Home

Developing a Testing Device for Total

1. 16 Figure 2 16 Field Calibration Curves for the Three Rolling Sensors eeeeeessssesssssssssssseeeereresees 16 Figure 2 17 TXDOT s FSF at Austin Bergstrom International Airport ABIA 17 Figure 2 18 Continuous Deflection Profiles at 0 5 and 1 mph Using the Center Rolling Sensor in the Towing Frame along Path E at TXDOT Fab 18 Figure 2 19 Continuous Deflection Profiles at 0 5 and 2 mph Using the Center Rolling Sensor in the Towing Frame along Path E at TXDOT Fab 18 Figure 2 20 Continuous Deflection Profiles at 0 5 and 3 mph Using the Center Rolling Sensor in the Towing Frame along Path E at TXDOT Fab 19 Figure 2 21 Continuous Deflection Profiles at 0 5 and 1 mph Using the Front Rolling Sensor in the Towing Frame along Path E at TXDOT Fab 20 1X Figure 2 22 Continuous Deflection Profiles at 0 5 and 2 mph Using the Front Rolling Sensor in the Towing Frame along Path E at TXDOT Fab 20 Figure 2 23 Continuous Deflection Profiles at 0 5 and 3 mph Using the Front Rolling Sensor in the Towing Frame along Path E at TXDOT Fab 21 Figure 2 24 Continuous Deflection Profiles at 0 5 and 1 mph Using the Rear Rolling Sensor in the Towing Frame along Path E at TXDOT Fab 21 Figure 2 25 Continuous Deflection Profiles at 0 5 and 2 mph Using the Rear Rolling Sensor in the Towing Frame along Path E at TXDOT Fab 27 Figure 2 26 Continuous Deflection Profiles at 0 5 and 3 mph Using the Rear Roll
2. O CTR Measurements Actual Peak to Peak 0 5 mph 9 00AM 70F Dynamic Load 8 kips Ex 1 mph 9 30AM 71F NO O Deflection mils 10kips CO 0 100 200 300 400 500 600 700 Distance ft Figure 2 18 Continuous Deflection Profiles at 0 5 and 1 mph Using the Center Rolling Sensor in the Towing Frame along Path E at TxDOT FSF 16 CTR Measurements 0 5 mph 9 00AM 70F 14 Actual Peak to Peak 2 mph 9 30AM 72F Dynamic Load z 8 kips X O Deflection mils 10kips CO 6 4 2 0 0 100 200 300 400 500 600 700 Distance ft Figure 2 19 Continuous Deflection Profiles at 0 5 and 2 mph Using the Center Rolling Sensor in the Towing Frame along Path E at TxDOT FSF 18 CTR Measurements 0 5 mph 9 00AM 70F 14 Actual Peak to Peak 9 mph 9 38AM 72F Dynamic Load z 8 kips Deflection mils 10kips CO Distance ft Figure 2 20 Continuous Deflection Profiles at 0 5 and 3 mph Using the Center Rolling Sensor in the Towing Frame along Path E at TxDOT FSF The deflection profiles collected with the front rolling sensor at testing speeds of 0 5 1 2 and 3 mph are shown in Figures 2 21 2 22 and 2 23 respectively The deflection profiles collected with the rear rolling sensor at 0 5 1 2 and 3 mph are shown in Figures 2 24 through
3. A 2 Installation of TPADana System Requirements TPADana has been developed for computer systems that meet these requirements most new computers do e Windows 2000 with Service Pack 3 or Windows XP with Service Pack 2 e 1 0GHz Processor e 2GB RAM e 20GB free hard drive space e 17 inch monitor with at least 1440x1024 resolution The TPAD data processing software 1s installed by running the setup software as follows Double click the TPADana setup program with the icon 5 RDDanasetup exe then follow the instructions step by step If the following dialog box appears this means the version of the existing installed DLL file Dynamic link library file is older than the file already existing in your computer Always answer No to keep the latest version of the DLL file Conde systema POEL TDL Thus ile exits and ri a rore recent verizon than the file te metai Do you want to averwrie the installed version amnia The installation procedure also asks the user for the installation folder and you can choose the default folder or select create the folder you like The whole installation takes about 50 two minutes The installation program also loads one set of TPAD data into a folder called Testdata this is a 1000 foot set of data collected on US 287 After the installation you can find the installed program icon EN TPADana exe on the desktop or in the program folder Double click this icon to start the TPADana p
4. critic i 3 0 1 2 0 to move the averaged distance window by another ft Click the Save and Run button Again go to the Matlab main screen and click deflection and then copy and paste the array to the Excel spreadsheet creating column E Click distance and copy and paste the array to the Excel spreadsheet creating column F Note that the user needs to create the column headings 2 ft Deflection and Distance see Figure 13 Load Data File Button Szen Figure 8 Screenshot of Matlab Importing the F1 txt File 75 Ir bie Ww kapai Deka leds W ndow Pw Dg Get A fee ran ee amp m dud w Be i SSS Am HA ooo BS E ey mmus Lemma Lm ZE Remain 11 HER TEE Lais Kr ani ab ee EI k Jess iEn Leer Jours F4 txt Loaded to LAS 11 4 Fh dde enm L p HA D a Matlab Figure 9 Screenshot of Matlab after Importing the F1 txt File Inps 5 Debug Dinie Wade Hiig Greta JU Ha Add E a Fare LTE BEC os sel LA Te Fis E Teu Co Tasli Dale Dr rzg Write Hee GE M 1 hee Bar fi Ger di Zeg sp mam el 5 3 E e A i 7 B s i BR e E Ep qs num I I i L 13 I l l I VM Le Lei 7 L c EEUE ETETE p dm Figure 10 Screenshot of Matlab after Importing the Analysis m File pe aed Al Le ge det
5. 1 in wide treads and one 2 in wide tread e SB 9 5 in diameter wheels two 2 in wide treads and one 4 in wide tread e SC 12 5 in diameter wheels two 1 in wide treads and one 2 in wide tread and e SD 12 5 in diameter wheels two 2 in wide treads and one 4 in wide tread Photographs of the four wheel sets of the rolling sensor installed 1n independent the towing frame are shown in Figure 2 2 a SA 9 5 1n Diameter Wheels Two 1 b SB 9 5 in Diameter Wheels Two 2 in in and One 2 in Wide Treads and One 4 in Wide Treads a Lm s ou ety c SC 12 5 in Diameter Wheels Two 1 d SD 12 5 in Diameter Wheels Two 2 in in and One 2 in Wide Treads and One 4 in Wide Treads Figure 2 2 Four Combinations of Wheels on the Rolling Sensor Installed in the Independent Towing Frame In this study each wheel set on the same sensor cart body was installed in the independent towing frame The towing frame was pulled along a 125 ft testing path at the TxDOT FSF The testing path included seven joints on the JCP Rolling noise measurements were performed at three different speeds The exact speed in each test was not measured However the average walking running speed of the person pulling the towing frame was calibrated by measuring the time it took to walk run a known distance It was determined that testing speeds associated with slow walking fast walking and slow running were
6. 2 26 In all cases the deflection profile determined at 0 5 mph is also used as the reference As with the comparisons shown in Figure 2 18 through 2 20 for the center rolling sensor deflection profiles determined at testing speeds of 0 5 1 and 2 mph with the rear rolling sensor were very well matched see Figures 2 24 and 2 25 while the deflection profile at 3 mph showed some differences especially at the joints see Figure 2 26 On the other hand deflection profiles evaluated with the front rolling sensor at testing speeds of 0 5 1 and 2 mph exhibited good comparisons see Figures 2 21 and 2 22 but exhibited considerable differences at 3 mph especially at the joints see Figure 2 23 The poorer performance of the front rolling sensor compared with the rear rolling sensor is attributed to the chatter created by the front swiveling wheels on the towing frame compared with the fixed wheels on the back of the towing frame This chattering noticeably increased with speed 19 Deflection mils 10kips A Figure 2 21 Deflection mils 10kips A CTR Measurements 0 5 mph 9 00AM 70F 4 mph 9 30AM 71F Actual Peak to Peak Dynamic Load 8 kips Ln id 100 200 300 400 500 600 700 Distance ft Continuous Deflection Profiles at 0 5 and 1 mph Using the Front Rolling Sensor in the Towing Frame along Path E at TxDOT FSF 0 5 mph 9 00AM 70F CTR M
7. 284 3 284 6 284 9 285 2 285 5 285 8 286 1 286 286 6 286 9 287 2 287 5 287 8 esado l9 O 2M QOO 409 0850 hor N L3 2 ha e 2 Dm Ki o H e N 288 5 288 8 289 2 289 5 289 290 0 290 3 290 6 290 9 291 2 291 5 291 8 TETTT OQ M Q b OO 7 00 Q O N WAM oy N o dub ga Q I Q b 0 O NODO IN W d ol N EN N N 292 3 292 6 292 8 293 1 293 4 293 7 294 0 294 3 294 6 294 9 295 2 295 5 295 8 d W 296 2 296 5 296 8 29 297 4 29 298 0 298 3 298 6 298 9 299 2 299 5 299 8 Figure A 12 Raw Load Data Chart When Turning to Other Pages 57 A 6 Button for Showing the Raw Geophone Chart e As in Section A 3 when the SA button is selected the screen will show the raw geophone data chart shown in Figure A 13 The Y axis is in mils and X axis is in feet The geophone data is much more variable than the load signal this data comes from a section of US287 near Bowie a survey in the northbound direction running at 2 mph Work is continuing with the TPAD to determine the best sensor mounting arrangement and operating speed Figure A 13 Raw Geaphone D Data Chart Y Axis in mils X Axis in SS A 7 Button for Showing the Power Spectrum Chart of Raw Load Data Barr Power spectral density PSD the power spectrum of the time series describes how the power of a signal or time series is distributed with frequency It is a method of defining the SNR where the value at 30 Hz the input loading frequency should be significantly
8. 45 Division personnel Patching CRCP is extremely difficult and there is always a risk of damaging the adjoining concrete 4 The data collected was used to generate a list of 14 locations with their GPS sf coordinates that the District should inspect to determine if additional repairs are required prior to placing the concrete overlay TPAD testing of upcoming CRCP rehabilitation projects should be encouraged Conclusions from TxDOT FSF at Austin are as follows I 2 The TPAD works very well on JCP Areas of poor pavement support can be easily located based on the center slab deflections from the center rolling sensor The front and rear rolling sensors provide the best information for computing LTEs The TPAD will be more than 10 times faster than traditional FWD testing and it can also readily test every joint in the pavement Future efforts should be aimed at automatically generating statistics from the shape of the deflection pattern over joints this seems to be highly correlated to the FWD LTE Temperature has a large effect on LTE joints with poor LTE at lower temperatures see an increase in the LTE as the temperature rises and the joints lock up More work is needed to account for test temperature on LTE determination with the TPAD Conclusions from US 290 near Houston are as follows l The TPAD provided useful data on this full depth HMA pavement 2 Areas with high deflection could be e
9. U ui l ULLA ch A A A 2 mph n IMIA kk i Ad N Ul LU I Y va mat eil e 3 mph i j J Di 3 j nr j Fw id al il i I EAEN ad T a r Tw i Figure 4 2 Repeated Runs of the JCP at Different Speeds with the Center Sensor at TxDOT FSF 4 3 The TPAD Potential for Evaluation of LTE with Comparison to FWD Data The main purpose of the testing at this facility was to evaluate the TPAD s potential to rapidly assess LTE for any JCP The testing involved measurement of LTE with an FWD at two different surface temperatures The FWD set up and the definition of LTE is presented in Figure 4 3 In the upstream computation the load plate was placed on the arrival slab shown in Figure 4 3 LTE was calculated using Equation 4 1 LTE upstream 100 w2 w1 w2 w1 Equation 4 1 Normalized to 9 kips with wl Deflection of wl at joint W2 Deflection of w2 at joint wl Deflection of w1 at center slab W2 Deflection of w2 at center slab For the downstream computation the load plate 1s placed on the departure slab The computation then includes the deflection under the load plate W1 compared to W4 which is the deflection on the unloaded side of the joint For both the upstream and downstream calculations the ratio of unloaded versus loaded deflections is normalized to the same ratio measured at the mid slab 34 zm j Figure
10. 2 5 2 2 Rolling Noise Measurements with an Independent Towing Frame at TxDOT FSF Rolling noise generated at a level near the RDD dynamic loading frequency typically 30 Hz is mainly caused by pavement roughness In general lower rolling noise is desirable Successive testing of three wheel sensor carts to determine the optimum combination of wheel diameter and tread width necessary to minimize rolling noise was conducted on the JCP at the TxDOT FSF The independent towing frame at the TxDOT FSF is shown in Figure 2 1 The rolling sensor computer Data Acquisition DAQ system and battery for the computer and DAQ system installed on the independent towing frame are shown in Figure 2 1a In Figure 2 1b it 1s easier to see the rolling sensor and also the portable air supply used to inflate the hold down air springs b Figure 2 1 Rolling Sensor Computer DAQ and Battery Installed on an Independent Towing Frame on JCP at TxDOT FSF During earlier rolling noise tests in the third year Stokoe et al 2011 three different hold down weights of 20 40 and 90 Ib were evaluated In these earlier tests it was found that the heaviest hold down weight 90 Ib resulted in the least rolling noise Therefore only a hold down weight of 90 Ib was used in this study Different sets of rolling sensor wheels used during the rolling noise measurements with the designations for each set of wheels as follows e SA 9 5 in diameter wheels two
11. 4 3 FWD Set Up for LTE Measurements FWD data was collected on 16 joints known to represent a wide range of LTE values Table 4 1 shows the results obtained at the two surface temperatures Clearly a wide range of LTEs was found on this pavement ranging from 6 to 100 Table 4 1 FWD Determined LTEs at both 84 and 104 F LTESAF LTESAF LTE 104 F LTE 104 F oint Number E pe LO DS GN 35 When reviewing the RDD profiling data from all sensors the research team realized that the most valuable information about LTE would be found in analyzing the data collected with the rear sensor Figure 4 4 presents a schematic illustrating the passage of the three TPAD rolling sensors Over a joint with assumed poor LTE In Figure 4 4 the rolling sensors are the small circles at five different positions A through E with respect to the joint in the JCP The plots at the bottom of the figure are the expected deflection patterns for each of the three sensors The front sensor FS is the lead sensor the deflection it measures will increase as it rolls towards the joint reaching a maximum when it reaches the joint As soon as the front sensor passes over the joint the loading rollers will still be on the other slab and the measured FS deflection will drop substantially The degree that the deflection drops will be a function of the LTE of the joint The rear sensor RS will have a similar pattern but delayed in time It is expected that the
