6- 2 6- 3 6-22 6-24 6-26 6-29 6-30
Contents
Linear Contact Enhancements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-1 New Options to Improve Contact Solution Accuracy Support of Contact in SOLs 103, 111, and 112 . . . . Superelements with Contact . . . . . . . . . . . . . . . . Shell Element Z-Offset with Contact . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77772 6 6 6
Strength Ratio Output for Composites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-1 Laminate Strength Ratio Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8- 2 Miscellaneous Enhancements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-1 Thermal Expansion of Rigid Elements Bolt Preload Analysis . . . . . . . . . . . . Improvements to Surface Glue . . . . . . Parameter Specication Improvements Punch Output Corrections . . . . . . . . . PARAM K6ROT Update . . . . . . . . . . AUTHQUEUE Update . . . . . . . . . . . Documentation Improvements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9- 2 9- 3 9- 6 9- 9 9-10 9-12 9-12 9-12
Upward Compatibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-1 Updated and New DMAP Modules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10- 1 Updated and New Datablocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10- 2 Updated and New Subdmaps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10- 2 System Description Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-1
Chapter
Introduction
This section presents an overview of the features introduced in NX Nastran 5.
1-1
Chapter 1
Introduction
1.1
Dynamics
The constraint mode method of enforced motion, which was referred to as the relative method in NX Nastran 4.1, uses a formulation in which the response output is calculated with both the normal mode shapes, and the constraint mode shapes. A change in the default has occurred in NX Nastran 5. The absolute displacement formulation is used as the default SPC/SPCD method. A new mode acceleration formulation with improved performance has been added in this release to supplement the current algorithm. The improved performance makes mode acceleration a more attractive option when increased accuracy is desired. Up until the release of NX Nastran version 4.0, only auto spectrum PSD functions of response could be output for random results. It was required to use DMAP alters to obtain cross spectrum results. In NX Nastran 4.1, the option to output cross-power spectral density functions became available. Now with NX Nastran 5.0, it is possible to also request cross-correlation functions. The Dynamic Design Analysis Method (DDAM) is a list of procedures to determine modal shock response. To complete DDAM analysis procedures in previous releases of NX Nastran, the use of dmap alter les were required. Now in NX Nastran 5,these DDAM procedures have been streamlined and automated, thus eliminating the alter requirement. The element strain (ESE), kinetic (EKE), and damping (EDE) case control commands have been improved to allow the output of element energy information in SORT2 complex format. This format is suitable for plotting complex function data from frequency domain dynamic solutions. In previous releases, you could only output the real magnitude of element energy, thus no phase information was written. In addition, only SORT1 format could be output, which is not convenient for function plotting. Rotor dynamics can now be used in a modal frequency response solution, SOL 111, to calculate the dynamic response of a rotating system. This new response calculation adds to the existing rotor dynamics capability, direct complex mode solution, SOL 110, whose output is used to create Campbell diagrams. When the modal method is used for dynamic response calculations, understanding which modes contribute to the response helps understand the dynamic behavior of the simulated system, and can provide insight as to how to improve the dynamic performance. In this release a new capability has been introduced to compute the modal contributions in modal frequency response. A structural damping matrix can now be included with a model in NX Nastran 5 using the new K42GG case control command. This ability adds to the other direct matrix capabilities of stiffness, mass, and damping matrices using DMIG bulk entries.
1.2
Numerical Capabilities
A new sparse data recovery option is available for the modal frequency response analysis (SOL 111), modal transient response analysis (SOL 112), and optimization (SOL 200) as the new default. For SOL 111 and 112, this feature reduces the cost of matrix-multiplications inside DDRMM modules when large amount of data are requested to recover in the modal analysis. Similarly, SOL 200 utilize the partitioning of eigenvector matrix in order to reduce the cost of matrix-matrix multiplies.
1-2
Introduction
The parallel processing method Distributed Memory Parallel (DMP) has been enhanced in NX Nastran 5 with the following additions: The normal mode calculation portion of an optimization solution (SOL 200) can now use DMP. A new linear statics (SOL 101) DMP option, Load Domain Static Analysis (LDSTAT), now exists to decrease the solution times when large numbers of load cases exist.
This NX Nastran release includes support for a new sparse Cholesky decomposition. All users of Cholesky decomposition (including the Givens and Householder eigensolver methods) will utilize sparse decomposition by default, which can be considerably faster than the non-sparse Cholesky decomposition method used in earlier NX Nastran versions. NX Nastran Version 5.0 introduces the new REDMULT performance option for use when solving vibration problems with the Lanczos method. This option reduces the cost of matrix-vector multiplies inside the READ module when the mass matrix involved is relatively dense, which can occur when a large number of MFLUID is present in the model.
1.3
Multi-body Dynamic Software Interfaces
This release includes interoperability between NX Nastran and the FunctionBay RFI (RecurDyn Flex Input) le product. You can now create a RecurDyn Flex Input le (RFI) directly from NX Nastran. The RFI contains the reduced order matrices from the results of a NX Nastran non-restart SOL 103 analysis. The RFI can be imported into RecurDyn and used to represent a exible component in a multi-body dynamics analysis. This direct RFI export capability streamlines the process of creating exible components from FE models, making it possible to obtain more accurate results from multi-body simulations. NX Nastran 5 enhances the ADAMS MNF le creation process with a new results recovery capability. The results from an ADAMS multi-body dynamics analysis, along with an optional component modal denitions le (OUTPUT2 format), are used in a consecutive NX Nastran SOL 103 results recovery solution. The set-up of a ex body modal solution in NX Nastran for export to ADAMS MNF or Recurdyn RFI les requires special considerations for the modal solution. This is because ex bodies will be attached to other components in the multi-body dynamic (MBD) simulation and local exibility effects at the connection locations are thus important. A modal solution method called Component Mode Reduction of Residual Structure (CMR of RS) is recommended for ex body solutions because it includes both global and local effects. The details of this method are presented in the multi-body dynamic chapter.
1.4
New Optimization Option
NX Nastran 5 introduces a new optimizer option, UGS-ADS, which is based on public domain ADS code.
1.5
Advanced Nonlinear
The CTDISP option has been created on the NXSTRAT bulk entry to prevent contact conditions from updating.
1-3
Chapter 1
Introduction
A new option on the NXSTRAT bulk entry, CTDAMP, has been created to stabilize the portions of the model experiencing rigid body motion, thus helping the solution to continue and converge. A new option when INIPENE=3 is available which will include both penetration and now gaps. This option is particularly benecial when contact conditions are dened on the faces of concentric cylinders. In NX Nastran 4.1, a new contact segment option became available for SOL 601 which improved contact results. This improvement matches the order of the contact segments with the order of the elements. The restrictions which existed in NX Nastran 4.1 when CSTYPE=1 have been eliminated in NX Nastran 5. The contact algorithm has been improved in NX Nastran 5 to be more robust for contact with friction. The option to glue element faces together during a 601 solution is available in NX Nastran 5. The glue option connects predened surfaces together and prevents relative motion in all directions. Predened regions of element free faces are used to detect where the glue elements are created. An option to dene element birth and death times for a specic set of elements is available in NX Nastran 5. NX Nastran 5 offers an automated method to simplify the multisolution process of bolt preloads with SOL 601. A new shell thickness result output option is available when using advanced nonlinear to solve large strain problems. A new 3D-iterative solution option is now available in NX Nastran 5 for SOL 601 to efciently solve large models containing mainly higher order 3-D solid elements (e.g., 10-node CTETRA, 20-node CHEXA, etc.).
1.6

Linear Contact Enhancements
Two new contact solution parameters, INTORD and REFINE, are now available in NX Nastran 5 on the BCTPARM bulk entry to improve the accuracy of the contact solution. Now in NX Nastran 5, a contact denition can be included in a normal mode solution (SOL 103), and in an optional modal dynamic response calculation (SOLs 111 and 112). In NX Nastran 5, the use of superelements is now permitted in solution sequences which support contact (SOLs 101, 103, 111 and 112). The only requirement is that the contact denition must occur in the residual structure. Now in NX Nastran 5, the linear contact solution can include the shell element ZOFFS when evaluating the contact surfaces.
1-4
Introduction
1.7
Strength Ratio Output for Composites
Strength ratio is now output together with the failure index when using the PCOMP bulk entry. Strength ratio is a more direct indicator of failure than failure index since it demonstrates the percentage of applied load to the failure criteria.
1.8

Miscellaneous Enhancements
The option to include rigid elements in thermal expansion calculations is now available. When analyzing preloaded bolts, you may be interested in obtaining the stresses due to the preload condition alone, or due to a combination of the bolt preload and service load. You can manually determine the preloaded bolt condition by using equivalent thermal loads, although using this method is approximate and typically requires many solution iterations when multiple bolts exist. NX Nastran 5 offers an automated method to simplify this multisolution process. The surface-to-surface glue capability was supported by the global iterative solver in NX Nastran 4.1, but not by the element iterative solver. Now in NX Nastran 5, the surface-to-surface glue denitions can be included when using the element iterative solver. New glue parameters, INTORD and REFINE, are available on the new BGPARM bulk entry to improve the accuracy of the surface-to-surface glue condition. Now in NX Nastran 5, the surface-to-surface glue capability can include the shell element ZOFFS when evaluating the glue surfaces. NX Nastran 5 now allows the specication of PARAM statements in the nast5rc (UNIX) and nast5.rcf (WINDOWS). Parameters can also be assigned a user dened keyword. The new keyword can then be used to specify a value for the parameter on the command line or in the nastran resource le. The keywords are dened in the nastran.params le in the architecture directory. In previous releases of NX Nastran, the algorithm which writes punch les incorrectly assumed that the rst item in an entry was either an integer or real, and that all consecutive items were real. These were sometimes invalid assumptions since there are occasions in which the rst item in a data entry is a character string, and there are also occasions in which some consecutive items are integers and not real. This has been corrected in NX Nastran 5. The default for parameter K6ROT has been modied from 0 to 100 in NX Nastran 5. Assigning PARAM K6ROT to 100 has shown to be a good, general penalty stiffness value which will not adversely effect results, yet will properly constrain the rotation degrees-of-freedom. When an NX Nastran job fails because of a failed license request, an option to have the job retried automatically became available in NX Nastran 4.1. NX Nastran will retry a failed license request job every minute up to the value of the AUTHQUEUE keyword. The AUTHQUEUE keyword default was 20 minutes in NX Nastran 4.1. Based on customer feedback, the default has been modied to 0 in NX Nastran 5. You must now explicitly dene the AUTHQUEUE keyword in the nast5rc/nast5.rcf les to take advantage of this capability. The following documentation improvements have occurred in NX Nastran 5:
1-5
Chapter 1
Introduction
The NX Nastran documentation is now available in both PDF and HTML format. A single bookshelf.html provides the links to the PDF and the HTML documentation. HTML search is available. You can search a single, or all HTML documents by selecting the magnifying glass icon located at the top left corner of the bookshelf page. The new NX Nastran Parallel Processing Users Guide is included with the NX Nastran documentation. This guide provides information on NX Nastran parallel methods for Linux and UNIX systems including parallel processing basics, detailed computational methods, an overview of execution methods, and examples using various solution sequences. The NX Nastran Numerical Methods Users Guide has been rewritten in this release.
1-6
Chapter
Dynamics
2-1
Chapter 2
Dynamics
2.1
SPC/SPCD Enforced Motion
The SPC/SPCD method of enforced motion allows you to directly specify displacements, velocities or accelerations via SPC/SPC1/SPCD data, thus eliminating the need to employ the large mass method or Lagrange Multipliers. The SPC/SPCD method is preferred because it is more straightforward to use than other methods. Prior to NX Nastran 4.1, the SPC/SPCD method was based on the absolute displacement formulation which gives unintended consequences when used with modal damping. Specically, external and internal damping forces are created, when generally modal damping is intended to simulate only internal damping. The external damping forces can result in noticeable differences in low frequency response compared to systems with only internal damping. Another concern with the absolute displacement formulation is that residual vectors are generally needed for accurate stress, strain, and force response. A new constraint mode method was introduced in NX Nastran 4.1 that did not suffer from these effects. The new method, referred to as the "relative displacement method" in NX Nastran 4.1, uses a formulation in which the response output is calculated from both normal mode shapes and constraint mode shapes. In contrast, the original absolute method uses only normal mode shapes. Residual vectors can be used with the constraint mode method but are not critical for accuracy. The selection of the enforced motion method is enabled by a new system cell 422 (ENFMOTN). The original absolute method is invoked with ENFMOTN=1 and the new constraint mode method is invoked with ENFMOTN=0. Since many users are dependent on the original absolute method as part of their workow, it is the default for NX Nastran 5. You may prefer the constraint mode method when you use modal damping. The two methods give almost identical results when no modal damping is present and residual vectors are used with the absolute displacement method. However, if modal energy (via the MODALE case control command) or modal contributions (via the MODCON case control command) are requested, you must use the absolute method. MODALE and MODCON are not supported for the constraint mode method. Another issue with the constraint mode method is its sensitivity to very high acceleration inputs. For example, an enforced displacement that has a large change in value over a small time step (almost a discontinuity) will produce large acceleration inputs. This can produce a bump in the response calculations around the time that this discontinuity occurs. It is recommended that such large acceleration inputs not be used with the constraint mode method, especially occurring over a single time step. If such a step change is required, it is recommended that a ne time step be used to reduce the high acceleration affect and, thus lesson the bump that occurs in the area of the discontinuity. The describers ABS/REL were added in version 4.1 to the DISPLACEMENT, VELOCITY, and ACCELERATION case control entries. These describers allow you to request the total response output (ABS), or the response output relative to the enforced motion (REL). Having these options on each output entry allows you the exibility to request, for example, relative displacement output with absolute acceleration output in the same run. The new describers are also supported on a subcase level to allow for different output requests in different subcases. The new system cell 422, along with the updated DISPLACEMENT, VELOCITY, and ACCELERATION case control commands are included in the Input Updates section below. Sparse data recovery has also been introduced for the constraint mode method. This technique is analogous to the sparse data recovery for modal solutions discussed in section 3.1, and is activated by the same system cell. The difference is that this particular sparse recovery is applied much earlier in the analysis procedure to reduce the size of matrices and, thus, computation time. This recovery process keys off the output requests and currently supports the following case control entries: DISPLACEMENT, VELOCITY, ACCELERATION, OLOAD, STRESS/ELSTRESS, STRAIN/ELSTRAIN, FORCE/ELFORCE, EDE, EKE, ESE, XYPEAK, XYPRINT, XYPLOT, and XYPUNCH. The sparse data recovery capability is currently automatically deactivated if
2-2
Dynamics
the following output requests are made: SPCFORCE, MPCFORCE, GPSTRESS, GPSTRAIN, GPFORCE, OUTPUT(PLOT), and OUTPUT(POST). Also, if a PARAM,POST request is made, the sparse data recovery is deactivated since most post processors require the full set of results. If you plan on running a mini restart (scr=mini) in order to recover additional results later from a constraint mode enforced motion analysis, you must deactivate the sparse data recovery (i.e. system cell 421/SPARSEDR=0; see section 3.1 for more information). Otherwise, only results which were requested in the original run will be output; any additional result requests in the mini restart run will be zero. Since the sparse data recovery method is so much more efcient, it is advised to not use the mini restart method when sparse data recovery is active. Either request all necessary outputs in the initial run or use a regular restart.
Enforced Motion Example

The following example demonstrates the improvements of the new constraint mode method over the absolute method.
Figure 2-1. Input is enforced acceleration: The geometry is a cantilever which is modeled with shell elements. One end of the cantilever is made rigid with an RBE2 element. The independent grid (Grid 100) of the RBE2 is restrained by an SPC in all six DOF. In addition, the z DOF (direction normal to the plane of the shells) is an SPCD DOF which identies it as the location of input motion. A random PSD analysis was performed with a constant base acceleration input spectrum of 100.0 across the frequency range from 1 to 1000 Hz. Results have been computed three different ways: using the old absolute displacement formulation, the new constraint mode formulation, and the large mass method. Figure 2-2 shows the resulting PSD acceleration at the mid point, grid 32. The acceleration response is very similar for all three methods. Figure 2-3 shows the stress response at a location near the base motion input. The results for the new relative displacement formulation and the large mass method are very similar and accurate. However, the absolute formulation result is only accurate above 10 Hz. Below 10 Hz the predicted response deviates noticeably from the correct response. The new constraint mode formulation is always more accurate than the absolute formulation, and it will match the response of the large mass method. The absolute formulation method gives good accuracy for most responses, but it is not as accurate for responses involving relative displacement quantities (such as stress, strain, or force), particularly when signicant damping exists in the model.
2-3
Chapter 2
Dynamics
Vertical Axis = Acceleration Horizontal Axis = Frequency Absolute Displacement Method Constraint Mode Method Large Mass Method Figure 2-2. Acceleration Response
2-4
Dynamics
Demonstrates the effect of external damping Vertical Axis = Stress Horizontal Axis = Frequency Absolute Displacement Method Constraint Mode Method Large Mass Method Figure 2-3. Stress Response The le used in this example plate_111.dat can be found in the install location NX_Nastran_Install_Path/nxn5/nast/tpl.
2-5
Chapter 2
Dynamics
Mathematical Details for Constraint Mode Method

The original SPC/SPCD enforced motion formulation uses an absolute displacement formulation per the equation:
Equation 2-1. Here the set nomenclature is simplied with f representing the free physical degrees of freedom (DOF), and s representing the SPCD enforced motion DOF, thus uf = absolute displacements of the f-set DOF us = applied enforced motion of the s-set DOF M, B, K, and are the mass, viscous damping, stiffness, and structural damping matrices. The vector P is the applied force load. The structural damping matrix, , consists of the uniform and elemental structural damping components so that
Equation 2-2. where G is the uniform structural damping coefcient, GE are elemental structural coefcients, and KE are elemental stiffness matrices. The formulation for the constraint mode method is obtained by separating the absolute displacement into a static enforced motion component, displacement component, vf, as , and a dynamic relative
Equation 2-3. Substituting into the equation of motion, a new equation of motion in terms of the dynamic displacement component is obtained
Equation 2-4. The above equation is solved for vf and then the absolute displacement is obtained by adding the static enforced motion displacement. Since the constraint mode method uses static mode shapes, it is not as critical that residual vectors be computed as for the absolute displacement method. But it is still possible to use residual vectors with the constraint mode method. In that case, residual vectors are computed for inertial, damping, and structural damping loads. Thus the residual vectors are computed from the following terms
2-6
Dynamics
Equation 2-5.
Input Updates for New Constraint Mode Method

New system cell 422ENFMOTN: 422 ENFMOTN Controls which formulation is used for enforced motion response analysis (and mode acceleration, if requested). 0: Constraint mode method of enforced motion formulation (and new mode acceleration method) 1: Absolute displacement enforced motion formulation (and old mode acceleration method) (Default). (ABS/REL option on DISPLACEMENT, VELOCITY, and ACCELERATION case control entries is disabled). Updated DISPLACEMENT, VELOCITY, and ACCELERATION case control commands:
DISPLACEMENT
Displacement Output Request
Requests the form and type of displacement or pressure vector output. Note: PRESSURE and VECTOR are equivalent commands.
FORMAT:
EXAMPLES:
DISPLACEMENT=5 DISPLACEMENT(REAL)=ALL DISPLACEMENT(SORT2, PUNCH, REAL)=ALL

DESCRIBERS:
Describer SORT1
Meaning Output will be presented as a tabular listing of grid points for each load, frequency, eigenvalue, or time, depending on the solution sequence.
2-7
Chapter 2
Dynamics
Describer SORT2 PRINT PUNCH PLOT REAL or IMAG
Meaning Output will be presented as a tabular listing of load, frequency or time for each grid point. The printer will be the output medium. The punch le will be the output medium. Generates, but does not print, displacement data. Requests rectangular format (real and imaginary) of complex output. Use of either REAL or IMAG yields the same output. Requests polar format (magnitude and phase) of complex output. Phase output is in degrees. For enforced motion dynamic analysis, displacement results will be output as absolute displacement. For enforced motion dynamic analysis, displacement results will be output relative to the enforced motion input. Requests the power spectral density function be calculated for random analysis post-processing. The request must be made above the subcase level and RANDOM must be selected in the Case Control. See remark 8. Requests the autocorrelation function be calculated for random analysis post-processing. The request must be made above the subcase level and RANDOM must be selected in the Case Control. See remark 8. Requests the cumulative root mean square function be calculated for random analysis post-processing. Request must be made above the subcase level and RANDOM must be made in the Case Control. See remark 8. Requests all of PSDF, ATOC, and CRMS be calculated for random analysis post-processing. The request must be made above the subcase level and RANDOM must be selected in the Case Control. See remark 8. Writes random analysis results to the print le. (Default) See remark 8. Disables the writing of random analysis results to the print le. See remark 8. Writes random analysis results to the punch le. See remark 8.
PHASE ABS REL PSDF
ATOC
CRMS
RALL
RPRINT NORPRINT RPUNCH
2-8
Dynamics
Describer ALL NONE n
Meaning Displacements for all points will be output. Displacement for no points will be output. Set identication of a previously appearing SET command. Only displacements of points with identication numbers that appear on this SET command will be output. (Integer>0)
REMARKS:
1. Both PRINT and PUNCH may be requested. 2. The defaults for SORT1 and SORT2 depend on the type of analysis: SORT1 is the default in static analysis, frequency response, steady state heat transfer analysis, real and complex eigenvalue analysis, utter analysis, and buckling analysis. SORT2 is the default in transient response analysis (structural and heat transfer). SORT2 is not available for real eigenvalue (including buckling), complex eigenvalue, or utter analysis. If SORT1 is selected in a transient solution for one or more of the commands ACCE, DISP, ENTH, FORC, HDOT, MPCF, OLOA, SPCF, STRA, STRE, and VELO then the remaining commands will also be output in SORT1 format. If SORT2 is selected in a frequency response solution for one or more of the commands ACCE, DISP, FORC, MPCF, OLOA, SPCF, STRA, STRE, and VELO then the remaining commands will also be output in SORT2 format. XY plot requests will force SORT2 format thus overriding SORT1 format requests.
3. VECTOR and PRESSURE are alternate forms and are entirely equivalent to DISPLACEMENT. 4. DISPLACEMENT = NONE overrides an overall output request. 5. The PLOT option is used when curve plots are desired in the magnitude/phase representation and no printer output request is present for magnitude/phase representation. 6. The units of translation are the same as the units of length of the model. Rotations are in units of radians. 7. Displacement results are output in the global coordinate system (see eld CD on the GRID Bulk Data entry). 8. The option of PSDF, ATOC, CRMS, and RALL, or any combination of them can be selected for random analysis. The results can be either printed to the .f06 le, punched to the punch le, or output to both les using RPRINT and RPUNCH.
2-9
Chapter 2
Dynamics
9. When doing enforced motion dynamic analysis and relative output is requested (using the REL describer), the output will be relative to the input as described by the equation:
where uf = absolute displacement yf = relative displacement us = enforced motion.

REMARKS RELATED TO SOLS 601 AND 701:
1. Output is restricted to REAL format. IMAG, PHASE, PSDF, ATOC and RALL are ignored. 2. Displacements, velocities and accelerations must be output for the same set of grid points if requested. Output requested for set n in this command will be combined with the sets requested in the VELOCITY and ACCELERATION commands, and displacements will be output at the grid points of the combined set.
VELOCITY
Velocity Output Request
Requests the form and type of velocity vector output.
FORMAT:
EXAMPLES:
VELOCITY=5 VELOCITY(SORT2,PHASE,PUNCH)=ALL
DESCRIBERS:
Describer SORT1
Meaning Output will be presented as a tabular listing of grid points for each load, frequency, eigenvalue, or time, depending on the solution sequence. Output will be presented as a tabular listing of frequency or time for each grid point.
SORT2
2-10
Dynamics
Describer PRINT PUNCH PLOT REAL or IMAG
Meaning The printer will be the output medium. The punch le will be the output medium. Generates, but does not print, velocities. Requests rectangular format (real and imaginary) of complex output. Use of either REAL or IMAG yields the same output. Requests polar format (magnitude and phase) of complex output. Phase output is in degrees. For enforced motion dynamic analysis, velocity results will be output as absolute velocity. For enforced motion dynamic analysis, velocity results will be output relative to the enforced motion input. Requests the power spectral density function be calculated for random analysis post-processing. The request must be made above the subcase level and RANDOM must be selected in the Case Control. See remark 7. Requests the autocorrelation function be calculated for random analysis post-processing. The request must be made above the subcase level and RANDOM must be selected in the Case Control. See remark 7. Requests the cumulative root mean square function be calculated for random analysis post-processing. Request must be made above the subcase level and RANDOM must be made in the Case Control. See remark 7. Requests all of PSDF, ATOC, and CRMS be calculated for random analysis post-processing. The request must be made above the subcase level and RANDOM must be selected in the Case Control. See remark 7. Writes random analysis results to the print le. (Default) See remark 7. Disables the writing of random analysis results to the print le. See remark 7. Writes random analysis results to the punch le. See remark 7. Velocity for all solution points will be output. Velocity for no solution points will be output.
PHASE ABS REL PSDF
ATOC
CRMS
RALL
RPRINT NORPRINT RPUNCH ALL NONE
2-11
Chapter 2
Dynamics
Describer n
Meaning Set identication of a previously appearing SET command. Only velocities of points with identication numbers that appear on this SET command will be output. (Integer>0)
REMARKS:
1. Both PRINT and PUNCH may be requested. 2. Velocity output is only available for transient and frequency response problems. 3. See DISPLACEMENT case control card for a discussion of SORT1 and SORT2. 4. VELOCITY=NONE overrides an overall output request. 5. The PLOT option is used when curve plots are desired in the magnitude/phase representation and no printer request is present for magnitude/phase representation. 6. Velocity results are output in the global coordinate system (see eld CD on the GRID Bulk Data entry). 7. The option of PSDF, ATOC, CRMS, and RALL, or any combination of them can be selected for random analysis. The results can be either printed to the .f06 le, punched to the punch le, or output to both les using RPRINT and RPUNCH. 8. When doing enforced motion dynamic analysis and relative output is requested (using the REL describer), the output will be relative to the input as described by the equation:

1. Output is restricted to REAL format. IMAG, PHASE, PSDF, ATOC and RALL are ignored. 2. Displacements, velocities and accelerations must be output for the same set of grid points if requested. Output requested for set n in this command will be combined with the sets requested in the DISPLACEMENT and ACCELERATION commands, and velocities will be output at the grid points of the combined set.
2-12
Dynamics
ACCELERATION
Acceleration Output Request
Requests form and type of acceleration vector output.
FORMAT:
EXAMPLES:
ACCELERATION=5 ACCELERATION(SORT2, PHASE)=ALL ACCELERATION(SORT1, PRINT, PUNCH, PHASE)=17

DESCRIBERS:
Describer SORT1
Meaning Output will be presented as a tabular listing of grid points for each load, frequency, eigenvalue, or time, depending on the solution sequence. Output will be presented as a tabular listing of frequency or time for each grid point. The printer will be the output medium. The punch le will be the output medium. Computes, but does not print or punch, acceleration output. Requests rectangular format (real and imaginary) of complex output. Use of either REAL or IMAG yields the same output. Requests polar format (magnitude and phase) of complex output. Phase output is in degrees. For enforced motion dynamic analysis, acceleration results will be output as absolute acceleration. For enforced motion dynamic analysis, acceleration results will be output relative to the enforced motion input. Requests the power spectral density function be calculated for random analysis post-processing. The request must be made above the subcase level and RANDOM must be selected in the Case Control. See remark 7.
SORT2 PRINT PUNCH PLOT REAL or IMAG PHASE ABS REL PSDF
2-13
Chapter 2
Dynamics
Describer ATOC
Meaning Requests the autocorrelation function be calculated for random analysis post-processing. The request must be made above the subcase level and RANDOM must be selected in the Case Control. See remark 7. Requests the cumulative root mean square function be calculated for random analysis post-processing. Request must be made above the subcase level and RANDOM must be made in the Case Control. See remark 7. Requests all of PSDF, ATOC, and CRMS be calculated for random analysis post-processing. The request must be made above the subcase level and RANDOM must be selected in the Case Control. See remark 7. Writes random analysis results to the print le. (Default) See remark 7. Disables the writing of random analysis results to the print le. See remark 7. Writes random analysis results to the punch le. See remark 7. Accelerations at all points will be output. Accelerations at no points will be output. Set identication of a previously appearing SET command. Only accelerations of points with identication numbers that appear on this SET command will be output. (Integer>0)
CRMS
RALL
RPRINT NORPRINT RPUNCH ALL NONE n
REMARKS:
1. Both PRINT and PUNCH may be requested. 2. Acceleration output is only available for transient and frequency response problems. 3. See Remark 2 under DISPLACEMENT case control command for a discussion of SORT1 and SORT2. 4. ACCELERATION = NONE allows overriding an overall output request. 5. The PLOT option is used when curve plots are desired in the magnitude/phase representation and no printer request is present for magnitude/phase representation. 6. Acceleration results are output in the global coordinate system (see eld CD on the GRID Bulk Data entry). 7. The option of PSDF, ATOC, CRMS, and RALL, or any combination of them can be selected for random analysis. The results can be either printed to the
2-14
Dynamics
.f06 le, punched to the punch le, or output to both les using RPRINT and RPUNCH. 8. When doing enforced motion dynamic analysis and relative output is requested (using the REL describer), the output will be relative to the input as described by the equation:

