Falling Weight Deflectometer Usage (2008)

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.

29 of subgrades. A presentation by VDOT describes its FWD testing on its entire Interstate highway network. To provide a structural component to Virginia’s pavement management system (PMS), FWD data were collected on Interstate 77 every 0.3 km (0.2 mi) and at three load levels. Additionally, FWD data were used to determine project acceptance. On Interstate 64, FWD testing was done at early-age cracking sites. The data showed “weak structure” and the contractor was “asked to remove and replace unaccepted pavement sec- tions” (Habib 2006). TxDOT owned 15 FWDs in 2003 and used them to test 5%–10% of network-level highways for their PMS. On the project level, TxDOT collected FWD data for “load zoning, design, forensic studies, joint load transfer on Jointed Con- crete Pavement (JCP), and many projects for determining structural adequacy” (Beck 2003). The Pennsylvania DOT outlines its pavement design pro- cedures in its Pavement Policy Manual (2007), including new pavement designs. According to procedures outlined in chapter six, “Pavement Design Procedures,” new pavement design submissions must include a table of Mr values backed up by either FWD data or lab tests. Additionally, federal-aid pavement preservation projects require patching percentages generated by FWD and by visual inspection. Back-calcula- tion of Mr values is permissible only under five scenarios: • Full depth bituminous pavement sections, • Existing bituminous overlays on thin concrete pave- ments (original concrete pavements less than 8 inches in depth or any parabolic sections), • Existing bituminous overlays on concrete pavements which suffer from severe alkali silica reaction (ASR) degradation, [and] • Directly on subgrade and subbase (this situation is rare) (Pavement Policy Manual 2007). case 2. paVeMenT rehaBiLiTaTion and oVerLaY When contractors perform pavement resurfacing projects in the state of Alabama, an FWD test is required (“ALDOT Procedure 390 . . .” 2004, p. 14). This chapter discusses how FWD data has been applied to various agency activities. The case studies cover the follow- ing topics: • Data collection and analysis refinement • Pavement rehabilitation and overlay • PCC joint sealing evaluation • Pavement management systems • Load transfer efficiency • Void detection • Spring load restrictions • Nonresilient pavement layer behavior • Utility cuts • Experimental paving materials • Project acceptance and evaluation • Conversion of data from other NDT devices • International practices case 1. daTa coLLecTion and anaLYsis reFineMenT The Kansas DOT sponsored a study of LTE and tempera- ture, during which FWD planning lessons were learned. FWD data were collected at one site along Interstate 70 at various times through the year. Because temperature has such significant effects on LTE and other pavement proper- ties, the Kansas DOT drafted the following recommenda- tions (Corn 2005): • Plan FWD data collection operations with climatic conditions in mind. • Test during temperate climate months. • Test approach and leave slabs. • Do not test while the ambient temperature is higher than 27° C (80° F), per the AASHTO recommendation. • “Don’t expect the expected.” VDOT uses FWD pavement testing at the project level, at the network level, and for forensic investigation of pavement failures. AC, jointed concrete pavement (JCP), CRCP, and composite pavements are all subject to FWD testing. At the project level, VDOT derives PCC elastic moduli, composite modulus of subgrade reaction (k-value), LTE, and presence of voids from FWD rigid pavement testing. On flexible pave- ments, they derive SNeff, layer moduli, and resilient moduli CHAPTER EIGHT appLicaTions oF FaLLinG WeiGhT deFLecToMeTer daTa— case sTudies

