Abstract
The aim of this study is to characterise the fatigue life data of multiaxial strain signals of an automobile suspension coil spring under various road load condition. The automobile coil spring experiences cyclic load due to the variation of road conditions, resulting in compression-tension loads, and providing a direct shear force and a torsion in the coil spring’s internal tension. Currently, the uniaxial analysis overestimates the component’s fatigue life because the uniaxial fatigue excludes complex stress and strain states relating to the fatigue life. Therefore, multiaxial analysis was proposed in order to determine the variance in the magnitude of main stress effects. The critical or hotspot region of a quarter car model was determined using finite element analysis. The rosette strain gauge was located at the critical region in order to capture the maximum strain signals on various roads, i.e. highway, rural and campus road during data measurement analysis in time-domain. Global statistic features of kurtosis and root mean square were analysed to obtain random signals classification. Brown-Miller, Fatemi–Socie, and Wang-Brown models were proposed to predict the multiaxial fatigue life where these models include strain criteria. Multiaxial fatigue models were proposed when evaluating fatigue life due to the complex geometries and torsional effect on coil spring. The results indicated that multiaxial strain life from Brown-Miller model possessed the highest fatigue lives. Probability distribution function for fatigue lives from Brown-Miller, Fatemi–Socie and Wang-Brown model is plotted to characterise the distribution of fatigue lives under the time domain. In order to determine the acceptability of the data, the conservative fatigue life 1:2 and 2:1 data scattering approach is proposed. Hence, the proposed life models can be used to characterise multiaxial fatigue under random strain signals for the automobile coil spring.