1. Output is restricted to REAL format. IMAG, PHASE, PSDF, ATOC and RALL are ignored. 2. Displacements, velocities and accelerations must be output for the same set of grid points if requested. Output requested for set n in this command will be combined with the sets requested in the VELOCITY and DISPLACEMENT commands, and accelerations will be output at the grid points of the combined set.
2-15
Chapter 2
Dynamics
2.2

Mode Acceleration Method
Dynamic analysis using modal methods is typically performed in the following sequence: A set of normal modes are calculated and used to dene the dynamic model in modal coordinates. The dynamic response is calculated for each normal mode in modal coordinates. The physical response is recovered from the dynamic response.
The accuracy of the recovered response depends on the number of normal modes computed. A general rule is to compute modes to frequencies twice the range of interest. Still in some cases, the effect of the modes not computed (the truncated modes) can be signicant. This is especially true for element forces, stresses and strains. It may not be practical to solve for more modes and the mode acceleration offers an alternative approach for improving accuracy. Mode acceleration accounts for both the dynamic effects of the computed normal modes plus the static response of the truncated modes. There is an added computational expense for performing the mode acceleration computations but a new mode acceleration formulation with improved performance has been added in this release as an alternative to the older algorithm. The improved performance makes mode acceleration a more attractive option when increased accuracy is desired. The MODACC parameter will continue to be used to turn on this capability, and the default remains as -1 for off. The current requirement that mode acceleration must always use PARAM,DDRMM,-1 still exists, although a change in the software has occurred where the DDRMM will always be assigned to -1 when either the old or new mode acceleration method is used, regardless of the value assigned. The new mode acceleration will be used for both applied loads and enforced motion terms and is supported in SOLs 111, 112, 146, and 200 (modal transient and modal frequency response only). Also, the ability to restart from the same solution sequence as well as SOL 103 is supported. You will be able to choose between the old and new mode acceleration methods with the system cell 422 (ENFMOTN), which is the same system cell used to select the enforced motion method. This is because the old mode acceleration method is based on the absolute displacement method for the enforced motion component, whereas the new mode acceleration method is based on the new constraint mode method for the enforced motion component, which was introduced in NX Nastran 4.1. Setting system cell 422 to 0 (with PARAM, MODACC, 1) will choose the new mode acceleration method; setting it to 1 (default) will choose the old mode acceleration method. However, if modal energy (via the MODALE case control command) or modal contributions (via the MODCON case control command) are requested, the absolute method must be used. MODALE and MODCON are not supported for the constraint mode method. The new MODACC parameter, and the system cell 422 (ENFMOTN) are included below under Input Updates for Mode Acceleration.
Mathematical Details for New Mode Acceleration Method:

The dynamic equation of motion with the constraint mode method is
Equation 2-6.
2-16
Dynamics
Where the total response, uf, is computed as the sum of a static enforced motion component, , and a dynamic relative displacement component, vf , as
Equation 2-7. The other terms in the above equations have been previously dened in section 21. The solution to equation 26 can be obtained efciently in modal coordinates, modal transformation , by using the
Equation 2-8. Here is the mode shape matrix of normal modes. Substituting into equation 26 gives
Equation 2-9. where
Equation 2-10. The matrices m and k are modal diagonal matrices. The modal viscous damping matrix, b, and modal structural damping matrix, , are generally not diagonal. , and substituting back into
Standard modal analysis involves solving equation 29 for equations 28 and 29 to get the physical displacement uf.
In the mode acceleration method, the same approach is used again to compute the modal displacement. But a more accurate physical displacement response is computed by solving for the displacement vf from a direct solution of equation 26 with the inertia and damping terms taken to the right hand side and computed in modal terms. This gives the response computed as
Equation 2-11. Transforming back into total displacement coordinates, uf , and using properties of the modal matrices, the absolute displacements can be computed as
2-17
Chapter 2
Dynamics
Equation 2-12. where is a square modal size matrix with the modal frequencies on the diagonal. The terms on the right hand side of equation 212 are essentially different types of mode shapes. The names for shapes are = real normal modes = static constraint modes = applied force attachment modes = inertia attachment modes = viscous damping attachment modes = structural damping attachment modes
Input Updates for Mode Acceleration

New parameter MODACC: MODACC Default = -1 MODACC = 0 selects the mode acceleration method for data recovery in dynamic analysis. See Formulation of Dynamic Equations in SubDMAP GMA in the NX Nastran Users Guide for further discussion. MODACC = 1 is the same as MODACC = 0 except if SUPORTi entries are present then the displacements are corrected for rigid body motion and the local exibility at the SUPORTi degrees-of-freedom. MODACC 0 is not recommended for use in hydroelastic problems. New system cell 422ENFMOTN (repeated here for convenience): 422 ENFMOTN Controls which formulation is used for enforced motion response analysis (and mode acceleration, if requested). 0: Constraint mode method of enforced motion formulation (and new mode acceleration method) 1: Absolute displacement enforced motion formulation (and old mode acceleration method) (Default). (ABS/REL option on DISPLACEMENT, VELOCITY, and ACCELERATION case control entries is disabled).
2-18
Dynamics
2.3
Random Response Output
In many applications, the loading on a structure is not known in a denite (or deterministic) sense, and is instead quantied in a statistical sense with properties such as the mean load and standard deviation of load. Such loadings are known as random or stochastic loads. Examples of random loads include: rough road surface loads on automobiles, ocean wave loads on offshore platforms, wind loads on buildings, vibration loads from rocket engines, and earthquake ground motions. You should expect the response of the structure due to random loads to also be random in nature, and analysis methods that determine the statistical properties of response are appropriate. The most commonly used approach for random analysis uses Power Spectral Density (PSD) functions to dene loading and the calculated response. A PSD function describes how the variance of a time series of the product of two variables is distributed with frequency. If the two variables are the same, the PSD function is known as an auto spectrum; if the variables are different, the PSD function is called a cross spectrum. NX Nastran performs random analysis in a two-step procedure. In the rst step, a frequency response analysis is performed to calculate transfer functions which are the ratio of output to a unit input. In the second step, the PSD loading function is used to scale the transfer functions to compute the random PSD response. The input PSD functions represent the statistics of the applied loads. If you have several applied loads that are independent of each other, you only need to apply the auto spectra of the loads. If, on the other hand, the loads are not independent of each other, you must also input the cross spectrum of the loads to dene the correlation of the loadings. Note: Although the process is described as two steps, it can be performed with NX Nastran as only one solution submittal. Random response output consists of the response PSD, autocorrelation functions, number of zero crossings with positive slope per unit time, and RMS (root-meansquare) values of response. Prior to NX Nastran version 4.0, only auto spectrum PSD functions of response could be output for random results. You were required to use a DMAP alter if you wanted cross spectrum results. In NX Nastran 4.1, the option to output cross-power spectral density functions became available. Now with NX Nastran 5.0, you can also request cross-correlation functions. The Case Control command, RCROSS and Bulk Data entry, RCROSS were created in version 4.1 for this purpose. The new formats of these entries are included at the end of this section. The following examples have been provided in the install location NX_Nastran_Install_Path/nxn5/nast/tpl to demonstrate capabilities and to show the correct syntax for using the new RCROSS commands: rcross01.dat, rcross02.dat, rcross03.dat, rcross04.dat.
Theory for Cross-power Spectral Density and Cross-correlation Functions

Assuming that a random process is ergodic, the cross correlation function between a pair of time records ua(t) and ub(t) is dened as:
Equation 2-13. where ua(t) and ub(t) can be displacement, velocity, acceleration or single-point constraint force responses at the same or different grid points; or stress, strain, and force components in the same or different elements. The assumption of ergodicty means that as the length of the record is increased the average approaches the corresponding ensemble averages. The operation of integration over a time record of length T and dividing that quantity by T represents an averaging process over time.
2-19
Chapter 2
Dynamics
If we denote the Fourier transform of a time history as:
Equation 2-14. we can dene the cross-power spectral density (cross-PSD) as:
Equation 2-15. where is angular frequency in rad/sec and the symbol (*) denotes complex conjugate. Simply stated, the cross-PSD is the product of the Fourier transform of two records. It can be shown using equations 26 and 27 that the cross correlation function and cross spectrum are related as Fourier transform pairs given as:
Equation 2-16.
Equation 2-17. It is noted that the cross correlation function is always real valued and the cross spectrum function is in general complex valued. In addition, it can be shown to possess the following properties:
Equation 2-18.
Equation 2-19. It should also be noted that we have been referring to the general case where ua(t) and ub(t) are different responses. In many situations, only a single response, rather than the joint response is of interest. In that case, auto-correlation functions, Raa(),and auto-spectrum functions, Saa(), are computed. The auto-spectrum by denition is completely real. Returning to the more general case, we can rewrite in terms of its real and imaginary parts as:
Equation 2-20. From the properties in equation 210 it can be determined that Srab() is an even function of and SIab() is an odd function of .
2-20
Dynamics
Substituting equation 213 into the equation 212 and recalling that eit = cos(t)+isin(wt), it can be shown that:
Equation 2-21. A particular case of interest is the value of the cross-correlation function for =0. As evidenced by equation 26 it represents the mean cross product. Using equation 214 it can be computed as:
Equation 2-22. where () denotes mean average. The mean cross product is an important statistical quantity. For cross-correlation functions, is it known as the covariance. For auto-correlation functions the mean cross-product is known as the variance and its square root is the standard deviation. By using the standard assumption that loadings and response are Gaussian (normally distributed), knowing the covariance allows you to dene the probability distribution and thus determine likelihood of occurrence of response levels. For example, if the standard deviation of response ua(t) is determined to be a, the likelihood that at any time the response is less than 3a is 99.9%. If a system is subject to multiple sources of excitation, the cross-PSD of responses ua(t) and ub(t) is related to the cross-PSDs of the excitation sources and the frequency response functions by the equation:
Equation 2-23. where Sjk() is the cross-PSD of loadings Pj(t) and Pk(t). Haj() is the frequency response function of ua(t) corresponding to the excitation source Pj(t) and similarly Hbk()is the complex conjugate of the frequency response function of ub(t) corresponding to the excitation source Pk(t). If the two excitation sources are not correlated, we have Sjk().= 0.
New RCROSS Case Control Command
RCROSS
Cross-Power Spectral Density and Cross-correlation Function Output Requests
Requests computation and output of cross-power spectral density and cross-correlation functions in random analysis.
FORMAT:
2-21
Chapter 2
Dynamics
EXAMPLES:
RCROSS(PHASE)=10 RCROSS(PSDF,NOPRINT,PUNCH)=20
DESCRIBERS:
Describer REAL or IMAG PHASE
Meaning Requests rectangular format (real and imaginary) of complex output (for cross-power spectral density function). Use of either REAL or IMAG yields the same output. (Default) Requests polar format (magnitude and phase) of complex output (for cross-power spectral density function). Phase output is in degrees. Writes output to print le. (Default) Does not write output to the print le. Writes output to punch le. Requests the cross-power spectral density function be calculated and output for random analysis post-processing. (Default) Requests the cross-correlation function be calculated and output for random analysis post-processing. Requests both the cross-power spectral density and cross-correlation function be calculated and output for random analysis post-processing. Identication number of RCROSS bulk data entry to be used in random analysis. (Integer > 0)
PRINT NOPRINT PUNCH PSDF
CORF RALL
REMARKS:
1. The case control RCROSS entry must be used in conjuction with the case control RANDOM entry. See remarks under the RANDOM case control entry. 2. Response quantities, such as DISPLACEMENT, STRESS, and FORCE, must be requested by corresponding case control entries in order to compute cross-power spectral density and cross-correlation functions between the two response quantities specied by the RCROSS bulk data entry. It is recommended that the DISPLACEMENT, STRESS, and FORCE requests be put above the subcase level to ensure that these response quantities exist when the random analysis post-processing occurs. 3. The response quantities must belong to the same superelement. The cross-power spectral density and cross-correlation functions between two responses that belong to different superelements are not supported.
2-22
Dynamics
New RCROSS Bulk Entry
RCROSS
Cross-Power Spectral Density and Cross-Correlation Function Output
Denes a pair of response quantities for computing the cross-power spectral density and cross-correlation functions in random analysis.
FORMAT: 1 RCROSS 2 SID 3 RTYPE1 4 ID1 5 COMP1 6 RTYPE2 7 ID2 8 COMP2 9 CURID 10
EXAMPLE: RCROSS 20 DISP 50 2 STRESS 150 8 4
FIELDS:
Field SID RTYPEi IDi COMPi CURID

REMARKS:
Contents Case control RCROSS identication number for cross-power spectral density and cross-correlation functions. (Integer>0) Type of response quantity. At least one eld must be selected. See remark 2. (Character or blank) Element, grid, or scalar point identication number. (Integer > 0) Component code (item) identication number. See remark 3. (Integer > 0) Curve identication number. See remark 4. (Integer > 0 or blank)
1. This entry is required for computing the cross-power spectral density and cross-correlation functions. SID must be selected with the case control command (RCROSS=SID). Fields RTYPE1, ID1, and COMP1 represent the rst response quantity; elds RTYPE2, ID2, and COMP2 represent the second response quantity. 2. The keywords for eld RTYPEi are listed as follows: Keyword DISP VELO ACCEL OLOAD SPCF Meaning Displacement Vector Velocity Vector Acceleration Vector Applied Load Vector Single-point Constraint Force Vector
2-23
Chapter 2
Dynamics
MPCF STRESS STRAIN FORCE
Multi-point Constraint Force Vector Element Stress Element Strain Element Force
If either RTYPE1 or RTYPE2 is blank, then the blank eld takes the default from the dened eld. 3. For elements, the item code COMPi represents a component of the element stress, strain or force and is described in Tables Element Stress-Strain Item Codes and Element Force Item Codes. For an item having both a real and imaginary part, the code of the real part must be selected. This is required for computing both the cross-power spectral density function and cross-correlation function. For grid points, the item code is one of 1,2,3,4,5, and 6, which represent the mnemonics T1, T2, T3, R1, R2, and R3, respectively. For scalar points, always use 1. 4. Field CURID is optional. It is for your convenience to identify the output by using a single index.
2-24
Dynamics
2.4
DDAM Process Automation
The Dynamic Design Analysis Method (DDAM) is a list of procedures to determine the modal shock response of on-board ship equipment due to underwater explosions. To complete DDAM analysis procedures in previous releases of NX Nastran, you were required to use dmap alter les. In NX Nastran 5, these DDAM procedures have been streamlined and automated, thus eliminating the alter requirement. The NX Nastran 5 DDAM analysis has been implemented as a new solution sequence, SOL 187. This solution sequence has 3 separate phases which can all occur automatically from a single NX Nastran job submittal, or can optionally be completed interactively: Phase 1: A modal analysis (SOL 187) runs to calculate the natural frequencies. Then the participation factors and modal effective weights are calculated for each mode. The modes, participation factors, and modal effective weights are written to an ASCII OUTPUT4 le. Phase 2: The Naval Shock Analysis (NAVSHOCK) FORTRAN program is automatically invoked to compute the modal shock responses. NAVSHOCK uses the following les as input: The OUTPUT4 le created in phase 1 by NX Nastran. A required, user-created DDAM Control le storing various runtime options. An optional, user-created DDAM Coefcient le containing the weighting factors used for the response calculations, the directional scaling factors, as well as the modal mass cutoff value. This le must be listed in the DDAM Control le for NAVSHOCK to use it. An optional, user-created Shock Spectra le which denes the input shock spectrum as data pairs of frequency and displacement, velocity, acceleration. This le must also be listed in the DDAM Control le for NAVSHOCK to use it. Phase 3: The modal shock responses created in phase 2 are read by NX Nastran, and results are recovered and output for post-processing.
DDAM Input Requirements

The following are the input requirements needed to run a DDAM analysis. When you run DDAM from a single NX Nastran job submittal, the NX Nastran input le creates a wrapper around the entire DDAM analysis process. The details of this NX Nastran input le, plus all les used by the NAVSHOK program are discussed below. NX Nastran Input File Requirements File Management and Executive Control You must select solution sequence SOL 187 in the executive control section. Note: The analysis inputs and options for SOL 187 are the same as for SOL 103. The DDAM analysis requires 3 ASSIGN statements in the NX Nastran input le. These statements link physical les required by the FORTRAN NAVSHOK program to the NX Nastran OUTPUT4 and INPUT4 les. DDAM analysis in NX Nastran uses the units 11, 13 and 22 as described below.
2-25
Chapter 2
Dynamics
Unit 11 is assigned to an OUTPUT4 le and is where NX Nastran writes the matrices which are used as input to the FORTRAN program. The following ASSIGN statement assigns the le ddam.f11 to be used as input to the FORTRAN program. The DELETE qualier tells NX Nastran to delete any existing version of the le and replace it with the one generated in the current submittal.
ASSIGN OUTPUT4=ddam.f11, UNIT=11, FORM=FORMATTED, DELETE
Unit 13 is assigned to an INPUT4 le that stores the output from the FORTRAN program. Once the FORTRAN program has completed, NX Nastran reads the results from this le. The following ASSIGN statement assigns the le ddam.f13 to store the output of the FORTRAN program. The new DEFER keyword tells NX Nastran not to check for the existence of the le at the start of the NX Nastran job. This keyword is needed because the FORTRAN output le is not created until the FORTRAN program runs later in the job.
ASSIGN INPUTT4=ddam.f13, UNIT=13, FORM=FORMATTED, DEFER
Unit 22 is assigned to an INPUT4 le that stores the input control options for the FORTRAN program. The FORTRAN program reads from this le the keyboard responses which would normally be provided interactively. The following ASSIGN statement assigns unit 22 to the ddam.inp le:
ASSIGN INPUTT4=ddam1.inp, UNIT=21, FORM=FORMATTED
Case Control The Case Control section for SOL 187 is similar to SOL 103, but there are a few exceptions: To obtain the needed modal mass for nal response post-processing, it is important that you dene a large enough frequency range for the natural frequency solution. Superelement reduction procedures are supported, although NRL-summed results recovery only occurs for the residual. The number of supported case control output requests is limited as compared to SOL 103. The available output requests for SOL 187 are STRESS, FORCE, DISPLACEMENT, VELOCITY and ACCELERATION. Mode shapes are written by default.
Parameter DPREONLY The new DPREONLY parameter is used to signal the SOL 187 analysis that it is to stop execution immediately before the FORTRAN NAVSHOK program is run. This option is used when you want to run the FORTRAN program outside the NX Nastran job. Often for security reasons, you may be required to run the FORTRAN program on a separate, secure computer, or you may be required to manually enter the shock coefcients from the keyboard. Setting this parameter to YES will stop the analysis after the OUTPUT4 les are written, but just before the FORTRAN program is executed. Once the FORTRAN program has been run manually, you use a NX Nastran restart solution to continue the analysis. Note, that the restart input le must use the /, directive to remove the DPREONLY parameter from the restart database. This is done by looking at the sorted Bulk Data from the original run and nding the sequence number (n) of the DPREONLY parameter statement, then including a "/,n" directive after the BEGIN BULK in the restart input le. The sorted Bulk Data is output by default. If the ECHO Case Control command is specied in the original run, it should be set to output the sorted Bulk Data (for example, ECHO=SORT). See the Restarts chapter in the NX Nastran Users Guide for more details.
2-26
Dynamics
Bulk Data In addition to all the information needed for a typical SOL 103 analysis, the bulk data section requires the following: A SUPORT entry is required to dene the single xed base location for the 3 translational degrees-of-freedom. Rigid elements or MPCs can be used to connect this single location across a distributed region of the model. If superelements are dened, the xed base point must be part of the residual, and in the Aset. The directions of the SUPORT location should reference a rectangular coordinate system. Although the shock response is only computed in translational DOF, it is still important to SPC any of the 6 DOF directions at the xed point which are not included on the SUPORT entry. If a massless DOF is included on the SUPORT entry, the following error occurs: MR MATRIX has NULL column. The eigenvectors must be normalized to MASS. Failure to do so will result in incorrect modal masses and participation factors. The NX Nastran model must be dened in the units of length=inches and mass=lb-sec2/in since this is a requirement of the NAVSHOK program.
Required DDAM Control File The NAVSHOK program requires the ASCII control le which supplies runtime options. As described above in the le management descriptions, this is the le assigned to unit 21. The format of this control le will be one of three, depending on which options you request. Default user options: FF nsurf nstruc nplast pref f/a_axis vert_axis f11 lename f13 lename .ver lename User coefcient option: TF coef.dat lename nsurf nstruc nplast pref f/a_axis vert_axis .f11 lename .f13 lename .ver lename User spectrum Option: FT spec.dat lename
2-27
Chapter 2
Dynamics
pref f/a_axis vert_axis .f11 lename .f13 lename .ver lename Specic le formats: First Line - spectrum control - format a1,1x,a1 First Item - Coefcients from File or form compiled source T=coefcients from external le F=use built-in coefcients Second item - DDAM or general spectrum run ag T=General non-DDAM spectrum run F=DDAM Second Line - le name (if needed) -format a80 If 1st item on line 1 is T, Name of coefcient le If 2nd item on line 1 is T, Name spectrum le If neither are T, line is not needed Third Line - location ags - format i1,1x,i1,1x,i1 First Item - Surface or Submarine 1=Surface 2=Submarine Second Item - equipment location 1=Deck 2=Hull 3=Shell Third Item - coefcient type 1=Elastic 2=Elastic/Plastic 4th Line - Weight cutoff percentage - format F8.3 (0. To 100.) 5th Line - Axis Orientation - format a1,1x,a1 First Item - F/A axis,X,Y, or Z Second Item - Vertical Axis X,Y, or Z 6th Line - Input le name -format a80 7th Line - Output le name - format a80 8th Line - Verication le name -format a80
2-28
Dynamics
Note: The capitalization and spacing of the rst line and the axis denition line are important. Below is an example to demonstrate proper formatting: FF 111 80. XZ navs1.f11 navs2.f11 navs1.ver Optional DDAM Coefcient File The DDAM coefcient le, which is formatted similar to an NX Nastran input le, contains the weighting factors used for the response calculations, the directional scaling factors, and the modal mass cutoff value. The equations which describe these factors are:
(M is the modal weight in kips.) For surface ship, hull and shell mount, the equation is:
A complete set of AA, AB, AC, and (when needed) AD weighting factors for each of the possible analysis congurations (surface ship and submerged ship, deck mount, shell mount and hull mount) must be included. In addition, there is a set of three AF (acceleration factors) and VF (velocity factors) for each of these, one for each shock direction. Additional factors are also required for Elastic-Plastic conditions. The coefcients and factors are dened on a COEF entry: COEF nsurf VF(1) VA nstruc VF(2) VB nplast VF(3) VC AF(1) AA AF(2) AB AF(3) AC
AD
nsurf nstruc nplast
Ship type. Allowable values are SUB (submerged) and SURF (surfaceship) Mounting location. Allowable values are DECK, HULL, and SHELL Elastic or elastic-plastic factors. Allowable values are ELASTIC and ELPL
The (1), (2), and (3) after VF and AF refer to the directions: (1)=fore/aft, (2)=athwartship, and (3)=vertical. To use the default in the NAVSHOK program, enter a blank or an asterisk *. In addition to the COEF entry, there is a CUTOFF entry which denes the modal mass cutoff percentage: 1 2 pref CUTOFF 3 4 5 6 7 8 9 10
2-29
Chapter 2
Dynamics
pref is the cutoff weight percentage for the modal mass calculation. Enter as a percentage, not a decimal fraction (that is 85. instead of .85). # Example factors # - deck coefcients COEF SURF DECK .30 .40 10. 20. # hull coefcients HULL COEF SURF .4 .5 4. 8.
ELPL 1.0 45.
.27 13.
.60 31.6
1.0 6.
ELPL 1.0 38.
.25 38.
.30 53.4
1.0 6.5
15.
# shell coefcients COEF SURF SHELL ELPL .25 .50 1.0 * * *
.25 10.
.50 44.6
1.0 7.2
17
All elds must be 8 characters long, and can be achieved by adding blanks to the ends of lines. Optionally dened Shock Spectra The Shock Spectra le allows you to dene the spectrum as data pairs of frequency versus displacement, velocity or acceleration. To span several orders of magnitude, you can use either the logarithmic or linear scale for the data. The format of the ASCII le is as follows:
# = comments, anywhere in file DATATYP type data1 dir freq interp BEGIN DATA f1, data1 f2, data2 fn, datan [BEGIN DATA] [f1, data1] [f2, data2] [fn, datan] [BEGIN DATA] [f1, data1] [f2, data2] [fn, datan] END FILE
The data on the DATATYP entry are as follows: type data One of the following: DISP, VELO, or ACCE Units of displacement, velocity or acceleration data as described: G (acceleration data in Gs), F (Displacement, Velocity or Acceleration data in feet, ft/sec, or ft/sec2), I (displacement, velocity or acceleration in inches, in/sec or in/sec2), or M (displacement, velocity or acceleration in meters, m/sec or m/sec2)
2-30
Dynamics
dir freq interp
Number of spectra in the le. A value of 1 indicates a single spectrum will be used for all three shock directions. A value of 3 indicates there are three spectra, one for each direction. Units of frequency terms. RAD (radians) or HERTZ (frequency in hertz). Flags the interpolation scheme for the table values. Options are: LOGLOG (both axes are logarithmic), LINLIN (neither axis is logarithmic), LOGLIN (the frequency axis is logarithmic, the other is not), or LINLOG (the frequency range is linear, the other is logarithmic).
A BEGIN DATA entry should precede each frequency/motion data section. If dir=1, there will be one BEGIN DATA entry, if dir=3, there will be three. The data section is completed with the END FILE entry. Example:
# sample spectrum file DATATYP ACCE I 1 HERTZ LOGLIN # acceleration vs frequency file - acceleration in in/sec**2 BEGIN DATA 1. 1. 10. 10. 100. 1000. # point added to define range 500. 800. 1000. 500. END FILE
Running ddam.exe interactively

You may choose to end the NX Nastran job after the natural frequency solution has completed, then run the NAVSHOK program interactively. When running interactively, NAVSHOK will prompt you for all needed inputs. The NAVSHOK program is included in the NX Nastran installation at /nxn5/arch/ddam.exe.
Selection of Specic Modes

DMI (Direct Matrix Input) is a way to choose which modes go into the NRL summation. You can dene a partitioning vector (PARTNVEC) on the DMI entries which will break the eigenvector matrix and UHV matrix into specic pieces. PARTNVEC is a multi-row, 3-column matrix where the rows are mode numbers, and the columns are shock direction. If the matrix has a 1, that mode is retained for a shock direction. The example below demonstrates a PARTNVEC denition: 1 2 3 4 5 6 $ dene the matrix as a 12 row by 3 column matrix 2 1 DMI PARTNVEC 0 $ dene col 1 (keep all modes 1 -12) (f/a shock) 1 1. DMI THRU PARTNVEC 1 $ dene col 2 (keep modes 1, 2 and 3) (athw shock) 2 1. 2 DMI PARTNVEC 2 $ dene col 3 (keep only modes 5) (vert. shock) 5 1. 1 DMI PARTNVEC 3 7 12 12 1. 1 3 1 1. 1 8 3 9
2-31
Chapter 2
Dynamics
You will need to request all the modal mass (100.) if you use the PARTNVEC option, since it will perform its own partitioning on the output from the DDAM program.
OUTPUT4 Data Written in Phase 1

Phase I of the SOL 187 is a modal analysis to calculate the natural frequencies of the system. Then the participation factors and modal effective weights are calculated for each mode and are written to Output 4 les in ASCII format. Phase I computes and outputs the following: An Output4 le containing the following matrices: Natural frequencies OMEX Participation factors PAB MTOT Total mass in each direction MFRACT Percent of total mass from the modes used The program also prints the following for user verication: The 4 matrices written to the above Output4 le Total rigid body mass matrix MTOTC MEFC Total modal effective mass matrix Diagonal Terms of MEFFC MEFF Rigid body vector set referenced to R set PHBASE
2-32
Dynamics
2.5
Elemental Energy Output
The element strain (ESE), kinetic (EKE), and damping (EDE) case control commands have been improved to allow the output of element energy information in SORT2 complex format. This format is suitable for plotting function data from frequency domain dynamic solutions. In previous releases, you could only output the real magnitude of element energy, thus no phase information was written. In addition, only SORT1 format could be output, which is not convenient for function plotting. To control the output format, new describer sets [RMAG, REAL or IMAG, PHASE] and [SORT1, SORT2] have been added to the ESE, EKE, and EDE case control cards. These new sets are consistent with the other NX Nastran case control commands, but the following exception: The RMAG option allows you to output the energy magnitude as a real number only, which is needed to retain the same data output as used in previous versions.
The updated case control formats for EDE, EKE, and ESE are included at the end of this section. The table below summarizes the output result with different describer selections: RMAG AVERAGE Eavg REAL or IMAG (Eavg , 0.) PHASE (Eavg , 0.)
AMPLITUDE
Eamp_mag
(Eamp_real , Eamp_imag ) (Eamp_mag , amp )
PEAK
Epeak
AVGAMP
(Eavg , 0.) (Eamp_real , Eamp_imag )
(Eavg , 0.) (Eamp_mag , amp )
Table of Energy Output Results
Example Element Energy Output Formats

