A Case Study on Distresses of Concrete Pavements Supported on a Retaining Wall

Abstract

Embankments and retaining walls are integral parts of the bridge system and provide a smooth transition from lower elevations (i.e., roadways) to higher elevations (i.e., bridge decks). Performances of pavement structures supported on embankments or retaining walls are directly related to their conditions. This paper presents a comprehensive case study of the evaluation of pavement structures supported on an in-service mechanically stabilized earth (MSE) wall that showed significant distresses, such as lane separation, faulting, lane settlement, and tilting of the MSE wall. The conditions of the pavement structures were evaluated via visual observations, falling weight deflectometer (FWD) tests, and coring through pavement structures. The conditions of the MSE wall were evaluated through dynamic cone penetrometer (DCP) tests, cone penetration tests (CPTs), LiDAR surveys, and soil borings. Detailed analysis of the data obtained in this study provides valuable insights into potential distress mechanisms.

1. Introduction

According to the ASCE 2021 Infrastructure Report Card [1], the evaluation of bridge conditions across the United States has yielded an overall grade of C. This grading underscores the middling state of the nation’s bridges, necessitating targeted efforts to rectify a range of concerns. Notably, a pressing issue arises from the identification of 7.5% of the country’s bridges—equivalent to 46,154 structures—as being subpar or exhibiting structural deficiencies. Within the wide-ranging bridge network of the state of Texas, a substantial portion, namely 18% of Texas’ bridges, has been classified as being in unsatisfactory condition [2].

Embankments and retaining walls are integral parts of the bridge system and provide a smooth transition from roadways to bridge decks [3]. A bridge approach slab is another essential component of the bridge system that acts as an intermediate bridge to span the end portion of the pavement on the roadway embankment/retaining wall directly behind the bridge abutment. In Texas, drilled shafts are commonly used to support bridge abutments, while sleeper slabs are utilized to support the bridge approach slabs on the pavement side [4]. Such a difference in support system typically induces differential settlement between the pavement side and the abutment side, creating an uneven transition onto the bridge.

There are other factors that can cause unsatisfactory performances of pavement structures supported on embankments/retaining walls, including poor conditions of the subgrade soil, inadequate construction practices, compression of the embankment/retaining wall fill materials, and drainage issues [5]. Furthermore, many researchers reported that a considerable number of current bridge approach slabs contain voids underneath them resulting from water infiltration and consequent material erosion [6,7,8].

A pavement structure is a multi-layered system, and the subgrade serves as a foundation of the pavement structure. Evaluations of pavement performance can be broadly divided into two tasks: (a) the assessment of the conditions of pavement structures and (b) investigation of the quality of the subgrade. Assessing conditions of pavement structure often involves the identification of distresses through visual surveys, determination of slab performance through deflection testing [9,10,11,12,13,14,15,16,17,18,19,20], and checking the quality of concrete via coring [21,22,23,24,25]. Smith et al. [26] conducted a comprehensive investigation into the utilization of falling weight deflectometer (FWD) data for mechanistic–empirical pavement design and rehabilitation procedures, establishing the efficacy of FWD tests in pavement evaluation and rehabilitation.

Investigating the quality of the subgrade requires direct access to the subgrade, and the dynamic cone penetrometer (DCP) test has been widely used to determine the California bearing ratio (CBR) or resilient modulus (M_r) of the subgrade due to its rapid setup and simplicity [27,28,29,30,31,32,33,34,35,36]. Results from the DCP tests can be used not only for the quality assessment of newly prepared subgrade but also for in-service roadways when adopting maintenance strategies. For example, if the results of DCP tests performed for in-service roadways indicate the high quality of the subgrade, maintenance measures can include simply treating the surface pavement layer. However, if the DCP investigations show the poor quality of the subgrade, then extensive stabilization measures or the replacement of the subgrade may be needed [37].

