Fault detection method in subsea power distribution systems using statistical optimisation

Abstract

Developing automation solutions that enable remote communications, monitoring and control for subsea applications are key steps in designing subsea power distribution systems. These systems require fast local control to protect the multiple electrical loads and the capability of transferring prompt real-time trip signals. This study introduces a data-driven distributed fault detection and identification algorithm to monitor multiple subsea loads. The proposed scheme is divided into three steps. First, a stochastic hidden-Markov model (HMM) is developed to model the dynamic evolution of different potential conditions of multiple subsea loads. Simultaneously, the second step computes a model of the transition probability between the current operating condition and the potential response of an individual load. In the third step, using real-time measurements, the HMM is updated to predict an unobserved degradation of the load's current condition. This is achieved through an integrated perturbation analysis and sequential quadratic programming method. An assessment of case studies on subsea AC power system is presented, which includes different subsea motor loads for compressors and pumps. Results show robustness against uncertainty in measurement noise and changes in equipment mean time between failures, providing enhanced reliability.