Example 1: ESE(PUNCH, AVGAMP, REAL, SORT2 )=ALL PUNCH Output
$TITLE = MODAL FREQUENCY RESPONSE $SUBTITLE= $LABEL = $AVERAGE ELEMENT STRAIN ENERGY 1 2 3 4
2-33
Chapter 2
Dynamics
$REAL-IMAGINARY OUTPUT $SUBCASE ID = 1 $ELEMENT NUMBER = 201 1.000000E-01 4.918013E+00 1.100000E-01 5.975217E+00 1.200000E-01 7.143042E+00 1.300000E-01 8.424302E+00 1.400000E-01 9.822133E+00 1.500000E-01 1.133997E+01 . . . $TITLE = MODAL FREQUENCY RESPONSE $SUBTITLE= $LABEL = $AMPLITUDE ELEMENT STRAIN ENERGY $REAL-IMAGINARY OUTPUT $SUBCASE ID = 1 $ELEMENT NUMBER = 201 1.000000E-01 4.914062E+00 1.100000E-01 5.969384E+00 1.200000E-01 7.134707E+00 1.300000E-01 8.412709E+00 1.400000E-01 9.806374E+00 1.500000E-01 1.131896E+01 1.600000E-01 1.295405E+01
0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00
5 6 7 8 9 10 11 12 13
1.970973E-01 2.639417E-01 3.449702E-01 4.418095E-01 5.561835E-01 6.899223E-01 8.449761E-01
7993 7994 7995 7996 7997 7998 7999 8000 8001 8002 8003 8004 8005 8006
Example 2: ESE(PUNCH, AVGAMP, REAL, SORT1 )=ALL PUNCH Output

$TITLE = MODAL FREQUENCY RESPONSE $SUBTITLE= $LABEL = $AVERAGE ELEMENTAL STRAIN ENERGY $REAL-IMAGINARY OUTPUT $SUBCASE ID = 1 $FREQUENCY = 0.1000000E+00 201 4.918013E+00 202 1.939230E-05 203 2.092620E-03 204 4.920125E+00 $TITLE = MODAL FREQUENCY RESPONSE . . . $TITLE = MODAL FREQUENCY RESPONSE $SUBTITLE= $LABEL = $AMPLITUDE ELEMENTAL STRAIN ENERGY $REAL-IMAGINARY OUTPUT $SUBCASE ID = 1 $FREQUENCY = 0.1000000E+00 201 7.134707E+00 202 2.792866E-05 203 3.013540E-03 204 7.137749E+00 $TITLE = MODAL FREQUENCY RESPONSE 1 2 3 4 5 6 7 8 9 10 11 12
0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00
3.449702E-01 2.065627E-08 9.193350E-07 3.449712E-01
21924 21925 21926 21927 21928 21929 21930 21931 21932 21933 21934 21935
2-34
Dynamics
Updated Case Control Commands EDE, EKE, and ESE
EDE
Element Energy Loss Per Cycle Output Request
Requests the output of the energy loss per cycle in selected elements.
FORMAT:
EXAMPLES:
EDE=ALL EDE(PUNCH, THRESH=.0001)=19

DESCRIBERS:
Describer PRINT PUNCH PLOT AVERAGE AMPLITUDE PEAK AVGAMP RMAG REAL or IMAG PHASE
Meaning Write energies to the print le (default). Write energies to the punch le. Do not write energies to either the punch le or the print le. Requests average energy in frequency response analysis only. Requests amplitude of energy in frequency response analysis only. Requests peak energy for frequency response analysis only. PEAK is the sum of AVERAGE and AMPLITUDE. Requests both the average and amplitude energy in frequency response analysis only. Outputs the energy magnitude in real data format. Requests rectangular format (real and imaginary) of complex output. Use of either REAL or IMAG yields the same output. Requests polar format (magnitude and phase) of complex output. Phase output is in degrees.
2-35
Chapter 2
Dynamics
Describer SORT1
Meaning Output will be presented as a tabular listing of grid points for each load, frequency, eigenvalue, or time, depending on the solution sequence. Output will be presented as a tabular listing of frequency or time for each grid point. Suppresses energies for elements having an energy value of less than p% in all output les. THRESH overrides the value of TINY described in Remark 1. (Default=0.001) Computes energy for all elements. Set identication number. Energy for all elements specied on the SET n command will be computed. The SET n command must be specied in the same subcase as the EDE command or above all subcases. (Integer>0) Element energy will not be output.
SORT2 THRESH
ALL n
NONE
REMARKS:
1. If THRESH = p is not specied, then p defaults to the values specied by user parameter TINY. 2. The energy calculations include the contribution of initial thermal strain. 3. Energy density (element energy divided by element volume) is also computed in some solution sequences. It can be suppressed by use of PARAM,EST,-1. 4. The equations used to calculate elemental damping energy components are given below. Average Damping Energy:
where: E {ur} {ui} [Be] = = = = elemental energy displacement - real part displacement - imaginary part elemental damping
Real Part of Damping Energy Amplitude:
Imaginary Part of Damping Energy Amplitude:
2-36
Dynamics
Magnitude of Damping Energy Amplitude:
Phase of Damping Energy Amplitude:
Peak Damping Energy:
5. In SOL 111, EDE can only be requested if PARAM,DDRMM,-1 is used. 6. Only damping from the viscous dampers (e.g., CVISC, CDAMPi, etc.) are included. Structural damping is not included in the calculation.
EKE
Element Kinetic Energy Output Request
Requests the output of the kinetic energy in selected elements.
FORMAT:
EXAMPLES:
EKE=ALL EKE(PUNCH, THRESH=.0001)=19

DESCRIBERS:
Describer PRINT PUNCH PLOT AVERAGE
Meaning Write energies to the print le (default). Write energies to the punch le. Do not write energies to either the punch le or the print le. Requests average energy in frequency response analysis only.
2-37
Chapter 2
Dynamics
Describer AMPLITUDE PEAK AVGAMP RMAG REAL or IMAG PHASE SORT1
Meaning Requests amplitude of energy in frequency response analysis only. Requests peak energy for frequency response analysis only. PEAK is the sum of AVERAGE and AMPLITUDE. Requests both the average and amplitude energy in frequency response analysis only. Outputs the energy magnitude in real data format. Requests rectangular format (real and imaginary) of complex output. Use of either REAL or IMAG yields the same output. Requests polar format (magnitude and phase) of complex output. Phase output is in degrees. Output will be presented as a tabular listing of grid points for each load, frequency, eigenvalue, or time, depending on the solution sequence. Output will be presented as a tabular listing of frequency or time for each grid point. Suppresses energies for elements having an energy value of less than p% in all output les. THRESH overrides the value of TINY described in Remark 1. (Default=0.001) Computes energy for all elements. Set identication number. Energy for all elements specied on the SET n command will be computed. The SET n command must be specied in the same subcase as the EKE command or above all subcases. (Integer>0) Element energy will not be output.
SORT2 THRESH
ALL n
NONE
REMARKS:
1. If THRESH = p is not specied, then p defaults to the values specied by user parameter TINY. 2. The energy calculations include the contribution of initial thermal strain. 3. Energy density (element energy divided by element volume) is also computed in some solution sequences. It can be suppressed by use of PARAM,EST,-1. 4. The equations used to calculate elemental kinetic energy components are given below. Average Kinetic Energy:
2-38
Dynamics
where: E {vr} {vi} [Me] = = = = elemental energy velocity - real part velocity - imaginary part elemental mass
Real Part of Kinetic Energy Amplitude:
Imaginary Part of Kinetic Energy Amplitude:
Magnitude of Kinetic Energy Amplitude:
Phase of Kinetic Energy Amplitude:
Peak Kinetic Energy:
5. In SOL 111, EKE can only be requested if PARAM,DDRMM,-1 is used.
ESE
Element Strain Energy Output Request
Requests the output of the strain energy in selected elements.
2-39
Chapter 2
Dynamics
FORMAT:
EXAMPLES:
ESE=ALL ESE (PUNCH, THRESH=.0001)=19

DESCRIBERS:
Describer PRINT PUNCH PLOT AVERAGE AMPLITUDE PEAK AVGAMP RMAG REAL or IMAG PHASE SORT1
Meaning Write energies to the print le (default). Write energies to the punch le. Do not write energies to either the punch le or the print le. Requests average energy in frequency response analysis only. Requests amplitude of energy in frequency response analysis only. Requests peak energy for frequency response analysis only. PEAK is the sum of AVERAGE and AMPLITUDE. Requests both the average and amplitude energy in frequency response analysis only. Outputs the energy magnitude in real data format. Requests rectangular format (real and imaginary) of complex output. Use of either REAL or IMAG yields the same output. Requests polar format (magnitude and phase) of complex output. Phase output is in degrees. Output will be presented as a tabular listing of grid points for each load, frequency, eigenvalue, or time, depending on the solution sequence. Output will be presented as a tabular listing of frequency or time for each grid point. Suppresses energies for elements having an energy value of less than p% in all output les. THRESH overrides the value of TINY described in Remark 1. (Default=0.001)
SORT2 THRESH
2-40
Dynamics
Describer ALL n
Meaning Computes energy for all elements. Set identication number. Energy for all elements specied on the SET n command will be computed. The SET n command must be specied in the same subcase as the ESE command or above all subcases. (Integer>0) Element strain energy will not be output.
NONE
REMARKS:
1. If THRESH = p is not specied then p defaults to the values specied by user parameter TINY. 2. The energy calculations include the contribution of initial thermal strain. 3. Energy density (element strain energy divided by element volume) is also computed in some solution sequences. It can be suppressed by use of PARAM,EST,-1. 4. The equations used to calculate elemental strain energy components are given below. Average Strain Energy:
where: E = elemental energy displacement - real part {ur} {ui} [Ke] =
= =
displacement - imaginary part elemental stiffness
Real Part of Strain Energy Amplitude:
Imaginary Part of Strain Energy Amplitude:
Magnitude of Strain Energy Amplitude:
2-41
Chapter 2
Dynamics
Phase of Strain Energy Amplitude:
Peak Strain Energy:
5. In SOL 111, ESE can only be requested if PARAM,DDRMM,-1 is used. 6. Element data recovery for thermal loads is not currently implemented in dynamics. 7. Element strain energy is available for nonlinear static analysis (SOL 106). All other nonlinear solution sequences do not support element strain energy output. 8. The strain energy for nonlinear elements is calculated by integrating the specic energy rate, which is the inner product of strain rate and stress, over element volume and time.
Equation 2-24. where: = = V = t = stress tensor tensor of the strain rate element volume actual time in the load history
Loads from temperature changes are included in Equation 2-24. If we assume a linear variation of temperatures from subcase to subcase, then the strain energy in Equation 2-24for the special case of linear material and geometrybecomes
Equation 2-25. where Pet is the element load vector for temperature differences. For linear elements, the default denition of element strain energy is
2-42
Dynamics
Equation 2-26. where Pet is the element load vector for temperature differences and element deformation. In Equation 2-26, the temperatures are assumed to be constant within a subcase. The default denition of the strain energy for linear elements differs from the denition for nonlinear elements by a factor of 1/2 in the temperature loads. To request the strain energy for linear elements using Equation 2-25, set the parameter XFLAG to 2; the default value for XFLAG is 0, which uses Equation 2-26 for the strain energy of linear elements.
2-43
Chapter 2
Dynamics
2.6
Modal Frequency Response in Rotor Dynamics
Rotor dynamics can now be used in a modal frequency response solution, SOL 111, to calculate the dynamic response of a rotating system. This new response calculation adds to the existing rotor dynamics capability, direct complex mode solution, SOL 110, whose output is used to create Campbell diagrams. The dynamic response calculations can be done using an asynchronous solution in which the rotor speed is dened, but is independent of the applied excitation, or using a synchronous solution in which the external excitation is synchronous with the rotor speed. A new option on the ROTORD bulk entry, SYNC, has been created to select the calculation method. Asynchronous solution inputs: The asynchronous solution runs when SYNC is set to 0. The rotor speeds are dened by the RSTART entry for each rotor denition. The RSTEP and NUMSTEP entries in this case are ignored. The frequency denition comes from the standard entries for frequency response analysis such as FREQ, DLOAD, RLOAD, which also dene the dynamic loading scenario. Synchronous solution inputs: The synchronous analysis runs when SYNC is set to 1. The rotor speeds are dened by RSTART, RSTEP, and NUMSTEP on the ROTORD card. The frequency corresponding to these rotation speeds is computed from the rotation speed, which also denes the dynamic loading. See the Rotor Dynamics Input Updates section below for the version 5 updates.
Mathematical Details
The general equation of motion in the frequency domain for a rotating system with applied loads is:
In the above, is the frequency of applied loading, is the rotor rotation speed, u() is the displacement, and F() is the applied load. The matrices m, b, k are the modal mass, damping and stiffness. The c is the modal Coriolis matrix and the kb is the internal structural damping. There are two forms of the general equation which depend on whether the applied loading is speed dependent or not (synchronous or asynchronous). If the loading is speed dependent (synchronous), then = , and the general equation has the form:
The loading in the synchronous case comes from any mass unbalance on the rotating rotors (not from an external excitation). When the loading is not speed dependent (asynchronous), then , and the general equation has the form:
The loading in the asynchronous case comes from an external excitation other than the rotation of the rotor. Each rotor in the asynchronous case has a dened speed , which remains constant for the response calculation.
2-44
Dynamics
Rotor Dynamics Input Updates

Case Control The following new case control card, MODSEL=n, has been created to select or deselect the mode numbers in dynamic response and aero-elasticity solutions:
MODSEL
Selects/deselects mode numbers
Used to select or deselect mode numbers in dynamic response and aero-elasticity solutions.
FORMAT:
MODSEL = n
EXAMPLES:
MODSEL=3
DESCRIBERS:
Describer n
Meaning Points to a SET card that holds the list of selected and retained mode numbers. The default is to use all modes. The mode numbers omitted from the list are removed from the modal space. Numbers larger than the number of eigenvalues computed are ignored. (Integer)
Bulk Entries The ROTORD entry has been updated with the new SYNC option, in addition, the RSETi option will now reference the RSETID on the new ROTORG entry. The ROTORG is now the required method of selecting the grids which compose a rotor. Both the updated ROTORD and the new ROTORG are shown below:
ROTORD
Denes rotor dynamics solution options (SOLs 110 and 111).
Denes rotor dynamics solution options for SOLs 110 and 111.
FORMAT: 1 ROTORD + + + .... + RID10 RSET10 RSPEED10 RCORD10 W3_10 W4_10 RFORCE10 2 SID ZSTEIN RID1 RID2 3 RSTART ORBEPS RSET1 RSET2 4 RSTEP ROTPRT RSPEED1 RSPEED2 5 NUMSTEP SYNC RCORD1 RCORD2 6 REFSYS ETYPE W3_1 W3_2 7 CMOUT EORDER W4_1 W4_2 RFORCE1 RFORCE2 8 RUNIT 9 FUNIT 10 +
2-45
Chapter 2
Dynamics
EXAMPLE: ROTORD +r0 +r1 +r2 +r3 +r4 +r5 +r6 +r7 +r8 +r9 +r10 998 no 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 1 1 1 1 1 1 1 1 1 1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 10 1 0.0 5000.0 58 x -1.0 cps +r0 +r1 +r2 +r3 +r4 +r5 +r6 +r7 +r8 +r9 +r10
FIELDS:
Field SID RSTART RSTEP NUMSTEP REFSYS
Contents Set identier for all rotors. Must be selected in the case control deck by RMETHOD=SID. Starting value of rotor speed. (Real 0.0) Step-size of rotor speed. (Real > 0.0) Number of steps for rotor speed including RSTART.(Integer > 0) Reference system. (default=ROT) = FIX analysis is performed in the xed reference system. = ROT analysis is performed in the rotational reference system.
CMOUT
Rotor speed for complex mode output request. (Integer 0) = -1.0 output complex modes and whirl direction for all RPM. = 0.0 no complex modes and no whirl direction are output. (default) > 0.0 rotor speed value for which complex modes will be calculated and written to F06 or OP2.
RUNIT
Units used for rotor speed input (RSTART and RSTEP) and output (units for output list and Campbells diagram output). = RPM revolutions per minute. (default) = CPS cycles per second. = HZ cycles per second. = RAD radians per second.
2-46
Dynamics
Field FUNIT
Contents Units used for frequency output. = RPM revolutions per minute. (default) = CPS cycles per second. = HZ cycles per second. = RAD radians per second.
ZSTEIN ORBEPS ROTPRT
Option to incorporate Steiners inertia terms. (default=NO) Threshold value for detection of whirl direction. (default=1E-6) Controls .f06 output options. = 0 no print. (default) = 1 print generalized matrices. = 2 print eigenvalue summary and eigenvectors at each RPM. = 3 combination of 1 & 2.
SYNC
Species if a rotor dynamics response calculation (SOL 111) is synchronous or asynchronous. (Integer) =0 synchronous (default) =1 asynchronous
ETYPE
Excitation type. =1 Mass unbalanced (default). Specify mass unbalance = m x r on DLOAD bulk entry and the program will multiply by 2. =0 Force excitation. Specify force = m x r x 2 on DLOAD bulk entry.
EORDER
Excitation order = 1.0 (default) = 0.0 Forward whirl (modes crossing with 0P line) = 2.0 Backward whirl (modes crossing with 2P line) This value is always 1.0 for xed reference system.
RIDi RSETi
Rotor ID for output identication (default=i); must be unique with respect to all other RIDi values. Refers to a RSETID value on a ROTORG bulk entry. Each ROTORG bulk entry species the grids for a single rotor. (Integer)
2-47
Chapter 2
Dynamics
Field RSPEEDi (place holder) RCORDi W3_i W4_i RFORCEi
Contents Once supported, RSPEED will adjust the rotor speed for rotor i in a rotor dynamics response calculation. RSPEED is currently hard coded to equal 1. Points to CORD** entry specifying rotation axis or rotor i as z. (default=0 for global z axis) Reference frequency for structural damping dened by PARAM G for rotor i. (default=0.0) Reference frequency for structural damping dened on material card for rotor i. (default=0.0) Points to RFORCE bulk data entry for rotor i. (Default=0 for no rotational force applied; a rotational force is required for differential stiffness to be calculated.)
REMARKS:
1. Any entries where defaults exist are optional. Thus, if the defaults are acceptable for a particular model, then a continuation card would not be needed. 2. There is a maximum limit of 10 rotors (i.e. 11 continuation cards). 3. The Steiners term option (ZSTEIN) should only be used for analyzing the stiff part of the rotor. In the rotating system analysis, there is a stiffening effect of the centrifugal forces. The gyroscopic matrix is calculated in the xed system. 4. The W3 parameter denes the reference frequency for the structural damping dened by PARAM G. 5. The W4 parameter denes the reference frequency for the structural damping dened on the material cards. 6. The static centrifugal force is calculated for unit speed measured in rad/sec. On the RFORCE card, the unit of Hz is used, thus the conversion 1/2()=0.159155 must be used. 7. For calculating the modal frequency response using synchronous analysis, the rotation speeds are dened by the RSTART, RSTEP and NUMSTEP elds on the ROTORD bulk entry. The frequencies corresponding to these rotation speeds are computed and the dynamic loads are calculated accordingly. 8. For calculating the modal frequency response using asynchronous analysis, the unique rotation speed is dened by the RSTART eld on the ROTORD bulk entry. The RSTEP and NUMSTEP elds in this case will be ignored. The frequency and dynamic load denitions are taken from the standard FREQ, DLOAD, and RLOAD bulk entries.
2-48
Dynamics
ROTORG
Rotor Grids Selection
Species the grids that compose each rotor.
FORMAT: 1 ROTORG 2 RSETID 3 GRID1 4 GRID2 5 ... 6 GRIDn 7 8 9 10
or
ROTORG RSETID GRID1 THRU GRID2 BY INC
EXAMPLE: ROTORG 14 6 10 15 18
FIELDS:
Field RSETID GRIDi THRU BY INC
Contents Identication number of rotor. References a RSETi on the ROTORD bulk entry. (Integer>0) Grids dening a rotor. (Integer>0) Species a range of grid IDs (Optional) Species an increment when using THRU option. Increment used with THRU option. (Integer>0)
2-49
Chapter 2
Dynamics
2.7
Modal Contribution Computation
When you use the modal method for dynamic response calculations, understanding which modes contribute to the response helps understand the dynamic behavior of the simulated system, and can provide insight as to how to improve the dynamic performance. In this release, a new capability has been introduced to compute the modal contributions in modal frequency response. The new command has the form: MODCON = YES/NO with a default of no, meaning the modal contribution is not output. The option to PRINT, PLOT, or PUNCH the modal contribution data can be selected using the existing DISP, VELO, ACCE and the SDISP, SVELO and SACCELERATION case control methods. There are no new bulk data entries required, however, you can use the USET card with U3 to dene the l=1,....n degrees-of-freedom which modal contributions are computed. The new MODCON case control command is included below:
New MODCON Case Control Command
MODCON
Modal Contribution Request
Requests the output of modal contribution data for SOL 111.
FORMAT:
MODCON=YES/NO
EXAMPLES:
MODCON=YES
DESCRIBERS:
Describer NO YES
REMARKS:
Meaning Do not output modal contribution data. (Default) Modal contribution data will be output.
1. The option to PRINT, PLOT, or PUNCH the modal contribution data can be selected by the DISP, VELO, ACCE and the SDISP, SVELO and SACCELERATION case control methods. 2. Output of modal contribution can only occur on grids. Element output is not supported at this time. 3. No new bulk data entries are required, however, the USET card with U3 can be used to dene the l=1,....n degrees-of-freedom which modal contributions are computed. 4. The complex and actual modal contributions are assigned with dummy subcase IDs for each DOF in the f06 and punch les to distinguish each set of results
2-50
Dynamics
while post processing. It is recommended to use a large increment when assigning the SUBCASE IDs in the case control section to avoid a dummy ID duplicating a case control ID. 5. This capability is NOT supported for the constraint mode method of enforced motion. 6. This capability is not supported for models having rigid body or uid modes.
New Output Datablocks

The new datablocks, MODCONPS and MODALCBN, have been created to output the absolute and actual complex modal contributions. The computed results are printed under the following headers:
***************************************************************************************** THE TABLE(S) STARTING ON THE FOLLOWING PAGE CONTAIN(S) THE COMPLEX MODAL CONTRIBUTIONS ***************************************************************************************** ***************************************************************************************** THE TABLE(S) STARTING ON THE FOLLOWING PAGE CONTAIN(S) THE ACTUAL MODAL CONTRIBUTIONS *****************************************************************************************
Mathematical Description
The general equation for the calculation of the physical response using the modal method is:
In the above equation,
represents some type of physical response from a modal dynamic simulation (displacement, velocity, acceleration, stress, strain .). The subscript l represents a specic component of the physical response (e.g. x-direction of node 22). The variable: , is the modal displacement response of mode number m. There are a total of M modes. The variables:
are elements of the mode shape matrix. Finally,
are the modal contributions of the physical response. It is the terms:
2-51
Chapter 2
Dynamics
that are to be computed for the new modal contribution output. From the above equation, the modal contributions are calculated as:
2-52
Dynamics
2.8
Direct Structural Damping Matrix Input
A structural damping matrix can now be included with a model in NX Nastran 5 using the new K42GG case control command. K42GG is repeated below for convenience. This ability adds to the other direct matrix capabilities of stiffness, mass, and damping matrices using DMIG bulk entries. For more information on direct matrix inputs, see Matrix Assembly Operations in SubDMAP SEMG in the NX Nastran Users Guide.
New K42GG Case Control Command
K42GG
Direct Input Structural Damping Matrix Selection
Selects direct input structural damping matrices.
FORMAT:
K42GG=name
EXAMPLES:
K42GG=K4DMIG K42GG=K1, K2, K3 SET 100=K4DMIG, K1, K8 K42GG=100

DESCRIBERS:
Describer name
Meaning Name of a [k4,2gg] matrix that is input on the DMIG Bulk Data entry. (Character) See Matrix Assembly Operations in SubDMAP SEMG in the NX Nastran Users Guide.
REMARKS:
1. DMIG matrices will not be used unless selected. 2. Terms are added to the structural damping matrix before any constraints are applied. 3. The matrix must be symmetric and eld 4 on the DMIG,name entry must contain the integer 6. 4. A scale factor may be applied to this input via the PARAM, CK42 entry. 5. The matrices are additive if multiple matrices are referenced on the K42GG command.
2-53
Chapter
3-1
Chapter 3
3.1
Sparse Data Recovery for Modal Solutions
A new sparse data recovery option is available for the modal frequency response analysis (SOL 111), modal transient response analysis (SOL 112), and optimization (SOL 200) as the new default. For SOL 111 and 112, this feature reduces the cost of matrix-multiplications inside DDRMM modules when you want to recover large amounts of data in the modal analysis. Similarly, SOL 200 utilizes the partitioning of the eigenvector matrix in order to reduce the cost of matrix-matrix multiplies.
Inputs
The option is on by default, but can be deactivated with system cell 421: NASTRAN SYSTEM(421)=0 or by using the nastran keyword SPARSEDR: NASTRAN SPARSEDR =0.
Performance Data
For a SOL 111 model, the SPARSEDR option reduces the time spent by more than 80%: Version 4 103:09 61.5 128 Version 5 20:50 65.5 127 Units min:sec Mwords Gbytes
DDRMM time Memory HiW Disk HiWater
In a SOL 200 model, the SPARSEDR option reduces the time spent by more than 23 %: With sparsedr:
1:48:50 490:20 2042.1G .0 13008.4 .0 EXITOPT 73 EXIT BEGN
Without sparsedr:
2:51:18 636:22 2673.8G .0 16838.7 .0 EXITOPT 73 EXIT BEGN
3.2
Distributed Memory Parallel Improvements
Due to the increased size and complexity of models today, reducing solution times is becoming more important. The NX Nastran parallel processing methods allow simultaneous use of multiple processors on one or more machines to decrease solution times. The parallel processing method Distributed Memory Parallel (DMP) has been enhanced in NX Nastran 5 with the following additions: The normal mode calculation portion of an optimization solution (SOL 200) can now use DMP A new linear statics (SOL 101) DMP option, Load Domain Static Analysis (LDSTAT), now exists to decrease the solution times when large numbers of load cases exist.
The details of these improvements are as follows. See the NX Nastran Parallel Processing Users Guide for complete details on all parallel processing methods.
3-2
Optimization, SOL 200 with Normal Modes

When an optimization solution (SOL 200) requires normal mode frequencies as part of the optimization problem, the normal mode computation portion of the solution can now take advantage of DMP. When DMP is activated in this situation, SOL 200 will run in a serial environment on all processes, except the normal mode analysis. Input Requirements In a SOL 200 solution, the DMP option can be activated using nastran keywords similar to a SOL 103 DMP run. The DMP methods supported are Frequency Domain Normal Modes Analysis (FDMODES), Geometric Domain Normal Modes Analysis (GDMODES), and Hierarchic Domain Normal Modes Analysis (HDMODES). With GDMODES and HDMODES, the keyword gpart=1 is required. Depending on the desired DMP method, one of the following inputs should be included in the SOL 200 submittal: GDMODES: NASTRAN gpart=1 dmp=p HDMODES: NASTRAN gpart =1 dmp =p nclust=c FDMODES: NASTRAN dmp=p where p is the number of processors and c is the number of clusters.
Linear Statics Improvement

