Electrical load prediction method and system, storage medium and electronic equipment

Abstract

The invention discloses an electrical load prediction method and system, a storage medium and electronic equipment, and the method comprises the steps: constructing a power network map according to a power network topological relation; preprocessing each power utilization load unit according to the historical power utilization load data, and calculating to obtain a first feature, a second feature and a third feature of the power utilization load of the power utilization load unit; and for each electrical load unit, comprehensively using the first feature, the second feature and the third feature of the electrical load as model training and prediction input, and training by using a time sequence model, thereby obtaining an electrical load prediction model corresponding to each electrical load unit. And loading the electrical load prediction model corresponding to each electrical load unit, and predicting the future electrical load. Therefore, according to the method for predicting the electrical load, the future load prediction of the electrical load unit can be accurately carried out, the prediction result evaluation index is greatly improved, and the robustness of load prediction is greatly improved.

Description

technical field [0001] The invention relates to the technical field of electric power, in particular to a method, system, storage medium and electronic equipment for electric load forecasting. Background technique [0002] At present, the model training methods of electricity load forecasting mainly include independent training and mixed training. Independent training refers to multiple independent electricity load units (such as residential meters, commercial meters, industrial meters, etc.), each electricity Load units conduct training based on their own historical electricity load data to obtain their corresponding electricity load forecasting models, and then use the corresponding models to predict future electricity loads. And mixed training refers to multiple independent power load units, using all their historical power load data to train together to obtain a comprehensive power load forecasting model, and then use this model to separately predict the power consumptio...

AI Technical Summary

Problems solved by technology

However, because the historical power load data of all units are used at the same time, and the correlation of power load laws of most units is very small, once the number of power load units is large, the fitting speed of model training will be very slow. It is even difficult to achieve the ideal fitting effect, resulting in unsatisfactory evaluation indicators of prediction results during actual load prediction

Images