Damage Identification of Offshore Platform Structure Based on Earch Model and Optimized KPCA Under Environmental Interference

Abstract

The influence of damage on structural characteristic parameters is often obscured by the complex and changeable marine environment, which may lead to damage misjudgment. Aiming at this problem, a time domain structural damage identification method based on exponential autoregressive conditional heteroskedasticity (EARCH) and kernel principal component analysis (KPCA) is proposed. First, the vibration modal component of the structure is obtained by the reduced-order variational mode decomposition method. Then, the EARCH model is established and the damaged feature matrix of the structure is extracted. After that, KPCA after parameter optimization is used to eliminate the influence of environmental changes on damage identification results. Finally, the SPE damage index is constructed to complete the damage identification of the structure. The structural damage identification was carried out on 16 test conditions under different environmental disturbances, based on indoor offshore platform vibration tests. The results show that the accuracy recognized by the proposed method reaches 98%, which is 4% higher than the traditional KPCA method based on EARCH model features, demonstrating the effectiveness of this approach. Furthermore, the method was applied to real offshore platform data, and the results show that the structural damage state is accurately identified even in the presence of typhoon interference, which is 16% higher than the traditional KPCA method, proving its feasibility and robustness.