Durability prediction of coil spring through multibody-dynamics-based strain generation

Abstract

• A new acceleration-strain conversion model is proposed. • Realistic strain time histories are generated using multibody-dynamics model. • Previous acceleration-strain model is compared with current work. • Proposed model gives accurate durability predictions. • The NRMSE values of the proposed model is very close to zero. The purpose of this study is to develop an acceleration-strain conversion model that considers torsional strain and spring curvature effects in inducing strain for a suspension coil spring. Measurements of strain–time histories at the coil spring are limited by complex geometry and insufficient workspace. This condition increases the demand for strain signal generation through multibody dynamics (MBD) simulation, reducing the need for real strain measurement of coil spring. Road tests were conducted to obtain the unsprung mass acceleration and strain signals of a coil spring under four road conditions (rural, industrial, highway, and campus road). Quarter-car suspension MBD simulation was modelled to simulate the deflection of a spring excited under unsprung mass acceleration. By using this model, simulated strain data with similar properties as the experimental data were generated for fatigue life prediction. The predicted fatigue life from the generated strain indicated a good correlation with the experimental fatigue life within the boundary and showed very low normalised root-mean-square error (NRMSE) between 4 × 10 −6 and 2 × 10 −4 . Finally, it is suggested that the acceleration-strain conversion model showed an enhanced performance for producing realistic strain signals in accurately predicting the durability of coil spring. This can, therefore, further reduce the need for real strain measurement at the coil spring that can result in an erroneous signal.