A NOx Concentration Prediction Model Based on a Sparse Regularization Stochastic Configuration Network  

作　　者：Aijun Yan Shenci Cao 

机构地区：[1]School of Information Science and Technology,Beijing University of Technology,Beijing 100124,China [2]Engineering Research Center of Digital Community,Minstry of Education,Bejig 100124,China [3]Beijing Laboratory for Urban Mass Transit,Bejing 100124,China

出　　处：《Instrumentation》2024年第3期13-22,共10页仪器仪表学报（英文版）

基　　金：supported by the National Natural Science Foundation of China (62373017, 62073006);the Beijing Natural Science Foundation of China (4212032)。

摘　　要：For accurate prediction of nitrogen oxides(NOx) concentration during the municipal solid waste incineration(MSWI) process, in this paper, a prediction modeling method based on a sparse regularization stochastic configuration network is proposed. The method combines Drop Connect regularization with L1 regularization. Based on the L1 regularization constraint stochastic configuration network output weights, Drop Connect regularization is applied to the input weights to introduce sparsity. A probability decay strategy based on network residuals is designed to address situations where the Drop Connect fixed drop probability affects model convergence. Finally, the generated sparse stochastic configuration network is used to establish the model, and is validated through experiments with standard datasets and actual data from an MSWI plant in Beijing. The experimental results prove that this modeling method exhibits high-precision prediction and generalization ability while effectively simplifying the model structure, which enables accurate prediction of NOx concentration.

关 键 词：municipal solid waste incineration NOx concentration prediction stochastic configuration network sparse regularization 

分 类 号：X831[环境科学与工程—环境工程]