The Relationship between Roughness of Finished Wood Floors and Slip Resistance

The present study investigates the relationship between the roughness of beech wood and oak wood surfaces treated with oil and polyurethane coating and the slip resistance in dry, water-wet and oily conditions. Pendulum tests were conducted for slip resistance assessment, and roughness measurements were performed by stylus instrument using Ra, Rt, Rp, Rz and Rsm parameters for surface roughness evaluation. Slip potential in dry conditions was low for all finished wood floors studied. Contamination of the surface with water and oil reduced the slip resistance of finished oak and beech flooring. The strong negative correlation was found between slip resistance on dry finished flooring and roughness parameters Ra, Rz, Rt and Rp, and positive correlation between slip resistance on water-wet finished flooring and roughness parameters Ra, Rz, Rt and Rp. Moreover, the correlations between roughness parameters Ra, Rt, Rp and Rz and slip resistance were very similar, and the roughness parameters correlated more strongly with the slip resistance on dry and water-wet surfaces than with the slip resistance on oil-wet surface. Comparison of the slip potential classifications of finished wood floors based on pendulum data and based on Rz surface roughness parameters showed that in some cases the Rz parameter appeared to overestimate the slip potential of the floors in wet conditions. The results confirm previous research that roughness measurements should only be used as a guide and should not be used as the only indicator of the slip potential of wood flooring materials.

ured friction indices for wet surfaces and parameters R a and R z for dry surfaces. Kim (2018) conducted dynamic friction tests among three shoes and nine fl oor specimens under different slippery environments and showed signifi cant effects of fl oor surface roughness parameters (R a , R t , R tm ) on slip resistance performance under soapy and oily conditions. Li et al. (2004) found very high correlation (r=0.932 to 0.99) between the four roughness parameters (R a , R tm , R pm , R q ) of fi ve fl oors and the measured coeffi cient of friction under wet and water-detergent conditions. Shaw (2007) reported moderate correlation between the roughness parameters R p , R t , R q , R a , R z and R y and wet PTV (Pendulum Test Value, closely related to coeffi cient of dynamic friction) and strong correlation between a particular combination of parameters (R p /RS) and wet pendulum values on a small sample of data from a range of different fl oor surfaces. This study was extended to over 100 fl oor samples and it was established that R p (height of the roughness peak) roughness parameter formed the strongest relationship between any single parameter and wet pendulum values. A strong relationship between wet PTV and R p /RS was confi rmed with a larger sample of data (Shaw et al., 2009). Surface roughness measurements are widely used as a secondary indication of slip resistance potential.
According to UK Health and Safety Executive (HSE-GEIS2, 2012) and Health and Safety Laboratory (HSL), R z (R tm ) is a useful parameter for the prediction of the likely slip resistance of a fl ooring material under water (and other fl uid) contamination.
There have been very limited studies of the effects of wood fl oor roughness on slip resistance. It has been shown that the relationship between surface micro roughness and slip resistance of the pre-engineered wood fl oors is complicated and, in some cases, there was disagreement between surface roughness and pendulum results (Loo-Morrey, 2007).
The aim of this paper is to investigate the relationship between the roughness of different wood surfaces treated with oil and polyurethane coating and the slip resistance in dry and water-wet and oily conditions.

