Sei sulla pagina 1di 3

best practices

Modeling Threaded
Bolted Joints in
ANSYS Workbench
Although bolted joints are extremely common, they can be difficult to model
accurately without using some best practices from an industry specialist.
By Michael Oliver, Oliver Testing and Consulting, Warner Robins, U.S.A.

B
olted joints are extremely com- The joint could consist of a standard hex,
mon fasteners in construction heavy hex or socket head bolts using a nut
and machine design. However, or internal thread (all per applicable
creating a finite element industry specifications). The program
model (FEM) of a threaded allows adjustment of the fit for both the
bolted joint is a complicated task but internal and external threads within the
well worth the effort. The steps are: limits of geometric size per industry
develop a solid model, create contacts, standards. This is an important concept
mesh the model and then establish from a manufacturing standpoint because
boundary conditions. Each of these tasks the threads have geometric size toler-
has its own set of challenges and issues; Figure 1. 3-D model of threaded bolted ances and allowances. Figure 2 shows the
these all impact the ability of the software joint using standard hex bolt into internal maximum and minimum flank-to-flank
thread
not only to converge, but to give accurate contact that can exist within the specifica-
results. The model is used to determine tion limits. Create the flanks surface
not only what happens when torque is areas as coincident, because a gap here
applied and clamp load developed, but will likely cause a bisection in the first
what happens when external loads are substep of the solve.
placed on the joint: Does the joint remain
tight, or is there failure? Here are some FINITE ELEMENT MODELING
suggestions for best practices. Once the 3-D model has been created,
Using a 3-D modeling approach is pre- export it to SolidWorks design software
ferred for most applications that are via IGES in which the individual volumes
deemed critical or that involve tightening of each part of the joint are combined. The
against or into soft material (in this case, critical point in this operation is that
aluminum). Torque must be applied to when the volumes are combined, the area
the hex faces of the bolt to provide rota- segments disappear to create one contact
tion. This develops clamp load in the joint surface instead of four. Once edited,
and mimics reality. Avoid using simulated import the solid model into the ANSYS
nonthreaded fasteners (place keepers) or Workbench environment directly using
pretension elements due to lack of rota- the geometry interface module.
tion. The rationale for both approaches This model contains only three contact
will become evident. zones: bolt underhead, threads and bot-
tom of the joint. Figure 3 shows a cross
GEOMETRY CREATION section of the FEM of Figure 1, with its
The authors practice is to create the 3-D three contact zones. Note that there are
model (Figure 1) within ANSYS APDL contact and noncontact areas in the exter-
using a specifically created .mac or macro nal and internal threaded region. The con-
file. (Alternatively, the geometry could be Figure 2. Cross section of external and tact parameters for each zone are the
internal threads. Threads shown are for
created in a CAD tool or in ANSYS maximum flank-to-flank condition (top) and same (except for the frictional values), as
DesignModeler.) This .mac file is an inter- minimum flank-to-flank condition (bottom) shown in Table 1. The values input for
active program that allows creation of a allowed per industry specification. friction are critical and determine whether
joint of almost any thread size and pitch. the analysis is correct. The values used in

ANSYS ADVANTAGE | 2012


Under-head
contact regions

Bolt and nut


thread contact
Bottom of joint
contact regions

A
Threaded
contact region

No thread contact

Figure 3. Cross section of threaded joint Figure 4. Cross section of internal thread Figure 5. Meshed model from Figure 1
model showing three contact regions. showing developed mesh. The flanks on
Area A shows a magnified view of the the lower surface are loaded. The bolt
threaded contact region. head is up.

