Prediction of Main Steam Pressure of Power Plant Boiler Based on CEEMDAN Combined with WTD-XGBoost-LSTM
A method for the prediction of main steam pressure in power plant boilers is proposed,which is based on complete ensemble empical mode deocmposition with adaptive noise(CEEMDAN)and using wavelet threshold denoising(WTD)for noise reduction of the high-frequency component IMF1,followed by prediction using extreme gradient boosting(XGBoost)and long short-term memory(LSTM)network respectively,according to the fluctuating form of the component.Finally,the main steam pressure prediction is realized by su-perimposing the output values of each model.Multiple comparison models are selected and the prediction results with real data.The re-sults show that the model proposed is better than other models in terms of time lag and prediction accuracy,and has strong significance for engineering applications.
power plant boilerpressure predictionCEEMDANXGBoostWTDLSTM
万方数据
维普