Evaluating pavement structures built on roadway embankments/retaining walls poses additional challenges compared to typical pavement structures. Maximum investigation depths of DCP tests are typically limited to 9 to 10 ft (2.7 to 3 m) due to the bending of the driving shaft, and such investigation depths may be sufficiently deep to assess the subgrade quality for typical pavement structures. However, heights of roadway embankments/retaining walls in Texas are typically up to 20 to 30 ft (6 to 9 m), and, therefore, evaluations of backfill materials for the entire embankment/wall height are required to identify the extent of low-quality backfills and determine a proper repair depth.

Another challenge is that the evaluation methodology for roadway embankments/retaining walls should be inclusive of the entire components (i.e., pavement structure, backfill materials, drainage systems, movement of embankments/retaining walls, etc.), not just the pavement structure and subgrade [38]. Aldosari et al. [39] employed mobile Light Detection and Ranging (LiDAR) mapping for the precise monitoring of the performance of a mechanically stabilized earth (MSE) wall and showed the potential advantages of the LiDAR survey for the effective and reliable monitoring of MSE walls, as well as deriving global and local serviceability measures.

Due to the aforementioned challenges, studies of the evaluations of in-service roadways built on retaining walls are exceedingly rare in the literature. The purpose of this case study is to (i) evaluate conditions of concrete pavement supported on an in-service MSE wall, (ii) identify potential mechanisms that might have led to the observed distresses, and (iii) determine repair strategies based on the collected data. To perform comprehensive investigations into the entire components, the authors have employed various evaluation methods. Pavement structures were evaluated through visual observations, FWD tests, and the coring of pavement structures. Additionally, the conditions of the MSE wall were assessed using DCP tests, LiDAR surveys, and soil borings. To overcome the limitation of shallow investigation depths for DCP tests, the authors also performed cone penetration tests (CPTs) and successfully characterized the conditions of backfill soils for the entire wall height. Subsequent sections of this paper provide detailed accounts of observed distresses in the field, followed by a systematic evaluation of pavement structures and MSE walls, along with insights into potential distress mechanisms.

2. Description of the Project Site

2.1. Site Location and Details of Pavement Structure

The project site is located in the southern part of Texas and consists of a six-lane highway (three main lanes on the eastbound and westbound sides, respectively). The subject section is a part of the main lanes on the west side of the bridge on the westbound highway and extends from the start of the bridge approach slab to 600 ft (183 m) west near the exit to a frontage road (Station 1026+92 to Station 1020+92), as shown in Figure 1. The three main lanes on the westbound side are supported on the MSE wall, and those on eastbound side are supported on an embankment with a 4:1 slope (see Figure 2). The subject section showed significant distresses at the time of this study, whereas the pavement structures on the eastbound and east side of the bridge on the westbound highway showed no major distresses.

The subject section was constructed in 1997 and has been in service ever since. The plan set drawings show that the pavement structure consists of 12 in (30.5 cm) of continuously reinforced concrete pavement (CRCP), 1 in (2.5 cm) of asphalt stabilized base (ASB), 6 in (15.2 cm) of Portland cement-treated base (PCTB), and 6 in (15.2 cm) of lime-treated subgrade (LTS). Figure 2 shows the typical section of the pavement structure presented in the plan set drawings.

2.2. Observed Distresses at the Subject Section

The major distresses observed in the subject section were lane separation, faulting, settlement, and the lateral movement of the MSE wall. Figure 3 shows photos of these distresses. The lane separations were prevalent throughout the subject section with a maximum separation width of 5 in (12.7 cm) (Figure 3a). Furthermore, tie-bar failures were often found where the lane separations had occurred (Figure 3b). The elevation of the outside lane was lower than that of the outside shoulder, and the amount of faulting was observed up to 3.5 in (8.9 cm) (Figure 3c). The lateral movement of the MSE wall was also visibly noticeable (Figure 3d) and observed along the section from the end of the bridge slab to approximately 500 ft (152.4 m) west. Figure 4 presents a distress map showing an overview of the locations where these distresses were found.