Table 1. Connection and Contact Settings


this example were determined from
Connections Auto detection No experiments performed in a fastener test
Tolerance slider 0 lab. (See more below.) Tabular values
Tolerance value 0.15 mm from industrial specifications or hand-
Face/face Yes books will work, but be aware that friction
Face/edge Yes in the underhead region and threads are
Edge/edge Yes rarely the same. If experimental data is
not available, you can establish a range of
Contacts Type Frictional values and execute multiple solves. In
Frictional coefficient 0.059 / 0.147/ 0.1 reality, changes made to the underhead
Scope method Manual region will have a bigger impact on clamp
Behavior Auto symmetric load development than that of the
threaded region.
Advanced Formulation Augmented Lagrange
Interface treatment Add offset, no ramping
MODELING
Offset 0 mm
Consistency of size and shape of mesh ele-
Normal stiffness Manual
ments on the thread flanks is important
Normal stiffness factor 0.1
for evaluation of load and stress distribu-
Update stiffness Each iteration
Pinball region Radius tions. Figure 4 shows the meshed internal
Pin radius 0.1 mm threads of the model from within
Workbench. A fairly dense mesh on the
contact flanks and root is recommended,
Table 2. Analysis Settings
then a less-dense mesh on the noncontact
Step Controls Number of steps 1 flanks. The top flank of the first internal
Step end time 1 thread requires a fine mesh as well,
Scope method 1 because this thread is not complete and
Auto time stepping On can actually plastically deform. The mesh
Define by Substeps on the bearing surface is controlled
Initial substeps 30 through contact sizing; in this example,
Minimum substeps 1 element sizing was used on the threads of
Maximum substeps 100 the bolt. Figure 5 shows the entire meshed
model. Default settings were used with all
Advanced Solver type Direct mesh sizing definitions with a relevance
Weak springs Program controlled set at zero. The analysis settings are
Large deflection On shown in Table 2 for this static structural
solution. This example model had
Nonlinear Controls Force convergence On 365,427 elements (88,178 contact ele-
- Value Calculated by solver ments) and 607,233 nodes.
- Tolerance 0.5 percent
- Minimum reference 0.01 N

ANSYS.COM 2012 ANSYS, INC. ANSYS ADVANTAGE


best practices

Figure 6. Equivalent stress on the joint Figure 7. Equivalent stress on joint Figure 8. Deformation in Z direction for
bearing surface of the torqued solve. bearing surface of pretension solve. top of internal threaded region
Maximum stress is 192.12 MPa. Maximum stress is 177 MPa.

This model used experimental data: The above model was also used in SUMMARY
thread torque, clamp load and total torque a comparison study with pretension The goal of simulation is to simulate the
(torque in) with the aid of a torque/ten- elements (conducted in Workbench) part or assembly under actual operating
sion load cell and a torque transducer. applied to the shank of the bolt, using conditions. But even with a 3-D model,
Underhead torque as well as thread and the clampload value of 27,105 N. Figures results can be inaccurate and misleading
underhead coefficients of friction were 6 and 7 show the bearing surface from if you do not follow best practices. This
calculated using the following equations both the torqued and pretension solu- may not be an issue with steel on steel,
(which can either be solved or obtained tions, respectively. Note the magnitude of but if the joint has soft material, or
from any good mechanical engineering the pressure on both solves (A356 yields even a gasket within it, a rotating 3-D
reference book). at 179 MPa). The rationale for the higher model is a must, and frictional values
must be carefully considered.

Reference
moliverconsulting.com

The author thanks Mai Doan from


ANSYS for help in transitioning the .mac
file from solving in ANSYS Parametric
Design Language to solving in the
ANSYS Workbench platform.

Where:
dp = Pitch diameter Pitch = Thread pitch
Dkm = Average of dw and dh Tin = Total torque or torque in
dw = External flange diameter of Tthd = Thread torque
bolt or nut Tund = Under head torque
dk = Joint clearance hole diameter thd = Thread coefficient of friction
P = Clamp load und = Under head coefficient of friction

The material for the joint and internal pressure on the rotational model is obvi-
threads was A356 aluminum. Thread size ously because of rotational shear forces.
was M10 x 1.5 6g6g for the external Figure 8 shows the top of the internal
threads and M10 x 1.5 6H6H for the inter- threaded region from the rotational solve.
nal threads. The measured clamp load The high amount of deformation is a
value of the actual joint, at 40 N-m of direct result of the partial thread that
torque, was 27,105 N, while the reaction exists. The amount of deformation was
force from the simulation was clampload measured in the lab, matching that of
value of 27,367 N. This represents a 1 the model.
percent error.

ANSYS ADVANTAGE | 2012

Potrebbero piacerti anche