The Use of Autocorrelation Function in the Seasonality Analysis for Fatigue Strain Data

The seasonal dependency or seasonality is a general component of the time series pattern that can be examined via correlograms, where the correlogram displays graphical and numerical information in an autocorrelation function. This paper discusses the use of an autocorrelation function in the seasonality analysis for the fatigue strain data to help identify the seasonal pattern. The objective of this study is to determine the capability of this time domain method in detecting the seasonality component in the fatigue time series. A set of case study data consisting of the non-stationary variable amplitude loading strain data that exhibits random behaviour was used. This random data was collected in the unit of micro-strain on a lower suspension arm of a mid-sedan car. The data was measured for 60 seconds at the sampling rate of 500 Hz, which gave 30,000 discrete data points. The collected data was then analysed in the form of an auto-correlogram shape and characteristics. As a result, an autocorrelation plot is weak in identifying the seasonal pattern, but the autocorrelation coefficient values are statistically significant and have shown a positive serial correlation. Thus, the finding of this characteristic is expected for a non-stationary signal.