Investigating causal factors of road damage: a case study

High rainfall leads to flood vulnerability and potentially leads to the failure of road pavement which affects the economic loss for the people. However, such road pavement failures need to be studied more deeply to be able to distinguish whether the failure is solely due to flood natural disaster or other factors, such as overloading and poor construction quality. This research is urgently needed because up to now there is no tool for assessing pavement failure as a tool of decision-making system that is able to distinguish road pavement failure causal factors into category of natural disaster or human negligence factors in accordance with the definition of building failure in Indonesia Law no. 2 year 2017. Assessment of the causal factor of failure is very important because it is closely related to the determination of the parties who must be legally responsible. This research intends to develop methods for assessing the damage and failure of pavement as a tool of rapid and accurate decision-making system to determine the factors of pavement failure. To obtain accurate results, a case study was taken on Kendal Timur road to enable a deep evaluation of the factors that causing failure on the road. The results showed that many factors involved in causing the failure of the road, such as un-functional drainage channel, uncontrolled overload due to the closing of weigh stations around the road, and also low quality of road maintenance work so that repeating similar road damages occurred in the not-too-distant future. Based on functional and structural failure analysis, as much as 71%, 10% and 19% of the road were in failed, critical and not-failed conditions, respectively.


Introduction
A pavement structure is generally designed to meet its service life, by taking into consideration the quality of materials used, environmental factors such as the amount of rainfall that will occur and the ability of the pavement (in this case the drainage channel) to reduce/eliminate the threat of environmental factors, and the magnitude of the predicted load will occur. However, at the time the pavement performs its function in serving the traffic, there is damage to the pavement structure which is often beyond prediction, especially in terms of the magnitude of environmental and load factors. This pavement damage could lead to the failure of road pavement which affects the economic loss for the people.
The process of identifying the causes of pavement failure and the severity level of damage occurred is usually carried out by the agencies and then several maintenance/rehabilitation plans are proposed. However, the proposed plans are generally only normative so it often does not touch the root cause of the pavement damage.
Therefore, in this study, the deterioration or even failure of pavement structures occurred needs to be further investigated to be able to distinguish whether the failure is caused only by natural disaster (such as floods) or due to other causes, such as overloading and lack of construction quality. Based on this, this research becomes very important to do because at present there is no tool that can be used to evaluate the failure of pavement and decide whether the failure of pavement structure is caused by natural disaster or human negligence factors as in accordance with the definition of building failure in Indonesia Law no. 2 year 2017. Assessment of the causal factor of failure is very important because it is closely related to the determination of the parties who must be legally responsible. This research intends to develop methods for assessing the damage and failure of pavement as a tool of rapid and accurate decision-making system to determine the factors of pavement failure.

Causal factors
There are 3 factors that will be reviewed in this research, i.e. excessive load factor, environmental factor (i.e. flood and sufficiency of drainage channel), and road construction quality.

Overload
Excess load or overload has a significant impact on the achievement of the design life of the road. Overload leads to damaged roads prematurely and this will result in economic losses. Traffic load deviations occur when heavy vehicles carry loads beyond the allowable carrying capacity. An overloaded vehicle can cause its axle load in terms of ESAL (equivalent standard axle load) increase rapidly following the equation below. (1) in which: k = a constant that depends on the number of axle (i.e. single, double or triple), standard axle load = 8.16 ton.
The higher the number of repetitions of vehicles with excessive load, the lower the performance of road pavement, as indicated by the reduction of service life, will be.

Environmental factor
Environmental factor, water in this case, contributes significantly to pavement conditions [1]. In pavement structure design, water has been considered in terms of drainage coefficient, which is a function of drainage system quality and saturation condition of pavement structure within a year (see Table 1).  Table 1 indicates that the lower the drainage system quality and the greater the time in a year the pavement in saturation condition, the lower the drainage coefficient will be. The drainage coefficient will be used to estimate the magnitude of the structural number (SN) of the pavement structure as follows: in which: ai = layer coefficient; Di = layer thickness

Quality of road pavement
The quality assessment of the pavement is carried out using a combination of two functional condition parameters of the road, that is, International Roughness Index (IRI) and Surface Distress Index (SDI) [3]. The combination of the two parameters represents four functional conditions of the road (good, fair, slightly damage, and severely damage) as shown in Table 2.

