Assessment of Continuous Pavement Deflection Measuring Technologies (2013)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

8This chapter presents the approaches used for the following: • Identification of the most promising technologies; • Determination of user needs; • Selection of candidate devices; • Data collection, data analysis, and interpretation; • Development of example applications; and • Identification of recent developments and possible improvements. Identification and Assessment of Available Technologies The project started with a review of literature, case studies, and ongoing research in order to (1) identify the deflection measur- ing devices that had the potential to meet the project require- ments and (2) assess their potential to meet user requirements. The review was completed in two stages. The first stage con- sisted of an Internet search for relevant papers, performed by searching for related key words such as “continuous deflec- tion device”; the second stage consisted of using the same key words to search the English language International Transport Research Documentation (ITRD) and Compendex databases for any relevant papers stored therein. This was complemented with the extensive personal archives of the members of the research team who have had many years of experience in this research field. Once these searches were completed, the papers and documents were used to compile a list of available survey vehicles capable of measuring pavement deflection, whether at traffic speed or otherwise. Prescreening of Candidate Devices The list of survey vehicles was then summarized in a compre- hensive table detailing the important aspects of each device (e.g., survey speed, measurement interval, personnel requirements, etc.). After reviewing this table, some devices were disregarded as irrelevant to this project, and an abridged version was produced grouping similar devices (e.g., variations on the deflectograph). Information presented in this table includes equipment type, model, characteristics, survey speed, and development status (whether the device is a current or former production model or a working prototype). The devices were further subcategorized into three groups: static measurement devices, moving measurement vehicles with stationary mea- surement apparatus, and moving measurement vehicles with nonstationary measurement apparatus. Definition of Continuous Deflection Device The purpose of this project was to identify current measure- ment devices capable of continuously measuring pavement bearing capacity without the need for the vehicle or measure- ment equipment (relative to the vehicle) to remain stationary while surveying. For this report, a continuous deflection device has been defined as a deflection measuring device constantly moving that can collect data at intervals of approximately 300 mm (1 ft) or smaller using load levels typical of truck load- ing (i.e., 40 to 50 kN [9 to 11 kips] per wheel or load assembly); the ideal solution would be a device requiring no traffic con- trol. It should be noted that this definition covers the interval at which the data are collected, not reported. The latter is often much greater than the former because of the significant mea- surement noise level generated during data collection. Averag- ing over these longer lengths reduces the noise considerably while retaining the important pattern of the road’s deflection in response to strength variations. Determination of User Needs The demand and the potential value of continuous deflec- tion devices for use in developing optimum pavement reha- bilitation strategies for rapid renewal projects were evaluated through a survey of state and provincial DOTs. The survey C h A p T e r 2 Research Approach

9 included questions to assess technical needs and also endeav- ored to determine the value assigned by the agencies to the collected data. The survey of state and provincial DOTs was divided into two stages. In the first stage, a web survey was sent to the differ- ent DOTs. The survey was divided into five sections based on the type of information requested. The first section collected contact information about the respondents. The second sec- tion focused on current practices and uses of deflection testing by DOTs. The third section inquired about pavement reha- bilitation design procedures and how, if applicable, deflection testing results are used for this purpose. The fourth section focused on whether and how deflection testing results are used in pavement management applications. The fifth and final section gave respondents the opportunity to provide general comments. In the second stage of the survey, follow-up phone interviews were conducted with nine states where continu- ous deflection devices have already been used or the states are facing substantial renewal challenges on high-traffic-volume roadways. These states were identified based on the web survey results. Selection of Candidate Devices Following the literature review, an assessment of the capa- bilities of available continuous deflection devices (moving