Statistical characterization of acceleration levels of random vibrations during transport

Random vibration tests are an efficient way to simulate the mechanical vibratory effects caused by transportation. The usual method is only concerned with the frequency distribution pattern of the signal using the average power spectral density. This work offers an additional method based on detailed analysis of instantaneous acceleration levels of a real road transport, which enables modelling of the statistical distribution of these levels. Continuous recording of acceleration signal all along the journey permits confirmation that this statistical distribution is not a Gaussian distribution but a modified Gaussian distribution, for which parameters are estimated and discussed. Therefore, it is possible to evaluate the transport severity by working out the appearance probability of acceleration levels greater than a fixed threshold and also the statistical moments, i.e. second order moment which gives the root mean square value together with fourth order moment (kurtosis) which evaluates the difference between the experimental distribution and the Gaussian distribution. Copyright © 2011 John Wiley & Sons, Ltd.