The new computational technique for DMP linear static analysis (SOL 101), and Load Domain Static Analysis (LDSTAT) is now available. LDSTAT will split the analysis tasks such that each processor will solve the problem for a single load case. Solving multiple load cases simultaneously across multiple processors can shorten the solution time dramatically when large numbers of load cases are included in the run. Input Requirements The option is activated with the nastran keyword gpart and dmp: LDSTAT: NASTRAN gpart=1 dmp=p where p is the number of processors. Example A model with 70 load cases was solved on a single processor machine (SERIAL case), then was solved across 4 processors using the new DMP option LDSTAT. The solve time was reduced by more than 74 %: SERIAL 17:01:18 25:13 17:03:06 27:01 LDSTAT ( p=4 ) 53849.0 0.0 59732.0 5883.0 1323.4 1430.9 0.0 107.5 STATRS 96 STATRS 96 SSG3 SSG3 BEGN END
3-3
Chapter 3
13:16:42 17:46 13:17:10 18:14
38790.0 0.0 42896.0 4106.0
717.8 745.9
0.0 28.1
STATRS 102 STATRS 102
SSG3 SSG3
BEGN END
3-4
3.3
Sparse Cholesky Technique for Linear Solutions
This NX Nastran release includes support for a new sparse Cholesky decomposition. All users of Cholesky decomposition (including the Givens and Householder eigensolver methods) will utilize sparse decomposition by default, which can be considerably faster than the non-sparse Cholesky decomposition method used in earlier NX Nastran versions. For example, the decomposition cost of a typical structural stiffness matrix of order 8000 compares as follows: Decomposition type Non-sparse Cholesky Sparse Cholesky Wall time (min:sec) 1:40 0:16
With the sparse method, the DCMP module produces a sparse Cholesky factor, identied by form 14 in its trailer. To use the non-sparse Cholesky method, set system cell 424 to : NASTRAN SYSTEM(424)=0 or use the SPCHOL keyword: NASTRAN SPCHOL=0. Not all FBS options are available with sparse Cholesky; for details, see the Numerical Users Guide.
3-5
Chapter 3
3.4
Eigenvalue Analysis Improvement
NX Nastran Version 5.0 introduces the new REDMULT performance option for use when solving vibration problems with the Lanczos method. This option reduces the cost of matrix-vector multiplies inside the READ module when the mass matrix involved is relatively dense, which can occur when a large number of MFLUID is present in the model. In such cases, matrix-vector multiplies can be a considerable portion of the overall Lanczos time, as indicated by the following summary:
*** USER INFORMATION MESSAGE 5403 (LNNRIGL) BREAKDOWN OF CPU TIME (SEC) DURING LANCZOS ITERATIONS: AVERAGE TOTAL OPERATION REPETITIONS FBS (BLOCK SIZE= 7) 75 10.97 822.60 MATRIX-VECTOR MULTIPLY 173 10.15 1756.69 1699.75 8498.73 SHIFT AND FACTOR 5 LANCZOS RUN 3 9.40 28.21
Using the REDMULT option reduces the time spent in the READ module by more than 15%:
*** USER INFORMATION MESSAGE 5403 (LNNRIGL) BREAKDOWN OF CPU TIME (SEC) DURING LANCZOS ITERATIONS: AVERAGE TOTAL OPERATION REPETITIONS FBS (BLOCK SIZE= 7) 81 12.34 999.39 MATRIX-VECTOR MULTIPLY 187 .00 .69 1687.57 8437.84 SHIFT AND FACTOR 5 LANCZOS RUN 3 10.41 31.24
Inputs
The option is activated by setting system cell 426: NASTRAN SYSTEM(426)=1 or by using the REDMULT keyword: NASTRAN REDMULT=1. The REDMULT option may be used independently of or in conjunction with the REDORTH option, introduced in Version 4.0.
3-6
Chapter
4 Multi-body Dynamic Software Interfaces
4-1
Chapter 4
4.1
RecurDyn Flex Input File Creation
This release includes interoperability between NX Nastran and the FunctionBay RFI (RecurDyn Flex Input) le product. You can now create a RecurDyn Flex Input (RFI) le directly from NX Nastran. The RFI contains the reduced order matrices from the results of a NX Nastran non-restart SOL 103 analysis. The RFI can be imported into RecurDyn and used to represent a exible component in a multi-body dynamics analysis. This direct RFI export capability streamlines the process of creating exible components from FE models, making it possible to obtain more accurate results from multi-body simulations. The component modes determined by NX Nastran during the SOL 103 analysis represent a set of constraint modes for boundary coordinates plus a truncated set of xed-interface normal modes. This combination of mode types is referred to as the Craig-Bampton modes. The RFI le contains data of this representation that FunctionBay can use to perform exible component dynamics analyses. The RFI creation is initiated by the NX Nastran case control command RECURDYNRFI FLEXBODY=YES, along with the inclusion of the Bulk Data entry DTI,UNITS. You can dene the exible body attachment points in NX Nastran by either dening the component as a superelement or part superelement, in which case the physical external (a-set) grids become the attachment points. For a residual-only type model, standard NX Nastran ASET Bulk Data entries are used to dene the attachment points. The new RECURDYNRFI is included below for convenience. The documented descriptions and remarks on this command contain a comprehensive list of use requirements for RFI export. These include the equations representing the nine mass variants, and the DTI,UNITS bulk entry requirement for units control of the RFI.
New Case Control Command
RECURDYNRFI
Generates RecurDyn Flex Input File
Generates a RecurDyn Flex Input (RFI) le during SOL 103.
FORMAT:
4-2
EXAMPLES:
RECURDYNRFI FLEXBODY=YES
DESCRIBERS:
Describer FLEXBODY NO YES FLEXONLY YES NO OUTGSTRS NO YES OUTGSTRN NO YES MINVAR PARTIAL
Meaning Requests the generation of RFI (required). Standard NX Nastran solution without RFI creation (default). RFI generation requested. Determines if standard DMAP solution and data recovery runs or not after RFI creation is complete. Only RFI creation occurs (default). RFI le creation occurs along with standard DMAP solution and data recovery. Determines if grid point stress is written to RFI. Do not write grid point stress to RFI (default). Write grid point stress to RFI. Determines if grid point strain is written to RFI. Do not write grid point strain to RFI (default). Write grid point strain to RFI. Determines how mass invariants are computed. Mass invariants 6 and 8 are not computed.
CONSTANT Mass invariants 1,2,3 and 9 are computed. FULL NONE PSETID All nine mass invariants are computed. No mass invariants are computed. Selects a set of elements dened in the OUTPUT(PLOT) (including PLOTEL) whose connectivity is exported to face geometry into the RFI. (See Remark 16) No specic sets are selected, thus all grids, geometry and associated modal data are written to RFI (default). The connectivity of a specic element set is used to export face geometry.
NONE setid
4-3
Chapter 4
Describer ALL
Meaning The connectivity of all element sets are used to export face geometry.
REMARKS:
1. The creation of the RecurDyn Flex Input le is applicable in a non-restart SOL 103 analysis only. 2. The creation of the RecurDyn Flex Input le is initiated by RECURDYNRFI FLEXBODY=YES (other describers are optional) along with the inclusion of the bulk data entry DTI,UNITS. 3. The Data Table Input Bulk Data entry DTI,UNITS, which is required for an RECURDYNRFI FLEXBODY=YES run, is used to specify the system of units for the data stored in the RFI (unlike NX Nastran, RecurDyn is not a unitless code). Once identied, the units will apply to all superelements in the model. The complete format is:
DTI UNITS 1 MASS FORCE LENGTH TIME
All entries are required. Acceptable character strings are listed below. Mass: KG - kilogram LBM pound-mass SLUG slug GRAM gram OZM ounce-mass KLBM kilo pound-mass (1000.lbm) MGG megagram Force: N Newton LBF pound-force KGF kilograms-force OZF ounce-force DYNE dyne KN kilonewton KLBF kilo pound-force (1000.lbf) MN - millinewton Length: KM kilometer
4-4
M meter CM centimeter MM millimeter MI mile FT foot IN inch Time: H hour MIN-minute S second MS millisecond 4. Since DTI,UNITS determines all units for the RFI, the units dened in WTMASS, which are important for units consistency in NX Nastran, are ignored in the output to the RFI. For example, if the model mass is kilograms, force in Newtons, length in meters, and time in seconds, then WTMASS would equal 1 ensuring that NX Nastran works with the consistent set of kg, N, and m. The units written to the RFI would be: DTI,UNITS,1,KG,N,M,S. 5. You can create exible body attachment points by dening the component as a superelement or part superelement, in which case the physical external (a-set) grids become the attachment points; or for a residual-only type model, you can use standard NX Nastran ASET Bulk Data entries to dene the attachment points. 6. The eight mass variants are:
4-5
Chapter 4
sp = [xyz]T are the coordinates of grid point p in the basic coordinate system.
p=partitioned orthogonal modal matrix that corresponds to the translational degrees of freedom of grid p. Ip=inertia tensor p. p*=partitioned orthogonal modal matrix that corresponds to the rotational degrees of freedom of grid p. =skew-symmetric matrix formed for each grid translational degree of freedom for each mode. M=number of modes. N=number of grids. 7. To accurately capture the mode shapes when supplying SPOINT/QSET combinations, the number of SPOINTS (ns) should be at least ns=n+(6+p), assuming that residual exibility is on. In the above equation for ns, the number of modes (n) is specied on the EIGR or EIGRL Bulk Data entries; the number of load cases is p. In general, you cant have too many SPOINTs, as excess ones will simply be truncated with no performance penalty. 8. For FLEXBODY=YES runs, residual vectors for the component should always be calculated as they result in a more accurate representation of the component shapes at little additional cost. 9. OMIT or OMIT1 Bulk Data entries are not supported. 10. Lumped mass formulation (default) is required. Either leave PARAM,COUPMASS out of the input le or supply PARAM,COUPMASS,-1 (default) to ensure lumped mass. 11. P-elements are not allowed because they are always use a coupled mass formulation. 12. PARAM,WTMASS,value with a value other than 1.0 may be used with an NX Nastran run generating an RFI. It must have consistent units with regard to the DTI,UNITS Bulk Data entry. Before generating the RFI, NX Nastran will appropriately scale the WTMASS from the physical mass matrix and mode shapes.
4-6
13. There is a distinction between how an RECURDYNRFI FLEXBODY=YES run handles element-specic loads (such as a PLOAD4 entry) versus those that are grid-specic (such as a FORCE entry), especially when superelements are used. The superelement sees the total element-specic applied load. For grid-specic loads, the loads attached to an external grid will move downstream with the grid. That is to say, it is part of the boundary and not part of the superelement. This distinction applies to a superelement run and not to a residual-only or parts superelement run. 14. The loads specied in NX Nastran generally fall into two categories: non-follower or xed direction loads (non-circulatory) and follower loads (circulatory). The follower loads are nonconservative in nature. Examples of xed direction loads are the FORCE entry or a PLOAD4 entry when its direction is specied via direction cosines. Examples of follower loads are the FORCE1 entry or the PLOAD4 entry when used to apply a normal pressure. By default in NX Nastran, the follower loads are always active in SOL 103 and will result in follower stiffness being added to the differential stiffness and elastic stiffness of the structure. In a run with RECURDYNRFI FLEXBODY=YES and superelements, if the follower force is associated with a grid description (such as a FORCE1) and the grid is external to the superelement, the follower load will move downstream with the grid. Thus, the downstream follower contribution to the components stiffness will be lost, which could yield poor results. This caution only applies to a superelement run and not to a residual-only or a part superelement run. 15. OUTGSTRS and OUTGSTRN entries require the use of standard NX Nastran STRESS= or STRAIN= used in conjunction with GPSTRESS= or GPSTRAIN= commands to produce grid point stress or strain. GPSTRESS(PLOT)= or GPSTRAIN(PLOT)= will suppress grid stress or strain print to the NX Nastran .f06 le. 16. To reduce the FE mesh detail for dynamic simulations, PSETID (on the RECURDYNRFI Case Control command) dened with a SET entry is used to dene a set of PLOTELs or other elements used to select grids to display the components in RecurDyn. This option can signicantly reduce the size of the RFI without compromising accuracy in the FunctionBay simulation providing that the mass invariant computation is requested. With superelement analysis, for any of these elements that lie entirely on the superelement boundary (all of the elements grids attached only to a-set or exterior grids), a SEELT Bulk Data entry must be specied to keep that display element with the superelement component. This can also be accomplished using PARAM, AUTOSEEL,YES. The SEELT entry is not required with parts superelements, as boundary elements stay with their component. If the SET entry points to an existing set from the OUTPUT(PLOT) section, this single set is used explicitly to dene elements used to select grids to display the component in RecurDyn. If PSETID does not nd the set ID in OUTPUT(PLOT), it will search sets in the case control for a matching set ID. This matching set ID list then represents a list of OUTPUT(PLOT) dened elements sets, the union of which will be used to dene a set of PLOTELs or other elements used to select grids to display the component in RecurDyn. If the user wishes to select all of the sets in the OUTPUT(PLOT) section, then use PSETID=ALL.
4-7
Chapter 4
The following element types are not supported for writing to an RFI, nor are they supported as a type entry in a set denition in OUTPUT(PLOT): CAABSF, CAEROi, CDUMi, CHACAB, CHACBR, CHBDYx, CDAMP3, CDAMP4, CELAS3, CELAS4, CFLUIDi, CMASS3, CMASS4, CRAC2D, CRAC3D, CTRMEM, CTWIST, CWEDGE, CWELD, and GENEL. 17. Typical NX Nastran data entry requirements are described below. Typical Parameters: PARAM,RESVEC,character_value controls calculation of residual exibility (including inertia relief) modes. In SOL 103, residual exibility is on by default for only component modes (o-set). PARAM,GRDPNT, value - mass invariants 1I, 2I, and 3I will be computed using results of NX Nastran grid point weight generator execution in the basic coordinate system.
Typical Case Control: RECURDYNRFI FLEXBODY=YES is required for RFI generation. METHOD=n is required before or in the rst subcase for modal solutions. SUPORT1=seid is necessary to select a static support set for a residual only linear preload run. SUPER=n,SEALL=n is useful with multiple superelement models to select an individual superelement as a exible body. Cannot be used with a linear STATSUB(PRELOAD) run. OUTPUT(PLOT) is necessary to dene elements used to select grids to display the component in RecurDyn when PSETID=ALL or setid. SET n=list of elements (including PLOTELs) is used to select grids to display the component. OUTPUT(POST) is necessary to dene volume and surface for grid stress or strain shapes. SET n=list is a list of elements for surface denition for grid stress or strain shapes. Stress and strain data in the RFI is limited to the six components (i.e. 3 normal and 3 shear) for a grid point for a given mode. SURFACE n SET n NORMAL z3 is used to dene a surface for writing stress and strain data. Only one FIBER selection is allowed for each SURFACE, thus the use of the FIBRE ALL keyword on the SURFACE case control command will write stresses to the RFI at the Z1 ber location only. Since the FIBRE keyword only applies to stresses, strain data will always be written to the RFI at the MID location. Stress and strain data at grid points can only be written to the RFI for surface and volume type elements (e.g. CQUAD and CHEXA). VOLUME n SET n is a volume denition.
4-8
The default SYSTEM BASIC is required with SURFACE or VOLUME. STRESS(PLOT) is necessary for stress shapes. STRAIN(PLOT) is necessary for strain shapes. GPSTRESS(PLOT) is necessary for grid point stress shapes to be included in the RFI. GPSTRAIN(PLOT) is necessary for grid point strain shapes to be included in the RFI. Typical Bulk Data: DTI,UNITS,1,MASS,FORCE,LENGTH,TIME is required for RFI generation. For input les containing superelements, this command must reside in the main bulk data section. SPOINT,id_list denes and displays modal amplitude. SESET,SEID,grid_list denes a superelement (see GRID and BEGIN BULK SUPER=). The exterior grids will represent the attachment points along with the q-set. SEELT,SEID,element_list reassigns superelement boundary elements to an upstream superelement. RELEASE,SEID,C,Gi is an optional entry that removes DOFs from an attachment grid for which no constraint mode is desired. For example, this allows the removal of rotational degrees of freedom from an analysis where only translational degrees of freedom are required. SEQSET,SEID,spoint_list denes modal amplitudes of a superelement (see SEQSET1). SENQSET,SEID,N denes modal amplitudes of a part superelement. It must reside in the main Bulk Data Section. ASET,IDi,Ci denes attachment points for a residual-only run (see ASET1). QSET1,C,IDi denes modal amplitudes for the residual structure or modal amplitudes for a part superelement (see QSET). SUPORT1,SID,IDi,Ci denes the static support for a preload condition with a residual-only run. This entry is case control selectable. Do not use SUPORT. PLOTEL,EID,Gi can be used, along with existing model elements, to dene elements used to select grids to display the components in RecurDyn. EIGR,SID,METHOD, obtains real eigenvalue extraction (see EIGRL).
18. RECURDYNRFI and ADAMSMNF case control entries cannot be used in the same analysis run. In other words, a RecurDyn RFI le or an ADAMS MNF le can be generated during a particular NX Nastran execution, but not both
4-9
Chapter 4
les at the same time. Attempting to generate both les in the same analysis will cause an error to be issued and the execution to be terminated.
4.2
ADAMS Stress Recovery
NX Nastran 3 introduced the ability to write an ADAMS modal neutral le (MNF) directly from an NX Nastran SOL 103 analysis using the ADAMSMNF case control command. The resulting MNF is imported into ADAMS and used to represent a exible component in a multi-body dynamics analysis. NX Nastran 5 enhances this process with a new results recovery capability. The results from an ADAMS multi-body dynamics analysis, along with an optional component modal denitions le (OUTPUT2 format), are used in a consecutive NX Nastran SOL 103 results recovery solution. The component modal denitions le can be created by NX Nastran during the original solution in which the MNF is created (most efcient method), or the component modal denitions can be recalculated during the nal recovery solution. The nal solution recovers results such as stresses, strains, and element forces. These can be written to an OUTPUT2 le and imported into an NX Nastran compatible post-processor.
Procedure for Results Recovery

The following steps describe the new procedure of results recovery, and introduce the new input options. 1. Initial NX Nastran Sol 103, MNF creation In this step, the ADAMS MNF creation occurs, which is used in step 2. The optional component modal denitions OUTPUT2 le is also created for use in the results recovery solution in step 3. The input le used in step 1 is similar to a SOL 103, but with the addition of the ADAMSMNF case control command, and the DTI,UNITS bulk entry which is used to specify the system of units for the data stored in the MNF. If you choose to create the optional component modal denitions le in this step, you need the new ADMOUT=YES option on the ADAMSMNF case control command, along with the following ASSIGN statement in the le management section: ASSIGN OUTPUT2=name.out STATUS=UNKNOWN UNIT=20 FORM=UNFORM The updates to the ADAMSMNF case control command include the new ADMOUT option, plus the new remarks 17 and 18. The updated ADAMSMNF case control command is given below. 2. Flexbody Analysis in ADAMS The MNF created in step 1 is imported into ADAMS, along with any other MNFs which represent the components of a multi-body assembly. The ADAMS solution stores the dynamic responses into a modal deformations le (MDF) in binary format (OUTPUT2), or in ASCII format (Punch). 3. Final NX Nastran SOL 103, results recovery The MDF created by ADAMS in step 2 is used in another NX Nastran SOL 103 solution to recover results, which can be used to evaluate the stress condition on specic components of the assembly. The new case control command ADMRECVR has been created to enter the needed recovery options. The ADMRECVR command is repeated below for convenience.
4-10
If the component modal denitions were computed in step 1, or created using the original dmap alter mnfx.alt, the resulting OUTPUT2 le can be used in step 3. The following line in the le management section of the NX Nastran input le is required to reference the existing OUTPUT2 le: ASSIGN INPUTT2=<OUTPUT2_lename> UNIT=20 If no component modal denitions exist, there is an option to recompute them in the results recovery solution. It is more efcient to calculate them in step 1 and have the OUTPUT2 le ready for the results recovery step. See the MSRMODE option on the new ADMRECVR case control command below for more information on these options.
Updated/New Case Control Commands
ADAMSMNF
Generates ADAMS Interface Modal Neutral File
Generates ADAMS Interface Modal Neutral File (MNF) during SOL 103.
FORMAT:
EXAMPLES:
ADAMSMNF FLEXBODY=YES
DESCRIBERS:
Describer FLEXBODY NO YES
Meaning Requests the generation of MNF. Standard NX Nastran solution without MNF creation (default). MNF generation requested.
4-11
Chapter 4
Describer FLEXONLY YES NO OUTGSTRS NO YES OUTGSTRN NO YES MINVAR PARTIAL
Meaning Determines if standard DMAP solution runs or not after MNF creation is complete. Only MNF creation occurs (default). MNF le creation occurs along with standard DMAP solution. Determines if grid point stress is written to MNF. Do not write grid point stress to MNF (default). Write grid point stress to MNF. Determines if grid point strain is written to MNF. Do not write grid point strain to MNF (default). Write grid point strain to MNF. Determines how mass invariants are computed. Mass invariants 5 and 9 are not computed.
CONSTANT Mass invariants 1,2,6 and 7 are computed. FULL NONE PSETID All nine mass invariants are computed. No mass invariants are computed. Selects a set of elements dened in the OUTPUT(PLOT) section (including PLOTEL) or on a sketch le whose connectivity is exported to face geometry to be used in ADAMS. (See Remark 15) All grids, geometry and associated modal data is written to MNF (default). The connectivity of a specic element set is used to export face geometry. The connectivity of all element sets are used to export face geometry. The connectivity of element faces dened on a sketch le is used to export face geometry. Note that the value must be a negative number to distinguish it from a setid value. Requests that the FLEXBODY run output an NX Nastran OP2 le for use in post processing of ADAMS/Flex results. OP2 le will not be generated (default).
NONE setid ALL sktunit
ADMOUT NO
4-12
Describer YES CHECK NO YES

REMARKS:
Meaning OP2 le will be generated. Requests debug output be written to the f06 le when ADMOUT=YES (See Remark 19). No debug output will be written (default). Debug output will be written.
1. The creation of the Adams MNF, which is applicable in a non-restart SOL 103 analysis only, is initiated by ADAMSMNF FLEXBODY=YES (other describers are optional) along with the inclusion of the bulk data entry DTI,UNITS. 2. The Data Table Input Bulk Data entry DTI,UNITS, which is required for an ADAMSMNF FLEXBODY=YES run, is used to specify the system of units for the data stored in the MNF (unlike NX Nastran, ADAMS is not a unitless code). Once identied, the units will apply to all superelements in the model. The complete format is:
DTI UNITS 1 MASS FORCE LENGTH TIME
All entries are required. Acceptable character strings are listed below. Mass: KG - kilogram LBM pound-mass (0.45359237 kg) SLUG slug (14.5939029372 kg) GRAM gram (1E-3 kg) OZM ounce-mass (0.02834952 kg) KLBM kilo pound-mass (1000 lbm) (453.59237 kg) MGG megagram (1E3 kg) MG milligram (1E-6 kg) MCG microgram (1E-9 kg) NG nanogram (1E-12 kg) UTON U.S. ton (907.18474 kg) SLI slinch (175.1271524 kg) Force: N Newton LBF pound-force (4.44822161526 N) KGF kilograms-force (9.80665 N) OZF ounce-force (0.2780139 N) DYNE dyne (1E-5 N)
4-13
Chapter 4
KN kilonewton (1E3 N) KLBF kilo pound-force (1000 lbf) (4448.22161526 N) MN millinewton (1E-3 N) MCN micronewton (1E-6 N) NN nanonewton (1E-9 N) Length: M meter KM kilometer (1E3 m) CM centimeter (1E-2 m) MM millimeter (1E-3 m) MI mile (1609.344 m) FT foot (0.3048 m) IN inch (25.4E-3 m) MCM micrometer (1E-6 m) NM nanometer (1E-9 m) A Angstrom (1E-10 m) YD yard (0.9144 m) ML mil (25.4E-6 m) MCI microinch (25.4E-9 m) Time: S second H hour (3600.0 sec) MIN-minute (60.0 sec) MS millisecond (1E-3 sec) MCS microsecond (1E-6 sec) NS nanosecond (1E-9 sec) D day (86.4E3 sec) 3. Since DTI,UNITS determines all units for the MNF, the units dened in WTMASS, which are important for units consistency in NX Nastran, are ignored in the output to the MNF. For example, if the model mass is kilograms, force in Newtons, length in meters, and time in seconds, then WTMASS would equal 1 ensuring that NX Nastran works with the consistent set of kg, N, and m. The units written to the MNF would be: DTI,UNITS,1,KG,N,M,S. 4. You can create exible body attachment points by dening the component as a superelement or part superelement, in which case the physical external (a-set) grids become the attachment points; or for a residual-only type model, you can use standard NX Nastran ASET Bulk Data entries to dene the attachment points.
4-14
5. The nine mass variants are:
sp = [xyz]T are the coordinates of grid point p in the basic coordinate system.
4-15
Chapter 4
p=partitioned orthogonal modal matrix that corresponds to the translational degrees of freedom of grid p. Ip=inertia tensor p. p*=partitioned orthogonal modal matrix that corresponds to the rotational degrees of freedom of grid p. =skew-symmetric matrix formed for each grid translational degree of freedom for each mode. M=number of modes. N=number of grids. 6. To accurately capture the mode shapes when supplying SPOINT/QSET combinations, the number of SPOINTS (ns) should be at least ns=n+(6+p), assuming that residual exibility is on. In the above equation for ns, the number of modes (n) is specied on the EIGR or EIGRL Bulk Data entries; the number of load cases is p. In general, you cant have too many SPOINTs, as excess ones will simply be truncated with no performance penalty. 7. For FLEXBODY=YES runs, residual vectors for the component should always be calculated as they result in a more accurate representation of the component shapes at little additional cost. 8. OMIT or OMIT1 Bulk Data entries are not supported. 9. Lumped mass formulation (default) is required. Either leave PARAM,COUPMASS out of the input le or supply PARAM,COUPMASS,-1 (default) to ensure lumped mass. 10. P-elements are not allowed because they are always use a coupled mass formulation. 11. PARAM,WTMASS,value with a value other than 1.0 may be used with an NX Nastran run generating an MNF. It must have consistent units with regard to the DTI,UNITS Bulk Data entry. Before generating the MNF, NX Nastran will appropriately scale the WTMASS from the physical mass matrix and mode shapes. 12. There is a distinction between how an ADAMSMNF FLEXBODY=YES run handles element-specic loads (such as a PLOAD4 entry) versus those that are grid-specic (such as a FORCE entry), especially when superelements are used. The superelement sees the total element-specic applied load. For grid-specic loads, the loads attached to an external grid will move downstream with the grid. That is to say, it is part of the boundary and not part of the superelement. This distinction applies to a superelement run and not to a residual-only or parts superelement run. 13. The loads specied in NX Nastran generally fall into two categories: non-follower or xed direction loads (non-circulatory) and follower loads (circulatory). The follower loads are nonconservative in nature. Examples of xed direction loads are the FORCE entry or a PLOAD4 entry when its direction is specied via direction cosines. Examples of follower loads are the FORCE1 entry or the PLOAD4 entry when used to apply a normal pressure.
4-16
By default in NX Nastran, the follower loads are always active in SOL 103 and will result in follower stiffness being added to the differential stiffness and elastic stiffness of the structure. In a run with ADAMSMNF FLEXBODY=YES and superelements, if the follower force is associated with a grid description (such as a FORCE1) and the grid is external to the superelement, the follower load will move downstream with the grid. Thus, the downstream follower contribution to the components stiffness will be lost, which could yield poor results. This caution only applies to a superelement run and not to a residual-only or a part superelement run. 14. OUTGSTRS and OUTGSTRN entries require the use of standard NX Nastran STRESS= or STRAIN= used in conjunction with GPSTRESS= or GPSTRAIN= commands to produce grid point stress or strain. GPSTRESS(PLOT)= or GPSTRAIN(PLOT)= will suppress grid stress or strain print to the NX Nastran .f06 le. 15. To reduce the FE mesh detail for dynamic simulations, PSETID (on the ADAMSMNF case control command) dened with a SET entry (i.e. setid) is used to dene a set of PLOTELs or other elements used to select grids to display the components in ADAMS. This option can signicantly reduce the size of the MNF without compromising accuracy in the ADAMS simulation providing that the mass invariant computation is requested. With superelement analysis, for any of these elements that lie entirely on the superelement boundary (all of the elements grids attached only to a-set or exterior grids), a SEELT Bulk Data entry must be specied to keep that display element with the superelement component. This can also be accomplished using PARAM, AUTOSEEL,YES. The SEELT entry is not required with parts superelements, as boundary elements stay with their component. If the SET entry points to an existing set from the OUTPUT(PLOT) section, this single set is used explicitly to dene elements used to select grids to display the component in ADAMS. If PSETID does not nd the set ID in OUTPUT(PLOT), it will search sets in the case control for a matching set ID. This matching set ID list then represents a list of OUTPUT(PLOT) dened elements sets, the union of which will be used to dene a set of PLOTELs or other elements used to select grids to display the component in ADAMS. If the user wishes to select all of the sets in the OUTPUT(PLOT) section, then use PSETID=ALL. The following element types are not supported for writing to an MNF, nor are they supported as a type entry in a set denition in OUTPUT(PLOT): CAABSF, CAEROi, CDUMi, CHACAB, CHACBR, CHBDYx, CDAMP3, CDAMP4, CELAS3, CELAS4, CFLUIDi, CMASS3, CMASS4, CRAC2D, CRAC3D, CTRMEM, CTWIST, CWEDGE, CWELD, and GENEL. PSETID can also point to a sketch le using PSETID=-sktunit, where sktunit references an ASSIGN statement of the form
ASSIGN SKT=sketch_file.dat,UNIT=sktunit.
The grids dened for the elements faces in the sketch le, along with all external (i.e. boundary) grids for the superelements, will be the only grids (and their associated data) written to the MNF. The format of the sketch le, which describes the mesh as a collection of faces, must be as follows:
face_count
4-17
Chapter 4
face_1_node_count face_1_nodeid_1 face_1_nodeid_2 ... face_2_node_count face_2_nodeid_1 face_2_nodeid_2 ... <etc>
Faces must have a node count of at least two. For example, a mesh comprised of a single brick element might be described as follows:
6 4 4 4 4 4 4 1000 1007 1000 1001 1002 1003 1001 1006 1004 1005 1006 1007 1002 1005 1005 1006 1007 1004 1003 1004 1001 1002 1003 1000
Alternatively, the mesh might be described as a stick gure using a collection of lines (two node faces), as shown below:
8 2 2 2 2 2 2 2 2 101 102 103 104 105 106 107 108 102 103 104 105 106 107 108 109
16. Typical NX Nastran data entry requirements are described below. Typical Parameters: PARAM,RESVEC,character_value controls calculation of residual exibility (including inertia relief) modes. In SOL 103, residual exibility is on by default for only component modes (o-set). PARAM,GRDPNT, value - mass invariants 1I, 2I, and 7I will be computed using results of NX Nastran grid point weight generator execution in the basic coordinate system.
Typical case control: ADAMSMNF FLEXBODY=YES is required for MNF generation. METHOD=n is required before or in the rst subcase for modal solutions. SUPORT1=seid is necessary to select a static support set for a residual only linear preload run. SUPER=n,SEALL=n is useful with multiple superelement models to select an individual superelement as a exible body. Cannot be used with a linear STATSUB(PRELOAD) run. OUTPUT(PLOT) is necessary to dene elements used to select grids to display the component in ADAMS when PSETID=ALL or setid. SET n=list of elements (including PLOTELs) is used to select grids to display the component. OUTPUT(POST) is necessary to dene volume and surface for grid stress or strain shapes.
4-18
SET n=list is a list of elements for surface denition for grid stress or strain shapes. Stress and strain data in the MNF is limited to the six components (i.e. 3 normal and 3 shear) for a grid point for a given mode. SURFACE n SET n NORMAL z3 is used to dene a surface for writing stress and strain data. Only one FIBER selection is allowed for each SURFACE, thus the use of the FIBRE ALL keyword on the SURFACE case control command will write stresses to the MNF at the Z1 ber location only. Since the FIBRE keyword only applies to stresses, strain data will always be written to the MNF at the MID location. Stress and strain data at grid points can only be written to the MNF for surface and volume type elements (e.g. CQUAD and CHEXA). VOLUME n SET n is a volume denition. The default SYSTEM BASIC is required with SURFACE or VOLUME. STRESS(PLOT) is necessary for stress shapes. STRAIN(PLOT) is necessary for strain shapes. GPSTRESS(PLOT) is necessary for grid point stress shapes to be included in the MNF. GPSTRAIN(PLOT) is necessary for grid point strain shapes to be included in the MNF. Typical Bulk Data: DTI,UNITS,1,MASS,FORCE,LENGTH,TIME is required for MNF generation. For input les containing superelements, this command must reside in the main bulk data section. SPOINT,id_list denes and displays modal amplitude.SESET,SEID,grid_list denes a superelement (see GRID and BEGIN BULK SUPER=). The exterior grids will represent the attachment points along with the q-set. SEELT,SEID,element_list reassigns superelement boundary elements to an upstream superelement. RELEASE,SEID,C,Gi is an optional entry that removes DOFs from an attachment grid for which no constraint mode is desired. For example, this allows the removal of rotational degrees of freedom from an analysis where only translational degrees of freedom are required. SEQSET,SEID,spoint_list denes modal amplitudes of a superelement (see SEQSET1). SENQSET,SEID,N denes modal amplitudes of a part superelement. It must reside in the main Bulk Data Section. ASET,IDi,Ci denes attachment points for a residual-only run (see ASET1).
4-19
Chapter 4
QSET1,C,IDi denes modal amplitudes for the residual structure or modal amplitudes for a part superelement (see QSET). SUPORT1,SID,IDi,Ci denes the static support for a preload condition with a residual-only run. This entry is case control selectable. Do not use SUPORT. PLOTEL,EID,Gi can be used, along with existing model elements, to dene elements used to select grids to display the components in ADAMS. EIGR,SID,METHOD, obtains real eigenvalue extraction (see EIGRL).
17. ADAMSMNF and RECURDYNRFI case control entries cannot be used in the same analysis run. In other words, an ADAMS MNF le or a RecurDyn RFI le can be generated during a particular NX Nastran execution, but not both les at the same time. Attempting to generate both les in the same analysis will cause an error to be issued and the execution to be terminated. 18. The ADMOUT=YES option is used when you would like results recovery (using the ADMRECVR case control entry) from an ADAMS/Flex analysis. This option requires the following assignment command: ASSIGN OUTPUT2=name.out STATUS=UNKNOWN UNIT=20 FORM=UNFORM inserted into the le management section of the NX Nastran input le. It will cause an OP2 le with a .out extension to be generated, which then can be used as input into an NX Nastran SOL 103 run using the ADMRECVR case control capability to perform results recovery from an ADAMS/Flex analysis. FLEXBODY=YES is required with its use. The data blocks output are: MGGEW - physical mass external sort with weight mass removed MAAEW - modal mass KAAE - modal stiffness CMODEXT - component modes. This capability is limited to one superelement per NX Nastran model. Residual-only analyses are not supported. 19. Setting CHECK=YES (which is only available when ADMOUT=YES) is not recommended for models of realistic size due to the amount of data that will be written to the f06. 20. The ADAMSMNF data routines use the environment variable TMPDIR for temporary storage during the processing of mode shape data. As a result, TMPDIR must be dened when using ADAMSMNF. TMPDIR should equate to a directory string for temporary disk storage, preferably one with a large amount of free space.
ADMRECVR
ADAMS stress recovery.
Recovers stress results from an ADAMS/Flex analysis.
4-20
FORMAT:
EXAMPLES:
ASSIGN INPUTT2=adams_results.mdf UNIT=13 ... CEND STRESS(PLOT)=100 ADMRECVR

