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.

12 calibrated individually, but a new reference calibration proce- dure allows multiple sensors to be calibrated simultaneously. SHAs typically perform reference calibrations once per year (Appendix B, question 14). An annual reference calibration is also recommended by LTPP (Schmalzer 2006). caLiBraTion procedures relative Differences between FWD models, sensor manufacturers, and available technology have led to several relative calibra- tion methods. According to a survey conducted for this synthesis, 55% of SHAs use a relative calibration procedure developed by Strategic Highway Research Program (SHRP)/LTPP. This procedure is detailed in the Long-Term Pavement Perfor- mance Program FWD Reference and Relative Calibration Manual (Schmalzer 2006). Conversely, 35% of respondents said they follow their FWD vendor’s relative calibration procedure. If the ven- dor is Dynatest or JILS, then the LTPP procedure is being followed. According to information provided by KUAB, they recommend that their clients in the United States follow the LTPP procedure and that their Swedish clients fol- low the Swedish Road Administration (SRA) method. The SRA method places all of the FWD’s sensors into a holder and subjects them to five consecutive drops. Calibration is successful if the largest and smallest measured deflections differ by no more than 2 μm (0.0787 mils) plus 1% of the measured value. Section 7.3.1 of ASTM D4694-96 describes a relative calibration procedure, which uses a vertical sensor holding tower. In a manner similar to the SRA method, five deflec- tions must be measured per sensor, and if they differ by no more than 0.3% from the average deflection then no correc- tion is required. Section 7.3.2 recommends repeating the procedure some distance away from the load plate so that “if any differences in average deflection greater than 2 μm (0.08 mils) are found, the device should be repaired and recali- brated according to the manufacturer’s recommendations” (“Standard Test Method for Deflections . . .” 2005). This chapter discusses FWD calibration practices and recommendations. If FWDs are not calibrated, the conse- quences can be financially significant. According to a study by the Indiana DOT (INDOT) (yigong and Nantung 2006), overestimating a deflection by 0.0254 mm (1 mil) resulted in 26% more undersealing area. This error resulted in $20,000 in unnecessary drilling and $29,000 in additional asphaltic materials. By simulating a 0.0508 mm (2 mil) deflection overestimate, $37,000 of additional drilling and $54,000 of additional asphaltic materials were deemed necessary, although they were actually unwarranted. Similar trends were observed on an asphalt concrete (AC) overlay project; additional deflections of 0.0254 mm (1 mil) led to additional $11,187.50 per lane-km ($17,900 per lane-mi) for asphal- tic materials, and 0.0508 mm (2 mil) errors led to $23,625 per lane-km ($37,800 per lane-mi) of additional materials. Conversely, underestimated deflections led to significantly reduced pavement design life. Underestimating deflections by 0.0254 mm (1 mil) translated to an AC layer 25.4 mm (1 in.) thinner than needed, resulting in a decrease of 2.8 mil- lion equivalent single axle load of pavement life. caLiBraTion TYpes relative Relative calibrations ascertain sensor functionality and rela- tive accuracy. All sensors should produce the same output when in the same position at the same site location (“Stan- dard Test Method for Deflections . . .” 2005). To achieve this, SHAs typically perform relative calibrations once per month (Appendix B, question 15). A monthly relative calibration is also recommended by LTPP (Schmalzer 2006). Relative calibrations can be performed at any location, in situations in which pavement layers are adequately strong. For example, 44% of survey respondents stated that they perform relative calibrations on a “calibration pad,” a specially designed PCC floor, and 33% stated that relative calibrations are done on an “in-service pavement” (Appendix B, question 16). reference These calibrations are done at specially designed calibration centers. Reference calibrations aim to ensure sensor accuracy according to defined benchmarks. Occasionally, sensors are CHAPTER THREE FaLLinG WeiGhT deFLecToMeTer caLiBraTion

