Short-term load forecasting based on the grid method and the time series fuzzy load forecasting method

Short-term load forecasting refers to daily and weekly load forecasting, which plays a vital role in ensuring security and high quality of power system´s economic operation for power generation planning. In short-term load forecasting, load periodicity with the fixed regularity is generally analyzed and random factors like weather are taken into consideration to improve the accuracy of the forecast, however, when large fluctuations of load occur due to failure of high energy-consuming load, there would be a certain forecasting deviation and accuracy would get decreased. To make up for the above deficiencies, firstly the total load is classified into high energy consumption load, general energy consumption load, low energy consumption load and other load according to corresponding power load proportion in the total load, then fuzzy time series method is applied to select similar days of all categories of load, finally grid method and neutral network method are implemented to take different random factors into account to forecast all types of load. Based on those, short-term load can be acquired through superposition of forecast results of all types of load within a same time period. Finally, historical load data of a certain region are applied to load forecasting to verify the feasibility of the classification, and forecasting results demonstrate that the method can improve accuracy effectively.