DESCRIBERS:
Describer ADMFORM BINARY ASCII MSRMODE 0
Meaning Species the format of the ADAMS/Flex modal deformations le (see remarks 1 and 2) OUTPUT2 le (default) PUNCH le. Species stress recovery type (see remarks 6 and 7). Component modal denitions are stored in an OUTPUT2 le (specically, a *.out le was created by using ADMOUT=YES on the ADAMSMNF case control in a pre-ADAMS/Flex NX Nastran run). The OUTPUT2 les used in this case do not contain data blocks used for MNF creation (default). Same as option 0, except that the OUTPUT2 le will contain 10 additional data blocks used for MNF creation by an ADAMS pre-processor (specically, a *.out le created through use of the mnfx.alt DMAP alter capability). No le reference (specically, component modal denitions will be recomputed) Requests the addition of rigid body motion with modal deformations (see remark 5). Do not include rigid body motion (default).
2 RGBODY NO
4-21
Chapter 4
Describer YES MSGLVL
Meaning Include rigid body motion. Level of diagnostic output from Lanczos eigensolver when component modal denitions are determined (applies only when MSRMODE=2). No output (default). Warning and fatal messages. Summary output. Detailed output on cost and convergence. Detailed output on orthogonalization. Requests debug output be written to the f06 le (See Remark 9). No debug output will be written (default). Debug output will be written.
0 1 2 3 4 ADMCHK NO YES
REMARKS:
1. When modal deformations to be read are in binary (OUTPUT2) format (specically, ADMFORM=BINARY), the following statement needs to be specied near the top of the NX Nastran input le in the le management section: ASSIGN INPUTT2=<MDFilename> UNIT=13 where <MDFilename> is the name of the modal deformations le from ADAMS. 2. To input the modal deformations le from ADAMS in ASCII (Punch) format (specically, ADMFORM=ASCII), the following statement needs to be included in the bulk data section: INCLUDE <MDFilename> where <MDFilename> is the name of the modal deformations le. 3. Dynamic stress/strain output can either be in .f06, PUNCH, and/or OUTPUT2 according to standard NX Nastran functionality. However, stress recovery in NX Nastran from ADAMS/Flex results does not support XYPLOT output. 4. If displacements, stresses, and/or strains are to be available for post processing, one or more of the following statements must appear in the case control section of the NX Nastran input le: DISP(PLOT) = <set id> STRAIN(FIBER,PLOT) = <set id> STRESS(PLOT) = <set id> 5. Rigid body motions from an ADAMS simulation are included in the modal deformations le, but they are not applied unless the RGBODY keyword is set to YES and the SORT1 option is included in the DISP(PLOT) command in case
4-22
control. Including rigid body motion affects the display and animation of the exible component, but it has no effect on dynamic stresses. 6. For MSRMODE=0 or 1, stress recovery references the OUTPUT2 le obtained from the initial CMS analysis (specically, ADMOUT=YES on the ADAMSMNF case control entry or use of the mnfx.alt DMAP alter capability). No other les are required. The geometric data needs to be included in the bulk data of the NX Nastran input le because geometry is missing from the OUTPUT2 le. This mode of stress recovery is faster than the MSRMODE=2 mode. To reference this OUTPUT2 le the following line needs to be included in the le specication section of the NX Nastran input le: ASSIGN INPUTT2=<OUTPUT2_lename> UNIT=20 7. For MSRMODE=2, no les are referenced for stress recovery. Instead, a full CMS reanalysis is performed to build the reference data for the stress recovery analysis. Obviously, the analysis time is signicantly far greater compared to the MSRMODE=0 or 1 method, but this method frees up disk space. There is also risk in using this method. If the reanalysis generates slightly different component eigenvalues or eigenvectors than were generated during the creation of the ADAMS MNF in the initial NX Nastran run, then the ADAMS results in the ADAMS MDF (modal deformation le) will be inconsistent and incorrect results will be recovered. Something as simple as a sign change for one eigenvector will cause incorrect results to be recovered. It is, therefore, highly recommended that MSRMODE=0 or 1 always be used. 8. This capability must be performed in SOL 103 and is limited to one superelement per NX Nastran model. Residual-only analyses are not supported. 9. Setting ADMCHK=YES is not recommended for models of realistic size due to the amount of data that will be written to the f06.
4.3 Component Mode Reduction of the Residual Structure Procedure for Flexbody Solutions
The set-up of a ex body modal solution in NX Nastran for export to ADAMS MNF or Recurdyn RFI les requires special considerations for the modal solution. This is because ex bodies will be attached to other components in the multi-body dynamic (MBD) simulation and local exibility effects at the connection locations are thus important. In a standard normal mode solution, the solved modes give good representation of the global dynamics of the component, but the local stiffness effects at the connections are typically not captured because of modal truncation. A modal solution method called Component Mode Reduction of Residual Structure (CMR of RS) is recommended for ex body solutions because it includes both global and local effects. The CMR of RS method is a variation to the standard SOL 103 modal solution in NX Nastran. It has long been employed by dynamic analysts to include the local stiffness effects in critical areas of a model. The CMR of RS method is a two-step modal solution approach and is equivalent to a Craig-Bampton superelement reduction on the residual structure.
4-23
Chapter 4
CMR of RS Theory
The connection degrees-of-freedom (DOF) for the ex body are equivalent to the exterior DOF of a superelement. In Nastran set terminology, this means the connection DOF are in the t-set. The interior DOF of the ex body component are part of the o-set. The union of the t-set and o-set is the f-set. The rst solution step of the CMR or RS method is to compute normal modes and static constraint modes. The normal modes are computed with the specied t-set DOF restrained. The normal mode shape matrix, , can be used to transform the physical displacement uf (and its partitions ut and uo) into the generalized modal DOF, q. The transformation relation is
{u } = u = {q}
0
f
ut o
Equation 4-1. The static constraint modes are static deection shapes computed by applying unit deections at the t-set DOFs. This is the same thing as performing a Guyan reduction on the o-set DOF (see chapter 11 in the NX Nastran Basic Dynamic Analysis Users Guide for more information). The static constraint modes transform full physical displacement uf into just the physical DOF, ut. The transformation relation is
{u } = u = K
f
ut o
I
1 oo
{u } K ot t
Equation 4-2. The matrices Koo and Kot are partitions of the Kff stiffness matrix. The normal mode transformation is a good reduction for the global dynamics and the constraint mode transformation is a good reduction for the local stiffness at the connection DOF. These transformations can be combined into a single transformation as
{u } = u = K
f
ut o
I
1 oo
0 ut q
Equation 4-3. The combined transformation captures both the global dynamics and the local stiffness of the component in the reduction. In the second solution step, the transformation in equation (3) is used to mathematically reduce the model from the full physical DOF set to a reduced set of generalized modal DOFs and the t-set physical DOFs. The reduced DOF set size is equal to the sum of the number of normal modes and number of connection DOFs. Since the normal modes and static modes are not orthogonal to each other, the reduced mass and stiffness matrices are not diagonal as they would be with a pure normal mode reduction. However, a second modal solution is performed on the reduced system resulting in a new set of modes. The nal modes are orthogonal to each other and importantly, capture both global dynamics and local stiffness characterizations of the ex body. It is this second set of modes shapes that are exported to the ex body le. It should be noted that it is critical that all the modes of
4-24
the reduced system should be solved in the second modal solution so that full global and local effects embodied in the reduction are retained.
CMR of RS Specication
The CMR of RS method can be set-up in an NX Nastran solution by adding a few additional solution control cards to a standard SOL 103 solution. In the case control, an RSMETHOD card is needed in addition to the standard METHOD card. The purpose of each is: RSMETHOD Identies modal solution controls cards (EIGR or EIGRL) to be used for the rst modal solution. Typically the control cards dene a frequency range to compute the modes. Modes outside that range are neglected. METHOD Identies modal solution controls cards (EIGR or EIGRL) to be used for the second modal solution. This is the modal solution of the reduced system. It is important to solve for all the modes of the reduced system. This can be done by specifying a wide frequency range on the EIGR or EIGRL cards. The bulk data section of the input le needs denition of the t-set and q-set DOF. These are dened by these cards: ASET SPOINT QSET Denes the connection DOFs (these are in the t-set). Creates additional DOFs. Identies SPOINT DOFs as modal reduction DOFs and places them in the q-set.
For the rst modal solution (RSMETHOD), it is recommended that the solved frequency range be at least twice the frequency range of interest in the MBD solution. For example, if the MBD solution is to be accurate to 1500 Hz, the rst modal solution of the components should capture modes to at least 3000 Hz. The user needs to be sure that the number of q-set DOF created by QSET and SPOINT cards is larger than the actual number of modes computed from RSMETHOD. As mentioned already, all the modes of the reduced system need to be computed for the second modal solution (METHOD). The user can verify that all the modes have been computed by conrming that the number of modes is equal to the reduced DOF set size (the sum of the number of normal modes and number of connection DOFs).
Mixed Boundary Reduction

Another variation of the CMR of RS method is to use a mixed boundary reduction. In the standard solution, all the t-set DOF are restrained in the rst eigenvalue solution. With mixed boundary conditions, the user can designate that some of the t-set DOF will be unrestrained in this solution. All t-set DOF are treated the same in the constraint mode solution. The reason for using the mixed boundary approach is to add exterior DOF to the solution that are more appropriately considered as a free DOF rather than a connection DOF. For example if a component will have four nodes used as connection points in the MBD analysis, those points should be restrained exterior points. If there are other points on the component that will not be connections but may be a marker location to track response, or to apply a force, those nodes may best be treated as an unrestrained exterior point. Additional bulk data cards for the mixed boundary approach are: CSE Denes the exterior DOFs that will be unrestrained in the rst eigenvalue solution.
4-25
Chapter 4
BSET
Denes the exterior DOFs that will be restrained in the rst eigenvalue solution. BSET and ASET are treated the same in this case.
Example
Below is a sample input le set-up for doing a CMR of RS solution for a Recurdyne exible body.
$ SOL 103 CEND $ $ CASE CONTROL $ TITLE = Solution 1 ECHO = NONE RECURDYNRFI,FLEXBODY=yes, FLEXONLY=no $ Creates RFI file RSMETH=100 $ Points to 1st modal extraction method METHOD = 200 $ Points to 2nd modal extraction method $ DISPLACEMENT(PLOT) = ALL STRESS(PLOT,CORNER) = ALL $ $ BULK DATA $ BEGIN BULK DTI,UNITS,1,KG,N,MM,S $ Units used in MBD solve $ $ SOLUTION CARDS $ EIGRL, 100, , 400.00 $ Eigenvalue extraction method for 1st modes EIGR, 200, AHOU, , , , 150 $ Eigenvalue extraction method for 2nd modes SPOINT,100001,thru,100101 $ SPOINTS > 1st Modes QSET1,,100001,thru,100101 $ QSET points same as SPOINTS PARAM AUTOSPC YES PARAM GRDPNT 0 PARAM POST -2 $ $ Flex body connection ASET points ASET1,123456,26 ASET1,123456,35 ASET1,123456,44 ASET1,123456,53 ASET1,123456,62 ASET1,123456,71 ASET1,123456,80 ASET1,123456,89 $ $ Rest of model . $ . .
Case Control Section RECURDYNRFI,FLEXBODY=yes,FLEXONLY=no Used to request calculation of the ex body. RSMETH = 100 This card points to the EIGRL card used for the rst modal solution METHOD = 200 This card points to the EIGR card used for the second modal solution
4-26
Bulk Data Section DTI,UNITS,1,KG,N,MM,S This denes the set of units used and is required by the MBD solver EIGRL, 100, , 400.00 This is the extraction method for the step-1 modal solve. Recall that you want the upper frequency limit to be about twice the maximum frequency of interest in the structure EIGR, 200, AHOU, , , , 150 This is the extraction method for the step-2 modes The Automatic Householder option is recommended since it will assure that you get all the modes of the reduced component ND (or 150 in this example), is the desired number of eigenvectors, this number has to be greater that the number of ASET plus the number of step-1 modes In this example, there are 8 ASET points, each with 6 DOF, and about 74 step-1modes, so we need 122 step-2 modes (8 x 6 + 74, its OK to request more) SPOINT,100001,thru,100101 SPOINTs are scalar DOFs and describe the modal degrees of freedom, you need more SPOINTs than expected step-1 modes, it doesnt hurt to have too many The numbering range has to be outside the range of grid points QSET1,,100001,thru,100101 The QSET card puts the SPOINTs in the q-set Use the same range as the SPOINTs ASET1,123456,grid number The ASET identies the connection grids and the DOF at those grids. In this case all 6 DOF are used as connection DOF, which assumes that all will be connected by xed joints in the MBD model
4-27
Chapter
5-1
Chapter 5
NX Nastran 5 introduces a new optimizer option, UGS-ADS. The newly introduced optimizer is based on the code ADS which was developed by Dr. Garrett N. Vanderplaats around 1985, and includes modications to the original code. The original form of the code is described in NASA Contractor Report 177985, published in September 1985. With the addition of UGS-ADS into NX Nastran, there are now two choices for an optimizer: DOT, the original optimizer, which remains the default for NX Nastran 5 UGS-ADS, the new optimizer
5.1
Input Requirements
Only two points need to be noted to use UGS-ADS as the optimizer of choice with NX Nastran 5: 1. In NX Nastran 5, UGS-ADS supports only the modied feasible directions as the optimization method. Any other user selection with UGS-ADS will automatically revert to the modied feasible directions method. Other optimization methods will be supported in future versions. 2. UGS-ADS can be selected by assigning the value 1 to the NASTRAN system cell 425 (ADSOPT). In this case, the NASTRAN statement would look like: NASTRAN ADSOPT = 1 -orNASTRAN SYSTEM(425) = 1 Otherwise, the system cell 425 (ADSOPT) will remain as 0 (default), and the optimization will proceed using DOT. Everything else related to optimization remains the same as in previous versions of NX Nastran.
5-2
Chapter
Advanced Nonlinear
6-1
Chapter 6
Advanced Nonlinear
6.1
Contact Improvements
Contact Small Displacement Option

A typical nonlinear solution can iterate many times as the geometry and material conditions are updated. The contact conditions also update as the geometry conditions change, which adds solution time. There may be occasions when preventing the contact conditions from updating between nonlinear iterations will not adversely affect the nonlinear results, for example, when displacements are small in the region where contact occurs. The CTDISP option has been created on the NXSTRAT bulk entry to prevent contact conditions from updating. The CTDISP options are as follows: 0 or 2 Use large displacement formulation (contact conditions are updated) (default) 1 Use small displacement formulation (contact conditions are not updated) CTDISP is a global option since it applies to all contact denitions in the model. If you would like to prevent or allow a specic contact set from updating, use the DISP option on the BCTPARA bulk entry. The DISP options are as follows: 0 Use the formulation selected by CTDISP in NXSTRAT entry (default) 1 Use small displacement formulation (contact conditions are not updated) 2 Use large displacement formulation (contact conditions are updated)
Contact Stability Option with Rigid Body Motion

Rigid body motion in a contact solution can add instability making convergence difcult. A new option on the NXSTRAT bulk entry, CTDAMP, has been created to stabilize the portions of the model experiencing rigid body motion, thus helping the solution to continue and converge. The CTDAMP options on the NXSTRAT bulk entry are as follows: 0 No stabilization damping is applied (default). 1 Stabilization damping is applied at the rst time step only. The specied damping coefcients are applied and ramped down to zero by the end of the rst time step. 2 The specied stabilization damping coefcients are applied at all time steps.
Contact Initial Gap and Penetration

The INIPENE parameter on the BCTPARA bulk entry is used when the goal is for a pair of contact regions to be initially touching without interference, but due to the faceted nature of nite elements around curved geometry, some of the element faces may have a slight penetration. A new option is available when INIPENE=3 which will include both penetration and now gaps. This option is particularly benecial when contact conditions are dened on the faces of concentric cylinders. The INIPENE options on the BCTPARA bulk entry are as follows: 0 Initial penetrations are eliminated. 1 Initial penetrations are eliminated and the list of pentrating nodes is printed. 2 Initial penetrations are ignored. In successive steps, each contractor node is allowed to penetrate the target up to its initial penetration.
6-2
Advanced Nonlinear
3 Initial penetrations or gaps are overridden by GAPVAL. This option is not available for rigid target algorithm (TYPE = 2).
Contact Restrictions Removed

In NX Nastran 4.1, a new contact segment option became available for SOL 601 which improved contact results. This improvement matched the order of the contact segments with the order of the elements. When quadratic contact segments were created on quadratic elements, the resulting contact tractions were signicantly better. The new parameter CSTYPE was created on the bulk data entry NXSTRAT to support the new contact segment creation. CSTYPE=1 was more accurate and gave better contact traction results, especially when used with quadratic elements (for example, 10-grid tetrahedral elements). Restrictions Eliminated in Version 5 The following restrictions which existed in NX Nastran 4.1 when CSTYPE=1 was used have been eliminated in NX Nastran 5: Contact surface offset must be zero (default), that is, OFFSET=0.0 in BCTPARA Only frictionless contact (default) or regular Coulomb friction is allowed, that is, FRICMOD=0 or 1 in BCTPARA Consistent contact stiffness cannot be used, that is, CSTIFF=0 in BCTPARA Continuous normal must be used (default), that is, SEGNORM=0 or 1 on BCTPARA Initial penetration cannot be ignored (default), that is, INIPENE=0 or 1 in BCTPARA
Existing Restrictions These restrictions remain in NX Nastran 5 when CSTYPE=1 is used: Only constraint function algorithm (default) may be used, that is, TYPE=0 in BCTPARA Only single-sided contact (default) is allowed, that is, NSIDE=1 in BCTPARA No tied contact (default), that is, TIED=0 in BCTPARA
6.2
Improved Convergence for Contact with Friction
Contact models with friction are difcult to converge at times. The contact algorithm has been improved in NX Nastran 5 to be more robust for contact with friction. The new improved friction algorithm is used by default, but the old algorithm can be selected by assigning the new FRICALG parameter on the NXSTRAT bulk entry to 1. In addition, a new friction delay option is available by setting the new FRICDLY parameter on the BCTPARA bulk entry to "1". Delaying the application of friction may also improve the convergence of the solution. The default for FRICDLY is "0" meaning no delay occurs. The updated NXSTRAT and BCTPARA bulk entries are provided below.
NXSTRAT
Strategy Parameters for SOLs 601 and 701
NX Nastran 5.0 Release Guide 6-3
Chapter 6
Advanced Nonlinear
Denes parameters for solution control and strategy in advanced nonlinear structural analysis.
FORMAT: 1 NXSTRAT 2 ID Param4 3 Param1 Value4 4 Value1 Param5 5 Param2 Value5 6 Value2 -etc7 Param3 8 Value3 9 10
EXAMPLE: NXSTRAT 1 ATSNEXT AUTO 3 1 MAXITE 30 RTOL 0.005
FIELDS:
Field ID PARAMi
Contents Identication number. Currently not used. (Integer 0) Name of the NXSTRAT parameter. Allowable names are given in Table 8-33. See remark 1 for parameters applicable to SOL 701. (Character) Value of the parameter. See Table 8-33. (Real or integer)
VALUEi
Table 6-1. NXSTRAT Parameters

Name Description Analysis Control Parameters SOLVER Selects the solver to use. (Integer; Default = 0) 0 Direct sparse solver 1 Multigrid solver 2 3D iterative solver. This solver is effective for models with large numbers of higher order 3D solid elements, i.e., CTETRA and CHEXA elements with mid-side nodes. AUTO Indicates whether automatic incrementation scheme is enabled. (Integer; Default = 0) 0 No automatic incrementation scheme is used 1 Automatic time stepping (ATS) scheme is enabled 2 Automatic load-displacement control (LDC) scheme is enabled
6-4
Advanced Nonlinear

Name NPOSIT Description Indicates whether analysis continues when the system matrix is not positive denite. (Integer; Default = 0) 0 Analysis may stop 1 Analysis continues Notes: If NPOSIT=0, analysis stops unless the ATS or LDC scheme is enabled; i.e. AUTO=1 or AUTO=2 contact analysis is being performed It is not recommended to set NPOSIT=1 for a linear analysis. MASSTYP Selects the type of mass matrix to be used in dynamic analysis. (Integer; Default = 1) 0 Lumped mass is used 1 Consistent mass is used Analysis Options TINT* Integration order for the local t-direction (through thickness) of shell elements with elasto-plastic materials. By default, 5-point Newton-Cotes is used for single-layered shell and 3-point Newton-Cotes is used for multi-layered shell. Note that 2-point Gauss integration is always used for all shell elements with elastic materials. (Integer; Default = 0). 1 TINT 6 Gauss integration method with integration order TINT -3, -5, -7 Newton-Cotes integration with order -TINT ICMODE* Indicates whether incompatible modes are used for 4-node shell elements. (Integer; Default =1 for SOL 601 and 0 for SOL 701) 0 - Incompatible modes are not used 1 - Incompatible modes are used MSTAB Indicates whether the stiffness matrix stabilization feature is used. (Integer; Default = 0) 0 Matrix stabilization is not used 1 Matrix stabilization is used MSFAC DTDELAY* Matrix stabilization factor. (Real; Default = 1.0E-12) Element death time delay. (Real; Default = 0.0) When an element is too deformed and becomes dead, its contribution to the overall stiffness of the structure is removed. By specifying DTDELAY > 0.0, the contribution from the element stiffness is gradually reduced to zero over time DTDELAY instead of being suddenly removed. This may help in the convergence of the solution. SDOFANG* Angle used to determine whether a shell mid-surface node is assigned 5 or 6 degrees of freedom. (Real; Default = 5.0)
6-5
Chapter 6
Advanced Nonlinear