Pavement Serviceability Index (PSI)
In this study, PSI is expressed as a function of pavement condition index (PCI) index and structure condition index (SCI) as [4] as shown in equation (3) - (5).
in which: IRI = international roughness index, PSIt = terminal present serviceability index for design, SNef = effective structural number, dan SNmin = minimum required structural number Simamora dan Hatmoko [5] stated that PSIt values for non-toll national roads are in the range of 2.0 -2.5, while SNeff and SNmin values are determined using equation (2) for existing condition and required structural conditions, respectively.

Research method
This study consists of several steps, as follows. a. Data collection from case study location, i.e. Kendal Timur road. This is a national road that connects two cities in Central Java province, Semarang and Kendal. The data collected includes: (i) road damage data (the types, dimension and severity level of damages; and road roughness); (ii) road material properties; (iii) rainfall and drainage system data; (iv) traffic volume and axle load of various trucks; and (iv) existing road design. Especially for road damage data, the data was collected per 100 m long.

Analysis of overload condition
Overloaded truck data was collected from load measurement at 16 weighbridge station around Kendal Timur road within period 2012 -2015. In 2015, overloaded truck data could only be obtained from 8 weighbridge stations due to the closure of 8 other stations. This overload analysis was conducted based on overloaded truck accumulation trends that grouped the trucks into four overloaded categories, i.e. overloading (OL) less than 5%, 5-15%, 15-25% and greater than 25% of allowable weight amount (JBI) [6] (see Figure 1). In Figure 1, the number of trucks with excess loads (except overloaded trucks >25%) generally has a decreasing trend in the period of 2012 -2015, especially the number of overloaded truck in 2015. However, this is more due to only 8 weighbridge stations that provided a report of overload trucks in that year. On the other hand, the cumulative equivalent standard axis load (CESAL) of trucks through Kendal Timur road increased significantly to 91.3 million ESALs (or 50% higher than that in 2012, i.e. 66.5 million ESALs).
Using only a half (or 8 stations) of weighing bridges in operation and the increased number of truck axle loads, it can be estimated that the number of overloaded trucks will increase sharply, with patterns that resembling previous three-year pattern of trucks with overload less than 5%, overload between 15%-25%, overload between 5%-15% and overload more than 25%.

Analysis of road damage
As described in section 2.3, road damage represents the level of quality of a road that can be expressed in two parameters, namely the surface distress index (SDI) and road roughness (in terms of IRI). Although SDI and IRI both represent the quality of a road, but conceptually, the two parameters state two different things.
SDI is a composite parameter that represents the dimensions and severity of three most dominant types of road damage, namely cracks, potholes and rutting, while IRI expressed the unevenness of the road surface as a result of road damage. The unevenness calculated in IRI can be caused by one or a combination of 19 road damages, as said by Shahin [7]. Because of the difference of measurement method of the two parameters, both IRI and SDI become complementary parameters to be able to know functional condition of a road segment, as depicted in Table 2.
In this study, Kendal Timur road of approximately 2.3 km-long (Km. 25 + 600 to 27 + 823) can be divided into 23 segments with a length per segment equals to 100 m. For each segment, SDI and IRI data were plotted so that the illustration, as shown in Figure 2, could be obtained. Figure 2 represents the variation of SDI and IRI values for each segment of the road within 4 consecutive years (2012 to 2015). In that figure, different dot colour in Figure 2 is referring to the colour of the functional conditions in Table 2. From

Analysis of drainage condition
Rainfall data was collected from Ketapang rain station around Kendal Timur Road, as shown in Table 3. From Table 3 and using a 5-year rainfall period, the rainfall intensity and flood rate could be calculated as follows: I = 67.84 mm, Qt = 6.63 m 3 /s (for 1-hour raining period), I = 42.74 mm, Qt = 4.18 m 3 /s (for 2-hour raining period), respectively.
The drainage channels on both sides of Kendal Timur road have 2 different shapes, trapezium and rectangle, which have a capacity of Q = 5.24 m 3 /s and 0.51 m 3 /s, respectively. If the capacity of these two drainage channels is compared with the flood discharge, either for 1-hour or 2-hour raining time, it can be concluded that in general the existing drainage cannot accommodate rain-water (Q <Qt).
This indicates that there is a tendency that Kendal Timur road is always submerged when it rains. Beside the lack of drainage capacity, the occurrence of puddles was also contributed by the fact that drainage channels were often clogged by soil or waste (see Figure 3).

