International Journal of the Physical Sciences Vol. 7(30), pp. 5167-5174, 9 August, 2012 Available online at http://www.academicjournals.

org/IJPS DOI: 10.5897/IJPS12.256 ISSN 1992 - 1950 2012 Academic Journals

Full Length Research Paper

The study of coefficient of friction for light motorcycle sliding on asphalt road
Chih-Yung Lin1, Chun-Chsia Hsu2 and Chin-Ping Fung3*
Department of Mechanical Engineering, Lunghwa University of Science and Technology, Guai-Shan, 33306, R.O.C. Taiwan. 2 Department of Multi-media and Game Science, Lunghwa University of Science and Technology, Guai-Shan, 33306, R.O.C. Taiwan. 3 Department of Mechanical Engineering, Oriental Institute of Technology, Pan-Chiao, 22061, R.O.C. Taiwan.
Accepted 22 May, 2012
1

The very high number of motorcycles in Taiwan contributes to its high accident rate. Handling a twowheeled motorcycle during an accident is difficult, frequently resulting in falls and sliding along the road surface. Speed is often used to identify the party at fault. The friction coefficient and scrape mark length are typically used to estimate vehicle speed. Most literature focused on heavy motorcycles, and whether the friction coefficient is independent of speed. To resolve disputes about whether the friction coefficient is dependent upon speed, an accurate measurement methodology was applied to determine initial speed of a motorcycle sliding along a road surface. Experiments using 125cc scooters, that is, light motorcycles, were undertaken on dry and wet asphalt roads. Slide distance was measured to examine the relationship between a motorcycles scrape mark length and speed during slide tests. Experimental analysis showed that the square of speed is proportional to scrape mark length, and the linear relationship derived from the work-energy theorem indicates that the friction coefficient is independent of speed. Key words: Coefficient of friction, light motorcycle speed, scrape mark length.

INTRODUCTION The number of motorcycles in Taiwan has increased on average 2% annually. Statistical data published by Taiwans Ministry of Transportation and Communications (2010) show that Taiwanese residents own 14,697,277 motorcycles. This number is approximately two-third of Taiwans total population. Compared with the US and Japan, 2007 data show that the number of motorcycles in Taiwan (13,943,000) was slightly higher than that in Japan (12,938,000), and markedly higher than that in the US (7,138,000). The number of motorcycles per square kilometer in Taiwan (385.3) was significantly higher than that of Japan (35.5) and the US (0.8). The number of casualties caused by motorcycle accidents in Taiwan is typically high, as the number and density of motorcycles on roads are large. The A1 type death toll, that is, the number of people who died immediately in a motorcycle accident or within 24 h, provided by the National Police Agency (2010) has been as high as 40% in recent years. As motorcycle speed cannot be determined without the coefficient of friction for a motorcycle sliding on a road, assigning blame is difficult. When motorcyclists encounter danger, they first apply their brakes to reduce speed. Therefore, efficiency in applying brakes is the most important factor in motorcycle safety. The Nevada Automotive Test Center (2001) performed a series of motorcycle braking performance tests under six conditions: maximum effort braking on dry asphalt; maximum effort braking on wet asphalt; maximum effort braking on dry and on wet concrete

*Corresponding author. E-mail: cpfung@mail.oit.edu.tw. Tel: +886-2-77380145. Fax: +886-2-77386648.

5168

Int. J. Phys. Sci.

Belgian Block surface; braking in a corner; and maximum effort braking with wetted brakes. This report provides very detailed test data. A joint research program (Biokinetics and Associates Ltd., 2003) by the US Department of Transportation, National Highway Traffic Safety Administration (NHTSA), and the Road Safety and Motor Vehicle Regulation Directorate, Transport Canada (TC), compared three regulations for motorcycle braking systems to assess test severity. Based on test results, the Federal Motor Vehicle Safety Standard (FMVSS) No. 122 for the dry braking test, and the Economic Commission for Europe (ECE) Regulation No. 78 for the fade and recovery test and the wet braking test were the most difficult protocols to meet. Although these studies assessed motorcycle braking performance, due to the characteristics of driving twowheeled motorized vehicles, motorcyclists may lose their balance and crash. After falling a motorcycles kinetic energy dissipates gradually due to friction between the motorcycle and road surface, leaving scrape marks on the road. Slide distance is typically used to estimate a motorcycles speed. The simple work-energy principle is used to determine initial speed V as follows:

