ClarusC64/energy-grid-load-instability-v0.1

该数据集用于评估模型是否能从短期操作轨迹中检测出电网的不稳定性。每行数据代表一个电网场景在三个时间点的观测，任务是通过多变量交互推理判断系统是趋向稳定还是不稳定状态。数据集包含负载轨迹、储备裕度轨迹、频率漂移轨迹等测量值，以及变压器温度、存储响应、调度延迟等系统指标。此外，还包含了一些干扰变量如天气噪声和传感器噪声。预测目标为二分类，标签1表示电网不稳定风险，标签0表示稳定或受控状态。数据集是Clarus稳定性推理基准的一部分，支持跨域稳定性检测、基于轨迹的推理和基础设施韧性建模等研究。

This dataset evaluates whether models can detect emerging electrical grid instability from short operational trajectories. Each row represents a grid scenario observed across three time steps, and the task is to determine whether the system is moving toward a stable operating state or toward instability. The dataset includes measurements such as load trajectory, reserve margin trajectory, frequency drift trajectory, and additional system indicators like transformer temperature, storage response, and dispatch delay. It also contains decoy variables like weather_noise and sensor_noise. The prediction target is binary classification, with label 1 indicating grid instability risk and label 0 indicating a stable or controlled grid state. The dataset is part of the Clarus Stability Reasoning Benchmark and supports research into cross-domain stability detection, trajectory-based reasoning, and infrastructure resilience modeling.