Civl50827.qxd Vol. 3, No. 1 (2006) 75-78ring Researchrnal of EngineeThe Jou 1. Introduction The main function of a highway pavement is to trans- port people and goods in a safe, comfortable, and econom- ic manner. This implies that the pavement structure should provide two basic types of services. First, the pavement should provide a functional service by giving the users a safe and comfortable ride for a specific range of speed. Second, the pavement must provide a structural service by supporting traffic loading and withstanding environmental influences. One of the important factors that determine a pavement functional service is the level of skid resistance. Safe driv- ing depends on an adequate surface friction for vehicle maneuvering, turning and braking. Surface friction is gen- erally given by the following equation: F = fW (1) where, F = tractive force (horizontal force applied to the test tire at the tire-pavement contact), (pound-force, lbf); f = friction factor or friction coefficient; and W = dynamic vertical load on test wheel (lbf). ________________________________________ *Corresponding author’s e-mail: amansour@ksu.edu.sa The value of the coefficient f depends upon several factors including tire pressure; tire wear and inflation pressure; vehicle speed; environmental conditions (wet and dry); pavement temperature; aggregate angularity; and asphalt content. The effect of these factors will be dis- cussed later. A standard test procedure would apply the above test under preset values of the above mentioned fac- tors. An example of such standard procedures is that rec- ommended by ASTM E274, from which a Skid Number, SN is calculated as: SN = 100 f = 100F/W (2) where F is obtained in a precisely defined manner (for example, by sliding a locked standard sized tire from a constant speed, typically 40 mph, along an artificially wetted pavement surface). It is important at this point, and for subsequent discussions, to indicate that the higher is the SN value the more skidding resistant is the pavement surface. The importance of pavement friction is basically relat- ed to driver safety during wet weather. To ensure a safe highway travel in a wet weather, a pavement must have sufficient skid resistance to enable drivers to perform driv- ing tasks without the risk of skidding and loss of vehicle control. Pavement surfaces without sufficient skid resist- ance would endanger the safety of users in two ways. Effects of Pavement Skid Resistance on Traffic Accidents Abdullah I. Al-Mansour* Civil Engineering Department, King Saud University, P.O. Box 800, Riyadh, 11421, Saudi Arabia Received 27 August 2005; accepted 27 November 2005 Abstract: The Ministry of Transport (MOT) in the Kingdom of Saudi Arabia had collected a massive amount of friction measurements using a Mu-meter covering most of the major highway network in the kingdom. Traffic accident data of 89 high accident rate locations from four main different highway classes were extracted from the MOT accident records. Pavement skid resistance for the selected locations was determined from the pavement skid resistance records. The objec- tive of this paper is to utilize these data to investigate the effects of pavement skid resistance on traffic accidents. The analy- sis included establishing relationships between skid resistance and accident number, accident significance and accident den- sity. It was determined that a decreasing skid resistance leads to an increase in traffic accidents. A critical value of skid resist- ance was also established based on number, significance and density of accidents. Keywords: Pavement, Skid resistance, Traffic accidents, Density, Significance degree, Critical value ájQhôŸG çOGƒ◊G ≈∏Y ∞°UôdG ¥’õfG áehÉ≤e ÒKÉJ Qƒ°üæŸG ¬∏dG~ÑY@ ::áá°°UUÓÓÿÿGGádÉ°SôdG √òg ±~¡J .ájQhôŸG çOGƒ◊G äÉfÉ«H ™ªLh á«°ù«FôdG ¥ô£dG áµÑ°T ≈∏Y ¥’õf’G áehÉ≤Ÿ ájQhO äÉ°SÉ«b πª©H ájOƒ©°ùdG á«Hô©dG áµ∏ªŸÉH π≤ædG IQGRh Ωƒ≤àe øe ±Éæ°UG á©HQ’ ™bƒe 89 ∫ ájQhôŸG çOGƒ◊ÉH ¥’õf’G áehÉ≤e §HQ á°SGQ~dG â∏ª°T .ájQhôŸG çOGƒ◊G ≈∏Y ¥’õf’G áehÉ≤e ÒKCÉJ åëÑd äÉfÉ«ÑdG √òg øe IOÉØà°S’G ¤G çOGƒ◊G IOÉjR ¤G …ODƒj ¥’õf’G áehÉ≤e πeÉ©e ¢VÉØîfG ¿G á°SGQ~dG âæ«H ~≤d .çOGƒ◊G áaÉãch IQƒ£Nh O~Y øe πch ¥’õf’G áehÉ≤e ÚH äÉbÓY ôjƒ£J á°SGQ~dG â∏ª°T .¥ô£dG .¥ô£dG øe ∞æ°U πµd çOGƒ◊G áaÉãch IQƒ£Nh O~Y ≈∏Y AÉæH ¥’õf’G áehÉ≤Ÿ áLôM º«b á°SGQ~dG äO~M ɪc ájQhôŸG áá««MMÉÉààØØŸŸGG ääGGOOôôØØŸŸGG.