30 AC overlay, were developed at yONA Engineering Con- sulting & Management in Israel. This method reduces SNeff calculations into five equations. Previously, SNeff determi- nation was done through an empirical process developed by AASHTO. Deflection basins measured by FWDs form the first step in the yONAPAVE process. The area under a plot of FWD deflections versus distance forms the basis for a characteristic length 10 equation, whose value is then car- ried through a calculation of subgrade modulus of elasticity Esg. SNeff can then be determined from the 10 and Esg values. The SNeff values generated by the yONAPAVE process cor- respond “in generally good agreement” with values calcu- lated by the MODULUS or AASHTO NDT interpretation approach (Hoffman 2003). To accomplish the elimination of spring load restric- tions by 2002, the Alaska DOT and Public Facilities (Alaska DOT&PF) conducted FWD tests on all roads where spring load restrictions had been practiced. In all cases, recommen- dations were made for AC overlays on the existing AC pave- ments based on the FWD data (Bush and Tohme 2003). case 3. JoinT seaLinG eVaLuaTion The Concrete Pavement Technology Program Task 9 research study, Cost-Effectiveness of Sealing Transverse Contraction Joints in Concrete Pavements (Hall et al. 2004), utilizes FWD analysis of joint load transfer and voids to analyze the performance of PCC pavements with sealed and unsealed joints. According to the authors, the study is expected to answer the following questions: • What are the effects on long-term performance of unsealed transverse joints in concrete pavements with different pavement cross sections and slab dimensions, traffic levels, and climatic conditions? • What are the effects of different transverse joint seal- ant materials and configurations on the long-term per- formance of concrete pavement in various climatic conditions (climatic zones)? • Is sealing transverse contraction joints cost-effective for different pavement designs and materials over a range of climatic zones and traffic levels (Hall et al. 2004)? case 4. paVeMenT ManaGeMenT sYsTeMs Several states include FWD data in their PMSs. For exam- ple, Nebraska’s PMS stores FWD data that are collected by the Materials and Research Division. These data are used for “structural capacity analysis, evaluation of existing subgrade strength, and overlay analysis.” Deflection test locations and frequency will vary according to project conditions (State of Nebraska Pavement Management Systems 2007). In a technology brief by IDOT’s BMPR, specific uses of FWD data were listed. On rigid pavements, the brief lists the following uses (“Pavement Technology Advisory . . .” 2005): • Locating areas of poor support beneath jointed con- crete pavements • Determining load transfer across transverse and longi- tudinal joints • Estimating subgrade and pavement layer elastic moduli values (E1, E2, etc.) • Developing cost-effective maintenance and rehabili- tation alternatives On flexible pavements, the IDOT brief lists the following uses (“Pavement Technology Advisory . . .” 2005): • Determining the structural adequacy of a pavement and identify causes of failure • Determining uniformity of support along a project and identify weak areas • Estimating subgrade and pavement layer elastic moduli values (E1, E2, etc.) • Developing cost-effective maintenance and rehabili- tation alternatives When conditions warrant, California’s PCC roadways may be rehabilitated by replacing individual slabs. These slabs, which must measure at least 3.6 m (12 ft) in width and 2 m (6.6 ft) in length, are considered good candidates for replacement if a visual inspection shows two or more corner breaks; if they contain “third-stage cracking,” segments that move relative to each other, longitudinal or transverse cracks wider than 13 mm (0.5 in.), or cracks with 150 mm (6 in.) or more of spalling; or if they are no longer supported because of settlement, base failure, or excessive curling. Once a visual inspection is done, the guidelines recommend FWD usage, along with drainage analysis and coring, to determine pave- ment condition (“Slab Replacement Guidelines” 2004). The New york State DOT (NySDOT) evaluates pave- ment structural capacity using FWD, but not on the project level. Additionally, tests are done only in the following situa- tions (Comprehensive Pavement Design Manual 2000): • Deflection survey of 50 statewide sites as a part of pavement performance monitoring program • Deflection survey of 48 statewide sites as part of a SUPERPAVE performance monitoring program • Determination of load transfer efficiencies at joints and cracks of PCC pavements • Determination of the appropriateness and effectiveness of cracking and seating, and rubblizing operations The yONAPAVE algorithms for evaluating the effective structural number (SNeff), and thereby the thickness of an

