Blowout preventers (BOPs) play a critical role in the safety of offshore drilling rigs. The impact of a major BOP failure was fully demonstrated by the Deepwater Horizon blowout in the Mexican Gulf in 2010. Nonetheless, today, only rudimentary methods exist for condition monitoring of many BOPs. In this paper, we will present a novel system for verifying the correct function of the BOPs and the detection of anomalies in the operation of these. It uses specialised, deep-sea hydrophones to collect acoustic data during the BOP operation. These data are then used to create a compound statistical model, based on mixtures of hidden Markov models, which can subsequently be used for function verification and anomaly detection from new samples of acoustic data. The work to date has been based on data collected from dry and wet testing of individual BOPs and has yielded promising results both in terms of both function verification and anomaly detection. In this paper we present results from the development testing from individual BOPs and discuss the suitability of the approach for monitoring in the field.
展开▼
