Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Introduction
Structural impact involves events such as plastic flow at high strain rates, possible local increase of temperature, and material fracture. The usual approach in numerical simulations is to operate with two different models, one representing plastic flow and one representing fracture. These two models can be coupled or uncoupled. To describe the various phenomena taking place during ballistic penetration, it is necessary to characterize the behavior of materials under impact-generated high strain rate loading conditions. The characterization involves not only the stress-strain response at large strains, different strain rates and temperatures, but also the accumulation of damage and the mode of failure. Such complex material behavior involving fracture is difficult to describe in analytical models. Another important feature is the physical difference between plastic flow and fracture. In ductile metals, plastic flow may be viewed macroscopically as a visible shape change, microscopically as the appearance of slip lines, and at the atomic level as the movement of dislocations. It is also known that plastic flow is driven by the deviatoric stress state in the material. Initiation of damage or fracture is due to arrests of dislocations by micro defects or micro stress concentrations giving de-cohesion and subsequently nucleation, growth and coalescence of micro cracks and micro voids. The damage evolution is strongly influenced by the hydrostatic stress state in the material. Accordingly, different mathematical models are needed to describe plastic flow and fracture, Johnson-cook is one of them.
J-C model is a coupled material model capturing viscoplasticity (rate-dependent plasticity) and ductile damage, developed for impact and penetration problems. This model is analytically verified by Hopperstad et al. [1] within the framework of viscoplasticity and continuum damage mechanics [2], allowing for large plastic strains, high strain rates and adiabatic heating.
(1)
Where A, B, C, n and m are constants, is accumulated plastic strain and = / is dimensionless strain rate. and are plastic stain rate and user defined reference strain rate.
The non-dimensional temperature =
Where is the current temperature, is the ambient temperature, and is the melting temperature. Clearly, the von Mises equivalent flow stress is the product of three factors representing strain hardening, strain rate and temperature. This facilitates the calibration of the model because each of the parentheses in Eq. (1) can be handled separately in three series with uniaxial tensile tests.
= +
(3)
Where D is damage variable, r is damage accumulated plastic strain =(1-D) , D = 0 virgin material and D = 1 fully broken material and hence D=Dc Stress state is an important factor in determining when fracture occurs. In particular stress triaxiality plays an important role [4] in governing the tendency for ductile fracture. Stress triaxiality is used to describe the portion of stress tensor that is hydrostatic. It is defined as the ratio of hydrostatic stress to equivalent Von-Mises stress. In other words the stress state with high triaxiality approaches the completely hydrostatic stress while with lower triaxiality stress state deviatoric stress dominates. Several investigations have revealed that increased triaxiality reduces ductility and thus failure strain The model describes linear elasticity, initial yielding, strain hardening, strain-rate hardening, damage evolution and fracture all material constants can be identified from uniaxial tensile tests.
Step 2.(A) Again the Quasi-static tensile tests were performed at large strains. Diameter reduction was observed till fracture in steps by vernier caliper and plastic strain in necked region is obtained by
= 2 ln ( )
F = A
where F is the applied force and A is the cross-sectional area instantaneous steps
Before necking,
= f (F)
Step 3. (B, n, Dc) In a similar way as for the smooth specimen, the applied load and crosssectional area in the notch specimens are measured during testing. It is assumed that the stress triaxiality ratio is approximately constant during plastic straining in each notched specimen while it varies in smooth specimen tensile test.
The different notches used gives a concentration of hydrostatic tension in the test specimen.
+ (4)
Where , and are discrete values of the variables , and, and is the measured fracture strain for the different notched specimens and is the measured fracture strain for different notched specimen.
An artificial notch produces an initial triaxiality different from that in the case of a smooth specimen where triaxiality is initiated only after the commencement of necking. Bridgmans relationship was used to correlate initial notch radius R and maximum stress triaxiality ratio 1 = + ln 1 + 3 2 The damaged accumulated plastic stain is calculated as :
= 2
1 2
(5)
The method of least squares is then used to minimize the difference between the experimental determined curves and the model by varying B, n and Dc simultaneously As seen, good correlation is obtained for the notched specimens. However, the fit is somewhat poorer for the smooth specimen. This is probably caused by the variation in stress triaxiality during testing.
