Experimental study on decoupling control of multi-motor synchronous based on an optimized neural network inverse

Aimed at complex multi-motor composed of three motors, difficulty of control by traditional methods such as PID are obvious. An improved control method for three-motor synchronous system which is taken as the research object based on RBF NN inverse is put forward. and the three-motor system can be linearized in series with NN inverse. Here, the inverse can be constructed by combining the RBF NN with an integrator. according to the performance index given, RBF NN neurons are optimized through growing or pruning. In the end, the tested platform based on Siemens S7–300PLC is designed. The performance of decoupling controller using inverse of improved RBF NN is verified by this experiment. The results show that velocity and tension can be completely decoupled under the stabilization. Compared with PID controller, superior static and dynamic performance are obtained by the proposed algorithm.