31 The procedure is documented in ODOT’s Manual for Aban- doned Mine Inventory and Risk Assessment (1998). case 7. sprinG Load resTricTions The North Dakota DOT imposes limits on the per-axle weight trucks may carry during the spring thaw. These spring load restrictions are imposed to save the pavement layers from otherwise-avoidable and significant damage. A North Dakota DOT website (“Implementation of Spring Load Restrictions . . .” 2007) details the three main factors used to determine when the restrictions should be posted. Direct strength measurements, interpreted from FWD data, are combined with long-range weather forecasts and tem- perature probes. SDDOT adopted using FWD data for spring load restric- tions in 1996. SDDOT had recorded centerline miles sub- jected to spring load restrictions since 1969. In the years since spring load limits were instituted in South Dakota, the percentage of road network mileage requiring load restric- tions during the spring thaw has generally decreased. When FWD data became a criterion for spring load restrictions in 1996, the number of lane-miles subject to restriction increased temporarily (12.7% of the roadway network in 1996 versus 11.1% in 1994 and 1995), but continued their downward trend thereafter. By 2007, 3.5% of SDDOT’s roadway network was subject to spring load restrictions. SDDOT attributes the additional limits to FWD data (“2007 Spring Load Restriction Summary” 2007). NySDOT utilized FWDs to study the seasonal variabil- ity of pavement layer moduli. Regions experiencing winter- freeze, spring-thaw conditions in the soil undergo severely weakened subgrade layers during the thaw season. Because such seasonal differences in pavement layer moduli severely affect pavement surfaces, pavement designers must com- pensate for them. Six possible seasons were identified (Orr 2006): • Freezing, when frost is present in less than 100 mm (4 in.) of the subgrade layer. • Winter, when at least 100 mm (4 in.) of the subgrade layer contains frost and no thaw is present. • Spring thaw, when any thaw in the unbound layers is present and some portion remains frozen. • Spring recovery, when resilient modulus increases quickly because of drainage. • Spring, when all frost has thawed, but precipitation outpaces evaporation. • Summer, when evaporation outpaces precipitation. To identify when such seasonal parameters are necessary, Cornell University and NySDOT developed a geographical model that shows which portions of New york are subject case 5. Load TransFer eFFiciencY An example of a forensic study using LTE was conducted in Michigan (Peng et al. 2005). A time history analysis of the deflection data showed that the dowels were likely loose. Deflection testing showed that permanent loss of slab con- tact with the base (void) existed near the doweled joint. case 6. Void deTecTion FWD data are used to detect voids where pavement layers have no support. undersealing To fill voids under a PCC pavement, injection holes are drilled into the pavement and a grout of cement, fly ash, and water is pumped through the holes. This procedure is referred to as “undersealing” by the South Dakota DOT (SDDOT). Before and after drilling holes, voids in the pavement are detected using FWD data. The FWD loading plate is placed as close as possible to the slab corner, and the LTE to the adjacent joint is measured. If the measured deflections fall out of a range determined by the state engineer as acceptable, under- sealing procedures begin (Standard Specifications for Roads & Bridges 2004). In a research report, MoDOT disseminates their FWD void detection efficiency findings. Because voids under PCC- bridge approach slabs contribute to premature cracking, early detection of these voids is crucial to avoid costly replacement and rehabilitation measures. Based on the study’s findings, MoDOT recommends that FWD should be used to determine voids under PCC slabs. This recommendation assumes that the FWD and operator are available, undersealing is being considered as a preventive maintenance treatment, and one or more of the following conditions are met (“Void Detection with the Falling Weight Deflectometer” 2004): • Long lane closures for proof-roll testing are not desir- able (e.g., at bridge approaches with reduced shoulder widths and high-volume routes). • Fewer personnel than required with proof rolling are available for testing. • The pavement shoulder is unstable for accurate proof- rolling measurements. • More clear and quantifiable indications of undersealing improvements than proof rolling can provide are desired (i.e., AASHTO rapid void detection procedure). abandoned Mine detection The Ohio DOT is experimenting with the use of FWD data to supplement investigations of abandoned mine detection.

