Step Through the Noise: Insight into Resilience-Driven Power Asset Management

The amount of data accumulated by utility companies is growing in volume and variety each year. Such data can be a very valuable asset to utilities, but because it is generally viewed only through traditional tables and charts, it has not yet been used to its full potential. A utility can make better use of its data by employing machine learning techniques to recognize patterns that are hidden when traditional techniques are used. Association rule analysis is one of numerous machine learning techniques, that has been widely used in business and some engineering fields to identify important feature correlations previously hidden due to the volume of data under analysis, but it is not widely adopted in the asset management sector. This project applies association rule analysis to a historical transmission line outage event database through a resilience-based asset management lens, distinguishing patterns and developing rules that relate to the occurrence of input features together with long-duration outages.