Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Abbreviations, Acronyms, Initialisms, and Symbols AC asphalt concrete AI GPR Activity Index CMP common midpoint DMI distance measuring instrument DOT Department of Transportation FCC Federal Communications Commission FHWA Federal Highway Administration GPR ground-penetrating radar GPS global positioning system GSSI Geophysical Survey Systems, Inc. HMA hot mix asphalt IE impact echo NCAT National Center for Asphalt Technology NDT nondestructive testing PSPA portable seismic pavement analyzer SASW spectral analysis of surface waves 2-D two-dimensional 3-D three-dimensional 15
Appendix A1 GPR Vendor Survey Responses Feedback from Vendor Survey for GPR Question 1 Do you currently supply equipment that can be configured to meet the proposed requirements? 3d-Radar Yes GSSI Yes MALA Yes IDS Yes Question 1a If yes, please describe the components that you would use, and how they would be configured to meet these requirements. 3d-Radar GeoScope Mark 4 data acquisition and radar unit; Model VX3341 antenna (3.3 m wide antenna array with 41 channels, 7.5 cm spacing). Older system (3231) had 10 cm spacing. Power source, DMI, GPS, cables, software. Data is now stored on the laptop. Minimum laptop requirements include adequate disc space with solid state drive or 7800 rpm hard disc, and minimum I-5 processor with adequate memory. Power requirement - 12V power source. Can run off of vehicle power. Recommend 12V deep cycle marine battery (70 amp hour, good for one day). GSSI GSSI would use two synchronized SIR-30 data acquisition and control units. Each unit would run 4 antennas, so the total would be 8 antennas. Then 8 antennas, when spaced 1.5 feet apart, would cover the full 12 foot lane width. The two units would be controlled by the operator using a single touch screen monitor and an optional keyboard and mouse. Data acquisition would be controlled using a DMI, and GPS data would be directly acquired and merged with the GPR data using a standard GPS receiver. There are two antenna options: (a) 2-Ghz horn antennas (typically 20" Ã 20" Ã 8" wide); and (b) 2.6 GHz ground coupled antennas (typically 7" Ã 4" Ã 2" high). The horn would permit unlimited driving speeds to meet the stated data collection requirements. The ground coupled is less expensive, but is limited to approx. 20 mph. A1
Feedback from Vendor Survey for GPR MALA MALA has 2.3 GHz antenna, and can run 6 antennas at a time with ProEx system using multichannel expansion. Each expansion module runs 2 antennas. MALA also has a 16 antenna array of 1.3 GHz antennas, called MIRA. The swath width of the array is 1.5 m. It is a true array, in that it can transmit from one antenna and receive by others. Example: can go from T1 to R8 (1 meter apart). Firing sequence, T1 to R1-R4. set up pairs. Number of pairs? MIRA can go 12 MPH. Can run multiple, operate independently. Run with total station for high precision. IDS IDS has a multichannel control unit, powers up to 8 antennas. Up to four Control units can be connected together in order to power up 32 antennas. Can connect 8 horn antennas to a single control unit. 1st antenna connected to the control unit. Other antennas connected to each. (chain connection). HI-PAVE - 1GHz and 2 GHz antennas, Multichannel control unit, customizable. Can use multiple controllers to increase speed. Question â 1b What type of carrying/mounting assembly would you use to meet the target testing speed? 3d-Radar Use a portable car charger to raise/lower the system. Could use a hand winch. Need at least 2 people to set up on vehicle. Would be nice to fold antenna. Will consider. GSSI The deployment would depend on the antenna type. For the horn antenna array, each horn antenna would be cantilevered from a mounting structure. One option is to have all 8 antennas mounted behind the rear of the vehicle. Another is to have 4 in front and 4 in the rear, offset appropriately. The primary challenge is safely deploying the antennas that are outboard of the vehicle footprint. This would clearly need some design work. The ground coupled antennas are smaller and might be easier to deploy. They would ride on a replaceable skid plate which would drag on the ground. MALA MIRA Array has a cart. Cart can be mounted to a vehicle. There is also a road cart for 2.3 GHz. MALA has also put together a 6 antenna road cart as a special order for a client. IDS Nothing off the shelf for 8 antennas. Need to evaluate effect of proximity. Currently have setup for 4 antennas. A2
Feedback from Vendor Survey for GPR Question 1c Also, if yes, can you exceed the proposed requirements, and to what degree? 3d-Radar Design goal is true highway speed. Mark 4 has 2 receiver boards. Transmit from one and receive from 2 at the same time. This will be included in the next release of firmware. With this upgrade, Task 7 5 mph test could now be at 15 mph. New antennas have GPS chip which will allow GPS data to be embedded with the GPR data. GSSI The system proposed as Option (a) in Question 1a, with horn antennas, can generate 10 scans/foot travelling at 50 mph. The system with the ground coupled antennas, while limited to 20 mph, can also generate 10 scans per foot at that speed. MALA With 2.3 GHz, can do 55 mph at 2 scans/foot, or 4 scans/foot at 35 mph. IDS Data collection rate - 8 antennas - 32 kmh with 10 cm data spacing, 30 cm spacing at 100 kmh. With 2 control units double the spacing. Can use up to 4 control units. 100 kph 7.5 cm is the maximum rate using 4 control units. Question 1d If not, do you have plans to offer equipment that would meet these requirements? When would this equipment become available? 3d-Radar Question does not apply. GSSI Question does not apply. MALA Researching higher frequency and speed, details unspecified. IDS Question does not apply. Question 1e If no, what changes in the requirements would need to be made in order for your equipment to comply? 3d-Radar Question does not apply. GSSI Question does not apply. MALA Lower the frequency to 1.3 GHz. Or, use 2.3 with 2 foot lateral spacing. IDS Question does not apply. Question 2 What would be the approximate cost range for the system that you would propose in answer to question 1? 3d-Radar Estimated costs: Geoscope: $155,000 3 meter Antenna: $100,000 3-DR Examiner: $14,200 3-years support: $5400 Total: ~$270,000 1.8 meter Antenna alternative - requires 2 passes - $43,000 A3
Feedback from Vendor Survey for GPR GSSI Including all mounting hardware, estimated cost is $200K for the horn system and $150K for the ground-coupled system. MALA Pro-EX $125K max; MIRA has never been sold - strictly a research system. IDS Rough estimate of horn, 8 antennas plus control unit. $150K Question 3 Are there modifications to the proposed requirements that you would recommend based on your perception of the overall objectives of this system? 3d-Radar Agree with requirements GSSI With the horn antenna system, you can increase the speed requirement to 40â50 mph. This would make the system more suitable for highway work and would eliminate the need for a rolling closure. MALA None IDS IDS would suggest considering the alternative Hi BrigHT system with 16 antennas at 2 GHz, ground coupled, spacing 3 to 4 inches between antennas obtain delamination results. Speed 6 to 7 km/h. Question 4 Based on your experience, who are the probable buyers for this equipment? 3d-Radar DOT's, larger consultants for network-level survey and for multiple uses. GSSI Most likely large companies providing pavement evaluation services. MALA For MIRA it would have to be DOT. For pro-ex it could be DOT's or consultants who work for DOT's. IDS Engineering companies involved in asphalt pavement evaluation. Road authority, but often don't have expertise. Question 5 Based on your experience, how frequently do you anticipate this system would need to be upgraded, and what would be the cost implications? 3d-Radar Depends on routine SW upgrade, bug fixes. New SW release every 6 months. Also depends on how frequently customer wants to upgrade. Also firmware support - biannual basis. New models and enhancements. GSSI Hardware should not require upgrade. Software is typically upgraded once a year, and these updates are provided free of charge to equipment owners. Equipment has a two-year warrantee. Extended warrantees are available for a fee. MALA Hardware - 5 years; Software - 3 years; SW negligible cost. Mounting might need repair. IDS Multichannel arrangement since 2004, and don't plan to change hardware. Would expect a software upgrade every 2 years. SW upgrades are free for customers. Extended warrantee and support. Typically 7%. of purchase price. A4
