Modeling And Optimal Control Of Double Reheat Steam Temperature Based On Wavelet Neural Network

China is a big country of coal consumption,coal will still occupy the main position in the energy structure of our country in the future.Power generation is one of the ways of clean utilization of coal.The construction of high-efficiency and low emission large capacity thermal power generation unit is the future development trend.Compared with the primary reheat unit of the same capacity,the double reheat unit has obvious advantages in power generation efficiency and emission index.In recent five years,China has put into operation several ultra supercritical double reheat units,and the research on the double reheat has become a hot spot.In this paper,the steam temperature system of the reheat unit is analyzed,and the simulation data of the unit is used for modeling and control.Wavelet neural network is used to identify reheat steam temperature system of reheat unit.For the problem that wavelet neural network may fall into local minimum value and the training result is not ideal and the model accuracy is low,the improvement measures of training algorithm are put forward,that is to optimize the network weight value through particle swarm optimization.For the problem that the effect of particle swarm optimization depends on the position of initial population,a method of generating initial population by momentum gradient method is proposed.The immune algorithm mechanism is introduced to solve the problem that it may fall into the local optimal convergence in the iterative process.In order to solve the problem of poor local search ability in the later stage of iteration,momentum gradient method is introduced for local fine tuning.Experiments are designed in MATLAB to verify the superiority of the improved algorithm.In this paper,the neural network model identification is studied deeply,the characteristics of different structure models are analyzed,and the appropriate network structure is selected.The influence of the historical input structure on the training time and the accuracy of the model is studied,the appropriate data structure is determined,and the neural network model of the double reheat system is established.The validity of the model is verified by comparing the model prediction output with the system output.The multivariable coupling characteristic of reheat system is analyzed,and the coupling PID controller is designed according to the characteristics of controlled object.Immune particle swarm optimization is used to optimize the parameters of the controller.By adjusting the fitness function of particle swarm optimization,the optimized parameters can meet the control requirements,and the control effect is tested.In the actual operation of the reheat unit,many control systems are PID control,but considering the stable safety and economic benefits of the unit in normal operation,it is not easy to change the controller parameters to optimize the system regulation quality.Through the above series of experimental simulation research,the system modeling and controller parameter optimization can be realized only depending on the system operation data,which provides a new idea for the optimization of the steam temperature control system of the double reheat system.