基于MI-IDBO-LSTM的SCR脱硝系统出口NOx体积分数预测OA

To address the issue of large prediction errors in SCR outlet NOx concentration caused by significant time delays and multiple disturbances in coal-fired power plant SCR denitrification systems,a prediction model of NOx concentration at the outlet of SCR based on MI-IDBO-LSTM was proposed.The time delay estimation of each input variable was completed by using Mutual Information(MI),and the time series of the data was reconstructed.Based on the processed data,the prediction model was established by using LSTM.The dung beetle optimizer(DBO)was improved by introducing the Tent chaotic map,adaptive weight,and the secretary bird optimization algorithm integrated with adaptive weight,so as to enhance its optimization ability.IDBO was used to optimize the key parameters of LSTM to improve the prediction accuracy of the model.Based on the historical operation data of SCR denitration system in a domestic 350 MW coal-fired power plant,the simulation results of IDBO-LSTM were compared with those of LSTM and DBO-LSTM.The results showed that IDBO-LSTM achieved the optimal performance among all models,with a mean absolute error of 0.453,a coefficient of determination of 0.976,and a root mean square error of 0.621.Experiments showed that the prediction model based on MI-IDBO-LSTM could achieve accurate prediction.