Feedback from Vendor Survey for GPR Question 6 What are the features of your process that supports data transfer and analysis? 3d-Radar System uses a fast solid state hard drive. Viewing the data with Examiner for processing can be time consuming due to IFFT and background removal. Data volumes can exported. Examiner has an open door. Anyone can create an algorithm as a task using C++. Examiner provides a software development kit to facilitate this process. GSSI Data is recorded directly to a solid state hard drive built into the system. Pure raw data is recorded, and settings (filters, gains) are recorded separately. The advantage is that if there are operator errors in field settings, the proper data can be recovered during processing. Also, programs for real time data analysis can be built into the system. Data can be downloaded after the survey using a standard ethernet connection or a conventional USB external drive. User can mark various features in the data (during collection and during processing) using programmable function keys (F1âF10). MALA Laptop runs system. Data transfer is Ethernet standard. IDS Data goes directly to laptop. All data is raw (unfiltered). Acquisition parameters saved. 16 bits per sample. Question 7 What are the features of your analysis software that supports our pavement evaluation objectives? 3d-Radar Prototype delamination detection software has been developed, and needs some further testing. Software worked on the test track, but needs more correlation with ground truth in other situations. Examiner has 3D layer tracking. Can output the time, point by point, and thickness. It also has an algorithm for calculating the dielectric constant for each layer. This can be done by setting up the Geoscope for periodic common midpoint (CMP) array measurements. These measurements produce a second data cube, which is analyzed by Examiner to get a map of dielectric. Accuracy of this process degrades with depth. GSSI RADAN Software can do time-depth slices of the data. These slices can be converted to bitmaps with GPS coordinates so that they can be displayed on Google earth. Slices must be done on the entire data file - not easy to break it down into smaller length units. MALA No analysis tools. 3D imaging. 3D migration. 6 antennas gives 6 data files, individual profiles. A5
Feedback from Vendor Survey for GPR IDS RIS Hi-Pave layer tracking, filtering, can represent cores, like RADAN. Can also represent video data and GPS. The Post processing platform can work with 8 horn antennas but currently is mainly dedicated to bridge deck evaluation. Can be optimized for road evaluation. IDS requires non-disclosure agreement to allow outside groups access to their data. Is possible to post process IDS data with other proprietary software. Question 8 How do you analyze the data and report delamination conditions? 3d-Radar Prototype software - marks areas where it thinks there is delamination An export function could be included. Also, algorithm can do some distinguishing levels. Maybe/probably/definitely GSSI RADAN provides Time-depth slices and automatic layer picking for detecting asphalt layering. The layer picking information is saved as text files which can be reported in spreadsheets or plots. MALA MALA has no special software for this function. IDS Currently only done with bridge decks. Would have to develop adaptation for horn antennas. Maps and depth slices, contour plots. Question 9 What is the level of expertise needed to operate the equipment and analyze the data? 3d-Radar Operate equipment - not much expertise. Field technician â 2 to 3 weeks training. Need automatic tools for self-diagnostic. Analyzing data - need more expertise. GSSI Field work requires a field technician who can do repetitive work but has enough knowledge to handle potential problems and troubleshooting. Analysis requires more training and experience. MALA Field Technician for data collection. IDS Data acquisition is relatively easy. Post processing need experience, but don't need to be engineer or geophysical expert for pavement thickness. Delamination detection is semi-automatic in our post processing platform we need to evaluate if the algorithm have to be changed using horn instead of ground coupled antenna. Question 10 What do you consider to be the unique strength of your system in meeting the proposed objectives? 3d-Radar Full coverage, high speed, high bandwidth. GSSI Data collection speed, system flexibility, ease of use, software widely used, company experience with pavement applications, large installed customer base, and customer support. A6
Feedback from Vendor Survey for GPR MALA MIRA Array - true multipath array. ProEx has speed capability. Each antenna can have separate settings. IDS Modularity, speed, number of antennas. Good quality data. Question 11 Similarly, what do you consider to be potential limitations of your system? 3d-Radar Ground bounce issues revealed in Task 7 data; price GSSI None discussed MALA Ground-coupled, collision with objects. IDS Will need customization for putting together 8 horn antennas. Post processing software needs new parameter for delamination analysis. Question 12 What are the weather and pavement condition limitations for your equipment? 3d-Radar Severely damaged pavement making antenna bounce - bit of rain okay. Water on pavement okay. Get false positive from wet pavement due to local scattering. Water spray not a problem due to the shielded shape. GSSI Rain MALA Rain is a limitation, since the components are not weather proof. Heavy road salt presence could be a problem. Array is encased. IDS Same limitation of others, not during rain or snow. Question 13 Has the equipment that you propose been approved by the FCC? If not, is the equipment going through the approval process, and if so, what is the status of this process? Are there any regulatory restrictions to the use of your equipment, and, if so, what are they? 3d-Radar 3d-Radar has a waiver for a step frequency system that has the characteristics of the system that they are selling. They want to make a change to the waiver - it currently prescribes a fixed frequency step size. They would like to make it a minimum step size. Equipment configuration will be tested for FCC approval over the next 2 months. Right now, 3d-Radar has a 1-year permission to operate the equipment for marketing purposes. GSSI All equipment has been FCC approved except for the "smart" version of the 2.6 GHz ground-coupled antenna. (52600S). With this version, the SIR-30 automatically detects the antenna type and optimizes the allowable system speed for that antenna. Since the "non-smart" version of this antenna is FCC approved, approval of this new version is considered routine. MALA Pro-Ex is FCC approved. MIRA is not, and MALA is not pursuing FCC approval for this system. IDS Currently IDS 2 GHz horn is FCC certified, along with control unit. A7
APPENDIX A2 SASW/IE Vendor Survey Responses Feedback from Vendor Survey for Mechanical Wave Methods Question 1 Do you currently supply equipment that can be configured to meet the proposed requirements described in the GPR specifications? Olson YES Geomedia NO â The current single-point portable pavement seismic analyzer (PSPA) system cannot meet the testing collection rate. However, the company is working on three projects to develop similar systems for concrete, including automated software (Air Force), light-weight equipment (tunnels), and a transverse measurement array for robotic testing (Rutgers), that would achieve incremental steps toward an NDT system that may meet the proposed requirements for asphalt pavement delamination. These projects are in early development stages and are not ready for public sale. Question 1a If yes, please describe the components that you would use, and how they would be configured to meet these requirements. Olson Olson Engineering is receptive to selling the system and could have a commercial model available by 2013. The system consists of: ⢠Array of three sensor wheel pairs. Each wheel is capable of independent IE measurement. Each pair measures SASW. Spacing between the impact source and motion sensors is fixed for pavement delamination, but the spacing between the motion sensors could be adjustable for other NDT applications. ⢠Control module. Handles the test sequencing between multiple wheel pairs. ⢠Mounting hardware. The hardware is full lane width (approximately 10 ft) and allows for any transverse spacing of the wheel pairs. ⢠Data acquisition system. The Freedom Data PC is standard commercially available Olson equipment for numerous Olson Instruments NDT systems. ⢠Data analysis software. The software is still at an applied research stage of development but has been advanced since the AL, FL, and KS HMAC field tests. A8
Feedback from Vendor Survey for Mechanical Wave Methods Geomedia The equipment that is currently under development is an array of eight impact sources and fourteen motion sensors split onto two 3-ft beams with the capacity to measure IE every six inches transverse across the six-foot array. The sensors have a capacity of 42 kHz. The source has a range of 17 to 30 kHz to cover a range of pavement temperatures. The 6-ft transverse array can collect the eight sets of data in two seconds. The array can be folded for easy transport to pavement sections. All data collection and analysis are continuous and achieved on a single laptop computer. Other parties are building the testing vehicle with the DMI to carry and position the measurement array. The software developed for PSPA has all the needed analysis and mapping functions and will be upgraded to display all eight (four?) signal traces from the array. Question 1b What type of carrying/mounting assembly would you use to meet the target testing speed? Olson The mounting hardware is fixed to the vehicle and can accommodate any transverse location of the sensors. The target testing speed is controlled by the sensors mounted on the wheels. Geomedia The test vehicle is under development by others (Rutgers) for bridge deck applications. Conceptually, a rolling system could be developed at a later time to improve testing speed. Question 1c Also, if yes, can you exceed the proposed requirements, and to what degree? Olson The wheel pairs can test every six inches in the direction of travel. An equipment configuration with six independent wheels can provide full lane- width testing at 2-foot spacing for IE measurements. An equipment configuration with six wheel-pairs could be provided for single pass, full lane- width testing at 2-foot spacing between adjacent wheel pairs for SASW. Geomedia The system under development will exceed the proposed lateral spacing for testing. The IE layout is a six-inch lateral spacing and the SASW layout is nine inches. Once the software is ready, the flexural analysis will exceed the capability of the SASW and IE. Question 1d If not, do you have plans to offer equipment that would meet these requirements? When would this equipment become available? Olson Question does not apply. Geomedia The current three projects will take 12 to 18 months. Further development on HMA testing may be ready in late 2014. A9