3. Evaluation of the Performance of the Subject Section

To identify potential causes of the pavement distresses observed in the subject section, various types of field tests were performed under traffic control on 10 and 11 February 2023, hereafter referred to as Day 1 and Day 2, respectively. The slab support condition was assessed via deflection testing using FWD, and coring was conducted to assess the quality of the concrete. Displacements of the pavement grade and MSE wall were evaluated from the LiDAR survey. Conditions of backfills in the MSE wall were assessed via DCP tests and CPTs.

3.1. Evaluation of Concrete Pavement Structure

3.1.1. Slab Deflection

To minimize traffic closure time, the authors selected an FWD test because it is fast and does not involve the removal of pavement material. FWD tests have been widely used for characterizing the mechanical response of the pavement system and assessing the structural capacity [40,41,42,43]. In this study, FWD tests were conducted on the inside shoulder (IS), inside lane (IL), and middle lane (ML) on Day 1 and the outside lane left wheel path (OL LWP), outside lane right wheel path (OL RWP), and outside shoulder (OS) on Day 2. The deflections at 9000 lb (40 kN) loading were evaluated along all of the lanes, covering a total length of 600 ft (183 m) each. Figure 5 shows FWD drop locations performed in the subject section.

Figure 6a shows slab deflections versus DMI (distance measuring instrument) in feet along the subject section obtained via FWD tests. Slab deflections on the inside shoulder and inside lane ranged between 0.8 mils (0.020 mm) and 3.8 mils (0.097 mm), whereas slab deflections on the middle lane ranged between 0.4 mils (0.010 mm) and 24 mils (0.610 mm) (note that although the maximum deflection value of 24 mils (0.610 mm) was measured at about DMI 42 ft (12.8 m), slab deflections on the middle lane overall ranged between 4 mils (0.102 mm) and 9 mils (0.229 mm)). This clearly indicates that the inside shoulder and inside lane have better slab performance than the middle lane. Along the outside lane, slab deflections ranged between about 1 mils (0.025 mm) and 19.3 mils (0.490 mm), showing much greater overall deflections than the inside shoulder and inside lane. Slab deflections on the outside shoulder ranged between 1 mils (0.025 mm) and 26.8 mils (0.681 mm), but the larger deflection values were limited to a zone between DMI 41 ft (12.5 m) and 60 ft (18.3 m). Other than the localized zone, slab deflections on the outside shoulder were less than 3 mils (0.076 mm).

To summarize, the inside shoulder, inside lane, and outside shoulder showed much better slab support conditions than the middle lane and outside lane. Also, higher deflections occurred between DMI 0 and 350 ft (106.7 m), particularly between DMI 30 ft (9.2 m) and 60 ft (18.3 m). Figure 6b, showing a colored profile map of slab deflections, with the green color indicating the slab deflections less than 3 mils (0.076 mm) and the red and yellow colors indicating slab deflections greater than 12 mils (0.305 mm), summarizes the overall slab support conditions in the subject section.

According to Won et al. [44], the statewide average deflection of well-performing 12 in (30.5 cm) CRCP in Texas is about 1.7 mils (0.043 mm). Figure 6 clearly shows that slab deflections at the subject section well exceed this value, indicating poor performance. In particular, very large deflection values near DMI 40 ft (12.2 m) indicate the possible presence of voids beneath the slab.

3.1.2. Tie-Bar Failure Modes

In the design and construction of CRCP, short pieces of reinforcing steel (i.e., tie-bars) are used across the longitudinal construction joints to keep the two adjoining slabs from separating and keep the pavement surface across the joints flat. As previously mentioned, faulting at longitudinal joints and lane separation have been identified as major distresses in the subject section. Lane separations in CRCP have been reported in many projects in Texas, and efforts were made to resolve this issue [45]. One recommendation was the placement of additional tie-bars, particularly when constructing multiple lanes. In 2009, the Texas Department of Transportation (TxDOT) revised its CRCP Design Standards to incorporate this recommendation, decreasing tie-bar spacing from 4 ft (1.2 m) to 2 ft (0.6 m). However, the subject section in this study was constructed in 1997, and, therefore, tie-bars were placed with 4 ft (1.2 m) spacing.

