Floor slipperiness measurement: friction coefficient, roughness of floors, and subjective perception under spillage conditions

Abstract

Measurement of the coefficient of friction (COF) between the shoe/sole and the floor is essential in understanding the risk of slipping accidents. In this research, the COF of five floor materials commonly used on a university campus, under five surface conditions including dry and four liquid spillage conditions, were measured. The COF measurements were conducted using a Brungraber Mark II slip tester with four footwear materials: leather, neolite, ethylene vinyl acetate, and blown rubber. The results of the COF measurements showed that floor tile, footwear material, and surface conditions were all significant factors affecting the COF. Interactions between these factors were also significant. Four surface roughness parameters (R_a, R_tm, R_pm, R_q) of the five tiles selected in the friction measurement were measured using a profilometer. The roughness of the two ceramic tiles was significantly higher than the three non-ceramic tiles. The correlation between the four roughness parameters and the measured COF was very high (r=0.932 to 0.99) under both wet and water–detergent conditions. The tile and surface conditions in the friction measurements were presented to 24 subjects and the subjective evaluation of floor slipperiness was determined. The differences of the scores from the five surface conditions were statistically significant. The difference under floor tile conditions with the same spillage condition was, however, not significant. Spearman's rank correlation coefficients between subjective score and measured COF using neolite footwear were in the range of 0.8–0.975 for the five floors under all the surface conditions. This implies that subjective scores may reasonably reflect floor slipperiness measured with the Brungraber Mark II slip tester using neolite footwear pad.

Introduction

Accidents caused by slips and falls on slippery surfaces present a significant safety problem (Leamon, 1992; Swensen et al., 1992; Grönqvist, 1995; Leclercq et al., 1995). Foot slips on floors are due to insufficient friction between the sole and the floor. The control of slipping events requires the establishment of a friction standard for the shoe/floor combination and the use of materials that meet this standard. Friction between the shoe and the floor may be determined statically or dynamically, the former is the static coefficient of friction (SCOF) and the later is the dynamic coefficient of friction (DCOF). The DCOF is expected to be a determining factor affecting slipperiness, as the foot is in motion when the shoe comes in contact with the floor (Andres and Chaffin, 1985; Tisserand, 1985). In practice, during normal walking conditions, the contact time between the sole and the floor is so short that SCOF may not be relevant (Perkins, 1978). Brungraber (1967), on the other hand, claimed that SCOF was the most significant parameter affecting slip resistance of floors. Perkins and Wilson (1983) also suggested that SCOF is a better indicator of slipperiness since it determines whether a slip will be initiated. The measurement of SCOF is usually easier than that of DCOF, since the later involves complicated control of the motion between the two contact surfaces. A measured SCOF of 0.5 has been adopted as a safety standard in the USA (Miller, 1983).

It is generally accepted that smooth surfaces are more slippery than rough surfaces. The COF between the shoe sole and the floor has been shown to be highly dependant on the roughness of the floor surface (Chang et al., 2001a). The roughness of a surface may be determined by using various surface roughness parameters. The arithmetical average of surface heights (R_a) is a commonly used one. Stevenson et al. (1989) indicated that DCOF under contaminated conditions, measured with a dynamic setup to simulate human slips, increases almost linearly with R_a, and increases only somewhat beyond certain R_a values. Grönqvist et al. (1990) reported that Pearson's product-moment correlation coefficient between the DCOF for glycerol contaminated floors and R_a was 0.87, with p<0.001. They suggested that an adequate R_a value for a proper slip resistance should be about 7–9 μm.

In addition to R_a, other roughness parameters have also been discussed. Harris and Shaw (1988) reported strong correlation (ρ=0.83) between the average maximum peak to valley distance in each cut-off length (R_tm) and users' opinions of floor safety. Manning et al. (1990) and Manning and Jones (1994) reported that the rank correlation coefficients between the measured friction and R_tm of shoe surfaces were 0.64 and 0.757 for wet and oily surfaces, respectively. Chang (1998) reported that the DCOF and the average of the maximum height above the mean line in each cut-off length (R_pm) was as high as 0.97. Chang (2001) expanded his study to include three different footwear materials on porcelain tiles with four different contaminants. The results showed that the R_pm had a strong correlation (r=0.77 to 0.86) with DCOF measured on tiles contaminated by an 85% glycerol solution.

