Control and Identification of Mould Bath Level Using Neural Networks

In continuous casting, the stability of the mould bath level is an important index for quality of bloom. Mould bath level stability is strongly connected to the internal and external of product quality. Break out problem is also directly correlated with the mould bath level oscillation in the mould particularly for the bloom with a little section. In this case the conventional PID controller is insufficient to obtain high closed loop stability. We consider in this work an application of neural network to closed loop control of the mould bath level in continuous casting, learning based on the process data is used to model identification. Simulation results show that the neural approach improves the mould bath level stability than the conventional PID controller.