It should be noted that Abaqus/Standard may generate inaccurate external work (ALLWK) in a static geometrically non-linear analysis (finite hyperelasticity) and, consequently, will cause ETOTAL to be inaccurate.
For example, we assume that a hyperelastic solid is only subjected to body forces. It doesn't matter a linear or non-linear analysis will be, the external work of the body forces can be calculated over reference configuration
$\;\;\;\;\;\;\;W^{ext}=\int_{\Omega_0} \boldsymbol{f}_0(\boldsymbol{X})\cdot \boldsymbol{U} \;dV $
For linear static analysis, we obtain the almost identical values of the external work provided by Abaqus and using custom post-processing calculation. Turning to finite elasticity, the difference may be significant and leads to the erroneous results in the external work estimation.
P.S. Recall that for a static problem a value of external work reported by Abaqus is half of "true" external work.
For example, we assume that a hyperelastic solid is only subjected to body forces. It doesn't matter a linear or non-linear analysis will be, the external work of the body forces can be calculated over reference configuration
$\;\;\;\;\;\;\;W^{ext}=\int_{\Omega_0} \boldsymbol{f}_0(\boldsymbol{X})\cdot \boldsymbol{U} \;dV $
For linear static analysis, we obtain the almost identical values of the external work provided by Abaqus and using custom post-processing calculation. Turning to finite elasticity, the difference may be significant and leads to the erroneous results in the external work estimation.
P.S. Recall that for a static problem a value of external work reported by Abaqus is half of "true" external work.