measurement vehicles with nonstationary measurement appa- ratus), and considering the information obtained during the user needs survey, two devices that offer the most promising technologies to address the needs of end users were chosen for further study. These devices are the rolling wheel deflectome- ter (RWD) and the traffic speed deflectometer (TSD). Sample data from the two devices were collected and processed as part of the preliminary assessment of their capabilities. Data Collection To assess the capability of continuous deflection technology to support network-level pavement management decisions, rela- tively long sections with uniform and variable structural con- ditions were selected for site testing. These network-level sites included flexible, composite, and rigid pavement sections. All attempts were made to include subsections with good, fair, and poor functional conditions within each of these pavement types. This enabled the evaluation of the capabilities of the devices for network-level use. Some sections were measured several times in succession in a single day, as discussed in the following sections. The evaluation included testing on different types of pave- ments, under various operational conditions, and reference FWD deflection testing equipment where possible. The plan included sections that were evaluated following a protocol applicable for network-level data, and a subset of these sec- tions was subject to a detailed evaluation as discussed in the following sections (illustrated by Figure 2.1). The final exper- imental design focused mainly on evaluating the capabilities of the general technology of traffic-speed continuous deflec- tion measurements and the application of this technology for supporting pavement management decisions. From the network-level testing routes, a few carefully selected evaluation sections were also assessed using FWD measurements. The data collected in these locations were used to assess the ability of the systems for detecting weak spots. TSD Data Collection To answer the main evaluation question posed earlier (that is, the capability of the devices to support network-level pave- ment management decisions), researchers assessed the TSD in the United Kingdom by identifying a number of evaluation sec- tions. Each section measured approximately 2 to 4 km (1.25 to 2.5 mi) in length and generally incorporated weak subsections; researchers conducted repeat TSD surveys of these sections, as well as FWD surveys on most sections, at spacing of up to 20 m (60 ft). The accuracy and consistency with which the TSD iden- tifies any strong and weak sections was also evaluated. Sites were chosen by examining the construction and structural condi- tion of a significant sample of the English road network. These evaluation sections covered a variety of structural designs and ages. In general, road sections were selected to cover a range of structural conditions as shown by the deflection response. Table 2.1 lists these sites and a summary of the key parame- ters of each site. The nominal deflection responses are equiva- lent peak central FWD deflections at a load of 50 kN (11 kips). The table includes the following: • Flexible sites, where the main structural layers are of asphalt or granular construction and can be broadly classified as fully flexible (i.e., with asphalt upper layers and granular lower layers); • Composite sites, where asphalt and cement-bound layers are both structural layers (i.e., with asphalt upper layers, usually greater than 150 mm [6 in.] thick, hydraulically bound base layers without joints, usually termed “lean concrete” in the United Kingdom); and • Rigid sites, with pavement quality concrete as the primary load-bearing layer, sometimes with asphalt upper surfacing layers, less than 75 mm (3 in.) thick. Some of these sites, at least two from each pavement type, were each surveyed repeatedly during the course of one day. Comparisons with FWD measurements were made on three of the flexible sites, one of the composite sites, and two of the rigid sites.

10 Network-level routes Evaluation sections Sub-sections of interest (detailed project-level analysis) Figure 2.1. Schematic of the section selection. Table 2.1. Summary of Potential Network-Level TSD Sites Site Length mi (km) Asphalt Thickness in. (mm) Cement Bound Thickness in. (mm) Nominal Deflection Response mils (mm) Structural Variability Site Use Flexible UK_F1 1.25 (2.0) 3 (75) NA 8–32 (0.2–0.8) Very high R and C UK_F3 2.0 (3.2) 10.5 (270) NA 4–8 (0.1–0.2) Low R and C UK_F5 2.0 (3.2) 12 (310) NA 4–10 (0.1–0.25) High R and C UK_F6 2.2 (3.5) 12.6 (320) NA 4–12 (0.1–0.3) Low R Composite UK_C1 2.5 (4.0) 7 (175) 7 (175) 2–14 (0.05–0.35) High R UK_C2 2.1 (3.4) 6.25 (160) 8.25 (210) 2–8 (0.05–0.2) Low R UK_C3 1.3 (2.1) 6.25–11.5 (160–290) 6–14.2 (150–360) 2–14 (0.05–0.35) High R and C Rigid UK_R2 0.9 (1.4) 4 (100) 8 (200) 8–80 (0.2–2.0) High R and C UK_R3 1.9 (3.0) 3 (70) 9 (225) NA Low R and C Note: R = repeatability assessment; C = comparison with FWD measurements for comparability assessment; and NA = not available.

