Scholarly article on topic 'Effect of motion inputs on the wear prediction of artificial hip joints'

Effect of motion inputs on the wear prediction of artificial hip joints Academic research paper on "Medical engineering"

CC BY
0
0
Share paper
Academic journal
Tribology International
OECD Field of science
Keywords
{"Ultra-high molecular weight polyethylene" / "Artificial hip joints" / "Wear modelling" / "Motion inputs"}

Abstract of research paper on Medical engineering, author of scientific article — Feng Liu, John Fisher, Zhongmin Jin

Abstract Hip joint simulators have been largely used to assess the wear performance of joint implants. Due to the complexity of joint movement, the motion mechanism adopted in simulators varies. The motion condition is particularly important for ultra-high molecular weight polyethylene (UHMWPE) since polyethylene wear can be substantially increased by the bearing cross-shear motion. Computational wear modelling has been improved recently for the conventional UHMWPE used in total hip joint replacements. A new polyethylene wear law is an explicit function of the contact area of the bearing and the sliding distance, and the effect of multidirectional motion on wear has been quantified by a factor, cross-shear ratio. In this study, the full simulated walking cycle condition based on a walking measurement and two simplified motions, including the ISO standard motion and a simplified ProSim hip simulator motion, were considered as the inputs for wear modelling based on the improved wear model. Both the full simulation and simplified motions generated the comparable multidirectional motion required to reproduce the physiological wear of the bearing in vivo. The predicted volumetric wear of the ProSim simulator motion and the ISO motion conditions for the walking cycle were 13% and 4% lower, respectively, than that of the measured walking condition. The maximum linear wear depths were almost the same, and the areas of the wear depth distribution were 13% and 7% lower for the ProSim simulator and the ISO condition, respectively, compared with that of the measured walking cycle motion condition.

Academic research paper on topic "Effect of motion inputs on the wear prediction of artificial hip joints"

Tribology International ■ (■■■■) ■■■-■■■

Contents lists available at SciVerse ScienceDirect

Tribology International

journal homepage: www.elsevier.com/locate/triboint

Effect of motion inputs on the wear prediction of artificial hip joints

Feng Liua'*, John Fishera, Zhongmin Jina,b

a Institute of Medical and Biological Engineering, School of Mechanical Engineering, University of Leeds, Leeds LS2 9JT, UK b School of Mechanical Engineering, Xi'an Jiaotong University, PR China

ARTICLE INFO

ABSTRACT

Article history: Received 31 October 2011 Received in revised form 4 May 2012 Accepted 30 May 2012

Keywords:

Ultra-high molecular weight polyethylene Artificial hip joints

Wear modelling Motion inputs

Hip joint simulators have been largely used to assess the wear performance of joint implants. Due to the complexity of joint movement, the motion mechanism adopted in simulators varies. The motion condition is particularly important for ultra-high molecular weight polyethylene (UHMWPE) since polyethylene wear can be substantially increased by the bearing cross-shear motion. Computational wear modelling has been improved recently for the conventional UHMWPE used in total hip joint replacements. A new polyethylene wear law is an explicit function of the contact area of the bearing and the sliding distance, and the effect of multidirectional motion on wear has been quantified by a factor, cross-shear ratio. In this study, the full simulated walking cycle condition based on a walking measurement and two simplified motions, including the ISO standard motion and a simplified ProSim hip simulator motion, were considered as the inputs for wear modelling based on the improved wear model. Both the full simulation and simplified motions generated the comparable multidirectional motion required to reproduce the physiological wear of the bearing in vivo. The predicted volumetric wear of the ProSim simulator motion and the ISO motion conditions for the walking cycle were 13% and 4% lower, respectively, than that of the measured walking condition. The maximum linear wear depths were almost the same, and the areas of the wear depth distribution were 13% and 7% lower for the ProSim simulator and the ISO condition, respectively, compared with that of the measured walking cycle motion condition.

© 2012 Elsevier Ltd. All rights reserved.

1. Introduction

Artificial joint replacements are effective in providing normal function to many patients suffering from severe joint diseases [1]. The joint replacement treatment has been continuously evolved, from hips and knees to other major synovial joints [2-5]. However, the joint bearings are subject to wear. Wear debris generated mainly from the joint bearing surface accumulates in local tissues, causes adverse tissue reaction, and ultimately leads to implant fixation failure [6]. Wear-induced failure remains a major limiting factor affecting the long-term performance of the joint replacements, particularly for younger and more active patients. Recognition of the wear issue has led to extensive wear studies to predict wear performance, understand wear mechanism and evaluate design factors [7-10].

Wear studies of artificial hip joint bearings have been largely carried out experimentally. Wear tests based on a simple bearing configuration using pin-on-plate testers are useful for identifying wear properties. For example, conventional UHMWPE has been substantially tested to determine variations in wear rates with

* Corresponding author. Tel.: +44 113 343 7472. E-mail address: F.Liu@leeds.ac.uk (F. Liu).

