The capability to accurately predict friction (both sticking and sliding states as well the timing of the transitions between the two states) is critical to the design of flexible risers, umbilicals, and connectors. For example, the critical curvature at which the tensile armor wires of a flexible riser transition from sticking to slipping determines the maximum bending moment and plays a key role in strength and fatigue life predictions.
Friction is complex and difficult to predict, and its simulation within multi-body structures is typically relegated to Finite Element (FE) methods. The FE solver friction algorithms are perhaps the most sophisticated and highly guarded part of the solver. This is a predominant differentiator from one solver from another – without proprietary algorithms like friction, FE packages can quickly reduce to a library of standard element formulations, which are mostly common to all solvers.
In this study, the performance and accuracy of a number of nonlinear FE solvers in predicting friction are characterized. A case study is presented involving a small scale, dynamic, friction dominated system designed to isolate and challenge FE solver friction algorithms. The arrangement of interconnected bodies within the system is analogous to an isolated section of a complex structure, such as the tensile armor and surrounding layers of a flexible, or a series links in a mooring chain. The system is exercised through each solver to produce a set of nonlinear time-history outputs using a range of solution parameters, including both lagrangian and penalty enforcement methods. The fidelity of the produced response, as well as the sensitivity to solver parameters is analyzed in detail and evaluated with respect to a predefined set of performance criteria relating specifically to frictional behavior.
A number of important findings are made relative to the friction algorithm in the different FE solvers. These findings include accuracy/performance at the recommended (default) penalty settings, as well as an expanded discussion on the observed effects. Since the case study system can be thought of as representative of interfacing components within an assembly, the findings correlate directly to internal kinematics which governs both stress levels and wear rates. The performance criteria and findings established herein highlight several deficiencies which can impact design and analysis, and encourages due diligence in selecting FE solvers and parameters for systems involving friction.