Country mobile energy storage load prediction error compensation optimization system and method

By constructing a Transformer-based time-series forecasting model and multiple error compensation models, the problem of large forecasting errors in rural electricity load has been solved, achieving high-precision and stable load forecasting, and improving the safety of power grid operation and the capacity for renewable energy absorption.

CN122173962APending Publication Date: 2026-06-09WEIHAI POWER SUPPLY COMPANY OF STATE GRID SHANDONG ELECTRIC POWER COMPANY
0 Cites 0 Cited by

Patent Information

Application Number
CN202610280061.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-03-09
Publication Date
2026-06-09

AI Technical Summary

Technical Problem

Existing technologies are insufficient for accurately predicting the diverse electricity loads in rural areas. In particular, the load prediction error is large due to the volatility and seasonality of distributed renewable energy output, which affects the safe and stable operation of the power grid and the absorption of renewable energy.

Method used

We employ a Transformer-based time series prediction model combined with BiLSTM, GRU, and XGBoost error models. Through multi-dimensional dataset construction, DBSCAN anomaly detection, Lagrange interpolation, and data normalization, we construct multi-type error prediction models. Finally, we select the optimal error compensation model through LA-Kmeans clustering to achieve accurate compensation for prediction errors.

Benefits of technology

It has improved the accuracy of rural energy load forecasting, enhanced the model's adaptability and stability in complex scenarios, and achieved high-precision forecasting and intelligent error compensation for electricity load in multiple rural scenarios.

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

Abstract

The invention discloses a rural mobile energy storage load prediction error compensation optimization system and method, and relates to the technical field of mobile energy storage regulation, and the method comprises the steps: collecting the power load, power generation and meteorological data of a rural typical scene, constructing a multi-dimensional data set covering a complete year, and carrying out the compensation optimization of the prediction error of the rural mobile energy storage load; and carrying out abnormal value identification, missing value interpolation and normalization processing to form a feature matrix. And constructing a time sequence prediction model based on a Transform architecture for training and verification to obtain an initial load prediction value and a prediction error sequence corresponding to the initial load prediction value. And constructing a multi-type error prediction model in parallel, and performing independent prediction on the error sequence. And after the actual error sequence and each model prediction value are paired and interpolated according to time, clustering analysis is performed in a two-dimensional space according to the distance between the actual error sequence and an ideal decision line, an optimal error prediction model is dynamically selected, an optimal error compensation value is determined, and a final load prediction value after error compensation is obtained.
Need to check novelty before this filing date? Find Prior Art