0301-679X/$-see front matter © 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016lj.triboint.2012.05.029

changes in individual tribological system parameters such as contact pressure, cross-shear, or counterface roughness [11-14]. The tests have been extended to develop wear laws for polyethylene joint bearings [15]. For the full-scale wear simulation of an actual joint bearing, joint simulators have been developed to replicate the motion, loading and environment in vivo [16]. However, a precise reproduction of the complex operating conditions is generally difficult [17]. The hip simulation ISO standard for wear assessment defines a standard walking cycle for a standard patient [18]. In many hip simulator designs, kinematics and loading conditions have been further simplified such as by reducing the full three axes rotation to flexion/extension and internal/external rotation [16]. With increasing demands for the implant wear testing under various functional requirements, the simulators that can provide a balanced complexity and accuracy are necessary, and such a design requires a better understanding of the effects of various operating conditions on wear. As an alternative, computational wear modelling is a suitable means to provide a rapid and vigorous comparison between multiple variables in a parametric study [19-23].

Computational wear modelling has improved for polyethylene artificial joints. Brown et al. [24] have pioneered the computational wear simulation of joint implants based on Archard's law and finite element contact models. The approach has been widely

2 F. Liu et al. / Tribology International I (Ulli III-III

Notation CSR cross-shear ratio

L sliding distance

d thickness of polyethylene cup wall V wear volume

p average contact pressure over loading history WPi discretised frictional work in the PMO direction

t time of loading history in minutes Wti discretised frictional work perpendicular to the PMO

A nominal contact area direction

B creep constant d linear wear depth

C wear coefficient dcr creep deformation at a point

applied [25-29], and continually enhanced particularly in an attempt to develop a generic wear model for UHMWPE based on fundamental wear properties [15,30]. Recent advances include the quantification of multidirectional motion effect on polyethylene wear [13], and the contact area dependent wear law [15,31] in which the wear volume is proportional to the nominal contact area and sliding distance. The wear law assumes that the wear coefficient is constant over contact pressures but a function of the cross-shear resulting from the multidirectional motion. The use of the new model has therefore provided an independent wear prediction and improved the prediction agreement with that of experimental tests for both polyethylene hip and knee joints [15,31]. The wear modelling has been further improved to include polyethylene creep in a parametric study on the joint diameters and bearing clearances of polyethylene hip joints [32].

Motion inputs are a major factor in the simulator wear studies [16,17]. Saikko and Calonius [33] have shown considerable differences in slide tracks on acetabular cup bearings resulting from varied motions, which implies a possible reason for causing a large difference in polyethylene wear. The present study focussed on the computational wear prediction of polyethylene cup bearings based on the contact area dependent wear law, and considered three types of motion inputs, the measured walking cycle [34], ISO 14242 recommendation [35], and simplified ProSim simulator testing [36] to justify the simplification of the motion inputs to provide guidance to the simulator design and testing.

2. Materials and methods

Total hip joint replacements consisting of an acetabular cup and a femoral head were modelled as a simple ball-in-socket configuration, as schematically shown in Fig. 1, with respect to a fixed rectangular Cartesian coordinate system (OXYZ) [26,32]. The bearing

Inclination Lateral

Anterior

materials were chosen as the conventional UHMWPE GUR 1050 (non-cross-linked) for the cup and cobalt chrome alloy (CoCrMo) for the head. In the present study, the cup was inclined at 35° under a vertically applied resultant load, which replicates a general cup inclination of 45° in the pelvis with the load vector 10° medially [37]. A nominal 28 mm diameter hip joint with a radial clearance of 0.04 mm and the cup wall thicknesses of 8.0 mm and 4.0 mm, respectively, for the polyethylene bearing and its metallic backing shell was analysed [15]. The bearing materials used in this study are summarised in Table 1.

Three sets of motion and loading conditions were considered to represent a physiological walking cycle of hip joints; the full simulated condition based on a gait measurement [34,19], and the simplified conditions including the ISO [35] and ProSim simulator testing [36], as shown in Figs. 2 and 3 for motion and loading profiles over a cycle, respectively. All the rotation components given above were considered as Euler angles; the rotation transformation was performed following the sequence of flexion/ extension (FE), abduction/adduction (AA) and internal/external rotation (1ER), and the rotation movements were performed on different bearing components for different cases [33]. For the measured walking (Case 1), all three rotations of FE, AA and 1ER were carried by the head with the cup being stationary; for the 1SO condition (Case 2), the FE and AA rotations were conducted on the cup, and the 1ER rotation by the head; 1n Case 3, the simplified ProSim simulator testing, the FE and 1ER rotations were achieved on the cup and head, respectively. The loading was generally considered to be similar with the magnitudes of the averaged loads being, respectively, 1242, 1293 and 1057 N for the above Cases 1, 2 and 3, and the resultant load directions being similarly varied relative to the cup bearing surface [19,35,36]. The details of the range of motions and the associated moving bearings considered are summarised in Table 2 for the measured walking cycle, 1SO and ProSim simulator [35,36].

The details of computational wear modelling for conventional UHMWPE bearings have been previously published, which include the development of the contact area dependent wear law [15], the quantification of the cross-shear effect [15,31],

Table 1

Details of the range of motion and moving components for the three cases considered in this study, the measured walking, ISO 14242 and ProSim simulator conditions.