áLô◊G º«≤dG ,IQƒ£ÿG áLQO ,çOGƒ◊G áaÉãc ,ájQhôe çOGƒM ,¥’õf’G áehÉ≤e ,∞°UQ : 76 Vol. 3, No. 1 (2006) 75-78ring Researchrnal of EngineeThe Jou First, it increases the stopping distance, which is a direct function of the coefficient of friction, provided by the pavement surface. Secondly, it increases the risk of hydroplaning. When the water film on a pavement is of a certain thickness, vehicles may hydroplane; i.e. the tires may be separated from the pavement by the water wedge formed between the tire surface and the pavement surface. Total hydroplaning occurs when fluid pressure forces (generated as a result of change in momentum of the fluid particles) in the water-wedged region exceed the total downward load on the tire (Agrawal et al. 1977). The direct consequence of these two events is the high proba- bility of the driver being involved in an accident. The ultimate goal of skid resistance measurements is to identify the level of friction provided by the pavement sur- face in the field. This is typically done either directly by measuring the skid level of the pavement surface, or indi- rectly by measuring the surface texture. Furthermore, to assess the required skid resistance levels, laboratory tests should be performed to measure texture and relate those measurements to observed levels of skid-resistance on the road. There are many factors that affect pavement surface friction. The study by Al-Mansour et al. (2002) indicated that traffic level, highway class, pavement age and percent of air voids in the asphalt mix have significant effects on pavement friction. Percent of asphalt content in the mix has a marginal effect. Pavement skid resistance is reduced as a pavement gets older. Skid number increases as the percent of air voids in the mix increases and decreases by increasing asphalt content. Several studies have been conducted over the last three decades with solid conclusions indicating that the rate of accident occurrence during a wet weather is much higher than that during a dry weather. The study by Ryell et al. (1979) reported that the rate of wet accidents (number of wet-weather accidents per 100 Million vehicle-miles) increased from about 27 at an average skid number of about 55, to about 75 at an average skid number of about 25. The results of another study (Zipkes, 1976) reported the significant reduction in the percent of wet-pavement accidents from 78% to 30% due to increasing the pave- ment coefficient of friction from 0.20 to 0.45 by applying surface grooving. On the other hand, other studies ana- lyzed specific sites during wet and dry weather periods and concluded that the risk of being involved in an acci- dent during wet periods is significantly higher than that during dry periods. Salt (1976) concluded that whilst a surface with a friction coefficient of 0.60 and above may by chance be a scene of an accident in which a vehicle skids in wet weather, the risk that it will be the scene of repeated skidding accidents is extremely small. This risk first becomes measurable with a coefficient of 0.55 to 0.60 and increases sharply by more than 20 times as the coeffi- cient falls to values of 0.40 to 0.45 and by about 300 times when the coefficient is 0.30 to 0.35. There are no standards agreed upon which define min- imum acceptable skid resistance levels. Several studies have been conducted to develop criteria for critical skid resistance levels. In general such studies were either based on skid resistance and traffic accident analysis or based on calculating the required stopping sight distance. A study conducted on the streets of Muscat indicated minimum skid number of 0.45 on normal sites (Ali et al. 1999). Another study (Rizenbergs et al. 1977) concluded that pavement with skid number below 26 are very slippery and must be corrected. 2. Objectives The Ministry of Transport (MOT) had collected a mas- sive amount of friction measurements using a Mu-meter covering most of the major highway network in the Kingdom of Saudi Arabia. The main objective of this study is to utilize this data to investigate the effect of skid resistance level on traffic accidents. 3. Data Collection In order to achieve the objective, wet pavement skid resistance data and traffic accidents data for ten main highways with different classes and traffic levels were extracted from the MOT skid resistance and traffic acci- dents files (Al-Tamime, 2002). Four roadway classes were included: dual roads (A) roads that have double lanes in one-way direction; expressway (B) roads that have three lanes or more in one-way direction; expressway with frontage roads (C); and undivided roads (D). The exact locations of the traffic accidents were matched with pave- ment skid resistance measurements. This process resulted in determining the skid number of the pavement for acci- dent location. The traffic accident report does not specify clearly the cause of accident. Therefore, all accidents reported in the year 1999 for the selected locations were incorporated in the analysis, including