Isogeometric random rough surfaces contact analysis using piecewise local contact detection

Authors:
Publisher website
Google Scholar
Abstract
A random rough surface topography significantly influences the surface contact result. In existing studies that solve the contact problem with random rough surfaces by Isogeometric analysis (IGA), finding the slave point’s closest point on the master random rough surface during the contact detection process is often tackled using intelligent optimization algorithms, such as genetic algorithm (GA), which can be computationally expensive. In order to promote the efficiency of contact detection, we propose a piecewise closest point projection algorithm (PCPPA). Since the geometry in IGA is represented by Non-Uniform Rational B-Spline (NURBS), the PCPPA splits the master random rough surface into multiple local regions by its knot vectors’ element values, and defines an influence radius to determine the influenced local regions that should be considered when finding the slave point’s closest point. The traditional closest point projection algorithm is utilized to find the slave point’s local closest point on each influenced local region. The global closest point is selected from these local closest points by comparing their distances to the slave point. Based on the contact problem between an elastomer cube and a rigid cuboid, both of which have random rough surface, hyper-parameter optimization of the GA is carried out firstly to find the optimal hyper-parameters to minimize its computation time while ensuring correct contact detection results. Secondly, by comparing with the optimal GA, the effectiveness and efficiency promotion of the proposed PCPPA are investigated. It is shown that the computation time of the proposed PCPPA is 8%–20% of that of the optimal GA. Finally, the contact problem solving results obtained through the IGA using the PCPPA-based contact detection method, the IGA using the GA-based contact detection method, and the finite element analysis are compared with one another. It is validated that the discrepancies among these results are very small.
Funding Information
  • National Natural Science Foundation of China (51935009)
  • National Natural Science Foundation of China (52075480)

This publication has 43 references indexed in Scilit: