The Research On The RBF Neural Network Setting PID Controlling PMLSM

NC machine are mainstream equipment of manufacturing machinery. It has become the consensus of people from all walks of life that we use NC machine and equipment to upgrade the level of traditional manufacturing industries. Especially with the improving technical requirements for NC machining achieve high-speed and ultra high-speed, precision and ultra-precision, high-speed response capability of the linear servo feed system - namely, "zero transmission" approach arises, linear drive technology positioning has also been widely used in the field of precision.The article researches the linear motor direct-drive system, and we select permanent magnet synchronous linear motor for the control object. In view of the contradiction between the linear servo system's stability and rapidity, we mainly carries on the optimization design of the control plan, and realizes the fast track to the system which assigns and greatest degree's suppression to the parameter variation and resistance perturbation, increases the linear servo's stable state precision.The linear motor servo system compares with the tradition"the rotating motor + ball screw"feed way, although eliminating some adverse effects which the mechanical transmission chain brings, but actually increasing the difficulty of control. In the occasion of requesting high-precision and micro-feed, we must stand on a higher level, considering more perturbations and uncertainty factor to the feed motion's influence, otherwise, zero transmission will lose the significance which originally hoped. In order to realize the high-precision and high-speed control result, we need to seek the effective control strategy.Up to now, PID control is still the oldest and most vitality of basic control mode. The continuous development of computer technology has provided a good material base for the digitized PID control technology. Because the computer have a quick operating speed, high-precision, huge storage capacity, flexible programming and strong communication and so on, it is widely applied in different industrial control.Linear AC servo motor with non-linearity, coupling, load disturbance, time-varying uncertainty, hard to establish precise mathematical model, etc, all these bring difficulty to establish system model. The use of conventional PID controller can not achieve the desired control effect. Back-propagation network is one of common neural networks, but BP network fall easily into some disadvantage, such as local-minimum, slow training-speed, low efficiency. To some extent, the radial basis function network overcomes the shortcomings of BP neural networks, with a small amount of computation, fast convergence, no local minimum and so on. The combination of the RBF neural network and PID control form RBF neural network-setting PID control, will better , to some extent, the performance of traditional PID controller.Using MATLAB compile control procedure, we carried on the simulation experiment. The simulation result indicates that this optimization's RBF neural network setting PID control system comparing with the traditional PID controller's servo system has better track ability and robust performance. It shows that this control system using RBF setting PID is one effective method in the linear AC servo system.