Tutorial summary | |
---|---|

Project name | tuto_IPMSM_12 |

Machine name | machine_IPMSM_A |

Source project | tuto_IPMSM_01 |

Description | This project aims at improving the calculation of the airgap flux and resulting Maxwell stress by including the tangential flux density component in the electromagnetic model. This can be done using the powerful subdomain modeling techniques extended to interior permanent magnetic machines as shown in the article "E. Devillers, J. Le Besnerais, M. Hecquet and J-P. Lecointe, A new hybrid method for the computation of electromagnetic noise in open circuit IPM machine, EVER 2017". |

The project *tuto_IPMSM_01_test* is copied/pasted (or loaded and saved with the GUI) in a new project *tuto_IPMSM_12_test* will the following change:

`Input.Simu.type_Bmodel=1; `

To let MANATEE calculate the optimal number of harmonics in the subdomain model one should also put (available only in the script for now)

`Input.Simu.Nhag = 0; % number airgap space harmonics in subdomain models `

Input.Simu.Nhss = 0; % number stator slot harmonics in subdomain models

Input.Simu.Nhrs = 0; % number rotor slot harmonics in subdomain models

Input.Simu.Nhso = 0; % number stator slot opening harmonics in subdomain models

Input.Simu.Nhro = 0; % number rotor slot opening harmonics in subdomain models

One can also automatically add the following post processing commands:

`Input.Simu.SS_post_filename ='plot_init;plot_B_space;plot_VS_ASWL_modal_cont;plot_VS_ASWL_order_analysis;plot_VS_ASPL_spectrogram';`

The simulation is run in command line with

`run_MANATEE('tuto_IPMSM_12_test')`

The call to FEMM is very short and does not significantly change the overall computing time:

The noise as a function of speed is now given by (plot_VS_ASWL_modal_cont) :

The variable speed sound pressure level is given by plot_VS_ASPL_spectrogram:

These conclusions show that at no-load up to 6500 rpm, the Prius electrical machine acoustic noise shows two resonances at 12th and 24th the electrical frequency with the breathing mode of the stack. This is in full agreement with the conclusions of this FEA work:

Running **a full electromagnetic and structural finite element model at variable speed requires several hours of calculation, while MANATEE only takes a few seconds to obtain the same conclusions**.