Analysis of functional and structural failure
This analysis aims to evaluate the condition of the road whether in the category of failed, critical or un-failed. A failure of a road pavement could be based on two conditions: functional and structural. Functional condition of a road can be assessed using PCI parameter, while the structural one can be evaluated using SCI parameter.
The value of PCI could be determined using equation (4) and IRI data from Figure 2. On the other hand, to calculate the value of SCI, two components of SCI, i.e. SN ef dan SN min, need to be determined using equation (2).
The value Di in SNef dan SNmin equations were obtained from design data, in which the thicknesses of D1, D2 dan D3 equals to 5.12 in., 4.72 in. and 9.84 in., respectively. The value ai in SNmin equation was maximum value for each layer, i.e. 0.44, 0.44 and 0.11, respectively, for materials of surface, base and subbase layers (a1, a2 and a3).
The material quality value ai for SNeff equation was obtained by means of visual condition survey per 100 m long of road segment to detect the presence of road damage: alligator cracking, transverse cracking and pumping. The result of the survey showed that alligator cracking with low severity level and area less than 10% dominated the damage of surface layer on several segments, while pumping damage was not found on base and subbase layers. Therefore, the selected ai for all layers were as follows [2]: 0.35 -0.40, 0.25 -0.35 and 0.10 -0.14 for a1, a2 and a3, respectively.
The values of m2 and m3 in equation (2) were determined using the following equation: P = (Tj * Th / 8760) * (1 -C) * 100 (6) in which: P = Percent of time pavement structure is exposed to moisture levels approaching saturation; Tj = average rainfall in a day (hour); Th = average number of rainy days per year (day); C = flow coefficient. According to Ketapang weather station, the average rainfall in a day (Tj) is 3 hours. To determine average number of rainy days per year, a data set of rainy days for period of year 2006 -2014 was collected, it resulted Th equals to 70.89 ~ 71 days. In addition, the flow coefficient (C) equals to 0.7 based on Directorate General of Highway drainage design guide [8].
Using the values of Tj, Th and C above, P can be calculated, that is, equals to 0.73%. The values of m2 and m3 then could be determined using Table 1. The observation of Kendal Timur road during rainy season indicated that water will completely remove after 1 day. According to AASHTO 1993 [2], the drainage quality as described above is in good category. Furthermore, based on P value and drainage quality, it could be determined that the values of m2 and m3 equal to 1.35 [2].
Using the values ai, Di and mi above, SNef, SNmin and SCI (or rasio SNef to SNmin) of Kendal Timur road can be calculated. The next step was to determine PSI as function of PCI and SCI (see equation (3)). Simamora [9] defined the equation (3)  in which: α1 and α2 = weight age factor for road functional and structural conditions, respectively. Both factors are determined using analytical hierarchy process (AHP) method based on experts' judgements in assessing the failure condition of pavement structures under different conditions. According to the experts' judgment, the value of α1 and α2 was set to 0.3 and 0.7. Table 3 shows the values of PCI, SCI and PSI that calculated using equations (4), (5) and (7). Condition status in Table 3 compares the measured PSI and terminal PSI (or PSIt) and produces 3 pavement status, i.e. failed, critical and not-failed for PSI < PSIt, PSI = PSIt and PSI > PSIt, respectively. The value of PSIt equals to 2.5 due to Jalan Kendal Timur is an arterial (non-toll) road [10]. According to Table 3, there are 17 segments (71%), 2 segments (10%) and 4 segments (19%) categorized as segments with failed, critical, and non-failed conditions. The main factor of the low PSI value that makes a segment categorized as failed is high IRI value (as a component of PCI). The high value of IRI can be contributed by road damages such as potholes or ruts, as a result of excessive load and submerged pavement structure. On the other hand, SCI values are generally less varied because of an assumption that the material quality is almost similar for the entire road segment.

Conclusions
This paper presented some causal factors of road damage, and how these factors can be formulated to determine the contribution of each factor to the occurrence of damage of road infrastructure. To obtain a comprehensive research result, Kendal Timur road was selected as the location of case study.
The results of this study show that un-functional drainage channels, uncontrolled overload due to the closing of weighbridge stations around the road, and also low quality of road maintenance work altogether are the factors that cause road damage.
The road roughness parameter (expressed as IRI) in the PCI function contributes significantly to the pavement serviceability index (PSI), while SCI parameters are not sensitive to PSI value. Based on the assessment of both PCI and SCI parameters of the road segment being the case study site, it can be determined that 71%, 10% and 19% of the road were in failed, critical and not-failed conditions, respectively.