The role of transactional data in prognostics and health management work processes

Analytics supporting prognostics and health management (PHM) work processes traditionally leverage time-series data to monitor component states and predict fault progressions in order to positively impact performance related to safety, profitability and risk management. Developing analytical models for the purpose of monitoring is asset-specific and assumes that the data is captured and accessible. In practice, monitoring assets in real-time is reserved for highly critical assets, while all assets have transactional data stored in enterprise asset management (EAM) systems. This paper reviews methods for measuring transactional data quality and for measuring asset performance metrics and health indicators from historical maintenance records that can be used in PHM initiatives. Data from both transactional sources and from machine-measured sources should be used together to derive a complete picture of the maintenance strategies and actions in an industrial site.