Research On Intelligent Control Method Of Ceramic Shuttle Kiln Based On Deep Learning

Ceramic shuttle kiln is a nonlinear system with high degree of nonlinearity,serious hysteresis and disequilibrium of space temperature.At present,the control method of ceramic shuttle kiln is very backward in the operation process,and the control precision of furnace temperature fluctuates obviously,so that the quality of ceramic products produced is uneven.Therefore,it is necessary to study advanced intelligent control methods for ceramic shuttle kiln,so as to improve the temperature control effect of kiln and the quality of ceramic products.This paper combines the methods of deep learning,fuzzy control and PID control,and proposes an intelligent control method of ceramic shuttle kiln based on deep learning,thus,a new method is provided for the control of ceramic shuttle kiln.The main work and innovation points of this paper:1.According to the structure of ceramic shuttle kiln and the theory of heat transfer,the heat transfer model of ceramic shuttle kiln was established,and the optimal heating curve was fitted,which provided the basis for temperature control.By using the powerful data representation ability of the deep belief network to set the parameters of each link of the PID algorithm,this paper proposes the PID parameter optimization control strategy based on the deep belief network for ceramic shuttle kiln.The simulation results show that the proposed method is effective and feasible.2.Combining the advantages of fuzzy theory and deep belief network,this paper proposes a PID parameter optimization control method based on fuzzy deep belief network for ceramic shuttle kiln,and designs the corresponding controller structure.The simulation results show that this method is effective.3.For the parameter estimation problem commonly existing in deep neural networks.In this paper,an improved Levenberg-Marquardt algorithm is used to deal with parameter estimation of fuzzy deep belief networks,which greatly alleviates the data computing dimension and storage space,reduces the computational amount and training time of the neural network,and improves the generalization ability of the network,and thus optimizes the PID parameter optimization control method of the fuzzy deep belief network.Simulation results show that the proposed method is feasible and effective.