Study On The Water Level Robust Control Of Steam Generator Of Nuclear Power Station

As a key equipment for connecting primary circuit with second circuit,the nuclear power station steam generator is a hub for the operation of primary circuit and second circuit.And the water level stability of nuclear power plant steam generator is of great significance for the safe operation of nuclear power plants.Because the steam generator is a complex system with nonlinearity and a large time delay,there is a complex internal thermal process in the course of operation.Besides,there is parameter shift of the control object in the control system caused by load conditions change,equipment aging and noise interference.It is very difficult for the traditional control system designed by the linear approximation model to meet the requirements of nonlinear uncertain high robustness and the fast tracking control of the water level object.Based on the Irving's model of steam generator water level,because of its global optimization feature,the genetic algorithm is used to optimize the parameters of the PID controller so that the PID controller designed for all operating conditions can be obtained.Simulation analysis of the steam generator water level was carried out under the conditions of step change,steam flow step change and the addition of external interference.At the same time,based on the theory of sliding mode control,the traditional sliding mode controller,dynamic sliding mode controller and improved sliding mode controller were designed.And the water level control effects of above controllers were analyzed by simulation test,respectively.The water level can be tracked and controlled under all operating conditions by PID controller,but the water level stability time is about 1200 s.When there is a step increase in the steam flow,the overshoot of water level exceeds 50 mm,and the water level stabilization time increases to 1500 s when there is noise interference.Control effect is not ideal.Compared with the PID controller,the sliding mode controller can greatly improve the response speed of the water level regulation,and the water level stabilization time is within 600 s.But there was a high frequency buffeting phenomenon of the feeding water flow when traditional sliding mode control was adopted.The continuous control of input(feeding water flow)can be achieved when discontinuous item in the sliding mode control was added into the derivative term of control input by adding control input into the dynamic sliding control in the design of sliding surface,thus improving the high frequency buffering of the feeding water flow phenomenon.But when above method was used,there is a large flow of water flow mutation value about 60 kg / s,a certain degree of buffeting phenomenon of the water level control in the presence of steam flow step disturbance,and a large water level overshoot in the presence of noise interference.In order to overcome the shortcomings of buffering phenomenon and large overshoot of the sliding mode controller,the coupling proximity term in the sliding surface was added.Finally the water level stabilization time is about 600 s,and the water level of the buffeting phenomenon was effectively suppressed.Besides,the steam flow step disturbance was also inhibited and there is a good anti-interference.Finally,through the analysis of the water level step response of the control system and the maximum overshoot and mean square error of the water level under the condition of noise interference,the simulation results showed that the maximum overshoot decreases with the increase of the operating power,and the maximum overshoot is the smallest when the water level is controlled by improved sliding mode controller.The mean square error under the control of four controllers all were less than 0.2.The control system showed certain control stability and good control effect.The noise reduction ratio ?n of the improved sliding mode controller is less than 0.01,and the control effect ratio is less than 16.The improved sliding mode control system under the uncertain model was tested to be robust to the parameter perturbation and noise interference.