Figure 7 shows two types of tie-bar failure modes observed in the subject section. Figure 7a clearly shows a necking failure and suggests that the tie-bar failed in tension. On the other hand, Figure 7b shows that the tie-bar was sheared off with no sign of necking. These two failure modes indicate that the concrete slab experienced both lateral and vertical movements. These tie-bar failures were observed in most sections where lane separation had occurred.

3.1.3. Concrete Material Quality

Based on the visual survey, no major distresses attributed to concrete material properties were observed. Typically, pavement distresses related to concrete materials are rarely reported in Texas [46,47]. Observed distresses related to concrete materials in the subject section were limited to minor spalling and alkali–silica reaction.

To verify the quality of concrete materials, concrete cores samples were taken in areas where significant slab deflections were observed via FWD tests. A total of nine cores were extracted from the subject section, and none of them exhibited any signs of poor concrete material quality. Out of the nine core samples, one of them (see Figure 8) was extracted from the outside lane near the DMI 40 ft (12.2 m), where exceedingly high slab deflections were observed and the worst slab support conditions were expected. As shown in Figure 8, there was a layer of polyurethane foam beneath the concrete, suggesting that the base materials were washed out, creating a significant void, and the polyurethane foam was used to fill the void in the past. Nevertheless, the concrete quality has remained in excellent condition.

3.2. Light Detection and Ranging (LiDAR) Survey

To obtain quantifiable displacement data, a LiDAR survey was carried out across the entire 600 ft (183 m) section of the pavement surface, as well as the MSE wall. Since there is no survey data generated at the time of construction, the design details from the plan set drawings served as a reference condition in this study. Figure 9 shows the cross-section of the subject section at STA 1025+00 created using the LiDAR measurements (presented as a white line) overlaid on top of the cross-section from the plan set drawing (presented as a green line). As mentioned previously, pavement structures on the eastbound (EB) side did not show any major distress, and the cross-section of EB at STA 1025+00 created using the LiDAR measurements is in very good agreement with that derived from the plan set drawings, showing a difference of less than 0.01 ft (0.3 cm) between them; this proves the validity of the LiDAR measurements. On the other hand, the cross-section of the westbound (WB) side shows a notable difference from the design cross-section, showing a difference as large as 1.02 ft (31.1 cm).

Details of the cross-section of the westbound side at STA 1025+00 are presented in Figure 10. Lane settlements were observed from the edge of the MSE wall to 36.9 ft (11.2 m) toward the inside lane, roughly corresponding to the longitudinal construction joint between the middle lane and the inside lane. The recorded settlement was 0.42 ft (12.8 cm) at the center of the middle lane, whereas the settlement at the edge of the outside shoulder was 0.29 ft (8.8 cm). Additionally, lateral movements of 0.6 ft (18.3 cm) and 0.94 ft (28.6 cm) were observed at the bottom and top of the MSE wall, respectively; the outside face of the end barrier (i.e., traffic rail) showed a lateral movement of 1.02 ft (31.1 cm). These observations indicate that the MSE wall has experienced both translation and rotation. The LiDAR survey data suggest that the tilting degree of the MSE wall is about 1.4°.

The settlement and tilting of the MSE wall were also observed from the visual surveys. Figure 11 shows the riprap at the bottom of the MSE wall of the subject section and that on the east side of the bridge. The plan set drawing, presented as an inset drawing in Figure 11, shows that the riprap should be constructed with a 4:1 slope, having a higher elevation at the wall facing. The riprap on the eastside of the bridge indeed showed a 4:1 slope as designed, but the riprap at the subjected section showed a reversed slope due to the settlement and rotation of the MSE wall.

