基于反事实推理的工业能源监测网络结构可解释优化方法
By constructing a deep learning framework based on counterfactual reasoning, and combining causal inference and a hybrid masking mechanism, the problem of multi-monitor optimization in industrial energy monitoring networks was solved, the importance and contribution rate of monitors were analyzed, and the energy monitoring network structure was optimized.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHONGQING UNIV
- Filing Date
- 2026-05-20
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies lack effective means for "top-down" multi-monitor optimization in industrial energy monitoring networks. This is especially true in the context of multiple workshops and branch plants in manufacturing enterprises, where it is difficult to explain the complex spatiotemporal dependencies of energy monitoring networks and analyze the contribution rates of monitors.
A deep learning framework based on counterfactual reasoning is adopted. A hierarchical spatiotemporal graph is constructed through a dual-branch structure consisting of a fact learner and a counterfactual inferrer. Causal inference is used to mine the spatiotemporal dependencies of the monitoring network, generate interpretable monitor importance and contribution rate analysis, and optimize the energy monitoring network structure.
Under the constraint of ensuring the accuracy of energy consumption prediction, the causal dependencies between monitoring nodes are identified, and the spatiotemporal characteristics of multiple monitors are analyzed to provide interpretable structural optimization support for industrial energy monitoring networks.
Smart Images
Figure CN122242286B_ABST