摘要
针对磨机运行过程中的非线性、大惯性、随机干扰大,常规PID控制不能取得很好控制效果的问题,提出用改进的RBF神经网络智能控制方式来控制磨机负荷。根据磨矿工艺流程和操作经验,利用改进的RBF神经网络构建在线辨识的磨机控制系统模型,解决了磨机控制系统难以建模的问题。结合自寻优控制方法,自动寻找磨机最佳负荷,减少磨机负荷的扰动,使磨机负荷维持稳定。实验结果表明,该控制方法能够很好地适应外界因素的变化,消除运行过程中的干扰,增强磨机系统的鲁棒性,使磨机保持稳定运行,提高磨矿效率,改善磨矿分级效果。
Aimed at the problems of mill in running process, including nonlinearity, large inertia, strong random interference,and poor control effect＂ of conventional PID control, an intelligent control method of improved RBF neural network was pro- posed to control mill load. Based on the technological process and operation experience of grinding, the improved RBF neural network was used to build the online identification model of mill control system, so as to solve the problem that the mill control system was difficult to be modeled. Combined with the self optimizing control method, the optimum load of mill was automati- cally searched, the disturbance of mill load was reduced, and the mill load was kept stable. The experimental results showed that this control method could well adapt to the change of external factors, eliminate the interference in the operation process, enhance the robustness of the mill system, maintain the stable operation of mill, as well as improve the grinding efficiency and classification effect.
出处
《矿业研究与开发》
CAS
北大核心
2018年第2期89-94,共6页
Mining Research and Development
关键词
RBF神经网络
磨机负荷
最佳逼近
全局最优
磨矿分级
RBF neural network, Mill load, Optimal approximation, Global optimum, Grinding classification