3.3. Evaluation of Soil Conditions in MSE Wall

3.3.1. Soil Modulus from DCP Test and CPT

Performances of pavement structures are often directly related to the conditions of embankments or retaining walls on which the pavement structures are supported. To evaluate soil conditions in the MSE wall of the subject section, DCP tests were conducted at 17 locations. DCP tests are commonly used to assess the subgrade conditions beneath the pavement structure, but the maximum investigation depth of a DCP test is typically limited to less than 10 ft (120 inches or 3 m). Since the maximum height of the MSE wall of the subject section was about 25 ft, CPTs were performed at five additional locations to characterize soil conditions at deeper depths. Two DCP tests (DCP4 and DCP8) and one CPT test (CPT8) were performed as controls at locations where relatively less distresses were observed, and the rest of the DCP tests and CPTs were performed at locations where major distresses were observed. The locations of DCP tests and CPTs conducted in the subject section are shown in Figure 12.

To have direct access to the backfill soils, coring or drilling was performed through the pavement structures and DCP tests were conducted through the cored or drilled holes with investigation depths varying from 81.4 in (2.1 m) to 102.7 in (2.6 m) below the pavement surface. Out of 17 tests, 3 DCP tests (DCP1 ML, DCP1 IL, and DCP9 OS) had early refusals, perhaps due to the presence of the cement-treated backfills, and, therefore, the results from the three DCP tests are not presented in the subsequent sections.

The results of DCP tests are usually interpreted from the cumulative DCP blow counts vs. depth curves to determine the resilient moduli of soils. The slope of the cumulative DCP blow counts vs. depth curve represents a penetration rate (i.e., penetration depth per blow), and the steeper slope indicates a greater penetration rate or lower modulus of soil. Webster et al. [48] proposed the following equation to estimate the resilient modulus (M_r) of soil from the penetration rate of DCP:

$$M_r[ksi]=2.55{\left(\frac{292}{{PR}^{1.12}}\right)}^{0.64}$$

(1)

where PR = penetration rate in mm/blow. Figure 13 shows this process, taking results from DCP3 OL LWP as an example.

Resilient moduli of soils in the MSE wall, determined via the aforementioned procedure, at the locations of all 15 non-refusal DCP tests are presented in Figure 14.

Weighted average values of resilient moduli were then computed at each DCP test location using the thicknesses of sublayers of soils as weights, and they are presented in

Table 1. Weighted average value of resilient modulus at each DCP location.

Test Location	Investigation Depth below Pavement Surface [in (m)]	Weighted Average Value of Resilient Modulus [ksi (MPa)]
Along the outside lane left wheel path from the east to the west of the subject section
DCP5 OL LWP	101.4 (2.58)	10.3 (71.0)
DCP3 OL LWP	102.7 (2.61)	10.2 (70.3)
DCP6 OL LWP	102.0 (2.59)	5.3 (36.5)
DCP7 OL LWP	101.6 (2.58)	7.0 (48.3)
DCP4 ML (Control)	82.8 (2.10)	15.8 (108.9)
DCP8 OL LWP (Control)	81.5 (2.07)	17.8 (122.7)
At Location 2 from the outside lane to the inside lane
DCP2 OL RWP	93.3 (2.37)	14.0 (96.5)
DCP2 OL LWP	96.7 (2.46)	9.8 (67.6)
DCP2 ML	88.8 (2.26)	4.4 (30.3)
DCP2 IL	91.1 (2.31)	7.9 (54.5)
DCP2 IS	81.4 (2.07)	16.4 (113.1)
At Location 3 from the outside lane to the inside lane
DCP3 OL RWP	98.5 (2.50)	17.4 (120.0)
DCP3 OL LWP	102.7 (2.61)	10.2 (70.3)
DCP3 ML	98.8 (2.51)	4.8 (33.1)
DCP3 IL	102.7 (2.61)	6.8 (46.9)

. At control sections, weighted average values of resilient moduli were 15.8 ksi (108.9 MPa) at DCP4 ML and 17.8 ksi (122.7 MPa) at DCP8 OL LWP. On the other hand, results from DCP2 OL LWP, DCP2 ML, DCP2 IL, DCP3 ML, DCP3 IL, DCP6 OL LWP, and DCP7 OL LWP tests show that the weighted average moduli at these non-control locations are less than 10 ksi (69 MPa). The lowest value of the weighted average of resilient modulus was observed at DCP2 ML, with a value of 4.4 ksi (30.3 MPa), roughly one-fourth of the value observed at the control location, namely DCP8 OL LWP.

