Evaluation of the perceived elasticity of sheepskin using a three-axis force sensor system

Abstract

Sheepskin refers to the hide taken from a sheep and is composed of a layer of skin and the attached hair. It has unique fabric handle properties, and sheepskin quality assessments generally depend on a subjective hand evaluation via sensory test. Therefore, it is necessary to measure the fabric handle of sheepskin mechanically. For this study, we focused on perceived elasticity, and attempted to clarify the relationship between the tactile sensation and the mechanical properties of sheepskin through the development of a new three-dimensional tactile sensation measurement system (3D-TSMS) to measure the compression properties in three axial directions. For compression testing, the compression parameters were calculated as follows: compression linearity (LC), compressional energy (WC), compressional resilience (RC), compressional flexibility coefficient (α _max), pressure relieving energy (ΔE _PR), and continuous mean deviation (CMD), etc. As a result of compression testing, the sheepskin samples which were highly rated for their perceived elasticity showed a convex curve during the initial compression process. In the statistical comparison analysis, LC, CMD, and ΔE _PR showed a correlation to the perceived elasticity of sheepskin.

Keywords

compression property fabric handle human sensory perceived elasticity sheepskin three-axis force sensor

Sheepskin is used to produce sheepskin leather products and soft wool-lined clothing or coverings, including gloves, hats, automotive seat covers, mats for babies and the disabled, and pelts. In particular, it has been found to be effective in reducing the incidence of pressure ulcers¹ and in giving drivers a more comfortable seat in their vehicles.² Hand evaluation via sensory test is an important evaluation method when determining the quality and usage of sheepskin. The palpate methods used when evaluating the fabric are different depending on the attributes of the evaluator, such as hand evaluation experts versus users, and male versus female, as wellas for fabrication methods, materials, and uses. Moreover, the evaluation terms used for hand evaluation are different according to the evaluation objectives.³ ^– ⁷

The Kawabata Evaluation System has been developed to objectively measure fabric handle.⁸ This system is useful not only for fabric but also for paper, film, etc. However, with this system it is difficult to measure the mechanical properties of thick materials such as sheepskin and thick-piled fabric.

Sheepskin refers to the hide taken from a sheep and is composed of a layer of skin and the attached hair. It has unique fabric handle properties that distinguish it not only from other fabrics, but also from leather products and pile fabrics. Moreover, it is difficult to mechanically predict and control its texture based on its material properties since sheepskin is a natural material and varies according to each animal. So, quality assessment of sheepskin depends mainly on the hand evaluation via sensory test. Therefore, it is necessary to measure the fabric handle properties of sheepskin mechanically. Sheepskins are evaluated in Japan using unique evaluation terms such as DANRYOKU-KAN (perceived elasticity), KEGOMI-KAN (richness and fullness of hair), KESABAKI-KAN (smoothness and softness of hair), etc. During hand evaluation, the palm and areas between the fingers are used in addition to the finger tips.

For this study, we focused on perceived elasticity (DANRYOKU-KAN), which is one of the fabric handle properties of sheepskin. In general, when a human manually evaluates the elasticity of thick fabric, they push it with their finger tips and palm, and we assumed that elasticity is expressed in response to the multi-axial compression properties. Therefore, sensing technology (or a device) which is capable of measuring multi-axial load is necessary. However, existing compression testers can only measure vertical compression properties.

The purpose of this study was to clarify the relationship between the tactile sensation and the mechanical properties of sheepskin. Therefore, we developed a new three-dimensional tactile sensation measurement system (3D-TSMS) to measure the compression properties along three axes. The compression properties ofsheepskin were measured multi-axially, and mechanicalparameters describing these properties were calculated. The perceived elasticity of sheepskin was evaluated using the sensory test. Statistical analyses were carried out to determine the compression property parameters which had an influence on the perceived elasticity of sheepskin. In this thesis, however, the relationships between the perceived elasticity and the apparent details of sheepskin samples such as hair length, density, and count were not considered, even though they related to each other, since tests for the details of sheepskins are less efficient than the sensory test and instrumental test, and these tests cause damages to sheepskins.

Experiments

Samples

Five kinds of mouton with different levels of hair density and hair count were used: fine, thin, medium, dense, and thick. Sample levels are named in consideration of their hair density and hair count. Table 1 shows the details of the mouton samples.

Table 1.