32 case 10. experiMenTaL paVinG MaTeriaLs Crushed aggregate, a popular base course for pavements, became progressively more expensive. To save money on base courses, FDOT has sponsored recycling concrete aggregate (RCA) research. FWD data were used to test vari- ous RCA mixes and the results show RCA to be a viable base course for roadway pavements (“Guidelines and Specifica- tions for the Use of Reclaimed Aggregates . . .” 2001). Ultra-thin whitetopping (UTW) was evaluated using FWD data in Minnesota. FWD data were collected one year after an experimental UTW pavement test section was con- structed at the Minnesota Road Research test facility. PCC thickness varied from section to section; the study’s intent is to determine an ideal PCC thickness. Strain data captured by the FWD showed a good bonding condition between the lower bituminous surface and the new PCC wearing course. Although an optimal UTW overlay design is not yet deter- mined, “the dynamic strain measurements indicate that there is a better bond between the asphalt and the overlay in the thinner sections.” It was also observed that the magnitude of the strains in the thinner sections were more dependent on the stiffness of the asphalt than the number of equivalent single axle loads accumulated (Vandenbossche and Rettner 1998). case 11. proJecT accepTance and eVaLuaTion FWD tests have potential for use during construction. FWD data may be used for the following (Clark 2005): • Subgrade strength improvements before structural sec- tion construction. • Subbase and base layer monitoring after structural sec- tion construction. • LTE on jointed plain concrete pavement. • Baseline development. As an example, FWD tests on Virginia State Highway 288 showed where a cement-treated base needed to be placed during construction. A second example, where CRCP was placed, showed deflections greater than 0.14 mm (5.5 mils). These deflections indicated poor construction joints, and further investigation showed reinforcing steel at the wrong depth. FWD was again used as a diagnostic tool along U.S. Highway 29, where two stations showed poor support. Although FWD could be used as an acceptance criterion, contractors would have to be familiar with their use and be able to afford one. to significant seasonal variation. FWD data were gathered at varying sites throughout the year. The model was then built using backcalculated FWD moduli. Additionally, the seasonal model was used to design a 20-year AC overlay in Arcadia, New york. case 8. nonresiLienT paVeMenT LaYer BehaVior Because the FWD has replaced the Benkelman beam as the primary pavement analysis and design device, measured layer moduli now include plastic deformations as well as recoverable deformations. Mechanistic design practices assume that all layers behave resiliently. In the past, these additional plastic deformations were assumed negligible; however, nonresilient behavior may be observed given a load of significant magnitude. The practice of “16 (FWD weight) drops at four load levels with four replicates at each drop height or load level” may result in nonresilient behav- ior. Such behavior can be detected by statistical tests. Two statistical methods of nonresilient behavior detection were tested using FWD tests at Cornell University. Tests were performed from February until May 2003. No trends were observed through ANOVA (analysis of variance) tests but chi-squared variance tests on the center sensor data revealed nonresiliency during the spring-thaw season. ANOVA tests “will detect systematic variations; however, if the deflec- tions are not always generally increasing or decreasing for a given load level, the test does not detect when nonresilient behavior is occurring” (Orr 2003). case 9. uTiLiTY cuTs The Iowa DOT sought to improve utility cut repair tech- niques. Utility cuts often settle over time, which leads to “uneven pavement surfaces, annoyance to drivers and, ultimately, further maintenance.” Causes of the settlement include differing backfill material between jurisdictions, excessive volumes of backfill materials “placed at bulking moisture contents,” and the lack of quality assurance or con- trol. FWD data showed that backfill materials within utility cuts—as well as an area 0.6 to 0.9 m (2 to 3 ft) beyond the cut perimeter—were susceptible to settlement. The Iowa DOT will continue to monitor its utility cuts using FWD tests, as well as nuclear gauges, dynamic cone penetrometers (DCPs), Clegg hammers, and laboratory tests. These data “will be studied with the goal of increasing pavement patch life and reducing the maintenance of the repaired areas” (Research News 2007).

