Finite element analysis of a polymer to polymer wearing process

Lin, Hungyen and Lin, H. Y. and Kong, L. X. (2006) Finite element analysis of a polymer to polymer wearing process. In: Proceedings of the Fifth International Conference on Engineering Computational Technology :. Civil-Comp proceedings . Civil-Comp Press, pp. 456-463. ISBN 9781905088119

Abstract

The purpose of the study presented here is to establish a finite element (FE) model to simulate the wear process of a polymer-to-polymer sliding system. Two polymer components which are named as "slider" and "ground" are operated on the slant-horizontal contact region reciprocally. A total of 3500 cycles was conducted. The sliding wear motion of the real product is simplified into a two-dimensional model for FE simulation. According to the requirement of the actual application, we are interested in one of effects of the wear performance - the horizontal driving force (HDF) which is applied on the "slider" component to generate the sliding motion [1]. Two commercial polymer materials have been considered in this paper: 33% glass reinforced, heat stabilised black nylon copolymer resin and BK15945% glass reinforced modified polyethylene terephthalate. For the solution of the wear simulation and prediction, the FE analysis software ANSYS and some mathematic wear equations have been used to build an analysis procedure for this polymer contact structure. First, the FE simulation model is applied to gain the contact surface pressure distribution. Then, the mathematical wear equations are used to calculate the wear depth along the wear sliding surface. Next, the contact surface shape with the wear depth datum is modified. Now, a new model with the upgraded geometry is carried out. Repeating the above four analysis steps, the whole wear process will be represented in the simulation model. In this paper, a series of simulations are performed for every 500 operating cycles for a total of 3500 cycles. The major task of ANSYS is to plot the stress and pressure distribution of the fields of contact and the HDF variation with the sliding motion. Both components have been map-meshed and the two-dimensional structure solid element which is designed as PLANE42 in ANSYS is used for the solid body of the model. The contact pair is described by the flexible surface-to-surface contact elements CONTAC172 and the flexible surface-to-surface target element TARGE169. Twenty load steps across the contact surface are setup for gathering the necessary datum (stress, surface pressure, reaction force, etc). To get a better convergence of the iterative solution, two treatments are also configured. First, the calculation algorithm the pure penalty is switched to the augmented Lagrange. Second, the contact stiffness parameter is adjusted to balance the precise result and the acceptable convergence. Furthermore, an approximate method is introduced to simplify the contact geometry of the upgraded model which is generated by applying the wear depth to the original model. Wear equations employed in this paper are proposed by Archard [2] and Yang [3]. The Archard's equation considers the wear in a steady-state and Yang's equation involves the running-in (transient) wear phase. Both of them based on metal materials and have been extended for the application of polymer materials by Ashraf [4]. A number of simulation results are obtained from a sequence of finite element analyses by progressively removing material and are compared with physical testing results. The simulation results of the horizontal force in an operation cycle of 3500, shows that the force increasing along the slant sliding surface and the position where the peak force occurred are very similar to the experiment. The plot of the von-Mises stress distribution from the FE simulation indicates the maximum stress appears at the top of the slant contact surface. Therefore the maximum wear depth should occur at this region. The variation of the peak horizontal force during the 3500 wear cycles demonstrate a dramatic reduction in peak force over the first 1500 cycles, followed by a gradual decrease in the rate of reduction. For subsequent cycles, the force has decreased continually but at a slower rate. It is found that the FE results have the similar trend with the experimental measured results which support the possibility of using the approximate FE simulation model to estimate the wear process in this particular structure. There are several ways to gain more accurate results. First, refine the model mesh and the load steps. Second, abandon the approximate method which is used to simplify the contact surface after wear. It makes the simulation condition more similar to the actual wear status than the approximate model. Third, apply FE simulation for more wear operation cycles. Due to the model simplifications and the input data of the material properties, this simulation model is only effective for this particular product. Further research will be carried out to extend it to more general the application.