Name UPFORM* Description Indicates whether u/p formulation is used for elements. Note that u/p formulation is always used for hyperelastic elements and always not used for hyperfoam elements and elastic elements with Poissons ratio less than 0.48. It is also not used for gasket elements (Integer; Default = 0) 0 u/p formulation is not used 1 u/p formulation is used instead of displacement-based formulation ULFORM* Indicates which large strain formulation is used for 4-node shell elements. (Integer; Default = 0) 0 Updated Lagrangian-Jaumann (ULJ) formulation is used if rigid-target algorithm contact is used or SOL 701 is run. Otherwise, Updated Lagrangian-Hencky (ULH) formulation is used. 1 Use ULH formulation 2 Use ULJ formulation Note: For shell elements which are not 4-noded, the ULJ formulation is always used for large strains. For 3-D solid, plane strain and axisymmetric elements, ULH formulation is always used for large strains. In restarting from SOL 601 to 701 or vice versa, ULFORM needs to be specied such that both analyses use the same formulation. DISPOPT Indicates whether prescribed displacements are applied to the original conguration or the deformed conguration. This option is only applicable for a restart analysis or when a delay (or arrival) time is specied for the prescribed displacement. (Integer; Default = 0) 0 Applied to original conguration 1 Applied to deformed conguration LOADOPT* Indicates whether prescribed loads (pressure and centrifugal) are deformation-dependent, i.e. the direction and magnitude of the load may change due to large deformation of the structure. This option is only applicable for large displacement analysis, i.e. PARAM,LGDISP,1 (Integer; Default = 1) 0 Load is independent of structural deformation 1 Load is affected by structural deformation MAXDISP Species a limit for the maximum displacement that is allowed for any node during the analysis. This feature is generally useful for contact analysis where rigid body motion exists in a model. A value of 0.0 means there is no limit on displacements. (Real 0.0; Default = 0.0) Time Integration TINTEG Selects the time integration method to be used for nonlinear transient analysis. (Integer; Default = 0) 0 Use the Newmark method 1 Use the ADINA composite method ALPHA DELTA Alpha coefcient for the Newmark time integration method. (Real; Default = 0.25) Delta coefcient for the Newmark method. (Real; Default = 0.5) SOL 701 Time Stepping
6-6
Advanced Nonlinear

Name XSTEPo Description Selects time step method used in an explicit time integration analysis. (Integer; Default = 0) 0 - Time step size is calucated by the program based on the critical time step size. The data in the selected TSTEP bulk data entry is used to calculate the total solution time for the analysis. 1 - The number of time steps and the time step size as specied in the selected TSTEP bulk data entry is used. XDTCALo Calculation of the critical time step size may be computationally expensive. This parameter species that the critical time step size be re-calculated every XDTCAL time steps. (Integer > 0, Default = 1) The critical time step size is calculated based on certain assumptions. It is often necessary, especially for nonlinear analysis, to use a time step size smaller than the calculated critical time step size. The factor multiplied by the calculated critical time step size gives the time step size used in the analysis. (4.0 > Real > 0.0, Default = 0.9) Species the factor to scale the mass (densities) of the entire model (at the beginning of the analysis) to increase the critical time step size required for stability when the explicit time integration scheme is used. See warning in Remark 2. (Real = 1.0, Default = 1.0) The minimum time step size used to determine if mass scaling will be applied to elements (at the beginning of the analysis) whose critical time step size is smaller than DTMIN1. The amount of mass scaling is calculated for each element so that the critical time step size is equal to DTMIN1. See Remark 2 and warning in Remark 3. (Real = 0.0, Default = 0.0) The minimum time step size used to determine whether an element will be removed in an explicit time integration analysis. In explicit time integration, the smaller an element size is, the smaller will the critical time step size be. If the critical time step size for an element is smaller than XDTMIN2, the element will be removed in the analysis. See Remark 2 and warning in Remark 3. (Real > 0.0, Default = 0.0) Multigrid Solver ITEMAX EPSIA EPSIB EPSII Maximum number of iterations allowed for the multigrid solver to converge. (Integer > 0; Default = 1000) Convergence tolerance EPSIA. (Real; Default = 1.0E-6) Convergence tolerance EPSIB. (Real; Default = 1.0E-4) Convergence tolerance EPSII. (Real; Default = 1.0E-8) Equilibrium Iteration and Convergence LSEARCH Flag to indicate the use of line searches within the iteration scheme. (Integer; Default = 0) 0 Line search is not used 1 Line search is used LSLOWER LSUPPER MAXITE Lower bound for line search. (0.0 Real < 1.0; Default = 0.001) Upper bound for line search. (1.0 Real; Default = 2.0) Maximum number of iterations within a time step. If the maximum number of iterations is reached without achieving convergence (see CONVCRI parameter), the program will stop unless the automatic time stepping (ATS) or load displacement control scheme is selected (see parameter AUTO). (1 Integer 999; Default = 15)
XDTFACo
XMSCALEo
XDTMIN1o
XDTMIN2o
6-7
Chapter 6
Advanced Nonlinear

Name CONVCRI Description Convergence Criteria. (Integer; Default = 0)0 Convergence based on energy1 Convergence based on energy and force2 Convergence based on energy and displacement3 Convergence based on force4 Convergence based on displacement Relative energy tolerance. (Real; Default = 0.001) Relative force (and moment) tolerance (Real; Default = 0.01) Reference force. (Real) Reference moment. (Real) Relative contact force tolerance. (Real; Default = 0.05) Relative displacement (translation and rotation) tolerance. (Real; Default = 0.01) Reference translation. (Real) Reference rotation. (Real) Line search convergence tolerance. (Real; Default = 0.5) Reference contact force. (Real; Default = 0.01) Line search energy threshold. (Real; Default = 0.0) Automatic Time Stepping (ATS) Scheme ATSSUBD Number that limits the smallest time step size when the automatic time stepping (ATS) scheme is used. For a time step size of DT, the program will stop if convergence is not achieved and the next subdivided time step size is less than DT/ATSSUBD. (Integer 1; Default = 10) Factor that limits the maximum time step size when the automatic time stepping (ATS) scheme is used. The ATS scheme may increase the time step size after convergence is achieved. However, for a time step size of DT, the program will not use a time step size greater than ATSMXDT * DT. (Real; Default = 3.0) Flag controls what time step size to use once convergence is reached after an ATS subdivision. (Integer; Default = 0) 0 Automatically set by program. For contact analysis, ATSNEXT = 2, otherwise ATSNEXT = 1. 1 Use the time step size that gave convergence, i.e., the reduced time step that led to convergence is used again. 2 Return to the original time step size, i.e., the original time step size before any subdivision took place is used. 3 Use a time step size such that the solution time matches the original solution time specied by the user. ATSDFAC Division factor used calculate the sub-increment time step size. If current time step size is DT and convergence is not achieved, the next time step size will be DT/ATSDFAC. (Real > 1.0; Default = 2.0) Flag whether a low-speed dynamics analysis is performed instead of a static analysis. (Integer; Default = 0)0 Low-speed dynamics option is not activated 1 Low-speed dynamics is performed Damping factor used in low-speed dynamics analysis. (Real 0.0; Default = 1.0e-4) Load Displacement Control (LDC) Scheme
ETOL RTOL RNORM RMNORM RCTOL DTOL DNORM DMNORM STOL RCONSM ENLSTH
ATSMXDT
ATSNEXT
ATSLOWS
ATSDAMP
6-8
Advanced Nonlinear

Name LDCGRID LDCDOF Description Grid point id at which a displacement is prescribed for the rst solution step. (Integer > 0) Degree of freedom for prescribed displacement at grid point LDCGRID. (1 Integer 6) 1 X translation 2 Y translation 3 Z translation 4 X rotation 5 Y rotation 6 Z rotation LDCDISP LDCIMAX Prescribed displacement at grid point LDCGRID for the rst solution step. (Real) Displacement convergence factor used to limit the maximum incremental displacement during a solution step. (Real; Default = 3.0) Maximum (absolute magnitude) displacement (for the degree of freedom specied by LDCDOF) at the grid point LDCGRID allowed during the analysis. When the displacement reaches or exceeds LDCDMAX, the program will stop the analysis. See Section 6.2.4 in Advanced Nonlinear Theory and Modeling Guide for other criteria that determines when an LDC solution will stop. (Real) Flag whether the solution is terminated when the rst critical point on the equilibrium path is reached. (Integer; Default = 0) 0 Solution stops 1 Solution continues LDCSUBD Maximum number of arc length subdivisions allowed. (Integer 1; Default = 10) Contact Control IMPACT Impact control scheme (Integer; Default = 0) 0 No special treatment is applied for impact problems 1 Post impact adjustment of velocities and accelerations is applied 2 Modied parameters are used in Newmark time integration scheme NSUPP Number of iterations for pairing contactor node to target segment. If NSUPP > 0, during the rst NSUPP iterations, the pairing target segment is recorded for each contactor node. From iteration NSUPP+1, if a target segment in the recorded list is repeated, it is frozen to be the pairing target segment for the remaining equilibrium iterations in that time step. Specifying NSUPP > 0 may help in the convergence for certain problems. (0 Integer 99; Default = 0) Selects the subdivision scheme used in the implicit rigid-target contact algorithm when the tensile contact force is too large. (Integer; Default = 0) 0 Subdivision is based on the magnitude of the tensile contact force, i.e., the larger the magnitude, the smaller will be the subdivided time step size. 1 Subdivision is based on the global automatic time stepping (ATS) subdivsion settings.
LDCDMAX
LDCCONT
RTSUBD
6-9
Chapter 6
Advanced Nonlinear

Name CSTYPE Description Selects the type of contact segment to use. (Integer; Default = 1) 0 Use linear contact segment 1 Use element-based contact segment which gives better contact traction results CTDISP Selects the default displacement formulation used for contact analysis. A different formulation may be selected for each individual contact set via BCTPARA entry. (Integer; Default = 2) 1 Use small displacement formulation 2 Use large displacement formulation RTALG Selects the rigid-target algorithm to use. (Integer; Default = 0) 0 Use the current algorithm 1 Use the old (v4) algorithm FRICALG Selects the friction algorithm to use. (Integer; Default = 0) 0 Use the current algorithm 1 Use the old (v4) algorithm CTDAMP Indicates whether stabilization damping is applied and how it is applied for contact analysis. This feature is generally useful when rigid body motion exists in a model. (Integer; Default = 0) 0 No stabilization damping is applied 1 Stabilization damping is applied at the rst time step only. The specied damping coefcients are applied and ramped down to zero by the end of the rst time step. 2 The specied stabilization damping coefcients are applied at all time steps. CTDAMPN CTDAMPT Species the normal stabilization damping coefcient. (Real 0.0, Default = 0.0) Species the tangential stabilization damping coefcient. (Real 0.0, Default = 0.0) Restart Options MODEX* Indicates the mode of execution. (Integer; Default = 0) 0 Normal analysis run, i.e. not a restart analysis 1 Restart analysis The restart (.res) le from a previous run must exist to do a restart analysis. The lename and location of the restart le is determined by the dbs keyword. By default, dbs points to the current working directory with the prex of the current job name. Note that keyword scratch=no must be used when running a restart analysis. TSTART* Solution starting time. If MODEX=1, TSTART must equal a solution time in which data was saved in a previous run. If TSTART = 0.0, the last time step in the restart le is used. (Real, Default = 0.0)
6-10
Advanced Nonlinear

Name IRINT* Description Frequency of saving the analysis results in the restart le. (Integer; Default = 0) 0 IRINT is set to 1 when implicit time integration is used and set to the number of steps in the rst time step block when explicit time integration is used. > 0 Restart le is overwritten every IRINT time steps < 0 Restart le is appended every IRINT time steps Other Parameters NSUBGRP* Number of sub-groups to divide large number of elements with same property ID into. Normally, elements with same type and property ID are placed into a group. If a group contains more than 1000 elements and NSUBGRP > 1, the elements are placed into NSUBGRP sub-groups for more efcient processing. (Integer > 0; Default = 1) Indicates the coordinate system used for output of nonlinear 3D element stress/strain results. For linear results, the coordinate system used for output is specied by CORDM in PSOLID entry. (Integer; Default = 0) 0 Results are output in element coordinate system 1 Results are output material coordinate system Translation Options ELCV Convert 8-node to 9-node quadrilateral (plane strain, axisymmetric and shell) elements and 20-node to 27-node brick elements. Note that this also converts 6-node to 7-node triangular (plane strain and axisymmetric) elements and 10-node to 11-node tetrahedral elements. (Integer; Default = 0) 0 No conversion of elements 1 Convert elements as described above; nodal coincidence is not checked against existing nodes and new generated nodes are always created. 2 Convert elements as described above; nodal coincidence is checked against existing nodes and a new node will not be created at a location if a node already exist at that location. EQRBAR Indicates how RBAR elements are handled. (Integer; Default = 0) 0 RBAR is simulated using rigid option in small displacement analysis and using exible option in large displacement analysis. 1 RBAR is simulated using rigid option (i.e. simulated by rigid link or constraint equations as determined by program) 2 RBAR is simulated using exible option (i.e. simulated by spring or beam elements as determined by program) 3 RBAR is simulated by spring elements See Section 2.7 of Advanced Nonlinear Theory and Modeling Guide for details on how RBAR elements are handled.
ELRESCS
6-11
Chapter 6
Advanced Nonlinear

Name EQRBE2 Description Indicates how RBE2 elements are handled. (Integer; Default = 0) 0 RBE2 is simulated using rigid option in small displacement analysis and using exible option in large displacement analysis. 1 RBE2 is simulated using rigid option (i.e. simulated by rigid links or constraint equations as determined by program) 2 RBE2 is simulated using exible option (i.e. simulated by spring or beam elements as determined by program) 3 RBE2 is simulated by spring elements See Section 2.7 of Advanced Nonlinear Theory and Modeling Guide for details on how RBE2 elements are handled. SPRINGK Stiffness of spring elements that simulate RBAR or RBE2 elements. (Real, Default = 0.0) If SPRINGK = 0.0, program automatically sets SPRINGK according to the following calculations. SPRINGK = EMAX * LMODEL where EMAX = maximum Youngs Modulus of materials in the model and LMODEL = largest dimension of the model. If no material is specied in the model, EMAX is set to 1.0E12. BEAME Youngs Modulus of material assigned to beam elements that simulate RBAR or RBE2 elements. (Real, Default = 0.0) If BEAME = 0.0, BEAME is set to EMAX * 100.0 where EMAX = maximum Youngs Modulus of materials in the model. If no material is specied in the model, EMAX is set to 1.0E12. BEAMA Circular cross section area of beam elements that simulate RBAR or RBE2 elements. (Real, Default = 0.0) If BEAMA = 0.0, program automatically sets BEAMA according to the following calculation: BEAMA = (LMODEL * .01)2 where LMODEL = largest dimension of the model RBLCRIT Critical length for determining how RBAR and RBE2 elements are simulated when the rigid or exible option is used to simulate RBAR (see EQRBAR) and RBE2 (see EQRBE2). (Real, Default = 0.0) If RBLCRIT = 0.0, then if EQRBAR (or EQRBE2) = 1, RBLCRIT = LMODEL * 1.0E-6 if EQRBAR (or EQRBE2) = 2, RBLCRIT = LMODEL * 1.0E-3
REMARKS:
1. Parameters applicable to SOL 701 are: XSTEP, XDTCAL, XDTFAC, XMSCALE, XDTMIN1 and XDTMIN2 are only used for SOL 701. These parameters are indicated in the table with a superscript o.
6-12
Advanced Nonlinear
TINT, ICMODE, DTDELAY, SDOFANG, UPFORM, LOADOPT, MODEX, TSTART, IRINT and NSUBGRP parameters can be used for SOL 701. These parameters are indicated in the table with a superscript*.
2. XMSCALE, XDTMIN1 and XDTMIN2 may be used together. XDTMIN1 and DTMIN2 are applied after XMSCALE is applied. If XDTMIN1 and XDTMIN2 are both used, XDTMIN1 should be greater than XDTMIN2. If XDTMIN2 = XDTMIN1 is specied, XDTMIN1 will be ignored. 3. WARNING: Specifying XMSCALE > 1.0, XDTMIN1 > 0.0 or XDTMIN2 > 0.0 may change the model signicantly. Hence, extra care should be exercised in examining the results when any of these parameters are used.
BCTPARA
Contact Set Parameters (SOLs 601 and 701 only)
Denes parameters for a contact set.
FORMAT: 1 BCTPARA 2 CSID Param4 3 Param1 Value4 4 Value1 Param5 5 Param2 Value5 6 Value2 -etc7 Param3 8 Value3 9 10
EXAMPLE: BCTPARA 2 EPSN 1.0E-10 TZPENE 0.1 CSTIFF 1
FIELDS:
Field CSID PARAMi VALUEi
Contents Contact set ID. Parameters dened in this command apply to contact set CSID dened by a BCTSET entry. (Integer > 0) Name of the BCTPARA parameter. Allowable names are given in the parameter listing below. (Character) Value of the parameter. See Table 6-2 for parameter listing. (Real or Integer)
Table 6-2. BCTPARA Parameters for SOL 601 Name Description General Parameters
6-13
Chapter 6
Advanced Nonlinear
Table 6-2. BCTPARA Parameters for SOL 601 Name TYPE Description Selects the type of contact algorithm. (Integer; Default=0) 0 Constraint function algorithm is used. 1 Segment method algorithm is used. 2 Rigid target algorithm is used. NSIDE Flags single or double-sided contact. (Integer; Default=1) 1 Contact surfaces are single-sided. 2 Contact surfaces are double-sided. TBIRTH TDEATH INIPENE Birth time for contact set. (Real; Default=0.0) Death time for contact set. (Real; Default=0.0) If TDEATH TBIRTH, it is ignored. Flags how initial penetrations or gaps are handled. (Integer; Default=0) 0 Initial penetrations are eliminated. 1 Initial penetrations are eliminated and the list of pentrating nodes is printed. 2 Initial penetrations are ignored. In successive steps, each contractor node is allowed to penetrate the target up to its initial penetration. 3 Initial penetrations or gaps are overridden by GAPVAL. This option is not available for rigid target algorithm (TYPE = 2). GAPVAL Species a constant gap distance between the source region (contactor) and the target region when INIPENE = 3. Negative GAPVAL means initial penetrations which will be eliminated. (Real, Default = 0.0) Penetration depth for single-sided contact (i.e. NSIDE=1). (Real; Default=0.0)If PDEPTH > 0.0, then penetration is detected. When penetration PDEPTH, and if penetration > PDEPTH, penetration is deemed not to occur. Indicates whether a continuous (interpolated) contact segment normal is used for the contact surfaces. (Integer; Default=0) 0 SEGNORM=1 if NSIDE=1, SEGNORM=-1 if NSIDE=2. 1 Continuous segment normal is used. -1 Continuous segment normal is not used.
PDEPTH
SEGNORM
6-14
Advanced Nonlinear
Table 6-2. BCTPARA Parameters for SOL 601 Name OFFTYPE Description Type of offset for contact regions. (Integer; Default=0) 0 User specied offset value only for single-sided contact (i.e., NSIDE=1). Use offset of 0.001 for double-sided contact (i.e., NSIDE=2) 1 Use specied offset value for either single- or double-sided contact. 2 Half the shell thickness is used for contact regions on shell elements and no offset is used otherwise. OFFSET Default offset distance for contact regions. (Real; Default=0.0) Note: For contact algorithm TYPE=0 or 1, individual offset distances can be specied for each contact region using the BCRPARA entry to override the default offset distance specied here. Standard contact algorithm (TYPE=0 or 1) DISP Selects the displacement formulation used for this contact set. (Integer; Default = 0) 0 Use the default formulation selected by CTDISP in NXSTRAT entry 1 Use small displacement formulation 2 Use large displacement formulation TZPENE Time to eliminate initial penetrations. (Real 0.0; Default=0.0) If TZPENE=0.0 and INIPENE=0 or 1, then the initial penetrations are eliminated in the rst time step. This may cause convergence difculties for certain problems. By using TZPENE > 0.0, the initial penetrations are eliminated gradually over time TZPENE. Indicates whether consistent contact stiffness is used. (Integer; Default=0) 0 Consistent contact stiffness is not used 1 Consistent contact stiffness is used TIED Indicates whether contact regions in each contact pair are tied together. Currently, tied contact option assumes small rotations of the contact regions. (Integer; Default=0) 0 Not tied 1 Tied
CSTIFF
6-15
Chapter 6
Advanced Nonlinear
Table 6-2. BCTPARA Parameters for SOL 601 Name TIEDTOL Description Tolerance used to determine whether contactor nodes are tied to the target region when TIED=1 is specied. A contactor node is tied to its target region if the distance between them is less than or equal to TIEDTOL. (Real; Default=0.0) Factor for extending contact surfaces beyond their boundaries. The amount of extension is given by this factor multiplied by the length of the contact segments. (1.0E-6 Real 0.1; Default=0.001) Type of friction model (0 Integer 13; Default=0). See Remark 1. 0 Constant coefcient of friction specied for each contact pair, i.e. FRICi in the BCTSET entry. 1 Constant coefcient of friction specied by FPARA1. 2 Friction model 1; uses FPARA1 and FPARA2. 3 Friction model 2; uses FPARA1, FPARA2, and FPARA3. 4 Use different static and dynamic friction coefcients; uses FPARA1, FPARA2, and FPARA3. 5 Friction coefcient varies with sliding velocity; uses FPARA1, FPARA2 and FPARA3. 6 Anisotropic friction model; uses FPARA1, FPARA2, FPARA3, FPARA4, and FPARA5. 7 Friction coefcient varies with consistent contact force; uses FPARA1 and FPARA2. 8 Friction coefcient varies with time; uses FPARA1, FPARA2 and FPARA3. 9 Friction coefcient varies with coordinate values; uses FPARA1, FPARA2, FPARA3, FPARA4, and FPARA5. 12 Modied friction model 1; uses FPARA1 and FPARA2. 13 Modied friction model 2; uses FPARA1, FPARA2, and FPARA3. See Section 4.4 of the Advanced Nonlinear Theory and Modeling Guide for a description of friction models. FPARA1 FPARA2 FPARA3 Friction parameter A1 Friction parameter A2 Friction parameter A3
EXTFAC
FRICMOD
6-16
Advanced Nonlinear
Table 6-2. BCTPARA Parameters for SOL 601 Name FPARA4 FPARA5 FRICDLY Description Friction parameter A4 Friction parameter A5 Indicates whether the application of friction is delayed, i.e., friction is applied on a node one time step after the node comes into contact. Delay of friction application may improve the convergence of the solution. (Integer; Default = 0) 0 No delay 1 Delay Constraint Function Contact Algorithm (Type=0) EPSN EPST Parameter for normal constraint function, w. (Real; Default=1.0E-12) Parameter for frictional constraint function, v. (Real > 0.0; Default=0.001)
CFACTOR1 Compliance factor. (Real; Default=0.0) Rigid Target Algorithm (TYPE=2) NCMOD Normal contact modulus. (Real; Default=1.0E11)
Parameters used with current rigid-target algorithm (RTALG=0 in NXSTRAT) TFORCE SLIDVEL The maximum tensile contact force allowed for a converged solution. (Real 0.0, Default = 0.001) The maximum sliding velocity used in modeling sticking friction. When the velocity is smaller than SLIDVEL, sticking is assumed; when the velocity is larger than SLIDVEL, sliding is assumed. (Real > 0.0, Default = 1.0E-10) Species whether oscillation checking is performed and when it is done. (Integer >= 0; Default = 5) If OCHECK=0, no oscillation checking is performed. Otherwise, oscillation checking is performed after equilibrium iteration OCHECK. Oscillation checking consists of two checks: a) If a contactor node oscillates between two neighboring target segments during the equilibrium iterations, oscillation checking puts the contactor node into contact with the boundary edge between the target segments.
OCHECK
6-17
Chapter 6
Advanced Nonlinear
Table 6-2. BCTPARA Parameters for SOL 601 Name Description b) In analysis with friction, if the sliding velocity of a contactor node oscillates during the equilibrium iterations, oscillation checking puts the contactor node into sticking contact. GAPBIAS Contact is detected when the distance between the target and contactor (accounting for any offsets) is less than GAPBIAS. (Real; Default = 0.0) Selects the implementation of offsets. (Integer; Default = 0) 0 Program chooses the implementation based upon the shape of the target surfaces; if a target surface is at or convex, spheres are used, otherwise, normals are used. 1 A sphere of radius equal to the offset is placed around each contactor node, and contact is detected between the sphere and the target surface. 2 Two surfaces are constructed for each contactor surface: an upper surface and a lower surface. These surfaces are constructed using the offsets and the averaged contactor normals. Contact is then detected between points on the constructed contactor surfaces and target surface.
OFFDET
Parameters used with old (NX Nastran Version 4) rigid-target algorithm (RTALG=1 in NXSTRAT) which may be removed in a future release. PENETOL TCMOD RFORCE Penetration tolerance which gives the maximum penetration allowed into a rigid target surface. (Real; Default=1.0E-8) Tangential contact modulus. (Real; Default=0.0) Minimum tensile contact force required to change the state of a contact node from "node in contact" to "free node", i.e., if the normal tensile force is greater than RFORCE, a "node in contact" becomes a "free node". (Real; Default=0.001) Limit (maximum) for the sum of all contact forces for nodes changing from the state of "node in contact" to "free node". If the absolute value of the sum of the forces is bigger than LFORCE, then the automatic time stepping (ATS) method will be activated to subdivide the current time step into smaller time increments. (Real; Default=1.0)
LFORCE
6-18
Advanced Nonlinear
Table 6-2. BCTPARA Parameters for SOL 601 Name Description
RTPCHECK Species whether penetration is checked (in addition to checking the tensile contact force) against the maximum allowable penetration when the rigid-target algorithm is used. (Integer; Default=0) 0 Penetration is not checked. Note that with this setting there is a possibility that the rigid target surface may penetrate the contactor surface excessively. 1 Penetration is checked and subdivision of the time step occurs if the penetration exceeds (RTPMAX * maximum model length.) 2 Penetration is checked and subdivision of time step occurs if the penetration exceeds RTPMAX. RTPMAX Species the maximum allowable penetration when the rigid target algorithm is used. RTPMAX is either a factor of the model size or an absolute value depending on the RTPCHECK parameter. (Real > 0.0; Default=0.001)
Table 6-3. BCTPARA Parameters for SOL 701 Name Description General Parameters XTYPE Selects the type of contact algorithm. (Integer; Default=0) 0 Kinematic constraint algorithm is used. 1 Penalty algorithm is used. 3 Rigid target algorithm is used. NSIDE Flags single or double-sided contact. (Integer; Default=1) 1 Contact surfaces are single-sided. 2 Contact surfaces are double-sided. TBIRTH TDEATH Birth time for contact set. (Real; Default=0.0) Death time for contact set. (Real; Default=0.0) If TDEATH TBIRTH, it is ignored.
6-19
Chapter 6
Advanced Nonlinear
Table 6-3. BCTPARA Parameters for SOL 701 Name INIPENE Description Flags how initial penetrations are handled. (Integer; Default=0) 0 Initial penetrations are eliminated. 1 Initial penetrations are eliminated and the list of penetrating nodes is printed. 2 Initial penetrations are ignored. In successive steps, each contactor node is allowed to penetrate the target up to its initial penetration. PDEPTH Penetration depth for single-sided contact (i.e. NSIDE=1). (Real; Default=0.0)If PDEPTH > 0.0, then penetration is detected. When penetration PDEPTH, and if penetration > PDEPTH, penetration is deemed not to occur. Type of offset for contact regions. (Integer; Default=0) 0 Use specied offset value only for single-sided contact (i.e., NSIDE=1). Use offset of 0.001 for double-sided contact (i.e., NSIDE=2). 1 Use specied offset value for either single- or double-sided contact. 2 Half the shell thickness is used for contact regions on shell elements and no offset is used otherwise. OFFSET Default offset distance for contact regions. (Real; Default=0.0)Note: For contact algorithm XTYPE=0 or 1, individual offset distances can be specied for each contact region using the BCRPARA entry to override the default offset distance specied here. Contact Algorithm XTYPE=0 or 1 TZPENE Time to eliminate initial penetrations. (Real=0.0; Default=0.0)If TZPENE=0.0 and INIPENE=0 or 1, then the initial penetrations are eliminated in the rst time step. This may cause convergence difculties for certain problems. By using TZPENE > 0.0, the initial penetrations are eliminated gradually over time TZPENE. Factor for extending contact surfaces beyond their boundaries. The amount of extension is given by this factor multiplied by the length of the contact segments. (1.0E-6 Real 0.1; Default=0.001) Penalty Contact Algorithm (XTYPE=1)
OFFTYPE
EXTFAC
6-20
Advanced Nonlinear
Table 6-3. BCTPARA Parameters for SOL 701 Name XKNCRIT Description Selects the criterion for evaluation of normal penalty stiffness. (Integer; Default=0) 0 - The program calculates the normal penalty stiffness. 1 - The user species the normal penalty stiffness (XKN). XKN XKTCRIT Species the normal penalty stiffness to be used when XKNCRIT=1. (Real) Selects the criterion for evaluation of tangential penalty stiffness. (Integer; Default=0) 0 - The program calculates the normal penalty stiffness 1 - The user species the normal penalty stiffness (XKT). XKT XDAMP Species the tangential penalty stiffness to be used when XKTCRIT=1. (Real) Indicates whether damping is used when the penalty explicit contact algorithm is used. (Integer; Default=0) 0 Damping is not used, i.e., the XNDAMP parameter is ignored. 1 Damping is used and XNDAMP is a factor of the critical damping, i.e., the damping coefcient is given by XNDAMP multiplied by the critical damping. This is the recommended choice if damping is used. 2 Damping is included and the damping coefcient is specied directly by XNDAMP. XNDAMP Species the relative or absolute damping coefcient (for normal penalty stiffness) when the penalty explicit contact algorithm is used and XDAMP=1 or 2. (Real=0.0; Default=0.1) Rigid Target Algorithm (XTYPE=3) PENETOL TCMOD
REMARKS:
Penetration tolerance which gives the maximum penetration allowed into a rigid target surface. (Real; Default=1.0E-8) Tangential contact modulus. (Real; Default=0.0)
1. Multiple BCTPARA with the same CSID can be used. The parameters specied in all BCTPARA with the same CSID will be combined and assigned to the contact set. 2. If duplicate parameters are specied, the last parameter value will be used.
6-21
Chapter 6
Advanced Nonlinear
3. Field 9 on line 1 should be blank. Beginning on the continuation lines, elds 2 to 9 can be used for 4 pairs of PARAM/VALUE
6.3
Surface-to-Surface Glue Support
The option to glue element faces together during a 601 solution is available in NX Nastran 5. The glue option connects predened surfaces together and prevents relative motion in all directions. Predened regions of element free faces are used to detect where the glue elements are created. The regions where the solver creates glue elements are dened using the BSURFS bulk entry. A region is a collection of solid element free faces in a section of the model where you expect gluing to occur. The BSURFS entry is dened by its own unique ID and is a list of solid element IDs each followed by 3 grid points dening which face of the 3D element to include in the glue region. Once regions are created, they must be paired. A glue pair is a way to combine two regions, source and target, in which gluing is analyzed during the solution. The BGSET bulk entry is used to dene each glue pair. The GSID eld needs to match the value of n on the BGSET case control entry for the solution to recognize this glue denition. The SIDi and TIDi elds refer to regions created by the BSURFS entry, and are used to dene source and target regions respectively for a pair. The search distance eld (SDIST) and penalty (PEN) factor on the BGSET bulk entry are not used by SOL 601. SOL 601 uses a search distance that is equal to the largest element edge in the source and target regions. The BGSET case control command, the BGSET bulk entry, and the BSURFS bulk entry are provided below.
BGSET (Case Control)

