Application of Independent Component Analysis on the Data Validation of Thermodynamic System in Power Plant

A data validation method based on the independent component analysis (ICA) was proposed, which was applied to the thermodynamic system in power plant. The data detection and reconstruction can be implemented on the base of the separating of independent source signals from the measured data. The ICA algorithm as well as the IC selection, initialization and thresholds were analyzed. Then the framework of the method was presented, which can be used for online data validation since it has less computing complex. Then, 21 data sets from a 300 MW unit were validated by the proposed method. The results revealed its validity to the detection of bad data. The ability of signal separating of these ICs was also analyzed.