A time mixer-based regional power grid short-time voltage disturbance dominant cause identification method
By employing the TimeMixer method to perform time alignment and multi-scale feature extraction on multi-channel time-series data of power grid disturbances, combined with counterfactual condition analysis, the problem of locating the source of power grid disturbances was solved, enabling accurate location and quantitative assessment of power grid faults and improving the objectivity and consistency of the analysis.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHANGSHA RIOTTO ELECTRONIC TECHNOLOGY CO LTD
- Filing Date
- 2026-05-11
- Publication Date
- 2026-07-24
AI Technical Summary
Existing technologies struggle to accurately pinpoint the source of power grid disturbances. Traditional methods rely on experience-based judgment, which cannot effectively handle complex events involving cascading operations of multiple devices. Furthermore, they lack assessment of the causal relationship between electrical quantity fluctuations and device actions, leading to analysis results that depend on personal experience and are prone to bias.
Using a TimeMixer-based approach, this study identifies voltage change boundary moments by aligning and fusing multi-channel time-series data, combining multi-scale hybrid encoders and counterfactual conditional analysis, calculating intervention effect scores, quantitatively assessing the influence of each factor, and ultimately identifying the dominant inducing factor.
It enables precise location of short-term voltage disturbances in the power grid, eliminates errors introduced by data asynchrony, improves the objectivity and consistency of analysis, provides multi-dimensional disturbance assessment, and supports differentiated processing strategies.
Smart Images

Figure CN122153830B_ABST