Towards a data analytics framework for medium voltage power cable lifetime mangement

Power cables are critical assets for the reliable and cost-effective operation of nuclear power plants. The unexpected failure of a power cable can lead to lack of export capability or even to catastrophic failures depending on the plant response to the cable failure and associated circuit. Prognostics and health management (PHM) strategies examine the health of the cable periodically to identify early indicators of anomalies, diagnose faults, and predict the remaining useful life. Traditionally, PHM-related strategies for power cables are considered separately with the associated penalties involved with this decision. Namely, there is a lack of consideration of the interactions and correlations between failure modes and PHM tests, which results in scalability issues of ad-hoc experiments, and accordingly incapability to exploit the full potential for PHM strategies in an effective manner. An effective and flexible PHM strategy should be able to consider not only different PHM strategies independently, but also it should be able to fuse these tests into a cable health state indicator. The main contribution of this paper is the proposal of a PHM-oriented data analytics framework for medium voltage power cable lifetime management which incorporates anomaly detection, diagnostics, prognostics, and health index modules. This framework includes the characterization of existing data sources and PHM-oriented data analytics for cable condition monitoring. This process enables the creation of a database of existing datasets, identification of complementary PHM techniques for an improved condition monitoring, and implementation of an end-to-end PHM framework.