Time-temperature analysis methods are usually applied to predict the useful life of automotive components. Components life is affected by exposure to heat during vehicle service life. The extent of reduction in component life, which may be caused by material thermal degradation, depends on the component temperature and the time duration at that temperature. The rate of material thermal degradation of automotive components varies widely depending on material thermal stability, vehicle duty cycle, and the thermal environment that the component is exposed to. Thermodynamic properties such as the activation energy of each material are used to determine the rate of thermal degradation [1,2]. In this approach, material thermal degradation models are used to predict component life during the service life of a vehicle. As the rate of thermal degradation increases with increasing material temperature, the useful life of a component will be reduced as the material temperature increases. Therefore, it is desired to keep the rate of thermal degradation low enough so that a certain level of component performance can be maintained at the end of the vehicle life. The acceptable performance level may be component dependent and vehicle dependent. For example, a passenger car will require different performance than a heavy duty truck even if same material is used on both vehicles. To maintain the required component performance, the definitions of “long term temperature goal” and “short term temperature goal” are introduced. Therefore, the factors affecting the predicted component life can be summarized as follows: measured component temperatures, material long and short term temperature limits (goals), material activation energy, and vehicle duty cycle. All of these factors typically have an inherent uncertainty. These uncertainties will affect the overall confidence level in the predicted time-temperature calculations. Therefore, it is the main purpose of this paper to estimate the uncertainty in component life predictions and their sensitivity to each of the input factors. Given these uncertainties, it is statistically possible to determine the most influential parameters and the overall uncertainty in the predicted component life. Several examples are given where the sensitivity/uncertainty analysis for different vehicle components are presented.