accidents in both wet and dry weather conditions. In addition to the number of accidents, other data such as types of accidents (death, injury and damage), degree of accident significance, and density of accident were also recorded. Types of accident included accident that result- ed in death, injury or damage. An accident degree of sig- nificance is defined as a recurring accident site factor. Density of accidents is the number of traffic accidents annually per kilometer of road. 4. Analysis and Results 4.1 Basic Statistics A total of ten highways with 340 accidents occurred at 89 locations were used in the analysis. The number of acci- dents and accident locations for each road are shown in Fig. 1. Road Number 15, in the South Western Region of the Kingdom of Saudi Arabia, had 145 accidents which was the highest number of accidents of the roads included in the analysis. Roads Number 500 and 517 had the low- est number of accidents. Similarly, Road Number 15 has the highest number of accident locations and Roads Numbers 500 and 517 had the lowest number of accident 77 Vol. 3, No. 1 (2006) 75-78ring Researchrnal of EngineeThe Jou location. 4.2 Effects of Skid Resistance on Number of Accidents Although accident reports did not specify the main cause of accidents, an attempt was made to relate skid resistance measurements to the total number of accidents for the analysis period. The analysis indicated that the number of accidents increased as skid resistance decreased. A low level of skid resistance resulted in an increase in stopping distance and a loss of vehicle control at a high speed. The relationship between number of acci- dents and pavement friction measured as a skid number is shown in Fig. 2. This data were best fitted in the follow- ing form: NA = a SNb (3) where; NA = number of accidents; SN = skid number; and a, b = regression parameters The values of a and b were estimated to be 2.799 and 0.408, respectively. The coefficient of correlation (R2) of the regression equation was 0.425. This relationship indicated that the number of accidents approach almost a constant number at a skid number greater than 0.45. To further investigate this observation, the number of accidents were plotted as a bar chart against the ranges of skid resistance. The plot is presented in Fig. 3. Pavement with skid resistance less than 0.35 have the highest number of accidents The data also indicate the same observations, that the number of traffic accidents are almost the same for skid number greater than 0.45. These results lead to the conclusion that the critical skid number required at high speed roads is 0.45. A similar analysis to the one just presented was con- ducted for each highway class. As indicated earlier four classes of highway are included. These classes are A, B, C, and D. This analysis was conducted on highway class A, B and D. Highway class C was excluded because of the lack accident data. A summary of the results is presented in Table 1. It is clear that as the highway class gets high- er the critical skid number value increases. These results were expected since the operating speed on a higher high- way class is more than that of a less class highway. 4.3 Effect of Skid Resistance on Accident Density The density of accidents is defined as the number of traffic accidents made annually for each road kilometer. Accident density is calculated as number of accidents divided by the analysis period and multiplied by section length. The relationship between skid number and acci- dent density is presented in Fig. 4. It was found that as skid number decreases, accident density increased. The maximum accident density of 2.66 occurs at a skid num- ber less than 0.35. At a range of 0.75 - 0.80, accident den- sity was found to be 1.76. Accident density seems to be constant at about 1.85 for a skid number greater than 0.45 (critical value). The statistical regression procedure (Neter et al. 1985) was used to model the accident density based on skid resistance number. The best model was found to be in the following form: AD = a SNb (4) The values of the parameters a and b were estimated to be 1.59 and 0.278, respectively. The p-value for the model is 0.0023 and the adjusted coefficient of determination (R2) is 0.36. 