11 RWD Data Collection The original plan included the collection of RWD data fol- lowing a protocol similar to the one used for the TSD. The data analyzed in Phase I of the project were collected with an older version of the device; this allowed only a prelimi- nary assessment of the device’s capabilities. The device has since been upgraded, by modifying the lasers ensemble and truck suspension, and was expected to perform bet- ter than it had in the various applications reported in the literature. However, the upgraded device had some prob- lems and was not fully operational during the second phase of the project. Thus, the planned tests were not possible within the time frame available and only data collected in the first phase of the project were used for the preliminary assessment of the capabilities of the RWD. Data from only two routes, one in Virginia and one in New Mexico, were used for assessing the repeatability and comparability of the RWD, respectively. The equipment operator, Applied Research Associates (ARA), in cooperation with the Federal Highway Administration (FHWA), is working to correct the problems. Once they have been corrected, FHWA plans to support additional field testing. Data Analysis Results from previous testing were analyzed in Phase I of the project. This preliminary analysis was performed using com- mon methods such as correlation and regression analysis. These methods provide a quick analysis; they do not allow for accurate evaluation of equipment capabilities. For Phase II of the project, the data analysis was expanded to allow for accurate evaluation of repeatability and comparability. The collected data were analyzed to evaluate the repeatability of the TSD and RWD and the comparability of both devices by comparison with the FWD. Results of different studies that evaluated the RWD (conducted in the United States) and the TSD (conducted in Europe) were analyzed. It was noted in this analysis that repeatability and reproducibility were not uniformly defined across all those studies. Measures and methodologies used to evaluate the devices included corre- lation, regression analysis, standard deviation, and in some cases subjective visual inspection of plots. The drawbacks associated with the use of correlation and regression analysis to evaluate repeatability and comparability are discussed in this chapter. Then, repeatability and comparability analysis based on the limits of agreement (LOA) method suggested by Bland and Altman (1986) is recommended and used to evalu- ate the continuous deflection devices. A method of evaluat- ing repeatability from one run is presented and compared to the method based on the LOA. Finally, the use of smoothing splines as a tool to remove the noise from TSD deflection slope measurements is investigated. This smoothing splines denoising methodology shows potential to improve the fre- quency at which useful information can be obtained (i.e., data averaging distance). In this report, repeatability (or compa- rability) is defined as the 95% confidence interval of the dif- ference between repeated measurements (difference between measurements of TSD and FWD or RWD and FWD). example Applications Based on the data collected and the analysis, a number of exam- ple applications for the data produced by the devices have been developed and are presented in Chapter 4. These applications illustrate the use of continuous deflection measurements for network-level pavement management. They are not meant to be comprehensive, but rather show how deflections can be used to address a number of issues applicable to modern pavement management practices. The applications presented include using measurements obtained from continuous deflection measurement devices to do the following: • Segment pavement sections into homogeneous sections; • Estimate the strain at the bottom of the asphalt layer and the effective structural number (SN) of the pavement; and • Identify relatively weak pavement sections as well as weak pavement sections defined by absolute thresholds. The segmentation was performed using a statistically based binary segmentation algorithm. The strain at the bottom of the asphalt layer and the effective structural number were estimated using the difference between two deflection slope measurements from the TSD. The identification of relatively weak pavement sections was used to demonstrate the abil- ity of the device to locate anomalies or locally weak sections within a network, whereas the identification of weak pave- ment sections based on thresholds demonstrated the ability of the device to repeatedly identify weak sections. Current and Future Developments Since Phase I of this study was completed, the research team has discovered one additional device that might meet the original requirements of this study. However, as this informa- tion was discovered since the commencement of the study, it has not been included in the data acquisition and comparison Phase II of the project. Furthermore, both the RWD and the TSD have been developed further since Phase I was completed. A description of these developments is presented in Chapter 4, along with recommendations for further improvement.