Damage degree identification of offshore platform structure based on KPCA-HKELM under environmental interference

Changes in damage characteristic parameters caused by structural damage are often disturbed by changes in environmental factors, making it difficult to identify the damage degree. To address this problem, a time-domain structural damage degree identification method based on kernel principal component analysis (KPCA) and hybrid kernel extreme learning machine (HKELM) is proposed. Firstly, the exponential autoregressive conditional heteroskedasticity model and KPCA are used to extract the damage characteristics the squared prediction error and kernel principal components that eliminate environmental interference. Then, the HKELM neural network optimized by the sparrow search algorithm is used to realize the identification of structural damage degree. Based on the vibration experiment of the indoor offshore platform, the structural damage degree of four kinds of damage degree under different environmental interference is identified. The results show that the accuracy rate is 96.5%, and the effectiveness of the method is proved. Furthermore, the weighted voting method is used to fuse the decisions of different sensors, which further improves the accuracy of damage degree recognition, reaching 99.0%.