Research on the Control Strategy of Passive Torque Servo System for Loading at High Frequency

Abstract

In view of the problem of difficult to suppress surplus torque and to obtain high servo accuracy at high frequency in the passive torque servo system(PTSS), the neural network PID control strategy with surplus torque compensation based on dynamic fuzzy neural network (DFNN) and double-stator motor is proposed. Firstly the model of the PTSS is built and the mathematical model of surplus torque is derived. Then the model of surplus torque is identified by DFNN. The surplus torque is estimated in real time and converted directly to the control signals of outer stator of double-stator motor which produces torque to compensate the surplus torque. Finally the parameters of the neural network PID (NNPID) controller are adjusted in real time according to the Jacobian information and system error. The Jacobian information is obtained from the online identification of PTSS by RBF neural network. The simulation results show that surplus torque model is close to the actual surplus torque system, that compensation control largely eliminates the surplus torque, that system performance is improved and that the control strategy is successful.