Numerical durability analysis is the only approach that can be used to assess the durability of vehicles in early stages of development. In these stages, where there are no physical prototypes available, the road wheel forces (or spindle forces) for durability testing on Belgian PG (Proving Ground) must be predicted by VPG (Virtual Proving Ground) or derived from the measured forces of predecessor vehicles. In addition, the tuning parts and geometry are not fixed at these stages. This results in the variation of spindle forces during the development stages. Therefore, it is not reasonable to choose the forces predicted at a specific tuning condition as standard forces. It is more reasonable to determine the standard forces stochastically using the DB of the measured forces of predecessor vehicles.
The spindle forces measured or predicted on Belgian PG are typically stationary random. To treat the force signals stochastically, it is necessary that the forces are expressed in analytic formulation. Without analytic expression, it may be impossible to reproduce the forces reasonably.
In this paper, we propose a methodology to reproduce the spindle forces measured on Belgian PG as an analytic formulation with sufficient accuracy. In addition, the coefficient of the analytic format can be treated stochastically. Using the coefficients, we can reproduce standard forces to be used for numerical durability testing. The effectiveness of the methodology is verified by comparing the results of the fatigue lives of suspension parts of A-segment vehicles using the standard forces with those using the measured forces.