Sample details

Sample name	Apparent thickness* (mm)	Hair length (mm)	Hair density (/cm²)	Hair count (tex)
Fine	31.4	40.5	4262	15.8
Thin	33.9	42.6	2269	17.0
Medium	33.0	41.6	4494	15.5
Dense	32.5	41.2	4629	16.1
Thick	33.1	42.4	2220	25.3

Excluding skin thickness.

Sensory test

Twenty Japanese students participated in this test on a voluntary basis (male: 8, female: 12; age range: 22–23). Participants compared the perceived elasticity (DANRYOKU-KAN) of each mouton sample against a reference mouton in Japanese. The reference mouton was a medium mouton cut to the length of 25 mm, which is usually used in bedclothes in Japan. Participants determined how much greater or less the elasticity of the sample mouton was in relation to the reference mouton using a seven-point scale (see Figure 1). The sample moutons were presented to participants at random and sample names were not given to them tominimize any potential bias. Participants were instructed in their mother tongue to compress the mouton with their palm and finger tips. The surface of the mouton pairs was covered with a board with a hole 10 cm long and 20 cm wide in order to ensure consistent control of the palpate method. The test environment was at a temperature of 23 ± 2°C and a relative humidity of 60 ± 5%.

Figure 1.

Perceived elasticity (DANRYOKU-KAN) rating scales.

Compression testing with the three-dimensional tactile sensation measurement system

3D-TSMS was performed to measure the mechanical properties of sheepskin.

Three-dimensional tactile sensation measurement system

When a mouton is palpated to assess its fabric handle, the surface of the mouton is pushed and stroked with the finger tips and the palm. As this creates a multi-axial force, a three-axis force sensor was required to replicate and simulate this phenomenon. The 3D-TSMS was developed to simulate the palpate motions of a human hand (Figure 2). The 3D-TSMS is composed of a capacitance three-axis force sensor (PD3-32-05-015, NITTA Corp., Japan) and a 3D Cartesian manipulator (CCR-M14-0403025, NIDEC SANKYO Corp., Japan). The force sensor can measure the applied loads contacting the mouton in three axial directions. The 3D-TSMS is controlled by two computers: a manipulator controller (SC3000, NIDEC SANKYO Corp., Japan) and a load measurement computer. The load which is generated when the sensor tip comes into contact with the mouton is detected by the force sensor and the output of the sensor is converted from analog to digital by the DAQ card (National Instruments Corp., USA), and then transmitted to the measurement computer. Therecorded load value is then analyzed by a program designed with LabVIEW (National Instruments Corp., USA). The manipulator can be controlled by transmitting the control signal corresponding to the load value to the manipulator controller through RS232C. The sampling frequency of the force sensor is 50 Hz, and the signal is filtered through a low-pass filter of 0.3 Hz and input into the measurement computer.

Figure 2.

Schematic of the three-dimensional tactile sensation measurement system (3D-TSMS).

Measurement

A brass disk-shaped indenter (diameter 50 mm) was used as a sensor tip to evaluate the compression properties along the Z-axis as well as along the X-axis and Y-axis as shown in Figure 3. The mouton was compressed by the manipulator with an operation speed of 2 mm/s. Compression was maintained until the output value of the force sensor along the Z-axis reached 30 gf/cm². The pressure value 30 gf/cm² was set based on the mean value obtained from the hand evaluations of the mouton samples.

Figure 3.

Sensor and indenter.

Changes in pressure and displacement during the compression and recovery processes were measured. The loads along the X-axis and Y-axis were measured during these processes. We were able to calculate the generated force and direction in three dimensions by measuring the three axial loads along with the compression and recovery. The measurement was taken five times. The measurement environment was at a temperature of 20 ± 2°C and a relative humidity of 65 ± 4%.

Results

Sensory test

Figure 4 shows the results of the sensory test. This profile is a mean value taken from among 20 participants. The mouton samples were rated in the following order (from highest perceived elasticity to lowest): dense, medium, fine, thick, and thin. Also, with the exception of medium and fine, the samples displayed significant differences between them when tested by multiple comparisons in one-way ANOVA using Scheffe’s method (dense-medium: p < 0.05 and fine-thick-thin: p < 0.01).

Figure 4.

Perceived elasticity profile.

Compression testing with the three-dimensional tactile sensation measurement system

Figure 5 shows the pressure-displacement curves obtained from along the Z-axis of the sensor.