UVOD
Slip resistance is an important feature of fl oor safety and can be defi ned as the ability of a surface to substantially reduce or prevent the risk of someone slipping (CCAA, 2003). Slip resistance is very complex because the likelihood of slipping is a function of many factors such as fl oor surface, footwear, environmental conditions, physical condition, etc. Falling mainly happens due to insuffi cient friction between the shoe sole and the fl oor, and the coeffi cient of friction (COF) is commonly accepted as an indicator of fl oor surface slipperiness level. The higher the COF is, the higher the degree of anti-slippery (slip resistance effect) will be (Chen et al., 2015). According to literature, factors affecting the results of friction measurement are fl oor materials, fl oor roughness, liquid/solid contaminants on fl oor, the groove design of shoes and the friction measurement device used (Liu et al., 2010;Chen et al., 2015). There is no generally accepted method of measuring slipperiness.
The Pendulum Tester is the most widely used for measuring the slip resistance of fl oorings. The device relies on the measurement of the coeffi cient of friction between a rubber slider and the fl ooring to assess the resistance to slip (Mijović et al., 2008). This method is used in the standard HRN EN 13036-4:2012 and also in Technical Specifi cation HRS CEN/TS 15676:2010 for determining slip resistance of wood fl ooring.
It has been shown that the coeffi cient of friction between the shoe sole and the fl oor is highly dependent on the roughness of the fl oor surface (Stevenson et al., 1989;Chang et al., 2001;Li et al., 2004). Chen et al. (2015) reported that shoe materials, fl oor roughness and liquid viscosity signifi cantly affected slip resistance. Various surface roughness parameters were used in scientifi c papers to determine the relationship between the roughness of the surface and slip resistance. Stevenson et al. (1989) reported that slip resistance of concrete and steel surfaces measured with dynamic friction testing machine increased with the arithmetical average of roughness (Ra).
Good correlation between dynamic friction and roughness parameter R pm of unglazed quarry tiles surfaces was reported by Chang (1998). Chang (1999) used different slipmeters for investigation relationship among slip resistance of unglazed quarry tile, surface roughness and surface conditions. It has been shown that the effect of surface roughness on friction index depended on the slipmeter used, and that rougher surface generally led to a higher friction index. Among 21 evaluated surface roughness parameters, R pk and R pm parameters had the highest correlation with the meas-........Miklečić, Jirouš-Rajković: The Relationship between Roughness of Finished Wood... ter ( Figure 2) manufactured by Taylor-Hobson on ten marked locations on which the slip resistance of the surface was measured. The measuring speed, radius and angle of conical stylus tip were 1 mm/s, 5 mm and 90°, respectively. Roughness measurement was carried out in the direction perpendicular to the wood grain over traverse of 4 mm and roughness profi les were fi ltered with a cut-off value of 0.8 mm using Gaussian fi lter. For the evaluation of surface roughness, fi ve parameters were used: R a , R t , R p , R z and R sm . Defi nition of used roughness parameters can be seen in Table 1.

REZULTATI I RASPRAVA
Results of slip resistance of oak and beech wood samples were generally referred to the slip resistance of the coating on the wood surface ( Figure 3). It can be seen that the contamination of the surface with water and linseed oil reduced the slip resistance of fi nished oak and beech wood. The highest slip resistance was measured on dry surfaces, followed by water-wet surfaces and oil-wet surfaces. Lemon and Griffi ths (1997) reported that liquids with higher viscosity required higher levels of surface roughness to provide equivalent levels of slip resistance, as the thickness of a squeezed fi lm formed between the fl ooring and treads increased as liquids viscosity increased. Polyurethane coating eliminated the infl uence of structural unevenness of the wood surface on the slippage as it created a dry coating fi lm on the surface. This is the reason for very small differences in slip resistance between oak and beech fi nished with PU coating. Furthermore, the structural unevenness of the wood surface became prominent on the oil-fi nished specimens because the oil did not form a fi lm on the wood surface. Thus, the slip resistance of the water-wet surface was higher on the oil-fi nished samples than on the PU-fi nished samples. However, the slip resistance on a dry surface was higher on PU-fi nished samples than on oil-fi nished samples, which can be attributed to the additives for slip resistance in the coating. Furthermore, oil-fi nished samples showed a greater difference between slip resistance of water-wet and oil-wet surfaces than PUfi nished samples. This could be due to raised wood fi bers due to wetting of the surface with water. This is also the reason why the slip resistance of oil-fi nished beech wood samples contaminated with water was greater than the slip resistance of oil-fi nished oak wood sam-and (50±5) % relative humidity (RH) to the constant mass. Wood samples were fi nished with two-component, solvent-based polyurethane coating (PU) and two-component oil based on isocyanates. Before applying PU coating, wood samples were hand-sanded with paper grit size P80-P120-P180 and wood samples fi nished with oil were hand-sanded with paper grit size P120. For each type of coating, six samples (three radial-textured and three tangential-textured) were prepared. Coatings were applied with a brush in the amount of 110 g/m 2 for PU coating and 80 g/m 2 for oil per layer. PU coating was applied in three layers (one layer of base coat and two layers of top coat) with a 4 hours drying time between the base and top coat and 24 hours drying time between layers of top coat. The dried base coat was hand-sanded with paper grit size P240. Oil was applied in one coat with wiping excess oil from the wood surface after 15 minutes of application of the oil. Surface fi nished samples were conditioned for seven days at (23±2) °C and (50±5) % RH before testing of slip resistance and roughness.

