Short-term Power Load Forecasting Based on Grey Relational Analysis and Support Vector Machine

Short-term power load forecasting is an important guarantee to ensure the smooth and efficient operation of power systems, and an important basis for building new digital and intelligent power systems. Given that short-term power system load is affected by various factors (e.g., climate, time), power system load has strong randomness and volatility while being periodic. Hence, the traditional power load forecasting method is no longer applicable. To improve the accuracy of short-term power load forecasting, this paper proposes a support vector machine (SVM) short-term power load forecasting method based on grey relational analysis and K-means clustering. First, similar days in historical days are extracted by using the grey relational analysis method to form a rough set of similar days. Second, the rough set of similar days is classified by K-means clustering, and the final set of similar days is obtained. Third, SVM is trained to determine the final predicted daily load. Lastly, the proposed method is verified by the actual electricity consumption data of a city in China, and the results show the effectiveness of this method.