Glue Contact Set Selection
Selects the glue contact set.
FORMAT:
BGSET=n
EXAMPLES:
BGSET=5
DESCRIBERS:
Describer n
Meaning Glue contact set identication of a BGSET Bulk Data entry. (Integer>0)
BGSET (Bulk Entry)

3D Glue Contact Set Denition
Denes glued contact pairs of a 3D set.
6-22
Advanced Nonlinear
FORMAT: 1 BGSET 2 GSID 3 SID1 SID2 -etc4 TID1 TID2 5 SDIST1 SDIST2 6 PEN 7 8 9 10
EXAMPLE: BGSET 4 1 4 FIELDS: 2 3
Field GSID SIDi TIDi SDISTi PEN
Contents Glue set identication number. (Integer > 0). Source region identication number for glue pair i. (Integer > 0) Target region identication number for glue pair i. (Integer > 0) Search distance for glue surfaces (Real); (Default=100) Penalty value used in calculations. The value dened on the rst line will be used for all pairs. (Real);(Default=1.0E5)
REMARKS:
1. The default search distance will glue all overlapping sections of the source and target regions. This value can be used in special cases to control what subregions are to be glued. 2. The default penalty factor will be sufcient for most cases. Increase it if separation of the surfaces is observed. If dened too large, numerical problems may occur.
REMARKS RELATED TO SOL 601:
1. SDISTi and PEN are not used by SOL 601. SOL 601 uses a search distance that is equal to the largest element edge in the source and target region.
BSURFS
3D Contact Region Denition by Solid Elements (SOL 101, 601 and 701 only)
Denes a 3D contact region by the faces of the CHEXA, CPENTA or CTETRA elements.
FORMAT: 1 BSURFS 2 ID EID2 G1 G2 G3 3 4 5 6 EID1 EID3 7 G1 G1 8 G2 G2 9 G3 G3 10
6-23
Chapter 6
Advanced Nonlinear
2 -etc-
10
EXAMPLE: BSURFS 7 15 11 21 16 19 21 17 20 22 21 24 26 10 20 16 15 17 20 16 18 21 20 22 25
FIELDS:
Field ID EIDi G1 - G3
Contents Identication number of a contact region. See Remarks 2 and 3. (Integer > 0) Element identication numbers of solid elements. (Integer > 0) Identication numbers of 3 corner grid points on the face (triangular or quadrilateral) of the solid element. (Integer > 0)
REMARKS:
1. The continuation eld is optional. 2. BSURFS is a collection of one or more element faces on solid elements. BSURFS denes a contact region which may act as a contact source (contactor) or target. 3. The ID must be unique with respect to all other BSURFS, BSURF, and BCPROP entries.
6.4
Element Birth and Death
An option to dene element birth and death times for a specic set of elements is available in NX Nastran 5. The new EBDSET case control command selects an EBDSET bulk data entry to be used in the solution. The EBDSET bulk data entry includes the element IDs along with their birth and death times. Another bulk data entry, EBDADD, can be used to combine EBDSET bulk entries such that several of them can be selected using a single, new set ID. All of these new commands needed for the element birth/death option are included below.
EBDSET
Element Birth/Death Set Selection
Selects the element birth/death set (SOLs 601 and 701)
FORMAT:
EBDSET=n
EXAMPLES:
EBDSET=5
6-24
Advanced Nonlinear
DESCRIBERS:
Describer n
Meaning Element birth/death set identication of an EBDSET or EBDADD Bulk Data entry. (Integer>0)
EBDSET
Element Birth/Death Set Denition (SOLs 601 and 701)
Denes element birth and death times for a set of elements.
FORMAT: 1 EBDSET 2 BDID 3 TBIRTH 4 TDEATH 5 EID1 6 EID2 7 EID3 8 EID4 9 EID5 10
CONTINUATION FORMAT 1: EID6 EID7 EID8 -etc-
CONTINUATION FORMAT 2 (THRU OPTION IS ONLY AVAILABLE ON A CONTINUATION LINE): EID6 THRU EID7 BY INC
EXAMPLE: EBDSET 3 25 0.2 THRU 0.5 33 101 201 300 400 450
FIELDS:
Field BDID
Contents Element birth/death set identication number. See Remark 2. (Integer > 0)
TBIRTHElement birth time. (Real 0.0; Default = 0.0) TDEATH Element death time. (Real > TBIRTH; Default = 1.0E+20) EIDi Element identication numbers. See Remark 3. (Integer > 0)
6-25
Chapter 6
Advanced Nonlinear
REMARKS:
1. The continuation line is optional. 2. The THRU option is only available on a continuation line. 3. BDID may be selected by Case Control command EBDSET. If other element birth/death sets are dened, the EBDADD entry must be used to combine all the EBDSET entries. 4. Element birth/death may be used with all elements, i.e., with all elements dened with CROD, CONROD, CBAR, CBEAM, CQUAD, CQUAD4, CQUAD8, CQUADR, CQUADX, CTRIA3, CTRIA6, CTRIAR, CTRIAX, CHEXA, CPENTA, CTETRA, CELAS1, CELAS2, CDAMP1, CDAMP2, CMASS1, CMASS2, CGAP, or CBUSH1D entry.
EBDADD
Element Birth/Death Set Combination (SOLs 601 and 701)
Denes an element birth/death set as a union of element birth/death sets dened on EBDSET entries.
FORMAT: 1 EBDADD 2 BDID 3 BD1 4 BD2 5 -etc6 7 8 9 10
EXAMPLE: EBDADD 10 1 2 3
FIELDS:
Field
Contents
BDID Element birth/death set identication number. (Integer > 0) BDi Identication numbers of element birth/death sets dened via EBDSET entries. (Integer > 0)
REMARKS:
1. To include several element birth/death sets dened via EBDSET entries in a model, EBDADD must be used to combine the element birth/death sets. BDID in EBDADD is then selected with the Case Control command EBDSET. 2. BDi must be unique and may not be the identication of this or any other EBDADD entry.
6.5
Bolt Preload Capability
When components of an assembly are bolted together, a specied torque translates into an axial bolt preload. Bolts should be properly preloaded in this way before service conditions are
6-26
Advanced Nonlinear
applied to the assembly. When analyzing preloaded bolts, you may be interested in obtaining the stresses due to the preload condition alone, or due to a combination of the bolt preload and service load. You can manually determine the preloaded bolt condition by using equivalent thermal loads, although using this method is approximate and typically requires many solution iterations when multiple bolts exist. NX Nastran 5 offers an automated method to simplify this multisolution process. A bolt analysis can be selected by including the new case control command BOLTLD=n, where "n" selects the new BOLTFOR bulk data entry containing the bolt preload value. The bolts should be modeled as beam elements, and included in the new BOLT bulk entry. A single beam may be used to represent a bolt. The new BOLTLD case control command, and the new BOLTFOR and BOLT bulk entries are provided below. The bolt preload capability is not supported in SOL 701.
New BOLTLD Case Control Command

Bolt Load Set Selection
Selects the BOLTFOR Bulk Data entry for bolt pre-load processing.
FORMAT:
BOLTLD=n
EXAMPLES:
BOLTLD=5
DESCRIBERS:
Describer n
Meaning Selects the BOLTFOR Bulk Data entry for bolt pre-load processing. (Integer>0)
REMARKS:
1. Bolt pre-loading is supported in SOLs 101, 103, 105, 107 through 112 and 601.
New BOLTFOR Bulk Entry

Preload Force on Set of Bolts
Denes preload force on a set of bolts.
FORMAT: 1 BOLTFOR 2 SID B7 B9 3 LOAD THRU B10 4 B1 B8 -etc5 B2 6 B3 7 B4 8 B5 9 B6 10
EXAMPLE: BOLTFOR 4 1 1000.0 4 12 6 THRU 9 21 10 26 32 34
6-27
Chapter 6
Advanced Nonlinear
37
43
51
FIELDS:
Field SID LOAD Bi
Contents Bolt preload force set identication number. (Integer>0) Magnitude of the preload force. (Real) Bolt identication numbers dened by bulk entry BOLT (Integer>0), or using THRU (B7<B8 for THRU option).
REMARKS:
1. Multiple BOLTFOR entries with the same SID can be used and the data will be combined. 2. SID is selected by BOLTLD Case Control command. 3. Entering the same bolt id multiple times for the same SID will produce an error.
BOLT
Bolt denition
Selects the CBEAM or CBAR elements to be included in the bolt pre-load calculation.
FORMAT: 1 BOLT 2 BID EID7 -etc3 ETYPE THRU 4 EID1 EID8 5 EID2 6 EID3 7 etc. 8 9 10
ALTERNATE FORMAT: 1 BOLT EXAMPLE 1 SOLs 101, 103, 105, 107 through 112 Example (single EID denes bolt): BOLT 4 1 11 2 BID 3 ETYPE 4 EID1 5 THRU 6 EID2 7 8 9 10
EXAMPLE 2 SOL 601 Example (all EIDs required to dene bolt): BOLT 4 15 1 16 11 28 8 30 2 33 1 20 14
6-28
Advanced Nonlinear
FIELDS:
Field BID ETYPE EIDi
Contents Bolt identication number. (Integer>0) Element type (Integer=1 is required in this eld until other options become available in the future). Element identication numbers of CBEAM or CBAR elements to include in the bolt pre-load calculation.(Integer>0), or using THRU (EID7<EID8 for THRU option).
REMARKS:
1. One BOLT entry is required to dene each physical bolt. 2. Bolt preload is supported in SOLs 101, 103, 105, 107 through 112 and 601. 3. In a SOL 101, 103, 105, 107 through 112, only one EID is required to dene a bolt even if it was modeled as several elements. Any additional entries will be ignored. 4. In a SOL 105, both the bolt preload and service load will be scaled to determine the buckling load.
1. All CBEAM or CBAR elements representing a physical bolt must be identied in the BOLT entry.
6.6
Shell Thickness Result Recovery
A new shell thickness result output option is available when using advanced nonlinear solutions to solve large strain problems. The new case control command SHELLTHK is needed to request the shell thickness results. Since this capability is only supported for large strain solutions, you must include PARAM,LGSTRN,1 in the input le if thickness results are requested. The new SHELLTHK case control command is included below.
SHELLTHK Case Control Command

Shell Thickness Output Request
Requests the form of shell thickness output (SOLs 601 and 701).
FORMAT:
EXAMPLES:
SHELLTHK(PLOT)=ALL
6-29
Chapter 6
Advanced Nonlinear
DESCRIBERS:
Describer PRINT PLOT ALL NONE

REMARKS:
Meaning The printer will be the output medium. Computes and puts shell thickness results in op2 le only. Shell thickness results will be output for all applicable shell elements. See remark 1. Shell thickness results will not be output.
1. Shell thickness results are output only for large strain analysis, i.e., PARAM,LGSTRN,1. 2. Shell thickness results are output at nodes on elements.
6.7
Iterative Solution Improvement
A new 3D-iterative solution option is now available in NX Nastran 5 for SOL 601 to efciently solve large models containing mainly higher order 3D solid elements (e.g., 10-node CTETRA, 20-node CHEXA, and so on). The new iterative solver is invoked when SOLVER=2 in the NXSTRAT bulk entry. In addition to the higher order 3D solid elements, models can contain the other advanced nonlinear supported element types (for example, shells,rods, beams, rebars, and so on), and contact conditions. The SOL 601 3D-iterative solver option is effective in linear or nonlinear static analysis and in nonlinear dynamic analysis.
6-30
Chapter
7-1
Chapter 7
7.1
New Options to Improve Contact Solution Accuracy
NX Nastran uses pre-dened source and target regions of element free faces to detect contact conditions in the model. More specically, the solver uses the element faces from a source region to project normals, then checks if these normals intersect with other element faces on a target region. A contact element is created if: NX Nastran nds an intersection between element faces, and The distance between the two faces is equal to or less than a distance that you specify.
The number of contact elements created and their distribution determines the accuracy of the contact solution. Accuracy is particularly important when pressure results at the contact interface are required. Two new contact solution parameters, INTORD and REFINE, are now available in NX Nastran 5 on the BCTPARM bulk entry to improve the accuracy of the contact solution. The updated BCTPARM bulk entry is repeated below for convenience. Previously, the number of locations where normals were projected (contact evaluation points) from the source region was dependent on the element type. For example, the linear triangle face would always project a single normal, while the parabolic quad would project 4 normals. In NX Nastran 5, the number of contact evaluation points is dependent on the value assigned to the new INTORD parameter, and on the element face type. The following table summarizes how the INTORD value adjusts the number of contact evaluation points for a particular element face: Face Type Linear Triangle Parabolic Triangle Linear Quad Parabolic Quad INTORD=1 1 3 1 4 Number of Contact Evaluation Points INTORD=2 (default) INTORD=3 3 7 7 12 4 9 9 16
INTORD allows for control over the number of contact elements created over a given element face. Use of INTORD=2 (default) is recommended for increased accuracy with minimal increase in computational cost, where accuracy is dened as computing smoother stress and contact pressure results. This recommendation applies for at and curved contact surfaces as well as load path determination and detailed stress determination analysis scenarios. When the INTORD=1 is used, the number of contact evaluation points is the same as in previous releases. The new REFINE parameter increases the number of contact evaluation points by rening the mesh on the source region. Part of the renement process is to project element edges and grids from the associated target region back to the source region. The resulting renement on the source region is then more consistent with the target side, which then gives a better distribution of contact elements. The rened grids and elements are only used during the solution. The contact results are transferred back to the original mesh for post processing results. The use of the contact region renement option (REFINE=1) is primarily recommended for analysis where detailed stress results are required since an increase in the computational cost may occur. Smoother results are obtained for at contact regions versus curved. The renement process provides the added benet of bringing the target region accuracy closer to the accuracy level of the source region. In general, the accuracy level of the source region results tends to be greater than that of the target region.
7-2
By default, REFINE=1 and mesh renement occurs. REFINE=0 will turn off the renement capability. Since both INTORD and REFINE increase the number of contact elements, an increase in solution times may result.
Updated BCTPARM Bulk Entry
BCTPARM
Contact Parameters (SOLs 101, 103, 111 and 112).
Control parameters for the SOL 101 contact algorithm.
FORMAT: 1 BCTPARM 2 CSID Param4 3 Param1 Value4 4 Value1 Param5 5 Param2 Value5 6 Value2 -etc7 Param3 8 Value3 9 10
EXAMPLE: BCTPARM 1 PENN 1.0 CTOL 0.001
FIELDS:
Field CSID PARAMi VALUEi
Contents Contact set ID. Parameters dened in this command apply to contact set CSID dened by a BCTSET entry. (Integer > 0) Name of the BCTPARM parameter. Allowable names are given in the parameter listing below. (Character) Value of the parameter. See Table 7-1 for the parameter listing. (Real or Integer)
Table 7-1. BCTPARM Parameters Name PENN PENT CTOL MAXF MAXS NCHG Description Penalty factor for normal direction. (Default=10.0) See Remark 2. Penalty factor for tangential direction. (Default=1.0) See Remark 2. Contact force convergence tolerance. (Default=0.01) Maximum number of iterations for force a loop. (Default=10) Maximum number of iterations for a status loop. (Default=20) Allowable number of contact changes for convergence. (Default=0)
7-3
Chapter 7
Table 7-1. BCTPARM Parameters Name MPER SHLTHK Description Minimum contact set percentage. (Default=100) Shell thickness offset ag. 0 - Includes half shell thickness as surface offset. (Default) 1 - Does not include thickness offset. RESET Flag to indicate if the contact status for a specic subcase is to start from the nal status of the previous subcase 0 - Starts from previous subcase. (Default) 1 - Starts from initial state. AVGSTS Determines the averaging method for contact pressure/traction results. 0 - The averaging of Pressure/Traction values for a contact grid will include the results from ALL contact elements attached to the grid regardless of whether they are active or inactive in the contact problem (Default). 1 - The averaging of the Pressure/Traction values for a contact grid will exclude those contact elements which are not active in the contact solution and thus have a zero Pressure/Traction value. INIPENE Use when the goal is for a pair of contact regions to be initially touching without interference, but due to the faceted nature of nite elements around curved geometry, some of the element faces may have a slight gap or penetration. 0 or 1 - Contact is evaluated exactly as geometry is modeled. No corrections will occur for gaps or penetrations (Default). 2 - Penetrations will be reset to a new initial condition in which there is no interference. 3 - Gaps and penetrations are both reset to a new initial condition in which there is no interference. Remarks: See Understanding the Contact Control Parameters Used in SOL 101 - BCTPARM in the NX Nastran Users Guide for more information on the BCTPARM options.
7-4
Table 7-1. BCTPARM Parameters Name REFINE Description Determines if the mesh on the source region is rened during the contact solution. 0 - Do not rene the contact source region based on target surface denition. 1 - Rene the contact source region based on target surface denition (default). INTORD Determines the number of contact evaluation points for a single element face on the source region. The number of contact evaluation points is dependent on the value of INTORD, and on the type of element face. See the table in Remark 1 for specic values. 1 The reduced number of contact evaluation points is used. 2 - Use an increased number of contact evaluation points (default). 3 Use a high number of contact evaluation points.
REMARKS:
1. A higher number of contact evaluation points can be used to increase the accuracy of a contact solution. Inaccuracies sometimes appear in the form of nonuniform contact pressure and stress results. There may be a penalty associated with using more evaluation points since the time for a contact problem to converge may be longer. The table below shows how the number of contact evaluation points is dependent on the element type, and how it can be adjusted using the INTORD option. The Face Type column applies to shell elements, and to the solid element with the associated face type. Face Type Linear Triangle Parabolic Triangle Linear Quad Parabolic Quad Number INTORD=1 1 3 1 4 of Contact Evaluation Points INTORD=2 INTORD=3 3 7 7 12 4 9 9 16
2. Penalty factors have units of 1/(length). A physical interpretation is that it is equivalent to the axial stiffness of a rod (K=e*E*dA) with area dA, modulus E, and length 1/e. The defaults for the penalty factors generally work well, but in the event that meshes have very large or very small edge lengths, adjustments may be necessary. See the chapter Surface Contact for SOL 101 in the NX Nastran Users Guide for tips on adjusting penalty factors.
7-5
Chapter 7
7.2
Support of Contact in SOLs 103, 111, and 112
In NX Nastran 5, a contact denition can be included in a normal mode solution (SOL 103), and in an optional dynamic response calculation (SOLs 111 and 112). In the normal mode solution, NX Nastran adds the contact stiffness result from the end of the converged linear statics contact solution. The contact stiffness values in the normal mode solution represents the nal contact condition of the structure around the contact interface. Thus, it will appear that the resulting contact surfaces are attached during the normal mode analysis. Since the calculated normal modes include the nal contact interface conditions, the response calculation (SOLs 111 and 112) which uses these normal modes automatically includes the same conditions. The inputs for the normal mode solution are consistent with differential stiffness solutions which require a linear statics subcase. The difference is that the linear statics subcase should include the BCSET case control command. When dening the normal modes subcase, a STATSUB bulk entry must be included to reference the subcase id containing the contact denition. The contact solution in the linear statics subcase must fully converge before moving to the normal mode portion of the run.
7.3
Superelements with Contact
In NX Nastran 5, the use of superelements is now permitted in solution sequences which support contact (SOLs 101, 103, 111 and 112). The only requirement is that the contact denition must occur in the residual structure.
7.4
Shell Element Z-Offset with Contact
Shell elements such as the CQUAD4 can be offset relative to the mean plane of their connected grid points using the ZOFFS option. In NX Nastran 5, the linear contact solution can include the shell element ZOFFS when evaluating the contact surfaces. The value assigned to the new ZOFFSET option on the BCTPARM bulk entry determines if the contact solution recognizes the ZOFFS value. By default, ZOFFSET=0 and the ZOFFS value is used when evaluating the contact surfaces. When ZOFFSET=1, the ZOFFS value is not used when evaluating the contact surfaces.
7-6
Chapter
8 Strength Ratio Output for Composites
8-1
Chapter 8
8.1
Laminate Strength Ratio Output
Strength ratio is now output together with the failure index when using the PCOMP bulk entry. Strength ratio is a more direct indicator of failure than failure index since it demonstrates the percentage of applied load to the failure criteria. Strength ratio is dened as: Strength Ratio (SR) = Allowable Stress / Calculated Stress For example, a SR = .75 not only indicates that a failure has occurred, but also indicates that the applied load is 25% beyond the allowable. A FI = 1.25 on the other hand does not represent a percentage of failure; only that a failure condition exists. NX Nastran provides strength ratio output for the same four commonly used denitions of the failure surface: Hills Theory, Hoffmans Theory, Tsai-Wu Theory, and maximum strain theory. Since the strength ratio calculation is based on the equations for failure index, both of their details are presented here. Failure indices assume a value of one on the periphery of a failure surface in stress space. If the failure index is < 1, the lamina stress is interior to the periphery of the failure surface and the lamina is assumed to be safe. If the failure index is > 1, the lamina stress is exterior to the periphery of the failure surface and the lamina is assumed to have failed.
These failure indices represent a phenomenological failure criterion in that only an occurrence of a failure is indicated and not the mode of failure. In the present context, concern is with the analytical denition of a failure surface in stress or strain space for use with laminae under biaxial loading.
Selecting a Failure Criterion or Strength Ratio

If requested, NX Nastran will calculate a failure index and a strength ratio for each ply. This failure index is obtained by considering the failure criteria for unidirectional ber composites as in the commonly used failure theories. You can select one of the following failure criteria for composites: Hills Theory Hoffmans Theory Tsai-Wu Theory Maximum Strain
Note: If you specify a failure theory (using the FT option in eld 6 of the PCOMP entry), you must also specify the allowable shear stress of the bonding material (using the SB option in eld 5). In the analysis of isotropic materials, strength is independent of the orientation of the body under load and one may compare the largest computed principal stress with an allowable stress to establish the integrity of the structure. Laminated composites, on the other hand, are orthotropic materials and may exhibit unequal properties in tension and compression. Thus, the strength of these orthotropic laminae is a function of body orientation relative to the imposed stresses.
8-2
As the evaluation of the matrices of material moduli for laminated composites provides sufcient information to determine the actual stress eld sustained by the material, the determination of structural integrity will depend on the denition of an allowable stress eld. The basic ingredient of this denition is the establishment of a set of allowable stresses or strengths in the principal material directions. Xt Xc Yt Yc S = = = = = Allowable tensile stress in the principal x (or 1) direction of the material. Allowable compressive stress in the principal x (or 1) direction of the material. Allowable tensile stress in the principal y (or 2) direction of the material. Allowable compressive stress in the principal y (or 2) direction of the material. Allowable shear stress in the principal material system.
The strength ratio (SR) is calculated for each of the four failure theories by solving the quadratic equation with FI =1, and replacing the applied stress with (SR applied stress). The SR calculation detail is discussed below for each theory. Hills Theory (HILL) Hills failure theory for orthotropic materials that have the same strength in tension and compression, i.e., xt = xc and yt = yc can be expressed as:
X Y S
is allowable stress in 1-direction is allowable stress in 2-direction is allowable stress in shear
and X = Xt if 1 is positive or X = Xc if 1 is negative and similarly for Y and 2. For the interaction term 12/X2, X = Xt if 12 is positive or X = Xc if 12 is negative. A plot of the above equation obtained by setting the failure index to 1 on the 1-2 plane yields an ellipse and is the anisotropic yield criterion of Hill (modied later by Tsai, and hence also sometimes known as the Tsai-Hill theory). Therefore, if the failure index so calculated is less than 1, the ply stresses are inside the yield ellipse and the ply is said to be safe; conversely, if the failure index is greater than 1, the ply stresses are outside the yield ellipse and the ply has failed. Replacing the applied stress with (SR applied stress), the Hill Failure Criteria can be rewritten in terms of a strength ratio:
The above equation can be rearranged into quadratic equation format:
giving:
8-3
Chapter 8
The value of a in the quadratic equation format matches the value for the Hill failure index (FI). Substituting the values of a = FI, b = 0, and c = -1 into the general quadratic equation solution:
gives the Hill strength ratio:
Hoffmans Failure Theory (HOFF) Hoffmans theory for an orthotropic lamina in a general state of plane stress with unequal tensile and compressive strengths is given by
The failure index is obtained by evaluating the left-hand side of the above equation. Note that this theory takes into account the difference in tensile and compressive allowable stresses by using linear terms in the equation. To calculate the strength ratio, the following terms are dened: F1 F2 F11 F22 F66 = = = = =
Substituting above terms into Hoffman FI equation and setting FI = 1:
Replacing the applied stress with (SR applied stress), the Hoffman Failure Criteria can be rewritten in terms of a strength ratio:
8-4
Rearranging into quadratic equation format:
Using the general quadratic equation solution:
The Hoffman strength ratio is calculated using the following roots:
Tensor Polynomial Theory of Tsai-Wu (TSAI) The theory of strength for anisotropic materials proposed by Tsai and Wu specialized to the case of an orthotropic lamina in a general state of plane stress with unequal tensile and compressive strengths is
where: F1 F2 F11 F22 F66 = = = = =
and F12 is to be evaluated experimentally. The magnitude of F12 is, however, constrained by the following inequality called a stability criterion:
The necessity of satisfying the stability criterion, together with the requirement that F12 be determined experimentally from a combined stress state, poses difculties in the use of this theory. Narayanaswami and Adelman (Narayanaswami, R., and H. M. Adelman, Evaluation
8-5
Chapter 8
of the Tensor Polynomial and Hoffman Strength Theories for Composite Materials, Journal of Composite Materials, Vol. II, 1977, p. 366.) have suggested that F12 be set to zero and that use of Hoffmans theory or Tensor Polynomial theory with F12 = 0 is a preferred alternative to the experimental determination of F12. If you have a value for use with F12 in the theory, you can input that value in the MAT8 Bulk Data entry; otherwise, with this failure theory, NX Nastran sets F12 to 0.0. The left-hand side of the above equation will be evaluated as the failure index by this theory. NX Nastran calculates the failure index of bonding material as the maximum interlaminar shear stress divided by the allowable bonding stress. NX Nastran writes the failure indices for all the plies into the OEFIT (Output Element failure index Table) output if stresses are requested. The failure index for the element is the largest value of the failure indices for all plies of the element. Replacing the applied stress with (SR applied stress), the Hoffman Failure Criteria can be rewritten in terms of a strength ratio:
Rearranging into quadratic equation format:
Using the general quadratic equation solution:
The TSAI-Wu strength ratio is calculated using the following roots:
Maximum Strain Theory (STRN) The midplane strains and curvatures are available in the element coordinate system. From these, the stresses and strains in each individual lamina along the ber direction and transverse direction can be easily calculated. You can use the STRAIN case control command to request the output of lamina strains. The maximum strain criteria has no strain interaction terms. The strain allowables specied on the MAT8 entry for each lamina include Xt,Xc Yt,Yc S Allowable strains in tension and compression, respectively, in the longitudinal direction. Allowable strains in tension and compression, respectively, in the transverse direction. Allowable strain for inplane shear.
8-6
The failure index is calculated using
and
i.e., strains minus thermal strains).
where
1, 2,12
are the elastic strains (total
Because you must understand which mode of failure index is critical; i.e., longitudinal (1), transverse (2), or shear (12), NX Nastran prints the mnemonic 1, 2 or 12 alongside the FP value to indicate the critical direction. Note: There is no change in the way the failure index is calculated for interlaminar shear stresses. When you use the maximum strain theory, you may want to specify lamina stress allowables instead of strain allowables on the MAT8. To do this, leave the STRN eld on the MAT8 entry blank. For this case, NX Nastran calculates the failure indices using
and
that is,
To calculate the Maximum Stress (Strain) strength ratio, the failure index which is dened as FI = Calculated Stress / Allowable Stress is set to unity, and the applied stress is replaced with (SR applied stress). The result is:
8-7
Chapter 8
Inputs for Strength Ratio Output

