带钢对边纠偏系统的参考模型神经网络控制

Model Reference Neural Network Control of Steel Strip EPC System

为了消除带钢对边纠偏控制系统中由于不考虑检测器安装位置与卷取机卷筒之间的带钢延时而引起的系统模型误差,本文重建了系统数学模型,并提出了参考模型神经网络算法,用该算法对纠偏控制系统进行了研究。仿真结果表明:在参考模型参数发生小幅度变化的情况下,参考模型神经网络算法的控制效果仍保持良好,具有较好的鲁棒性。

The mathematical model of steel strip EPC system was rebuilt to eliminate model error, which is the result of neglecting the time-lag caused by detector mounting position. Based on the new model, model reference neural network control method was developed. The result of simulation indicates that model reference neural network method can achieve good performance under small parameter changing of reference model, which shows good robustness of the method.