Research On Cement Calciner Outlet Temperature Forecasting Based On Neural Network

With the large-scale development of cement equipment,it is the development direction of China’s cement industry to improve the traditional cement production mode with pre decomposition technology.Precalciner is the core part of precalciner system,which undertakes the tasks of pulverized coal combustion,gas-solid heat transfer and carbonate decomposition.The effective decomposition of carbonate is an important factor that restricts the quality of cement,and its effective decomposition requires a relatively stable temperature,so the temperature control of calciner is very important for the thermal distribution of the whole pre decomposition system and the stability of thermal system.In this thesis,based on the research status of temperature prediction technology of cement calciner at home and abroad,firstly,based on the new dry cement pre decomposition process,the main characteristics of calciner temperature control system,such as nonlinearity,coupling,large inertia,disturbance and time delay uncertainty,are studied,and the shortcomings in the process of temperature control are analyzed.This thesis discusses the main factors that affect the decomposition rate of raw meal and the stable operation of the temperature control system of the calciner,obtains the main characteristic variables that affect the temperature of the calciner,leads to the principal component analysis(PCA)and the kernel principal component analysis(KPCA)method,selects the main characteristic variables,reduces the dimension of the data,and puts the reduced variables into the neural network,so as to put forward the neural network model based on bidirectional long-term and short-term memory(Bi LSTM),The model can effectively utilize the ability of long-distance dependence on information of sequence data,fully excavate the hidden rules behind the sequence data,normalize the data,and then use the mean square error and mean absolute percentage error as the evaluation index.Finally,the optimized model is used to predict the outlet temperature of calciner.Through empirical analysis,the thesis found that the mean square error and average absolute percentage error of the combined model based on KPCA-Bi LSTM in the test set were compared and analyzed with Bi LSTM,KPCA-LSTM,and PCA-Bi LSTM,and found that the combined model of KPCA-Bi LSTM both the mean square error and the average absolute percentage error obtained on the test set are lower,and the prediction fitting curve fit is much better than other methods.It shows the feasibility of the combined model based on KPCA-Bi LSTM.At the same time,the prediction of the outlet temperature of the calciner can also provide the cement industry with an optimized control method for the outlet temperature of the calciner,which can effectively reduce the intervention of other influencing factors,and make the temperature of the calciner more stable during decomposition.The role in the calciner is more thorough,effectively improving the quality of the finished cement.