Power consumption complementation and prediction method and system for missing power consumption data

By combining generative adversarial networks and long short-term memory networks, the problem of missing electricity consumption data in low-voltage distribution networks was solved, achieving stable electricity consumption prediction and data recovery, and improving prediction accuracy.

CN122243025APending Publication Date: 2026-06-19GUANGDONG POWER GRID CO LTD DONGGUAN POWER SUPPLY BUREAU
0 Cites 0 Cited by

Patent Information

Application Number
CN202610229191.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-02-26
Publication Date
2026-06-19

AI Technical Summary

Technical Problem

In low-voltage distribution networks, due to unstable communication links, failure of data acquisition terminals, and other reasons, electricity consumption data is often missing. Existing technologies struggle to accurately learn the temporal variation patterns of electricity consumption under incomplete data conditions, leading to a decrease in prediction accuracy.

Method used

Generative adversarial networks are used for data completion, and long short-term memory networks are combined for electricity consumption prediction. By alternating training of the generator and discriminator, the complete electricity consumption time series is reconstructed, and the prediction error is used to optimize the network to ensure the accuracy of the prediction results.

Benefits of technology

It effectively restored the temporal patterns disrupted by missing data, significantly improved the accuracy of electricity consumption forecasting in scenarios with missing data, and even surpassed the forecasting accuracy of complete data, especially in seasons with severe data loss.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention provides a missing power consumption data-oriented power consumption complementing and predicting method and system, and the method comprises the steps: firstly obtaining the historical power consumption data containing random missing and a mask for the historical power consumption data with random missing, carrying out the data filling through employing a generative adversarial network, and carrying out the prediction of the data; and future electricity consumption prediction is carried out through the long-short-term memory network. And finally, performing fine adjustment on the generative adversarial network under the constraint of a prediction result to realize joint optimization of complementation and prediction. According to the invention, through a cooperation mechanism of the data completion module and the prediction module, the adverse effect of data missing on prediction precision is effectively relieved; a generative adversarial network is adopted to recover time characteristics of missing data, and it is ensured that a completion result is kept consistent with real data in statistical distribution and time rules. According to the method, joint optimization of complementation and prediction is realized through the prediction error reverse constraint generator, so that good prediction stability can be kept under different missing rates and different seasonal conditions.
Need to check novelty before this filing date? Find Prior Art