The switched reluctance motor has many advantages such as simple construction, large starting moment, and capable of being used in systems requiring high speed. However, the disadvantage of this motor is that the ripple torque is large and difficult to control. In order to apply the control methods, the motor's mathematical model needs to be precisely defined. Unlike conventional motors, the mathematical model of the switched reluctance motor requires knowing the flux characteristic, which can only be accurately determined based on experimental measurements. The experimental measurement process requires a real equipment system, complicated installation process, not suitable in research, design, and simulation of control algorithms. To solve this problem, the paper proposes a mathematical function of the magnetization characteristic, using a neural algorithm to check the accuracy of the proposed function. The magnetization characteristic function for the switched reluctance motor is built considering the influence of mutual inductance phases and magnetic saturation, magnetization function is used in the development and evaluation of control algorithms for switched reluctance motor.

Keywords: Switched reluctance motor; magnetization; identification; finite element method; neural.