An integrated energy load forecasting method based on multi-scale graph conditioned state space model

By using a multi-scale graph-conditionalized state-space model and dynamic graph learning, combined with a three-dimensional attention mechanism, the nonlinear characteristics and coupling relationships of load forecasting in the integrated energy system of the park were solved, achieving high-precision load forecasting and improved robustness.

CN122178304APending Publication Date: 2026-06-09NANJING UNIV OF INFORMATION SCI & TECH
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
NANJING UNIV OF INFORMATION SCI & TECH
Filing Date
2026-05-12
Publication Date
2026-06-09

AI Technical Summary

Technical Problem

Existing load forecasting methods for integrated energy systems in industrial parks are unable to capture the nonlinear characteristics in load data and the coupling relationship between electricity, cooling, and heating loads. They lack the ability to collaboratively model three-dimensional features, and the state transition matrix cannot be dynamically adjusted, resulting in insufficient prediction accuracy and robustness.

Method used

A joint prediction method based on a multi-scale graph conditional state space model, dynamic graph learning, and a three-dimensional attention mechanism is adopted. By modulating the state transition matrix through the adjacency matrix of the dynamic graph, spatiotemporal endogenous fusion is achieved, capturing high-frequency and medium-to-long-term features, and performing three-dimensional feature collaborative optimization.

Benefits of technology

It improves the accuracy and robustness of load forecasting, adapts to changes in the topology of the park's energy network, enhances the model's real-time performance and adaptability, and is suitable for online application scenarios of park energy dispatch.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122178304A_ABST
    Figure CN122178304A_ABST
Patent Text Reader

Abstract

This invention discloses a comprehensive energy load forecasting method based on a multi-scale graph conditional state-space model. The method includes constructing a comprehensive energy load forecasting dataset for a park, performing data preprocessing, and then using Pearson correlation analysis to select highly correlated features; dividing the dataset into training, validation, and test sets, and standardizing the data; constructing a joint prediction model based on dynamic graph learning, a multi-scale graph conditional state-space model, and a three-dimensional attention mechanism; training the joint prediction model using the training set; adjusting hyperparameters and selecting the optimal model using the validation set; inputting the test set into the trained model, and outputting the predicted electricity, cooling, and heating loads; restoring the actual predicted values ​​through inverse normalization; and evaluating the model performance using multiple indicators. This invention ensures the real-time requirement of the forecast and is suitable for online application scenarios in park energy dispatching.
Need to check novelty before this filing date? Find Prior Art