Static and transient performance prediction for CFB boilers using a BP neural network

摘要

This paper discusses the use of neural netowrks for prediction of static and transient performances of Circulating Fluidized Bed (CFB) blers.In both cases,the Back-Propagation neural Network (BPNN) is used successfully to prediect the performances of the boilers.The selection of the topology and training parameters for the BPNN is also touched upon.Good perform-ance predictions are essential for the operation guide of CFB boilers.Besides,the good transient performance prediction is crucial for model predictive control of CFB boilers.