Review and comparison of dry friction force models

Abstract

Friction force models play a fundamental role for simulation of mechanical systems. Their choice affects the matching of numerical results with physically observed behavior. Friction is a complex phenomenon depending on many physical parameters and working conditions, and none of the available models can claim general validity. This paper focuses the attention on well-known friction models and offers a review and comparison based on numerical efficiency. However, it should be acknowledged that each model has its own distinctive pros and cons. Suitability of the model depends on physical and operating conditions. Features such as the capability to replicate stiction, Stribeck effect, and pre-sliding displacement are taken into account when selecting a friction formulation. For mechanical systems, the computational efficiency of the algorithm is a critical issue when a fast and responsive dynamic computation is required. This paper reports and compares eight widespread engineering friction force models. These are divided into two main categories: those based on the Coulomb approach and those established on the bristle analogy. The numerical performances and differences of each model have been monitored and compared. Three test cases are discussed: the Rabinowicz test and other two test problems casted for this occurrence.