Predicting method for short-term load of power system

Abstract

The invention discloses a predicting method for short-term load of a power system. Active power of load data is used as a data basis; through dividing the data decomposing frequency to every stable frequency component from high to low, and then predicting a low frequency component and a medium high frequency component by a multiple linear regression and LSTM neutral network; finally, predicted results are overlapped to obtain a complete predicting result; thus the shortcoming that a traditional load predicting method is hard to effectively predict local features of the high frequency ; meanwhile, the change trend of the load can be exactly predicted, the predicting effectiveness under a complex condition is improved.

Description

Technical field [0001] The present invention belongs to the technical field of power load forecasting, and more specifically, relates to a method for forecasting short-term load of a power system. Background technique [0002] With the development of the electricity market and the gradual increase in user demand, the safe and economic operation of the power grid has become critical. Accurate short-term forecasting of power load can effectively ensure the safe operation of the power grid, reduce power generation costs, meet user needs, and improve social and economic benefits. Since the production, transmission, distribution, and supply and consumption of electric energy are almost completed at the same time, electric energy, as a relatively special energy source, is difficult to store in large quantities. This makes the power generation output of the power system consistent with the changes in system load at any time to achieve dynamic balance. Satisfy the relationship between s...

AI Technical Summary

Problems solved by technology

In order to fully mine data information and improve prediction accuracy, traditional mathematical models often use methods of increasing data dimensions or subdividing data, which can achieve fast and accurate predictions under certain conditions, but when the data dimension is low and the amount of data is small At that time, for complex and changeable power loads, due to its random and non-periodic effects, it is difficult for traditional mathematical models to obtain effective prediction accuracy

Although artificial intelligence-based algorithms can achieve higher prediction accuracy in comparison, they often have the disadvantages of difficult establishment of hyperparameters, high resource consumption, and relatively slow calculation speed. For low-dimensional time series data, due to its The mutual influence between local features makes it difficult for the model to clearly find out the rules and establish rules from the data

Images