As shown in Table 1, the maximum DCP investigation depth was 102.7 in (2.6 m) at Location 3, whereas the height of the MSE wall at Location 3 was about 17.8 ft (5.4 m). Figure 15 presents a plan set drawing that shows the height of the MSE wall at the CPT locations. To characterize soil conditions in the MSE wall for the entire height, as well as soils beneath the MSE wall (i.e., founding soils), CPTs were conducted through drilled holes (the depths of drilled holes were about 31 in (0.79 m)) at Locations 3, 5, 6, 7, and 8 in the close proximities of the corresponding DCP tests, typically within 2 ft (0.6 m) of each other. The maximum CPT investigation depth was 46.7 ft (14.2 m) below the pavement surface at Location 5.

Cone resistances from all five CPTs are plotted in Figure 16. It is observed that cone resistance values from a depth of about 2.5 ft (0.8 m) to 20 ft (6.1 m) below the pavement surface are typically less than 15 tsf (1.4 MPa). Also, CPT8 (control section) shows the largest cone resistance values, and CPT7 shows the smallest values. Variations in cone resistance up to a depth of 9 ft (2.7 m), which was about the maximum depth of the DCP tests, are also presented in Figure 16.

The results of CPTs are commonly used to estimate the engineering properties of soils, such as friction angle and relative density for sandy soils and undrained shear strength for clayey soils [49]. Unfortunately, very limited research was conducted into the relationship between cone resistance and resilient modulus, and there are no well-accepted correlations between them in the literature. Mohammad et al. [50] proposed a statistical model for predicting resilient modulus from a miniature cone resistance but acknowledged that more tests were needed to develop a general model because their study was limited to a specific soil type. Similarly, Dehler and Labuz [51] examined the feasibility of using CPT data to estimate resilient modulus through CPTs and laboratory tests and concluded that the statistical analysis model used by them was not sufficient to allow the use of CPT to measure resilient modulus in the field.

Due to the lack of a well-established model for the relationship between the CPT cone resistance and resilient modulus, a correlation study between the resilient moduli estimated from DCP results and cone resistances from CPTs was carried out for the project site to empirically determine the resilient moduli using the CPT results. At locations where both DCP tests and CPTs were performed (i.e., locations 3, 5, 6, 7, and 8), the resilient moduli from DCP tests were compared to the average cone resistance from CPTs for the corresponding soil layers. Then, regression analysis, as shown in Figure 17, was performed to establish a site-specific correlation between them for the subject section.

The resilient moduli of backfill soils were then determined using the established correlation from the cone resistance measurements for the entire CPT investigation depths. Figure 18 presents the resilient modulus estimated via CPTs vs. depth, and the bottom of the MSE wall at each CPT location is indicated in the figure (note that CPT8 is a control location). Figure 18 shows that resilient moduli of backfill soils in the MSE wall from the non-control CPTs are, overall, less than 10 ksi (69 MPa) for the entire MSE wall height, except for near the wall base, suggesting the poor conditions of backfill soils in the MSE wall. In particular, CPT6 and CPT7, locations where severe distresses were observed in the outside lane, showed resilient moduli even smaller than 8 ksi (55.2 MPa) (note that the weighted average values of resilient moduli estimated from DCP at DCP6 OL LWP and DCP7 OL LWP were also smaller than 8 ksi (55.2 MPa), 5.3 ksi (36.5 MPa), and 7.0 ksi (48.3 MPa), respectively, as presented in Table 1).

