Mallorca 1999 - Identification and Characteristics of Surface Components

M05 The Influence of Friction in Turning Movements

Juan V. Durá, Instituto de Biomecánica de Valencia

Introduction

The friction between shoe sole and surface is necessary for walking and running without slipping. In the context of industry, the issue of friction is a safety problem. In this case, the more friction the better, but in the sports case the problem is different. In sports, high friction avoids slipping and permits one to grip a surface better, and this normally permits faster movements. But, if friction is excessive, overload is produced in joints and injuries may occur, especially in sports with fast turning movements.

It might be that the ideal combination of footwear and surface can permit adequate traction in such a way that the athletes or players can accelerate and decelerate while keeping the balance, rate and co-ordination of the movements of the body, arms and legs. During sport practice, frequent changes of direction take place, which cause rotational frictional forces, i.e. friction moments or torques. Friction torques generated at the interface between the footwear and surface should be such that the knee and ankle ligaments are not subjected to excessive tension when the player performs rapid direction changes.

High friction has been related with stress injuries, muscular overload and ligament injuries. This problem has been studied specially in American football and tennis. Powell and Schootman (1992) found that artificial sports surfaces with high friction increase the number of injuries of knee cruciate ligament in American football. Nigg and Segesser (1988) showed that injuries in tennis increase when the surface has more friction.

Although some authors recommend friction coefficients around 0.8 (Frederick 1993, Valiant 1990) and others around 0.5 - 0.7 (Nigg 1988), there are open questions about the effect of friction in performance and safety. The question is to decide what level of friction is adequate for avoiding injuries and if this level will affect performance.

An additional problem is to select a method for measuring friction. The test machines used for measuring friction coefficients have problems for simulating what happen in sports. In the case of sports movements like stops or turnings, people change their movementsas a function of the friction (μ) existing between the footwear and surface. The differences of friction found between different materials when tested with machines is not found in tests with subjects, because they adapt to the surface. For this reason it is necessary to record the movement of people (Frederick 1993).

The aim of this paper is to study the effect of different surfaces in turning movements, in the changes of the movements and in performance, considering the relation with standard measurements of the coefficient of friction (). For measuring the coefficient of friction in a standard way the test machine defined in the German standard DIN18032-2 for sports surfaces has been used. This procedure has been selected because it has been adopted in different European countries, including Spain.

Methods

Five different surfaces were selected and tested with the test machine defined in standard DIN18032-2 (Table 1).

Table 1: Tested surfaces and friction coefficients.

SURFACE	μ	DESCRIPTION
A	0.43	Wood surface, beech parquet
B	0.58	Aluminium
C	0.73	Synthetic surface (PVC)
D	0.77	Synthetic surface (Synthetic rubber)
E	0.92	Asphalt surface with highly abrasive resin coating, used in outdoors tennis courts

This method involves the use of the sliding test apparatus shown in Figure 1. A vertical shaft of diameter 20mm is arranged in a frame, the lower part of which is designed as threaded spindle (of pitch 12mm/turn). The total weight of the shaft, weight and test foot is set to 20Kg ± 1Kg. And the polar moment is set to 3000 ± 200Kg cm2. The test foot has three skids covered with leather.

Figure 1: DIN 18032 Friction Machine

Five healthy young persons, non-elite sportsmen, were selected with the following characteristics:

Age:	From 17 to 24 years
Weight:	From 64 to 69.5 kg
Height:	From 1.75 to 1.82 m
Shoe size:	42 (French scale)

   
The movement performed consisted of a crouched stance as if starting a 100 m sprint, and at the first step, with the right foot, they turn and run in the opposite direction. The subjects were allowed to perform as many repetitions as they felt necessary to adapt themselves to the surfaces, because the objective was to detect the adaptation of the subject to the pavement.

A total of 125 turnings were registered (5 subjects x 5 repetitions x 5 surfaces).

The samples of the surfaces were fixed over a force platform (DINASCAN-IBV) for measuring the forces during the stance phase.

Motion analysis was carried out and the lower limb was divided into four body segments and the axis fixed in each segment following the model defined by Vaughan et al (1992) in the following manner: pelvis, thigh, leg and foot. Three joints were defined as the hip, knee and ankle. Over each segment 3 markers were fixed for 3-D movement analysis and the movement was recorded using three video cameras (50 Hz). In practice, the hip joint data was discarded from the study since the subjects placed their right arm in front of the marker during almost all of the turning motion (Figure 2).

Figure 2: Camera position

The three-dimensional co-ordinates were calculated by means of a DLT algorithm and a smoothing process using the estimated variance of the error defined by Woltring (Woltring 1986) and fifth order B-splines.

The subjects were recorded in standing position and their joint angles calculated. The joint angles were used as the origin of the measure (0 degrees). The relative angular position between segments was calculated using the attitude vector (Woltring 1994).

Different parameters were obtained, and with each of these parameters a multifactor analysis of variance of repeated measures was performed. Subject and surface were considered as factors. A multiple range test of Least Squares Differences (LSD) at 95% was used for post hoc analysis to determine on which surfaces the differences were significant.