Figure 5.

Pressure–displacement curve from compression testing.

Figure 6a, b, and c show the load-displacement curves obtained from along the X-axis and Y-axis of the sensor and their resultant force. In the Figures, ‘displacement’ represents Z-axis displacement.

Figure 6.

Load–displacement curves. (a) Load applied along the X-axis. (b) Load applied along the Y-axis. (c) Resultant force in the X-axis and Y-axis.

In Figure 6c, the curve for the dense sample shows a convex load change with a displacement of approximately 5 mm to 13 mm. The curves of other mouton samples show similar load changes, with the exception of the thin mouton. This shows that there is directionality on the hair of the mouton, and that the load applied along the Z-axis when the mouton is compressed is transferred to the X–Y plane. During initial compression the hair is raised, and significant directional transfer of the force applied to the mouton was observed. The hairs start to bend and deform under the progressing compression, and the vertical force is redirected and thus lessened under this state. Then the redirection of the vertical force on the mouton increases again.

Compression property parameters

Eight parameters were derived to evaluate the mouton compression properties based on the results of the compression testing. The following four parameters were measured from the Z-axis pressure-displacement curve using the parameters defined with the KES-F system compression tester (Figure 7).

Figure 7.

Pressure–displacement curve.

Compression linearity LC defines compressive hardness as:

(1)

where the second parameter WC is the compressional energy:

(2)

where P is the compression pressure and l_x is the compressed displacement at the pressure xgf/cm² and WOC is defined by

WOC = P_{30} (l_{30} - l_{1}) / 2

Compressional resilience RC is the extent of recovery, or the regained thickness, when the force is removed:

(3)

where

{WC}^{'}

is the recovery cycle energy:

W C' = \int_{l_{1}}^{l_{30}} P' d l

and

P^{'}

is the pressure in the recovery cycle. Also, the thickness T is measured at the pressure of 1 gf/cm² and reported in millimeters.⁸

In addition to these parameters, the following three parameters were measured:

Compression distance L is measured at the pressure of 1 gf/cm² to 30 gf/cm².

(4)

The compressional flexibility coefficient $α_{max}$ represents ease of compression and is defined by:

(5)

The pressure relieving energy $Δ E_{PR}$ is the area of thehysteresis loop and shows the amount of energy lost during a compression and recovery cycle. It is represented by the margin between compression and recovery energy:

(6)

Furthermore, the continuous mean deviation CMD was measured from the load-displacement curve with the resultant X–Y plane force and Z-axis displacement (Figure 8):

(7)

where,

\bar{F}

is the mean of F, and F is the resultant X–Y plane force, which is calculated as:

F = \sqrt{F_{X}^{2} + F_{Y}^{2}}

where F_X is the X-axis load and F_Y is the Y-axis load.

Figure 8.

Resultant force (X–Y plane)–displacement (Z-axis) curve.

Table 2 shows the results of compression testing by the 3D-TSMS. The measurement results are the mean value of five samples. Standard deviations are reported in parentheses.

Table 2.

Result of the compression test

	LC (-)	WC (gf • cm/cm²)	RC (%)	T (mm)	L (mm)	α _max (gf/cm²)	ΔE _PR (gf • cm/cm²)	CMD (-)
Fine	0.79 (0.02)	24.8 (0.26)	54.5 (0.64)	37.2 (0.31)	21.0 (0.45)	0.68 (0.01)	8.1 (0.43)	0.33 (0.03)
Thin	0.60 (0.01)	19.8 (0.50)	43.6 (0.67)	33.8 (0.58)	21.7 (0.63)	0.86 (0.01)	4.5 (0.45)	0.21 (0.02)
Medium	0.90 (0.02)	26.2 (0.56)	50.8 (2.10)	39.6 (0.10)	19.5 (0.08)	0.54 (0.01)	12.5 (0.81)	0.05 (0.02)
Dense	0.97 (0.03)	25.4 (0.83)	47.8 (1.29)	39.9 (0.26)	17.3 (0.13)	0.48 (0.01)	13.2 (0.77)	0.19 (0.05)
Thick	0.73 (0.08)	26.3 (0.38)	57.0 (2.57)	40.1 (0.14)	24.0 (2.53)	0.61 (0.06)	8.6 (2.09)	0.27 (0.08)

Discussion

Correlation analysis and multi-regression analysis were performed to determine the compression property parameters which had an influence on the perceived elasticity of sheepskin. Excel statistics 2006 for Windows (SSRI Co., Ltd, Japan) was used as the statistical package.

