System identification and model-free adaptive control of once-through steam generator in nuclear power

Due to the increasingly complexity of the nuclear power device system, the demands for its performance and security become rather high. Once through steam generator(OTSG) has been widely researched and applied because of its practical advantages like simple structure and small volume, good static performance and mobility. It can produce superheated steam, can work without dehumidification device, and can improve the thermal efficiency of a system. However OTSG has evidently shortages on tight coupling, nonlinearity, large hysteresis and strong interference. That makes its controller designing faces with great challenges. The model of `feed value-steam pressure' channel is established by system identification method. Model Free Adaptive Controller is independent from the concreting mathematical model of the control plant. It contains a universal model and some control modules. The controller will model and control the plant after obtaining new observation data. With the controlling continues, the `generic model' will be more accurate for each time. Thus the performance of the controller will be improved. There are many successful application of MFAC in some complicated control processes. It's necessary to improve the calculation and control precision of it. For this practical reason, a new type of OTSG Controller based on model-free adaptive technology is proposed in this paper Compared with the simulation result of PID, the convergence rate of MFAC is evidently faster than PID's. In brief, MFAC has obvious superiorities in tracking, adapting, anti-interference and overcoming large hysteresis when compared with PID controller.