Investigation of Friction Coefficients of Veneers as a Function of Fibre Direction and Moisture Content

During the manufacture of veneer based moulded parts, veneers move against one another. Friction is caused due to this movement. Different conditions, such as gluing or fi bre direction, could infl uence the friction coeffi cients and thus the moulding process. For a better understanding of the manufacturing process of veneer based moulded parts, it is important to know which parameters infl uence friction and friction coeffi cients. In this paper, results of friction investigations are presented. Thereby the moisture content of the used veneers was varied as well as the fi bre direction. Considering the manufacture conditions, the investigations were also conducted with glue-coated veneers. The results prove an infl uence of fi bre direction but this infl uence depends on moisture content.


INTRODUCTION 1. UVOD
Moulded parts made of laminated wood are used as chair shells, built-in parts in sophisticated interior furnishing, loudspeaker enclosure or the like.In general, the manufacture of three-dimensional veneerbased moulded parts is conducted in heatable shaping presses.Thereby several veneer layers are stacked.During the moulding process, veneers shift against one another, causing friction.Depending on the layering of veneers (parallel or crosswise stacked veneer), there are different shift scenarios.The relative movement between veneer layers can also take place at an angle of 45°, depending on the used mould.
In the referred studies, different infl uencing variables were considered, such as sliding speed, normal load, temperature, moisture content or fi bre direction.In each study different wood species were used.All scientists in the referred studies have determined the static (μ st ) and sliding (μ sl ) friction coeffi cient.
The infl uence of sliding speed was investigated by McKenzie and Karpovich (1968)  The infl uence of the normal load was investigated by Ramananantoandro et al. (2007) and Murase (1984).Murase (1984) has conducted friction tests with Western hemlock.Murase has varied the normal load between 2.3 kPa and 65 kPa, and Ramananantoandro et al. (2007) between 23 kPa and 228 kPa.According to both studies, the friction coeffi cients are independent of the normal load.
The infl uence of moisture content was the object of investigations conducted by McKenzie and Karpovich (1968) and Murase (1984).McKenzie and Karpovich (1968) investigated wood conditioned to 12 % moisture content and wet, water soaked and dripping wood.The main conclusion concerning the woodwood friction resulting from their study was that water soaked samples have clearly higher friction coefficients than wood with 12 % moisture content.They did not differentiate between the varieties of wood species they had investigated.Murase (1984) has varied the moisture content of wood between air dry and water saturation in four steps.He has confi rmed the results of McKenzie and Karpovich (1968) and found that both friction coeffi cients increase when moisture content increases.Murase (1984) explained that the increase in friction coeffi cients with increasing moisture content is due to a higher adhesion between wood and wood.The hydrogen bonding is considered to be responsible for this adhesion.
The fi bre direction, as an infl uencing variable, was investigated by McKenzie and Karpovich (1968) and Ramananantoandro et al. (2007).McKenzie and Karpovich (1968) have investigated three sliding situations: a) surface parallel to the fi bre direction, sliding parallel to the fi bre direction; b) surface parallel to the fi bre direction, sliding perpendicular to the fi bre direction (cross-cut); c) surface perpendicular to the fi bre direction (cross-cut), sliding perpendicular to the fi bre direction (cross-cut).They have found only small differences due to fi bre orientation.Thus, all results were averaged.Ramananantoandro et al. (2007) have determined friction coeffi cients at different anatomical planes and directions, but the fi bre and sliding direction from unmoved and moved surfaces were always the same.They have concluded that the fi bre direction has no infl uence on the static frictional behaviour, while infl uence was observed on the sliding frictional behaviour.Thus, in the considered fi bre directions of both studies, no infl uence of the fi bre direction on the static friction coeffi cient could consistently be found.Considering the sliding frictional behaviour, Ramananantoandro et al. (2007) have found an infl uence, while McKenzie and Karpovich (1968) could not fi nd any infl uence.However, both authors have investigated different fi bre directions and situations of unmoved and moved planes.
The literature review consistently confi rms the independence of friction coeffi cients on the normal load.Also, the dependence of moisture content was consistently confi rmed.There are contradictions about the infl uence of fi bre direction and sliding speed.
Considering the above mentioned background of these investigations, in this study friction coeffi cients for the moulding process of stacked veneer layers were determined.Subsequently, these coeffi cients shall be used for modelling this process.Thereby, the investigated infl uencing parameters were chosen to be similar to conditions occurring during the moulding process.Different combinations of movement directions were investigated considering the manufacture of three dimensional, veneer-based moulded parts.Such combinations of fi bre and sliding direction are not known from literature.
To achieve higher degrees of deformation, veneers are mostly moistened.Furthermore, the veneers have to be glued for manufacturing moulded parts.That is why different moisture conditions and glued veneers were investigated.

