In this project a new stochastic process, Laplace moving average process (LMA), has been used to simulate the load signals recorded from a stiff-structured cultivator machine while it is driven on a rough field. Fatigue damage estimation has been the main focus and the rainflow cycle count method has been used for this purpose. To simulate the tine vibration when it is not in the working state a Gaussian process has been used, while the skewed load signals with random extreme values produced by working tines have been simulated by the Laplace moving average process. This turns out to be a very appropriate method to simulate load signals, which generates slightly-overestimated damage values of great importance for successful machine designs. Furthermore, a Monte Carlo study has been conducted to evaluate the completeness and robustness of the LMA simulation. A variability analysis with bootstrap has also been applied to assess the difference of sample characteristics among different experiments. Finally, we have investigated the correlation between loaded signals from different tines and signals corresponding to different physical properties. The rainflow damage of a virtual synthetic signal with maximized variance has been evaluated by principal component analysis.
Photo credits: Nic McPhee