Abstract

The modern power grids are vulnerable to faults. Without proper management and control methods, a single fault could result in the entire malfunction of the power grid, resulting in huge economic loss. Power grid simulation offers a cost-effective way to simulate possible faults. By analyzing the propagation patterns of different faults, analysts can make corresponding plans for different situations in advance. However, fault propagation analysis relies heavily on human expertise and is always performed manually based on the simulation results. To address this problem, we introduce FaultTracer, a visual analysis system that supports interactive exploration of fault propagation patterns in power grid simulation data. We propose an advanced statistical quality control theory-based anomaly detection method to identify the abnormal bus status caused by the fault. Novel visual representations for multi-dimensional time series are also introduced to display the simulation details and the anomaly detection process. Propagation of the fault is revealed by partitioning the simulation into different stages according to the abnormal status of buses. The effectiveness of FaultTracer is demonstrated by the case studies.

Graphic abstract

中文翻译：

---

FaultTracer：电网仿真数据中故障传播模式的交互式可视化探索

摘要

现代电网容易出现故障。如果没有适当的管理和控制方法，单个故障可能会导致整个电网故障，从而造成巨大的经济损失。电网仿真提供了一种经济有效的方法来仿真可能的故障。通过分析不同故障的传播方式，分析人员可以提前针对不同情况制定相应的计划。但是，故障传播分析在很大程度上依赖于人类的专业知识，并且总是根据仿真结果手动执行。为了解决这个问题，我们引入了FaultTracer，这是一种可视化分析系统，支持交互式探索电网仿真数据中的故障传播模式。我们提出了一种基于统计质量控制理论的高级异常检测方法，以识别由故障引起的异常总线状态。还介绍了多维时间序列的新颖视觉表示形式，以显示仿真详细信息和异常检测过程。通过根据总线的异常状态将仿真划分为不同的阶段来揭示故障的传播。案例研究证明了FaultTracer的有效性。

图形摘要