12. 5 the Gd button is used for showing the geophone s Power Spectrum chart as given in Figure A 15 59 Geophone Power Spectrum from point 65000 to 70000 or DMI from 126 49 ft to 141 09 ft Power Spectrum in Loc frequency Hz Figure A 15 Power Spectrum Chart of Raw Geophone Signal A 9 Button for Showing the Chart of Raw and Filtered Load Data Hn This button s operation is very similar to button Dk discussed in Section A 3 except when clicking this button the chart will show the raw and filtered data together 1n one chart as shown in Figure A 16 Notice that the filtered data appears as blue lines and green lines represent the raw load data Since we use the 29 31 Hz band filter the static load component is removed and the filtered data oscillate around the zero On the filtered load signal the red dot specifies the peak location This peak will be used for calculating the continuous deflections 60 129504 to 124900 319 6 319 9 320 2 320 5 320 8 321 1 321 4 321 7 322 0 322 3 322 6 322 9 323 2 323 5 323 6 323 9 324 2 324 5 324 8 325 1 325 4 325 7 326 0 326 3 326 6 326 9 327 2 327 5 327 6 327 9 328 2 328 5 328 8 329 1 329 4 329 7 330 0 330 3 330 6 330 9 331 2 331 5 Figure A 16 Chart of Raw and Filtered Load Data A 10 Button for Showing the Chart of Raw and Filtered Geophone Data m This button s operation is very similar to that of button discussed in Section A 7 except this b
13. 56 Figure A 11 Raw Load Data Chart DMI vs Loadmkmsi 57 Figure A 12 Raw Load Data Chart When Turning to Other Pages see 27 Figure A 13 Raw Geophone Data Chart Y Axis in mils X Axis in feet eeseeeessssse 58 Figure A 14 Power Spectrum Chart of Raw Load Suenal 59 Figure A 15 Power Spectrum Chart of Raw Geophone Signal ccssssseeeeeeeeeeeceeeeeeaeeees 60 Figure A 16 Chart of Raw and Filtered Load Data 61 Figure A 17 Chart of Raw and Filtered Geophone Data 62 Figure A 18 The Comprehensive Results View Screen View Option li 63 Figure A 19 Comprehensive Results View Screen View Option 2 eese 64 Figure A 20 Infrared Sensor for Getting the Surface Temperature 65 Figure A 21 GPS Antenna Mounted at the Top of the Loading Frame 65 Figure A 22 Other Information Charts Surface Temperature Sea Elevation and AR UENO EE 66 Figure A 23 Video Frames Showing the TPAD Cross the Bride eeseeeeeesssss 67 Eisure A 24 Opn Fie E 68 Xl xli List of Tables Table 3 1 Areas Needing to be Inspected Prior to Placing the Overlay 30 Table 4 1 FWD Determined LTEs at both 84 and JOE 35 Table AI File Size Tor the US287 Field Testni eet o ro dE 49 Table A 2 Function of Each Toolbar Button 52 xii X1V Chapter 1 Introduction 1 1 Brief Overview o
14. E BM D GE ar HO i m Dim i Figure 1 Flow Chart of the 6 Step Data Analysis Procedure Used to Calculate Normalized Peak to Peak Dynamic Displacements from the Rolling Sensor Outputs Time Based Data Analysis IGOR software is used for the time based data analysis performed in Steps 1 2 3 and 4 The initial programming was completed by Dr James Bay details are in TxDOT report 1422 3F Bay and Stokoe 1998 Figure 2 shows a screen shot of when IGOR is opened The steps for the time based analysis are as follows 1 Raw data needs to be loaded To load raw data click Load general binary file in the Load waves submenu under the Data tab as shown in Figure 3 For the CTR data processing raw data consists of two parts File name A DAT and File name C DAT A data collector assigns the file name when recording the raw data File name_A DAT consists of one force and up to three geophones and File name C DAT is data from the distance measurement instrument The data type of A DAT is single float and C DAT is double float When Load General Binary Data is clicked the screen shown in Figure 4 appears Both raw data files A DAT and C DAT need to be loaded separately 70 i Aman O aeiia T d Lys T Upgex Junges Les keen Lea 01182511 ee TRAD RD better ad TR FEF 20185405 ra TO FS Ri Den mph 512 7 Codi Cini d ulum Teen 512 1 Cw 7600 otal beten sn largi ET CRIT 182 12202 EA duel dE disini
15. Houston are discussed in Chapter 5 Finally conclusions are presented in Chapter 6 In addition a user s manual of the TPAD data analysis software developed by TTI personnel is presented in Appendix A Appendix B provides a manual for post processing software for calculating peak to peak dynamic deflections from the TPAD RDD rolling sensors developed by CTR personnel Chapter 2 Rolling Noise Measurements Design and Installation of Towing Frame and Field Evaluation of the Complete TPAD RDD System 2 1 Introduction Rolling noise measurements to optimize the wheel configurations on the rolling sensor carts were performed at TxDOT s FSF An independent towing frame that was pulled by hand was used in these tests The pavement at the TxDOT FSF is a JCP Many different combinations of rolling sensor wheels were tested These tests are described in Section 2 2 The final towing frame configuration that 1s attached to the TPAD is also presented in this chapter CEM personnel designed and installed this towing frame on the TPAD The design and installation of the final towing frame are discussed in Section 2 3 The field calibration of the rolling sensors and initial TPAD testing with the final towing frame were performed at the TxDOT FSF Field calibration procedure results and initial TPAD testing results analyzed using the CTR data processing methodology are presented in Section 2 4 Finally a brief summary of Chapter 2 1s discussed in Section
16. a zero in the selected box in Figure A 8 RDD Continuous Deflection Profiles Chart in mils 10 kips Deflection mits TO kipss Figure A 10 Plot of Raw Data Blue Line and Filtered Data Using the Options Selected in Figure AS 56 A 5 Button for Showing the Raw Load Chart After you click the button and the raw data is loaded selecting the button will display the raw load data shown in Figure A 11 In the top left corner the information shows the data point range displayed on this screen The buttons a and P can be used to turn to other pages This chart shows the raw load signal vs the DMI in feet This 1s the beginning of the data file the TPAD is not moving Therefore all the DMI values in Figure A 11 are zero and the load signal is a perfect sine wave When turning to other pages using the button Figure A 12 is displayed then the load signal changes with time and the DMI is no longer zero from 0 to 5396 SET O 2M Q OO 0 QO 2I Ob O o o o e o o o o o o o o o o o o o o o o o o o o o o o o o o o o 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 gt ag cd ad ad dad d d ad ad ad ad eg 93N9o 500 0002N9050 OANWADANODO NWA o o o o o o o o o o o o o o o o o o o o o 39532343 ON Ob 00 0q00O02l oho 0 0 0 0 0 0 0 0 Figure A 11 Raw Load Data Chart DMI vs Load in Kips 118712 to 124108 4 12 11 a T SANWLOAAW9HOO who H o
17. along pavements at speeds in the range of 2 to 5 mph The test functions include those associated with a Rolling Dynamic Deflectometer RDD ground penetrating radar GPR Distance Measurement Instrument DMI high precision differential GPS surface temperature and digital video imaging of the pavement and right of way conditions The towing frame system used to position and pull the RDD rolling sensors was developed and installed on the TPAD The towing frame system minimizes the transmission of vibrations from the TPAD mobile platform and loading system to the rolling sensors and allows for the incorporation of three rolling sensors front center and rear on the TPAD The front and rear rolling sensors have 12 5 in diameter wheels and a 90 Ib hold down weight while the center rolling sensor has 9 5 in diameter wheels and a 40 Ib hold down weight Initial deflection measurements using the new towing frame sensor cart arrangement were performed at the TxDOT Flight Services Facility FSF Based on these tests it is recommended that most deflection profiling be done with the center and rear rolling sensors at a testing speed of 2 mph Three case studies with the TPAD were performed at a testing speed of 2 mph on US 297 near Wichita Falls at TxDOT FSF on US 290 near Houston The pavements at the three sites are continuously reinforced concrete pavement CRCP jointed concrete pavement JCP and full depth hot mix asphalt HMA respectively The
18. and the final system was added to the TPAD Concurrently rolling noise measurements were performed using an independent towing frame on a jointed concrete pavement JCP at the TxDOT FSF to optimize the wheel diameter and tread width of the 3 wheel sensor cart Earlier in this project wheels with tread widths of 1 2 and 4 in and diameters of 9 5 and 12 5 in were fabricated to allow a variety of three wheel carts for the rolling sensor to be studied These sensor carts were installed on the independent towing frame pulled by hand over a range of speeds The RDD function of the TPAD with the final towing frame and rolling sensor system was tested on three different types of pavements in Texas presented as three case studies in this report The pavements were a JCP at the TxDOT FSF a continuously reinforced concrete pavement CRCP on US 287 near Wichita Falls and a hot mix asphalt HMA pavement on US 290 near Houston At the TxDOT FSF Falling Weight Deflectometer FWD testing was also performed and the results from the TPAD and FWD are compared Personnel from TTI conducted these three case histories with the assistance of CTR personnel The rolling noise measurements and the final design of the towing frame on the TPAD are described in Chapter 2 The three case studies are described as follows TPAD testing near Wichita Falls is presented in Chapter 3 TPAD and FWD tests at the TxDOT FSF are presented in Chapter 4 and the testing results near
19. approximate 1 3 and 5 mph respectively Rolling noise levels were recorded for all wheel sets for each of these speeds The results of the rolling noise level measurements in mV at the three testing speeds are shown in Figures 2 3 2 4 2 5 and 2 6 The figures show the mean and plus and minus one standard deviation of the rolling noise level for the sensors As the testing speed increases with a given wheel set the rolling noise level increases as expected In general 12 5 in diameter wheels SC and SD showed better performances than 9 5 1n diameter wheels SA and SB The results from narrow SA and SC and wide SB and SD showed similar results 35 30 9A 9 5 in Diameter two 1 in and one 2 in Wide Tread 25 20 15 10 Noise Level mV Approximate Speed mph Figure 2 3 Rolling Noise Levels Averaged over 125 ft Long Testing Path at Three Different Testing Speeds Measured with Wheel Set SA SB 9 5 in Diameter two 2 in and 40 30 20 Noise Level mV 10 Approximate Speed mph Figure 2 4 Rolling Noise Levels Averaged over 125 ft Long Testing Path at Three Different Testing Speeds Measured with Wheel Set SB Noise Level mV Approximate Speed mph Figure 2 5 Rolling Noise Levels Averaged over 125 ft Long Testing Path at Three Different Testing Speeds Measured with Wheel Set SC 35 30 SD 12 5 in Diameter two 2 in 25 20 15 10 Noise Level mV Approximate Speed
20. as device had to be lifted over failures flat line Im N 33 4528141 Not on patch localized high deflection no 3106 W 97 7159548 apparent cause check for close crack spacing I m m l l 0 2m N 33 4612626 RDD had to be lifted over failure at each end of 1321 W 97 7818424 pavement no patch 2 3 4m N 33 4919789 Partial Patch with failure at leave side RDD 3366 W 97 8020680 had to be lifted 4 Examples of the GPR data collected on this project are shown in Figure 3 7 The upper collected closer to the middle of the lane than to the open joint between the CRCP lane and color at the mid slab depth is at the level of the rebar The deflection pattern should be Im N 33 4537038 347 W 97 7765700 Multiple peaks failure at end of patch om 4 781 N 33 4599434 1 RDD had to be lifted over failure at end of W 97 7809276 patch 4m N 33 4881118 High deflections but no clear cause from 1651 W 97 7991886 4m N 33 4881118 Three peaks around patch not sure if this 1s an 1816 W 97 7991886 issue 1 4m N 33 4928505 1 Multiple patches back to back check joint 3666 W 97 8027162 between patches 3 4 GPR Pavement Temperature and GPS Data Collected During TPAD Profiling plot Figure 3 7a is from close to the problem area described above and shown in Figures 3 2 and 3 3 From the GPR data no problems were detected at the top of the base as the data were HMA shoulder Different conclusions would possibly have been found if the GPR had
21. been performed close to the pavement edge In Figure 3 7a the strong and continuous reflection red contrasted with that from the solid area shown at the bottom of Figure 3 7 No verification coring was performed at this location but this strong reflection could be associated with issues at mid 30 depth in the concrete possibly midslab delamination where stresses tend to concentrate at the rebar concrete interface surface Rebar Top base Figure 3 7 GPR Data a Upper Plot is from a Problem Area Around 9000 feet and b Lower Plot from a Sound Area The TTI data collection package also monitors the surface temperature during the test as well as project elevation GPS coordinates and the speed of the TPAD Figure 3 8 shows the surface temperature upper plot and elevation lower plot measurements from the GPS The spikes in the data were caused by travelling under bridges along the project where a the concrete temperatures are significantly cooler in the shade and b the system temporality lost the GPS signal Pavement Surface Temperature in F RDD test track Level chart Y Level in feet obove the sea IF Elevation 3 5 Conclusions from US 287 TPAD Testing l The TPAD loading and automotive systems and data acquisition system worked well at the test speed of 2 mph All sensors were operational The only issue encountered was crossing failed areas similar to Figure 3 6 where the load sensors wer