MATERIJALI I METODE
Industrial sliced beech veneer (Fagus sylvatica L.; 1.2 mm thick) was used for the investigations because it represents the most commonly used species for manufacturing veneer based moulded parts.Industrially produced veneers do not have a strictly anatomically oriented surface.The created plane is a cross of the longitudinal plane in one direction and the radial and tangential plane in the other direction.Hence, the tests were conducted with veneers of the L-RT plane.Friction tests were conducted referring to DIN EN 14882 (2005) (Figure 1) in a standard atmosphere at 20 °C, 65 % rh.
The bottom veneer (veneer 1) was the stationary counterpart.Its dimensions were 70 mm x 130 mm.The moving veneer (veneer 2) was pulled across the stationary veneer with a constant normal load and sliding speed.The arrows show the direction of movement.The moved veneer had dimensions of 50 mm x 50 mm and was fi xed to a block.This veneer was pulled over a distance of 100 mm across its counterpart, veneer 1.During this movement, the load F was measured.This load is considered to determine the friction force.The load peak (F st ), occurring just before the movement, was used for calculating the static friction coeffi cient  st (Eq.1). (1) The normal load F N amounted to 59 N. In order to calculate the sliding friction coefficient μ sl , all load values between 10 mm and 100 mm during the sliding process were averaged.The values below 10 mm were not considered in order to exclude the infl uence of the sliding start.The coeffi cient was calculated in analogy to Eq. 1, as follows: (2) According to different scenarios of movements and fi bre directions in a veneer batch during the moulding process, different sliding situations were investigated (Figure 2).The sliding situation describes the fibre directions relative to the sliding direction.The large arrow in Figure 2 indicates the direction of movement (sliding direction).The lines in the veneer illustrations indicate the fi bre direction.0° means the fi bre direction of the veneer is the same direction as the sliding direction; 90° means the fi bre direction runs 90° in relation to the sliding direction, 45° means the angle between fi bre direction and sliding direction amounts to 45°.
Every sliding situation was conducted 5 times, whereas each veneer was used only once.The veneer surfaces were damaged during the friction tests thus each test had to be conducted with new veneer samples.The sliding speed amounted to 400 mm/ min.Basically, the static friction coeffi cient is higher than the sliding friction coeffi cient.That was to be expected, considering the literature.McKenzie and Karpovich (1968) determined a coeffi cient of static friction (μ st ) of 0.6 and a coeffi cient of sliding friction (μ sl ) of 0.45 for wood with a moisture content of 12%.Concerning the water soaked samples, μ st was determined to be 0.85 and μ sl varied between 0.38 and 0.64, depending on sliding speed.Murase (1984) determined the static friction coeffi cient μ st to be at 0.8 and the sliding friction.The comparison of different sliding situations shows an infl uence of fi bre direction on friction coeffi cients.The ANOVA-test (p<0.05)results in the proof of signifi cant differences between the mean values of the friction coeffi cients of the tested sliding situations.This statistical signifi cance could be proven for static as well as for sliding friction for air-dried, water-saturated and adhesive-coated samples.Thereby, the airdried samples show the clearest differences between the various sliding situations.The highest friction values were obtained by movement of two surfaces in the fi bre direction of 90°.

