Elsevier

Applied Ergonomics

Volume 32, Issue 2, April 2001, Pages 173-184
Applied Ergonomics

The effect of surface roughness and contaminant on the dynamic friction of porcelain tile

https://doi.org/10.1016/S0003-6870(00)00054-5Get rights and content

Abstract

Surface roughness affects friction, but it is not clear what surface roughness characteristics are better correlated with friction. The average of the maximum height above the mean line in each cut-off length (Rpm) and the arithmetical average of surface slope (Δa) had the highest correlation with dynamic friction coefficient in a previous study. The previous study was expanded to two different footwear materials and four different contaminants on a porcelain tile in the current investigation. The results showed that dynamic friction decreased as the interface speed and glycerol content in the contaminant were increased due to the hydrodynamic lubrication effect. Δa had the highest correlation with friction for most of the test conditions with neolite. For Four S rubber, friction coefficient appeared to have the highest correlation with the parameters related to the surface void volume at 30% glycerol content, related to the surface slope at 70 and 85% glycerol contents, and related to the peak to valley distance at 99% glycerol content. A good indicator of surface slip resistance probably should consist of the surface parameters representing the surface slope, the surface void volume and the surface peak-to-valley distance with the coefficients determined by the system parameters.

Introduction

Slips and falls are a serious problem. Based on the information provided by an earlier study (Leamon and Murphy, 1995), the annual direct cost of occupational injuries due to slips and falls in the US may be as high as 7 billion dollars. The total cost due to slips and falls for the whole population in the US is enormous. The common perception of fall injuries might be related to falls from elevation. However, falls from elevation, usually resulting in a higher claim cost, represented only 35% of all claim cases related to the problem. On the contrary, falls on the same level accounted for 65% of claim cases and, consequently, 55% of claim cost in the total direct workers’ compensation for the occupational injuries due to slips and falls (Leamon and Murphy, 1995).

Human gait related to slips and falls was investigated by Strandberg and Lanshammar (1981), Lanshammar and Strandberg (1983), Perkins and Wilson (1983), Leamon (1988), Leamon and Li (1990), and Myung et al. (1992). Their findings led to the conclusion that a higher friction could potentially improve slip resistance as discussed previously (Chang, 1998).

The applicability of static and dynamic friction to human walking was addressed by Ekkubus and Killey (1973), Lanshammar and Strandberg (1983) and Tisserand (1985). Based on the evidence provided by Lanshammar and Strandberg (1983) and Tisserand (1985), it is likely that dynamic friction is more applicable to human walking than static friction.

Surface roughness also plays a role in floor slipperiness. Using the hydrodynamic squeeze-film theory, Proctor and Coleman (1988) demonstrated that certain surface roughness is needed to improve slip resistance. Harris and Shaw (1988) assessed the subjective opinions of 10 contaminated floor surfaces. In their study, the Spearman rank correlation between the opinion ranking in slip resistance and the surface parameter Rtm (the averaged peak-to-valley distance) was 0.83 with p<0.1. Stevenson et al. (1989) varied the roughness on steel and concrete surfaces, and measured their slip resistance with a dynamic setup to simulate human walking. They concluded that dynamic friction in contaminated conditions generally increases almost linearly with the arithmetical average of floor surface roughness (Ra), and increases only slightly beyond certain Ra values. Manning et al. (1990) and Manning and Jones (1994) used a traction walking test to assess shoe slip resistance on a contaminated floor. They used several methods of abrasion to simulate the polishing effect of walking. They reported that the Spearman rank correlation coefficients between Rtm on shoe surfaces and the measured friction were 0.64 (p<0.05) and 0.757 (p<0.025) for wet and oily surfaces, respectively. Grönqvist et al. (1990) used another dynamic apparatus to simulate human walking for measuring the slip resistance of several contaminated floor materials. They reported that the Pearson's product-moment and the Spearman rank correlation coefficients between the measured friction and Ra of floor surface roughness were 0.87 and 0.86, respectively, with p<0.001. Lloyd and Stevenson (1992) used the same setup as Stevenson et al. (1989) to measure the slip resistance of several contaminated floor materials. They introduced a roughness index (RI) which contained Rq (the root mean square of surface heights), λq (the root mean square of wave length) and Rsk (the skewness of surface heights). They reported that the multiple correlation coefficient between friction and RI was 0.983 (p<0.001). Chang (1998), Chang (1999)) used sand blasting process to alter the surface roughness on quarry tiles. He reported (1998) that the correlation coefficient between dynamic friction and the average of the maximum height above the mean line in each cut-off length (Rpm) was as high as 0.97 for the averaged surface parameters generated from eight roughness measurements. He also reported (1998) that the correlation coefficient between dynamic friction and the arithmetical average of surface slope (Δa) was as high as 0.89 for the surface parameters generated from a roughness measurement. Chang (1999) used several slipmeters to measure the slip resistance of quarry tiles. He reported that the correlation coefficients between the surface parameter Rpk, and the slip resistance measured under wet condition with the Ergodyne and the Brungraber Mark II were 0.85 and 0.63, respectively.