Simulation Flexion/extension Abduction/ conditions adduction

Internal/external rotation

Range Moving (deg.) bearing

Range Moving Range Moving (deg.) bearing (deg.) bearing

Fig. 1. Schematic of three-dimensional total hip joint with load and motion of flexion/extension about the x-axis, abduction/adduction about the y-axis and internal/external rotation about the z-axis.

Case 1: 30/ - 15 Head 5/-4 Head 6/ - 8 Head

Walking

Case 2: ISO 25/- 18 Cup 4/ - 7 Cup 2/-10 Head

Case 3: 30/ - 15 Head 10/ -10 Cup

ProSim

F. Liu et al. / Tribology International I (Ulli III-III

a 30 -1

<u p 10-

wi a -10-

b 30 -1

¡U 10-

wi a -10-

<u ¡U 10-

Time (s)

> ................

Time (s)

Time (s)

Fig. 2. Motion profiles of (a) flexion/extension, (b) abduction/adduction and (c) internal/external rotation for the measured walking cycle, ISO standard and ProSim simulator (two-axis) conditions.

and the full numerical calculation of wear and creep [32]. The methods were directly applied in this study, and for clarity, the major equations are briefly described in this paper. The volumetric wear was expressed as

V = CAL (1)

where A is the contact area, L is the sliding distance and C is the wear coefficient [15]. The linear wear depth (d) was derived from Eq. (1) by dividing the contact area as

d = CL (2)

The wear coefficient (C) was assumed to be constant over the contact pressure range as experienced in the joint replacements, but dependent on the cross-shear resulting from the multidirectional sliding motion of the bearings. A general theory on the cross-shear effect of polyethylene wear is based on the polyethylene molecular reorientation and strain hardening [38,39]. The polyethylene molecules on the bearing surface align with the principal direction of sliding, defined as principal molecular orientation (PMO), and lead to strain hardening and enhanced wear resistance; the sliding against the principal direction weakens the molecular reorientation and causes more wear. The cross-shear was quantified using a cross-shear ratio, CSR, given by

CSR =J2 WH/J2 W + Wpi)

where Wti and Wpi are the discretised frictional work resolved in two directions, perpendicular to and along with the PMO of the bearing, respectively. The frictional work was calculated as the

3500 3000 2500 2000 1500 1000 500 0

-500 -1000 -1500 3500 3000 2500 2000 1500 1000 500 0

-500 -1000 -1500 3500 3000 2500 2000 1500 1000 500 0

-500 -1000 -1500

Superior-inferior force

0.2. v 0.4" ,0..6\>''0.8

Medial-lateral force

Time (s)

0.2 0.4 0.6 0.8 Time (s)

0 0.2 0.4 0.6 0.8

Time (s)

Fig. 3. Loading profiles of (a) the measured walking cycle, (b) ISO standard and (c) ProSim simulator conditions.

Table 2

Details of the predicted maximum linear depth, wear distribution area, maximum cross-shear ratio, and sliding distance averaged over whole bearing surface for a motion cycle, of a 28 mm diameter bearing surface after 5 million cycles for the three conditions considered, the measured walking, ISO and ProSim simulator conditions.

Max. linear Wear Max. cross- Average

wear depth distribution shear ratio sliding

(mm) area (mm2) distance (mm)

Case 1: Walking 0.1 971 0.35 20

Case 2: ISO 0.09 902 0.3 18

Case 3: ProSim 0.09 842 0.35 21

product of frictional force and sliding distance, and contact pressure determined from the bearing contact model was used to calculate the frictional force. The PMO was considered to be the direction along which the maximum amount of frictional work is released, and was numerically determined by iteratively comparing all the possible PMO directions [15,32].

The ram-extruded non-cross-linked UHMWPE (0 Mrad, GUR 1050) pin and the cobalt-chromium plate with an average surface roughness Ra of approximately 0.01 mm was used to determine the wear coefficient with pin-on-plate wear testing under conditions of the multidirectional motion, varied contact pressures and lubricated environment replicating those of full joint simulation [19]. The cross-shear was achieved by varying the rotation of polyethylene pins and the translation of metal plates in tests. The test was carried out at normal laboratory room temperature (approximately 20°) with 25% solution of new-born calf serum as the lubricant. The details of the pin-on-plate wear test can be

F. Liu et al. / Tribology International l (lllli lll-lll

found elsewhere [19]. The curve-fitted general wear coefficient (C) as a function of cross-shear ratio for the conventional non-cross-linked UHMWPE (GUR 1050) for contact pressure range considered (1-10 MPa) was given by

C = (8.5 x 10"5 + 9.3 x CSR)015 x 10"

A finite element contact model was constructed with linear hexahedral elements for the cup and an analytical rigid body for the

-ProSim

......ISO

- - Gait

1 2 3 4 5 Number of cycles (Million)

Fig. 4. Comparison of the accumulated volumetric wear predicted for the measured walking, ISO standard, and ProSim simulator conditions over the number of simulated cycles for a 28 mm diameter polyethylene cup bearing with a radial clearance of 0.04 mm. (The results for the ProSim simulator condition is based on previous calculation [32].)

head and solved to obtain contact pressure using ABAQUS (version 6.8-1, S1MUL1A, Rhode Island) [15]. Approximately, 2000 elements were used for the polyethylene cup with a typical element length of 0.5 mm to ensure converged contact stress. The polyethylene with Young's modulus, 500 MPa, and Poisson's ratio, 0.4, was modelled as an elastic-plastic material, with initial Von Mises yield stress of 10.8 MPa as tested by Barbour et al. [12] using isotropic GUR412 UHMWPE (molecular weight 4 x 106 g/mol). This data was assumed to be similar to that of the GUR 1050 in terms of elastic-plastic properties considered in the present study.