Slip resistance 2.2. Klizavost
The slip resistance measurement was made using pendulum test equipment ( Figure 1) and slider 55 on a dry surface, surface contaminated with distilled water and on surface contaminated with linseed oil according to HRS CEN/TS 15676:2010. For each type of wood texture, surface fi nishing and surface contamination, ten measurements on different places along the grain on the wood surface were made, and average slip resistance was calculated. For measuring slip resistance on wet surface, each measuring place on the sample was moistened evenly with the test fl uid and rubber slider was wiped and cleaned after each measurement.

Surface roughness 2.3. Hrapavost površine
Three samples for each type of wood species, texture and coating were evaluated. Roughness was measured with Surtronic S-126 stylus-type profi lome-  tively small, and a greater difference in slip resistance between radial-textured and tangential-textured samples could only be seen on oil-wet surface of oak wood samples. It can be assumed that oil-wet surface of tan-ples contaminated with water. Beech wood has a higher swelling coeffi cient than oak, so the fi bers on beech wood are more raised than on oak wood. Differences in slip resistance due to the texture of wood were rela-  gential-textured oak wood samples had a lower slip resistance compared to oil-wet surface of radial-textured wood samples due to a higher share of latewood, which has smaller pores compared to earlywood. Table 2 shows the means and standard deviations of roughness measurements of the eight fi nished wood fl oor surfaces. For the R a roughness parameter, there are no variations of means among different wood species fi nished with PU coating. The average R a value of the oil fi nished radial oak wood surface is greater than the average R a value of the oil fi nished beech wood surface, which can be explained by the differences in the anatomical structure of these two types of wood.
For R z and R t the differences between the oil fi nished surfaces and those fi nished with polyurethane are much greater. Since oil is a penetrating fi nish that does not form a fi lm on the surface, the substrate itself greatly affects the results of roughness measurements. The results of measuring the roughness of radial and tangential surfaces differ much more for oil fi nished samples than for polyurethane fi nished samples. The average R p values of the radial and tangential wood surfaces do not differ much for the surfaces fi nished with polyurethane coating, while for the oiled surfaces there is a difference in the values of R p between the radial and tangential surfaces. It can be seen from Table 1 that the average R sm values are much higher for polyurethane fi nished wood surfaces than for oil fi nished surfaces. Roughness parameter R sm is the measure of the spacing between the peaks of the surface profi le and the R sm values are infl uenced by the thickness of the fi lm, that is, the application of the polyurethane coating.
Spearman rank correlation coeffi cients between slip resistance and roughness parameters of surface fi nished beech and oak wood samples are presented in Tables 3 and 4. It can be seen that there is a signifi cant negative correlation between slip resistance of a dry surface and the type of coating. This was expected because the PU coating forms a fi lm on the wood surface, while the wood absorbs oil and thus the structure of the oil-fi nished wood surface also affects the slip resistance. Furthermore, a strongly negative correlation between slip resistance of dry surfaces of oak and beech wood and roughness parameters R a , R t , R p and R z was observed. However, the correlation of slip resistance of a dry surface with R sm parameter was signifi cant and negative and was higher on beech than on oak wood. This can be attributed to the higher standard deviation   of roughness parameters R sm on oak than on beech wood ( Table 2). Slip resistance of water-wet surface had a positive correlation with roughness parameters R a , R t , R p and R z on beech and oak wood, and this correlation was higher on beech than on oak wood. Furthermore, a signifi cant negative correlation of slip resistance on water-wet surface and roughness parameter R sm on beech wood was found, whereas on oak wood this correlation was not signifi cant. It can also be seen that the correlation of slip resistance and roughness parameters R a , R t , R p and R z was higher on water-wet beech wood surface than on a dry surface. For slip resistance on oil-wet surface, no correlation was found with the investigated roughness parameters on beech wood, while on oak wood there was a small correlation between slip resistance on oil-wet surface and roughness parameters R a , R t , R p and R z . However, slip resistance on oil-wet surface was in a strong correlation with wood texture on beech and oak wood. The obtained correlation between the parameter R a and the slip resistance on water-wet and oil-wet surfaces is less than the correlation obtained by Lie et al. (2004) on the ceramic fl oors.
According to the results shown in Table 3, it can be seen that the correlations between roughness parameters R a , R t , R p and R z and slip resistance are very similar and it can be said that no roughness parameter deviate. Furthermore, it can be seen that the roughness parameters correlated more strongly with the slip resistance on dry and water-wet surfaces than with the slip resistance on oil-wet surface.
The technical specifi cation (HRN CEN/TS 1567) prescribes a pendulum test for determining slip resistance of wood fl ooring but does not provide slip resistance ratings (or does not provide interpretation of slip resistance data, or classifi cation). The results of slip resistance measurements and R z roughness measure-ments in this study were interpreted according to the UKSRG Guidelines (HSE-GEIS2, 2012). The interpretation of pendulum results is shown in Table 5 (HSE-GEIS2, 2012). According to UK Slip Resistance Group, Rz roughness parameter gives a good indication of fl oor slipperiness in water contaminated conditions. However, the roughness measurement should be considered as a complementary measurement to be used in conjunction with pendulum test values. Slip potential classifi cation, based on R z microroughness values, is shown in Table 6 (HSE-GEIS2, 2012).
Pendulum results on dry and wet wood surfaces and slip potential in dry and wet conditions are given in   Table 7. Mean average values of the R z parameter and slip potential in water-wet conditions predicted by R z parameter (according to UKSRG Guidelines) are also given in Table 7. It can be seen that slip potential in dry conditions was low for all studied fi nished wood fl oors. The oil-fi nished wood fl oors exhibited moderate slip potential in water-wet conditions, while PU-fi nished wood fl oors showed high slip potential in water-wet conditions. Kim (2018) showed that the fl oor fi nishes require different levels of surface roughness for different types of environmental conditions to effectively control slip potential. Slip potential in oily conditions was shown to be high for all fi nished wood surfaces except oiled oak wood radial surfaces, where slip resistance was shown to be moderate. Comparisons of the slip potential classifi cations of fi nished wood fl oors based on pendulum data and based on R z surface roughness parameters show that in two cases the R z parameter appears to overestimate the slip potential of the fl oors in wet conditions. This result, as well as results reported by Lo-Morrey (2007), indicates that the parameter R z is not recommended as the sole selection criteria for selecting a new fl oor. The parameter R z should be considered together with the pendulum measurements in both wet and dry conditions before making a specifi cation decision.

