Jump to content

Mandy Kaiser

  • Content Count

  • Joined

  • Last visited

Everything posted by Mandy Kaiser

  1. Dear Alejandro, thank you for your explanations on the loss models. Within the presentation you attached was a screenshot of an excel file for determining the coefficients of the bertotti equation. Could you also provide this file or is it somewhere in the FLUX installation folder? Best regards, Mandy Edit: I just saw, there is the path to the file: flux\Flux\DocExamples\Tools\BertottiLossesCoefficients
  2. Dear Alejandro, thank you for your fast and helpful reply! You brought light into the darkness! According to your explanations, the determination of the resistance by FLUX is correct (something else would have also surprised me... ;-) ). Because of the symmetry I have 2 non-meshed coils and when calculating by hand I also get about 6 mOhm. So, I will set the value to zero and let FLUX calculate the resistance by giving the material parameters and the filling factor. I also looked in the documentation of FLUX, but I couldn't find the explanation to this point. Maybe I didn't look in the right spot, but maybe it isn't described explicitly? If necessary, can Altair revise the documentation at this point? Best regards, Mandy
  3. Dear all, I have a question about stranded coil conductors using non-meshed coils. I have performed two transient simulations with non-meshed coils in 3D for a transverse flux machine. For the first simulation I entered 100 mOhm in the field "Resistance Formula" of the stranded conductor. For the second simulation I entered 5 mOhm. The geometry as well as the material parameters and number of turns were the same for both simulations. For both simulations, the non-meshed coil was part of an electric circuit. I assumed that the resistance value is calculated via FLUX using the resistance formula R=rho*l/A and that the value entered in the "Resistance Formula" field (stranded coil conductor) does not matter if the specific resistance as well as the fill factor are specified for the non-meshed coil. However, the result for the evaluated resistance is different for both simulations. Preset value of 100 mOhm --> evaluated resistance = 106 mOhm Preset value of 5 mOhm --> evaluated resistance = 11 mOhm The evaluated resistance of the non-meshed coil is the same for both simulations: 3 mOhm. Why is there a difference between the evaluated value of the non-meshed coil and of the stranded conductor? I assumed, this should be the same? Therefore, could you please tell me, how the evaluated resistance of the stranded coil conductor is determined? Is it important, to enter the "right" value in the field "resistance formula"? Thank you in advance and best regards, Mandy
  4. Dear Alejandro, thank you very much for your reply and the paper's recommendation. It seems to be exactly what I need. According to my understanding it is especially important to have a fine mesh in the area of the air gap. Here I generally choose a 4-layer hybrid mesh, whereby I assign the volumes in the two central layers to the mesh generator "mapped" and use an automatic mesh for the outer layers in the air gap. I also assign the mesh generator "Mapped" to all stator and rotor surfaces facing to the air gap. Additionally, I limit the edge length of the mesh cells in the air gap by assigning a mesh line with minimum distance between the nodes to the corresponding edges. The number of mesh cells with poor quality is below 0.1%. I have only been working with FEM simulations for a few months, but I would say that the mesh is very good. Do you have any other tips for improving the mesh? Nevertheless, I will have a closer look into the paper you suggested. At first sight only the Frozen Permeability method is required, which I have already implemented in 3D. Best regards, Mandy
  5. Dear all, I would like to implement a macro for the separate calculation of the different torque components (cogging torque, reluctance torque and interaction torque) in 3D simulation of motors using the Frozen Permeability Method. I already derived the Frozen Permeability method for the 3D case from the existing 2D macro. I found a suitable method in Baserrah, S. Theoretical and Experimental Investigations of a Permanent Magnet Excited Transverse Flux Machine with a Segmented Stator for In-Wheel Motor Applications. Dissertation, Universität Bremen, Bremen, 2014. (page 238). In the first step I would like to calculate the cogging torque using the stored magnetic energy (the Frozen Permeability Method is not yet required for this). With the help of a sensor I can determine the stored magnetic energy. Since I take advantage of the symmetry and periodicity of the motor and simulate only a part of it, I have to multiply the result by a factor. The derivation according to the rotor angle roughly gives the course of the cogging torque, but with strong fluctuations compared to the calculation via 'TorqueElecMag(ROTOR)' (see image below). Does anyone have any idea what may be the cause of this deviation? I have concerns about using this procedure to calculate the reluctance torque, because large deviations are also to be expected here. Can anyone give me a hint on how to extract the reluctance torque from the simulation? Many thanks in advance! Mandy
  • Create New...