The motion/loading cycle of 1 s was equally discretised into 16 time points. At each time point, the contact model was solved. For each nodal point on the cup bearing surface, the contact pressures obtained over the loading cycle were used to determine the wear coefficient, and subsequently the linear wear depths at each time point and the accumulated linear wear over the cycle as

d = £ C(CSR) Li

where n is the number of discretised time points in a loading cycle, the cross-shear ratio (CSR), and Li is the sliding distance in the ith interval. The wear depth calculated for a single loading/ motion cycle was then scaled by a geometry update factor that is the number of loading/motion cycles in the interval after which the bearing geometry was modified. For the initial 64,000 cycles

10 15 -15 -10

Fig. 5. Comparison of the contact pressure distributions predicted for (a) the measured walking cycle, (b) ISO standard, and (c) ProSim simulator conditions for a 28 mm diameter polyethylene cup bearing after 5 million cycles, with the averaged contact pressure (MPa) values being illustrated. (The results for the ProSim simulator condition is based on previous calculation [32].)

F. Liu et al. / Tribology International l (lllli lll-lll

(the first 2000 min), 10 intervals evenly selected using a logarithmic scale were considered in order to calculate the high creep strain rate [32], and the geometry update factor was then fixed at 250,000 throughout the rest of simulation [24,26]. The linear creep deformation at a geometric point (Scr) was calculated as

Scr = Bplog(t)d (6)

where B is a creep constant, 7.97/[log(min)) MPa], based on the extruded, unirradiated GUR 4150HP UHMWPE rod stock under constant pressures of 2.0-8.0 MPa in a 37 °C bovine serum reservoir for the polyethylene used [40], p is the average contact pressure over loading history, t is the time length of the loading duration in minutes, and d is the thickness of the polyethylene bearing. The total linear penetration at a geometric point was considered to be the superposition of the linear wear and creep. The bearing geometry was modified by correcting the nodal coordinates of the contact model after each geometry update interval, and the finite element contact model was then recalculated based on the new geometry. The total simulation was carried out for 5 million cycles.

3. Results

The accumulated volumetric wear of the polyethylene cup bearings predicted with the three different motion inputs is compared in Fig. 4, as a function of the number of simulated

cycles, with the wear rates being 14.0, 13.4 and 12.2 mm3 per million cycles, respectively, for the measured walking, ISO and ProSim simulator conditions. The predicted volumetric wear rates of the ISO and ProSim simulator conditions were 4% and 13% lower, respectively, compared with that of the measured walking condition.

Fig. 5a-c shows the comparison of contact pressure distributions predicted on the bearing surface corresponding to the first peak loads of the three simulation conditions after 5 million cycles. The contact pressures predicted were generally comparable in terms of distribution areas and values, with slightly varied peak values of 6.0, 7.5 and 8.0 MPa for the measured walking, ISO and ProSim simulator conditions, respectively, and slightly different locations on the bearings (Fig. 5a-c). The comparisons of wear depths and cross-shear ratios after 5 million cycle simulation are shown in Figs. 6 and 7, respectively. The worn area for the measured walking condition was found to be larger by 7% and 13% (Table 2), respectively, compared with those of the ISO and ProSim simulator conditions, but the maximum linear wear depths were approximately 0.1 mm for all cases (Fig. 6). A good agreement in linear wear depths occurred between the three axes rotation motions of the measured walking and ISO conditions (Fig. 6a and b), while the use of the two axes of rotation led to the greater wear depths at two separate locations on the bearing (Fig. 6c). The above trends in linear wear depths were mainly attributed to the cross-shear ratios, and good agreement was also

0 5 10 15

Fig. 6. Comparison of the wear depth distributions predicted for (a) the measured walking cycle, (b) ISO standard, and (c) ProSim simulator conditions for a 28 mm diameter polyethylene cup bearing after 5 million cycles, with the averaged wear depth (mm) values being illustrated (The results for the ProSim simulator condition is based on previous calculation [32].)