ZAKLJUČAK
According to the results obtained in this study, it can be concluded that contamination of the surface with water and linseed oil reduces the slip resistance of fi nished oak and beech fl ooring. Furthermore, the viscosity of the contaminant has a greater effect on reducing the slip resistance on fl ooring fi nished with penetrating coating materials, while on fl ooring fi nished with fi lm-forming coating materials, the viscosity of contaminant has little effect on changing the slip resistance. Based on the results of the roughness measurement, it can be concluded that the oil-fi nished surface has a greater infl uence on the roughness than the surface fi nished with polyurethane varnish. Moreover, the correlations between roughness parameters R a , R t , R p and R z and slip resistance are very similar and the roughness parameters correlate more strongly with the slip resistance on dry and water-wet surfaces than with the slip resistance on oil-wet surface. According to Table 7 Slip potential classifi cation, based on pendulum test values (PTV) a and R z microroughness values (applicable for water-wet pedestrian areas) Tablica 7. Vjerojatnost poskliznuća prema testu klatnom (PTV) a i R z vrijednosti mikrohrapavosti (odnosi se na vodom zalivene površine za hodanje)

Roughness
R z , mm Hrapavost R z , mm HSE-GEIS2 (2012), the slip potential in dry conditions is low for oil and PU-fi nished wood fl oors, while the oil-fi nished wood fl oors exhibited moderate slip potential in water-wet conditions and PU-fi nished wood fl oors showed high slip potential in water-wet conditions. Furthermore, slip potential in oily conditions was shown to be high for all fi nished wood surfaces except oiled oak wood radial surfaces, where slip resistance was shown to be moderate. The slip potential based on R z surface roughness parameter indicates that the parameter R z should be considered together with the pendulum measurements in both wet and dry conditions before making a specifi cation decision.