一种水文序列缺失数据补全与趋势预测系统及方法
By constructing a hydrological sequence missing data completion and trend prediction system, and combining dynamic graph neural networks and physical constraint verification, the problem of hydrological data completion under high missing rate was solved, achieving high accuracy and reliability of data processing, and meeting the decision-making needs of water conservancy management.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- TIANJIN UNIV
- Filing Date
- 2026-03-31
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies, when processing missing hydrological sequence data, especially in scenarios with high proportions or continuous missing data, struggle to effectively utilize the spatiotemporal coordination within the watershed. This leads to physical distortions or logical contradictions in the completion results, and a lack of credibility and transparency, failing to meet the decision-making needs of water conservancy management.
A system for completing and predicting missing hydrological sequence data is constructed. Through data acquisition and preprocessing, spatiotemporal feature fusion, physical constraint completion, uncertainty prediction and credible evidence storage modules, combined with blockchain technology, the system achieves credible evidence storage and transparent traceability of data. Dynamic graph neural networks and physical constraint verification are used to improve the robustness and interpretability of the completion results.
It significantly improves the accuracy and robustness of completion in scenarios with high missing rates, ensures that the completion results conform to the physical consistency of natural hydrological processes, solves the problems of data trust and traceability, enhances the interpretability and auditability of the system, and adapts to complex environmental changes.
Smart Images
Figure CN121958786B_ABST