[FEMM] Finite Element Magnetic Model of the SRM2 Electric Synchronous Reluctance Motor being considered for the Hybrid Electric Vehicle Toyota Prius (2023)
Feel free to use this FEMM model as you wish. Below are some points that I had to assume in order to get the model running as close to the published data [1,2], as possible. If you find new improvements, feel free to adjust the model and share your findings with the rest of the community, possibly here via Figshare. I hope this is useful. Have fun !
-----------------------------------------------------------------------------------------------------------------------------------------------------------
BEGINNER: if you are just starting, and wish to access a simple step-by-step example of an electric motor modelled in FEMM, then I would recommend the following short thesis, before tackling the above simulation.
FEMM: official homepage and open source software download is here. With it, you can open, modify and rerun the above FEMM files.
DESIGN: the 2D drawing of the model was extracted (to the best of this authors ability) from the original publication [1] The DWG files are in enclosed, so you can make your own modifications. These were created with the open source software nanoCAD, however other free CAD softwares like FreeCAD can be used too.
ANGLES/CURRENTS: The same model with various angles are provided here. To make your own angle, take the aligned case, select the arc curves by mouse right-click (they become red colored), and press in the top down menu Edit > move. There select rotation (the default of 0,0 is correct) and type your angle (negative moves CW, and positive CCW). To alter the current i, go to the top down menu Properties > circuits. Select R and press modify, followed by typing the new current. Do the same thing for B. Rerun the model, and you are done.
MATERIALS: Not all of the required material properties of the Super Core 10JNEX900 necessary for the stator and rotor (like for instance the non-linear B-H curve) were available to program it into the FEMM, thus an alternative (lesser performant) steel was used (namely M-15 which was available in the FEMM material library). This possibly accounts for the fact that the predicted flux-linkage values are slightly lower (around on average 5%) than those originaly published in [1,2].
COILS: As per this author's understanding, the experiment had each stator pole coiled with 17 parallel-connected wires (each with a diameter of d=0.6mm) with 22 turns (around the pole) each, stacked on top of each other. The approach used to model this in FEMM was to lump the 22 turns into a single turn of an equivalently larger diameter wire (D=sqrt(N)d=sqrt(22).0.6=3mm, having the same group crossectional area and thus transporting the same current ---- skinning effects are not considered here), and coil it 17 times around the pole. The Flux-Linkage (Wb) is extracted from the results (.ans file), by pressing the button that looks like a coil. The aforementioned short thesis presents an analytical way to compute the flux-linkage, based on the individual path of each magnetic circuit. If you wish to model a different coil length other than 135mm, then this can be altered in the top down menu Problem (keep the name of the coil material short like W x 22, otherwise they overlap and become ilegible).
OBSERVATIONS: In my experience, if (when you press the compute button) the error "Cannot compute" appears, it is almost always because you forgot a point not connected to a line, or an undefined source point somewhere, that only needs deletion for it to work (a recurrent problem comes when you press somewhere in the FEMM window as you transit from another window, and automatically leave a point on that spot that prevents the computation --- FEMM does not have a mouse pointer, and the default is the point creation/highlighting pointer. This has hapenned to me many times, and thus it may happen to you too). So, the error probably has nothing to do with the numerics, and does not necessarily mean that it will not compute, it simply means that you didn't tidy up your geometry (and it will not run until this is resolved). Problems with lines show in red are easy to see, but a missed point somewhere is harder to spot. Usually, if indeed the solver cannot compute, what you will see is the progress bar starting and getting stuck (either in the 2nd or 3rd bar of the first sweep), or the progress bar never moves (meaning that the solver got stuck somewhere in the mesh, trying to solve something it cannot compute like a numerical anomaly --- for instance, a singularity on a cell predicting infinite magnetic flux).
[2] - Average Rated Torque Calculations for Switched Reluctance Machines Based on Vector Analysis
-------------------------------------------------------------------------
For more public data, please visit my Figshare profile: https://figshare.com/authors/Luis_Teia/10811244
