Online Learning Algorithm-Based High-Efficient Control of IPM motors

An Interior Permanent Magnet Synchronous Motor(IPMSM) has a structure in which a permanent magnet is arranged inside the rotor, and reluctance torque can be used combined with magnet torque. It is easy to achieve high efficiency and has been widely used in home appliances, industrial equipment, automobiles, etc. This paper presents a learning method that optimizes the efficiency of IPMSMs by using a multilayer neural network structure without using identification methods or estimation methods even when the parameters of the controlled object varied. The effectiveness of the proposed method is verified by simulation in comparison with the conventional method.