Vibration induced fatigue life estimation based on non-stationary stochastic processes segmented to piesewise stationary segments

At excavations of quarries, construction sites or in mining, etc. articulated haulers are commonly used for the transportation of material. Thus, articulated haulers are exposed to a diversity of different worksites that all will load them differently. With a continuously growing market and added exhaust emission regulations, the machines need to be continuously optimized to meet durability and weight (fuel efficiency) requirements. In the structural design of any given component in an articulated hauler, the assumed loads used will reflect on its design. However, it is rarely the case that the loads used in the design procedure strictly represents the dynamic loads a hauler will be subjected to. As a result, the machines are generally oversized. To address the issue of oversized components, a statistical approach for the estimation of their fatigue life may be utilized. With the aid of power spectral density estimates of a components dynamic loads in combination with a finite element model of the component its vibration response in terms of power spectral densities may be estimated. Based on the vibration response spectra the component’s fatigue life may be estimated assuming that the components dynamic loads may be considered as weakly stationary stochastic processes. To investigate if the dynamic loads used for fatigue life estimation may be considered as weakly stationary stochastic processes; stationarity tests have been carried out as well as estimates of the kurtosis and the mean square value as a function of time have been produced. The results indicate that the dynamic loads may not be considered as weakly stationary stochastic processes. Furthermore, by segmenting the dynamic load time series into segments that can be considered as weakly stationary the estimated fatigue life seems to decrease slightly if compared to the case when assuming that the dynamic loads are weakly stationary.