Correlation analysis between sensory test and compression properties

Table 3 shows the simple correlation coefficients between the sensory test results and the compression property parameters measured by the 3D-TSMS. Perceived elasticity showed a strong positive correlation with LC, WC, T, and ΔE _PR, and a strong negative correlation with L and α _max. However, in the results of a test for no correlation, which was carried out due to the small number of samples, only LC, α _max, and ΔE _PR showed a statistically significant difference.

Table 3.

Simple correlation coefficient between perceived elasticity and the compression properties

Characteristic value	Simple correlation coefficient for perceived elasticity	p
Compression linearity, LC	0.97	< 0.01
Compression energy, WC	0.76
Compression resilience, RC	0.28
Thickness at pressure 1 gf/cm², T	0.79
Compression distance, L	−0.71
Flexibility coefficient, α _max	−0.93	< 0.05
Pressure relieving energy, ΔE _PR	0.93	< 0.05
Continuous mean deviation, CMD	−0.17

LC was strongly correlated with perceived elasticity and had a simple correlation coefficient of 0.97 (Figure 9). A higher LC value indicates the linearity of the compression curve due to compression and translates to greater perceived elasticity in the mouton when touched by a human hand as reflected in the 3D-TSMS measurement results. This is usually an indication of compressive hardness. It shows that LC corresponds to perceived elasticity because the convex load change that appears during the initial compression process contributes to the linearity of the compression curve similar to the dense mouton sample. Therefore, we have determined that one of the notable features of compression behavior for moutons which are perceived elastic is a convex shape curve during initial compression.

Figure 9.

Relationships between perceived elasticity and compression linearity (LC).

α _max was employed to express the ease of compression of a viscoelastic material. This is equal to the ratio of the compression strain divided by the pressure. α _max shows a high value for a large compression distance, as is seen at low pressure. The thin mouton samples have the highest α _max value, and this parameter corresponds to the tactile sensation when it comes into contact with a mouton sample with poor perceived elasticity. α _max was negatively correlated with perceived elasticity and had a simple correlation coefficient of -0.93. Therefore, it is believed that α _max is a negative mechanical parameter indicating the perceived elasticity of mouton.

ΔE _PR is the area of the hysteresis loop on the pressure-displacement curve which shows the amount of energy lost during a compression and recovery cycle. So, ΔE _PR was believed to be a negative indicator of the mouton recovery properties. Also, compressional resilience, RC, is known to be a parameter which represents the recovery properties of fabric. However, there was weak correlation between RC and perceived elasticity, while ΔE _PR was strongly correlated with perceived elasticity and displayed a simple correlation coefficient of 0.93 (Figure 10). In the case of sheepskins, unlike the viscoelastic properties of general fabrics, it is considered that humans perceive elasticity when the gap between the compression energy and recovery energy is wide.

Figure 10.

Relationships between perceived elasticity and pressure relieving energy (ΔE _PR).

From the results of the above-mentioned correlation analysis, it is believed that humans perceive elasticity when the mouton is difficult to compress and has a low elastic restoration force. In other words, if energy applied to the mouton is absorbed by the deformation, friction, or interference of the mouton hair and elastic energy is reduced, the elasticity of the mouton is evaluated as being high.

Multiple regression analysis between sensory test and compression properties

Multiple regression analysis was carried out to determine which variables influence perceived elasticity. The dependent variable was the perceived elasticity and independent variables were the compression parameters provided by compression testing. A forward selection method was used to select variables, since the number of samples was lower than the number of variables. For the variable selection F-value, both F-in and F-out were 2.0, and the confidence interval was 95%. Table 4 shows the results of the multiple regression analysis.

Table 4.

Result of multiple regression analysis (independent variable: all compression parameters; variable selection: forward selection method; F-in: 2.0, F-out: 2.0, confidence interval: 95%)

Predictor	PRC	SPRC	F	T	P	SE	PCC	SCC
LC	5.34	0.98	1771.7	42.1	0.02	0.13	1.00	0.97
CMD	1.83	0.24	869.1	29.5	0.02	0.06	1.00	−0.17
α _max	−0.49	−0.09	16.2	−4.0	0.16	0.12	−0.97	−0.93
Constant	−4.76		722.5	−26.9	0.02	0.18

PRC: partial regression coefficient, SPRC: standard partial regression coefficient, SE: standard error, PCC: partial correlation coefficient, SCC: simple correlation coefficient.