The limitations on the surface parameters and test materials used in the majority of the published studies were discussed previously (Chang (1998), Chang (1999)), and only key points are summarized here. The surface parameters Ra and Rtm were used by Harris and Shaw (1988), Stevenson et al. (1989), Manning et al. (1990), Manning and Jones (1994) and Grönqvist et al. (1990). However, these two surface parameters are highly location dependent and reflect very limited surface characteristics although they might have a good correlation with the measured friction. Several floor materials were used to cover the desired range of surface roughness by Harris and Shaw (1988), Grönqvist et al. (1990) and Lloyd and Stevenson (1992), and most of these friction measurements reflected the combined effect of floor materials and surface roughness. Only Stevenson et al. (1989) varied the surface roughness on identical floor materials, but they used Ra to represent the surfaces. Manning et al. (1990) and Manning and Jones (1994) varied the surface roughness on identical shoe materials, but they only used Rtm to represent the shoe surfaces. Although the Rtm values of three floor surfaces were reported, they did not investigate the correlation between friction coefficient and floor surface roughness parameter. Footwear materials are usually not as hard as floor materials. Shoe materials are likely to have a much larger deformation than floor materials when in contact. Therefore, the surface roughness on footwear is less critical than that on floors.

In this paper, a previous study (Chang, 1998) is broadened to vary the footwear materials used and the viscosity of contaminants. Also, the quarry tiles used in the previous study are replaced by porcelain tiles to eliminate potential problems with the porosity in the tiles in retaining a consistent amount of contaminants on the surface during the measurements. A commercially available pin-on-disk tester is used to measure the dynamic friction between a footwear material and a tile. Also, a commercially available profilometer is used to measure the surface profiles of tiles. Twenty-two surface parameters are correlated with the measured friction. The objective of this study is to identify the surface parameters that have the highest linear correlation with the measured friction with different footwear materials and contaminants used in this experiment.

Section snippets

Test apparatus and design of experiment

The experiment was run in an environmental chamber. The relative humidity was kept between 48 and 52% and the temperature was 21°C (70°F). The contact interface between footwear and floor can be generally divided into the squeeze-film, draping and traction zones (Strandberg, 1985). The film thickness of the contaminant in the squeeze-film zone is usually much larger than the magnitude of floor surface roughness. In the draping and traction zones where the footwear has a direct contact with the

Results

To investigate whether the sand blasting processes resulted in differences in surface parameters, principal component analysis on 22 surface parameters distilled 22 variables into three columns of component scores that accounted for 87% of the variance of 22 variables. A 7×6 between subject ANOVA (analysis of variance) was performed with sand blasting process and tile as independent variables. Three component scores were used as the dependent measures. The results demonstrated the main effect