Feedback from Vendor Survey for Mechanical Wave Methods Question 1e If no, what changes in the requirements would need to be made in order for your equipment to comply? Olson Question does not apply. Geomedia See response 1. Question 2 What would be the approximate cost range for the system that you would propose in answer to question 1? Olson The commercial price for a six pair system is estimated to be in the range of $100,000 to $150,000. Geomedia The approximate cost of a 6-ft array with software is anticipated to be $80,000 to $90,000. The density (spacing) of the sensors ($5000 per set) will directly affect the total price. Question 3 Are there modifications to the proposed requirements that you would recommend based on your perception of the overall objectives of this system? Olson In the long term, when technology is available, increase the array from three wheel-pairs to six wheel-pairs for full lane-width, single-pass testing. Add video logging to tie the field measurement location (series of video frames for slow speed). Include a GPS location system since the wheel-based calibrated DMI can accumulate error over long distances. Develop a unit to automate lifting/lowering the test system OR trailer assembly. Include an infrared temperature sensor (for adjusting material modulus to a target analysis temperature). Place a laser surface texture ahead of the motion sensors (to identify rough/raveled surface). Geomedia The proposed requirements are appropriate for SASW and IE, but a new analysis approach, based on the flexural properties of the pavement, has the potential to exceed the capability of SASW and IE to identify delamination. It is possible that the 50 kHz frequency range is too high and could be reduced to 42 kHz with a substantial savings in hardware cost and no loss of measurement quality. It should be noted that the energy source pins should be capable of multiple frequencies to account for changes in pavement stiffness as temperature changes. A10
Question 4 Based on your experience, who are the probable buyers for this equipment? Olson Past DOT customers for NDT testing were DOT bridge maintenance. Probable customers for the new array system included both pavement and bridge departments for project development. Sales to consulting firms depend on the quantity of units. These sales would create competition with Olson Engineering consulting service. If the market is large, consulting sales would be okay. Geomedia The first tier buyers will be research consultants. The second tier buyers will be highway agencies with active pavement management programs. Purchase of this technology is generally limited by the resistance to âblack boxâ perception. Question 5 Based on your experience, how frequently do you anticipate this system would need to be upgraded, and what would be the cost implications? Olson Purchase of a system would probably include an option for a software maintenance agreement at a cost of approximately $1000 to $2000 per year. The hardware technology is maturing and will not require major changes. Geomedia Software will need to be upgraded every 3 years as base-software changes ($1,000 to $2,000). Hardware will need to be upgraded every 5 to 10 years as computer hardware upgrades ($5,000?). Sensors and source technology is mature and will not require upgrading. Question 6 What are the features of your process that supports data transfer and analysis? Olson The SASW/IE data files are not large and download to windows âon the fly.â Field data is 50 Kbyte per foot (about 25 MByte per mile). The field data is easily moved to a laptop for analysis. The field data is archived as it is downloaded to the laptop. Geomedia All data is transferred by USB. Ethernet is not cost-effective at this time. A11
Question 7 What are the features of your analysis software that supports our pavement evaluation objectives? Olson Automation of the analysis software as primarily advanced through bridge deck studies. IE data analysis is well automated. The more complex SASW data analyses needs further automation. The SASW software can automatically locate the surface wave portions of the signal. Automatic masking to locate cycle 1 is complete. Pattern recognition of the phase data will be implemented in the near future. It is planned to implement a modeling algorithm to match up with the experimental phase data to more fully automate SASW data analyses. The results of theoretical modeling will be the actual compression wave velocity profiles (depths) of the pavement. Geomedia The SASW software is robust and current development to automate the analysis will strengthen it. The IE software is not as strong because it has more limitations for HMA delamination detection. The developing flexural mode analysis is expected to exceed both SASW and IE and should be reported in 2013. Question 8 How do you analyze the data and report delamination conditions? Olson In general, a significant drop of surface wave velocity (from the SASW tests) in the dispersive curve is a great indication of asphalt debonding. When asphalt is cold, the Impact Echo test can detect the delamination deeper than 4 inches. Currently the surface wave velocities are plotted on a gray scale by slice depth. The analysis software generates 2D color-scale slices at depths and tabulates the planar summary values. A plot of each depth is presented side by side so that comparison of velocities can be made and delamination can be visually located. Locations/areas of delamination are also reported in tables that can be read into spreadsheets. Geomedia Current focus is on PCC testing. Further development for HMA is still in concept only. A12
Question 9 What is the level of expertise needed to operate the equipment and analyze the data? Olson The technician and engineer need Windows-PC and digital data processing skills similar to operating GPR NDT technology. Training will be required for field testing. The engineer needs physics, geotechnical, and materials understanding for data analysis. Both the field technician and analysis engineer need to recognize âbadâ data. Development of expertise should be similar to expertise development during the implementation of the falling weight deflectometer. Geomedia The military is training young (high school graduates) technicians to operate the equipment. The analysis is automated. More expertise is needed to review the quality of the results. Question 10 What do you consider to be the unique strength of your system in meeting the proposed objectives? Olson The system can identify delamination reasonably well in the top 6 inches of the pavement based on the accuracy from testing at the NCAT Test Track. SASW technology has the potential to give more information, like material properties (moduli vs. wavelength/depth) and crack development. Geomedia The test protocol and analysis are firmly guided by a deep understanding of the theory. The tight lateral spacing of the sensors exceeds the requirements. A13
Question 11 Similarly, what do you consider to be potential limitations of your system? Olson There is noise at the sensor pavement contact due to rolling wheel movement and the size of the sensor contact area. Olson has recently built and successfully tested a pair of sensor wheels on an extremely rough asphalt overlaid bridge deck for the Colorado DOT and the new sensor design greatly reduced the rolling sensor contact noise. It is difficult to control the speed of the tow vehicle at slow speeds of 1 to 2 mph. Reasonably uniform testing speed is important for smoother rolling of the sensor wheels. Area of coverage is about 25% in transverse direction. A dirty road surface influences impactor and sensors contact. It may be possible to brush the surface ahead of the wheels and/or use compressed air to blow dirt away. The data acquisition system had limited speed. This limitation has been addressed with a better data controller. Currently the three wheel-pair axles are not tied together to maintain a 1-inch testing offset sequence which results in occasionally testing at a 1-ft interval instead of the 6-inch interval. Geomedia Energy wave propagation is a complex subject. The analysis and confidence in the results will improve with a greater density of measured data. Question 12 What are the weather and pavement condition limitations for your equipment? Olson Temperature of the pavement is critical. High material temperature (100F) reduces the test response resolution. This is more of a problem with IE. A higher material temperature also impacts the SASW measurement because it reduces the wave velocity change at the delamination interface. More field experience on warm to very hot asphalt pavements is anticipated this summer to further understand operating temperature limits. The hardware system could be waterproofed to permit field testing in rain conditions. SASW and IE can be operated on wet pavement. Tests on a raveled pavement surface would likely need to filter out the excessive roughness. Geomedia PSPA can be operated in 20F to 120F and is not affected by wet conditions. The flexibility of the array frame will allow up to 1-inch vertical (rutting) difference. A14
