Research On Optimal Control Of Control System Based On Online Performance Evaluation

With the rapid development of my country’s economy and society and the continuous improvement of the level of industrial automation,the control loops of the production process layer have also become more and more complex.Under this background,the online performance evaluation of the control system is of great significance for improving the control performance and improving the production efficiency of the enterprise.This paper takes the deterministic index and random index of the control system as the main technical breakthrough point,and takes some important control loops of the thermal power plant as the research object to carry out the evaluation and optimization research on the thermal power plant system.The main contents are as follows:Firstly,for the deterministic performance evaluation of the control system,an online deterministic performance evaluation method of the control loop based on operating data is proposed,which uses the method of moving average filtering and setting the dead zone of the steady-state value to solve the problem of data glitches and fluctuations.overcome.Then,the IAE index under the internal model PID control law is used as the ideal reference value to compare with the IAE index of the actual control loop,and a new dimensionless evaluation index is given.The above methods are verified by taking the high water level control loop and primary air volume control loop of a power plant as objects.Secondly,for the stochastic performance evaluation of the single-loop system,a minimum variance index evaluation method based on the LR algorithm with forgetting factor is proposed,and it is compared with the traditional MV benchmark method.It is more stable and real,verifying its superiority.Aiming at the evaluation of the control loop of a thermal power plant,a calculation method of the minimum variance on-line evaluation index based on operating data is proposed,and the reheat steam temperature control loop of a thermal power plant is taken as the object to be verified.Afterwards,the MV index is extended to the evaluation of feedforward-feedback loop,cascade loop and multivariable system which are widely used in thermal power plants.For the cascade control system,its generalized minimum variance evaluation index is given.The evaluation and verification of the boiler feedwater control system of a thermal power plant is carried out.The results show that the randomness performance can be significantly improved after the feedforward control is introduced into the loop.Finally,combining the online performance evaluation of the loop with the control optimization,a comprehensive deterministic index is proposed as the objective function of the genetic algorithm,and the parameters of the PID controller for the desulfurization slurry pH value of a power plant are optimized online.It is confirmed that the tuning effect is better than the traditional method.better.The minimum variance control(MVC)method is introduced,and the generalized minimum variance control(GMVC)after it is extended to non-minimum phase systems is given.The above two control strategies are simulated and studied.In order to overcome the shortcomings of the control strategy that simply pursues the MV index,an online optimization strategy of cascade loop control driven by the minimum variance index is then proposed based on the improved virtual reference iterative tuning algorithm(FRIT).The steam temperature control system is used as the experimental object to verify the effectiveness of the method.