Spillage is common in many public and working areas. Water and detergent solutions are common due to leakage and/or floor maintenance. On the university campus where the first author served, the floors of the kitchens of the 15 restaurants are almost always covered with liquids during most meal serving periods. Spillage is the primary source of these contaminants. Spillage may occur when cooking oils are poured into and/or removed from the vat or the pan. The oil on the floor immediately after the leakage may be very thick. It may then be spread to the adjacent area and mixed with water by the workers' shoes after repeatedly walking over the area. Spillage also happens in the serving areas of the campus restaurants, and even walkways in other buildings, when people take their meals to certain locations to eat. Spillage not only occurred in the restaurants and food serving areas but also in the machine shops, laboratories, garages, and certain equipment storage areas. Liquid leakage from a container, a machine or a vehicle is the primary source of these contaminations. A pool of contaminant may accumulate on the floor if the leakage is not removed immediately. The friction of a floor surface is altered when covered with liquids. Liquids of varying viscosities produce varying lubricating effects between the shoe and the floor. The effect is to separate the shoe sole and the floor, thus reducing the friction available (Grönqvist, 1995; Leclercq et al., 1995). Manning and Jones (2001) pointed out that oil contamination is the most dangerous because the DCOF values on such floors are invariably lower with oil contamination than with water.

In addition to tribological effects, a subject's perception of floor slipperiness is also essential in slip prevention as the subject may then manipulate gait patterns when walking on a slippery surface to reduce the probability of a slip. The floor slipperiness is initially judged by the subject's visual perception. Myung et al. (1993) compared the subjective ranking of slipperiness and the DCOF of ceramic, steel, vinyl, plywood, and sandpaper. Their results indicated that humans have a promising ability to subjectively differentiate floor slipperiness with a reliable confidence rating for the tested surfaces, even though the slipperiness difference might not be large. They concluded that humans were reliable, but risky, discriminators of floor slipperiness. Cohen and Cohen (1994a) asked their subjects to visually compare 23 tested tiles to a standard tile with a SCOF of 0.5 and judge whether the tile was more slippery. They found a significant number of disagreements between subjective responses and the SCOF values of the tiles, in contrast to the findings of Myung et al. (1993). In a follow-up study, Cohen and Cohen (1994b) exposed 8 subjects to 10 outdoor walking surfaces under both dry and wet conditions. The subjects observed and then walked over each surface under each condition before rating their perception of floor slipperiness on a one-to-seven scale. Pearson's correlation coefficients between the DCOF of the surfaces and the subjective ratings were calculated. The authors found that the correlation was weak for the dry condition (r=0.045 and 0.241 for `observed' and `experienced' ratings, respectively) and moderate for the wet condition (r=0.407 and 0.677 for the two ratings, respectively). The results from both of the studies (Cohen and Cohen, 1994a, Cohen and Cohen, 1994b) indicated that humans' perceptions of floor slipperiness might be quite different from the actual traction of the floor as measured by COF. A false perception of floor slipperiness may result in an inappropriate gait pattern and result in slippage of the foot on the floor.

Friction has been commonly adopted as an indicator of slipperiness. Measurement of the COF between footwear material and floor has been the subject of much research (Stevenson et al., 1989; Manning et al., 1990; Grönqvist, 1995; Leclercq et al., 1995; Chang and Matz, 2001). Extensions of friction measurement to roughness measurement have also been reported (Grönqvist et al., 1990; Manning and Jones, 2001; Chang, 1998, Chang, 1999, Chang, 2001, Chang, 2002a; Chang et al., 2001a). In addition to friction and roughness measurements, subjective measurement has also been discussed (Swensen et al., 1992; Myung et al., 1993; Cohen and Cohen, 1994a, Cohen and Cohen, 1994b; Grönqvist et al., 2001). However, most of the studies involving tribology, surface geometry, and subjective measurements addressed at most two factors at a time. Field studies that combine friction, roughness measurement and subjective scoring are rare. In addition, spillage of oil has not been addressed in friction measurements in the previous studies. The objectives of this study were:

• to measure the COF of five commonly used floor tiles on a university campus under one dry and four liquid-spillage conditions using four footwear materials;

• to measure the roughness of the selected floor tiles;

• to investigate the perceived floor slipperiness by human subjects for the floor-spillage conditions; and

• to discuss the correlations between the measured COF and perceived floor slipperiness.