APPENDIX B1 GPR User Guidelines B1
Use of Ground-Penetrating Radar (GPR) for Identifying Asphalt Pavement Delamination: User Guidelines Prepared by the SHRP R06D Research Team 1 General Theory Ground-penetrating radar (GPR) operates by transmitting short pulses of electromagnetic energy into the pavement using an antenna attached to a mobile platform or survey vehicle. These pulses are reflected back to the antenna with an arrival time and amplitude that are related to the location and nature of discontinuities in the material (air/asphalt or asphalt/concrete, reinforcing steel). The reflected energy is received in the form a series of pulses that are referred to as the radar waveform. The waveform contains a record of the properties and thickness of the layers within the pavement. GPR measures changes in the dielectric properties of pavement layers and the velocity of wave propagation within those layers. In a study on Texas highways, Scullion and Rmeili (1997) found that GPR technology was effective for detecting stripping in HMA layers where the deterioration was at a moderate or advanced stage. Stripped HMA typically has higher moisture contents or higher air voids, or both. The dielectric constant of the material is affected by both moisture content and air voids, as is the velocity of wave propagation. A GPR system can be implemented using one or more antennas, with four or more antennas considered a multi-antenna array. The purpose of an array is to collect equally spaced parallel lines of data simultaneously so that coherent areas of delamination can be identified and mapped. Data is collected continuously while the system is driven along the surface of the pavement. The data collection is typically triggered using a distance measuring instrument (DMI) mounted to the vehicle wheel or to an external distance wheel. The array produces three- dimensional data, Amp(x, y, z), where x is longitudinal distance, y is transverse offset, z is GPR time in nanoseconds (equivalent to depth), and Amp is the amplitude of the GPR signal. Testing conducted under SHRP 2 R06D using actual or simulated arrays has shown that debonded asphalt layers with trapped moisture and stripped asphalt layers will produce detectable reflections not normally seen with intact pavement layers. Semi-automated software to detect these reflections and map the areas of potential damage has been developed and tested under this project. Two methods of measuring these damage-related reflections were explored. Method 1 uses variations in waveform amplitudes in the time domain, while Method 2 uses a delamination detection algorithm in the frequency domain. B2
Method 1: Activity Index (time domain) A GPR indicator defined as the GPR activity index (AI) has been explored to identify the anomalies associated with wet delamination interfaces and moisture damage. This approach is based on the increased reflections from affected layers producing localized reflection anomalies within otherwise homogeneous layers. It is this additional reflection activity that the method seeks to quantify. Because these deterioration processes tend to occur non-uniformly in the pavement, a measure of the homogeneity of the electromagnetic properties was found to be useful in segmenting the roadway into features which can subsequently be used to plan more localized seismic testing and coring, as well as defining areas of deterioration. As a GPR indicator, the AI can be defined as the normalized average absolute amplitude of the GPR scan as follows in Equation 1: ), 2 ( ),( ),( yLx yx yx A AAI ± = (1) where A is the average absolute reflection amplitude for the scan at (x, y) and L is the normalization length. When compared to the values from neighboring locations, the index shows changes in reflection activity, which, if sufficient, may be related to delamination or moisture damage. Normalization allows for an AI that varies around 1.0 and thus permits lane-to-lane and site-to-site comparison without concern for the absolute values. For the initial site screening and segmentation, the normalization length can be selected as the entire project length. For detailed mapping of areas with potential moisture damage, a smaller normalization length may be appropriate to highlight local variability. Where scan magnitude variation occurs, the scan is scaled before AI normalization by dividing A by the amplitude of the direct coupling or end reflection of the scan. Another key parameter in the AI computation is the depth range over which this scan amplitude is calculated. The depth range should be selected to highlight the depth in which delamination or moisture damage is believed to be occurring. If this is not known, then a larger range can be used. Once cores are obtained and other data are available, this range can be reduced and the index recalculated. The depth range is bounded in the GPR scan, with the upper boundary defined by the interface between the air and the road surface, and the lower boundary layer as the interface at the bottom of the asphalt. To avoid including the amplitude of the reflection caused by the layer boundaries, the analyzed trace section should begin below the upper boundary layer and above the lower boundary layer. Method 2: Delamination detection algorithm (frequency domain) A delamination detection algorithm was prototyped based on the data collected at the NCAT facility. The algorithm concept incorporates the fact that delamination can occur at a relatively B3
wide range of depths and show a variety of amplitude characteristics in the recorded data. An energy-based study of frequency intervals was performed in areas of known delamination in the three-dimensional radar data. As represented in Figure B1.1, the time window of interest is extracted from every trace and converted to the frequency domain. The frequency spectrum is then divided into frequency intervals, or bins. The energy contained in each bin is calculated and the bins are sorted by energy values. Figure B1.1. Representation of the delamination detection algorithm: operations on the trace at position x = j, y = i. When an area is damaged, the waveform propagates in a different way through the roadway structure. This change can be used to generate a damage detection procedure. Because the analysis takes place in the frequency domain, every sample will carry information about the whole depth range to be analyzed. Sorting the energy values takes into account the varying amplitudes of signatures due to delamination. The operator chooses the bin of interest (e.g., bin 3, the red bin in Figure B1.1), the minimum size of delamination width and the cutoff threshold for the energy value that represents delamination, to produce a plot of the damaged areas (Figure B1.2). The rectangles in the picture delineate areas of delamination based on the final output of the algorithm and the statistical analysis based on parameter inputs. B4
Figure B1.2. Areas of delamination based on final algorithm and statistical analysis. 2 Equipment Specification Table B1.1. GPR Equipment Specifications Component Specification System type Array of multiple antenna elements lined up transverse to the direction of travel Frequency range: Impulse radar systems Center frequency of pulse > 2.0 GHz â10 db limits: 0.5 to 5.0 GHz Frequency range: Frequency sweep radar systems Up to 3.0 GHz Lateral spacing of antenna elements < 1.5 feet Lateral coverage per pass 12 feet (full lane width) Longitudinal data collection rate > 2 scans per foot per antenna element Travel speed during data collection >20 mph Travel speed during mobilization Posted speed limit Real-time display B-scan for selected antenna elements System monitoring and control From within the survey vehicle Data collection rate Data collection should be triggered on distance using a DMI Spatial reference Vehicle DMI and/or global positioning system (GPS) Detection depth range 2 to12 in. 3 Proposed Data Output and Display Requirements The field operation and playback software should be capable of the following displays: ⢠Direct time domain waveform (A-scan); ⢠Longitudinal profile for a given transverse offset (B-scan); ⢠Time/depth slice for a given time range; and ⢠Transverse profile for a given location or station. Examples of these displays are shown in Figure B1.3. Output Format Output should be a volume of data with amplitude as a function of x (longitudinal distance), y (transverse offset), and z (time). B5