22. code Er Le areas Las OUTDO eeh 112041 namg ar TOT FGF 20520522 TRAD TTT FEF TTI DAC Rare Den taph Kad 1 C VG DAT 131 val ivtra 890031 02042 Figure 5 Screenshot of IGOR When AnalysisProc Is Clicked dde Te Des o d bie eee M n Iesie rr anta ze ER Eder Wet Se Usar equa Law ogni Lon RTS Fete 11112908 TPA TPAD prit ad TADOT FSF 261284572 TRAD Test FSF TIA Rae Das o Eent 1C Agen aep ber sa wn ani 1 COAT 712 total bytes m i e MIA EY Et deal deurt E AL ett Figure 6 Screenshot of IGOR When New Table Is Clicked 73 N Zeep Pr eg ees TD kees IY Figure 7 Screenshot of IGOR When New Table Is Made Distance Based Moving Average Analysis After IGOR finishes the time based analysis Matlab performs the distance based moving average analysis as follows 1 The time based analyzed data saved in the text file F1 txt needs to be imported into Matlab Click the Load data file button to import the text file FI as shown in Figure 8 By importing the FI txt file the F1 table consisting four columns is created as shown in Figure 9 The first second third and fourth columns of F1 are time based analyses of the center front and rear rolling sensors and distance measurements respectively 2 Open the Analysis m file by clicking Open under the File tab Figure 10 presents a shot of the screen after opening the Analysis m file 3 Note The center rolling ce
23. f 12 end 13 sumet 14 number st 15 avg zerosiengtb dun 1 1b numnberofdata length dest 17 18 Iim 5vanable 19 Le J20 I2 whe I2 lt numberofdata 2 7 of stl lt cenbel 23 stim som ampiT2 24 mznber number 1 d avg 11 sum number 26 I2 I2 1 27 ere dit T2 gt cre 28 mm 6 3 A H Ki idm Dea ap Window lei Dg i58 609 RS P ose c enara regen Law DIS Henger ch 1 1 12081 PDC Lanes qr arsi tate Tesoro Fem Maius V Ire im hracia El imee b di E g f P Jus par LL n Hugues t B Reg i Amn Bee Iess Hesg Se VE ST SS L a E EE E 4 Figure 12 Screenshot of Matlab after Clicking the Save and Run Button TI 6 In the spreadsheet create two final deflection and distance columns columns G and H adding the column headings All Deflection and Distance Columns G and H will be populated with data from Columns A through F First copy the data from Columns A and B and paste under the corresponding headings of G and H Then copy the data from Columns C and D and paste directly under the pasted data from A and B Under that paste data copied from Columns E and F Lastly select Columns G and H and sort by the Distance column to arrive at the final distance based moving average data The plot can be drawn in the Excel spreadsheet Figure 13 presents the Excel spreadsheet with one example of a
24. f x mn i az la 3 TI de Lg II FA 3 x An am B nz Lan IS Enem 3 ii E Ha EE ns Le Up t3 38 dg i HE it 2 degt jii 2 2 i13 Ta n M 4 ki Sheer hast Seti amp EE EIOS d tar ere Amm sta a Figure 13 Excel Spreadsheet after Sorting and Drawing Deflection Profile 78 Reference Bay J A and Stokoe II K H 1998 Development of a Rolling Dynamic Deflectometer for Continuous Deflection Testing of Pavements Publication Report No FHWA TX 99 1422 3F FHWA Texas Department of Transportation Center for Transportation Research 79
25. is somewhat similar to the reporting window used in CTR s earlier data analysis software 25 ROD Continuous Deflection Profiles Chart in mils 10 kips iv AC 12000 L n ice Eco Figure 3 2 Deflection Profile in mils 10 kips vs Distance for the Worst Areas on US 287 It became clear in reviewing the deflection data that some of the high deflection peaks were found at the ends of recently patched areas The photograph from the location with the highest deflection in Figure 3 2 is shown in Figure 3 3 The spikes in defection were measured at either end of the patch shown These high deflections were a real concern as bonded concrete overlays will perform well as long as they have uniform support In other locations the concrete next to the patched area has started to fail 26 Figure 3 3 Area of Highest Deflection on US 287 3 3 Poor Pavement Conditions Identified in RDD Deflection Profile Based on the results shown above the TTI team with the help of the project director Joe Leidy of the Construction Division reviewed all of the localized high deflection locations from the complete data set Two examples of typical data are shown in Figures 3 4 and 3 5 Figure 3 4 is the classic case of high deflections at the end of patches The plot below the main deflection plot Figure 3 4a shows 100 ft of deflections around the location selected by the user and marked with the vertical red line in Figure 3 4a High deflections were found a
26. of larger diameter wheels 12 5 in resulted in lower rolling noise but little difference was found between wide one 4 in and two 2 in and narrow one 2 in and two l in tread widths through the rolling noise measurements Concurrently the final towing frame used to position and tow the rolling sensors was designed fabricated and installed by CEM personnel The transmission of vibrations from the TPAD dynamic loading system to the rolling sensors was effectively prevented with the towing frame system The original plan for the towing frame system was to have it designed so that three 12 5 in diameter wheels and 90 Ib of hold down weight could be used by all three rolling sensor carts However 9 5 in diameter wheels and a 40 lb hold down weight had to be used for the center rolling sensor due to space limitations between the loading rollers on the TPAD mobile platform This wheels weight setup was thought to be acceptable because the center rolling sensor generally has a highest SNR because of the proximity of the sensor to the loading rollers Initial deflection profiling to check the performance of the rolling sensors with the final towing frame was conducted on Testing Path E at the TxDOT FSF see Figure 2 16 In this testing the deflection profile performed at a testing speed of 0 5 mph was used as a reference and the deflection profiles determined at 1 2 and 3 mph were compared for all three rolling sensors the center rear and front
27. sa Lcd a id d Figure 2 Screenshot of IGOR When Opened Figure 3 Screenshot of IGOR to Load Raw Data 2 After choosing the data type in the Input File dialog box see Figure 4 the path to the raw data needs to be determined To determine the path click the File button and select a general binary file to load After loading the two types of raw data File name A DAT and File name C DAT separately click the the Do It button 71 iri Fla eme E genee Bee ba dp g dei vm Mbauum of mau m Bhai cd ped em tdem Fe mo amj Lm Dad at uu Fe we n T e 0 0m 0 Dom rus arm rr Bl na TN va vii Den lagh HETI egen reel GEL pp ik egcTett eet eer keenu Les brau Lee C1132031 React TPAD RDO beet ad TO FSE 20895405 eo DNK FSF Hp Caia 5mph 512 1 ae Gansta teuary fe ka born a ter 512 1 dot TOO itai been haie ait ace wb era ka gi S SE Figure 4 Screenshot of IGOR When Load General Binary Data Is Clicked 3 After loading the two types of raw data click analysisProc under the Macros tab the screen shown in Figure 5 will appear Testing parameters such as sampling frequency operating frequency nominal peak to peak dynamic force and time resolution can be selected Among parameters the sampling frequency should be the same as the frequency used while collecting the raw data default is 1000 Hz The time resolution varies with the testing speed In general the time resolutions of 1 0 5
28. the user To convert simply click the UT button the open file dialog box will appear as in Figure A 24 same as Figure A 9 After clicking the Open button the conversion process will 67 start Based on the size of the file it takes a while for converting the file When the toolbar button begins to respond again to the mouse actions the conversion process is finished As a check the user can browse the raw data folder and find the two new files for the UT Austin analysis system For example if the user selects USSINBTB RDD file as the file to convert as in Figure 2 23 then the two new files names are the following e US8INBTB_A DAT is the sensors and DMI raw data file 32bit floating point data format e US8INBTB_C DAT is the DMI processed data format two columns time and distance in feet in 64 bits EX Open Binary EDD test data file AOD Lack jb RDDUS TET sl e Ier Es Name e Date maddied Ji Fw 5 5 2002 430 PM dE Tom 4 26 2012 6 25 AM LJ VSSINBTA EDD 4 25 2012 12557 PM LL USSINBTE RDO A 25 2017 149 EM US amp LSETA SD 4 25 2012949 AM Pepe USEINETERDD L gem Pen g ipe RDD el Cancel Figure A 24 Open File Dialog Box 68 Appendix B CTR Post Processing Software for Calculating Peak to Peak Dynamic Deflections from the Voltage Signals Generated by the RDD Rolling Sensors Kenneth H Stokoe II and Jung Su Lee Introduction to the Software As part of the development of the Tot
29. we noticed four locations where the pavement surface temperature is low compared with other locations and the temperature drops nearly 25 F These four locations are marked with red circles When we looked at the video frames at these locations we find that the temperature drops occur when the TPAD passes under a bridge Figure A 23 shows the video frames for the four low surface temperature locations 64 The middle chart of Figure 2 21 shows the elevation above sea level in feet Since the elevation data is obtained from the GPS system when the TPAD passed under the bridge the GPS had poor signal reception Compared with the top temperature chart we noticed that when the temperature drops the GPS reading under the bridge has an error and the elevation data point fluctuates much more than at other locations The bottom chart just shows the TPAD running speed with DMI offsets We can see the TPAD s speed was close to 2 0 mph and this reading matches the specified test speed of 2 mph Figure A 20 Infrared Sensor for Getting the Surface Temperature Figure A 21 GPS Antenna Mounted at the Top of the Loading Frame 65 Favement Surface Temperature in F doner agga e Liu r000 Tee Tames P ODD TES 202058 22900 24000 20000 26000 RDD test track Level chart Elevation above Sea Level H SCH angu eun bong TUDOD 120 Eed egug TES ZO uuu Anu zmonu RDD Running speed chart zuuun D 200 anno poop p Tana 1708 MN mans 1
30. whereas those with poor LTE have a distorted pattern where the deflections drop and then rise again as the load moves to the second slab a H a m m K Figure 4 5 Deflection Patterns Collected with the Front Sensor Transitions We _ 33 11 35 Figure 4 6 Blow Up of the Front Sensor Deflections with Several of the Joints Tested with the FWD 27 The next step in the analysis 1s to generate a statistic that captures the non uniformity of the deflection as the sensor and load passes over the joint Many parameters can be generated based on the shape of the deflections Figure 4 7 shows the patterns over two adjacent joints The joint in the middle of the plot has the characteristic shape proposed in Figure 4 4 with a clear large rise 4 72 mils drop 2 56 mils and rise 3 25 mils as the sensor rolls over the joint This pattern indicates possible low LTE The second joint with a deflection of 3 03 mils is completely symmetric indicating possible good LTE 4 72 4 03 Figure 4 7 The Front Sensor Deflection Patterns over Two Joints with Different LTEs The numbers on the plots are the deflections in mils per 10 kips at the maximums and minimums For this analysis a simple parameter was generated as follows TPAD LTE 100 max min max 100 4 72 2 56 4 72 45 7 For the symmetrical case where there is no minimum value the min and max are selected to be the same value
31. 3000 20000 22055 K NN 20000 DEMA Figure A 22 Other Information Charts Surface Temperature Sea Elevation and Running Speed 66 Whemeross the kst ENEG When Cross the 2nd Bridge DM Si oon pm mn Cross the 3rd Bridge DIVI Lade 5i t gr Ce a ml 34 e stl Hl A et Figure A 23 Video Frames Showing the TPAD Cross the Brid A 13 Button for Saving Output of the Normalized Deflections Results If the user wants to import the data into other software such as Excel this button E is used to save the results to an ASCII format The output format has two columns one is the DMI offset and the other 1s the normalized 10 kip deflections The DMI offsets are in feet and the deflections are in mils IlOkips Executing this operation is straightforward and will not be covered further here A 14 Button for Saving Output of the Raw Data to the UT Austin Format To UT Since UT Austin already has a history of evaluating continuous deflection data using their own process this feature allows sharing the data with their system by changing TTI s TPAD data format to the UT Austin format The TTI researcher went to great lengths to fully understand the UT Austin format The main problem is that the UT Austin s data acquisition system outputs the data according to an old Apple Mac format The sequence for saving the raw sensor and DMI data is reversed from the normal Windows format Now however the process has been made simple for
32. Austin 1616 Guadalupe St Suite 4 202 Austin TX 78701 www utexas edu research ctr Copyright c 2013 Center for Transportation Research The University of Texas at Austin All rights reserved Printed in the United States of America Disclaimers Author s Disclaimer The contents of this report reflect the views of the authors who are responsible for the facts and the accuracy of the data presented herein The contents do not necessarily reflect the official view or policies of the Federal Highway Administration or the Texas Department of Transportation TxDOT This report does not constitute a standard specification or regulation Patent Disclaimer There was no invention or discovery conceived or first actually reduced to practice in the course of or under this contract including any art method process machine manufacture design or composition of matter or any new useful improvement thereof or any variety of plant which is or may be patentable under the patent laws of the United States of America or any foreign country Engineering Disclaimer NOT INTENDED FOR CONSTRUCTION BIDDING OR PERMIT PURPOSES Project Engineer Kenneth H Stokoe II Professional Engineer License State and Number Texas No 49095 P E Designation Research Supervisor Acknowledgments The research team wishes to express its sincere gratitude to TxDOT for supporting this research project The interaction and support from the Research Monitoring Comm