Discussions

As expected, friction coefficient decreased as glycerol content was increased based on the results shown in Table 4, Table 5, and Fig. 2, Fig. 3, Fig. 4. As the speed was increased, friction coefficient also decreased and became less sensitive to the footwear material as shown in Fig. 2. As the glycerol content was increased, friction coefficient became less sensitive to the velocity as shown in Fig. 4. This indicates that more normal load was supported by the hydrodynamic lubrication of the

Conclusions

As expected, dynamic friction coefficient decreased as the speed at the interface and the glycerol content in contaminant were increased due to the hydrodynamic lubrication effect. The correlation between the surface parameter and the measured friction at the glycerol contents of 70 and 85% was higher than that at 30 and 99% glycerol contents. Among 22 surface parameters evaluated in the correlation, Δa had the highest correlation with friction for most of the test conditions with neolite. For

Acknowledgements

The author likes to thank Mr. Ilya Bezverkhny, Ms. Joanne Gouin, Mr. Gary Hampel and Mr. Richard Holihan for their technical assistance, and Ms. Lobat Hashemi for her help in statistical analysis during the course of this study. Four S rubber used in this experiment was provided by RAPRA Technology Limited.

References (28)

  • American Society for Testing and Materials F-609-79, 1984 (Reapproved). Standard method of test for static slip...
  • G.M. Bartenev et al.

    The actual contact area and friction properties of elastomers under frictional contact with solid surfaces

    Rubber Chem. Techno.

    (1972)
  • R. Brough et al.

    Measurement of coefficient of friction of floors

    J. Phys. D

    (1979)
  • W. Chase et al.

    General Statistics

    (1986)
  • Cited by (52)

    • Effects of roughness parameters on slip resistance for different methods used to determine the coefficient of friction for ceramic floor tiles

      2021, Ceramics International
      Citation Excerpt :

      Then, to break the adhering areas, a friction force is created [13]. Although the contact interface between a shoe sole and floor will be complex, the surface roughness plays a crucial role in determining the surface friction [14]. According to Chang [15], the surface roughness affects the friction, but it is not clear which characteristic of surface roughness presents the best correlation with the COF.

    • Investigation of Floor Surface Finishes for Optimal Slip Resistance Performance

      2018, Safety and Health at Work
      Citation Excerpt :

      However, further increases in floor surface roughness did not provide additional benefits to the slip resistance performance under all three environments. This result was consistent with other studies reporting functional relationships between the floor surface roughness and DFCs against contaminated conditions [13,19,24–27]. It was also found that there was a lack of correlation between the floor types and DFCs under the wet environment.

    • Effectiveness, durability and wear of anti-slip treatments for resilient floor coverings

      2015, Safety Science
      Citation Excerpt :

      Both these factors are related to the viscoelastic properties of the elastomer and depend on the surface roughness of the contacting materials. Clear evidence that surface roughness is a key factor influencing the slip resistance of floor coverings came from laboratory experiments in which the surface roughness state of different materials was varied artificially (Chang, 2001, 2002; Kim and Smith, 2000; Wieder, 1988; Wilson, 1996). Derler et al. (2005) found that the smoothing of rough floor coverings is a common surface wear mechanism in practice, changing the surface and frictional properties faster and stronger than those wear processes that lead to the roughening of initially smooth surfaces (for the same materials and equivalent mechanical influences).

    • Functional levels of floor surface roughness for the prevention of slips and falls: Clean-and-dry and soapsuds-covered wet surfaces

      2013, Applied Ergonomics
      Citation Excerpt :

      The significant correlation between the surface roughness and DFC values under the lubricated condition clearly supports the main theory of the role of floor surface roughness on slip resistance performance in this study. This result is consistent with previous studies reporting functional relationships between the surface roughness and DFC values under contaminated conditions (Chang, 1999, 2001; Manning and Jones, 2001). However, the lack of correlation between the surface roughness and DFC values under the clean-and-dry surface condition indicates involvement of multiple mechanisms of slip resistance properties and possible strong effects of other material characteristics such as wear development during the dynamic friction tests.

    View all citing articles on Scopus
    View full text