A Penalty-Based PSO Algorithm for Demagnetization Risk-Free Design of Slotless Halbach PM Machines

The versatility of analytical models is limited by the fundamental assumption of having linear material properties. The performance of the popular Halbach PM rotor is bounded to a careful optimization of the array; thus, a thorough demagnetization risk assessment is already required under no-load conditions. For this reason, finite element analysis (FEA) is typically the preferred choice for the inherent capability to handle non-linearities at the expense of time-intensive computations. This paper fully exploits the computational efficiency of analytical modeling into a penalty-based particle swarm optimization (PSO) algorithm to design demagnetization-free slotless machines. The demagnetization risk is assessed through the analytical field solution and used as a penalty in the PSO to target demagnetization-less topologies in the design phase. The optimization routine without the penalty is shown to heavily overestimate the performance compared to the one employing the penalty, suggesting the latter method to be more reliable.