Study of the Effect of Statistical Properties on the Random Vibration Analysis
Fatigue failure caused by the random loads (coming from the road load conditions) is of major concern for automotive industries. Random Vibration analysis is often performed to predict the life of the component, as in case of dynamic loads failure occurs by fatigue. Fatigue analysis can be done either in frequency domain or time domain (Transient Analysis). But Transient Analysis is computationally expensive as it consumes lot of time to solve the problem when compared with the Frequency Domain analysis. Hence, Transient Analysis is used mostly for the validation purpose. To convert time domain data into frequency domain FFT is performed and PSD profile is generated. This then becomes equivalent load in the Frequency Domain. FFT assumes data to be Gaussian meaning FFT is same for a Gaussian and non-Gaussian data. So, it is required to check the statistical properties of the data. Mean, standard deviation, Kurtosis and Skewness are evaluated to check the Gaussianity of the data. This paper mainly focusses on the impact of statistical properties on Random Vibration analysis.
Keywords - Random Loads, Stastical Properties, Gaussian, Fatigue failure, Frequency domain, FFT.