33. C Dynamic peak peak Load 5 16 fedt Da Load kips Static Load chat hart of ADD Static Load in Kb 5 J fedt Stat Load kips Dynamic Geaphone chart Lief of RDD Dynamic peak peak Delle 0 1 fedt Deflection mils Final TPAD analysis result RDD Continuous Deflection Profiles Che 0 15 fedt Deflection mils 10 kip Figure A 6 Output Chart Format Parameters 55 A 4 Load the Raw Data Button 4 This button is used for loading the TPAD raw test data file collected using the TTI data acquisition system When the user clicks this button the open file dialog box shown in Figure A 9 appears Open Binary RDD test data file RDD Look in BN Test Data fe Ex Marne Date modified Ty USSLSBTB RDD 4 25 2012 1 26 PM RI 4 m F Files of type RDD Cancel Figure A 9 Open File Dialog Box The TPAD raw field data file is given the extension name of RDD this file is in binary format and cannot be edited or viewed with any other software In the background the software actually reads three files the RDD geophone file the video file and the GPS file After the user selects the raw data file from the dialog box shown in Figure A 9 then clicks the Open button data processing will begin A graphic is automatically displayed as a plot of sensor deflection in mils versus distance in feet Figure A 10 The blue line is the raw deflection data and the red line is the smoothed data The red line can be removed by putting
34. DD Deflection Profiling at the TxDOT FSF As with all TPAD testing deflections were measured with the three rolling sensors The center sensor CS is located between the loading rollers with the front and rear sensors approximately 2 feet before and after the loading rollers see Figure 2 13 The deflection pattern from the center sensor for this site is shown in Figure 4 1 The first seven joints are known to be on longer thicker slabs 16 in thick slabs Joint 8 1s a transition joint to a thinner set of slabs 8 to 10 in thick slabs The deflections are substantially lower in the first few slabs both the center slab and joint deflections are low UH Si DL i d oF se SI SP RARe B UI Lat F788072031308510t TEQTUSGASTIA DAUCISESOB Value TOBAGA Expanded Profiles in b and c d Figure 4 1 Typical TPAD Deflection Profile for the JCP at the TxDOT FSF During this testing TPAD data were collected at several different speeds The deflection profiles are shown at speeds of 1 2 and 3 mph in Figure 4 2 These profiles are discussed in Section 2 4 2 by the CTR team From these data and other data sets collected in Project 6005 the 33 TPAD gave stable readings up to 2 miles per hour but above 2 miles per hour significant additional noise was introduced into the signals especially in the front sensor as discussed in Section 2 4 2 a i i WRITTEN E Ly i i VN VUN UVU WY M A UU vuU LE
35. Figure 5 3 The GPR data from the embankment with increasing deflections is shown below The red line at the top of the figure is the surface of the pavement The lower red line is 4 from the interface at the bottom of the HMA and top of the stabilized subgrade layer This section has over 20 inches of HMA over a lime treated soil At this location no major deflects were noted in the HMA layer and the thicknesses were reasonably uniform throughout the section The one major difference is the change in intensity of the reflection from the top of the subgrade A strong reflection is marked in Figure 5 3 shown as a strong red line in the areas where the surface deflections increase This strong reflection can only be associated with a change in moisture content of the lime stabilized layers Where the deflections are low the reflections are fainter implying a drier subgrade and better support conditions d Dk e DL a Dg 1 e fe SIR AR f DEM Tea eae H TEE EE P FS BCEE i cam Lat 5530 038 Lot BENZESOTEAT ni 14440 2 vitas BENE Expanded Profiles in b and c a a Figure 5 2 US 290 TPAD Data from a Location of Varying Deflection po o Seed Leg KC i 1 i i L i bal Ad AO CN CH A i MM i bell Figure 5 3 GPR Data Collected by TPAD Showing Variation in Subgrade Moisture Content As with other runs the elevation above sea level and TPAD speed were monitored throughout the test The results are shown i
36. In this case the TPAD LTE statistic would be 0 note many different parameters could be computed to describe the shape of the deflection pattern other parameters should be evaluated For each of the joints tested with the FWD the TPAD LTE was calculated at both temperatures The relationship between the TPAD LTE and the FWD LTEs is shown in Figures 4 8 and 4 9 38 LTE 84F 40 60 80 Load Transfer from FWD Figure 4 9 Comparison of FWD LTE to a Statistic from the Front Sensor of the TPAD at 104 F In both cases the correlation between the FWD and TPAD evaluation of LTE were reasonable Recall that the LTE is the critical parameter in making overlay design decisions on all pavement types it is required input to TxDOT s new ME Overlay design program The TPAD even at 2 mph will be approximately 10 times faster than the stop and go FWD This potential should be built upon in future TPAD implementation efforts 4 4 GPR Data Collected at the TxDOT FSF GPR data was also collected on the runs at the airport see Figure 4 10 39 i i i I i 62 125 188 251 314 378 441 504 587 e29 Feet Figure 4 10 GPR Data from the TxDOT FSF GPR signals are more highly attenuated by rigid pavement slabs compared to flexible pavement systems The thicker the slab the more the signal will be attenuated At the TxDOT FSF there is a noticeable improvement in data quality after the transition to the thin slabs In the GPR data the slab bottom
37. PR collection rate is one trace per foot collected in the distance mode GPS location is collected once per second as is the infrared temperature data Table A 1 File Size for the US287 Field Testing E Comment Speed Length GPR file Moie SH m name file file file mph miles Kbytes Kbytes Kbytes Kbytes SBTA 51400 491700 76000 49 From this table we noticed that based on the quoted settings the video file size is more than 10 times the GPR data file size and for a 2 0 mph running speed the RDD data file size is around 1 5 times the size of the GPR data file Compared to these files the GPS file is much smaller Since the TPAD geophone data are processed in the frequency domain we cannot collect the deflection data in the distance mode So the TPAD data file size is only dependent on the sampling rate user defined used for collecting data In contrast the GPR and video data are collected in the distance mode so the file size is only tied to the tested distance and trace collection interval From Table A 1 we noticed that for a 5 mile data set the file size 1s quite large Just one geophone sensor provides around 9 267 000 data points The current data processing software TPADana processes and displays the deflection video GPS and temperature data Currently the GPR and video data are processed with the Pavecheck software Future versions of this software will integrate the evaluation of both data sets
38. TPAD data the TTI team and Joe Leidy generated Table 3 1 listing locations that should be inspected prior to placing the concrete overlay These locations were thought to be areas where reflection cracks could potentially appear in the concrete overlay A total of 14 locations were identified for inspection The Priority Rating PR column in Table 3 1 is the priority researchers assigned to each problem area Additional verification testing was recommended with the FWD The District staff needs to inspect each area to determine what if any maintenance action should be taken prior to placing the bonded concrete overlay 29 Table 3 1 Areas Needing to be Inspected Prior to Placing the Overlay Distance GPS from Comments Coordinates start Om N 33 4400731 i i 2671 W 97 7644745 Patch check either side with FWD Om N 33 4408793 1 High deflection only on 1 side of patch Test 3041 W 97 7652607 with FWD 3 Om N 33 4413374 1 Two adjacent small patches high localized 3246 W 97 7657174 deflections Patch within a patch N 33 4464208 Partial lane width patch clear high deflections 4 Im 351 W 97 7707743 l at either end of patch good one to test with l FWD to confirm location of high deflections 5 Im 4 786 N 33 4473576 Deflection pattern not as clear many peaks T W 97 7716964 close together End of patch looks to have problems Major failures at either end of patch RDD Im N 33 4483291 1221 W 97 7725820 1 pattern unclear
39. TPAD proved durable during the testing period and provided valuable information for all three types of pavements 17 Key Words 18 Distribution Statement Continuous Deflection Profiling Testing Speed GPR No restrictions This document is available to the Imaging GPS Measurements Video Imaging public through the National Technical Information Pavement Surface Temperature Measurements Service Springfield Virginia 22161 www ntis gov Distance Measurements and Single Moving Platform 19 Security Classif of report 20 Security Classif of this page 21 No of pages 22 Price Unclassified Unclassified 94 Form DOT F 1700 7 8 72 Reproduction of completed page authorized a lexas 48 Transportation a institute Developing a Testing Device for Total Pavements Acceptance Final Report CTR Kenneth H Stokoe II Jung Su Lee Mike Lewis Richard Hayes TTI Thomas Scullion Wenting Liu CTR Technical Report 0 6005 3 Report Date November 2012 Project 0 6005 Project Title Developing a Testing Device for Total Pavement Acceptance Sponsoring Agency Texas Department of Transportation Performing Agency Center for Transportation Research at The University of Texas at Austin Texas A amp M Transportation Institute Project performed in cooperation with the Texas Department of Transportation and the Federal Highway Administration Center for Transportation Research The University of Texas at
40. Technical Report Documentation Page 1 Report No 2 Government 3 Recipient s Catalog No FHWA TX 13 0 6005 3 Accession No 4 Title and Subtitle 5 Report Date Developing a Testing Device for Total Pavements November 2012 Revised April 2013 Acceptance Final Report Published June 2013 6 Performing Organization Code 7 Author s 8 Performing Organization Report No Kenneth H Stokoe II Jung Su Lee Mike Lewis Richard 0 6005 3 Hayes CTR Thomas Scullion and Wenting Liu TTI 9 Performing Organization Name and Address 10 Work Unit No TRAIS Center for Transportation Research 11 Contract or Grant No The University of Texas at Austin 0 6005 1616 Guadalupe St Suite 4 202 Austin TX 78701 Texas A amp M Transportation Institute The Texas A amp M University System 3135 TAMU College Station Texas 77843 3135 Sponsoring Agency Name and Address 13 Type of Report and Period Covered Texas Department of Transportation Technical Report Research and Technology Implementation Office 9 1 2008 8 3 1 2012 P O Box 5080 Austin TX 78763 5080 14 Sponsoring Agency Code Supplementary Notes Project performed in cooperation with the Texas Department of Transportation and the Federal Highway Administration Abstract During the fourth year of Project 0 6005 construction of the Total Pavement Acceptance Device TPAD was completed The TPAD is a multi function pavement evaluation device used to profile continuously
41. Towing Frame Arm A a vue te Figure 2 14 Photograph of the RDD Portion of the TPAD Showing the Towing Frame Front Rolling Sensor Location of Center Rolling Sensor and One Loading Roller 2 4 TPAD Testing with Towing Frame System at TxDOT FSF After installation of the towing frame with the three RDD rolling sensors the TPAD was brought to TxDOT FSF and the performance of the RDD rolling sensors with towing frame was evaluated 2 4 1 Rolling Sensor In Situ Calibration As part of work conducted on earlier RDD projects CTR personnel have developed a procedure for in situ field calibration of the rolling sensors This in situ calibration was performed on the pavement at the TxDOT FSF This calibration is done using two laboratory calibrated 4 5 Hz geophones borrowed from the Soil Dynamics Laboratory at UT The reference transducers are used to measure the motion on the pavement surface while the TPAD loading system is applying both static and dynamic forces to the pavement The rolling sensor and reference transducers set up in the field calibration procedure are shown in Figure 2 15 The in situ calibration is performed by changing the excitation loading frequency sweeping of the loading rollers on the pavement typically between 20 to 50 Hz while the TPAD is stationary Usually uncracked mid slab areas are selected as a testbed for JCP to reduce the influence from any cracks or joints For this site the mid slab area on an 8 in thic
42. al Pavement Acceptance Device TPAD CTR created post processing software for calculating the dynamic deflections of the pavement from the output of the rolling dynamic deflectometer RDD rolling sensors To analyze the data the software performs these steps 1 load raw voltage signals of the RDD rolling sensors 2 apply the composite infinite impulse response and finite impulse response filters 3 apply field or lab determined calibration factors to calculate peak to peak deflections over a time interval determined by a testing speed 4 normalize the dynamic deflections to a force level selected by the data processor with dynamic force time data normally a peak to peak force level of 10 kips 5 average the converted rolling sensor deflections over a selected distance and 6 apply a moving average Figure presents a flow chart of the post processing software Note that running this software requires the MS Windows operating system Windows XP Windows Vista or Windows 7 and software packages IGOR PRO and Matlab The CD accompanying this manual contains the CTR software to load onto the user s computer TPAD ana_ver 2 0_TxDOT pxp and Analysis m 69 Desrnodukalson force the dspacement or force signal x cos Zi Filter with i an2nf t ic Decimation Zeck dala over cdebennined fine jenn fir scing bc tt m Credo ih a xen D ou b m amp i d H UM OUS us UV
