

Dear community,
As shown in the figure, the left panel presents the initial mesh generated in Trelis, while the right panel shows the mesh obtained after conversion using exo2nek. During this conversion, each HEX8 element is subdivided into eight HEX20 elements. How does this subdivision affect the numerical accuracy of the computation? If the effective mesh resolution required is on the order of 80,000 elements, is it feasible to generate only about 10,000 HEX8 elements in Trelis and obtain a high-order mesh that meets the accuracy requirements through the exo2nek conversion?
Any suggestions and criticisms are appreciated.
Student W. T. Ma
|
You don't often get email from mmwwtt...@gmail.com.
Learn why this is important
|
Dear Haomin
(Kirk) Yuan
Thanks for your answer, it really helps.
Student Ma