Comprehensive Performance Assessment And Optimization Of Thermal Power Plant Control System

Control system performance monitoring and assessment is an important part to ensure the production effectiveness and safety in the process of industrial and environment protection.With increasing scale of production and control loops,there is much difficulty to monitor control performance by operator.The control performance assessment has to be accomplished automatically to ensure efficacy.Nowadays,thermal power generation accounts for a large share of total electricity generation,about 75%.The equipment of thermal power generation is highly complicated and it is difficult to achieve the desired performance of the control loop.It is necessary to evaluate the performance of the control loop by certain methods,to guide the optimization of controllers.At the same time,due to a variety of internal and external factors,the performance of control loop will be worse after a period of operation.So,it is important to monitor the control performance.Therefore,in order to ensure a good operating condition,it is necessary to study the performance assessment and optimization of thermal power plant control system.Regarding the Control system performance assessment and optimization,the main work and contribution of this paper is concluded as follows:First,the classic performance assessment of the minimum variance benchmark is introduced regarding the Guassian disturbance in single variable system and extended to multivariable system.A performance assessment method based on equivalent open-loop transfer function in multivariable system is proposed.The problems in multivariable minimum variance benchmark,including difficult to calculate and too much restrain,are solved.Simulation experiments are made to testify the feasibility and accuracy of the proposed method.Second,the minimum information entropy performance assessment index is introduced regarding the system under non-Guassian disturbance.Rational entropy is brought in to modify the original index and the minimum rational entropy performance assessment index is proposed.Meanwhile,a data driven performance assessment index based on Hellinger distance is introduced.Simulations are made to compare the effectiveness of these three indexes.As a result,the Hellinger distance based index is more suitable in practical industry control loop,but not objective enough.Compared to minimum information entropy performance assessment index,minimum rational entropy performance assessment index is more accurate in telling the trend of system performance.Third,a dimensionless integral of the absolute value of the error performance assessment index is proposed to assess the deterministic performance of control system.Combining the index with stochastic index by sub goal multiplication and division method,a comprehensive performance index is generated.The comprehensive index serves as an objective function to guide parameters optimization in power plant unit coordinated control system controllers.The optimal parameters and the corresponding performance indexes of the auto disturbance rejection controller,the generalized predictive controller and the internal model PI controller are gained.The performance of these three controllers are compared and monitored.Simulation results show that the GPC has the best comprehensive performance and the best robustness in model parameter perturbation.Finally,regarding an actual thermal plant with unknown model parameters,a data driven comprehensive performance assessment method is concluded.And a performance assessment is made for the steam drum water level control loop in the third unit in Pannan Power Plant,Guizhou.