6 F. Liu et al. / Tribology International l (lllli lll-lll

Fig. 7. Comparison of the cross-shear ratio distributions predicted for (a) the measured walking cycle, (b) ISO standard, and (c) ProSim simulator conditions for a 28 mm cup bearing after 5 million cycles, with the averaged cross-shear ratio values being illustrated.

found between the measured walking and ISO conditions (Fig. 7a and b), and the two axes of rotation of the ProSim simulator led to the large cross-shear ratios at separated locations (Fig. 7c). However, the cross-shear ratios were generally comparable in terms of distribution sizes and value ranges between 0 and 0.35 for all conditions despite of the double peaks under the ProSim simulator condition. Moreover, for the non-cross-linked conventional UHMWPE, wear coefficients increased with the cross-shear ratio raised to the power of 0.15, as given in Eq. (4), which showed a sharp increase (3.2-7.7 x 10~10) over low cross-shear ratios (CSR=0-0.01) and were steady (1.0-1.6 x 10~9) over large ratios (CSR=0.02-0.5). The major cross-shear ratio range of 0.02-0.4 was predicted on the polyethylene cup bearings (Fig. 7a-c) and the resulting wear coefficients were generally constant (1.0-1.6 x 10_ 9) over the bearing surfaces.

Fig. 8a-c shows the comparison of sliding distance distributions computed for the cup bearing surfaces. The sliding distances around the main loading areas on cup bearing surfaces were found to be in a comparable range, 9-22, 8-21, and 11-22 mm, respectively, for the three cases. The maximum linear wear depths, the areas of wear distribution, the maximum cross-shear ratios, and the average sliding distances are summarised in Table 3.

The slide tracks of some representative points on the cup bearing surface were computed with respect to the head bearing surface and shown in Fig. 9a-c, for the three conditions. In Fig. 9,

the hemispherical bearing surface was flattened in order to show the slide track accurately without much distortion in a projection plot [33]. Therefore, the radial distance of the flattened surface equals to the distance measured from the pole along the spherical surface to the equator of the hemisphere; for example, the radius of the flattened surface is 22 mm in the present plot (Fig. 9). The plotted slide tracks are similar to those of Saikko and Calonius [33]. The tracks were less smooth for the measured walking condition due to the less smooth motion inputs from the measurement (Fig. 9a).

4. Discussion

Previous experimental studies carried out to investigate the effect of motion inputs on wear using ProSim simulators showed comparable volumetric wear. Barbour et al. [12] reported a typical wear rate of 38 mm3 per million cycles for the three-axis rotation motion and a slightly lower rate of 32 mm3 per million cycles by Bigsby et al. [41] for the two-axis rotation motion. In this study, the corresponding wear rates were 14 and 12 mm3 per million cycles, respectively, for the three-axis and two-axis rotation conditions (Fig. 4). In addition to the wear rate, the trend that the wear is comparable between the full and simplified rotation conditions was predicted by the present model. As pointed out in previous studies on model development [15], the lower wear

F. Liu et al. | Tribology International I (Ulli III-III

-15 -10 -б 0 б 10 15 -15 -10 -б

5 10 15

Fig. 8. Comparison of the sliding distance distributions predicted for (a) the measured walking cycle, (b) ISO standard, and (c) ProSim simulator conditions for a 28 mm cup bearing over a motion cycle, with the averaged values being illustrated (mm).

Table 3

Materials for the joint bearings considered in experimental wear tests including joint simulators and pin-on-plate tester, and material properties used for computational modelling.

Components Material Young's modulus (MPa) Poisson's ratio

Acetabular cup/ Conventional

polyethylene pin UHMWPE

Femoral head/metal Cobalt-chrome 210,000 0.3

plate alloy

Cup backing shell Titanium 210,000 0.3

prediction was mainly due to the underestimation of the wear coefficients. The pin-on-plate wear tests were conducted under constant load other than dynamic as in joint simulators. However, the multidirectional motion of the bearings, the resulting cross-shear and its effect on wear were fully modelled, quantified and incorporated based on pin-on-plate wear tests to study the effects of motion inputs. The contact pressure distributions predicted with respect to the peak loads at 5 million cycles were generally comparable (Fig. 5) under all the three simulation conditions. The linear wear depth distributions generally followed the contact pressure distributions (Fig. 6). The greater wear depths, for example, 0.08-0.1 mm, occurred at the major loading areas on the bearing surfaces (Fig. 6a-c) showing the dominant effect of loading in determining the contact area and consequently wear.

The distributed linear wear for the three simulation conditions was generally comparable in terms of the distribution area and the range of wear depth values, which led to the comparable volumetric wear (Fig. 4).

Based on the multidirectional theory and the assumption that both the molecular reorientation and polyethylene wear are a result of dissipated frictional energy, the cross-shear was quantified as the proportion of frictional work perpendicular to the PMO to the total frictional work [13,15]. For UHMWPE cup bearings, in addition to loading, variations in motion inputs, which can affect cross-shear ratio, wear coefficients, and sliding distance, would be the major factor affecting wear as discussed below.

The cross-shear ratios for all the three simulation conditions markedly varied between 0.01 and 0.35 over the bearing surfaces (Fig. 7), indicating that complicated cross-shear sliding occurred for both the three axes rotation and the simplified two axes rotation. The cross-shear ratios resulting from the three axes rotation of the measured walking and ISO were in good agreement (Fig. 7a and b), while the two axes rotation of the ProSim simulator led to a major variation in cross-shear ratios with the double-peak distribution (Fig. 7c) compared to that with single peak for the three axes rotation conditions. The double-peaked cross-shear ratio led to the double peak distribution of wear depths as demonstrated by the maximum wear depths being predicted at two separate locations on the bearing surface (Fig. 6c). However, the cross-shear ratios for all the three simulation conditions were generally in the range of 0.02-0.2 for the

Fig. 9. Comparison of the slide tracks of the points on the cup bearing surface relative the head bearing surface calculated for (a) the measured walking, (b) ISO standard, and (c) ProSim simulator conditions, for a 28 mm diameter cup bearing (the hemisphere bearing surface is flattened).

points at the main contact areas on the bearing, and the larger cross-shear ratios (CSR > 0.2) were generally distributed close to the peripheries of the main contact areas (Fig. 7a-c). Therefore, the effect of the larger cross-shear ratios (CSR > 0.2) on wear was considerably reduced due to the lack of contact in generating wear compared to those points within the main contact loading areas.