Results

During the turning movement two phases were observed, first a braking phase and after this another phase of starting or flying. These phases were separated by a minimum in the vertical force (Figure 3).

Figure 3: Typical vertical force in the turning movement

Thus, different parameters were defined for each of the phases.

The instant that separates the braking and start phases is taken when the vertical force reaches the minimum (F_Vmin). The initial and final instants are when the vertical force is greater and lower than 40 N, respectively.

From the force platform measures, the parameters selected for statistical analysis were:

• The times spend for braking (t_brake) and starting phase (t_start), and total time (t_total).
• The maximum of the vertical force for each phase in body weight units (F_VmaxB and  F_VmaxS Respectively).
• The mechanical impulse in each phase and differentiating in each phase the vertical force (l_VB for braking, l_VS for starting), the horizontal force (l_HB, l_HS) and the resultant force (l_B, l_S). The impulses were calculated from the forces, in body weight units.
• The mechanical impulse without considering phases for vertical force (l_TV), horizontal force (l_TH) and resultant force (l_T).
• The maximum moment (M_max) and the maximum coefficient of friction (μ_max) measured during the movement. The coefficient of friction is calculated dividing the vertical force by the horizontal force at each instant.

The results obtained with the variance analysis are summarised in Table 2. The surface factor had not significant influence on forces (F_VmaxB and F_VmaxS) and mechanical impulses, considering the total time (l_B, l_S).

Table 2: Multifactor Variance Analysis results

					MEANS
PARAMETER	p	ERROR	A	B	C	D	E
ttotal (s)	0.029	±0.009	0.578	0.602	0.573	0.567	0.561
tbrake* (s)	0.000	±0.011	0.088	0.239	0.251	0.236	0.258
tstart* (s)	0.000	±0.013	0.389	0.362	0.321	0.331	0.302
lVB* (s)	0.000	±0.016	0.260	0.304	0.361	0.332	0.363
lVS* (s)	0.000	±0.017	0.486	0.444	0.394	0.402	0.363
lHB* (s)	0.000	±0.009	0.155	0.168	0.207	0.197	0.208
lHS* (s)	0.000	±0.009	0.276	0.233	0.226	0.233	0.208
lB* (s)	0.000	±0.018	0.304	0.348	0.418	0.388	0.421
lS* (s)	0.000	±0.020	0.560	0.503	0.455	0.465	0.420
Mmax (Nm)	0.000	±1.878	14.921	18.117	19.409	20.805	29.465
μmax*	0.000	±0.023	0.845	0.701	0.907	0.880	1.029
θkmin* (°)	0.000	±1.672	-27.904	-57.501	-66.349	-64.250	-66.222

* The interaction surface-subject has been significant.

The curves of joint angles do not show well defined minimum or maximum values that could be used as parameters for the statistical analysis, except the knee and ankle flexion-extension angle during the braking phase. The knee and ankle have a maximum flexion instant during the braking phase. But only the knee minimum extension, θ_kmins, shows significant differences (see Table 2).

θ_min has negative values due to the orientation of the axis. More negative values mean more flexion.

Although the interaction between subject and surface is significant, the tendency observed in the five subjects is similar. Anyway in future research more subjects should be used because the interaction could have an influence in the results obtained.

The results for the multiple range test of Least Squares Differences (LSD) at 95% is shown in Table 3. Only 3 or 2 homogenous groups were found in function of the parameter selected. The groups are numbered from minor to major, for example, surface B belongs to group 2 for t_brake and surface A belongs to group 1, this means that t_brake is significantly lower in surface A. When a surface could belong to two or more groups, more than one number appears in the cell, for example surface A could belong to groups 1 and 2 for the parameter t_total.

Table 3: Homogenous groups. LSD method at 95% 

	A	B	C	D	E
ttotal	1,2	2	1	1	1
tbrake	1	2	2	2	2
tstart	3	3,2	1	2,1	1
lVB	2	1,2	3	2,3	3
lVS	3	3,2	2,1	2,1	1
lHB	1	1	2	2	2
lHS	2	1	1	1	1
lB	1	1,2	2,3	3	3
lS	3	2	2,1	2,1	1
Mmax	1	1,2	1,2	2	3
μmax	2	1	2	2	3
θkmin	2	2	1	1	1

The times have different tendency in each phase (Figure 4). The time spent in the braking phase (t_brake) increases if the coefficient of friction is higher. And time for starting phase (t_start) is lower with higher friction. These different tendencies produce a compensation effect and reduce the differences when performing the turning movement (t_total).

Figure 4: Tendencies of the time means.

The impulses behave in a similar fashion as the times. The impulses increase when the coefficient of friction is high in the braking phase, and decrease in the starting phase. In this case the compensation is so high that significant differences were not found in the total impulses, for the total time.

The maximum moment (M_max) is higher in the surfaces that have a higher coefficient of friction, but the mean values are in a range considered as safe by other authors, i.e. less that 40Nm (Valiant, 1990).