Figure B1.3. Example GPR system output displays. station time (depth) time of slice in Figure (d) station of profile in Figure (b) pavement surface (a) longitudinal profile for a given offset ("b-scan") time (depth) offset amplitude (e.g., volts) surface (depth=0) offset of profile in Figure (a) (b) transverse profile for a given station (c) time domain waveform (a-scan) for a given station and offset station offset (d) time/depth slice (c-scan) B6
4 Equipment Calibration and Verification 4.1 DMI Calibration All systems will use some sort of DMI to control the data collection rate and to record linear distance. The DMI will either be attached to the wheel of the survey vehicle, or to a separate wheel associated with the antenna array. In either case, the DMI needs to be calibrated at regular time intervals to ensure accurate distance measurement. The DMI is calibrated by running the system over a pre-measured distance (usually ranging from 1,000 feet to one mile). The measured distance expressed by the DMI is compared to the known distance, and the DMI calibration factor is adjusted so that the two will agree. It is advisable to repeat the calibration after the calibration factor has been adjusted to confirm that the calibration has been carried out correctly. The measured distance should be within a foot (0.1% accuracy) of the pre-measured distance after the repeat calibration run. 4.2 Radar System Calibration The manufacturers of GPR systems do not provide user calibration procedures. If there are any concerns about the signal and receiver antenna system, the system should be returned to the manufacturer for evaluation. Field verification of radar system function will depend on the type of equipment system. Horn antenna systems usually require initial base-line measurement using a metal plate âbounceâ test to provide data for calculating dielectric properties and pavement thickness. Other systems, such as the 3d-Radar array and ground-coupled antenna systems, do not use a metal plate. These systems require other means for calculation of dielectric properties and thickness. Consult with the manufacturer for acceptable methods of field verification. 4.3 Verification of Radar Operation Awareness of problems in the GPR electronic signal is something that generally improves with the experience of the operator. Some simple checks are suggested below. One type of verification is to ensure that the signal is properly positioned; that is, the surface of the road is near the beginning of the signal (shortly beyond zero time). For a horn antenna or any other air-launch antenna, this can be quickly checked by sliding a metal plate onto the pavement under the antenna. The signal from the top of the pavement should increase to a very large value, highlighting the location of the pavement surface. For ground-coupled antennas, the same effect can be produced by lifting the antenna a few inches off of the pavement. In either case, the observed signal change should be located within 2 to 3 nanoseconds of time zero. Sometimes, when the equipment is set up in the vicinity of strong radio transmission sources, the data display will have the appearance of malfunctioning equipment. This problem generally disappears when the equipment is driven to a location away from the radio B7
transmission source. Degree of radio transmission sensitivity will vary with the type of antenna system. A practical method to check the radar operation would be to have an identified pavement test section of well-documented asphalt thickness and possibly with well-defined flaws, similar to that built at NCAT under this project. Periodically, the equipment can be brought to this section for data collection, to ensure that the results of the analysis are consistent with previous analyses and known conditions. 5 Climate and Pavement Conditions for Testing All pavement temperatures above freezing are acceptable for this type of system. Below freezing, the moisture in the pavement can freeze. Since frozen water has dielectric properties very similar to asphalt, the anomalies that are normally produced by moisture infiltration and moisture damage would not be present. Radar measurements on wet pavement are not recommended. The water will produce anomalous reflections which might interfere with the subsurface condition detection process. Testing in rain is not recommended since (a) the water can accumulate on the antenna elements and distort the GPR signal and (b) the wet pavement can produce a distorted signal as discussed in the previous paragraph. All other conditions are acceptable. 6 Testing Modes and Required Settings For a given array setup, with a given number of antennas at a set spacing, the variables to consider for the testing setup include time range (in nanoseconds), the number of samples per scan, and data rate (scans per foot of travel). Other settings such as gains and filters tend to be specific to the equipment, and the operator should have experience with these settings. The time range relates directly to the depth of detection. The specified 2- to 12-in. detection depth range translates to a time range of approximately 6 nanoseconds (ns) for the typical range of asphalt properties. For an air-launched system the time to reach the pavement surface must be added, so a minimum time for that type of system would be 10 ns. The stepped frequency systems specify a frequency range, not a time range, and the operator has less direct control over the resulting time range. The number of samples per scan defines the resolution in time. Typically a 10- to 20-ns waveform will be digitized into 256 or 512 samples, resulting in resolutions ranging from 12.8 and 51.2 samples per ns. Higher sampling rates show more detail and may facilitate processing, but there is a point beyond which there is no benefit. Based on past experience, rates on the order of 20 samples per ns should be sufficient for this application. The data rate determines the spatial resolution of the detected subsurface features. The more scans per foot, the more scans there will be in delaminated areas. This greater density usually increased the rate of detection. On the other hand, the speed of most systems is affected by the data rate; the more scans per foot collected, the slower the system has to go. B8
For project-level work, which generally demands greater detail, data can be collected with a moving closure, and therefore one should use the maximum practical data collection rate for that arrangement. Network-level applications, where the required results are usually less detailed, can get by with a reduced data rate, one which would be suitable for driving speed data collection. 6.1 Non-Array GPR Systems While this guideline recommends an array of antennas spaced laterally across the pavement, many organizations own single or dual antenna systems which they may wish to apply for this delamination application. This is possible by collecting multiple passes of data, so that the resulting series of data lines is equivalent to that which would be generated using an array. For example, if an organization has a single antenna system, they would need to collect, for a given lane, data lines at 1.0, 2.5, 4.0, 5.5, 7.0, 8.5, 10.0, and 11.5 feet offset from the shoulder line. Some arrangement needs to be made to ensure the accuracy of these offsets, and the ability to collect data while maintaining each line parallel to the shoulder line. The above data collection protocol would generate eight data files which can be combined to create the data volume discussed in Section 1 and in the following sections. 7 Test Output Data Formats The goal of the GPR field data collection is to produce a three-dimensional volume (x, y, z) of reflection amplitudes, with x (distance) as the direction of vehicle travel, y (distance) the offset between antennas, and with z (time) being the arrival time of the GPR pulse reflections. x and y can be obtained from field specifications or from attached GPS coordinates. The format of this data volume will vary with each equipment manufacturer. In some array systems, each antenna will generate a data file, and then these individual files will be combined to create the three-dimensional volume described above. In this case, each file will represent x and z for one particular value of y (the offset of the antenna). Software provided by the supplier can generally be used to combine these files into the 3-D volume. In other array systems, the 3-D volume will be generated directly during data collection. Note that this application will generate very large data volumes. Consider, for example, a one-mile survey, with an eight-antenna array collecting two scans per foot at 512 samples per scan. This collection will result in a file size of 8 Ã 2 Ã 512 Ã 2 (bytes per sample) Ã 5,280 = 43 MB, a size that is well within the range of standard jump drives. Use of more antenna elements, more scans per foot, and more bytes per sample can produce data files on the order of 4 to 5 GB per mile. At this high end, data storage and transfer can become a problem, and accommodations need to be made to deal with this. B9