33 pavement product is achieved after the pavement work has been performed” (Frabizzio et al. 2002). In Kentucky, a 5.1 km (3.17 mi) section of Interstate 265 was examined following a pavement reconstruction project. In addition to FWD data, ground-penetrating radar (GPR) testing and coring were completed along the segment. The PCC slabs showed transverse cracks and differential settle- ment. The FWD data were used to determine layer stiffness and LTE. Although the Kentucky Transportation Cabinet accepts LTE values of 90% or greater, all slabs in the study area had less than 90% LTE (Rister et al. 2003). case 12. conVersion oF daTa FroM oTher nondesTrucTiVe TesTinG deVices FDOT elected to replace its entire Dynaflect fleet with FWDs. This decision is attributable to FWD providing a more accurate simulation of actual traffic loads, its use as a pavement research tool, and its adoption by LTPP. Because FDOT testing data were collected by means of Dynaflect before FWD adoption, conversion from Dynaflect to FWD was needed. A linear correlation was found to make the conversion, and this study refines this correlation. FWD, Seismic Pavement Analyzer, and Dynaflect data were col- lected at pavement sites throughout Florida, and statistical correlations were determined. Additionally, the researchers conducted a state-of-the-practice literature review, as well as a survey of SHAs. FWD data were processed into Mr and soil support value using MODULUS, EVERCALC, the AASHTO method, and a finite-element modeling program. FWD Mr data backcalculated through MODULUS showed a strong correlation to Mr values collected by the Dynaflect (R2 = 0.867). Similarly, Dynaflect Mr values correlated strongly with FWD data processed through EVERCALC (R2 = 0.742) and through the AASHTO method (R2 = 0.925) along 483 km (300 mi) pavement sections. In cases in which pavement testing had been performed by LTPP, the LTPP database was “found to be the best database available to deduce general patterns of the pavement behavior during field testing.” The researchers reached 16 conclusions. The following 11 con- clusions were relevant to this synthesis (Tawfiq 2003): • FWDs accurately simulated vehicle loads on pavements. • Other NDT devices did not accurately simulate vehicle loads. • Thick AC layers, very thin AC layers, shallow bedrock, and heavier loads may have given unrealistic data and should be compensated for before performing back- calculation. • Calibration was crucial. In a research report for VDOT (Diefenderfer and Bry- ant 2005), pavement warranty contracts are suggested for future rehabilitation projects. VDOT considered requiring pavement contractors to enter into warranty contracts. Such warranties ensure quality pavements over the course of a pavement’s design life. In some cases, however, competition between contractors was reduced. As a potential study case, an AC overlay project was chosen. FWD data were employed in the AC overlay design phase and before acceptance. FWD data collection included using four load levels spaced at loca- tions 22.9 m (75 ft) apart. After a jointed reinforced concrete pavement rehabilita- tion was completed along Interstate 287 in New Jersey, FWD data were used for “assessing the existing condition of the mainline pavement, investigating the causes of premature distresses in the mainline pavement, and monitoring the effectiveness of slab undersealing at joint locations.” For example, after a pavement rehabilitation project was com- plete, low- to medium-severity transverse cracks appeared. “The FWD, DCP, and compressive strength test results were used to evaluate the condition of the various pavement lay- ers. A normalization load of 40 kN (9,000 lbf) was used for the analysis of the FWD test results. That is, the FWD deflections from the actual applied loads were normalized or adjusted to the values that would have resulted if a 40 kN (9,000 lbf) loading had been applied.” The FWD and compressive strength test results revealed that the PCC layer was in fair-to-good condition. The average backcalculated PCC layer modulus (EPCC) was almost 34,500 MPa (5,000 ksi), whereas the average compressive strength of the PCC layer was 60 MPa (8,700 psi). The FWD results indicated that the support to the PCC layer was adequate at midslab locations, because the average backcalculated modulus of subgrade reaction (k) value was 5.5 kg/cm3 (200 pci), a “fair value.” However, the DCP test results indicated low Cali- fornia Bearing Ratio (CBR) values (average CBR = 46%) for the nonstabilized open graded (NSOG) layer. Notwith- standing the LTE values (92% on average), the FWD joint test results indicated that the pavement was not performing well at joint locations. The joint deflection (i.e., deflection directly beneath the center of the FWD load plate during joint testing) and joint intercept values (indicative of slab support) were fairly high—average values of 9.6 mils and 2.1 mils, respectively. These results suggest that voids likely exist beneath the slabs near joints, and excessive vertical slab movement consequently occurs at these locations. These voids were promptly undersealed. Additionally, high degrees of nighttime slab curl were confirmed by FWD testing, exac- erbated by the nonstabilized open graded base layer instead of a more densely graded base material. The researchers concluded that “the FWD can thus act as an evaluative and investigative tool during the early stages of a project and as a quality control instrument to ensure that the desired final