REZULTATI I RASPRAVA
According to the available literature, the combinations of fi bre and sliding direction (the sliding situations) measured in this study have never been measured before.In this respect the results cannot be compared with literature values (McKenzie and Karpovich, 1968; Ramananantoandro et al., 2007).
Basically, the results of this study prove differences of friction coeffi cients when different fi bre directions slide over each other.Previous studies did not result in such clear differences.
The friction coeffi cients of water-saturated samples are clearly higher than the coeffi cients of the airdried samples.That was basically also to be expected, considering the literature, but the known fi ndings could be extended for the regarded sliding situations.
An interesting fact is that the results indicated only small differences between the sliding situations.
In contrast to the air-dried samples that were sliding continuously, all water-saturated samples showed a slip-stick effect.That means a continuous jerking occurred, which might be regarded as alternating between static and sliding friction.Figure 5a shows an example of measuring values of a smooth-running air-dried sample, and Figure 5b  coeffi cient is averaged using all load values.McKenzie and Karpovich (1968) and Murase (1984) explain the higher coeffi cients of wet samples with some additional hydrogen bonding between the two counterparts.The additional hydrogen bonding might be one reason, but other things might increase the friction of watersaturated samples, as well.Due to the high moisture content of wet samples, the samples are swollen.Because of the raised fi bres, the roughness is increased (Csanády et al., 2015).Although Ramananantoandro (2007) could not prove a correlation between roughness and friction coeffi cients, an obstruction of the movement due to raised fi bres is conceivable.Furthermore, the water softens the veneers.Thus, the upper (moving) veneer can be pressed into the counterpart veneer by the normal load, generating a dent.Therefore, a higher drag is created against the movement.That might also explain the slip-stick: the sample sinks in causing it to stop and high power is necessary to restart the movement (power peak); then the sample stops because of sinking in again.
The friction coeffi cients of the samples coated with adhesive behave similarly to the water-saturated samples in terms of differences of fi bre directions, but the friction coeffi cients are lower.The differences between the sliding situations are clearly reduced.80 % of the tests showed a slip-stick.Figure 5c shows a sample diagram.The adhesive layer is very thin.Part of the adhesive will certainly be absorbed by the surface and soften the surface like the water-saturated samples.Probably, therefore, the adhesive cannot undertake the function of a lubricant.A possible reason for the slipstick effect of these samples is the additional adhesion between wood and adhesive that the UF resin elicits.In contrast to water, the adhesive is sticky and thus inhibits the motion.
The results show that neither the water nor the glue act as lubricants.The friction is increased by using either of the two fl uids.Independently of the kind of the fl uids between the friction counterparts, the differences between the fi bre directions are clearly reduced due to the presence of fl uids.Nevertheless, the differences are statistically signifi cant.

ZAKLJUČAK
The following conclusions can be drawn: -The sliding situation and the fi bre direction have an infl uence on the friction.This infl uence depends on moisture content.-The differences between the friction coeffi cients are clearly reduced by water-saturated samples and samples coated with adhesive.Thus, the moisture equalizes the coeffi cients, but it also increases the friction.
-The glue layer does not work as a lubricant.The friction is increased due to the fl uid adhesive.The presence of adhesive elicits a slip-stick-effect by most of the investigated samples.
and Ramananantoandro et al. (2007).McKenzie and Karpovich (1968) have presented a basic study concerning the friction of wood.They investigated a variety of wood species.A milling machine was used for applying friction to the wood.The use of a milling machine indicates the background of these investigations: friction and wear during the wood machining process.The sliding speed was varied between 0 and 2000 mm/min.Ramananantoandro et al. (2007) conducted friction tests with oak with a newly developed friction testing machine that performs the friction by a linear movement.Whereas McKenzie and Karpovich (1968) found an infl uence of sliding speed on friction coeffi cients, Ramananantoandro et al. (2007) could not ascertain an infl uence.

Figure
Figure 3a-c and Figure 4a-c show the test results.Basically, the static friction coeffi cient is higher than the sliding friction coeffi cient.That was to be expected, considering the literature.McKenzie and Karpovich (1968) determined a coeffi cient of static friction (μ st ) of 0.6 and a coeffi cient of sliding friction (μ sl ) of 0.45 for wood with a moisture content of 12%.Concerning the water soaked samples, μ st was determined