f mg S mV 2 / 2
where S is the length of the scrape mark and f is the sliding coefficient, which is assumed constant. The sliding coefficient, according to some studies (Becke, 1985; Becke and Golder, 1986), is dependent on speed. However, Day and Smith (1984), Lambourn (1991) and Medwell et al. (1997) obtained different findings. Sliding coefficients and motorcycle speed at the start of a slide are the primary goals of motorcycle accident reconstruction. To study scrape mark length and the friction factor, Day and Smith (1984) conducted a series of slide tests, in which motorcycles were dragged over moderately worn asphalt and on loose surfaces at a very low speed and at 40 km/h. They demonstrated that friction factors differed significantly on different surfaces. The friction factors were in the range of 0.45 to 0.58 on asphalt and in the range of 0.68 to 0.79 on gravel. However, they did not generate conclusions for the relationship between friction factor and speed. Lambourn (1991) performed tests to investigate how average deceleration by a motorcycle sliding to rest from a high speed compares with friction measured by a low-speed drag test. Friction measured in low-speed drag tests was close to that for high-speed sliding. When a motorcycle was dropped onto a road surface in trailer tests, the coefficient of friction had little or no dependence on speed; however, when a motorcycle failed on the road from an upright position, the coefficient of friction may have some dependence on speed. Medwell et al. (1997) conducted testing to determine the sliding friction factor

on asphalt pavement for a sport motorcycle equipped with a full coverage fairing. Based on test results, the friction factor for fully-faired motorcycles on asphalt was in the range of 0.28 to 0.38 when the motorcycle, initially upright in the bed of a pickup truck, was released sideways onto the road surface at test speed. As only four tests were conducted, data were insufficient to determine whether the friction factor depends on motorcycle speed. McNally and Bartlett (2007) performed three motorcycle drop tests to simulate a motorcycle falling on pavement under real-world conditions. These tests were performed on a gridded test surface pre-lined with colored tape, and one camera was placed in an elevated lift for video analysis of the motorcycle s trajectory. An average coefficient of friction of 0.41 to 0.48 was obtained from video data, indicating that the motorcycle tends to decelerate at high rates during its initial impact with pavement; however, data were limited by the small number of tests performed. Lynch (1984) conducted slide tests using two Honda 350cc motorcycles. These motorcycles were either dropped sideways from a towed trailer or pushed on their side, causing the motorcycle to hit the pavement when the motorcycle was rolled up to speed. The friction coefficients were from 0.28 to 0.55. A brief review of literature for the motorcycle slide-tostop tests reveals that data were primarily obtained for heavy motorcycles. Studies of light motorcycles, with engines 125 cc, however, are rare. Previous study results also have limited application, as the number of tests was insufficient to generate confidence when determining the speed of a motorcycle at the start of an accident sequence, and when deriving conclusions via statistical analysis. Most studies merely presented experimental results, and did not investigate relationships between initial motorcycle speed, scrape mark length, and the sliding friction factor. In addition, a motorcycles speed prior to falling over was not calculated accurately, resulting in inaccurate conclusions. Therefore, to determine a light motorcycles speed prior to falling over in slide-to-stop tests on dry asphalt road during sunny days and wet asphalt road during rainy days, this work applies a novel and accurate measurement methodology. For measuring a motorcycles speed and scrape mark length accurately, the coefficient of friction in each test was determined using the work-energy principle. Experimental data were analyzed using a statistical method to confirm that the coefficient of friction is independent of motorcycle speed.
EXPERIMENTAL WORK Most previous studies used heavy motorcycles. Data for accident reconstruction for light motorcycles are lacking. Thus this work investigates the sliding coefficient of light motorcycles on dry and wet asphalt roads. Twenty-five 125 cc scooters, that is light

Lin et al.

