Research on BOF steelmaking endpoint control based on neural network

BOF (Basic Oxygen Furnace) steelmaking is a very complex industry process and the endpoint control is an important operation in the later period of BOF steelmaking. Owing to the environment is so harsh and the temperature is so high that it is very difficult to measure the temperature and contents continuously and accurately, therefore, we can not use normal method of process control. This paper makes use of the advantage of the RBF (Radial Basis Function) neural network (NN) that has quick convergence velocity and can avoid the relative extreme value effectively; the prediction model of endpoint steel temperature and the endpoint carbon content is established. On the basis of prediction model, the dynamic endpoint control method based on BP NN is proposed to calculate the need of coolant and blown oxygen during the reblowing. The simulation result shows that the new method overcomes the shortcoming that the control model based on the heat balance and the oxygen balance is not accurate in the traditional method, and increases the endpoint hitting ratio.