Researchers have proposed a standardized framework for evaluating the performance of machine learning-based power system protection systems. The framework defines seven key dimensions that must be specified in each study, including the protection objective, physical scope, and evaluation outputs. A case study using a public benchmark dataset showed that a multi-layer perceptron achieved high accuracy for fault classification and localization, but was outperformed by a conventional locator under certain conditions. --- Why it matters: This matters to engineers working on power system protection because it provides a standardized way to evaluate the performance of machine learning-based systems. This can help ensure that these systems are reliable and effective in real-world applications. Source: https://arxiv.org/abs/2608.20181