基于改进时空生成对抗网络的电力系统缺失数据修复方法
By using an improved spatiotemporal generative adversarial network, combined with the GAU attention mechanism and the TCAN network, the problem of missing data repair in power systems was solved, achieving fast and accurate data repair and improving the reliability and observability of the system.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- YANSHAN UNIV
- Filing Date
- 2022-09-01
- Publication Date
- 2026-07-17
AI Technical Summary
Existing generative adversarial network models struggle to effectively extract the complex spatiotemporal characteristics of power measurement data in power systems, resulting in poor data restoration performance, training difficulties, and a tendency for model collapse.
An improved spatiotemporal generative adversarial network is adopted, which combines the GAU attention mechanism and the TCAN network. By extracting the temporal and spatial features of the data, and using Wasserstein distance as the loss function, data repair is performed by combining context constraints and authenticity loss.
It enables rapid and accurate repair of missing data in the power system, improves the system's reliability and observability, and solves the problem of the power system's self-healing capability after being attacked.
Smart Images
Figure CN115525472B_ABST