基于自适应律的RBF磨矿分级优化控制过程

Based on Adaptive of RBF Grinding Optimizing Control Process

针对磨矿过程中被控对象存在的大惯性、非线性、滞后性等特点,将球磨机简化为一个连续控制过程。根据磨矿特性和料级颗粒分布影响,采用连续磨矿采样数据确定的参数,提出一种基于自适应律的径向基(Radial Basis Function,RBF)神经网络控制系统。采用自适应律对RBF神经网络的初始权值进行优化,采用优化后的RBF神经网络辨识主控制回路被控对象。通过matlab仿真对比实验结果,验证了优化后的RBF网络控制更加逼近理想状态。

In view of the controlled object exists in the process of grinding characteristics of large inertia,nonlinear and hysteresis,the ball mill is simplified to a continuous control process.According to the characteristics of grinding and particle distribution,there is a control system of Radial Basis Function based on adaptive.Use this method to optimize the initial weights and use the optimized RBF neural network to identify the main control circuit controlled object.The optimized RBF network control is verified by MATLAB is more close to the ideal state.