A novel approach based on meta-modeling technique and time transformation function for reliability analysis of upgraded automotive components

Highlights: • A meta-model is combined to a time transformation function for reliability analysis. • The meta-model is built from a dimensional decomposition and Lagrange polynomials. • A power model is shown suitable to build the time transformation function. • Efficiency and accuracy of the approach are demonstrated through an application. • The proposed approach allows early reliability estimation at design stage. Early reliability estimation is still a challenging task. The paper presents a novel approach to deal with early reliability estimation of upgraded automotive components. The key idea is to combine reliability analysis based on efficient surrogate models and time transformation function principle. The surrogate model, built using Dimensional Decomposition Method and projection throughout a Lagrange polynomial basis, is used to substitute a time consuming implicit model initially used to compute the fatigue lifetime. The time transformation function is represented by a parametric power law model where the corresponding parameters are obtained through statistical analysis based on both numerical and experimental reliability results of a reference design. The reliability of an upgraded design is easily obtained by applying the time transformation function to the reliability estimation given by performing Monte-Carlo simulations on the surrogate model corresponding to the upgraded design. An application to a mechanical component, used in car seats, clearly illustrates the efficiency and the accuracy of the proposed approach.