The sliding distances of the points on cup bearing surfaces over a single motion cycle varied in a wide range of 6-24 mm with the averages being approximately 20, 18, 21 mm, respectively, for the measured walking, ISO and ProSim conditions (Fig. 8a-c). However, the larger sliding distances were found to occur mainly at the main contact areas with comparable values ranging from 15 to 22 mm for all the motion inputs, and consequently the larger linear wear depths in these areas in addition to the effects of loading and cross-shear. The computed slide tracks of cup bearing surfaces (Fig. 9a-c) were found to be closely linked with the cross-shear ratios (Fig. 7a-c). In contrast to the linear tracks leading to lower cross-shear ratios, the open slide tracks resulted in larger cross-shear ratios. Therefore, particularly for the ProSim condition, the distributed slide tracks on the bearing surface gradually changing shapes from elliptical to linear in the central region of the distribution (Fig. 9c) led to the decreased cross-shear ratios for those points near the centre and consequently the double peak distribution (Fig. 7c). Additionally, for the maximum cross-shear ratios under all the three conditions, 0.3 or 0.35 (Fig. 7) the corresponding slide tracks are comparable in the elliptical shapes as highlighted in the example in Fig. 9a-c.

In hip joint wear simulation of a normal walking condition, motion inputs can be simplified to lead to generally comparable

wear as illustrated by the three cases considered in this study. The ISO condition closely replicated the three rotations of the measured walking, despite slight differences in the phases of AA and 1ER rotations; the simplified ProSim simulator motion was considered to be generally comparable with three rotation motion with AA rotation being removed and 1ER rotation being slightly increased in amplitudes approximately by 30% (Fig. 2b and c). The computational modelling clearly demonstrated the complex cross-shear effect of bearing motion on major parameters in affecting wear such as slide tracks, sliding distances, cross-shear ratios, and wear coefficients. However, the resulting variations in major parameters were generally in a comparable range. The overall volumetric wear was therefore not largely affected by motion inputs. Normal walking is obviously only one of the joint activities. Even under similar activities, both motion and loading conditions can be significantly different for different patients, which can be expected to have more significant impact on wear and should be analysed further. Again, the wear coefficient determination based on the pin-on-plate wear testing to represent the full testing conditions including loading and motion was considered with some limitations. A recent study has shown the possible dynamic loading effect on UHMWPE wear [42]. It has also been recognised that UHMWPE-on-metal hip implants are working under boundary and mixed lubrication condition [43]. The effect of lubrication on wear was only generally incorporated in the wear coefficient by using the same lubricant in the pin-on-plate testing as in the simulator. The Young's modulus of 500 MPa used for UHMWPE in this study was lower than that of UHMWPE after in vivo oxidation as considered in other studies [24,25]. The

F. Liu et al. / Tribology International l (Ulli lll-lll

use of lower Young's modulus can generally lead to larger bearing deformation prediction, but its effect on wear was considered to be small in the present modelling. Additionally, the strain softening effect on wear has been improved recently in which Lee et al. [ 44] pointed out that each sliding component may have an individual softening factor to provide better quantification of the cross-shear effect. All these limitations need further considerations.

5. Conclusions

The effect of motion inputs on UHMWPE wear was quantitatively analysed based on the major parameters in determining wear including cross-shear ratios, sliding distances, slide tracks and wear coefficients. Both the full simulated motion condition using the walking measurement, and the simplified motion simulation, including ISO standard and ProSim simulator conditions generated markedly varied but comparable cross-shear effect on wear over the cup bearing surfaces. The major range of cross-shear ratios (0.02-0.2) was found in the main contact areas on the bearings, and the larger cross-shear ratios (CSR > 0.2) were generally at the peripheries of the main loading areas. The predicted volumetric wear and linear wear depths were generally comparable under all the three sets of simulation conditions. The present study supports the use of the simplified motions of the ISO and the ProSim simulator to simulate the physiological walking conditions in wear prediction.

Acknowledgements

1t was partially funded through WELMEC, a Centre of Excellence in Medical Engineering funded by the Wellcome Trust and EPSRC, under grant number WT 088908/Z/09/Z and additionally supported by the N1HR (National 1nstitute for Health Research) as part of a collaboration with the LMBRU (Leeds Musculoskeletal Biomedical Research Unit). Prof. John Fisher is a N1HR senior investigator.

References

[ 1] Fisher J, Jin Z, Tipper JL, Stone MH, Ingham E. Presidential guest lecture: tribology of alternative bearings. Clinical Orthopaedics and Related Research 2006;453:25-34.

