Design Improvement of Dual Pole Synchronous Reluctance Motor

This paper presents design improvements and optimization of the dual-pole line start Synchronous Reluctance Motors (SynRM). Such motors are capable of operating in two different synchronous speeds without a drive unit, based on the pole-changing concept. The motor is started, accelerated and decelerated from one speed to another using squirrel cage bars. Thus, this paper aims to optimize the cage bars geometry to ensure a smooth starting as well as a smooth transition from one speed to another without compromising the steady state performance of the motor. The Particle Swarm Optimization (PSO) algorithm coupled with Finite Element analysis (FEA) is used to optimize the motor's dynamic and steady-state performance. FEA is employed to generate data for training a system, in particular an Artificial Neural Network (ANN), which models the objective functions. Then, the objective functions are minimized via PSO.