Investigation of chip seal performance under cold climate conditions

2.1 General characteristics of the test roads

In this study, non-destructive tests were performed at regular intervals on the roads in Erzurum Region of Turkey. Erzurum-Karagöbek test road (A₂) is located on the Erzurum-Tortum State Highway and 500 m in length. The test road is undivided and consists of completely filled cross-sectional profile. Also, the road is double-chip-sealed pavement, and it has base layer of 13.6 cm thickness. Tortum-Uzundere (B₂) test road is located on Tortum-Artvin State Highway and 500 m in length. The test road is also undivided and double-chip-sealed pavement, and the base layer is 12.4 cm. Akşar-Göle test road (C₁) is located on Akşar-Ardahan State Highway and 800 m in length. The test road are undivided and it has a cutting cross-sectional profile also double chip-sealed pavement and base layer with 16 cm thickness. Test section properties and traffic surveying results are given in Tables 1 and 2, respectively.

2.2 Material

A series of laboratory tests were performed in order to determine the aggregate and bitumen’s engineering properties. Accordingly, the aggregate test results and binder characteristics used in test roads are shown in Tables 3 and 4, respectively.

2.3 Method

The study consists of assessments made as a result of in situ non-destructive tests and observations. The non-destructive tests were performed on the right lanes, with five different test zones and three different test points for each zone. These points are right wheel path (1), interval of the wheel path (2) and left wheel path (3) (Figure 1). The non-destructive tests included sand patch [19], British pendulum [20], light weight deflectometer (LWD) [21], and dynamic cone penetration (DCP) [22] tests as well as electromagnetic density measurements [23], temperature measurements with thermal camera and base layer thickness measurements with ground-penetrating radar (GPR) [24]. Additionally, each of the test roads were assessed and scored by specialists according to the seven common deterioration types (potholes, ravelling, loss of texture, flushing, polishing, rutting and settlement), and general deterioration indexes were created in relation to the test roads.

Figure 1:

Schematic plan of nondestructive test points.

3.1 Deterioration indexes of test roads

Deterioration indexes were created as a result of performance monitoring with non-destructive tests for three different test roads. The A₂ and B₂ test roads were observed for 2 years and resealed in this period. The C₁ test road was observed for 1 year. No resealing operations were performed on C₁. While the deterioration indexes were created, common deterioration types were considered. These included potholes, ravelling, loss of texture, flushing, polishing, rutting and settlements. The general deterioration index was formed on the basis of non-destructive test results conducted and visual assessment in the field. Scoring range was selected as 0–100. If the score (deterioration index) decreases, the pavement condition is getting worse. To compute the deterioration index, deterioration tables were prepared for each test section. According to the results of non-destructive tests and field observations, final deterioration types and severities of test sections at the end of 1 year are given in Table 5.

For each test section, seven different deterioration types were observed. Texture loss and polishing were measured with sand circle and British pendulum tests, respectively. Others were scored by the expert according to field observation. If the skid number is <50, polishing deterioration is assessed as poor (P). Similarly, if the surface texture depth is <2 mm, loss of texture deterioration is assessed as poor (P). According to the number of test zones and the number of test points on the zone, the deterioration score varies. Final deterioration index is calculated according to Equations (6)–(8).

where ∑TP=total number of test points, TS=number of test sections, TP=number of test points, DS=number of one test point deterioration score, NDT=number of deterioration types, NP=number of “poor condition” points, NM=number of “medium condition” points and ScDI=score of deterioration index.

Accordingly, if the deterioration index score of test road is between 100 and 60 or between 60 and 50 or between 50 and 0, road condition has been assessed as good, middle and poor, respectively. Variation of deterioration indexes of test roads is shown in Figure 2.

Figure 2:

General deterioration index of test roads.

As a result, the final condition of the A₂ was found to be better than B₂ and C₁. Besides, all of the test roads’ performances significantly decreased at the end of the test period.

