Figure 4.5 "Uniaxial Behavior for Multilinear Kinematic Hardening" in the
ANSYS 10.0 Theory Reference shows a default MKIN tangent slope of zero
after the fifth (last) MKIN defined point. However, it looks like ANSYS is
just extending the last defined slope instead of using a zero tangent slope.
Actually, ANSYS should be using a slope that is 1.0e-6 times the Young's
modulus, which is essentially zero, but a little more stable than perfectly
plastic. Due to the error, the new generation elements are carrying much
more load than they should be.
The problem does NOT occur for the old generation elements (like SOLID45).
Further, an equivalent KINH material model works fine for both the older
elements and the new generation elements (like SOLID185 and BEAM188).
For BEAM188 elements, the error does NOT occur if the quadratic shape
function (KEYOPT(3)=2) is NOT used. The section properties also make a
difference. These issues cause the reaction loads to be in error, too. Is
this behavior due to a known error?
Yes, Defect 42575 was filed and fixed in the daily build of ANSYS 11.0.
The MKIN behavior now matches the equivalent KINH behavior. ANSYS
build 11.0BETA UP20060119 was verified on a Windows XP Professional PC.