Vehicle-Mounted Bridge Deck Scanner

Transcription

1 Highway IDEA Program Vehicle-Mounted Bridge Deck Scanner Final Report for Highway IDEA Project 132 Prepared by: Yajai Tinkey, Larry D. Olson, Olson Engineering, Inc. August 2010

2 INNOVATIONS DESERVING EXPLORATORY ANALYSIS (IDEA) PROGRAMS MANAGED BY THE TRANSPORTATION RESEARCH BOARD (TRB) This NCHRP-IDEA investigation was by Research & Technology Corp. completed as part of the National Cooperative Highway Research Program (NCHRP). The NCHRP-IDEA program is one of the three IDEA programs managed by the Transportation Research Board (TRB) to foster innovations in highway and intermodal surface transportation systems. The other two IDEA program areas are TRANSIT-IDEA, which focuses on products and results for transit practice, in support of the Transit Cooperative Research Program (TCRP), and ITS-IDEA, which focuses on products and results for the development and deployment of intelligent transportation systems (ITS), in support of the U.S. Department of Transportation s national ITS program plan. The three IDEA program areas are integrated to achieve the development and testing of nontraditional and innovative concepts, methods, and technologies, including conversion technologies from the defense, aerospace, computer, and communication sectors that are new to highway, transit, intelligent, and intermodal surface transportation systems. For information on the IDEA Program contact IDEA Program, Transportation Research Board, th Street, N.W., Washington, D.C (phone: 202/ , fax: 202/ , The project that is the subject of this contractor-authored report was a part of the Innovations Deserving Exploratory Analysis (IDEA) Programs, which are managed by the Transportation Research Board (TRB) with the approval of the Governing Board of the National Research Council. The members of the oversight committee that monitored the project and reviewed the report were chosen for their special competencies and with regard for appropriate balance. The views expressed in this report are those of the contractor who conducted the investigation documented in this report and do not necessarily reflect those of the Transportation Research Board, the National Research Council, or the sponsors of the IDEA Programs. This document has not been edited by TRB. The Transportation Research Board of the National Academies, the National Research Council, and the organizations that sponsor the IDEA Programs do not endorse products or manufacturers. Trade or manufacturers' names appear herein solely because they are considered essential to the object of the investigation.

3 VEHICLE-MOUNTED BRIDGE DECK SCANNER IDEA Program Final Report Sponsored by NCHRP 132 Prepared for the IDEA Program Transportation Research Board The National Academies Prepared By Principal Investigator Yajai Tinkey, Ph.D., P.E. Associate Engineer Olson Engineering, Inc. Co-Principal Investigator Larry D. Olson, P.E. President Olson Engineering, Inc. A report from Olson Engineering W 49 th Ave. Wheat Ridge, CO Phone: Fax: August 2010

4

5 Table of Contents 1.0 EXECUTIVE SUMMARY PROBLEM STATEMENT CONCEPT AND INNOVATION LITERATURE REVIEWS Non-Contact Transducers Used In Nondestructive Evaluation Microphones Laser Vibrometers Microwave Sensors Background of Nondestructive Evaluation Methods Applicable for Bridge Decks Sounding Impact Echo Spectral Analysis of Surface Waves Slab Impulse Response Rolling Contact Transducers Used In Nondestructive Evaluation INVESTIGATION APPROACH Introduction Preliminary Investigation of Non-Contact Transducers Preliminary Investigation of Non-Contact Microphones Preliminary Investigation of Laser Vibrometer Preliminary Investigation of Microwave Transducer Development of the Bridge Deck Scanner Prototype Description of Test Structures and Test Procedures Douglas Bridge in Douglas, WY st Street Bridge in Casper, WY BRIDGE DECK SCANNER HARDWARE AND SOFTWARE IMPROVEMENTS Hardware Original Hardware Design First Iteration BDS Improvements Second Iteration BDS Improvements (Current Design) Software TEST SETUP AND RESULTS FROM 1 st STREET BRIDGE (CASPER, WY) Test Setups and Results from Traditional NDE Test Methods Test Setup and Results from Sounding Using Chain Drags Test Setup and Results from Ground Penetrating Radar (GPR) Tests Test Setup and Results from Point by Point Impact Echo Tests Test Setup and Results from Infrared Thermography Test Setups and Results from the Bridge Deck Scanner Prototype Test Setup Using the BDS Prototype Findings from Impact Echo Scanning Tests from the BDS Prototype Findings from Spectral Analysis of Surface Waves Tests from the BDS Prototype Findings from Automated Acoustic Sounding with the BDS Prototype Findings from Slab Impulse Response Tests from the BDS Prototype...75

