Torque Calculation Error Compensation for Reluctance Spherical Motor Based on Extreme Learning Machine

To reduce the torque calculation error caused by the analytical model of reluctance spherical motor (RSM), this paper proposes a novel calculation error compensation method. Firstly, the calculation error between the calculated torque obtained by analytical method and the simulated torque is analyzed. Then, extreme learning machine(ELM) is adopted to compensate the torque calculation error according to the error characteristics. The simulation results show that this error compensation method has high compensation accuracy, strong stability and good real-time performance.