5169

Figure 1. The mechanism designed for towed upright motorcycle.

Figure 2. The diagrammatic sketch for the relative positions of camera, test field and trucks driving direction.

motorcycles, were used in this work. All motorcycles were used and in good mechanical condition. Before all tests, oil tanks were emptied. Field tests were conducted during winter under temperatures of 10 to 20C. The test procedure was designed to simulate motorcycle driving conditions during a traffic accident. During the test, a motorcycle was released from a pickup truck in an upright position and made to fell onto its side before it started sliding. A bracket fitted to the rear of the small pickup truck held a motorcycles front wheel, while its rear wheel rolled freely on the road (Figure 1). Two ropes attached to the motorcycles handle bar were held by an assistant sitting in the rear of the truck to keep the

motorcycle upright. When the desired speed was reached, the ropes were released, and the motorcycle fell onto its side on the road. A scrape mark was left on the road, spanning the distance from where the motorcycle first touched the road surface to where it stopped sliding. The length of the scrape mark was then measured using a tape measure. Video analysis was employed to determine accurately the initial speed of the motorcycle sliding freely along a road surface. A highspeed video camera, the Casio EX-F1, was set up on the sidewalk facing the test field. Tests of the light motorcycles sliding were performed on dry and wet asphalt roads. Figure 2 shows a diagrammatic sketch for the positions of the camera, the test field,

5170

Int. J. Phys. Sci.

Figure 3. The image of virtual lines on the test field.

and the direction in which the truck was travelling. The speed of the pickup truck towing a test motorcycle varied. The driver tried to keep truck speed constant at intervals of 5 km/h in the range of 25 to 50 km/h. When the truck entered the video cameras range, the assistant released the ropes holding the motorcycle. The motorcycle then fell onto its side on the road. The speed of the motorcycle when it touched down and was sliding freely on a road surface was determined for further analysis. The high-speed video camera obtained six megapixel stills at up to 60 frames/sec. It also started recording images in a memory buffer that it continually cleared. When the shutter was fully pressed, the buffer content was saved along with all subsequent images. Therefore, important images of a motorcycle sliding on the road were retained, even when the assistant did not press the shutter at the exact moment of contact. The initial speed when the motorcycle first started sliding along the road surface was determined using virtual lines on the road surface (Figure 3), which indicated the position of a motorcycle sliding on the road at different times. Two fire hoses were placed on the two sides of the test field to make a runway for the test truck. The two fire hoses were marked with colored tape at 10 cm intervals. Then, the image of the test field was post-processed using PhotoImpact, a digital image program. Virtual lines were drawn on each image to link the corresponding marks on the two fire hoses. The image of virtual lines on the test field was then superimposed upon images obtained by the high-speed camera. Figure 4 shows continuous images of a motorcycle sliding on the road with virtual lines. These continuous images were shot at 60 frames/s; thus, a time lap of 1/60 s existed between any two images. The point at which the motorcycle first touched on the road surface was the reference point, and the distance a motorcycle slid along the road was determined easily by comparing reference points in two continuous images with virtual lines. By dividing the distance by the time gap, initial speed of a motorcycle sliding on a road surface was obtained.

RESULTS AND DISCUSSION Coefficient of friction on dry and wet asphalt roads Eighteen experiments were conducted on dry asphalt at velocities in the range of 25 to 50 km/h. The motorcycles initial kinetic energy when it fell was gradually consumed by friction force between the motorcycles body and the road surface. The motorcycle stopped sliding after its kinetic energy was totally exhausted. During the period of a motorcycle sliding on a road surface, frictional force, R, is the product of the friction coefficient, f , and the motorcycles gravity, mg , that is, R f mg . The work, W, by friction force is the product of friction force, R, and scrape mark length, S, that is W R S or W f mg S . A motorcycle is at rest when W equals the motorcycles initial kinetic energy. The work-energy theorem gives the
2

equation

W mV 2 / 2

or f mg S mV / 2 . Following the work-energy theorem, the friction coefficient is:

f V 2 /( 2 g S)

(1)

By substituting the values of the motorcycles initial sliding speed and scrape mark length in each test into Equation (1), the friction coefficient, f, was determined. As mentioned, 18 experiments were performed on dry asphalt road. The friction coefficients were in the range of

Lin et al.