3.3.2. Soil Properties from Laboratory Tests

A total of four soil borings were performed on 30 and 31 March 2023 to take soil samples in the MSE wall and beneath it, and a series of laboratory tests were conducted on the soil samples. Soils in the MSE wall were high-plastic clays, with average values of LL of 70 and PI of 37, and classified as CH per USCS and A-7-6 per AASHTO. Soils beneath the MSE wall were also high-plastic clays, with average values of LL of 70 and PI of 40, and classified as CH per USCS and A-7-6 per AASHTO. The water contents of the soils inside the MSE wall ranged between 30% and 39.4%, with an average value of 34.4%.

Table 2. Summary of laboratory test results.

Sample Location	Soil Classification Per USCS	Soil Classification Per AASHTO	Water Content (%)	% Passing #200	LL (a)	PI (a)
In MSE wall	CH	A-7-6	30–39.4 (34.4) (b)	90–93.6 (92.6) (b)	64–75 (70) (b)	32–43 (37) (b)
Beneath MSE wall	CH	A-7-6	22–40.6 (32) (b)	90.7–99.7 (96.7) (b)	38–84 (70) (b)	22–48 (40) (b)

Notes: ^(a) LL = liquid limit; PI = plasticity index.; ^(b) Values in parentheses indicate averages.

summarizes the results of the laboratory tests conducted on the soil samples.

The existence of high-plastic clays (CH) with high water contents in the MSE wall suggests that the high shrink–swell potential of CH soils, combined with repeated traffic loading, might have contributed to the differential movement among lanes and facilitated the water infiltration. Furthermore, the large amount of water infiltrated into the MSE would have applied large lateral pressure onto the MSE wall due to the low hydraulic conductivity of CH soils, explaining the tensile failure of tie-bars presented in Figure 7a. More details of potential distress mechanisms are presented in the following section.

4. Potential Distresses Mechanisms

Although there is a limited understanding of the mechanisms underlying lane separation and faulting occurrences in concrete pavement, Fowler et al. [52] suggested the following mechanism:

“Over time environmental changes due to temperature and moisture cause curling and warping in the slab. In areas with poor drainage and expansive soils, differential movement of the pavement lowers or lifts the outside edge of the shoulder slab, creating a lever arm that is restrained by the tie-bars. These cyclic changes allow free water and non-compressible debris into the joint, increasing the stress on the tie-bar with every significant cycle. Meanwhile, the free moisture in the joint continues to slowly corrode the tie-bar until its reduced section ruptures or pulls out of the concrete. Once several tie-bars in a row have failed, the slabs are free to move apart from each other. This can allow more water into the joint, resulting in a softening of the sub-base support.

After vertical friction is lost between the two slabs at the longitudinal joint, the slabs will no longer effectively transfer wheel loads across the joint. Then, as traffic wheel loads cross the longitudinal joint, the two adjacent slabs will move independently. The slab carrying most of the wheel loads will deflect more than adjacent slabs, creating a pumping action from traffic in wet weather. As water infiltrates the failed, separated joint, pumping removes waterborne fines from the sub-base, replacing them with more water. This cycle continues until enough sub-base is softened and removed to create a large enough void beneath that the pumping slab begins to permanently subside, or fault” (Pages 1 and 2, Fowler et al. [52]).

