Application of hybrid strategy for NOx volume fraction prediction in cement kiln calcination
NOx volume fraction is a key environmental indicator that responds to the nitrogen emissions in the cement kiln calcination process.Due to the complex characteristics of the cement calcination process such as large noise,large time lag,and non-linearity.To cope with these characteristics,this study proposes an algorithm based on Complementary Ensemble Empirical Mode Decomposition(CEEMD),Mutual Information(MI)with entropy principle,Max-Relevance and Min Redundancy(mRMR),and Beetle Antennae Search(BAS)optimization Back Propagation Neural Network(BPNN)hybrid strategies for NOx volume fraction prediction.Firstly,the CEEMD method is used to decompose the original data and calculate the correlation to select the characterized high-frequency data using median average filtering for processing,which can effectively cope with large noise and provide modeling high-quality data,making the model accuracy improved.Besides,the MI of the entropy principle is used for temporal matching,and the matched sample set is selected using the mRMR algorithm for relevant variables,which reduces the coupling between variables and eliminates the influence of large time lags on prediction accuracy.On the other hand,the BAS algorithm and BP neural network are fused into one system,the former is used for system initial weight optimization and the latter is used for system weight training,both of which promote each other and solve the nonlinear working condition problem.Finally,the strategy is applied industrially,and the Root Mean Square Error(RMSE)and Mean Absolute Error(MAE)are only 0.302 4,0.205 9 and 0.215 3,0.201 3 in 25 900 industrial test samples,which are much better than other models,verifying the effectiveness of the strategy.The NOx volume fraction predicted in advance can provide reference for SNCR ammonia flow control,and adjust the ammonia flow control in advance to effectively stabilize NO,emission,while saving ammonia consumption,effectively reducing denitrification costs and improving enterprise quality development.
万方数据
