A heat load prediction method and system for district heating

By combining random forest split gain and mutual information for feature clustering screening, diffusion graph convolution and multi-head attention mechanisms are used to generate a gated adaptive adjacency matrix, the problem of high-dimensional features and nonlinear coupling relationship of regional heating data is solved, and more accurate and fast thermal load prediction is achieved.

CN120317455BActive Publication Date: 2025-08-26BEIJING KINGFORE HV & ENERGY CONSERVATION TECH CORP
2 Cites 1 Cited by

Patent Information

Application Number
CN202510797020.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2025-06-16
Publication Date
2025-08-26
Estimated Expiration
2045-06-16

AI Technical Summary

Technical Problem

In the existing thermal load prediction methods, the regional heating data has high-dimensional characteristics and nonlinear coupling relationships between the features, which are susceptible to noise interference, resulting in increased data processing complexity and difficulty in accurately capturing the changes in thermal loads. In addition, traditional methods have poor adaptability to complex heating scenarios, resulting in large deviations or lags in prediction results.

Method used

The random forest split gain and mutual information are combined to calculate the global significance of the feature and cluster it to screen key features; the gated adaptive adjacency matrix is ​​generated through diffusion graph convolution and multi-head attention mechanism, and the spatiotemporal features are extracted, and the adjacency weight is dynamically adjusted to adapt to different heating scenarios.

Benefits of technology

It improves the accuracy and speed of thermal load prediction, reduces noise interference, enhances adaptability to complex heating scenarios, and reduces prediction deviation.

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

Abstract

The present invention discloses a heat load prediction method and system for district heating, belonging to the field of data processing technology. The method includes: data collection, data preprocessing, district heating data optimization, construction of a heat load prediction model, and intelligent prediction. This solution combines random forest split gain and mutual information to obtain the global significance of features and select the initial centroid, fuse Euclidean distance and mutual information, calculate weighted distance and perform clustering, calculate association purity based on mutual information and information entropy, and filter and delete redundant features based on redundancy conditions; generate a bimodal base adjacency matrix through a gating mechanism, generate a hidden feature matrix and a dynamic propagation adjacency matrix based on diffusion graph convolution, fuse the two adjacency matrices to generate a gated adaptive adjacency matrix, extract spatial features based on diffusion graph convolution, and extract temporal features based on spatiotemporal embedding and multi-head attention mechanism, thereby reducing prediction bias and improving the heat load prediction speed while ensuring accuracy.
Need to check novelty before this filing date? Find Prior Art