5171

(a)

(b)

(c)

(d)

(e)

(f)

Figure 4. The continuous images of the motorcycle sliding on the road with virtual lines.

Motorcycle speed (km/h)

Figure 5. The friction coefficients at various initial sliding speeds for light motorcycles on dry asphalt road.

0.36 to 0.53 for motorcycle sliding speeds of 25 to 50 km/h. Figure 5 shows the friction coefficients at various

initial sliding speeds. The friction coefficient data did not change much as motorcycle sliding speed changed, and

5172

Int. J. Phys. Sci.

Motorcycle speed (km/h)

Figure 6. The friction coefficients at various initial sliding speeds for light motorcycles on dry asphalt road.

data were scattered within a narrow range. In addition, the average friction coefficient in the 18 tests was 0.428, and its standard deviation was 0.046. Statistical data also indicate that the scatter range of the friction coefficient was small. To determine the friction coefficient on wet asphalt road, seven experiments were undertaken on rainy days. The experimental process for the wet asphalt was the same as that for dry asphalt. Experimental results show that the friction coefficients were in the range of 0.29 to 0.40. The average friction coefficient on wet asphalt was 0.340, lower than that for dry asphalt. The difference between these friction coefficients is reasonable, as water lubricates surfaces between a motorcycle and the road. In addition, the values of the friction coefficient on dry and wet asphalt were consistent with data ranges in previous studies (Day and Smith, 1984; Lambourn, 1991; Lynch, 1984; McNally and Bartlett, 2007; Medwell et al., 1997). Figure 6 shows a scatter plot of friction coefficients on wet asphalt road. The scatter range of friction coefficients is small. Statistical analysis generated a standard deviation of 0.037, indicating that the friction coefficient did not change much with the motorcycle sliding speed.

a motorcycle sliding to rest, the friction coefficient is considered constant because mechanical work done by friction force consumes kinetic energy. Moreover, the friction coefficient should be constant when the condition of the two solid surfaces in relative motion does not change. However, whether the friction coefficient is independent of speed must still be confirmed, even although previous studies (Day and Smith, 1984; Lambourn, 1991; Medwell et al., 1997) underscored the importance of the friction coefficient for a sliding motorcycle. Based on experimental results, this study investigates by statistical analysis whether the friction coefficient with is independent of sliding speed. Equation (1), derived from the work-energy theorem, yields the friction coefficient. From Equation (1), the friction coefficient is constant when a linear relationship exists between scrape mark length, S, and the speed variable, V / 2 . To elucidate the relationship between scrape mark length, S, and the speed variable, V / 2 , regression analysis was employed to determine how the typical value of scrape mark length changed when initial sliding-speed varied. Statistical analysis indicates that the relationship between scrape mark length, S (y), and the speed variable, V (x), can be expressed by the following linear regression model:
2 2 2

Speed-independent coefficient of friction From the work-energy theorem, a fast motorcycle sliding speed produces a large amount of kinetic energy ( mV 2 / 2 ), such that a large mechanical work done by friction force ( f mg S ) is expected. During the period of

y 0.0047x 2.1559

(2)

Equation (2) elucidates the linear relationship between scrape mark length, S (y), and the speed variable,

V 2 (x),

Lin et al.

5173

Square of velocity (km/h)2

Figure 7. The scatter plot of scrape mark length S and the speed variable

V2

on dry asphalt road.

Square of velocity (km/h)

Figure 8. The scatter plot of scrape mark length S and the speed variable

V2

on wet asphalt road.