6 7.3 Comparison of Test Results CONCLUSIONS AND RECOMMENDATIONS INVESTIGATOR PROFILES REFERENCES...86

7 1.0 EXECUTIVE SUMMARY The objective of the research project was to develop a Bridge Deck Scanner (BDS) that can be mounted behind a vehicle for comprehensive condition evaluation of concrete bridge decks with nondestructive evaluation methods (Impact Echo-IE, Slab Impulse Response-SIR, Spectral Analysis of Surface Waves-SASW and Acoustic Sounding-AS). In addition, the research explored and compared ground contact transducers to non-contact transducers for a vehicle mounted scanning system. The non-contact transducers explored in this research project include microphone, laser vibrometer and microwave transducer. The results from this research are to provide information on top/bottom delamination, internal cracks, vertical crack depths, thickness profile, and the stiffness of the bridge deck. Contacting vs. Non-Contacting Transducers. Such non-contacting transducers as a laser displacement vibrometer, microwave velocity transducer, directional and non-directional microphones were compared with contacting displacement, velocity (geophone) and accelerometer transducers for the above nondestructive test methods. The non-directional microphones were found to have the most potential application for leaky Lamb surface waves and impact echo and for acoustic sounding. However, at this time, the contacting transducers were determined to be more robust for use in the IE, SASW and SIR tests. Problems with rolling noise limited the use of the laser displacement vibrometer for IE tests and sensitivity of the microwave velocity transducer was found to be poor for SIR tests. Prototype BDS Unit. A prototype BDS unit was developed for this research project as shown in Photos 1 and 2 below. The prototype BDS is composed of one unit with two transducer wheels connected by an axle and an automatic nail gun impulse hammer. Each transducer wheel is identical and has six built-in displacement transducers and six automatic solenoid impactors. The Impact Echo test can be performed from either of the two transducer wheels. The Spectral Analysis of Surface Waves test uses both transducer wheels in a synchronized fashion. The non-contact microphone mounted near the transducer wheel (the one with the active impactor) is used to listen to shallow delaminations. Note that all three tests (Impact Echo, Spectral Analysis of Surface Waves and Automated Acoustic Sounding) are performed simultaneously (see Photo 1). The Slab Impulse Response test is performed using an automatic nail gun to drive a 3lb impulse hammer mounted on a separate frame and a non-contact geophone mounted to the axle between the two transducer wheels (see Photo 2). Overview of Field BDS Bridge Deck Test Program. The first BDS prototype was used on the Douglas bridge located in Douglas, Wyoming. The bridge deck of the Douglas Bridge was a silica fume overlay on a concrete deck. It was not the objective to perform a full investigation of this concrete bridge deck, but rather to initially test the prototype bridge deck scanner system. BDS performance feedback from the experiment on the Douglas Bridge were used to improve both the hardware and software of the BDS prototype. After the hardware modifications were completed, the BDS prototype was used on the concrete deck of the 1 st Street Bridge in Caster, Wyoming at slow rolling speeds of 1 to 1.5 mph maximum. The 1st Street Bridge Investigation was conducted as a full investigation of the two east-bound lanes of the bridge. Test runs for IE, SASW and AS were performed the full length of the bridge every 0.5 ft along the bridge length at 1 foot transverse 1

8 spacings to provide for these tests every 0.5 sq ft of the bridge. The SIR tests were conducted every 3 ft along the length of the bridge at 1 ft transverse spacings to provide a test every 3 sq ft. This extensive day of real world field testing again led to several hardware improvements and a better understanding of the BDS system. Discussion of Bridge Deck Scanner Results from 1 st Street Bridge. The results from the Impact Echo tests every 0.5 ft showed areas with top and bottom delaminations with excellent precision. The BDS IE results showed good agreement with the previous results from acoustic sounding by chain dragging and Ground Penetrating Radar methods with less correlation with Infrared Thermography tests for shallow delaminations by the University of Wyoming. In addition, the results from the IE tests were able to determine thinner sections and bottom delaminations of the bridge deck versus AS or GPR. The test results from the SASW tests indicated concrete quality was good, but were not so applicable to the 1 st Street Bridge since the bridge deck is a one layer system with no significant freeze thaw cracking damage. The data obtained from the Slab Impulse Response tests with the BDS unit were poor due to deck coupling/vibration problems between the impulse hammer deck impact and geophone (on the axle), and vibrations as a result of rolling. The Acoustic Sounding tests did detect delaminations with the microphone as well, but the Impact Echo tests also provided more information on the deeper concrete deck conditions. In comparing nondestructive testing results from all methods used on the 1 st street deck, the BDS Impact Echo tests provided the most detail on bridge deck concrete conditions in terms of top/bottom delaminations in comparison to Ground Penetrating Radar, point by point Impact Echo, chain drag Acoustic Sounding and Infrared Thermography test results as presented in Section 7 herein. In addition, BDS Surface Wave and Acoustic Sounding tests were found to provide useful information on the bridge deck conditions. Bridge Deck Scanner (BDS) Status. As the research team wrapped up the project, the hardware and software of the BDS system has continued to be improved; in particular the Slab Impulse Response components have been improved. The BDS unit has been used for a project demonstration for the SHRP II R06 research project on detection of debonded hot mix asphalt pavement layers being conducted by Dr. Michael Heitzman of the National Center for Asphalt Technology in Auburn, Alabama. Consulting projects are also being discussed and proposed for evaluation of bridge and parking deck conditions. 2

