A regional network supply load fine prediction method based on two-stage reduction

Abstract

A regional power grid supply load fine prediction method based on two-stage reduction comprises the following steps that 1, historical data of regional power consumption loads, small hydro-thermal power, new energy power generation and regional mutual supply power and historical and prediction data of time-sharing zoning meteorology are collected; Step 2, decomposing the grid supply load into a combination of a regional power consumption load, small hydro-thermal power, new energy power generation and regional mutual supply power by adopting a two-stage reduction method; Step 3, introducing time-sharing partition meteorological information according to characteristics and influence factors of different types of loads, constructing different prediction models, and obtaining a prediction result of each type of load curve based on each type of load prediction model and a refined meteorological prediction value of a prediction day; And step 4, combining the load prediction curves of various types, and restoring to obtain a network supply load prediction curve of the prediction day.

Description

technical field [0001] The invention belongs to the field of power systems, and in particular relates to a refined prediction method for regional network supply load based on two-stage restoration. Background technique [0002] Short-term load forecasting of power system is an important basis for dispatching plan and operation mode arrangement. Its accuracy rate is not only an important index to measure dispatching operation level, but also a comprehensive reflection of dispatching work level. The research on short-term load forecasting of power system has a long history. At present, there are mainly the following types of load forecasting methods at home and abroad: classical method, traditional method (including time series analysis method, regression analysis method, gray forecasting method) method), emerging methods (including expert systems, fuzzy prediction methods, artificial neural networks and other artificial intelligence methods, and emerging wavelet analysis meth...

AI Technical Summary

Problems solved by technology

With the increasing temperature adjustment load of residents, the proportion of wind, light, water and other clean energy power generation continues to increase, and the proportion of meteorologically sensitive loads in the total load of the region continues to rise. Traditional load forecasting methods, when dealing with complex meteorological conditions and multi-load component coexistence network supply load forecasting, its accuracy will not meet the requirements

Images