8 Test Output Data Quality Control Checks It is highly desirable to check the quality of the data prior to demobilizing from the field site. Quality control can be carried out during lunch breaks or other pauses in the normal data collection process. Such checks should include ⢠Checking all of the recorded data files to confirm that their size is consistent with the amount of data collected; ⢠Correlating the field notes with the data files to ensure that all noted files actually exist; ⢠Scrolling through each data file to ensure that the data looks reasonable and that there are no problems with the data; and ⢠Checking the length of each file (recorded distance) and confirming that it agrees with the intended length. If problems are encountered or there are data of questionable quality, the data collection should be repeated. 9 Data Analysis Data analysis can be performed by both automated and manual methods. The goal of both methods is the same: to define areas of increased scan activity and thus delineate areas of potential deterioration. A 3-D volume of amplitude data can be obtained by extracting the reflection amplitude values from a three-dimensional radar file or by collating the data stored in files collected by individual antennas. The surface reflection and bottom of asphalt reflections are identified in each trace. If the analysis will be by depth slices, the individual GPR traces at each x, y location need to be adjusted so that the surface is horizontal. z can be converted to depth by GPR dielectric calculation using GPR data, or by estimation, of the materialâs dielectric properties. The thickness of each pavement layer can be computed according to Equation 2: ( ) ε2 tch âÃ= (2) where h is the layer thickness, tâ is the two-way travel time of the electromagnetic pulse wave through the layer, c is the speed of light in free space (c = 3 à 108 m/s or 11.8 in./ns), and ε is the dielectric constant of the layer. Three sites were analyzed for damaged areas using the AI measure of radar activity and the delamination detection algorithm discussed in Section 1. A control site, the NCAT test track, was manufactured with two lifts. The lower lift was 3 in. thick, and the upper lift was 2 in. thick. A layer of unbound asphalt (1 in. thick) was placed in small areas on the lower lift, creating areas of low density to represent moisture-damaged (stripped) asphalt. Measurements at the control site were used to develop algorithms which were applied at two field sites (Kansas and Florida). B10
The activity index at the test site was determined by using time slices through the C-scan at regular, overlapping intervals. At the two field sites, the scan sections for activity analysis were chosen by picking the layers of interest in the three-dimensional view and exporting the results to be used as bounding regions. At these sites, the ground surface and asphalt bottom layers were picked to constrain the time slices. The delamination detection algorithm was developed at the test site; it was then utilized at both the test site and field sites by using time slices through the C-scan at distinct intervals. 9.1 Data Analysis: Automated Tools Method 1: activity index (time domain) To automate analysis of the activity index, slices can be extracted from a cube of 3-D radar data. The depth and thickness of each slice needs to be defined for each location based upon the depth of pavement and area of interest. NCAT Site Activity was determined in 2 overall slices and in multiple overlapping slices between 2.1 and 5.5 ns from the start of the scan. These times represent locations at the air/surface interface and below the asphalt/base interface (from 0 in. to 8.5 in. from the ground surface). Figure B1.4 shows activity analysis results for depth slices at various depths below the ground surface at the NCAT test area with manufactured voids. Areas with increased reflection are shown in green (lower) to blue (higher). The x axis is distance in the direction of travel, the y axis is the offset distance, and the color shading represents the AI for the layer depth interval. These slices are roughly equivalent to the following depth intervals. ⢠Activity 18â26 (0 to 2.4 in.) ⢠Activity 21â30 (0.9 to 3.6 in.) ⢠Activity 25â34 (2.1 to 4.7 in.) ⢠Activity 29â38 (3.3 to 5.9 in.) ⢠Activity 33â42 (4.4 to 7.1 in.) ⢠Activity 38â46 (5.9 to 8.3 in.) Note that the damaged areas were placed at 1-in. to 2-in. depths. However, when a void is encountered in damaged pavement, the signal below the damaged area will reverberate and show a ringing pattern. High activity in the 29â38 (3.3 to 5.9 in.), 33â42 (4.4 to 7.1 in.), and 38â46 (5.9 to 8.3 in.) plots shows the effect that ringing beneath the void causes in AI calculation. This reverberation can cause confusion as to the actual depth extent of the voids. The ringing pattern is also noted in Figure B1.5, in the radar image below the void located at approximately 3 ns below the ground surface. As with the delamination detection algorithm, this method suffers in field application due to variation in pavement structure. For example, when defining slice intervals at the test track, B11
we are able to define slices based upon knowledge of the layer depths. It is possible to slice through the boundary between layers and map the boundary activity, or to create slices which are the thickness of the material between adjacent boundaries to examine the activity in each layer. In real world conditions, layer boundaries vary in depth, and it is necessary to avoid the layer reflections in creating the slice interval. Otherwise, slicing through a dipping layer produces reflections at the boundaries which appear to be areas of increased activity. Figure B1.4. Activity index for various slices from the NCAT test track. B12
Figure B1.5. Three-dimensional radar representation of NCAT site showing âstrippedâ asphalt layers at 3 ns (1 to 2 in. below the ground surface). Method 2: Delamination detection algorithm (frequency domain) The delamination detection method provides a way to automate the delineation of damaged areas using frequency domain radar data. This method is particularly convenient for stepped frequency GPR systems, since the raw data is in the frequency domain format and conversion to time domain can be skipped. The operator chooses the bin of interest (Figure B1.1), the minimum size of delamination, and the cutoff threshold for the energy value that represents delamination. Results are a plot of the damaged areas as seen in Figure B1.2. However, this method was not found to be reliable in field investigation. Differences in pavement layers and surface variation served to thwart the method. Since the algorithm was developed based on the experience at the NCAT facility, where simulated delamination provided solid and clean characteristics, further study of the data collected in real situations, combined with ground truth, is needed to refine the signature given by delaminations, both qualitatively and quantitatively. Further enhancement is necessary to factor in characteristics of the road substructure where variation in depth of targeted layers and lateral discontinuities are present. 9.2 Data Analysis: Manual (Semi-automated) Manipulation Since actual pavement layer boundaries vary in elevation, software was developed to extract depth slices from the 3-D radar model between asphalt layer interfaces. First, each layer is picked in the 3-D radar program and exported to a text file (x, y, z). These text files are used to bind the region of interest. To avoid including the amplitude of the reflection caused by the layer boundaries, the analyzed trace section began 0.5 ns below the upper layer and 0.5 ns above the lower layer. void ringing B13
Figures B1.6 and B1.7 show the picked surface reflection and bottom of asphalt reflection. The upper part of Figure B1.6 shows the longitudinal and transverse profile and the waveform defined earlier in Figure B1.3. The red and blue trace lines in the profile plots represent layer interfaces that were defined as limits for the activity analysis. For this example, the top of the pavement was selected as the upper boundary and the bottom of the asphalt was selected as the lower boundary. The lower part of Figure B1.6 is a color-shaded area representation of the depth of the asphalt bottom. The orange shades represent a shallower asphalt bottom and the light blue shades represent a deeper asphalt bottom (see 3-D version in Figure B1.7). For the 3d-Radar system data collected during this project, these layers can be picked using the 3d-Radar system Examiner software (3dr-Examiner). The layer depths are exported and then used as boundaries in the activity index calculation for the layer slices. When individual antennas are used, the individual GPR files are combined (as shown in Figure B1.8); the layers are picked in each antenna pass and again used as boundaries in the activity index calculation. Figure B1.9 shows a contour plot of the activity index that was calculated for the region between the picked layers. Figure B1.6. Three-dimensional radar analysis. Asphalt Bottom Surface B14
Figure B1.7. Three-dimensional radar analysis sections. Figure B1.8. C-scan time/depth slices from NCAT test section. (a) 3d-Radar array (b) Depth slice data created from combined data from individual radar antenna scans station (ft) 115 140 165 B15
