Weather Sensitive Load Forecasting with Factor Analysis and Multivariate Nonlinear Regression Analysis

Abstract

Load forecasting is essential for the power system to meet the demand and supply equilibrium. Previous researches mainly focus on the improvement of forecasting algorithms. However, few researches consider the impact of weather condition which influences not only the average load per day, but also the maximum load, minimum load and the load peak-to-valley difference. Moreover, huge computational burden exists with precision verification of traditional algorithms. Therefore, this paper proposes a weather sensitive load forecasting model to comprehensively evaluate the impact of weather condition. A detailed analysis about load indexes is demonstrated, and the relationships between weather condition and load indexes are evaluated by factor analysis. Then, multivariate nonlinear regression analysis is applied to represent the forecasting equations. Based on the above analysis, a more effective and convenient precision verification is designed. Case studies are conducted to verify the proposed technique.