Structural Model Updating Method of Medium-Deep Water Jacket Platform Based on Sensitivity Clustering

As offshore oil and gas exploration extends into deeper waters and platforms operate over extended service lives, accurate and timely structural health monitoring of deepwater jacket platforms becomes increasingly critical. Rapid construction of a high-fidelity numerical twin model is therefore essential, yet existing structural model correction methods are largely designed for simple or low-dimensional systems and are inadequate for large-scale marine structures with high-dimensional parameters. To address these challenges, this study proposes a sensitivity-based clustering method for structural model correction of jacket platforms. First, a novel numerical sensitivity index of correction parameters is established, and these parameters are clustered into representative groups. An optimization objective function based on modal characteristics is then formulated, converting the model correction problem into a minimization problem, which is solved using a multi-objective optimization algorithm. To verify the effectiveness and noise robustness of the proposed method, numerical simulations of a jacket platform in the Weizhou Oilfield were performed. Here, maximum errors are reduced to 0.05% and 0.12% under low- and high-noise conditions, respectively, showing significant reductions in relative errors and improved modal confidence. The proposed methodology effectively enables rapid and accurate updating of numerical twin models for deepwater jacket platforms.