Rotary Kiln Calcination Zone Temperature Soft Measurement Of LiESN

Rotary kiln is widely used in the production process of many industries such as metal smelting,chemical industry,building materials,refractory materials,etc.,and plays an important role in the production process of these fields.Rotary kiln is the core equipment in the entire production process.The temperature of the calcination zone reflects the burning state of the material in the kiln,which directly affects the quality and yield of the product.However,in the current production process,because the temperature of the calcination zone in the rotary kiln is high,there is no way to directly measure the temperature change in the kiln.This dissertation proposes a soft measurement approach based on the modeling of leaky integral echo state network.The specific research work is as follows:(1)The main research object of this dissertation is the calcination process of rotary kiln that is affected by many variables,and there are serious coupling and interference problems between these variables.The method of principal component analysis is used in this dissertation to reduce the dimensionality of the collected data.The coupling problem between variables is eliminated via the dimensionality reduction for high-dimensional space and the complexity of the model input problem is reduced.(2)An improved echo state network is proposed.The leakage integral type echo state network is based on the standard echo state network,and an additional leak rate parameter is added to the reserve pool state equation.It further improves the short-term memory capacity of the reserve pool,enhances the performance of the network,and raises the predictability and adaptability of the network.In the calculation of output weights,Owing to the linear regression algorithm is easily causing the problem of "ill-conditioned" solutions in the reserve pool,a ridge regression algorithm is proposed to eliminate the problem of "ill-conditioned" solutions in the process of solving the linear regression algorithm by applying regular coefficients.(3)Aiming at the problem that the internal parameters of the network reserve pool are randomly generated rather than optimal parameters,a binary particle swarm algorithm is used to determine the parameters in the reserve pool: the size of the reserve pool(N),the input pool shrinkage factor(IS),the reserve pool spectral radius(SR),the reserve pool sparsity(SD)is optimized.An optimized leaky integral echo state network model is obtained.(4)Simulating the data after dimension reduction processing by principal component analysis,and using the method of this dissertation,the standard echo state network optimized by particle swarm optimization and BP neural network to predict the temperature of the rotary kiln calcination zone,the simulation results show that the proposed method has higher prediction accuracy.