Long-Term Pavement Performance Automated Faulting Measurement

CHAPTER 2. LITERATURE REVIEW

An extensive literature review was conducted concerning joint detection and faulting computation on jointed concrete pavements. Twelve documents were reviewed, and key findings are listed in the following summaries.

Saghafi et al. researched "Artificial Neural Networks and Regression Analysis for Predicting Faulting in Jointed Concrete Pavements Considering Base Condition."⁽⁷⁾ The objective of the research was to predict jointed concrete pavement faulting by considering base layer and other parameters of the LTPP database using artificial neural networks (ANN) and multivariate linear regression (MLR) analysis. Results showed that the ANN approach predicted JPCP faulting more accurately than a MLR analysis, with higher coefficient of multiple determination (R square) values and a very low error level.

Chang et al. studied the "Practical Implementation of Automated Fault Measurement Based on Pavement Profiles."⁽⁸⁾ The authors developed an AFM module in the ProVAL to allow practical application of the AFM method for State department of transportation users. The ProVAL AFM method detects transverse joints and cracks on JPCP and computes joint faulting. The ProVAL AFM serves as the basis for AASHTO R36-12, "Standard Practice for Estimating Faulting on Concrete Pavements."⁽³⁾

Miller and Bellinger developed the Distress Identification Manual for LTPP Program (Fourth Revised Edition).⁽⁴⁾ This manual provides information required for accurate, consistent, and repeatable distress evaluation surveys, including measurement of faulting. It provides graphics illustrating distresses found in three basic pavement types: asphalt concrete-surfaced, jointed (plain and reinforced) Portland cement concrete, and continuously reinforced concrete. These graphics provide a reference to assess distress type, severity, and measurement. Moreover, they provide a method to measure faulting and instructions on how to calibrate and operate fault measurement devices.

Khazanovich et al. worked on LTPP Data Analysis: FAQs About Joint Faulting with Answers from LTPP.⁽⁹⁾ (This Federal Highway Administration TechBrief was developed from the study "Common Characteristics of Good and Poorly Performing Portland Cement Concrete (PCC) Pavements.") The objective was to examine the LTPP database and identify site conditions and design features that significantly affect transverse joint faulting. The emphasis of the study was to find what works and does not work to control the development of joint faulting. The document provides answers to questions regarding design features and site conditions that lead to "good" (better than expected) and "poor" (worse than expected) performance of jointed concrete pavements with regard to joint faulting. In addition, guidelines are provided to assist highway agencies with what works and what does not work in the design of transverse joints to control joint faulting.

Sayers and Karamihas published The Little Book of Profiling.⁽⁵⁾ The objective of the book was to provide basic instructions on measuring and interpreting road profiles.

Nazef et al. worked on A Semi-Automated Faulting Measurement Approach for Rigid Pavements Using High Speed Inertial Profiler Data.⁽¹⁰⁾ The study objectives were to determine an appropriate profiler sampling interval to accurately locate transverse joints and to determine how well faulting estimated from profile elevations using an AFM algorithm compares with faulting measured with the GFM. The AFM algorithm accurately detects, on average, 95 percent of transverse joints from profile data collected at highway speed using a 17.3-mm sampling interval. This algorithm was also adapted to estimate faulting measured with the GFM in accordance with the AASHTO R36-04 protocol. Although the algorithm results are repeatable, the algorithm over-estimated the faulting at joints by 1.3 mm to 1.5 mm compared with faulting measured with the GFM.

Nazef et al.conducted a validation study, "Alternative Validation Practice of an Automated Faulting Measurement Method."⁽¹¹⁾ The objective of the study was to evaluate the accuracy and precision of an HSIP-based AFM system using a two-phase approach. The first phase evaluated the HSIP's ability to produce reliable faulting measurements under controlled conditions. The second phase tested the validity of the automated method to produce repeatable and reproducible results under normal field conditions. The goal was to use the results from this study to support the implementation of the AFM system in FDOT's Annual Pavement Condition Survey process. Except for one HSIP, all profilers achieved a minimum profile repeatability cross-correlation of 92 percent. Under controlled conditions, the HSIP has a faulting measurement accuracy and repeatability of 0.60 mm and 0.65 mm, respectively. The HSIP has a positive joint detection rate (JDR) ranging from 80 to 94 percent. Under controlled conditions, the HSIP has accuracy, repeatability, and reproducibility rates of 1.2 mm, 1.1 mm, and 0.5 mm, respectively.

Selezneva et al. worked on Preliminary Evaluation and Analysis of LTPP Faulting Data—Final Report."⁽⁶⁾ The objective was to evaluate the quality of LTPP faulting data, including identification of missing and questionable data. Faulting data indexes (average joint faulting for each visit) and related statistical parameters were developed. Subsequently, these parameters were used to determine the impact of joint faulting and related data in identifying factors that affect joint faulting. Analysis indicated that doweled joints exhibit very little faulting even after many years of service and that the effects of design features, such as drainage, tied-concrete shoulder use, and joint spacing, are not as significant when doweled joints are used. For non-doweled JPCP, the following design features were found to significantly reduce faulting: widened lanes, an effective drainage system, a stabilized base/subbase, and narrower joint spacing. The effect of faulting on ride quality was also investigated on JPCP sections with three or more faulting and IRI surveys. A strong correlation was found between the rate of change in faulting values and the rate of change in IRI values for JPCP sections. The results indicate that faulting is a major component of increased roughness of JPCP.

Perera et al. worked on LTPP Manual for Profile Measurements and Processing.⁽¹²⁾ The objective was to provide detailed information on operational procedures for measuring longitudinal pavement profiles for the LTPP Program using the ICC road profiler, Face® Dipstick®, and the rod and level. The manual also explained calibration of equipment, data collection, record keeping, and data processing.

Karamihas and Senn conducted a study, Curl and Warp Analysis of the LTPP SPS-2 Site in Arizona.⁽¹³⁾ The authors examined the roughness and roughness progression of 21 test sections over the first 16 years of the experiment to analyze slab curl and warp effects on pavement roughness. As part of the study, the authors conducted faulting analysis (detecting joint locations from multiple profiles) using profile data to examine the effect of joint faulting on IRI values. All test sections except 0262 and 0265 produced average faulting of less than 1.27 mm. For sections 0262 and 0265, the severity of faulting grew throughout the experiment, and the increase in IRI with time was primarily due to faulting.

Vedula et al. published "Adaptability of AASHTO Provisional Standards for Condition Surveys for Roughness and Faulting in Kansas."⁽¹⁴⁾ In this study, profile data were collected on about 346 km of Kansas highways following AASHTO provisional standards PP-37-00 for quantifying roughness and PP-39-00 for faulting, and the Kansas Department of Transportation (KDOT) standard for condition surveys. The comparison of statistical analysis results from the algorithms following the KDOT Network Optimization System and the AASHTO provisional standards (PP-37-00) indicated that roughness measurements tended to produce statistically similar results. However, fault values computed from AASHTO PP-39-00 and KDOT automated faulting procedure were significantly different even after some modification to PP 38-00 following current practices in Kansas.

Watkins of the Mississippi Department of Transportation developed a joint/crack-location algorithm based on a brute-force method by determining the elevation difference between adjacent samples greater than 2.03 mm using a profile sampling interval rate of 12.7 mm.⁽¹⁵⁾