9 Photo 1: Bridge Deck Scanner (BDS) Test Setup for Impact Echo (IE), Spectral Analysis of Surface Waves (SASW) and Automated Acoustic Sounding (AS) on the 1 st Street Bridge over the North Platte River in Casper, Wyoming Photo 2: Bridge Deck Scanner Test Setup for Slab Impulse Response Tests on the 1 st Street Bridge over the North Platte River in Casper, Wyoming 3

10 2.0 PROBLEM STATEMENT Most of the reinforced concrete bridges in the nation were built between 1955 and 1970 (Concrete Society 1996). After 1970, the proportion of prestressed concrete bridges has been increasing steadily (Concrete Society 1996). As traffic flow increases and heavier truckloads are permitted, older bridges can become deficient. In addition, environmental attacks including freezethaw degradation and intrusion of chloride ions from deicing salts can cause active corrosion of reinforcing. Cracks which can be caused by shrinkage, poor curing, moisture and temperature changes and loading, provide numerous open pathways for water and deicing salt to infiltrate the concrete bridge deck (Woodward et al 1988). Further, the porous microstructure of the cement and aggregate provides additional avenues through which water and chemicals migrate into uncracked concrete initiating the cracking process, typically due to reinforcing steel corrosion and/or freezethaw cracking damage. Although current concrete mix designs and components are much more resistant to the forces of deterioration than older concrete, there are still problems with older bridges (Woodward et al 1988). Chase and Washer showed that there were more than 19,000 structurally deficient concrete bridges in the US in 1997 and the most serious types of deterioration include decks, superstructure or substructure (Concrete Society 1996). Corrosion of reinforcement leading to concrete deck delaminations is a major maintenance repair/replacement cost for state DOT s and accurate mapping of top and bottom delaminations is needed for repair/replacement decisions. The Federal Highway Administration (FHWA) requires all bridges to be inspected at least every two years (Woodward et al 1988). The inspection of concrete bridge decks typically includes a delamination survey (with chain dragging for acoustic sounding that detects top rebar delaminations only not deck bottom delaminations), chloride sampling and core sampling. The drilled cores can be used to determine the soundness, strength and thickness of existing deck concrete. This research focused on the development of technologies for rapid inspection that can provide the following information about the bridge deck: 1. Top delamination mapping 2. Internal conditions; including cracks, crack depth, concrete deterioration and bottom deck delamination mapping 3. Thickness profiling 4. Stiffness/structural integrity of the deck 4

11 Although the proposed techniques do not provide information about the chloride content in a bridge deck, they do provide critical structural integrity data such as information on both top and bottom delaminations as well as cracks and crack depth/severity. For example, the Automated Sounding (Acoustic) and Impact Echo tests provide information on top delamination. The Impact Echo test also provides additional information on the existence of cracks parallel to the testing surface, bottom delamination and the thickness profile. Cracks perpendicular to the testing surface can be detected and the depth can be measured with the Spectral Analysis of Surface Waves technique. Last, the Slab Impulse Response test provides the stiffness of the deck. The current practice (using acoustic sounding, visual inspection or Ground Penetrating Radar) is not able to provide information on bottom delaminations and the internal condition of the bridge deck without destructive coring of the concrete deck. The prototype BDS system will save time and cost by minimizing the need for coring and accurately map deck areas in need of repair/replacement, thus improving safety for the public. 5

