Research of temperature control of hot blast furnace based on RBF neural network tuning

Abstract

In order to improve the combustion efficiency of the hot blast stove and improve the automation degree of the hot blast stove temperature control system, a PID control strategy based on RBF neural network tuning is proposed. First, through the combination of the RBF neural network algorithm and the incremental PID controller, the powerful self-learning ability of the neural network is used to adjust the parameters of the incremental PID. Then, based on the conventional hot-blast stove temperature control system, the outer loop was changed to PID control using RBF neural network tuning. In the hot-blast furnace temperature control system, the inner ring takes the opening degree of the gas valve as a variable, and the outer ring takes the dome temperature as a control variable. The improved cascade control is used to optimize the combustion of the hot-blast stove. Matlab simulation analysis and practical application results show that the PID control curve set by the RBF neural network has almost no overshoot, and the anti-interference ability of the system is increased by 50% compared with the traditional PID control. Compared with the traditional manual control, the proposed control strategy makes the original system's ability to suppress interference significantly stronger and more robust. It has good research and application value in hot air furnace temperature control.

Keywords

• control system simulation technology
• hot blast stove
• temperature control
• rbf neural network
• pid incremental control
• conventional pid control