The Control Algorithm Research of the Warp Tension Based on Neural Network

This paper put forward to use neural network control algorithm to control warp tension, and introduced the single neuron control method and multi-neuron control method respectively. And first proposes a modeling method that fits non-linear system with multiple single neuron models, introduction of multi-model reference trajectory to acquire a new multi-model prediction control. Then use the SIMULINK of MATLAB to simulate. The transition time of system which based on multi-neuron PID control is about 0.5s, and there is no overshoot exists, system stability, and multi-neuron PID controller has a better control effect.