A data-driven LSTM-based management and control approach for fatigue life of subsea wellhead system

Subsea wellhead system (SWS) is the foundation of oil and gas well construction and the crucial channel of intervene operation. As the safety barrier of well control, the inappropriate management and control of SWS fatigue life would lead to catastrophic safety accidents. This paper puts forward a data-driven LSTM-based fatigue life management and control methodology for SWS, integrating data monitoring, a data-driven LSTM model, probability density function (PDF) and reliability assessment. In detail, monitored parameters are mapped into the load spectrum assisted by a data-driven LSTM model to assess the SWS fatigue damage. Considering fatigue uncertainty, the PDF of fatigue damage is constructed to analyze possible accumulation damage evolution paths of SWS. Further, the relationship among reliability, the accumulated fatigue and the main load amplitude is predicted to determine the health index characterizing the fatigue magnitude and the operation risks of SWS. The analysis results illustrate that the main bending stress amplitude can be deemed as SWS health index, as it can reflect the fatigue status and the reliability of SWS. Thus, the real-time health index provides a vital reference for operators to take timely and effective measures, achieving the management and control of SWS fatigue life. •A data-driven LSTM-based fatigue life management and control methodology is proposed for subsea wellheadd system.•A data-driven LSTM model is developed to map the monitored parameters into the load spectrum.•The possible accumulation damage evolution paths are established using thePDF of fatigue damage for SWS.•The health index characterizing the fatigue magnitude and the operation risks is determined by reliability assessment.