The variables LC, CMD, and α _max all turned out to influence the perceived elasticity. CMD showed a higher standard partial regression coefficient than α _max, even though it was not strongly correlated with the perceived elasticity. Also, as a result of the partial regression coefficient test, CMD was shown to be a significant variable at the 0.05 level. However, α _max made no significant difference on the dependent variable (p > 0.05), even though it was negatively correlated with the perceived elasticity, and there was a multicollinearity between LC and α _max because both variables correlated strongly with each other (R: −0.94) and also α _max was inversely similar to the mechanical parameter LC. So, it was determined that α _max was inappropriate as a variable to explain the perceived elasticity.

Conclusions

In this study, we developed a new hand evaluation device, namely, the three-dimension tactile sensation measurement system (3D-TSMS), which consists of a Cartesian manipulator and a three-axis force sensor. From the vertical pressure-displacement curve, seven parameters were calculated to indicate the compression properties of sheepskin as follows: compression linearity, compressional energy, compressional resilience, thickness, compression distance, compressional flexibility coefficient, and pressure relieving energy. The continuous mean deviation was calculated from the load-displacement curve in the resultant X–Y plane force and the Z-axis displacement. The sensory test results of the perceived elasticity of five kinds of sheepskin were statistically compared with eight compression property parameters measured by the 3D-TSMS.

In statistical analysis results, compression linearity turned out to be the most influential mechanical parameter to perceived elasticity, as highly elastic sheepskin displayed a convex shape curve during the initial compression process which contributed to the linearity of the compression curve. The load applied vertically on the mouton hair was transferred horizontally. During this process, the applied energy was absorbed through interaction with the mouton hairs via bending deformation, friction, and interference. These behaviors disturbed the recovery of hairs. This phenomenon was reflected in continuous mean deviation and pressure relieving energy. Therefore, the perceived elasticity of sheepskin could be evaluated by measuring compression linearity, continuous mean deviation, and pressure relieving energy.

Furthermore, it was verified that the 3D-TSMS was a useful instrument as a hand evaluation device for the compression properties of thick-piled fabrics that accurately reflects the sensorial properties of the human hand.

In the future, further research into the relationship between compression parameters and cross-cultural sensory test results with more varied samples will be needed to estimate the perceived elasticity of sheepskin, since the sensory test was carried out only with Japanese participants in their mother tongue and limited to only five kinds of sample.

Footnotes

Funding

This work was supported by Grant-in-Aid for Scientific Research (B) (no. 22300075) and Global COE Program of the Ministry of Education, Culture, Sports, Science, and Technology of Japan.

Conflict of interest statement

The authors declare there is no conflict of interest.

Acknowledgements

We would like to give special thanks to Mr Jun Saito, who worked with us on 3D-TSMS, and Mr Atsushi Sato, who worked with us on the sensory test.

References

Jolley

Wright

McGowan

Hickey

Campbell

Sinclair

. Preventing pressure ulcers with the Australian Medical sheepskin: an open-label randomized controlled trial. Med J Aus 2004; 180: 324–327.

Garrow

Wooller

. The use of sheepskin covers on vehicle seats. Text Res J 1970; 13: 255–263.

Miwa

Ishida

. An analysis of ‘SOFTNESS’ of fabrics. J Text Mach Soc Japan 1978; 31: 403–410.

Stearn

D’Arcy

Postle

Mahar

. A statistical analysis of subjective and objective methods of evaluation fabric handle. J Text Mach Soc Japan 1987; 40(8): T75–T80.

Harda

Saitoh

. Handle information and planning system. SEN’I GAKKAISHI 1990; 46: 259–264.

Matsudaira

Kiuchi

Kubo

. Objective evaluation of hand of futon cloth. SEN’I GAKKAISHI 1993; 46: T207–T214.

Lee

Kamijo

Honywood

Nishimatsu

Shimizu

. Analysis of finger motion in evaluating the hand of a cloth using a glove-type measurement system. Text Res J 2007; 77: 13–19.

Kawabata

. The standardization and analysis of hand evaluation, 2nd ed. Osaka, Japan: The Textile Machinery Society of Japan, 1980.