A Data-Driven method for Lead-Lag analysis in Coal-Fired power plants using an improved DTW algorithm

The dynamic characteristics of coal-fired power plants during load change processes are essential for safe, clean, and efficient operation. However, existing approaches struggle to capture lead–lag relationships among key operational parameters, particularly during load change processes, where Dynamic Time Warping (DTW) often suffers from local alignment errors. This study proposes a data-driven method, FSSS-DTW, which integrates filtering, smoothing, standardization, shape-based feature extraction, and an improving step pattern to reduce alignment errors and improve temporal alignment accuracy. The effectiveness of the proposed method is validated using operational data from a 330 MW subcritical coal-fired power plant, where it demonstrates superior capability in extracting accurate lead-lag relationships and interpreting control-response dynamics. The results reveal that fuel input consistently lags behind steam flow during load changes, while high-pressure valve opening exhibits minimal lag. For energy flows, enthalpy changes across boiler heat exchangers lag behind energy input in sliding-pressure mode but lead in constant-pressure mode. The proposed method provides a scalable and interpretable tool for dynamic analysis, offering new insights into coordination mechanisms between turbine and boiler subsystems and supporting intelligent monitoring in complex energy systems.