The maximum of knee flexion (or minimum extension θ_kmin) is higher for the surfaces with more friction (C, D and E).

Discussion

Considering first the method used for measuring the coefficient of friction using, the DIN 18032-2 method, the surfaces have a friction coefficient (μ) from 0.4 (A) to 0.9 (E). The maximum coefficient of friction measured in this study is from 0.7 to 1.0. The DIN method does not, therefore, reproduce the values that appear with people. The material used in the DIN method is leather, which is clearly different from the materials used in sports footwear soles. Also the method probably does not simulate the forces and velocities that appear in human movements.

The differences of coefficient of friction between surfaces is lower in the tests with people (0.3 between the surface with the lowest friction and the one with the highest friction) than with DIN method (0.5). This could be explained considering the adaptation and modification of movements. The adaptation appears in the different times taken in each of the two phases of the movement. When the friction force is higher people spend more time in the braking phase and need less time for starting, and this produces the same effect in the mechanical impulses for each phase. In this manner, people use the extra time for braking in more knee flexion, that could be interpreted as a protective mechanism. People try to maintain forces and torques within acceptable limits; when the friction force is higher, the torque increases, but this effect is reduced by people using more time for braking and with more knee flexion. For the people who participated in this study, the upper limit of acceptable torque was around 30 Nm, and this value is coincident with the recommendation of Valiant (Valiant, 1990).

Conclusion

The DIN18032 test machine does not reproduce the forces and friction coefficients that appear in sports movements.

People adapt their movements according to the friction coefficient that appears between the shoe and the surface. This adaptation implies changes in joint flexion and times and, for this reason, the friction coefficient measured with machines is different to the one measured with people.

When the knee is flexed during the braking phase an eccentric contraction of the quadriceps is produced. This movement could produce injuries if the sportsmen are tired. Considering that the time for doing the turning is a parameter that measures the performance of the movement, and that this parameter is not very affected by the different surfaces because there is a compensation between the two phases of the movement, then it is more recommended that the surface have a low coefficient of friction, around 0.4 as measured with DIN18032 method.

Acknowledgements

This work was supported by the Spanish Interministry Commission for Science and Technology (Reference Number SAF94-0518) and JUNCKERS INDUSTRIER A/S.

References

Andréasson, G.; Lindenberger, U.; Renström, P.; Peterson, L.; (1986) Torque developed at simulated sliding between sport shoes and an artificial turf. The American Journal of Sports Medicine, 14 (3), 225-230..
Bonstingl, R.W.; Morehouse, A.C.; Niebel, B.W.; (1975) Torques developed by different types of shoes on various playing surfaces. Medicine and Science in Sports, 7 (2), 127-131.

DIN(1991) DIN 18032-2: Sport halls; halls for gymnastics and games; floors for sporting activities, requirements, testing. Deutsches Institut für Normung.

Frederick, E.C.; (1993) Optimal frictional properties for sport shoes and sport surfaces. In: Biomechanics Sports XI, (eds J. Hamill; T. Derrick & T. Elliott), pp. 15-21.

Gheluwe, B.V.; Deporte, E.; (1992) Friction measurement in tennis on the field and in the laboratory. International Journal of Sport Biomechanics, 8, 48-61.

Nigg, B.M.; Segesser, B.; (1988) The influence of playing surfaces on the load on the locomotor system and on football and tennis injuries. Sports Medicine., 5, 375-385.

Powell, J.W.; Schootman, M.; (1992) A multivariate risk analysis of selected playing surfaces in the National Football League: 1980 to 1989. The American Journal of Sports Medicine, 20 (6), 686-694.

Rheinstein,D.; Morehouse,C.A.; Niebel,B.W.; (1978). Effects on traction of outsole composition and hardnesses of basketball shoes and three types of playing surfaces. Medicine and Science in Sports, 10 (4), 282-288.

Stucke, H.; Baudzus, W.; (1984) On Friction Characteristics of Playing Surfaces. In: Sport shoes and playing surfaces: Biomechanical properties (eds. E.C. Frederick), pp. 89-97. Human Kinetic, Champaign, Illinois.

Torg, J.S.; Stilwell, G.; Rogers, K.; (1996) The effect of ambient temperature on the shoe-surface interface release coefficient. The American Journal of Sports Medicine, 24 (1), 79-82.

Valiant, G.A.; (1990) Traction characteristics of outsoles for use on artificial playing surfaces. In: Natural and artificial playing fields: characteristics and safety features, ASTM STP 1073 (eds. R.C. Schmidt et al.), pp. 61-68.

Vaughan, C.L.; Davis, B.L.; O'Connor, J.C.; (1992) dynamics of human gait Ed.: Human Kinetics Publishers.

Wilson, M.; Peet, M.; (1990) Performance requirements of sports footwear. SATRA Bulletin, September, 96-99.

Woltring, H.J. (1986) A Fortran package for generalised cross-validatory spline smoothing and differentiation. Adv. Eng. Software, 8, 104-113.

Woltring, H.J.; (1994) 3-D attitude representation of human joints: a standardization proposal. J. Biomechanics