Structural Reliability Of Jacket Platform Based On Non-Probabilistic Method

In the practical engineering application of jacket platform,there are usually many uncertain factors involved,so the uncertainty analysis technology of structure and reliability research have become an effective solution to solve this problem.Therefore,it is of great significance to study the method of structural uncertainty analysis on jacket platform to verify the reliability of engineering structure,optimize the design and control the cost.Monte Carlo Simulation(MCS)is a common method to deal with structural uncertainty analysis and reliability research of jacket platform.It realizes effective measurement of structural uncertainty and reliability analysis.However,the Monte Carlo simulation method requires too many samples and too many cycles,so the calculation accuracy increases with the increase of sample points and cycles,and the calculation workload is very complicated.So,based on Monte Carlo simulation,two methods of Polynomial Chaos Expansion(PCE)and constructing Polygonal Convex Set model(PCS)are proposed in this paper.Based on these three methods,uncertainty modeling and reliability analysis are discussed.At the same time,chaotic polynomial model and multilateral convex set model are applied to the reliability analysis of jacket platform structure.Not only the multilateral convex set model is applied to the reliability evaluation of jacket platform structure,but also the non-probabilistic reliability index based on the model is proposed.In general,the research covers the following three aspects:Study the structural reliability of jacket platform based on Monte Carlo simulation method.The finite element model of jacket platform structure is constructed,and the coupling effect of wave,wind and sea current load is combined.The load effect values and ultimate structural resistance in 8 directions were determined,and the reliability of the jacket platform structure and its corresponding failure probability were calculated by Monte Carlo simulation method.According to the analysis,the reliability of jacket platform is the lowest in the direction of 90° under load,which indicates that its resistance is weak in the direction of90°.Carry out the structural reliability research of jacket platform based on chaotic polynomial expansion method.First,the environmental load equivalent of the extreme I distribution was normalized into normal distribution,and the normal distribution random variables were expanded by Hermite orthogonal polynomials,and the PCE model of jacket platform structure was established.Considering the accuracy of the PCE model,the 1-8 order PCE model is constructed and its polynomial coefficients are calculated using the orthogonal matching tracking algorithm in the field of compressed sensing.Finally,the failure probability and reliability are calculated,and the model converges at the 8th order.Compared with the results of 105 MCS cycles,it takes less time to achieve the same accuracy,and the convergence rate is faster and the efficiency is higher.Study on structural reliability of jacket platform based on multilateral convex set model.Firstly,the multi-dimensional uncertain parameters of the jacket platform structure were analyzed by Principal Component Analysis(PCA).Firstly,the whole parameter sample was pre-processed in this paper,and the reference parameter vector was orthodontized through covariance matrix operation to obtain a new feature vector.These new vectors are used as the model base,and the original parameters are transformed to the model base direction by linear transformation,and then the upper and lower limits of uncertainty of parameters in the new model space are calculated.Using these information and the uncertainty modeling coordinate system,the PCA interval model is established successfully,and the uncertainty of structural parameters is evaluated.Finally,combined with the intersection domain between the traditional interval model and PCA interval model,a multilateral convex set model suitable for jacket platform structure was constructed.The PCS model is completely contained in the PCA interval model and the traditional interval model,and the model is more compact with less blank area.The multilateral convex set model is introduced into the structural non-probabilistic reliability analysis,and the non-probabilistic reliability index is calculated using the ratio of two norms.Comparing with the results of MCS cycle,the results show that the PCS model can measure the multi-dimensional parameters of jacket platform structure more reasonably and effectively than the MCS method.The model can reasonably reflect the correlation information between parameters,which is realized by shape and size,and can accurately calculate the failure probability and reliability.It provides an effective solution for the reliability analysis of jacket platform structure in high dimensions.