Since strength ratio is output together with failure index, the output requirements are the same. The only exception is that the new SRCOMPS parameter has been created to turn on the strength ratio output. By default, the parameter is set to NO and strength ratio is not output. SRCOMPS Default = NO SRCOMPS controls the computation and printout of ply strength ratios. If SRCOMPS = YES, ply strength ratios are output for composite elements that have failure indices requested. See the PCOMP bulk entry in the QRG for information on the failure index and strength ratio input requirements.
8-8
Chapter
9-1
Chapter 9
9.1
Thermal Expansion of Rigid Elements
The option to include rigid elements in thermal expansion calculations is available in NX Nastran 5. The new case control command, RIGID, should be assigned to LAGRAN to use the new method. In addition, the coefcient of thermal expansion (ALPHA) has been added as an additional entry on the rigid elements: RBAR, RBE1, RBE2, RROD, RTRPPLT. This capability is supported in solutions 101 through 112. When the RIGID case control command is set to the default of LINEAR, the original linear elimination method is used. This method treats rigid elements as an MPC equation without thermal loading effects. The new RIGID case control command is duplicated below for convenience.
New RIGID Case Control Command
RIGID
Rigid Element Method
Selects the rigid element processing method for RBAR, RBE1, RBE2, RROD and RTRPLT elements.
FORMAT:
EXAMPLES:
RIGID=LAGRAN
DESCRIBERS:
Describer LINEAR LAGRAN

REMARKS:
Meaning Selects the linear elimination method. Selects the Lagrange multiplier method.
1. The RIGID command must be above the SUBCASE level. 2. The LAGRAN method allows for the thermal expansion of the rigid elements. 3. The RIGID command can only be used in SOLs 101 through 112. For all other solution sequences, RIGID command is ignored and RIGID=LINEAR is used. 4. If the RIGID command is not specied, RIGID=LINEAR is used. 5. LINEAR processing will not compute the thermal loads. Also, in SOLs 103 through 112, LAGRAN method must be used to compute the differential stiffness due to the thermal expansion of the rigid elements. 6. When using RIGID=LAGRAN, K6ROT must be dened as non-zero.
9-2
9.2
Bolt Preload Analysis
When components of an assembly are bolted together, a specied torque translates into an axial bolt preload. Bolts should be properly preloaded in this way before service conditions are applied to the assembly. When analyzing preloaded bolts, you may be interested in obtaining the stresses due to the preload condition alone, or due to a combination of the bolt preload and service load. You can manually determine the preloaded bolt condition by using equivalent thermal loads, although using this method is approximate and typically requires many solution iterations when multiple bolts exist. NX Nastran 5 offers an automated method to simplify this multisolution process. A bolt analysis can be selected by including the new case control command BOLTLD=n, where "n" selects the new BOLTFOR bulk data entry containing the bolt preload value. The bolts should be modeled as beam elements, and included in the new BOLT bulk entry. One BOLT entry is required to dene each physical bolt. The service loads can also be included and are selected using the LOAD case control command. Superelements are permitted with preloaded bolts but the elements which dene the bolts must be in the residual structure. NX Nastran uses the following two solution process to automate the preloading of bolts: The beam elements which represent the bolts are reduced in stiffness by the value of the parameter BOLTFACT which makes their stiffness insignicant. The bolt preloads are applied at the ends of the bolts in the axial direction. A linear statics solution runs to get the relative displacements, U2 and U1, for each pair of grids. The bolt strains are calculated as: Bolt Strain = (U2-U1)/L P/AE where U1 and U2 are the deections at the ends of the bolts, P is the preload, and A is the area of the bolt. In the nal solution step, the bolts are treated as they were modeled (beam elements), then the calculated bolt strains are applied plus any service loads (if dened). The new BOLTLD case control command, and the new BOLTFOR and BOLT bulk entries are provided below for convenience.
New BOLTLD Case Control Command
BOLTLD
Bolt Load Set Selection
Selects the BOLTFOR Bulk Data entry for bolt preload processing.
FORMAT:
BOLTLD=n
EXAMPLES:
BOLTLD=5
9-3
Chapter 9
DESCRIBERS:
Describer n
Meaning Selects the BOLTFOR Bulk Data entry for bolt preload processing. (Integer>0)
REMARKS:
1. Bolt preloading is supported in SOLs 101, 103, 105, 107 through 112 and 601.
New BOLTFOR Bulk Entry
BOLTFOR
Preload Force on Set of Bolts
Denes preload force on a set of bolts.
FORMAT: 1 BOLTFOR 2 SID B7 B9 3 LOAD THRU B10 4 B1 B8 -etc5 B2 6 B3 7 B4 8 B5 9 B6 10
EXAMPLE: BOLTFOR 4 1 37 1000.0 4 43 12 6 51 THRU 9 21 10 26 32 34
FIELDS:
Field SID LOAD Bi
Contents Bolt preload force set identication number. (Integer>0) Magnitude of the preload force. (Real) Bolt identication numbers dened by bulk entry BOLT (Integer>0), or using THRU (B7<B8 for THRU option).
REMARKS:
1. Multiple BOLTFOR entries with the same SID can be used and the data will be combined. 2. SID is selected by BOLTLD Case Control command.
9-4
New BOLT Bulk Entry
BOLT
Bolt denition
Selects the CBEAM or CBAR elements to be included in the bolt pre-load calculation.
FORMAT: 1 BOLT 2 BID EID7 -etc3 ETYPE THRU 4 EID1 EID8 5 EID2 6 EID3 7 etc. 8 9 10
ALTERNATE FORMAT: 1 BOLT 2 BID 3 ETYPE 4 EID1 5 THRU 6 EID2 7 8 9 10
EXAMPLE 1 SOLs 101, 103, 105, 107 through 112 Example (single EID denes bolt): BOLT 4 1 11
EXAMPLE 2 SOL 601 Example (all EIDs required to dene bolt): BOLT 4 15 1 16 11 28 8 30 2 33 1 20 14
FIELDS:
Field BID ETYPE EIDi
Contents Bolt identication number. (Integer>0) Element type (Integer=1 is required in this eld until other options become available in the future). Element identication numbers of CBEAM or CBAR elements to include in the bolt pre-load calculation.(Integer>0), or using THRU (EID7<EID8 for THRU option).
REMARKS:
1. One BOLT entry is required to dene each physical bolt. 2. Bolt preload is supported in SOLs 101, 103, 105, 107 through 112 and 601.
9-5
Chapter 9
3. In a SOL 101, 103, 105, 107 through 112, only one EID is required to dene a bolt even if it was modeled as several elements. Any additional entries will be ignored. 4. In a SOL 105, both the bolt preload and service load will be scaled to determine the buckling load.
1. All CBEAM or CBAR elements representing a physical bolt must be identied in the BOLT entry.
9.3
Improvements to Surface Glue
Element Iterative Solver Support of Surface Glue

In NX Nastran 4.1, a new option to glue element faces together during a solution became available. Glue denitions, which are supported in all solution sequences except for SOL 701, create stiff springs to connect the pre-dened surfaces. Glue elements connect components together in such a way as to prevent all relative motion. Glue denitions in a SOL 153 heat transfer analysis are treated as near innite conductivity connections. The surface-to-surface glue capability was supported by the global iterative solver in NX Nastran 4.1, but not by the element iterative solver. Now in NX Nastran 5, the surface-to-surface glue denitions can be included when using the element iterative solver. To utilize the element based iterative solver, add the keyword ELEMITER=YES to the Nastran card, so that you have: NASTRAN ITER=YES, ELEMITER=YES $ When using glue denitions with the element iterative solver in NX Nastran 5, it is recommended you use the primal preconditioner which was introduced in NX Nastran 4. It can signicantly reduce the number of iterations required, thus decreasing the solution time. To turn this option on, enter PRIMAL in the PRECOND eld on the ITER bulk data card.
Glue Accuracy Improvement

NX Nastran uses pre-dened source and target regions of element free faces to detect glue conditions in the model. More specically, the solver uses the element faces from a source region to project normals, then checks if these normals intersect with other element faces on a target region. A glue element is created if: NX Nastran nds an intersection between element faces, and The distance between the two faces is equal to or less than a distance that you specify.
The number of glue elements created and their distribution determine the accuracy of the glued interface. Two new glue solution parameters, INTORD and REFINE, are now available in NX Nastran 5 on the BGPARM bulk entry to improve the accuracy of the glue solution. The updated BGPARM bulk entry is repeated below for convenience.
9-6
Previously, the number of locations where normals were projected (glue points) from the source region was dependent on the element type. For example, the linear triangle face would always project a single normal, while the parabolic quad would project 4 normals. Now in NX Nastran 5, the number of glue points used is dependent on the value assigned to the new INTORD parameter, and on the element face type. By default, INTORD=2 and the number of glue points increases as compared to the previous release. When INTORD=1, the number of glue points is the same as the previous release. The following table summarizes how the INTORD value adjusts the number of glue points for a particular element face: Face Type Linear Triangle Parabolic Triangle Linear Quad Parabolic Quad Number of Glue Points Used in Glue Element Evaluation INTORD=2 (default) INTORD=1 INTORD=3 1 3 7 3 7 12 1 4 9 4 9 16
The new REFINE parameter increases the number of glue points by rening the mesh on the source region. Part of the renement process is to project element edges and grids from the associated target region back to the source region. The resulting renement on the source region is then more consistent with the target side, which then gives a better distribution of glue elements. The rened grids and elements are only used during the solution. The glue results are transferred back to the original mesh for post processing results. By default, REFINE=1 and mesh renement occurs. REFINE=0 turns off the renement capability.
9-7
Chapter 9
New BGPARM Bulk Entry
BGPARM
Glue Parameters.
CONTROL PARAMETER FOR THE GLUE CONTACT ALGORITHM. FORMAT: 1 BGPARM 2 GSID 3 Param1 4 Value1 5 6 7 8 9 10
EXAMPLE: BGPARM 4 INTORD 2
FIELDS:
Field GSID PARAMi VALUEi
Contents Glue set ID. Parameters dened in this command apply to glue set GSID dened by a BGSET entry. (Integer > 0) Name of the BGPARM parameter. Allowable names are given in the parameter listing below. (Character) Value of the parameter. See below for the parameter listing. (Real or Integer)
Table 9-1. BGPARM Parameters Name ZOFFSET Description Determines if the shell element z-offset is included in the glue solution. 0 - Includes the shell z-offset when determining the glue surfaces (Default) 1 - Does not include the shell z-offset when determining the glue surfaces (as in NXN4.1 and prior) INTORD Determines the number of glue points for a single element face on the source region. 1 - Low order 2 - Medium order (default) 3 - High order
9-8
Table 9-1. BGPARM Parameters Name REFINE Description Determines if the mesh on the source region is rened during the glue solution. 0 - Do not rene the glue source region based on target surface denition. 1 - Rene the glue source region based on target surface denition (default).
Shell Element Z-offset with Surface Glue

Shell elements such as the CQUAD4 can be offset relative to the mean plane of their connected grid points using the ZOFFS option. In NX Nastran 5, the surface glue solution can include the shell element ZOFFS when evaluating the glue surfaces. The value assigned to the new ZOFFSET option on the BGPARM bulk entry determines if the glue solution recognizes the ZOFFS value. By default, ZOFFSET=0 and the ZOFFS value is used when evaluating the glue surfaces. When ZOFFSET=1, the ZOFFS value is not used when evaluating the glue surfaces.
9.4
Parameter Specication Improvements
NX Nastran 5 allows the specication of PARAM statements as follows: 1. Parameter values can be specied in the nast5rc (UNIX) and nast5.rcf (WINDOWS). 2. Parameters can be assigned a user dened keyword. The new keyword can then be used to specify a value for the parameter on the command line or in the nastran resource le. The keywords can be dened in the nastran.params le in the architecture directory.
Specifying Parameters in the Resource (rc) le

Parameters can be specied using the following syntax in the rc le: PARAM,name,value For example: PARAM,POST,-2 PARAM,AUTOSPC,NO
Using the nastran.params le

The parameter dening keywords are assigned in the nastran.params le located at: UNIX: install_dir/nxn5/arch_dir/nastran.params Windows: install_dir\nxn5\i386\nastran.params You can override this default le using the 0.params keyword on the command line. For example, if you have a custom dened parameter le named my_specs.param, the command line submittal would be:
9-9
Chapter 9
nastran input_le.dat 0.params=full_path_name/my_specs.param where full_path_name is the fully qualied path name of the le my_specs.param. The parameters can be assigned to keywords in the nastran.params le as follows: keyword name : param name : param value where Keyword name the name to be assigned as a mnemonic for the parameter name. This should be a single word without embedded white space. Also, this name should not be the same as any of the internal keywords. The keyword name can be the same as the parameter name. Parameter Name the actual name of the parameter, for example, autospc. Parameter Value used for syntax checking when the keyword name is used in the command line. The parameter value is either a number (integer/real) or an acceptable value list. For example, {yes,no}.
Comments can be included in the nastran.params le by starting the line with #, $ or ; as the rst character. Below is an example nastran.params le: The punch le algorithm has been corrected in NX Nastran 5 to interpret data values correctly, thus data entries which are character, real or integer will each be written to the punch output as such. The new system cell 210 has been created to control the punch formatting. By default, system cell 210 is set to 1 which correctly interprets data values. To revert back to the old, incorrect method of punch output, the system cell can be set to 0.
# # example nastran.params file mypost : POST : number nxautospc: AUTOSPC : {yes, no} alpha1: alpha1 : number,number
In the nastran submittal, these keywords can be used to assign values to the parameters:
nastran input_file.dat mypost=-2 nxautospc=no alpha1=0.1,0.2
The order of processing the parameter values follows the same precedence rules as in the rc le. That is, parameters in the input le have the precedence over the command line specication which has precedence over the rc le specication.
9.5
Punch Output Corrections
In previous releases of NX Nastran, the algorithm which writes punch les assumed that the rst item in an entry was either an integer or real, and that all consecutive items were real. These were sometimes invalid assumptions since there are occasions in which the rst item in a data entry is a character string, and there are also occasions in which some consecutive items are integers and not real. This has been corrected in NX Nastran 5. The example below demonstrates how an incorrectly written punch might look, and how it is corrected in NX Nastran 5:
9-10
The characters highlighted in red were incorrectly written using the old punch algorithm. The characters in green demonstrate how the new algorithm interprets and writes the correct data type. The system cell 210 is available to revert back to the old punch algorithm. System cell 210 settings are summarized as follows: System Cell 210 Value 0 1 Result OLD Punch algorithm used NEW Punch algorithm used (default)
9-11
Chapter 9
9.6
PARAM K6ROT Update
The formulation of the CQUAD4 and CTRIA3 elements are based on the Mindlin-Reissner shell theory. These elements do not provide direct stiffness for the rotational degrees-of-freedom which are normal to the surface of the element. This zero stiffness results in a singular stiffness matrix, preventing NX Nastran from solving. To avoid this problem, you can: Constrain the rotational degree-of-freedom either manually with an SPC entry or automatically with the AUTOSPC parameter. Apply an articial stiffness term to the degrees of freedom using PARAM K6ROT.
Assigning PARAM K6ROT to 0 may overly constraint, particularly when curvature exists, or when shell and solid elements are connected to common grids. To avoid overly constraining shell rotations, it is better to assign a penalty stiffness value to PARAM K6ROT versus the xed condition when 0 is used. Assigning PARAM K6ROT to 100 has shown to be a good, general penalty stiffness value which will not adversely affect results, yet will properly constrain the rotation degrees-of-freedom. Therefore, the default for parameter K6ROT has been modied from 0 to 100 in NX Nastran 5. In addition to the default update, K6ROT will now be assigned a value of 0 when a model is made entirely of membrane elements. The recommendation to not use K6ROT with membrane elements has always existed. This change simply makes the recommendation a requirement. K6ROT Default = 100.0 K6ROT species the stiffness to be added to the normal rotation for CQUAD4 and CTRIA3 elements. This is an alternate method to suppress the grid point singularities, and is intended primarily for geometric nonlinear analysis. A value between 1.0 and 100.0 is recommended to suppress singularities. A large value may be required in nonlinear and eigenvalue analyses. This parameter is ignored for CQUADR, CTRIAR, CQUAD8, and CTRIA6 elements. K6ROT is forced to 0 when only membrane elements exist.
9.7
AUTHQUEUE Update
When an NX Nastran job fails because of a failed license request, an option to have the job retried automatically became available in NX Nastran 4.1. NX Nastran will retry a failed license request job every minute up to the value of the AUTHQUEUE keyword. The AUTHQUEUE keyword default was 20 minutes in NX Nastran 4.1. Based on customer feedback, the default has been modied to 0 in NX Nastran 5. You must now explicitly dene the AUTHQUEUE keyword in the nast5rc/nast5.rcf les to take advantage of this capability. authqueue authqueue=number Default: 0
9.8
Documentation Improvements
The following documentation improvements have occurred in NX Nastran 5: The NX Nastran documentation is now available in both PDF and HTML format. A single bookshelf.html provides the links to the PDF and the HTML documentation.
9-12
HTML search is available. You can search a single, or all HTML documents by selecting located at the top left corner of the bookshelf page. The web browser must have a Java plugin installed, and the Java plugin option must be turned on in the browser by checking the Java plugin box under: Tools->Internet Options->Advanced. The pdf documents can be searched using the Adobe Reader search options when a PDF is opened. To search across multiple PDF documents, the pdf le must open directly within the Adobe Reader, not from within a web browser, and the advanced search options selected. To recongure your preferences in Adobe Reader 7.0 to launch les outside of the web browser, select on the Reader menu Edit>Preferences>Internet, then deselect Display PDF in browser.
The new NX Nastran Parallel Processing Users Guide is included with the NX Nastran documentation. This guide provides information on NX Nastran parallel methods for Linux and UNIX systems including parallel processing basics, detailed computational methods, an overview of execution methods, and examples using various solution sequences. The NX Nastran Numerical Methods Users Guide has been rewritten in this release. It now reects the software improvements since NX NASTRAN 1.
9-13
Chapter
10
Upward Compatibility
See the books NX Nastran 5 DMAP Updates and Additions and NX Nastran DMAP Programmers Guide for the updated formats.
10.1
Updated and New DMAP Modules
Updated Modules
BCDR DDRMM DSCMCOL EFFMAS EMG FOCOEL FOELCS FRLG GP1 GPSTR2 MATMOD MTRXIN NXNADAMS PARAML RANDOM SDR2 SDRCOMP SDRX SOLVIT XYTRAN
New Modules
ADDVM
10-1
Chapter 10
BOLTFOR BOLTSF DDR2N EMR GP1LM LRFORCE NXNRFI
10.2
Updated and New Datablocks
Updated Datablocks
CASECC CONTACT DSCMCOL DYNAMIC OEE OEF OES OPG OQG OUG
New Datablocks
OCCORF OCPSDF OSDISP2 OSHT
10.3
Updated and New Subdmaps
The following is the list of subdmaps where calling arguments and/or the input/output data blocks have changed: attmods.dat (new) cforce.dat cforcerd.dat cforcrd1.dat (new) cforcrd2.dat
10-2
conattm.dat (new) conmods.dat ddamexe.dat (new) ddampost.dat (new) ddampre.dat (new) disopts.dat disprs.dat genmodes.dat (new) ifpl.dat largrgb.dat (new) mkpxout.dat (new) modcon.dat (new) phase1a.dat phase1b.dat phase1c.dat phase1d.dat (new) phase1dr.dat phase1e.dat (new) phase1f.dat (new) poolit.dat (new) resddam.dat (new) sdr2stat.dat sdree1.dat (new) sedrcvr.dat semg.dat semg1.dat skewmat.dat (new) super1.dat super3.dat tranmat.dat (new) vdr1.dat xform.dat (new) xmtrxin.dat
10-3
Chapter
11
System Description Summary
Table 11-1. System Description HP9000 HP-UX

Item Supported Model(s) Congurations for Installed Timing Constants Build Operating System Other Supported Operating Systems Word Length Build Type MPI required for DMP PA-RISC 250, 710, 712, 715, 720, 730, 735, 778, 800, 819, 889, 2200, 2600, 2733, 3700, 4000, 4900, 6750 PA-RISC: HP-UX B.11.00 HP-UX B.11.11 32 bits LP-64, LP-64 DMP HP MPI 2.0.2 (comes with OS) Description
Table 11-2. System Description Intel Itanium HP-UX

Item Supported Model(s) Congurations for Installed Timing Constants Build Operating System Other Supported Operating Systems Word Length Build Type MPI required for DMP LP-64: 32 bits; ILP-64: 64 bits LP-64, LP-64 DMP, ILP-64, ILP-64 DMP HP MPI 2.0.2 (comes with OS) Intel-Itanium-HP-UX 4900, 5300, 5400, 5600 HP-UX B.11.23 Description
Table 11-3. System Description Intel Windows (32-bit)

Item Supported Model(s) Congurations for Installed Timing Constants Build Operating System Other Supported Operating Systems Word Length Build Type Intel and Intel-compatible Pentium II 400 MHz, P4 1.5GHz, Pentium Pro., P4 3 GHz Windows 2000, SP3 WXP SP2, WXP-64 SP1 (on EM64T/Opteron) 32 bits ILP-32 Description
11-1
Chapter 11
Table 11-4. System Description Intel (EM64T/Opteron) Windows (64 bit)

Item Supported Model(s) Congurations for Installed Timing Constants Build Operating System Other Supported Operating Systems Word Length Build Type Intel and Intel-compatible Pentium II 400 MHz, 733 MHz Itanium 32 Bit, P4 1.5GHz, Pentium Pro. Windows Server 2003 64 SP1 Windows XP64 SP1 32 bits LP-64 Description
Table 11-5. System Description Intel Linux

Item Supported Model(s) Congurations for Installed Timing Constants Build Operating System Other Supported Operating Systems Word Length Build Type MPI required for DMP Intel and Intel-compatible P4 2.8Ghz Redhat 7.3 Suse 9.2, Redhat 9, Redhat EL 3.0, Redhat EL 4.0 32 bits ILP-32, ILP-32 DMP HP MPI 2.0.2 (included with NX Nastran install) Description
Table 11-6. System Description X86_64 Linux (AMD Opteron/EM64T)

Item Supported Model(s) Congurations for Installed Timing Constants Build Operating System Other Supported Operating Systems Word Length Build Type MPI required for DMP X86-64 Linux 8664 Suse 9.0 Suse 9.1, Suse 9.3, Redhat EL 3.0, Redhat EL 4.0 LP-64: 32 bits; ILP-64: 64 bits LP-64, LP-64 DMP, ILP-64, ILP-64 DMP HP MPI 2.0.2 (included with NX Nastran install) Description
Table 11-7. System Description Intel Itanium Linux

Item Supported Model(s) Congurations for Installed Timing Constants Build Operating System Other Supported Operating Systems Word Length Itanium II IA-64 800 Mhz & 733Mhz Redhat EL3.0 Redhat EL4.0 LP-64: 32 bits; ILP-64: 64 bits Description
11-2
Table 11-7. System Description Intel Itanium Linux

Build Type MPI required for DMP LP-64, LP-64 DMP, ILP-64, ILP-64 DMP HP MPI 2.0.2 (included with NX Nastran install)
Table 11-8. System Description Sun SPARC Solaris

Item Supported Model(s) Congurations for Installed Timing Constants Build Operating System Other Supported Operating Systems Word Length Build Type UltraSPARC UltraSPARC (75 & 95) UltraSPARC: Solaris 8 Solaris 9, Solaris 10 32 bits LP-64 Description
Table 11-9. System Description IBM RS/6000 AIX (64 bit)

Item Supported Model(s) Congurations for Installed Timing Constants Build Operating System Other Supported Operating Systems Word Length Build Type MPI required for DMP Power3, Power4, Power5 303, 320H, 370, 375, 390, 397, 530, 530h, 550, 560, 570, 580, 590, 591, 950, 980, 990, 4316, 9133 AIX 5.1 AIX 5.2, AIX 5.3 LP-64: 32 bits; ILP-64: 64 bits LP-64, LP-64 DMP, ILP-64, ILP-64 DMP POE 3.2.0.0 (add on from IBM) Description
Table 11-10. System Description SGI R8K, R10K, R12K IRIX64

Item Supported Model(s) Congurations for Installed Timing Constants Build Operating System Other Supported Operating Systems Word Length Build Type MPI required for DMP R8K, R10K, R12K, R16K IP7, IP19, IP20, IP21, IP22, IP27, IP28, IP30, IP35, 240, 510 IRIX 6.5.21 IRIX 6.5.24m, IRIX 6.5.27m 32 bits LP-64, LP-64 DMP MPI comes with OS, version depends on OS level Description
Table 11-11. System Description SGI Altix

Item Supported Model(s) SGI-ALTIX Description
11-3
Chapter 11
Table 11-11. System Description SGI Altix

Congurations for Installed Timing Constants Build Operating System Other Supported Operating Systems Word Length Build Type MPI required for DMP 6402 SGI ProPack 3 sp3 SGI ProPack 4 LP-64: 32 bits; ILP-64: 64 bits LP-64, LP-64 DMP, ILP-64, ILP-64 DMP SGI MPT 1.10 (comes with OS)
11-4

NX Nastran 5 Release Guide

Caricato da

Informazioni sul documento

Descrizione originale:

Titolo originale

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

NX Nastran 5 Release Guide

Caricato da

Copyright:

Formati disponibili

NX Nastran 5 Release Guide

Proprietary & Restricted Rights Notice

NX Nastran 5.0 Release Guide

NX Nastran 5.0 Release Guide

NX Nastran 5.0 Release Guide

This section presents an overview of the features introduced in NX Nastran 5.

NX Nastran 5.0 Release Guide

NX Nastran 5.0 Release Guide

Multi-body Dynamic Software Interfaces

New Optimization Option

NX Nastran 5.0 Release Guide

Linear Contact Enhancements

NX Nastran 5.0 Release Guide

Strength Ratio Output for Composites

NX Nastran 5.0 Release Guide

NX Nastran 5.0 Release Guide

NX Nastran 5.0 Release Guide

SPC/SPCD Enforced Motion

NX Nastran 5.0 Release Guide

Enforced Motion Example

NX Nastran 5.0 Release Guide

NX Nastran 5.0 Release Guide

NX Nastran 5.0 Release Guide

Mathematical Details for Constraint Mode Method

NX Nastran 5.0 Release Guide

Input Updates for New Constraint Mode Method

DISPLACEMENT=5 DISPLACEMENT(REAL)=ALL DISPLACEMENT(SORT2, PUNCH, REAL)=ALL

NX Nastran 5.0 Release Guide

Describer SORT2 PRINT PUNCH PLOT REAL or IMAG

PHASE ABS REL PSDF

RPRINT NORPRINT RPUNCH

NX Nastran 5.0 Release Guide

Describer ALL NONE n

NX Nastran 5.0 Release Guide

where uf = absolute displacement yf = relative displacement us = enforced motion.

NX Nastran 5.0 Release Guide

Describer PRINT PUNCH PLOT REAL or IMAG

PHASE ABS REL PSDF

RPRINT NORPRINT RPUNCH ALL NONE

NX Nastran 5.0 Release Guide

where uf = absolute displacement yf = relative displacement us = enforced motion.

NX Nastran 5.0 Release Guide

ACCELERATION=5 ACCELERATION(SORT2, PHASE)=ALL ACCELERATION(SORT1, PRINT, PUNCH, PHASE)=17

NX Nastran 5.0 Release Guide

RPRINT NORPRINT RPUNCH ALL NONE n

NX Nastran 5.0 Release Guide

where uf = absolute displacement yf = relative displacement us = enforced motion.

NX Nastran 5.0 Release Guide

Mode Acceleration Method

Mathematical Details for New Mode Acceleration Method:

NX Nastran 5.0 Release Guide

Equation 2-9. where

NX Nastran 5.0 Release Guide

Input Updates for Mode Acceleration

NX Nastran 5.0 Release Guide

Random Response Output

Theory for Cross-power Spectral Density and Cross-correlation Functions

NX Nastran 5.0 Release Guide

If we denote the Fourier transform of a time history as:

NX Nastran 5.0 Release Guide

New RCROSS Case Control Command

NX Nastran 5.0 Release Guide