An Adaptive Forecast-Based Chart for Non-Gaussian Processes Monitoring: With Application to Equipment Malfunctions Detection in a Thermal Power Plant

In order to ensure power quality and keep supplying power in a thermal power plant, early detection of equipment malfunctions is a critical issue. This study attempts to develop an adaptive forecast-based chart so as to enhance the fault detectability in a thermal power plant. In the proposed monitoring statistic, the exponentially weighted moving average is adopted to preserve the information of past observations. Simultaneously, independent component analysis (ICA) is used to extract non-Gaussian information. The advantages of the proposed statistic include the fact that it is capable of monitoring non-Gaussian processes, the detection of small process shifts is improved, and the traditional ICA chart is a special case of the proposed one. The efficiency of the proposed method is verified by a simulated process and a real case of thermal power plant of Taiwan Power Company. Results demonstrated that the proposed method outperforms conventional monitoring methods, especially for detecting small process changes.