基于Widrow-Hoff算法的轧机单独传动系统同步控制

Synchronous control of single drive system of rolling mill based on Widrow-Hoff algorithm

为了提高轧机单独传动系统中电机同步速度的精度,提出了基于Widrow-Hoff学习算法的交叉耦合同步控制策略。该策略将速度同步误差输入到同步控制器中,通过权值的变化规律调整权值,直至达到性能指标的要求。由仿真结果可以看出:与PI并行加补偿同步控制系统、PI交叉耦合同步控制系统相比,该方案的抗干扰能力最强,跟踪响应速度最快,同步精度最高,达到速度同步的时间最短,而且速度同步误差小于0.0001 rad/s。

Aiming at the problem that the precision of speed synchronization for motors driving the upper and lower rolls is not high in the single drive system of rolling mill,taking DC speed control system as the research object,a neural network cross coupling synchronization control strategy based on Widrow Hoff learning algorithm is proposed.In this strategy,speed synchronization error is the input of neural network synchronization controller.According to the weight constantly adjust the change rule of weight,until to achieve performance index.The simulation results and analyses show that in this control strategy,the speed of tracking response is the fastest,the anti-disturbance ability is much stronger when the steel is rolled,and the time required for speed synchronization is shorter,compared with the PI synchronization control system including parallel compensation and the PI cross-coupling synchronization control system.The synchronization precision is high,the speed synchronization error can be stable below 0.0001 rad/s.