Note that to create the 3-D cubes of radar data at the test track, the 3d-Radar system used five parallel passes with a 5.5-foot-wide, 21-channel swept-frequency antenna array (140 to 3,000 MHz), and the individual antenna system used 25 parallel passes with a single 3 GHz antenna. Figure B1.9. Contour plot of activity index between the two asphalt layer boundaries at NCAT. 9.3 Data Analysis Samples from Field Sites Florida Site At the Florida field site, the activity index was determined by performing calculations between the picked layers. The upper layer was the interface between the air and the road surface, and the lower layer was the interface at the bottom of asphalt. To avoid including the amplitude of the reflection caused by the layer boundaries, the analyzed trace section began 0.5 ns below the upper layer, and 0.5 ns above the lower layer. The end reflection was picked for this site, but it had an abnormal appearance, and normalization of the amplitudes by the end reflection did not work. Activity index results on either side of the roadway centerline were statistically different from one another, so each side was normalized by the average mean positive amplitude for that side before the activity index was calculated. Figure B1.10 shows a 3-D radar sample from the Florida site, and Figure B1.11 shows activity index plots for this data. For the activity index plot, a threshold of 20% above the mean is established; below this threshold, no colors are shown. Therefore, the only the areas that show color on the plot are those that represent potential delamination and moisture damage. B16
Figure B1.10. Florida radar data at the location of Cores 1, 2, and 3. B17
Figure B1.11. Florida site activity index plot of region between picked layers. Kansas Site At the Kansas field site, activity was determined by performing calculations between the picked layers. The upper layer was the interface between the air and the road surface, and the lower layer was the interface at the bottom of asphalt. To avoid including the amplitude of the reflection caused by the layer boundaries, the analyzed trace section began 0.5 ns below the upper layer, and 0.5 ns above the lower layer for each waveform. The end reflection was picked for this site, and the amplitudes were normalized by the end-reflection amplitude before the activity index was calculated. This step assures that the energy in each scan is identical. Figure B1.12 shows a GPR sample from the Kansas site. B18
Figure B1.12. GPR radar data from the Kansas site. 10 Test Reporting Reporting of the GPR test results can take place at various levels of detail, ranging from raw data to summary descriptions of the processed data. The level of detail should be dictated by the information needs. Based on the data presented in these guidelines, the following options are available: 1. Depth and profile slices of raw data This is the most detailed level of reporting. Examples of this type of reporting can be found in Figures B1.3, B1.4, B1.8, and B1.10. This type of output is useful for examining local detail of potential delamination conditions and for locating cores for confirmation of delamination conditions. It provides both the spatial location and depth of potentially delaminated areas. 2. Contour/area plots of delamination indicators, such as activity index Figures B1.2, B1.5, B1.9, and B1.11 show examples of this type of reporting. This presentation gives the user a quick visual assessment of the extent and location of delamination conditions, and is particularly suitable for project-level evaluation. It provides location but not necessarily depth of delaminated areas. Depending on the scale of the plot, this representation is suitable for project lengths of 1 to 10 miles. 3. Line plot of delamination indicators Figure B1.13 shows an example of a line plot that presents a delamination indicator, such as activity index, versus milepost. In this presentation, the areas where the activity index exceeds 1.0 are designated as areas where delamination is likely, and these are shaded blue. This presentation presents a concise summary over many miles, and can be useful at the network level to identify areas for future investigation. B19
Figure B1.13. Line plot showing delamination indicators versus milepost (blue fill = delamination). 4. Tabular summary of delamination indicators At the network level, a tabular summary of delamination conditions would also be appropriate for pavement management applications. This data can be imported into the pavement management system along with other variables for each pavement segment and can be incorporated into the rehabilitation planning process. A sample of such a tabular summary is shown in Table B1.2. 159 160 161 162 163 164 165 166 167 168 169 0.5 1.0 1.5 Ac tiv ity 169 170 171 172 173 174 175 176 177 178 179 180 181 182 Milepost 0.5 1.0 1.5 A ct iv ity B20
Table B1.2. Example Tabular Output at the Network Level MP Likelihood of Delamination (%) 159 100 159.1 100 159.2 94 159.3 92 159.4 100 159.5 94 159.6 92 159.7 100 159.8 93 159.9 72 160 55 160.1 47 160.2 44 160.3 52 160.4 31 160.5 48 160.6 59 160.7 100 160.8 100 160.9 100 Reference Scullion, T., and E. H. Rmeili. Detecting Stripping in Asphalt Concrete Layers Using Ground-Penetrating Radar, Research Report 2964-S. Texas Transportation Institute, College Station, TX, 1997. B21
APPENDIX B2 GPR Vendorsâ Features 3d-Radar .................................................................................................................................B23 Geophysical Survey Systems, Inc. .........................................................................................B28 IDS .........................................................................................................................................B33 B22
3d-Radar High Speed, Full Coverage Ground-Penetrating Radar Solutions (prepared by 3d-Radar, February 2013) 3d-Radar provides ground penetrating radar hardware and software for data collection across an entire swath width at high speeds with high resolution at all resolvable depths. Below are the key components of the system. 1. Equipment â Hardware GeoScope⢠Mk IV The Geoscope Mk IV generates the waveform and sends this continuous electromagnetic wave to the antenna for transmission and measures the return. It enables high-density high speed 3-D data capture with a combination of deep subsurface penetration coupled to high resolution. GeoScope Mk IV Features ⢠Resolution at all depths: Step-frequency technology enables the users to achieve good resolution at each investigation depth. Penetration and high resolution are simultaneously achieved with one single antenna array. No need to employ different frequency antennas to adapt to different depths. ⢠Area survey speed (work rate): Very high scan rates and an efficient sampling method enable the GeoScope Mk IV to provide full resolution 3-D imagery at highway speeds. ⢠High resolution 3-D sub surface imagery: 7.5 cm channel spacing in the antenna array combined with 3 GHz bandwidth enables high-density sampling as required by utility mapping, military applications, and archaeology prospecting. ⢠Wide range of antenna arrays with uniform response across the elements: The GeoScope Mk IV is compatible with all 3d-Radar VX and DX Series antenna arrays ranging up to 330 cm in width. B23
The GeoScope Mk IV connects to any laptop via a standard GBit Ethernet connection and the primary interface is via a web browser. No software needs to be installed on the host computer to communicate with the system. Via an impedance matched antenna cable, the GeoScope Mk IV connects to 3d-Radar antennas to complete the communications path between the user and collected GPR data. DX Series Antenna 3d-Radar DX Series antenna arrays allow scanning of up to 41 channels of GPR data over a continuous 200 MHz to 3 GHz frequency range. Near surface signal fidelity with the DX antenna is enhanced by orienting the sending and receiving antenna elements in opposite directions, minimizing antenna ringing while delivering high resolution imagery in collected GPR data. The air-coupled antenna design offers clear impulse response with low ringing and high suppression of the direct wave from transmitter to receiver. Operated with the GeoScope Mk IV step-frequency radar, the DX Series antennas are capable of collecting 3-dimensional GPR data with dense line spacing, allowing 3-dimensional data processing. The antenna has support for multi-offset recording and CMP, and has a built in GPS receiver. DX1821 antenna, trailer mounted. The DX Series of antennas can be used for applications such as road/bridge deck inspection, utility mapping, archaeology, railway ballast inspection and military uses, the DX Series of antennas are available in 1.8, 2.1, 2.4, and 3.3 meter widths. Accessories While it is recommended that the antenna is mounted semi-permanently to the front or rear of the survey vehicle, a trailer with an integrated DMI (distance measurement instrument) is available for lower speed data collection. B24
Equipment â Mounting and Climate Restrictions 3d-Radar step frequency antennas require no warm-up or settling time in cold weather. Operational in both subfreezing and hot climates, the antennas are sealed from the elements and can be exposed to snow and rain without damaging the electronics. The GeoScope Mk IV should be kept dry at all times. If at all possible the instrument should be housed and operated in an environment no colder than 32°F (0°C). 2. Software â Data Processing 3dr-Examiner 3dr-Examiner is a software application that enables users to process, analyze, and inspect data from 3d-Radar ground penetrating radar systems. 3dr-Examiner is configured to handle large amounts of high resolution data on normal PCs. High resolution radar surveys result in large amounts of data, which can make them difficult to process, navigate, and analyze. 3dr-Examiner includes techniques and tools to utilize the speed of modern PCs. The software is able to process radar data as fast as it can be collected, enabling surveyors to analyze in 3-D an area while still in the field. 3dr-ExaminerâOverview mode. B25