Step 5. (m) Effect of thermal softening on true stress is shown in figure. yield stress shows a linier decrease with increasing temperature while ultimate stress shows a local minima at about 300 deg. C By assuming adiabatic condition at high strain rates, the effect of thermal softening is included in the model by fitting the material constant m to the decreasing yield stress for ( = 0, = /0 = 1 ) = +
where a and R are the initial radius of the specimen in the neck and the initial neck radius, respectively, since the triaxiality ratio is assumed constant during plastic straining.
The three material constants, , , in the model have been fitted to the given data, and the curve has been extrapolated into the hydrostatic compression region as seen in Figure.
cannot assume negative values because it corresponds to the purely hydrostatical stress state.
Since the strain rate and temperature effects on the flow stress are uncoupled ( eq. 1) , the Johnson-Cook model is easy to calibrate with a minimum of experimental data.
Damage evolution can be coupled with the model via strain equivalence principle of damage mechanics [2]. Strain rate sensitivity (step 4) is found experimentally to increase with increasing temperature, while the Flow stress is decreasing as observed by Harding [7] which is in contrast to the model. According to Harding [7] this uncoupling between thermal and viscous effects (step 7 and 8) used in the J-C model may not capture the correct physics observed in experiments.
Negative aspects:
Closing Remarks
One of the important parameter to accurately calibrate J-C model is stresstriaxiality. As seen, to obtain stress triaxiality for different notched specimen is a tedious and costly affair. Alternatively, it can be evaluated in Abaqus-CAE by requesting the TRIAX command in field output for different notched specimen. J-C elasticity model is : = +
and all the parameters A, B, etc. can be effectively evaluated in a single shot just by having a engineering stress strain data (may be or may not be time dependent) for the material in use. How ? By using Mcalibration software developed by Jorgen Bergstrom [8]. It uses the regression analysis and fits the full equation in a single go to the given material data.
Borvik [5] suggested some improvements in the plasticity model like approaches minus infinity for very small strain rates. So he modified the corresponding bracket by +
Value of quantifies . As is related to deviatoric state of stress during the failure it governs the deformed shape of the material as shown in figure in which creation of petals during an impact by conical projectile is governed by deviatoric state of the Cauchys stress tensor [9]. Changing the value of by keeping all others constants same will change the way material deforms.
References
1. 2. 3. 4.
Hopperstad OS, Berstad T, Borvik T, Langseth M. Computational model for viscoplasticity and ductile damage. Proceedings of "fifth International LS-DYNA User Conference, Michigan, USA, 21-22 September 1998. Lemaitre J. A Course on damage mechanics. Berlin: Springer, 1992. Johnson GR, Cook WH. A constitutive model and data for metals subjected to large strains, high strain rates and high temperatures. Proceedings of Seventh International Symposium on Ballistics, The Hague, The Netherlands, April 1983 Johnson GR, Cook WH. Fracture characteristics of three metals subjected to various strains, strain rates, temperatures, and pressures. Eng. Fracture Mech. 1985; 21: 31-48.
5.
6. 7. 8. 9.
Borvik T, Langseth M, Hopperstad OS, Malo KA. Ballistic penetration of steel plates. Int J Impact Eng 1999;22:85586.
Bridgman PW. Studies in large plastic fow fracture. Cambridge, MA: Harvard Univ. Press, 1964. Harding J. The development of constitutive relationships for material behaviour at high rates of strain, Inst Phys Conf Ser No. 102: Session 5, Oxford, UK, 1989. http://polymerfem.com/content.php?9-MCalibration M.A. Iqbal, S.H. Khan, R. Ansari, N.K. Gupta, Experimental and numerical studies of double-nosed projectile impact on aluminum plates, International Journal of Impact Engineering, Volume 54, April 2013, Pages 232-245, ISSN 0734-743X, 10.1016/j.ijimpeng.2012.11.007.
Thank you