Hybrid FEM–Kalman Digital Twin for Damage Detection and Remaining Useful Life Prediction of Offshore Jacket Platforms

Abstract

Offshore water jacket structures in oil/gas production and wind energy face continuous wave-induced fatigue, corrosion, and degradation. Conventional inspection methods detect only 35% of cracks <10mm with 4.2-month detection lag, costing $2.5M annually per platform (73% mobilization). This research develops a validated digital twin integrating real-time sensor data with predictive simulations for proactive maintenance. Methodology: systematic review of 156 case studies and 847 bibliometric records; sensor sensitivity analysis identifying 18 strain gauges and 8 accelerometers at 1Hz; hybrid FEM-Kalman filter architecture implemented in ANSYS for BZ 25-1 jacket (48,750 elements, 287,440 DOF); validated against 20 inspection records (2020-2024) and one-year buoy wave data. Results: stress prediction accuracy 89% (R² = 0.91; r = 0.954); fatigue damage detection improved from 42% to 85%; detection time reduced from 4.2 months to <7 days; RUL predictions 16% MAE (all within factor-of-two); Node K8-9 retrospective validation issued alert 3-4 months before 22mm crack discovery. Corrosion-fatigue modeling shows absence of cathodic protection reduces fatigue life by 50%; effective protection limits ABS reduction to 16%. System achieves 34% inspection cost reduction. This bridges theoretical digital twin frameworks and practical offshore SHM, providing validated, computationally efficient, openly documented model applicable to jacket platforms in oil/gas and offshore wind industries.