Short-term load prediction method based on long short-term memory network

Abstract

The invention discloses a short-term load prediction method based on a long short-term memory network, and the method comprises the steps: collecting short-term load data, taking the collected short-term load data as a training sample, and carrying out the VMD decomposition, and obtaining n random component IMFs; for the n decomposed components, constructing a corresponding WOA-LSTM model for each component, optimizing LSTM parameters in the models by using WOA, and finding out the most suitable parameters; training the component data IMF by using the LSTM with the optimal parameter to obtain a prediction result of each component; based on the LSTM models constructed after respective optimization, the LSTM models are used to measure the short-term load data, and each LSTM model outputs a corresponding prediction result; and adding all prediction results to obtain a final load prediction result.

Description

technical field [0001] The invention relates to the technical field of short-term load forecasting of power systems, in particular to a short-term load forecasting method based on a long short-term memory network (LSTM). Background technique [0002] Load forecasting is an important basis for energy planning, economic operation and energy management, and usually includes long-term load forecasting, mid-term load forecasting and short-term load forecasting. The characteristic of short-term load is that it will be greatly affected by factors such as weather, equipment conditions, and major social activities. Therefore, it is difficult to accurately predict short-term load. The characteristics of industrial, civil, and public utilities are very different, and the power load often has great fluctuation and seasonality due to changes in weather. Accurate prediction of the power load is the basis for the power system to formulate capacity expansion, operation, maintenance, etc. b...

AI Technical Summary

Problems solved by technology

[0005] In order to solve the deficiencies in the existing technology, this application proposes a short-term load forecasting method based on the long-term short-term memory network (LSTM), by introducing the whale algorithm to optimize the parameters of the long-term short-term memory network and the variational modal signal decomposition technology Decomposing load data can effectively solve the problem of poor short-term load forecasting effect in the prior art using neural networks; at the same time, introducing VMD decomposition can decompose nonlinear and non-stationary data into several linear and stationary data

Images