13 TxDOT’s FWD fleet. Because TxDOT’s FWD units are of varying ages, one FWD may produce differing results than another. To provide better reproducibility, a three-phase cali- bration plan was created. These phases are as follows (Rocha et al. 2003): • Physical inspection and component replacement • Preliminary calibration—a relative calibration is performed and sensors not passing calibration are identified • Comprehensive calibration—sensors not passing calibration are calibrated more thoroughly and data- gathering issues are troubleshot These new protocols greatly improved consistency from one FWD to another, and the researchers recommended that “TxDOT implement the new protocol as soon as possible.” Section 7 of ASTM D4694-96 acknowledges the UTEP method, which is “more complementary than interchange- able” with the SHRP/LTPP method (“Standard Test Method for Deflections . . .” 2005). portable Falling Weight deflectometer calibration Carl Bro’s PRIMA 100 PFWD was designed to mimic their existing PRI 2100 trailer-mounted FWD. The design employs three geophones, compared with PRI 2100’s nine geophones. Because both models use the same geophones, calibrating the PFWD’s geophones uses an identical proce- dure to the PRI 2100. A time-history system serves as the backbone of Carl Bro’s calibration software, which uses a fast Fourier transformation algorithm. Carl Bro calibration equipment employs a test cell connected to an LVDT, and the procedure is verified by means of the SHRP 1994 protocol (Clemen 2003). caLiBraTion requireMenTs calibration Frequencies For calibration, ASTM D4695-03 recommends that impulse- loading type devices be calibrated “at least once per year using the procedure in Appendix A of SHRP Report SHRP- P-661” for reference calibration and “relative calibration once a month during operation” (“Standard Guide for General Pavement Deflection Measurements” 2005). Additionally, Section 7 of ASTM D4694-96 recommends that deflection sensors be calibrated “at least once a month or in accordance with the manufacturer’s recommendations” (“Standard Test Method for Deflections . . .” 2005). According to the LTPP manual, reference calibration is required once per year, unless the FWD is based in Alaska, Hawaii, or Puerto Rico. Similar requirements are detailed for Georgia (Pavement Design Manual 2005) and Florida, reference Reference calibrations, per LTPP, must be performed at a specialized facility. A pooled fund study was commissioned by the FHWA in 2004 to improve the reference calibration process. Improvement to the original 1994 SHRP reference calibration procedure was needed for the following reasons (Orr et al. 2007): • The 1994 procedure was designed around Dynatest and KUAB FWDs, the only commercially available FWDs in the United States at the time. Because of differences between FWD manufacturers, the original procedure was not completely compatible with equipment from other manufacturers. • The 1994 procedure required individual sensor calibra- tion, and took six hours to complete as a result. • The 1994 procedure used a linear variable displace- ment transducers (LVDT) for reference deflections, the accuracy of which was occasionally compromised by movement of the mass and beam to which it was mounted. Accelerometers were viable replacements for LVDTs, because they are self-referencing. • The 1994 procedure used DOS-based software. DOS is no longer the state-of-the-art PC operating system. The new procedure addresses each of the pooled fund study’s points. Universal compatibility is achieved through modified triggering mechanisms. Time is saved by placing all FWD sensors into a single support stand and calibrat- ing them simultaneously. New accelerometer-based control board and data acquisition systems were designed. A new program, WinFWDCal, was written in Microsoft Visual Basic to provide a graphical user interface (GUI) for calibra- tion. Calibration is recommended once per year, but it takes only about two hours to complete (Orr et al. 2007). By 2006, the updated FWD calibration procedure included an FWD calibration results database, conversions were made to the DOS-based FWDCAL software to work with Micro- soft Windows, and software was adapted to work with accel- erometers and modern data acquisition boards. Additionally, WinFWDCal was augmented with a utility to convert FWD file formats from the different equipment types to the PDDX format adopted by AASHTO. For sensor calibration, a single support stand was designed so that sensor position was not significant. With such a support stand, all of an FWD’s sen- sors may be tested simultaneously, as opposed to the one- sensor-at-a-time calibration method put forth by the 1994 procedure. The finalized calibration procedure is discussed in a draft final report (Irwin 2006). Independent of the FHWA-pooled fund study, another calibration procedure was developed at the University of Texas at El Paso (UTEP). With the support of the Texas DOT (TxDOT), UTEP developed a new calibration protocol for

14 caLiBraTion cenTers To implement the 1994 calibration procedure, LTPP opened four calibration centers. Currently, those centers are oper- ated by the DOTs in Colorado, Minnesota, Pennsylvania, and Texas. Additionally, the privately operated Dynatest calibration center in Florida and JILS calibration center in California provide calibration services. These calibration centers have since put the 2007 FHWA calibration process into practice. At the Texas calibration center, the TxDOT method can also be used. On average, the Texas and Penn- sylvania DOT calibration centers see about 30 FWDs per year, the Colorado calibration center sees about 20, and the Minnesota calibration center sees about 8. None of the cali- bration centers charge SHAs for calibration services; how- ever, the Colorado, Pennsylvania, and Texas centers charge at least $300 per session to private firms. Training requirements differ slightly from one calibra- tion center to another. For example, Pennsylvania calibration center technicians are trained by Cornell University and are certified by MACTEC. Minnesota, on the other hand, has not implemented an in-house training program as of the date of this report, but it plans to do so in the future. Sixty-three percent of survey respondents support the construction of additional calibration centers; however, 76% of survey respondents stated that they are not willing to sponsor such a calibration center (Appendix B, questions 20–21). which also require that “manuals describing the relative calibration procedure and other aspects of deflection test- ing should be kept in the vehicle and office” (Holzschuher and Lee 2006). Relative calibrations are required either monthly, if the FWD sees regular usage, or within 42 days of any given data collection operation. Checklists are given for before transit, before operations, and after operations (Schmalzer 2006). Travel distances and costs associated with calibration Because relative calibrations can be performed on any sur- face where pavement layers are adequately strong, travel distances tend to be short. According to survey data, 52% of SHAs travel 0.62 km (1 mi) or less to a relative calibration site. Additionally, relative calibrations tended to be inex- pensive. The same survey showed that 52% of respondents spend less than $100 per relative calibration (Appendix B, question 19). Reference calibrations, on the other hand, involve signifi- cant expenditures of travel time and money. Because most states require reference calibration once per year at one of the FHWA-certified calibration centers—four LTPP cen- ters, one privately operated by Dynatest, and one privately operated by JILS—57% of survey respondents reported that they travel 805 km (500 mi) or further for their reference calibrations. Similarly, 64% of survey respondents reported that, including total labor, materials, travel, and other inci- dental expenses, a single reference calibration costs more than $1,000.