3dr-Examiner converts data collected from the frequency to time domain, applies filtering and tuning and can migrate in a self-contained, menu-driven environment. As the data are processed, they can be viewed in three dimensions, with the ability to compare processing settings to determine how to optimize the collected data. 3. Software â Data Analysis Displayed data can be annotated with user-defined groups. Pipes, subsurface variations, and other structures can be mapped, displayed, and moved inside the data. Layers can be traced semi-automatically and displayed in 3dr-Examiner or along with annotations, exported to .kmz format for display in Google Earth, .dxf for AutoCAD, or to other user-defined formats. Annotated pipe varying in depth from 50 to 2.5 m depth. 4. Field Data Output and Displays During data collection, collected data are displayed on a controlling laptop in any web browser. As objects are seen during data collection, they can be marked for future review when post-processing in 3dr-Examiner. Because it can take advantage of quad-core processors 3dr-Examiner can be used between data collection while in the field to quickly process a subsection for QA and onsite analysis. 5. Analysis Data Output and Displays 3dr-Examiner can view any data captured with a 3d-Radar system in two modes, an Explore view and an Overview mode. Overview can be integrated with drawings or satellite imagery while the more traditional Explore view enables the data to be viewed in an inline, cross-line, or horizontal slice simultaneously. 6. Equipment Calibration and Field Validation At the factory, antennas are calibrated and this information is stored in firmware inside the antenna itself. No further calibration is necessary in the field. To insure the system is functioning properly before data collection, the system can be initially configured to capture data on time triggers instead of distance. This enables stationary data collection to ensure good data is captured before the actual survey starts. B26
7. Equipment Upgrades and Service As improvements to 3d-Radar hardware and firmware are implemented, firmware upgrades to both the GeoScope Mk IV and DX Series antennas can be downloaded from 3d-Radar over the Internet. Future Equipment Developments New antenna designs will be compatible with the GeoScope Mk IV, ensuring that continuous improvements will be available for only a fraction of the purchase price of the original system. B27
Geophysical Survey Systems, Inc. GPR System Specification (prepared by GSSI, February 2013) Geophysical Survey Systems, Inc. 12 Industrial Way Salem, NH 03079 System Type Specification: Array of multiple antenna elements lined up transverse to the direction of travel Response: The GSSI RoadScan 30 system, based on the SIR-30 Multichannel Radar Controller, can support up to 4 antennas, oriented transverse to the direction of travel. The system can be extended by adding a second SIR-30 controller (operating as a slave controller) to support up to 4 additional antennas, for a total of up to 8 antennas. Model 42000S Horn Antenna, Model SIR-30 Multichannel Radar Controller Frequency Range (impulse radar systems) Specification: center frequency of pulse > 2.0 GHz, â10 db limits: 0.5â5.0 GHz Response: The RoadScan 30 system is based on the model 42000S, an air-launched antenna that is suitable for determining asphalt thickness, delamination, and void detection. The 42000S has a center frequency of 2.0 GHz and a bandwidth of 0.6â2.8 GHz @ -10 dB. Frequency Range (frequency swept radar systems) Specification: up to 3.0 GHz Response: Not applicable B28
Lateral spacing of antenna elements Specification: <1.5 feet Response: The minimum lateral spacing for the Model 42000S antenna is 0.75 feet. Lateral coverage per pass Specification: 12 feet (full lane width) Response: Using eight 42000S antennas with a lateral spacing of 1.7 feet, RoadScan 30 will provide full lane coverage of 12 feet. Longitudinal data collection rate Specification: >2 scans per foot per antenna element Response: RoadScan 30 is capable of a longitudinal data collection rate in excess of 2 scans per foot per antenna. The scan density at 23 mph (@ 512 samples per scan) is specified at 12 scans per foot. The scan rate is not affected by the number of antenna elements employed. Travel speed during data collection Specification: >20 mph Response: RoadScan 30 can collect data at speeds in excess of 20 mph. The maximum data collection speed (@ 4 scans per foot and 512 samples per scan) is 70 mph. Travel speed during mobilization Specification: posted speed limit Response: RoadScan 30 can be mobilized at posted speed limits without requiring a mechanical adaptation to the system. Real time display Specification: B-scan for selected antenna elements Response: The RoadScan 30 system can be configured to display a B-scan in real- time. System monitoring and control Specification: from within the survey vehicle Response: The RoadScan 30 system provides monitoring and control from within the survey vehicle via a monitor and keyboard, or laptop computer. Data collection rate Specification: data collection should be triggered on distance using a DMI Response: The RoadScan 30 system is configured with a standard DMI used to trigger data collection based on traversed distance. Spatial reference Specification: vehicle DMI and/or GPS B29
Response: The RoadScan 30 system supports a vehicle-mounted DMI and/or a GPS for spatial referencing. Detection depth range Specification: 2â12 inches Response: The RoadScan 30 system, configured with the 42000S 2 GHz air-launched antenna, has a detection depth range from less than 1 inch up to 20 inches, depending on the dielectric properties of the medium under test. An advantage to RoadScan 30 is that the air-launched antenna calibration is performed once prior to data collection (using a metal plate), eliminating the need for repeated coring to determine the radar propagation velocity (and by computation, the depth) in the medium under test. Additional Information System Hardware, Environmental Specifications RoadScan 30 consists of the following: - SIR-30 Multichannel Radar Controller w/transit case - Model 42000S Horn Antenna - Wheel-mounted Distance Measuring Instrument (DMI) - 7-meter Control Cable - SIR-30 Mounting Kit - AC Adaptor - User Manual Additional antennas, control cables, and radar controllers can be added to the base RoadScan 30 system. Operating Temperature: â10oC to 50oC Relative Humidity: <95% noncondensing Mounting Hardware The model FGVFHM-SL is a permanent multiple antenna mount that can support up to 3 antenna arm mount assemblies. The FGVFHM-SL is permanently mounted to the data collection vehicle. One model FGVFHM-SARM antenna arm mount assembly is required for each 42000S horn antenna deployed. System Software, Data Processing/Analysis The system software embedded in the SIR-30 mainframe provides full instrument control: data acquisition, positioning, and data storage and output. Data is stored on an internal solid state drive in 32-bit RADAN .dzt format, compatible with GSSIâs RADAN 7 post processing and data analysis package. The SIR-30 system software provides data acquisition control over ⢠Scan rate; ⢠Number of samples/scan; and ⢠Real-time filters. B30
The SIR-30 system software is compatible with all GSSI antennas and provides automatic recognition and setup of GSSI Smart antennas. RADAN 7 is a data processing and analysis package designed for use with GSSI 32-bit data acquisition systems. RADAN 7 is modular and allows users to add functionality based on a specific application. Configured for operation under Windows 7, RADAN 7 provides GPR post-processing functionality under a Windows-based user interface. The RoadScan Module is optimized for road asphalt layer and bridge deck structure and deterioration assessment and is capable of semi-automated layer picking and target mapping to the basic RADAN 7 module. The following dataset was collected with RoadScan 30 and two antennas: a 42000S 2GHz horn antenna and a 50400S 400 MHz antenna. Processed in RADAN 7, the data depict asphalt overlay and sub-base layer information. Data Example The following dataset was collected using RoadScan 30 with four 42000S antennas and processed using RADAN 7. The data depict asphalt layer information across a full lane. B31
Calibration, Field Validation The only calibration required by RoadScan 30 is the metal plate calibration used to determine the radar propagation velocity of the medium under test. This calibration is performed daily prior to data collection. There is no other calibration required by the system. System Upgrades, Service Should service be required, the system can be returned to the GSSI Factory Service Center, located in Salem, New Hampshire. As a matter of routine, hardware and/or firmware updates that address known performance issues are completed when a system is returned to GSSI for repair or for preventative maintenance. B32