(scatter plot as shown in Figure 7). The coefficient of 2 determination, R , was 0.8469; therefore, the correlation coefficient, R, was 0.920. A correlation coefficient of 0.8 is generally described as strong, whereas a correlation coefficient of 0.5 is generally characterized as weak. The correlation coefficient was 0.8 in this study, indicating that a strong positive linear correlation exists between scrape mark length, S, and speed variable,

friction coefficient is independent of speed. To investigate the linear relationship between scrape mark length, S, and the speed variable, V 2 / 2 , on wet asphalt, regression analysis was applied again. Statistical analysis indicates that the relationship between scrape mark length, S (y), and the speed variable, V (x), on wet asphalt can also be expressed by a linear regression model, y 0.0088x 3.0905 (Figure 8). The coefficient 2 of determination, R was 0.8253; therefore, the
2

V 2 . This linear correlation indicates that the

5174

Int. J. Phys. Sci.

correlation coefficient, R was 0.908, indicating that scrape mark length, S, and the speed variable, V , have a strong positive linear correlation on wet asphalt. Statistical analysis of experimental results for motorcycles sliding on wet asphalt indicates that the friction coefficient on wet asphalt is also independent of sliding speed.
2

REFERENCES Becke M (1985). Zweiradrutschverzgerungen bei hohen Geschwindigkeiten (Deceleration at high speed of two-wheelers). Verkehrsunfall und Fahrzeugtechnik 2:37-40. Becke M, Golder U (1986). Rutschversuche mit Zweirdern auf nasser Fahrbahn und Gras (Skit tests with two-wheelers on wet roads and grass). Verkehrsunfall und Fahrzeugtechnik 4:91-95. Biokinetics and Associates Ltd (2003). Review of Motorcycle Brake Standards. Transport Canada and National Highway Traffic Safety Administration, Report No. R03-07. Day TD, Smith JR (1984). Friction Factors for Motorcycles Sliding on Various Surfaces. SAE Paper No. 840250. Lambourn RF (1991). The Calculation of Motorcycle Speeds from Sliding Distances. SAE Paper No. 910125. Lynch GF (1984). Conduction Test Slides: Motorcycles on Asphalt. Law and Order November 1984 pp. 11-12. McNally BF Bartlett W (2007). Motorcycle Sliding Coefficient of Friction Tests. Accident Reconstruction J., 17(2): 47-49. Medwell CJ, McCarthy JR, Shanahan MT (1997). Motorcycle Slide to Stop Tests. SAE Paper No. 970963. Ministry of Transportation and Communications ROC (2010). Number of Registered Motor Vehicles [online]. Available from: http://www.motc.gov.tw/mocwebGIP/wSite/lp?ctNode=550&CtUnit=9 4&BaseDSD=16&mp=1 [Accessed 26 August 2011]. National Police Agency ROC (2010). Traffic accident (A1 type) [online]. Available from: http://www.npa.gov.tw/NPAGip/wSite/lp?ctNode=12595&nowPage= 1&pagesize=15&mp=1[Accessed 26 August 2011]. Nevada Automotive Test Center (2001). Evaluation of Motorcycle Braking System Performance and Characteristics Relative to Current Regulatory Standards. National Highway Traffic Safety Administration, Report No: 122-NATC-00-001.

Conclusions To resolve the dispute about whether the friction coefficient is dependent upon speed, a novel and accurate measurement methodology was applied to determine initial speed when a motorcycle started sliding freely on a road surface. Experiments using the 125cc light motorcycles were undertaken on dry and wet asphalt roads. Experimental results are summarized as follows. 1. The friction coefficients on dry and wet asphalt roads were in the range of 0.36 to 0.53, and 0.29 to 0.40, respectively. The average friction coefficient on dry and wet asphalt roads was 0.428 and 0.340, respectively. The average friction coefficient on dry asphalt road was larger than that on wet asphalt road. 2. The relationship between scrape mark length, S, and the speed variable, V , on both dry and wet asphalt roads can be expressed by a linear regression model. The linear relationship indicates that the friction coefficient is independent of sliding speed.
2

ACKNOWLEDGEMENT The authors would like to acknowledge the assistance of Chun-Hung Lin for his help with experiments.