钢包精炼炉的电极系统智能建模及控制

Intelligent Modeling and Control Strategy for the Electrode System in Ladle Fur- nance

针对钢包精炼炉电极控制系统具有非线性、时变、模型不确定、大滞后、多输入多输出耦合的特点,提出一种基于神经网络实时在线辩识的内模控制方案.控制器采用神经网络解耦,将混沌机制引入到BP算法中,用以加快学习的收敛速度.仿真结果证实了控制策略的有效性.

In accordance with such characters of the electrode control system in ladle furnace as the high non-linearity, time-variant, uncertainty of the model, output response time delay serious, and multivariable input and output coupling, an internal model control strategy based on real-time identification on line by neural network was presented. The control strategy applies neural network decoupline control and the chaos algorithm to the improved BP algorithm and speeds up the training of neural network. The validity of the control strategy is verified by simulation analysis.