43. and 0 33 seconds are used respectively at testing speeds of 1 2 and 3 mph default is 0 5 seconds Normally an operating frequency of 30 Hz vibrating frequency a normalization peak to peak dynamic force of 10 kips and a minimum dynamic force of 2 kips are used and designated as default values The attenuator is installed to avoid sending voltages that are too large for the data acquisition system and the default is in After setting all parameters click the Continue button 4 The data analyzed in IGOR needs to be imported as a txt file for distance based moving average analysis Click New Table under the Windows tab and choose danf1 danf2 danf3 and doutl They are measurements for center danfl front danf2 and rear danf3 rolling sensors and the distance dout1 measurement respectively Figures 6 and 7 present screen shots of when New Table is clicked and a new table is made The table can be copied and pasted in a text file by using keyboard shortcuts Ctrl A Ctrl C and Ctrl V The name of the text file should be F1 as designated by an operator so that the text file can be imported to Matlab and be used in Analysis m 12 at D i 1 5 Ek ee Ce hPa CIL cae a ae C Users Aega Les ge Lee SRS Ane 11112010 TRAD TRAD teeing ai OT PSP 206905727 TAD Tet FSP TD GAS Rew Den mph Kd 1 A Tetka Fer Ee ad kom 1 ADAT H4 52000 boli ben Daa tength 41713000 wave
44. angular loading frame which is just over 3 in when full load is applied as shown in Figure 2 7 The second constraint is the width space available for the towing frame As seen in Figure 2 8 even though the total width of the rectangular loading frame of the loading rollers is 22 in the width available for the towing frame is just 14 in The third constraint is the vertical and horizontal distances between the loading frame and the TPAD mobile platform when the loading frame is fully lifted The measurements from the rectangular loading frame to the TPAD frame when the loading frame is fully lifted are shown in Figure 2 9 They are about 3 and 8 in respectively abautsin m Figure 2 7 Vertical Space Constraints under the TPAD for Towing Frame 10 Figure 2 8 Lateral Width Constraints under the TPAD for the Towing Frame Figure 2 9 Vertical and Horizontal Distances from the Loading Frame to the TPAD Frame As a trial in designing the towing frame a wooden mock up was built and then test fitted to define the acceptable size of the towing frame After the trial with the mock up the design shape of the towing cart frame was determined The towing frame consists of three connected pieces which are a center bridge rectangular box beam and two end sections bolted to the center bridge beam as shown in Figure 2 10 Vertical channel assemblies are bolted to each section The 11 vertical channels are used to attach the rolling s
45. asily identified 3 The GPR data was useful 1n determining the possible causes of the higher deflections and identifying locations where validation coring or additional testing would be beneficial 46 References Bay J A and Stokoe II K H 1998 Development of a Rolling Dynamic Deflectometer for Continuous Deflection Testing of Pavements Publication Report No FHWA TX 99 1422 3F FHW A Texas Department of Transportation Center for Transportation Research Stokoe K H Kallivokas L F Nam B H Carpenter C K Bryant A D Weeks D A Beno J H Scullion T and Liu W 2010 Progress during the First Year towards Building the Total Pavement Acceptance Device TPAD CTR Technical Report 0 6005 1 Center for Transportation Research University of Texas at Austin Stokoe K H Lee J S Nam B H Lewis M Hayes R Scullion T and Liu W 2011 Developing a Testing Device for Total Pavements Acceptance Third Year Report CTR Technical Report 0 6005 4 Center for Transportation Research University of Texas at Austin 47 48 Appendix A TPADana 1 0 User s Manual for the TPAD Data Analysis Software Dr Wenting Liu A 1 Summary The TPAD is a continuous pavement deflection testing device enhanced with several additional survey features Since the device is designed for total acceptance of pavements we have combined the deflection testing with Ground Penetrating Radar GPR dig
46. ation The last button is used to transfer the raw TTI formatted data to the raw UT Austin format allowing the UT researchers to analyze the data with their existing software In the following sections we will discuss the function of each button in detail A 3 Program Options and Configuration Button de Click this button and the screen shown in Figure A 2 will be displayed this shows the options available to process and display the TPAD data TPAD analysis options Signal Analysis Filter Control Select Geophone Sensor Sensor Calibration Factor Total Data Segments From Ta Load Factor H Hz Geophone 1 Dynamic analysis Bandpass filter 23 E Center Geophone H Cert ete Serene Front Geoph Static Load analysis Lowpass filter 2 Hz E 0 73 Rear Geophone m Video offset means video in Chart Cantral front of Geophone feet Title Chart Min Chart Mas Label Y Label Dynamic Load chart Chat of ROD Dynamic peak peak Load B BO fect Dynamic Load kips 575 Static Load chart Chart of ROD Static Load in lb B mm fect Static Load kips Dynamic Geaphone chart Chart of RDD Dynamic Geck Geck Delle mE moo feet Deflection mis Final TPAD analysis result RDD Continuous Deflection Profiles Che bo o 5 ooo fet Deflection mils D kip Smooth D ata e Moving 4verage method 20 CH D wa H WI Average by Number 30 A AN Average by distance ft Cancel Figure A 2 Program Setup or Options Sc
47. ble of Contents Chapter E ini Ce apo ee 1 1 1 Brief Overview or All Components of the TPAJD estote tae eo eee otto pea eee Rer marie l 1 2 INCUVIUICS QUEIDS Year EN 3 Chapter 2 Rolling Noise Measurements Design and Installation of Towing Frame and Field Evaluation of the Complete TPAD RDD System ccssccccssssssssssssssccccseceesees 5 PAX MAMET O GUC OM E 5 2 2 Rolling Noise Measurements with an Independent Towing Frame at TxDOT FSF 5 2 3 Design of Towing Frame by CEM Personnel ient ioo abe cea tos i eerie ode a podran 9 2 4 TPAD Testing with Towing Frame System at TXDOT Ab 15 De SUMMA Siento atem neces ss eae either ad MI ELI MEE A 22 Chapter 3 TPAD Case Study 1 US 287 near Wichita Falls CRCP 25 DL let te EE 23 3 2 TPAD Operating and Data Collection Functions essere 25 3 3 Poor Pavement Conditions Identified in RDD Deflection Profile 27 3 4 GPR Pavement Temperature and GPS Data Collected During TPAD Profiling 30 25 2 Conclusions from US 28 7 TPAD Testing cniinne ienei a Denia EREE 32 Chapter 4 TPAD Case Study 2 TxDOT FSF in Austin JCP csccccccsscccsssssccccccsees 33 Z ol nere E 33 4 RD Detl cuon Pronmhime at the TXDOT ere ee 33 4 3 The TPAD Potential for Evaluation of LTE with Comparison to FWD Doata 34 4 4 GPR Data Collected at the TXDOT FSF wivwisesnssussuscasuavadesaa
48. bt At En meet 16 DA E lor ei berger Air cela WH mater sf erg zerlegt D feb dels e eth dest Da L press bs CL em pueden dana G e eget 17 pap rant eet Teh Faber mper verge TL anf reer Belt Ho artic rr acubare 3 averaged H EXXH ben brum yg Artragt Lj ig ong ram Shr path Paci al 76 m n m F Dhe piaga me gis Lai ink OUI Pernt iO IT Lae bn dra io ide ai d i a E AA 00 9798N D pelz hes am erage ds bel tho pr s Eu ari eges Loe ge une 87 er ee gem ed tet ree Pf ties Vi Le eee pe Bros CAUsersijungsa LeetJungsu Lee 08172010 Research 11 11201 DEE Labview Igor and Matlab Distance Profile Matlab M fileAnslysis m lex SI wma n qs E ipm m eee n ES Eda Text Cell Tools Debug Desktop Window Help x Og a LB er d ASF AD Beal GE Steck Ir EIIBesD i sa a Veliobe this m hie is to determine an averaged def echon prehle P 2 Verepeated EDD tering along same wheel path Pancally the deflerton 3 peche are recalculated averaged over 38 to obtam the same distance 4 Vurtereal not time mterval bw ft 3 d 5 belit step import data tyt Eelanalyzed by Igor save and Run E And step run the Matlab Button 8 dest A 95E Tis input data from Igor t fle Rename for ferent Re 9 mwnpsFl 1y oF c put data from Igor tt fe Rename For diferent Bes 10 fer rellengthidist 11 entice yet DH D Se Average Date over every 3 f and Shalt 00
49. center sensor CS which is between the loading rollers will be severely impacted by the presence of the joint However the center sensor measures the overall pavement response to load if there is a void beneath the slab then the total deflection measured at the center sensor will be high TRANSFER Front Sensor FS Center Sensor CS LEI A _ RearSensor RS i A B C D E Figure 4 4 Schematic of the Proposed Change in Sensor Deflections over a Poor Joint The next step is to review the shape of the actual deflections observed in the front sensor for the joints with known poor LTE Figure 4 5 shows the normalized deflection pattern for the front sensor at the 84 F test temperature Three of the joints are labeled Joints 29 32 and 33 36 for which the measured LTE were 21 100 and 6 The center plot shows a close up of the deflections measured over these joints Joints 29 and 33 have a pattern very similar to the pattern shown in the front sensor plot in Figure 4 4 with a significant drop in deflection whereas joint 32 has a very symmetric deflection pattern with very little change as it rolls over the joint The deflection patterns for the front sensor are also shown in Figure 4 6 where several of the joints tested with the FWD are also marked There is evidence of the relationship between shape of the deflection pattern and measured LTE The joints with close to 100 LTE have very symmetric deflections
50. ctly with sufficient ruggedness for longer projects and to determine if useful information could be obtained for the District staff with regard to the upcoming construction project Figure 3 1 shows the start location of testing performed in the southbound outside lane mm M E lt p pur A T i Figure 3 1 US 287 US 81 Start Location 3 2 TPAD Operating and Data Collection Functions No major problems were found with the TPAD operating and data collection functions At a few locations where severe faulted punch outs were found the RDD geophone carriage was stopped and lifted over the faulted areas to avoid the risk of damaging the rolling sensors which are not designed to clear vertical offsets of more than about 0 5 in A total of 26 800 feet of continuous deflection data was collected The results below were obtained from TTI s TPADana software package which is described in Appendix A The deflection data from the worst areas 1n terms of both condition and high deflections are shown in Figure 3 2 This figure presents a plot of deflections normalized to a 10 kip load level versus distance along the pavement in feet The worst area was found to exist between 9000 and 10000 ft from the start of the data collection In Figure 3 2 the blue line represents a continuous deflection profile based on every cycle while the red line shows the average deflection collected for every 2 feet of pavement tested The 2 ft average
51. deflection profile 7 For a front rolling sensor type these specifications into the Analysis m box distzF1 4 amp F1 2 Then repeat steps 3 4 5 and 6 For a rear rolling sensor type distzF1 4 ampzF1 3 and repeat steps 3 4 5 and 6 Mesaebt Leni Bande EEUU Mee ee ems Ze fe Mem Bn i o X d i 7 ech CR ub i E Gl i x base sl 3955 T 2 D RII j mL l I Ein D zn E i J KZ i LSTA IE H 1 LHN ra z2zml TH H io Ar a zm 2 7 5 il TOM i4 gt r3 i m H BE e EE SC z ri 3 E 13 gang EN mi LE id S ep 3 Step 4i T E 8 ID a WT F I j H 3 rmm n r5 i LER H H E I E EEA U Le o am T5 d B A 1i RES Ok DIN di aH Ti TES PN n 19 Lem 23 Tag o Lm CL re Hj a 20 EE 5 zm Ka 2g 53 gba A el LI H LTIS Kei 2288 K art a 2 en 5 LITE W R75 3 rm d n Lei gu Lens H 278 EZ FHE a F zd TE J elt ld a Lig L CR ji I 25 KEI 6 2 p 2298 pj EEEN zi n o ao un TY ech que Cem tat O 100 200 400 500 EX o0 at Ch 1 AIT I Lg d eer Distance It m zer ED E Leen D nore 25 zu Les d EmA 9 TEJ Cr Ds d i 3X z ch Bl g rel E MI Dr m 2 KI LETE 8a 2 085 5 L Bi me 28 x L Si Act B Lem Ce wor 28 D LE KH Imi Fi Lau ES diti X CH panim 33 1 5156 5d Lang 7 wF 3 Lem 16379 9 Lee E i TEA 2 j Jh La EI L5 D I 38 191 E IJ Hn Cl xm I2 Lael B e d i Ma 54 LS un LE t ii i a
52. different from that observed at other frequencies background noise The power spectrum shows the strength of the variations energy as a function of frequency It shows at which frequencies variations are strong and at which frequencies variations are weak The PSD units are energy per frequency width and it is possible to obtain energy within a specific frequency range by integrating PSD within that frequency range Computation of the PSD is done directly by the method called Fast Fourier Transform FFT or by computing the autocorrelation function and then transforming it When the user clicks the button MI the PSD chart of raw load time series will show on the screen as in Figure A 14 The analysis is currently performed for each set of 5000 data points this segmenting option is under review at this moment Figure A 14 shows data points from 65 000 to 70 000 accordingly the DMI is from 126 49 feet to 141 09 feet The chart s Y axis 1s 58 in log10 units and the X axis is in frequency This chart shows the dynamic load frequency is exactly 30 Hz Buttons a and P can be used to turn to other pages Each page covers 5000 data points Load Power Spectrum from point 65000 to 70000 or DMI from 126 49 ft to 141 09 ft Power Spectrum in Loa frequency Hz Figure A 14 Power Spectrum Chart of Raw Load Signal A 8 Button for Showing the Power Spectrum Chart of Raw Geophone Data a cp Similar to the button discussed in Section A
53. e lifted to avoid the risk of damage Very limited continuous deflection data has previously been collected on CRCP pavements but the data generated on this project point out some important issues with the pavement structure The data highlighted problems with the current full depth patching operation that the District has been using This should be reviewed by the District staff and Construction Division personnel Patching CRCP is extremely difficult and there is always a risk of damaging the adjoining concrete The data collected was used to generate a list of 14 locations with their GPS coordinates that the District should inspect to determine if additional repairs are required prior to placing the concrete overlay IPAD testing of upcoming CRCP rehabilitation projects should be encouraged 32 Chapter 4 TPAD Case Study 2 TxDOT FSF in Austin JCP 4 1 Introduction TxDOT s FSF in Austin is an ideal test section for evaluating TPAD s capabilities on JCP Previous testing at this site has shown a wide range in load transfer efficiencies LTE for the various joints along Testing Path E shown in Figure 2 16 Testing with both the TPAD and FWD was conducted at two time periods during one day early morning and close to mid day to evaluate the effect of temperature on the LTEs Testing was conducted with the TPAD at various speeds The results below are based on TPAD data collected at a profiling speed of 2 mph 4 2 R
