A lifespan prediction method for commercial vehicle seat slide rails based on parameter simulation and probabilistic learning

The prediction of the lifespan of commercial vehicle seat slide rails is plagued by complex modeling and low accuracy. To address these problems, this paper proposes a slide life prediction method integrating parametric finite-element simulation with probabilistic learning. Firstly, a fully parametric finite element model is constructed. With this model, multi-condition simulations are set up to obtain stress and life data. Secondly, orthogonal experimental design is combined with analysis of variance to identify the primary factors influencing lifespan. Thirdly, Monte Carlo stochastic augmentation based on the Weibull distribution assumption is performed to build a highly representative training dataset. Finally, a joint model integrating an Optimized Kolmogorov–Arnold Network (OKAN) and Gaussian process regression (GPR) is developed. OKAN captures high-dimensional non-linear mapping relationships. GPR quantifies the prediction uncertainty, enabling point prediction and confidence interval estimation for slide-rail life. Experimental verification shows that the error between the finite element simulation results of the new method and the physical test data is only 3.9%. The orthogonal experimental design reduces the 729 simulations required for full-factor experiments to only 27. Through comparative analysis, the prediction accuracy of the new method is higher than that of traditional methods.