34 included deflection data taken at 200 m (656 ft) dis- tances in both directions and moduli are backcalcu- lated using an “equivalent thickness approach.” • Ireland—FWD data are collected in 200 m (656 ft) sec- tions “with measuring distances of 25 to 50 m (82 to 164 ft).” These data are used to classify pavement layer bearing capacity, subgrade layer bearing capacity, and AC overlay thickness. In the state of Western Australia, FWD testing has gradu- ally replaced the Benkelman beam as the standard deflection testing mechanism since the 1990s. Two FWDs are present in Western Australia. “With the arrival of FWDs, Main Roads Western Australia (MRWA) commenced conducting network level FWD deflection survey together with profi- lometer survey for roughness, rutting, and texture measure- ments in annual basis.” Annual calibration was required and a calibration center based on the SHRP 1994 protocol was built in Perth. Traffic control was performed by a “driver operator” with an escort vehicle warning sign behind the FWD trailer. Test methods followed the ASTM D4694-96 protocol. Sensors were located at 0, 203, 305, 406, 508, 610, 762, 914, 1,524 mm (0, 8, 12, 16, 20, 24, 30, 36, and 60 in.) from the load plate. Tests were conducted at 50 kN (11,200 lbf) to simulate Western Australia tire pressure loads. Data were collected every 0.8 km (0.5 mi) in the outer wheel path. The data gathered were used “together with Rutting, Roughness, Surface Texture, and Skid Resistance to the key pavement performance indicators for the Road Network Maintenance Contracts,” and thereby to determine payment to contractors. Furthermore, “based on the network deflec- tion data the Contractors select the sections requiring project level pavement investigation for rehabilitation works.” The presence of FWDs has greatly increased testing efficiency. “Average production rate was 250 deflection tests per day. Typical production rate of Benkelman Beam tests is 80 to 100 per day.” Testing integrity is validated through calibra- tion, data auditing, marking the test point, and accounting for environmental factors. Additionally— Deflection data from the tests carried out on or close to the outer wheel path after the rainy season were generally high. This reflected wetting of subgrade and weakening pavement edges. Seasonal effects on deflection data between the successive surveys can be reduced if deflection testing is carried out around the same time of the year (Sapkota 2003). • MODULUS gave no indication of invalid FWD data, whereas EVERCALC gave an error message. • MODULUS and EVERCALC gave similar results, but EVERCALC worked within the Windows GUI and MODULUS required a DOS command line interface. Additionally, EVERCALC was more sensitive to seed moduli. Furthermore, EVERCALC was accompanied by software for overlay design and stress simulation but had poor user manuals. • Sensor D6—placed 36 in. from the load plate— better measured subgrade response than sensor D7. This observation may have been unique to Florida. • While finite-element analysis was generally reliable, bedrock and subgrade moduli were occasionally overestimated. • Soil moisture was not considered for back-calculation strategies, but it can drastically change soil properties. • Bedrock depth was not considered for back-calculation techniques, but it can be determined through finite- element analysis. • FWD data should be coupled with other data to be useful. Such data include bedrock depth and layer thickness. case 13. inTernaTionaL pracTices A nation-by-nation assessment of FWD usage and nonde- structive testing was provided in a report by the European Cooperation in the Field of Scientific and Technical Research (COST), under the auspices of the European Union. The fol- lowing are some of the findings (Beuving 2000): • Spain—At the project level, FWDs were used exclu- sively on rigid pavements; flexible pavements were tested using either Lacroix deflection measurement devices or FWDs. Measurements are taken every 200 m (656 ft), and surveys are completed every 4 years. • Finland—Structural assessments are performed using KUAB FWD data. Measurements are taken “not fur- ther apart than 500 m (1,640 ft)” and are completed every 3 to 5 years. The primary parameter derived from FWD data is the spring Bearing Capacity Ratio; how- ever, plans were under way to switch to the Structural Condition Index. • Denmark—Deflection data from FWD and average daily traffic are used to determine structural pave- ment capacity. Denmark’s PMS, in place since 1988,