Adjoint Sensitivity Analysis of Radial Force Components of Switched Reluctance Machines

Switched reluctance machines (SRMs) are getting more popular due to their uncomplicated and rugged structure. However, acoustic noise and vibration resemble two main issues with this type of machine. Radial forces acting on the machine stator contribute significantly to acoustic noise generation. These forces can be reduced by optimizing the machine geometry using deterministic optimization techniques. These techniques need the sensitivities of the cost function with respect to the design variables. In this work, the sensitivities of the radial force density components of SRMs are estimated with respect to various geometrical parameters using the adjoint variable method (AVM). The AVM method reduces the computational burden as compared to the more precise but computationally costly Central Finite Differences method (CFD).