I remember that two years ago when I did the case studies and exercises by using 2013 most results were the same as the book.
Hi Dongwei, results will differ slightly from service pack to service pack, and from version to version.
Even if you run the same study twice, you won't get exactly the same results. Simulation utilises an automesher which means the mesh is never exactly the same for the same study when run.
Having said that, there should not be significant differences in results between 2013 and 2015 (>10%). There were enhancements made to Simulation in 2014 and 2015, perhaps the study setup is slightly different in 2015, have you tried redefining the contact sets, and loads etc?