AMPS: Collaborative Research: Efficient Algorithms for Ultra-Fast Detection of Power System Contingencies in the Transient Regime

电力系统的稳定运行对电力系统的可靠运行至关重要。这项工作侧重于大扰动的影响，例如由于短路造成的输电线路损失，或者由于不可预测的故障而导致的火力发电机组突然断开。当这种故障发生时，系统将经历一个高度非线性的暂态事件。可再生能源在电网中的大规模整合对传统的暂态稳定分析提出了挑战，并降低了系统的惯性。与此同时，电网上的相量测量单元(PMU)激增。这些传感器提供有关系统周围不同位置的电压、电流和频率的准确、高频信息，有助于提高我们对系统级别的情况感知。然而，将输入数据转换为可操作的信息仍然是一个开放的研究问题。在暂态制度的背景下，重要的是尽快确定，即在几个电气周期内，哪里出了问题，以便预测故障后一段时间内系统动态的演变。这种超高速检测是至关重要的，因为它可能允许广域监测和控制系统采取适当的行动，这些广域监测和控制系统可以启动电网中的各种资产，例如灵活的交流输电系统(FACTS)设备，甚至可再生资源的变流器。这项工作的主要目标是开发可在计算上可扩展的、统计上有效的实时算法，以便根据PMU输出在电力系统暂态过程的早期阶段检测和识别意外事件。这些算法的有效性和有效性将基于电力系统在事故后的暂态稳定区域内的随机模型。这是一个非常具有挑战性的问题，由于电网动态系统的高维和计算复杂性，现有的不确定性传播方法无法解决这一问题。一个关键的组成部分是不确定性传播问题的重新表述，它克服了维度灾难，使得能够随着电力系统运行点的变化而快速地重新估计统计量，并允许自适应地选择电力系统仿真。这项工作将依赖于多种技术的组合，如不确定性量化、电网建模和顺序检测。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

The stability of the electric power system is of the utmost importance for its reliable operation. This work focuses on the impact of large disturbances, such as the loss of a transmission line due to a short-circuit, or the sudden disconnection of a thermal generating unit due to an unpredictable malfunction. When such faults occur, the system undergoes a transient event of highly nonlinear nature. The ongoing large-scale integration of renewable energy resources in the grid challenges conventional transient stability analysis, and reduces the 'inertia' of the system. At the same time, there is a proliferation of phasor measurement units (PMUs) on the grid. These sensors provide accurate, high-frequency information about voltage, current, and frequency at various locations around the system, and are instrumental in increasing our system-level situation awareness. Nevertheless, the transformation of the incoming data to actionable information is still an open research problem. In the context of transient regimes, it is important to establish as soon as possible, i.e., within a few electrical cycles, what has gone wrong, in order to predict the evolution of the system dynamics over the period of time following the fault. Such 'ultra-fast' detection is critical, since it may allow appropriate actions to be taken by a wide-area monitoring and control system that can actuate various assets around the grid, such as flexible ac transmission system (FACTS) devices or even the converters of renewable resources.The main objective of this work is to develop computationally scalable, yet statistically efficient, real-time algorithms in order to detect and identify contingencies in the early stages of the transient regime of the power system based on PMU outputs. The validity and efficacy of these algorithms will be based on a stochastic model for the power system during the transient stability regime following a contingency. This is a very challenging problem that cannot be addressed with existing uncertainty propagation methodologies, because of the high-dimensionality and the computational complexity of the power grid dynamical system. A key component is the reformulation of the uncertainty propagation problem which overcomes the curse of dimensionality, enables the fast re-estimation of the statistics as the power system operating point changes, and allows for the adaptive selection of power system simulations. This work will rely on combination of diverse techniques, such as uncertainty quantification, power grid modeling, and sequential detection.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.