Stack type self-encoding multi-model load prediction method and system based on LSTM

Abstract

The invention belongs to the technical field of load prediction, and particularly relates to a stack type self-encoding multi-model load prediction method and system based on LSTM. The method comprises the following steps: acquiring a data set, and preprocessing the data set; establishing a probability model of the charging start time of the electric vehicle and reconstructing a data set; constructing an LSTM stack type self-coding structure and training the LSTM stack type self-coding structure; and predicting a short-term load by utilizing the XGBoost model, and carrying out index evaluation. The system comprises a data set acquisition module, a preprocessing module, a probability model and reconstruction module, an LSTM stack type self-coding structure construction and training module,a prediction module and an index evaluation module. The method provided by the invention can consider the influence of the charging load of the electric vehicle, utilizes the original data to the greatest extent, deeply learns the internal features, and effectively improves the short-term load prediction precision.

Description

Technical field [0001] The present invention belongs to the field of load prediction, and specifically, the present invention relates to the LSTM stack self-coding multi-model load prediction method and system. Background technique [0002] With the development of smart grid, intelligent terminal measurement equipment has been popularized, and my country has installed more than 500 million smart meters. The index growth of power data and complexity is effective, and the huge data set will gain great value, the most representative of which is to use historical data to predict future load. Considering the increase in flexible load such as electric vehicles, heat storage, etc. . Therefore, it is possible to say that today's load prediction has become a problem based on multivariate timing data, how to effectively utilize large databases based on data mining technology, and improve load prediction accuracy is today research hotspot. In load prediction, short-term load prediction is a...

AI Technical Summary

Problems solved by technology

[0007] The present invention is to overcome the defect of destroying the hidden value of the data caused by manually extracting the characteristics of the influencing factors in the conventional forecasting method, and proposes a multi-model charge forecasting method and system based on LSTM stacked self-encoding, and its purpose is to solve the problem that the multi-variable time series data cannot be fully utilized problem, and the implementation of short-term load forecasting considering electric vehicles from a data mining perspective

Images