3.2 Non-destructive test results

Non-destructive tests were performed on three different points for each zone, for a total of five different sections on three different test roads. The test results showed that deterioration at points 1 and 3 which were most affected by vehicle loads was greater than at point 2. Secondary tests and measurements were carried out after resealing. This is the main cause of an increase in macro texture and decrease in density demonstrated in the second tests.

When A₂, B₂ and C₁ were compared to each other, C₁ had the worst condition with respect to deteriorations at the right wheel paths. B₂ and A₂ followed, respectively. The lowest elasticity modulus values were obtained from the C₁ test road. B₂ and A₂ followed, respectively. According to the LWD test results, the lowest elasticity modulus values were obtained from the C₁ test road. However, the deformation values are high (see Table 6).

Indeed, CBR values were obtained from DCP test as A₂<C₁<B₂. However, CBR values of A₂ and C₁ are so close to each other. Moreover, Alderson [25] reported that a relationship exists between the hardness of the surface of the base layer and the occurrence of flushing. Consequently, flushing is the most common deterioration type found in the C₁ test road (see Figure 3).

Figure 3:

Flushing deterioration at C₁ test road.

The non-destructive test results at point 2 (interval of the wheel path) are better than at points 1 and 3 (see Tables 6–8).However, the embedding of aggregate and flushing are much higher at these points. Although all of the test roads have low total traffic volume, the deterioration rate is not slow. The final test results showed that the final macro texture depth value of the B₂ test road’s point 2 is under 1 mm. According to the TNZ P17 New Zealand performance based specification [26], chip seal should be resealed once the macro texture depth value decreases to under 0.9 mm. Similarly, the values of polishing at point 2 are better than at points 1 and 3; however, the values are close to the appropriate limit of 50. High density values at these points show that flushing started.

Deterioration at point 3 of the A₂ test road, which has height of embankment of more than 2 m, was less than that of other test roads at point 3.

Low CBR values obtained from the DCP test performed at three different zones for each of the test roads showed that the bearing capacity of test roads was inadequate (see Figure 4). Therefore, the lack of bearing capacity of the base layer in each of the three test roads was one of the most important causes of flushing. However, higher temperature values of road surfaces during summer and the use of unmodified bitumen in chip seal construction also caused higher rates of flushing on the roads (see Figure 5). First and second DCP tests were performed in the sixth and eighth months, respectively. The sixth month was thawing and rainy season for Erzurum region. Therefore, the first CBR values of A₂ and B₂ test roads obtained lower compression than the second DCP test results. The bearing capacity of C₁ test road was decreased, and this also led to decrease in CBR value (see Figure 4).

Figure 4:

Variation of base layer’s CBR according to the DCP test results.

Figure 5:

Flushing deteriorations at A₂ and B₂ test roads.

Water and moisture from melting snow caused the weakening of sublayers on the shoulder of chip-sealed surfaces. This resulted in low values of CBR and elasticity modulus, which indicate flushing. Although the A₂ test road’s CBR and elasticity modulus values were low, its deterioration rate was lower than those of the other test roads. The reason might be the embankment height of the A₂ road which was higher than 2 m (see Figure 6).

Figure 6:

Snow shoulder surface at A₂ and C₁ test roads (1, 2, 3); high slope of embankment and permeable zones of shoulder’s surface at A₂ test road (2).

Indeed, Towler and Ball [27] reported that chip seals applied with a binder of at least 1.5 l/m² rate were believed to be completely impermeable in the past, but studies indicated that this was not true. Rain and snow water on chip-sealed road pavements penetrate the chip-sealed pavement by way of tire pressure due to high traffic. After evaporating, this water leads to bitumen rising to the road surface by passing through small holes as bitumen bubbles and consequently causes flushing (see Figure 7).

Figure 7:

Binder bubbles due to the tendency of vaporization of moisture at base layer during summer season.