[ 2] Fisher J, Jennings LM, Galvin AL, Jin Z, Stone M, 1ngham E. Knee society presidential guest lecture: polyethylene wear in total knees. Clinical Orthopaedics and Related Research 2010;468(1):12-8.

[ 3] Hall RM, Brown TD, Fisher J, 1ngham E, Mendoza SA, Mayer HM. 1ntroduction to lumbar total disc replacement: factors that affect tribological performance. Proceedings of the 1nstitution of Mechanical Engineers: Part J, Journal of Engineering Tribology 2006;220:775-786.

[4] Gill LH. Challenges in total ankle arthroplasty. Foot & Ankle 1nternational 2004;25(4):195-207.

5] Nam D, Kepler CK, Nho SJ, Craig EV, Warren RF, Wright TM. Observations on retrieved humeral polyethylene components from reverse total shoulder arthroplasty. Journal of Shoulder and Elbow Surgery/American Shoulder and Elbow Surgeons 2010;19:1003-12.

[ 6] 1ngham E, Fisher J. Biological reactions to wear debris in total joint replacement. Proceedings of the 1nstitution of Mechanical Engineers: Part H, Journal of Engineering in Medicine 2000;214:21-37.

[ 7] Clarke 1C. Wear of artificial joint materials 1V Hip joint simulator studies. Engineering in Medicine 1981;10:189-98.

8] MeKellop H, Clarke 1, Markolf K, Amstutz H. Friction and wear properties of polymer, metal and ceramic prosthetic joint materials evaluated on a multi-channel screening devices. Journal of Biomedical Materials Research 1981;15:619-53.

9] Hammerberg EM, Wan Z, Dastane M, Dorr LD. Wear and range of motion of different femoral head sizes. Journal of Arthroplasty 2010;25(6):839-43.

10] Livermore J, 1lstrup D, Morrey B. Effect of femoral head size on wear of the polyethylene acetabular component. Journal of Bone & Joint Surgery 1990;72A:518-28.

11] Saikko V. Effect of contact pressure on wear and friction of ultra-high molecular polyethylene in multidirectional sliding. Proceedings of the 1nsti-

tution of Mechanical Engineers: Part H, Journal of Engineering in Medicine 2006;220:723-31.

12] Barbour PSM, Barton DC, Fisher J. The influence of contact stress on the wear of UHMWPE for total replacement hip prostheses. Wear 1995;181-183:250-7.

13] Kang L, Galvin AL, Brown TD, Jin Z, Fisher J. Quantification of the effect of cross-shear on the wear of conventional and highly cross-linked UHMWPE. Journal of Biomechanics 2008;41(2):340-6.

14] Cooper JR, Dowson D, Fisher J. Macroscopic and microscopic wear mechanisms in ultra-high molecular weight polyethylene. Wear 1993;162-164A:378-84.

15] Liu F, Galvin AL, Jin Z, Fisher J. A new formulation for the prediction of polyethylene wear in artificial hip joints. Proceedings of the 1nstitution of Mechanical Engineers: Part H, Journal of Engineering in Medicine 2011;225:16-24.

16] Barbour PSM, Stone MH, Fisher J. A hip joint simulator study using simplified loading and motion cycles generating physiological wear paths and rates. Proceedings of the 1nstitution of Mechanical Engineers: Part H, Journal of Engineering in Medicine 1999;213:455-67.

17] Goldsmith AA, Dowson D. Development of a ten-station, multi-axis hip joint simulator. Proceedings of the 1nstitution of Mechanical Engineers: Part H, Journal of Engineering in Medicine 1999;213:311-6.

18] 1SO Standard 14242-1:2002. 1mplants for surgery-wear of total hip joint prostheses. Part 1: loading and displacement parameters for wear testing machines and corresponding environmental conditions for test (1nternational Standardization Organization, Switzerland).

19] Kang L, Galvin AL, Jin Z, Fisher J. A simple fully integrated contact-coupled wear prediction for ultra-high molecular weight polyethylene hip implats. Proceedings of the 1nstitution of Mechanical Engineers: Part H, Journal of Engineering in Medicine 2006;220:33-46.

20] Bouchard SM, Stewart KJ, Pedersen DR, Callaghan JJ, Brown TD. Design factors influencing performance of constrained acetabular liners: finite element characterization. Journal of Biomechanics 2006;39(5):885-93.

21] Jourdan F, Samida A. An implicit numerical method for wear modeling applied to a hip joint prosthesis problem. Computer Methods in Applied Mechanics and Engineering 2009;198(27-29):2209-17.

22] Ong KL, Rundell S, Liepins 1, Laurent R, Markel D, Kurtz SM. Biomechanical modeling of acetabular component polyethylene stresses, fracture risk, and wear rate following press-fit implantation. Journal of Orthopaedic Research : Official Publication of the Orthopaedic Research Society 2009;27(11):1467-72.

23] Strickland MA, Dressler MR, Taylor M. Predicting implant UHMWPE wear in-silico: a robust, adaptable computational-numerical framework for future theoretical models. Wear 2012;274-275:100-8.