54. eak 3 mph 9 38AM 72F Dynamic Load z kips f a V d n M n UU i Di Deflection mils 10kips D gt 0 100 200 300 400 500 600 700 Distance ft Figure 2 26 Continuous Deflection Profiles at 0 5 and 3 mph Using the Rear Rolling Sensor in the Towing Frame along Path E at TxDOT FSF 2 5 Summary Rolling noise measurements were performed with the rolling sensors using an independent towing frame that was pulled by hand hence no additional noise created by the TPAD or loading rollers These measurements were performed on JCP at the TxDOT FSF In general the rolling sensors with larger diameter wheels showed better performance in terms of lower rolling noise while almost no difference was found between wide and narrow treads on 27 the sensor wheels Therefore the researchers elected to use larger diameter wheels 12 5 in compared to 9 5 in wherever space beneath the TPAD allowed A final towing frame for the rolling sensors was designed and fabricated by CEM personnel The towing frame system is isolated from the TPAD mobile platform as much as possible to prevent transmission of vibrations from the dynamic loading system to the rolling sensors Initially the towing frame was intended to incorporate 12 5 in diameter wheels on the rolling sensor carts and a heavier hold down weight 90 Ib However 9 5 in diameter wheels and a 40 Ib hold down weight were u
55. easurements 2 mph 9 30AM 72F Actual Peak to Peak Dynamic Load 8 kips 100 200 300 400 500 600 700 Distance ft Figure 2 22 Continuous Deflection Profiles at 0 5 and 2 mph Using the Front Rolling Sensor in the Towing Frame along Path E at TxDOT FSF 20 CTR Measurements eee 7 Actual Peak to Peak EE Dynamic Load 8 kips Deflection mils 10kips D gt 0 100 200 300 400 500 600 700 Distance ft Figure 2 23 Continuous Deflection Profiles at 0 5 and 3 mph Using the Front Rolling Sensor in the Towing Frame along Path E at TxDOT FSF CTR Measurements RP 7 Actual Peak to Peak 1 mph 9 30AM 71F Dynamic Load 8 kips Deflection mils 10kips D 0 100 200 300 400 500 600 700 Distance ft Figure 2 24 Continuous Deflection Profiles at 0 5 and 1 mph Using the Rear Rolling Sensor in the Towing Frame along Path E at TxDOT FSF 2 CTR Measurements 0 5 mph 9 00AM 70F 7 Actual Peak to Peak 2 mph 9 30AM 72F Dynamic Load z 8 kips Deflection mils 10kips A 0 100 200 300 400 500 600 700 Distance ft Figure 2 25 Continuous Deflection Profiles at 0 5 and 2 mph Using the Rear Rolling Sensor in the Towing Frame along Path E at TxDOT FSF CTR Measurements 0 5 mph 9 00AM 70F 7 Actual Peak to P
56. ensesnwavdsnasudenseetsaveeadaWavasmenuaencanctes 39 E onelusions rom T xDOT E EE 40 Chapter 5 TPAD Case Study 3 US 290 Houston Full Depth Hot Mix Asphalt RE E E 41 GEET 4 5 2 RDD Deflection GPR Elevation and Testing Speed Profiling sssss 4 3 9 C OnCIustons From US 290 T CS Une ecce a esae is nU reto e ea QU do RAE Edu e dedecus ds 43 Chapter 6 Conclusions oe een ee 45 RelIereliCes cicossecessies usto EE 47 Appendix A TPADana 1 0 User s Manual for the TPAD Data Analysis Software 49 Appendix B CTR Post Processing Software for Calculating Peak to Peak Dynamic Deflections from the Voltage Signals Generated by the RDD Rolling Sensors 69 Vil Vill List of Figures Figure 1 1 Photograph of the TPAD Preparing to Evaluate the JCP at TxDOT FSF I Figure 1 2 Cross Sectional View of the TPAD RDD Loading System Which Applies Static and Sinusoidal Dynamic Loads to the Pavement with Two Loading Rollers 2 Figure 1 3 Photograph of Rolling Sensor Composed of a Three Wheel Cart 2 Hz Geophone with 40 Ib Hold Down Wecht nnn 3 Figure 2 1 Rolling Sensor Computer DAQ and Battery Installed on an Independent rowing Frame on JCP at ROTES E aa a 6 Figure 2 2 Four Combinations of Wheels on the Rolling Sensor Installed in the Independent Loving Ge EEN 7 Figure 2 3 Rolling Noise Levels Averaged over 125 ft Long Testing Path at Three Di
57. ensor carts Four 8 in diameter pneumatic wheels are used to support the towing frame when it 1s on the pavement the two front wheels swivel while the two back wheels are rigidly attached to the towing frame The drawing of the towing frame design is shown in Figure 2 10 Two End Sections Center Bridge Beam Vertical Channel Assemblies Figure 2 10 Towing Frame Design Drawing by CEM Personnel Since the rolling sensor carts with 12 in diameter wheels SC and SD and the higher hold down weight generally performed better in the rolling noise measurements it was determined to use larger diameter wheels wherever physical constraints would permit The drawings of one rolling sensor package are shown in Figure 2 11 The drawing of three rolling sensors positioned in the towing frame is shown in Figure 2 12 12 b Front View c Side View Figure 2 11 RDD Rolling Sensor Used for Towing Frame System Figure 2 12 Three RDD Rolling Sensors Attached to the Towing Frame 13 Fabrication of the towing frame attachments of rolling sensors to the frame and installation of the towing frame with three rolling sensors were all completed in January 2012 As mentioned above the TPAD currently has three RDD rolling sensors These sensors are arranged in a linear array as shown in Figure 2 13 The array is oriented along the longitudinal axis of the TPAD and is centered mid way between the loading rollers The sensors are named acc
58. f All Components of the TPAD The objective of Project 0 6005 is to develop the Total Pavement Acceptance Device TPAD The TPAD is a multi function nondestructive testing device that will be used to continuously profile along pavements to assess their structural conditions The multi functions include 1 Rolling Dynamic Deflectometer RDD measurements 2 Ground Penetrating Radar GPR imaging 3 a Distance Measurement Instrument DMI 4 high precision differential GPS measurements 5 pavement surface temperature measurements and 6 digital video imaging of the pavement surface and right of way conditions A photograph of the TPAD is presented in Figure 1 1 m P F er 7247 AA Ew NE C Figure 1 1 Photograph of the TPAD Preparing to Evaluate the JCP at TxDOT FSF The TPAD is a hydraulically operated mobile platform with a total weight of about 18 kips The TPAD is 20 ft long 7 5 ft wide and has a maximum height of 7 8 ft While conducting pavement profiles the TPAD has a control system for profiling at a constant speed within a range of 0 5 to 10 mph The TPAD RDD loading system is capable of generating static forces of 3 4 to 14 kips and dynamic sinusoidal forces with a peak to peak amplitude of 2 to 24 kips over a frequency range of about 7 to 200 Hz Static hold down forces are measured based on hydraulic pressure measurements and dynamic forces are measured with accelerometers installed on the reaction ma
59. fferent Testing Speeds Measured with Wheel Set SA esseeeeeeeeeneeeee 8 Figure 2 4 Rolling Noise Levels Averaged over 125 ft Long Testing Path at Three Different Testing Speeds Measured with Wheel Set SR 8 Figure 2 5 Rolling Noise Levels Averaged over 125 ft Long Testing Path at Three Different Testing Speeds Measured with Wheel Set SC sse 9 Figure 2 6 Rolling Noise Levels Averaged over 125 ft Long Testing Path at Three Different Testing Speeds Measured with Wheel Set Sal 9 Figure 2 7 Vertical Space Constraints under the TPAD for Towing Frame 10 Figure 2 8 Lateral Width Constraints under the TPAD for the Towing Frame 11 Figure 2 9 Vertical and Horizontal Distances from the Loading Frame to the TPAD Eegeregie 11 Figure 2 10 Towing Frame Design Drawing by CEM Berzonnel 12 Figure 2 11 RDD Rolling Sensor Used for Towing Frame System eseeeeeeeesss 13 Figure 2 12 Three RDD Rolling Sensors Attached to the Towing Frame 13 Figure 2 13 Arrangement of Sensor Array with Three RDD Rolling Sensors 14 Figure 2 14 Photograph of the RDD Portion of the TPAD Showing the Towing Frame Front Rolling Sensor Location of Center Rolling Sensor and One Loading Roller 15 Figure 2 15 Set up Used for In Situ Calibration of a Rolling Sensor
60. he user clicks on This view of the data is potentially the most useful to the user Once the mouse is clicked in the top graph two lower graphs are generated that zoom in on the area around the cursor location in the top graph The upper of these graphs shows the deflection plot for a total of 100 62 feet 50 feet either side of the selected location The lower graph zooms in even closer with a deflection plot of approximately 15 ft 7 5 ft either side of the selected location en Haw data file hac JOST SO pele aed 1547 grees Currant read segment bci R B Ub wa Ul w Ul em TZ RAR oe zB up Lat 83 488107510t 97 8053541 DMI T44L7 Value ONES 110p dh Peri k bk i DPHIL ech Mach E S S E S S i i 3 ard top toD Z N d d start TPAC Ru cro z B ee cla Fl Fri untitled Paint EE JS Figure A 16 The Comprehensive Results View Screen View Option I If the user moves the cursor to any of these three charts the DMI offset and the Value field will show the normalized deflection magnitude of the mouse location The overlap buttons on the video frame can be used to control the play mode of the video The function of each button is listed below IE Go to the first frame of video EJ Play the video backwards ES Go to previous video frame E Go to next video frame ES Play the video forwards ESI Go to the last frame of video ISl Save the current frame to a JPEG image file W
61. hen the user plays the video by using button J or Il the vertical red line showing the current location on the normalized deflections trace chart moves along with the video 63 Data A 12 Comprehensive Results View Option 2 View This data view shows the normalized continuous deflections plot in the upper part of the screen and an enlarged video image of the highway surface Figure A 19 Clicking on any location in the deflection plot will display the image from that location i a alb ai Caii cene Lat 388 4957440 Lot 08708051188 DMI ESBSTIB Value EEBEBE Figure A 19 Comprehensive Results View Screen View Option 2 A 12 Buttons for Showing Other Information Tt As shown in Figures A 20 and A 21 the TPAD is equipped with an infrared surface temperature sensor and an accurate GPS antenna mounted on top of the loading frame A GPS allows the user to permanently record the test location in the data file In cases where the user wants to go back to the site to further investigate problematic locations the GPS is the best tool for locating problem areas Also a concrete pavement s behavior under loading especially at Joints and working cracks is related to slab temperature when tested and it s necessary to account for temperature effects when evaluating deflections The button P is used for showing this information By clicking this button the screen shown in Figure A 22 will be shown From this chart
62. igh deflections Currently a default value of 15 75 feet 1s the input value for distance from the loading rollers to the GPR antenna This value may be revised at a later date if an alternate camera aim point is selected With the proper value the program automatically lines up the data This offset is automatically added to each display chart Video offset means video in front af Geophone feet 15 75 Figure A 6 Offset and Moving Average Parameters Input Screen The user is provided with three options to smooth the deflection data Figure A 7 When selected a smoothed data curve is plotted on top of the raw data as a red line as shown in Figure A 9 For the TPAD travelling at 2 mph with a data collection system operating at 1000 Hz using a 30 Hz loading frequency approximately one deflection point is collected for every 1 inch of travel Figure A 7 provides the user with options to down sample and smooth these data plots The down sampled data can also be output to Excel for further data analysis and display Button 19 Smooth Data e Moving Average method 20 f Average by Number 30 Average by distance 1 d Figure A 7 Options to Smooth Raw Deflection Data From the center of the options dialog box Figure A 8 the user can format the output data charts defining the chart title minimum maximum and x y labels of each chart Chart Control Title ChartMin Uert Ma amp Label Y Label Dynamic Load chart Chart of RO
63. ing Sensor in the Towing Frame along Path E at TXDOT Fab 22 Pisure 5 1 0526 7705 Ole Start LOC GEO EE 25 Figure 3 2 Deflection Profile in mils 10 kips vs Distance for the Worst Areas on US pr T 26 Figure 5 5 Area ot Highest Deflection on US 28 epes eege Eee epos Een 27 Figure 3 4 Typical Problems Found at Either End of Concrete Patches 28 Figure 3 5 Area High Deflection Area Not Associated with a Patch ssseeeeeessssesssssseeeeereresses 28 Figure 5 0 Problem Areas on US 287 s osodeeedecterses eccutag qub sua na eee cotes degkeet ege 20 Figure 3 7 GPR Data a Upper Plot is from a Problem Area Around 9000 feet and b Mower PIOLTrOH ac 9 OUDO XE EE 31 Figure 3 8 Additional Data Items Collected in a Typical TPAD Run Temperature and PNG Y AUTOM DEE 31 Figure 4 1 Typical TPAD Deflection Profile for the JCP at the TxDOT FSF 33 Figure 4 2 Repeated Runs of the JCP at Different Speeds with the Center Sensor at EAR AR E 34 Figure 4 5 FWD Set Up for LTE Meas remlernts u ie tete erae e Eta E Pe FEL utet ae te tod EeRE 35 Figure 4 4 Schematic of the Proposed Change in Sensor Deflections over a Poor Joint 36 Figure 4 5 Deflection Patterns Collected with the Front Sensor oJ Figure 4 6 Blow Up of the Front Sensor Deflections with Several of the Joints Tested EE 37 Figure 4 7 The Front Sen
64. ing frame is used to position the rolling sensors and pull them along with the TPAD while the TPAD moves along the pavement The towing frame is designed to minimize the transmission of vibrations from the dynamic loading system through the machine and to the rolling sensors The towing frame has the potential to be modified so that additional sensors can be incorporated as the need arises The other TPAD testing functions GPR GPS video camera and pavement surface temperature sensors are described in the Year 3 report by personnel from the Texas A amp M Transportation Institute TTI who were responsible for developing these systems and integrating all facets of the data collection and processing into one computer system Transducer 2 Hz Geophone Hidden by Cart Frame and Hold Down Weight Figure 1 3 Photograph of Rolling Sensor Composed of a Three Wheel Cart 2 Hz Geophone with 40 lb Hold Down Weight 1 2 Activities during Year 4 In Year 4 a towing frame used to position the rolling sensors on the pavement was designed and installed on the TPAD by personnel from UT s Center for Electromechanics CEM The towing system is configured to hold three rolling sensors Initial prototype tests with the towing frame and three rolling sensors were performed at the Texas Department of Transportation TxDOT Flight Services Facility FSF After the initial tests modifications were made to parts of the towing frame as discussed in Chapter 2
