关键词: 卷积神经网络, 自适应, 监督预测, 热能循环, 预测控制

Abstract:  In order to improve the precision and stability of thermal cycle control in power plant, a new model based on convolution neural network was proposed. Firstly, according to the thermal cycle control problem, design the adaptive thermal cycle learning model using convolutional neural network(CNN), the problem that the identification model can not be adjusted adaptively according to the actual working condition is solved, improve the prediction accuracy of the model; Secondly, the state feedback adaptive controller is designed for the designed control system, and the asymptotic stability of the controller is proved, which provides a theoretical basis for the application; Finally, the simulation test on the boiler and turbine combined cycle control in the power plant shows that the proposed method is superior to the traditional PID control algorithm and the generalized predictive control algorithm. And the convergence rate is relatively faster, with better control performance.

Key words: Convolutional neural network, adaptive, supervised prediction, thermal cycle, predictive control