Design, implementation and speed estimation of three-phase 2 kW out-runner permanent magnet BLDC motor for ultralight electric vehicles

Abstract

Brushless DC motor has specifications such as high efficiency, high startup moment and silent running. Thanks to its low inertia, high torque/size and power/size ratio, it can be used in specially designed vehicles such as electric vehicles (EVs), spacecraft and submarines. As there is no brush and commutator that in its structure can cause arc forming, it can be used in fire-sensitive areas. In this study, a 2 kW three-phase out-runner permanent magnet brushless DC (BLDC) motor was designed to be used in ultralight EV. The size of the motor, the magnetic equivalent circuit and the electrical equivalent circuit parameters required for the BLDC motor design process were analytically calculated. The finite element method was then used to evaluate flux density, flux distributions, torque and motor efficiency and was approved for analytical design. The BLDC motor, which has about 89% efficiency, has been manufactured and mounted on an ultralight EV. Finally, the motor speed was estimated using a new robust hybrid metaheuristic model called artificial neural networks (ANNs) trained with particle swarm optimization (PSO) and radial movement optimization (RMO). A genuine and unconventional technique was used to examine the model's performance. That is, using three distinct input variables such as output torque, efficiency and output power, the output variable of motor speed was estimated. And then, the results were compared using the other three hybrid models. In all performs, it was seen that ANNs trained with PSO + RMO model achieved the most successful results with the lowest errors.