While I am aware that Strudl has three different error Estimating Methods, I mostly relied on changing mesh size to test for convergence or near convergence of the solution – obviously this approach could be time consuming. I would rather use Error Estimate approach, if interpretation of the results in either of methods is manageable.
Here are the questions I want to know to get started:
- Which one is the better estimating method?
- Are these methods case specific? – are the methods restricted to be employed under certain conditions only? If so what are these conditions for each method?
- What are the acceptable errors using each method. Could you locate me any literature with information about acceptable errors when each method is used?
Last but not least, I appreciate any info explaining the analytical roots of error estimate methods used in Strudl.
Here are the questions I want to know to get started:
- Which one is the better estimating method?
- Are these methods case specific? – are the methods restricted to be employed under certain conditions only? If so what are these conditions for each method?
- What are the acceptable errors using each method. Could you locate me any literature with information about acceptable errors when each method is used?
Last but not least, I appreciate any info explaining the analytical roots of error estimate methods used in Strudl.