Numerical identification of position-dependent friction coefficients from measured displacement data in a bolt-nut connection

Abstract

Friction is a complex system affected by microscopical effects and multidisciplinary phenomena. Coulomb's simple friction model with a constant friction coefficient cannot account for all these tribological effects. Nevertheless, this model is still widely utilised for calculations of mechanical applications. In order to reflect the importance of friction as a parameter for functionality, we need more realistic and sophisticated calculations. This is particularly relevant for bolt-nut connections, which serve as motivating example for our study. Our approach is to introduce position-dependent friction coefficients by dividing the contact surface into different friction areas, each characterised by a constant friction coefficient. These coefficients are then adapted to measured displacement data. To this end, we develop a numerical parameter identification tool. The tool combines calculations in Ansys Mechanical, an established Finite Element software, and Microsoft's Visual Basics for Applications for optimisation purposes. We verify the parameter identification tool using the simple model of a block on a planar surface. Within this test scenario, the algorithm converges and provides a good approximation of the friction coefficients. Subsequently, we apply parameter identification to the model of a bolt-nut connection. We perform optical measurements to acquire experimental displacement data. The parameter identification tool demonstrates its functionality. Finally, we discuss future modifications of the procedure, that will enable more realistic and reliable results.