Abstract Frequent peaking operations of coal-fired units, driven by renewable energy intermittency, impose significant stress on induced draft fans. This study develops an automatic monitoring technique for induced draft fans to enhance operational reliability. A PCCs-CNN-LSTM-AM (PCLA) model is proposed, including Pearson Correlation Coefficients (PCCs) for feature screening, Convolutional Neural Network (CNN) for local feature extraction, Long Short-Term Memory (LSTM) network for temporal dependency modeling, and Attention Mechanism (AM) for dynamic weight allocation. The model was trained on one year of operational data (26,352 samples) from 300MW, 600MW, and 1000MW units. Comparative experiments demonstrated PCLA’s superiority over benchmark models (XGBoost, PCCs-LSTM, PCA-LSTM, CNN-LSTM), achieving optimal performance in current prediction (MAE=1.5920, RMSE=2.0325, R²=0.9970). Ablation studies confirmed critical contributions from all four components. The deployed online monitoring system reduced induced draft fans’ failure rates by establishing dynamic health thresholds. Its implementation significantly improves equipment reliability and operational efficiency, providing an intelligent solution for coal-fired unit maintenance under peaking demands.
Xie et al. (Tue,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: