Regional Wind Power Forecasting Method Based On Temporal And Spatial Distribution Characteristics

With the rapid wind power technology development and growing wind power capacity, large-scale wind power connected to the grid has brought a lot of problems and adverse effects for power system. Intermittent and random characteristics have become a major restriction factor for the development. If wind power can be effective forecasted, it would be good for scheduling department to set out day operation mode and adjust scheduling plan, reducing the situation of abandoning wind and power rationing, and to improve wind power acceptation ratio for the whole system, which would make clean energy more efficient to use. Regional wind power forecasting can not only has good practical significance, providing wind power forecasting information of the whole grid to scheduling department, but also reduce prediction error to some extent. Therefore, regional wind power prediction is a subject worthy of further study.This paper proposed and implemented data preprocessing of wind power abnormal data, missing data and inconsistent data, and analyzed wind power volatility, complementarity and correlation. Based on wind power prediction error research status, wind power prediction error skewed distribution is proposed, and conditional distribution characteristics is analyzed, and wind power prediction error evaluation index is summarized and enriched. According to wind power correlation, spatial upscaling method was used as core algorithm of regional wind power prediction. In order to improve accurate of prediction, this paper preprocessed wind farm prediction power based on wind power conditional distribution characteristics first, divided the whole region into sub-regions and used equivalent capacity to consider wind farms that have strong correlation with sub-region wind power, but lower prediction accuracy. Aimed at practical application of the prediction system, this paper analyzed design principles and database design, determined forecasting system structure and development platform and analyzed functions of each module.