Research on Improved Spatial Power Load Forecasting Method Based on Land Utility and Development Time

Abstract

Traditional spatial load forecasting method involves large volume of information, high model complexity. As a result, prediction accuracy and speed are difficult to guarantee. Considering land utility and development time, this paper proposes an improved spatial load forecasting method based on the logistic regression. In this improved method, regional load forecasting problem is splited into two subproblems: parameter estimation of gridded partition and integrated forecasting of logistic model. Parameter estimation scheme is based on development speed and the median year. The complexity of load forecasting is reduced by maximum likelihood estimation of the improved logistic regression. Simulation result shows that proposed method (PM) is superior to the traditional method (TM) in simplifying calculation complexity and improving prediction accuracy. PM can be effectively applied to regional load forecasting.