Voltage Instabilities in Large-Scale Power Systems

本文探讨了与大规模电力系统动力学相关的“电压崩溃”的概念。虽然在过去的几年里对静态电压不稳定问题进行了大量的研究，但动态电压不稳定问题在分析中才刚刚开始出现。这里的前提是，虽然静态指标很重要，但只有通过对动态的理解，才能推导出防止崩溃的有效控制措施。这项建议包含三个主要思想。首先，任何对电压动态的理解都需要一个相关的负载模型。精确的第一性原理动态负荷模型即使不是不可能实现，也是很难实现的。提出了一种基于动态负荷数据训练的神经网络模型，以代替不可能完成的负荷微分方程的编写任务。第二种思路是修改最小奇异值静态性能指标，使之包含动态现象。第三个思想是将中心流形动态电压崩溃模型推广到一般慢时变多参数系统。

In this proposal the notion of "voltage collapse" associated with large-scale electric power system dynamics is explored. While there has been significant research done during the past few years on the static voltage instability problem, the dynamic problem is only now beginning to emerge in analysis. The premise here is that while static indicators are important it is only through an understanding of the dynamics that effective control measures against collapse can be derived. This proposal contains three main ideas. First, any understanding of voltage dynamics requires a relevant load model. An accurate first principle dynamic load model is difficult if not impossible to achieve. A neural network based model trained on dynamic load data as a substitute for the impossible task of writing the load differential equations is proposed. The second idea is to modify the minimum singular value static performance index to include dynamic phenomena. The third idea is to extend a center manifold dynamic voltage collapse model to general slowly time-varying multi-parameter systems.