24] Maxian TA, Brown TD, Pedersen DR, Callaghan JJ. The Frank Stinchfield Award. 3-Dimensional sliding/contact computational simulation of total hip wear. Clinical Orthopaedics and Related Research 1996;333:41-50.

25] Bevill SL, Bevill GR, Penmetsa JR, Petrella AJ, Rullkoetter PJ. Finite element simulation of early creep and wear in total hip arthroplasty. Journal of Biomechanics 2005;38:2365-74.

26] Liu F, Leslie 1, Williams S, Fisher J, Jin Z. Development of computational wear simulation of metal-on-metal hip resurfacing replacements. Journal of Biomechanics 2008;41(3):686-94.

27] Matsoukas G, Willing R, Kim 1Y. Total hip wear assessment: a comparison between computational and in vitro wear assessment techniques using 1SO 14242 loading and kinematics. Journal of Biomechanical Engineering 2009;131:1011-22.

28] Goreham-Voss CM, Hyde PJ, Hall RM, Fisher J, Brown TD. Cross-shear implementation in sliding-distance-coupled finite element analysis of wear in metal-on-polyethylene total joint arthroplasty: intervertebral total disc replacement as an illustrative application. Journal of Biomechanics 2010;43(9):1674-81.

29] Zhao D, Sakoda H, Sawyer WG, Banks SA, Fregly BJ. Predicting knee replacement damage in a simulator machine using a computational model with a consistent wear factor. Journal of Biomechanical Engineering 2008;130(1) 011004-1_10.

30] Kang L, Galvin AL, Fisher J, Jin Z. Enhanced computational prediction of polyethylene wear in hip joints by incorporating cross-shear and contact pressure in additional to load and sliding distance: effect of head diameter. Journal of Biomechanics 2009;42(7):912-8.

31] Abdelgaied A, Liu F, Brockett C, Jennings LM, Fisher J, Jin Z. Computational wear prediction of artificial knee joints based on a new wear law and formulation. JBiomechanics 2011;44(6):1108-16.

32] Liu F, Fisher J, Jin Z. Computational modelling of polyethylene wear and creep in total hip joint replacements: effect of the bearing clearance and diameter. Proceedings of the 1nstitution of Mechanical Engineers: Part J, Journal of Engineering Tribology 2012;226(6):552-63.

33] Saikko V, Calonius O. Slide track analysis of the relative motion between femoral head and acetabular cup in walking and in hip simulators. Journal of Biomechanics 2002;35:455-64.

34] Johnston RC, Smidt GL. Measurement of hip-joint motion during walkingevaluation of an electrogoniometric method. The Journal of Bone and Joint Surgery 1969;51-A:1083-94.

35] Kaddick C, Wimmer MA. Hip simulator wear testing according to the newly introduced standard 1SO 14242. Proceedings of the 1nstitution of Mechanical Engineers: Part H, Journal of Engineering Tribology 2001;215:429-41.

36] Galvin AL, Tipper JL, Jennings LM, Stone MH, Jin Z, 1ngham E, et al. Wear and biological activity of highly crosslinked polyethylene in the hip under low serum protein concentrations. Proceedings of the 1nstitution of Mechanical Engineers: Part H, Journal of Engineering in Medicine 2007;221:1-10.

F. Liu et al. / Tribology International l (lllli lll-lll

[37] Goldsmith AA, Dowson D. A multi-station hip simulator study of the performance of 22 mm diameter zirconia-ultra-high molecular weight polyethylene total replacement hip joints. Proceedings of the Institution of Mechanical Engineers: Part H, Journal of Engineering in Medicine 1999;213:77-90.

[38] Wang A. A unified theory of wear for ultra-high molecular weight polyethylene in multi-directional sliding. Wear 2001;248(1-2):38-47.

[39] Turell M, Wang A, Bellare A. Quantification of the effect of cross-path motion on the wear rate of ultra-high molecular weight polyethylene. Wear 2003;255:1034-9.

[40] Lee K, Pienkowski D. Compressive creep characteristics of extruded ultra high molecular weight polyethylene. Journal of Biomedical Materials Research 1998;39:261-5.

[41] Bigsby RJA, Hardaker CS, Fisher J. Wear of ultra-high molecular weight polyethylene acetabular cups in a physiological hip joint simulator in the anatomical position using bovine serum as a lubricant. Proceedings of the

Institution of Mechanical Engineers: Part H, Journal of Engineering in Medicine 1997;211:265-9.

[42] Saikko V, Kostamo J. RandomPOD—a new method and device for advanced wear simulation of orthopaedic biomaterials. Journal of Biomechanics 2011;44(5):810-4.

[43] Jin Z, Medley JB and Dowson D. Fluid film lubrication in artificial hip joints. Tribological research and design for engineering systems. In: Dowson D, Priest M, Dalmaz G, Lubrecht AA, editors, Proceedings of the 29th Leeds-Lyon symposium on tribology. Amsterdam: Elsevier; 2003. p. 237-256.

[44] Lee RK, Korduba LA, Wang A. An improved theoretical model of orientation softening and cross-shear wear of ultra high molecular weight polyethylene. Wear 2011;271:2230-3.