Investigation of Model Simplification and Its Influence on the Accuracy in FEM Magnetic Calculations of Gearless Drives

Publication: Research - peer-reviewJournal article – Annual report year: 2012

View graph of relations

Finite-element models of electrical motors often become very complex and time consuming to evaluate when taking into account every little detail. There is therefore a need for simplifications to make the models computational within a reasonable time frame. This is especially important in an optimization process, as many iterations usually have to be performed. The focus of this work is an investigation of the electromagnetic part of a gearless mill drive based on real system data which is part of a larger project building a multiphysics model including electromagnet, thermal, and structural interactions. This multiphysics model will later on be used for simulating and parameter optimization of a gearless mill drive with the use of Evolution Strategies which necessitates the reduction in computation time. What has been investigated is how model simplifications influence the accuracy on the calculated forces and torque coming from the drive where each simplification made is described and justified. To further reduce the evaluation time, it is examined how coarse the mesh can be, while still predicting the results with a high accuracy. From this investigation, it is shown that there are certain ratios between the mesh size in the air gap and the iron core, which will result in an optimal determination of the forces and torque. It will be shown that it is possible, just through simplifications and choosing the correct mesh size, to reduce the computational time by 98%, keeping an accuracy for the torque and forces of less than 0.3% and 1.2%, respectively.
Original languageEnglish
JournalI E E E Transactions on Magnetics
Issue number7
Pages (from-to)2166-2177
StatePublished - 2012
CitationsWeb of Science® Times Cited: 1
Download as:
Download as PDF
Select render style:
Download as HTML
Select render style:
Download as Word
Select render style:

ID: 9770252