|
> Hello McGee, David and All:
> It seems great idea. I have the experence in calculating > hysteresis losses in electric machine. The way I used is > same as you, a time stepping approach. But I still have > problem in how to get a satisfying and correct result by > interpolating in each group of data. Can you recommand > some references, I hardly to find this kind of source > here. > > Thank you. > Si hang > > ----- Original Message ----- >> From:McGee >> To: femm@xxxxxxxxxxxxxxx >> Sent: Tuesday, April 10, 2001 10:20 PM >> Subject: RE: [femm] Hysteresis >> David >> >> Using the approximate approach for modeling hysteresis >> that you have implemented in your model has worked for me >> in the past as a "history matching" parameter. As a >> predictive tool an approach using Preisach's Theory might >> be useful. I would like to take a stab at implementing >> hysteresis into your code. It would be tremendously >> beneficial to me (hysteresis losses in steel pipes) and I >> am sure others might find it useful. Have you looked at >> Preisach's Theory. I know several researchers have >> implemented this approach into FEM code, however I do not >> know the success of this approach. I used a time >> stepping approach around the BH loop as H varied with >> time; in FDTD this was a very direct approach. Any >> suggestions. >> >> Bruce On a mesh of 3-node triangles, I would be consistent to take the average flux density over the element for use in your hysteresis model. For example, this is how things are interpreted for the purposes of the B-H curve in nonlinear problems. I am aware of the various flavors Preisach model and I have seen papers about people implementing this hysteresis model in the context of finite elements, but I haven't had the occasion to implement one yet myself. It would be interesting to make a time-transient solver for femm and include a reasonable model of hysteresis effects. However, this is a non-trivial extension to the program. Really, it would be a major revision to the way things work, and it would probably end up as a separate program all together. It would also be nice to hook things into a more general simulation engine (may something like Scicos, the Simulink-like add on to Scilab (a free matlab "clone") ). Anyhow, I think I have just about everything that I want in the version 3 program, with the exception of completing the documentation--I have to draw the line somewhere. But that just means that it's time to plan what the next version will look like. I'd welcome those of you who are interested in developing the next version of femm. To help coordinate the next version of the program, I've recently opened up a project at Berlios (http://developer.berlios.de), which is sort of like SourceForge but not so big. If you are interested in being a developer for the next version, you can join the project on Berlios. Dave Meeker
|
begin:vcard n:Meeker;David tel;fax:781-890-3489 tel;work:781-684-4070 x-mozilla-html:TRUE url:http://members.aol.com/dcm3c org:Foster-Miller, Inc.;Electrical and Electronic Systems Group version:2.1 email;internet:dmeeker@xxxxxxxxxxxxxxxxx title:Senior Engineer adr;quoted-printable:;;350 Second Avenue=0D=0A;Waltham;MA;02451-1196;USA fn:David Meeker, Ph.D. end:vcard