In older versions of LS-DYNA (950 and lower) , when mass-scaling was active and over the course of the simulation when the percentage of added mass was greater than ENDMASS in *CONTROL_TERMINATION , LS-DYNA would terminate that job with no additional message. However, recent versions of LS-DYNA terminate the job AND provide a list of 20 nodes into D3HSP file which are sorted by the magnitude of their nodal mass (LS-DYNA uses a lumped mass approach). This usually helps since we can now review these 20 nodes to identify local instability especially in a large model.
What I wanted to share with this post is how can we use those 20 nodes more effectively. The list of 20 nodes provided by LS-DYNA could be misleading since only the ABSOLUTE is used instead of the incremental mass (possibly just over a certain last few cycles). So reviewing the ENTIRE list of 20 nodes could contain nodes which had no increase in added mass since the first cycle. When viewing graphically, this could contain nodes away from where the local instability could have occured. The recommended method would be view just the first 4-5 nodes OR normalize the mass with respect to the maximum mass of the 20 node list.