65. ital video and GPS technologies and the final system will be suitable for testing both new pavements in a quality assurance capacity and those scheduled for rehabilitation to determine suitable strategies This user s manual describes the data processing system and uses the data collected on a 2012 survey of US 287 in the Wichita Falls District for illustrative purposes The executable load module for this software and the associated data from US 287 are supplied with this manual The TPADScan data acquisition software developed by TTI to collect TPAD raw data is run on one laptop computer After each field test TPADScan will create four files e GPR raw data also has DMI and GPS information file extension is DAT e RDD raw data also has Infrared temperature and DMI information file extension is RDD e Digital video data file also has DMI data imbedded file extension is IMG e GPS raw data file also has DMI on trace number and time data file extension is GPS The detailed data structure inside each file type is given in the TPAD data acquisition user s manual An example of output file volume based on test survey settings for the US287 field testing 1s given in Table A 1 The basic settings for the test surveys were as follows The deflection data collection rate is 1000 sample s with four channels of data Video collection rate is 5 feet per frame and the video resolution is 1280x800 collected in distance mode G
66. ittee Mr Joe Leidy Mr Ed Oshinski Dr Dar Hao Chen and Dr German Claros has been very important to the success of the project The UT Staff Mr Cecil Hoffpauir Mr Curtis Mullins Mr Andrew Valentine and Dr Farn Yuh Menq helped significantly in the field tests with the UT RDD which was used in developing improved rolling sensors and an improved data analysis software as well as evaluating new data collection hardware We owe thanks to Industrial Vehicles International ivi for developing the prototype ivi RDD that was used in Year 1 The generosity of Mr Jay Bird Mr Elmo Christensen and Mr Mike Grady of ivi in sharing their ideas the prototype vehicle and prototype rolling sensors with us in Year 2 at no cost to the project is appreciated Further the willingness of ivi to loan the hydraulic system associated with the front rolling sensor for prototype testing in Years 3 and 4 was greatly appreciated as was the constant interaction and strong support between ivi CTR and TTI during acceptance testing Finally we are grateful to Mr Don Ramsey and the personnel of the TxDOT Flight Services Facility for their kindness in allowing us to continue to use a pavement testbed at their facility during Years 3 and 4 of the project Products Appendix B contains product 0 6005 P2 CTR Post Processing Software for Calculating Peak to Peak Dynamic Deflections from the Voltage Signals Generated by the RDD Rolling Sensors vi Ta
67. k slab was used During application of the forces over a range in frequencies to the pavement the signals of 2 Hz geophone in the rolling sensor package and the two 4 5 Hz geophones were recorded and then averaged movements from the two reference geophones and rolling sensor transducer were compared This procedure was performed for all three rolling sensors Field calibration curves for the three RDD rolling sensors are shown in Figure 2 16 As seen in the figure the front and 15 rear sensors showed similar values while the center sensor showed a slightly different value likely because the center sensor has different sized wheels Calibrated 4 5 Hz Geophones Calibrated 4 5 Hz Geophones d j L i VK V E ee 2 Hz Geophone Hidden Pavement Surface Layer by Sensor Package Base a Front View b Side View Figure 2 15 Set up Used for In Situ Calibration of a Rolling Sensor 1 6 9 14 EN 1 2 is gt 1 S og Calibration Factor 30Hz 06 NW SCH CS 0 65 V in sec so FS 0 73 Vi in sec s 02 RS 0 75 Vi in sec 0 I 20 25 30 35 40 45 50 Frequency Hz Figure 2 16 Field Calibration Curves for the Three Rolling Sensors 2 4 2 Deflection Profiling with the Rolling Sensors In March 2012 personnel from CTR and TTI performed TPAD testing at the TxDOT FSF This testing was conducted to check the performance of the rolling
68. l depth flexible pavements so this project presented the project team with an opportunity to explore this potential 5 2 RDD Deflection GPR Elevation and Testing Speed Profiling The basic deflection data from the east bound test run is shown in Figures 5 1 and 5 2 Figure 5 1 is typical of most of this project The HMA thickness is over 20 inches and the deflections are all low in the range of 3 to 6 mils The exceptions are very close to the bridges or at the bridge approach slabs In these cases the deflections increase substantially The gaps in the deflection profile are over bridges because no testing was done on the bridge decks In these cases loading rollers and rolling sensors were raised when crossing each bridge deck d E Ub ga DL a UL guo a B nui SAAR zf emeemmlDpMI 7 5 we A Lat 513020875930 558878818304 Du r pTR Vaive A0 Expanded Profiles in b and c b Figure 5 1 Typical TPAD Data from Full Depth Asphalt Section of US 290 The one case where deflections were found to increase is shown in Figure 5 2 In this case the high deflections are associated with a bridge approach on an embankment that was 14000 ft from the start of the data collection operation As can be seen right of the vertical red line the deflections increase by a factor of 3 5 from 2 to 7 mils on average To explain the cause of this large increase in deflections it 1s necessary to view the synchronized GPR data shown in
69. lled by buttons 3 to 8 Shows the continuous normalized deflections trace chart deflections for 1 ok bad each 10 kip dynamic loading Use when wishing to return to analysis of deflections after using other features Data Shows Comprehensive Results view important this displays the deflection 12 View i and video data on a single screen j3 Data Shows Comprehensive Results view similar to above but with different View summary graphs ONE Turns to previous segment of data each max segment size is 3 200 000 as R gt Turns to next segment of data each segment size 1s 3 200 000 16 Shows the chart of altitude above sea level surface temperature and survey speed vs DMI Saves the final result to disk for viewing in other software like Excel Changes the TTI format to UT Austin format for analyzing UT data 52 The first and second buttons are used for loading and setting the parameters in the analysis Buttons 2 8 are used for viewing the raw and filtered data in chart form Buttons 9 and 10 are used with 3 8 buttons for scrolling through the data file Buttons 11 13 are used for showing the analysis results in different formats ButtonS 14 and 15 are for scrolling to earlier or later sections in the data file Button 16 is used to display the temperature vehicle speed and elevation data Button 17 is for saving a table of deflections vs DMI data into ASCII format and this data can be used by Excel for report gener
70. mph Figure 2 6 Rolling Noise Levels Averaged over 125 ft Long Testing Path at Three Different Testing Speeds Measured with Wheel Set SD 2 3 Design of Towing Frame by CEM Personnel The new format of the TPAD towing system for the rolling sensors was designed constructed and installed during this year The towing frame will replace the IVI designed sensor system described in report 0 6005 4 The new towing and lifting frame on the TPAD is used to position and tow three rolling sensors The towing lifting system was designed by CEM personnel at UT The functions of the towing frame are to 1 pull all rolling sensors from a single frame that is towed by the TPAD mobile platform 2 make the towing frame bridge the loading roller assembly so that the frame will automatically be lifted when the loading rollers are raised 3 make the towing frame compatible with all rolling sensors and 4 maintain rolling sensors at a uniform spacing with respect to the loading rollers and each other The towing frame was designed to accommodate three constraints that exist mainly due to space limitations under the TPAD The open space under the TPAD varies with the position of the loading rollers with respect to the loading frame The open space when the loading rollers are raised and lowered as well as when the static and dynamic loads are fully applied to the pavement must be considered The first constraint is the vertical space between the round bar and rect
71. n Figure 5 4 The humps in the elevation data are as 42 the TPAD passes over the bridge approaches For this data collection effort the TPAD was profiling at 2 mph However when crossing bridges with the loading rollers and rolling sensors in the raised position the speed of the TPAD was increased to an average of about 5 mph as it crossed the bridge decks TPAD Test Track Elevation Chart Figure 5 4 Elevation and Speed Data from the US 290 Data Collection 5 3 Conclusions from US 290 Testing 1 The TPAD provided useful data on this full depth HMA pavement 2 Areas with high deflections can be easily identified 3 The GPR data can be used to determine the possible causes of the higher deflections and identify locations where validation coring or additional testing would be helpful or required 43 44 Chapter 6 Conclusions The activities during Year 4 the final year of project 0 6005 have been successful and productive Rolling noise measurements with four different sets of wheels on the rolling sensor cart were evaluated to determine the wheels that created the least rolling noise To perform these tests an independent towing frame was utilized The optimum wheel diameter and tread width for the sensor carts was determined by a series of tests on the JCP at the TxDOT FSF The independent towing frame was pulled by hand to determine the rolling noise created by the pavement roughness and joints The use
72. nter analysis 1s conducted first Step 7 provides the specific inputs for then running an analysis of the front and rear sensors To analyze the center rolling sensor go to the Analysis m dialogue box and type dist F1 4 amp F1 1 and critic 1 23 0 1 0 0 and then click the Save and Run button as shown in Figure 11 The deflections are averaged over a 3 ft distance Figure 12 shows a screenshot of the Matlab main screen after clicking the Save and Run button Many arrays are generated as the figure indicates Under the list of arrays click deflection and copy and paste the array to an Excel spreadsheet thus creating column A Click distance and copy and paste the array to Excel spreadsheet to create column B Note that the user needs to create the column headings 0 ft Deflection and Distance see Figure 13 4 Return to the Analysis m box and then change critic 1 3 0 1 1 0 to move the averaged distance window by ft Click the Save and Run button As in the previous step go to the Matlab main screen s list of arrays and click deflection and then copy and paste the array to 74 the Excel spreadsheet to create column C Again click distance and copy and paste the array to the Excel spreadsheet to create column D Note that the user needs to create the column headings 1 ft Deflection and Distance see Figure 13 5 Return to the Analysis m dialogue box and then change
73. ording to their locations relative to the loading rollers The center sensor CS is located mid way between the two loading rollers while the front sensor FS and rear sensors RS are located 0 635 m 25 in forward of and to the rear of the center sensor CS respectively Figure 2 14 shows a photograph of the front rolling sensor FS attached to the towing frame and the location of the center sensor CS In an initial plan all three RDD rolling sensors had 12 5 in diameter wheels and heavier hold down weight 90 Ib because this sensor cart format had the lowest rolling noise However 9 5 in diameter wheels and a 40 Ib of hold down weight were used for the center sensor CS because of space limitations between the loading rollers In general since the center sensor is located closest to the loading rollers it will have the largest signal and a good signal to noise ratio SNR Therefore use of 9 5 in diameter wheels and the lighter hold down weight 40 Ib for the center sensor are considered acceptable l l l Direction of Travel l 0 635 m oading Front Sensor FS 2 1 ft Loading 5 Roller Center Sansor CS 0 365 m 1 2 ft 0 635 m Loading i 2 1 ft Roller Contacts Pavement Rear me RS 0 533 m 1 75 ft 0 533 m 1 75 ft Figure 2 13 Arrangement of Sensor Array with Three RDD Rolling Sensors 14 Center Sen sor dei L C wem
74. reen This software is supplied with a data set collected on US 287 stored in the folder TestData under the name USSISBTB RDD This is a short run of just over 1000 feet in length Loading the raw data is described in Section A 2 53 Once the file is loaded the total number of segments in the selected data set is displayed 1 for the provided short data set in the upper right dropdown edit box Figure A 3 The maximum number of data points that can be displayed in one segment is 3 2 million points At a TPAD speed of 2 mph using the default sampling rate this corresponds to roughly 9000 feet of data For long runs such as were made on US 287 over 5 miles of data was collected in a single complete run In this case the data is broken into three segments for viewing Total Data Segments Current Data Segment Figure A 3 Showing the Total Number and Current Segment Selected In the top left dialog box the user can input the filter parameters to control how the raw data will be processed using a digital software filter Inside the software a band pass filter 1s used to remove unwanted noise and to get the geophone and load peak to peak amplitudes The TPAD currently uses a 30Hz loading frequency a band pass filter range of 29Hz to 31Hz is used to remove unwanted components of other frequencies Only signals in this frequency range will be accepted for analysis A 2Hz low pass filter 1s recommended for getting the static load Both filter
