Design Of Control System And Research On Temperature Control Methods For Tunnel Kiln

Refractory is a very important basic raw material in metallurgical,chemical,petroleum,machinery manufacturing,silicate and other industrial fields.Tunnel kiln is an important process equipment for producing refractory materials.Only reasonable configuration of the relevant parameters in tunnel kiln can guarantee the excellent quality of refractory materials.With the development of the times,people’s requirements for the automation of the production process have been constantly raised.A consummate automatic control system is of great help to green production,energy consumption reduction,manpower saving and overall management.This thesis designs a tunnel kiln automatic control system and does research on the control method of tunnel kiln temperature based on the above purposes.In this thesis,a tunnel kiln automatic control system with monitoring screen is designed.System-based control level achieves data acquisition,data processing,loop control and other functions.System monitoring and management level achieves the furnace running status of real-time monitoring,data recording,status alarm,production management and other functions.The temperature of the sintering zone of tunnel kiln is a controlled object with complicated characteristics such as nonlinearity,time-varying and pure hysteresis.The traditional mechanism model and approximate line model largely lose characteristics of the system.This thesis analyzes the Least Squares(LS)identification strategy to autoregressive model with exogenous variables(ARX),and use Akaike’s Information Criterion(AIC)to determine the order of the system.RBF-ARX model which combined Radial Basis Function Neural Network(RBF)with the ARX model is introduced.The model preserves the time-varying property of the parameters and it is autoregressive.The thesis improves the optimization strategy of RBF-ARX model and optimizes the internal parameters of RBF network by Particle Swarm Optimization(PSO)algorithm.The simulation results show that the RBF-ARX model has better fitting accuracy to the historical data of the controlled object and can accurately estimate the temperature of the sintering zone of tunnel kiln.Because RBF-ARX model parameter optimization time is too long,it can not be used as the online prediction model of generalized predictive control(GPC).This thesis uses the Recursion of Least Square(RLS)with forgetting factor to identify model parameters.In the process of calculating GPC optimal control law,there are a large number of high-dimensional matrix inversion and Diophantine equations to solve.The calculation is complex.In this thesis,an alternative strategy is adopted,which is to use the DE-PSO hybrid optimization method based on PSO and Differential Evolution(DE)to calculate the optimal control law.The simulation results show that the performance of DE-PSO algorithm is better than that of single DE algorithm and PSO algorithm.Based on DE-PSO-GPC,the RBF-ARX model control task can be finished and the expected value changes can be tracked quickly.