The mechanism proposed by Fowler et al. [52], in conjunction with the results of the investigations performed in this study, provide valuable insights into the potential distress mechanism of the subject section, which is graphically illustrated in Figure 19. Figure 19a presents the appearance of the subject section immediately after construction. The presence of high plastic clays inside and beneath the MSE wall, combined with repeated traffic loading and environmental factors, causes the settlement of the MSE wall (Figure 19b), perhaps partially dragging down the riprap. Differential movements of pavement due to cyclic changes in temperature and moisture can allow free water to enter the joint, and the water infiltration into the backfill soils of the MSE wall induces greater lateral pressure behind the MSE wall, causing lateral movement and tilting (Figure 19c). Such lateral movement and tilting of the MSE wall causes not only the further settlement of riprap but also lane separations, thus imposing very large tensile stress on tie-bars. The free moisture in the joint continues to slowly corrode the tie-bars until its reduced section fails in tension due to the large tensile stress induced via the lateral movement and tilting of the MSE wall. Once several tie-bars have failed, more water enters the joint, weakening sub-base support. Repeated traffic loading with the softened sub-base creates voids beneath the pavement (Figure 19d). With the greatly reduced vertical friction between slabs due to lane separation and tie-bar failures, as well as loss of support due to the voids, the pavement slabs can now undergo significant downward movement under heavy wheel loads, causing shear failure of tie-bars and faulting (Figure 19e,f). Although it is not certain whether the subject section experienced a sequence exactly the same as that presented in Figure 19, the distresses observed in the subject section well support the potential mechanism of lane separation and faulting proposed by Fowler et al. [52].

5. Repair Strategy Adopted

As previously mentioned, to properly determine repair strategies, the depths of low-quality backfills in the MSE wall should be identified. The authors employed CPTs to achieve this goal and were able to evaluate the backfill conditions for the entire MSE wall height through the site-specific correlation between DCP tests and CPTs. The resilient moduli estimated from the correlation indicated that the entire backfill materials were in poor conditions, except for near the MSE wall base, as shown in Figure 18. Accordingly, it was decided to remove the materials from the outside shoulder to the middle lane, from the pavement surface to about 2.9 ft (0.9 m) above the wall base, and replace them with new backfill soils, which will be cement stabilized. The new pavement structure will consist of 12 in (30.5 cm) of CRCP, 1 in (2.5 cm) of asphalt stabilized base (ASB), and 12 in (30.5 cm) of cement stabilized base (CSB).

6. Conclusions

This case study presented a comprehensive evaluation of continuously reinforced concrete pavement (CRCP) supported on an in-service MSE wall. The subject section showed various types of distresses, such as lane separation, faulting, settlement, and lateral movement of the MSE wall. Extensive field investigations were performed to assess the conditions of the pavement structures and MSE wall, including FWD tests, coring, DCP tests, CPTs, LiDAR surveys, and soil borings. Extensive analyses of results from field and laboratory tests led to the following major conclusions:

• The distresses observed in the subject section are believed to be primarily caused by the high plasticity of the backfill soils, which are prone to shrink and swell with moisture variations. Additionally, the increased lateral pressure behind the MSE wall due to water infiltration and the low hydraulic conductivity of CH soils appears to have accelerated the progress of distress. Furthermore, the continuous water infiltration behind the MSE wall is attributed to the weakened embankment soils and material washout.
• The resilient moduli of the backfill soils, estimated via DCP tests conducted at the locations where major distresses were observed, were significantly smaller than those at control locations. In particular, the resilient moduli of the backfill soil were less than 8 ksi (55.1 MPa) where severe pavement distresses were observed.
• FWD testing results revealed that the overall slab deflections in the area where distresses were observed were significantly greater than the typical deflection values for well-performing 12 in (30.5 cm) CRCP in Texas.
• Core samples taken through the pavement structure showed excellent concrete quality. For one core sample extracted from a location where exceedingly high slab deflections were observed, there was a layer of polyurethane foam directly beneath the concrete, suggesting the existence of a significant void in the past due to the washout of the base materials.
• The cone penetration tests (CPTs) were utilized to evaluate the conditions of backfill materials for the entire MSE wall height and identify the boundary of the soft backfill soils. The resilient moduli of the backfill soils, estimated through the correlation between the results of DCP tests and CPTs, indicated that the entire backfill materials were in poor conditions, except for near the bottom of the MSE wall. Accordingly, the adopted repair strategy was to remove the existing materials and replace them with cement-treated backfills from the pavement surface to 2.9 ft (0.9 m) above the MSE wall base.