75. rogram and the first screen shown in Figure A 1 will appear Figure A 1 Main Menu Screen of the TPADana Software We can see that all the functions can be accessed from the toolbar buttons Table A 2 describes the function of each button The correlation of icon symbols to TPAD functions is as follows icon means the geophone D icon means the raw data of geophone Ll icon means the load related and UE icon means the raw data chart of load signal 51 Table A 2 Function of Each Toolbar Button Icons Functions Program setup or options clicking this button will show the options dialog box used to set up the basic parameters for running this software Loads the analysis file to be processed the first step required to analyze test data 3 Shows the raw load chart similar buttons shown in 9 and 10 are used for s turning to other pages of data to be displayed Shows the raw geophone chart 5 H Shows the power spectrum chart for each set of 5000 raw load data points d button on 9 10 are used for turning to other pages of data to be displayed e e Shows the power spectrum chart for each 5000 raw geophone data points 7 H Shows the raw and filtered load chart button on 9 10 are used for turning to other pages of data to be displayed DOIT Shows the raw and filtered geophone chart DOEN Turns to previous page for the operations controlled by buttons 3 to 8 nmm Turns to next page for the operations contro
76. rolling sensors designated as CS FS and RS respectively The CS and RS showed good agreement up to a testing speed of 2 mph while the FS exhibited somewhat poorer agreement compared with the CS and RS All three rolling sensors showed generally poorer results at 3 mph combined with significant rolling noise at the joints The poorer performance of the FS at 2 and 3 mph is due to the chatter created by the front swiveling wheels on the towing frame A testing speed of 2 mph and use of the CS and RS for pavement performance analysis are recommended If however a testing speed of mph is used then all three sensors work well Three case studies with the TPAD were performed on US 297 near Wichita Falls CRCP at TxDOT FSF in Austin JCP and on US 290 near Houston HMA Conclusions from US 297 near Wichita Falls are as follows 1 The TPAD system and data acquisition system worked well at the testing speed of 2 mph All sensors were operational The only issue encountered was crossing failed areas similar to Figure 3 6 the load sensors were lifted to avoid the risk of damage 2 Very limited continuous deflection data has previously been collected on CRCP pavements but the data generated on this project point out some important issues with the pavement structure 3 The data highlighted problems with the current full depth patching operation that the District has been using This should be reviewed by the District staff and Construction
77. s are clear and there is also a layer of steel at mid depth in the concrete On the first few thicker slabs it 1s very difficult to see the bottom of the slabs and there does not appear to be any steel in these slabs other than at the joints 4 5 Conclusions from TxDOT FSF Testing l P The TPAD works very well on JCP Areas of poor pavement support can be easily located based on the center slab deflections from the center sensor The front and rear sensors provide the best information for computing LTEs The TPAD will be more than 10 times faster than traditional FWD testing and it can also readily test every joint in the pavement Future efforts should be aimed at automatically generating statistics from the shape of the deflection pattern over joints this seems to be highly correlated to the FWD LTE Temperature has a large effect on LTE joints with poor LTE at lower temperatures will see an increase in LTE as the temperature rises and the joints lock up More work is needed to account for test temperature on LTE determination with the TPAD 40 Chapter 5 TPAD Case Study 3 US 290 Houston Full Depth Hot Mix Asphalt Pavement 5 1 Introduction The Houston District wanted to test a full depth asphalt section of US 290 to determine if it is a suitable candidate for a concrete overlay The key requirement is that the deflections pavement support should be uniform Very little TPAD data has been collected on ful
78. s are set in the Signal Analysis Filter Control box shown in Figure A A Signal Analysis Filter Control From Ta Dynamic analysis Bandpass filter E E Hz Static Load analysis Lowpass filter 2 Hz Figure A 4 Filter Parameter At the top center of the options dialog box Figure A 5 the user can select which geophone sensor is to be analyzed and displayed as well as input the sensor calibration factors The unit of the load calibration factor is 5 0 kIb V and the geophone calibration factor is V in sec The units of the geophone calibration factor are tied to velocity the final deflection results are reported in mils one thousandth of inch The software converts the geophone output into mils Geophone 1 is the sensor between the loading rollers of the TPAD whereas sensor 2 is the lead sensor and sensor 3 is the trailing sensor Select Geophone Sensor Sensor Calibration Factor Load Factor E Geophone 1 MM Center Geophone RE Front Geophone p II Rear Geophone p faa Figure A 5 Sensor Selection and Calibration Factor Input Screen 54 In the right middle section of the options dialog box Figure A 6 the user is asked to input the offset distance from the loading wheels to the video image Since the video camera 1s mounted on front of the TPAD vehicle the image from the camera does not match the TPAD s deflection test location The goal is to provide an offset so the user can view the pavement condition at the location of h
79. sed for the center rolling sensor because of the space limitations around the loading rollers Since the center sensor generally has a higher SNR because of the proximity to the loading rollers this wheels weight set up was considered to be acceptable Three rolling sensors in the towing frame system were tested at TxDOT FSF under similar surface temperature conditions around 70 F The center and rear rolling sensors showed good agreements in deflections profiles determined at testing speeds up to 2 mph when compared with the deflection profile at 0 5 mph In contrast the front rolling sensor showed poorer performance in the same comparison 0 5 vs 1 and 2 mph It seems that this poorer performance of the front sensor results from the chatter in the front swiveling wheels on the towing frame At a testing speed of 3 mph deflection profiles collected with the center rear and front rolling sensors exhibited comparable results in general but showed considerable differences mainly at the JCP joints 23 24 Chapter 3 TPAD Case Study 1 US 287 near Wichita Falls CRCP 3 1 Introduction The US 287 data collection effort was on a CRCP section near Wichita Falls that was scheduled to receive a bonded concrete overlay The total project length was around 5 miles with data collected at a profiling speed of 2 mph This project was the first project level test conducted with the TPAD The primary goal was to check that the system is working corre
80. sensors with the towing frame system Testing Path E at the TxDOT FSF was used for this work Testing Path E is shown in the aerial photograph of the TxDOT FSF in Figure 2 17 In an earlier study the effect of temperature on the rolling dynamic deflections were well established joint 33 deflection dropped from 38 mils IOkips at 91 F to 12 mils 10 kips at 128 F Stokoe et al 2010 To avoid 16 this effect testing was performed during the morning time when pavement surface temperature was quite stable and around 70 F Figure 2 17 TxDOT s FSF at Austin Bergstrom International Airport ABIA Continuous profiling along Path E with the final system of rolling sensors and the towing frame was performed at average testing speeds of 0 5 1 2 and 3 mph The deflection profiles collected with the center sensor CS at testing speeds of 0 5 1 2 and 3 mph are shown in Figures 2 18 2 19 and 2 20 respectively In these figures the deflection profile evaluated at a testing speed of 0 5 mph was used as the reference It is well known that rolling noise level generally increases as testing speed increases Upon comparing the figures it can be seen that profiles determined at 0 5 1 and 2 mph show similar results see Figures 2 18 and 2 19 Figure 2 20 presents the comparison between profiles determined at 0 5 and 3 mph and the rolling noise has increased considerably at the joints but there 1s still a reasonable match between the two profiles
81. sor Deflection Patterns over Two Joints with Different LTEs 38 Figure 4 8 Comparison of FWD LTE to a Statistic from the Front Sensor of the TPAD at o ee E E es T A eo E E N A A E ee A E T 39 Figure 4 9 Comparison of FWD LTE to a Statistic from the Front Sensor of the TPAD at Eeer 39 Figure 4 10 GPR Data from the TXDOT bouge ota co saseedesadvessanseaveatiandendeabadeaedeoadsectenneasenls 40 Figure 5 1 Typical TPAD Data from Full Depth Asphalt Section of US 290 4 Figure 5 2 US 290 TPAD Data from a Location of Varying Deflecnon 42 Figure 5 3 GPR Data Collected by TPAD Showing Variation in Subgrade Moisture Eea E a eres m r 42 Figure 5 4 Elevation and Speed Data from the US 290 Data Collection 43 Figure A 1 Main Menu Screen of the TPADana Software 51 Figure 3 2 Program EENEG 53 Figure A 3 Showing the Total Number and Current Segment Selected 54 Fee ees ae ol EE EE 54 Figure A 5 Sensor Selection and Calibration Factor Input Screen 54 Figure A 6 Offset and Moving Average Parameters Input Screen 55 Figure A 7 Options to Smooth Raw Deflection Data 55 Figure As Output Chart Format Parameters o btt eee A AE 55 Figure A9 Open File Dialog BOX enges petitus ases an a E T R 56 Figure A 10 Plot of Raw Data Blue Line and Filtered Data Using the Options Selected giu A E
82. ss and support frame of the loading rollers Two loading rollers are used to apply the static and dynamic forces to the pavement Each loading roller is 1 5 ft in diameter 1 2 ft wide and has a tread that is about 2 5 in thick and is made of a 92 durometer shore A polyurethane material A cross sectional view of the TPAD RDD loading system is shown in Figure 1 2 The air conditioned cab at the front houses the driver operator of the data collection activities and hardware and software systems The cab has an internal volume of approximately 142 cubic feet Electrical power is generated with an onboard 2 000 watt pure sine wave inverter All movements of the mobile platform the RDD loads imparted to the pavement and the raising lowering portion of the rolling sensor system are hydraulically operated I 30 Hz Sinusoidal Air Spring Isolation 0 36 m 07m 0 36 m 1 2 ft 2 3ft 1 2 ft Figure 1 2 Cross Sectional View of the TPAD RDD Loading System Which Applies Static and Sinusoidal Dynamic Loads to the Pavement with Two Loading Rollers Currently the TPAD has three RDD rolling sensors Each rolling sensor is composed of a three wheel cart on which a 2 Hz geophone 1s attached One sensor is shown in Figure 1 3 The 2 Hz geophone is used to measure dynamic deflections of the pavement created by the sinusoidal dynamic loads The sensors are arranged in a linear array that is oriented along the longitudinal axis of the TPAD A tow
83. t 5650 feet and 5680 feet at either end of the 30 ft long patch as shown in Figure 3 4b Zi vacuos 1 8 88 Lat 33 48483870 BT TSBBAOBA 7 5B588 5 in b and c e Figure 3 4 Typical Problems Found at Either End of Concrete Patches 4 G Ub on UL ow DA gr a fe I Se RARE UT Lat 533 478 0428 ET TIR Che RR yeu EER WI in b and c Figure 3 5 Area High Deflection Area Not Associated with a Patch Figure 3 5 shows a different pavement condition from the one shown in Figure 3 4 In Figure 3 5a a high deflection area in highlighted by the vertical red line at this location the pavement has closely spaced cracks 2 to 4 feet apart but no patching Photographs showing other problem areas are presented below in Figure 3 6 The photograph on the left Figure 3 6a shows the consequences of problems at the end of the patches where the adjoining concrete fails next to the new patch The photograph on the right Figure 3 6b shows the lack of a seal between the CRCP main lanes and the HMA shoulder It is thought that the lack of the seal is the origin of most of the distress on this highway The lack of the seal results in base erosion which 28 leads to edge failures and patching The proposed concrete overlay will not address the fundamental problem with this pavement section To summarize the recommendations to the District on the results from the
84. utton is used for showing the geophone s raw and filtered data chart as given in Figure A 17 61 64752 to 70148 NAJ VA J VV d AAA j V V V VV F VA f N Ay UV i f VV A i J X X MN A Ay H s V JV V v V V VY A Ad Y hy NVAAANAAAWVVAAY AY Yh A Wh 37 6 137 9 138 2 138 5 138 8 139 1 139 3 139 6 139 9 140 2 140 5 140 8 141 1 141 4 Figure A 17 Chart of Raw and Filtered Geophone Data A 11 Button for Showing the Normalized Continuous Deflections Trace Chart ok bad When the user clicks the 4 button as shown in Section A 2 all the data is loaded to the memory and the chart from Figure A 10 1s displayed The user can also return to evaluating deflections after using other program features by clicking the SS bad button This chart shows the deflection normalized to a 10kip load The X axis is in units of feet and the Y axis is in mils lOkips Data A 12 Comprehensive Results View Option 1 Vie By clicking view d the comprehensive results screen shown in Figure A 18 will be displayed The top part of this screen shows the normalized deflections results as in Figure A 10 On the top right side of this chart you can find the GPS coordinates DMI offset and deflection value in mils that are updated when the user moves the mouse inside the chart In the bottom left side of the screen the video frame shows the video with the DMI offset in miles and feet corresponding to the location on the deflection trace that t

Developing a Testing Device for Total

Contents

Download Pdf Manuals

Related Search

Related Contents