Fatigue Evaluation of Multi-Degree of Freedom, Frequency Domain, Stochastic, Truck Road Load Models

A number of semi-deterministic and stochastic formulations of multi-degree of freedom, frequency domain load models for heavy truck chassis are proposed and evaluated. The semi-deterministic models aim at reproducing the damage of a specific vehicle, while the stochastic ones aim to describe a collection of vehicle loads. The stochastic models are divided into two groups: Monte Carlo based and models based on single spectrum matrices. In both cases, the objective is to provide a load model that may be used to produce a design with a certain probability of survival. The goodness of the models is evaluated through a comparison of their damage outcomes with the corresponding damages of a set of time domain loads. This original time domain load set consists of chassis accelerations collected from seven physical trucks. Model parameters investigated included spectrum formulation (averaged vs damage equivalent spectra), polar vs rectangular formulation, and the effect of sorting of outcomes in the Monte Carlo models. For semi-deterministic load models, damage equivalent formulations performed well independent of number of load degrees of freedom, while averaged spectra here consistently under estimated the damage. No probabilistic model was found that consistently yielded good damage correlation over a large range of load degrees of freedom and for designs with a significant range of desired survival probability.