0 20 40 60 80 100 120 140 160 R15 R40 R50 R60 R65 R80 R85 R279 R500 R517 Roadway Number N um be r o f A cc id en t a nd A cc id en t L oc at io n Number of Accidents Number of Accident Locations Figure 1. Number of accidents and accident locations 0 1 2 3 4 5 6 7 8 9 10 11 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 Skid Number A cc id en t N um be r/L oc at io n Figure 2. Relationship between skid number and traffic accidents 0 20 40 60 80 100 < 0.35 0.35 - 0.45 0.45 - 0.55 0.55 - 0.65 0.65 - 0.75 0.75 - 0.80 Skid Number N um be r o f A cc id en ts Figure 3. Effect of skid number of traffic accidents Highway Class Critical SN Model Adj R2 A 0.48 NA = 2.79 SN-0.41 0.425 B 0.42 NA = 2.76 SN-0.38 0.265 D 0.38 NA = 2.73 SN-0.36 0.588 Tble 1. Critical skid number values 78 Vol. 3, No. 1 (2006) 75-78ring Researchrnal of EngineeThe Jou 4.4 Effect of Skid Resistance on Accident Degree of Significance The accident degree of significance is defined as recur- ring accident site factor. This factor was determined by multiplying number of accidents by accident seriousness divided by analysis period multiplied by average annual traffic and section length. The seriousness of accident is calculated as: [1 + (number of death accident + number of injury accidents + number of damage accident)/ total num- ber of accidents]. The accident degree of significance is strongly bound- ed with the number of accidents. Whenever skid resist- ance level is low, accident degree of significance increas- es. The effect of skid resistance on accident degree of sig- nificance is shown in Fig. 5. Pavement sections with skid number less than 0.35 has an accident degree of signifi- cance of 2.51. The results also indicated that the accident degree of significance is almost the same for all sections with skid resistance number greater than 0.45. This is another indi- cation that the critical skid resistance number on high speed road can be considered as 0.45. At this critical skid resistance level, the degree of accident significance for all highway classes is 1.7; highway class "A" is 1.7; highway class "B" is 2.1; and for highway class "D" is 2.2. 5. Conclusions Within the scope of this investigation and based on the results obtained, the following may be concluded: 1. The most common method of evaluating pavement skid resistance was the locked-wheel skid test using ASTM E274 test method. 2. There are no standards defined for minimum accept- able skid resistance levels. 3. Road Number 15, in the southwestern region of the Kingdom has the highest number of accidents and accident locations of all the roads included in the analysis. 4. Pavements with a skid resistance number less than 0.35 have the highest number of traffic accidents. The number of traffic accidents seems to be constant for all sections with a skid resistance number equal to or greater than 0.45. 5. The maximum accident density of 2.66 occurred at pavement sections with a skid resistance less than 0.35. Accident density remained constant on all sec- tions with a skid resistance number equal to or greater than 0.45. 6. Whenever skid resistance number is low, accident degree of significance increases. Pavement sections with skid resistance number less than 0.35 have 2.51 accident degree of significance. Accident degree of significance remains constant for all sections with skid resistance number equal to or greater than 0.45. References Agrawal, S.K. and Henry, J.J, 1977, "Technique for Evaluating Hydroplaning Potential of Pavements," TRRL Report LR633. Ali, G., Al-Mahrooq, R. and Tahqa, R., 1999, "Measurement, Analysis, Evaluation, and Restoration of Skid Resistance on Streets of Muscat," J. of The Transportation Research Board, Transportation Research Record No. 1655, Washington, D. C. , USA Al-Mansour, A., Al-Suhaibani, A. and Al-Syyari, S., 2002, "Pavement Skid Resistance Analysis and Evaluation of Highway Network in Saudi Arabia," First Gulf Conference on Roads, Kuwait, March 2002. Al-Tamime, M., 2002, "Pavement Skid Resistance and Traffic Accidents," Senior Project, College of Engineering, King Saud University, Saudi Arabia. Neter, J., Wasserman, W. and Kutner, M., 1985, Applied Linear Statistical Models, Irwin Inc., Homewood, Illinois, USA. Rizenbergs, R. L., Burcheit, J.C. and Warren, L., 1977, "Relation of Accidents and Pavement Friction on Rural Two-Lane Roads," TRRL Report LR633. Ryell, J., Corkill, T. and Musgrove, C., 1979, "Skid Resistance of Bituminous Pavement Test Sections: Toronto By-Pass Project," TRRL Report LR712. Salt, G.F., 1976, "Research on Skid Resistance at The Transport and Road Research Laboratory," TRRL Report LR622. Zipkes, E., 1976, "The Influence of Grooving of Road Pavement on Accident Frequency," TRRL Report LR623. 0 0.5 1 1.5 2 2.5 3 < 0.35 0.35 - 0.45 0.45 - 0.55 0.55 - 0.65 0.65 - 0.75 0.75 - 0.80 Skid Number A cc id en t D en si ty Figure 4. Effect of skid number level on accident density 0 0.5 1 1.5 2 2.5 3 < 0.35 0.35 - 0.45 0.45 - 0.55 0.55 - 0.65 0.65 - 0.75 0.75 - 0.80 Skid Number D eg re e of S gn ifi ca nc e Figure 5. Effect of skid nuber level on accident degree of significance