12 3.0 CONCEPT AND INNOVATION This research project proposed to develop an effective and reliable system using nondestructive evaluation methods (stress waves and acoustic) to quickly and simultaneously determine the concrete bridge deck thickness profile, stiffness, and internal condition of the deck including top and bottom delamination, crack locations, crack depths and deterioration of the concrete deck. This device is attached behind a vehicle so a controlled rapid survey can be undertaken in a continuously rolling fashion. In addition to microphone based acoustic sounding, the stress wave techniques include Impact Echo (IE), Spectral Analysis of Surface Waves (SASW) and Slab Impulse Response (SIR - sometimes called Impulse Response). Multiple channels of non-contact transducers are also used as receivers for the NDE tests. The non-contact transducers used in this prototype include airborne microphones. The Ground Penetrating Radar (GPR) method has been extensively researched and developed for pavement or bridge deck thickness surveys (Maser et al 1990, Azevedo et al 1996, Davidson et al, 1998, and Mast 1993). GPR systems are commercially available that can be used to determine pavement layer thickness and base and sub-base evaluations. The GPR surveys can determine the top delamination of the concrete bridge deck (Romero et al 2009 and Parrillo et al 2009) and GPR surveys were done for comparison purposes in this research as reported herein However, the GPR test is heavily dependent on a pre-select threshold to determine the areas with shallow delamination which can be subjective. Recent research has shown that the use of both GPR and IE methods can be complementary for condition assessment of bridge decks (Gucunski et al 2009). The result of the research project is the first product that provides a complete scanning of bridge decks including mapping the thickness profile, evaluation of the stiffness and the internal condition of the bridge deck (top and bottom delaminations, internal cracks and general concrete deterioration). This is the first time that all four NDE techniques have been combined in the same system and performed simultaneously. Results from the IE test provide a thickness profile of the bridge deck (Sansalone et al 1997). In addition, the IE test can detect top and bottom delaminations, location of cracks and general deterioration of concrete (Sansalone et al 1997). Results from the SASW test provide surface wave velocity that can be used to theoretically 6

13 calculate the compressional wave velocity used to calibrate the Impact Echo test. The SASW test can also detect cracks perpendicular to the surface of the bridge deck and evaluate the crack depth (Kalinski 2004). Most importantly, results from the SASW test provide the depth of concrete deterioration in the bridge deck (Kalinski 2004). The Slab Impulse Response (SIR) test can be used in the evaluation of concrete conditions to provide secondary information from the IE and SASW tests (Davis et al 2003). In addition, information from the SIR test can be used to determine the dynamic stiffness of the bridge deck (Davis et al 2003). Non-contact microphones are used to listen to the hollow sound for shallow delamination detection. Data from all the three NDE tests plus information from the automated sounding with a microphone will not only compliment each other but also still provide redundancy to increase the confidence level of the data interpretation. Excluding the information on the chloride content of the bridge deck, the results from the proposed technologies provide comprehensive information that typical routine bridge inspections acquire on a bridge deck. 7

14 4.0 LITERATURE REVIEWS 4.1 Non-Contact Transducers Used In Nondestructive Evaluation Several types of non-contact transducers were studied throughout the research presented herein. Non-contact transducers are of significant interest because they may allow the test methods to be performed more rapidly, which would allow greater speeds of a vehicle mounted bridge deck scanner. Non-contact transducers may also eliminate noise sources associated with rolling wheels and other contact points. Based upon the research team s experience and extensive knowledge of the test methods and governing wave mechanics, as well as knowledge gained from discussions with other researchers, it was determined that the most promising non-contact transducers to pursue were microphones, laser vibrometers, and microwave transducers. Below is a review of the current literature available discussing the implementation of these non-contact transducers in measuring vibrations similar to those inherent in the proposed test methods Microphones The physical basis of utilizing non-contact microphones to measure surface waves is a phenomenon known as Leaky Lamb Waves (LLW). This phenomenon is essentially the coupling of wave energy from the surface of the excited medium (in our case concrete) into the fluid in contact with that surface (in our case air). A detailed discussion of the LLW phenomenon as well as information regarding the development of the method can be found in Bar-Cohen et al (2001) and in Holland and Chimenti (2003). Since the method s development for use on thin composite materials with high frequency excitation and response, several researchers have applied the same principles in performing both the Surface Wave and similarly the Impact Echo test methods on concrete slabs using non-contacting microphone receivers. It is these recent studies pertaining to Surface Wave and Impact Echo testing that are most pertinent to our research investigation. Note that noncontacting excitation of the concrete surface has been unsuccessful in past studies (Cetrangolo and Popovics 2006) but is of little concern due to the relative ease of employing contacting solenoid impacts for excitation. In 2001 Zhu and Popovics implemented air-coupled surface wave testing using directional microphone receivers to detect the LLW from the concrete surface. This study was supported by additional studies by Zhu (2005), Zhu and Popovics (2005), as well as a study by Ryden et al (2006) 8

15 in which non-directional audio microphones, which are much less expensive than directional microphones, were utilized in surface wave testing. Ryden et al (2006) mentions external noise sources such as wind noise but reports good results. There is also concern of interference from the direct air wave arrival from the impact source, however for surface wave testing the distance between the impact source and receiver can be made large enough that the two arrivals occur at significantly different times due to the differences in velocities of the air born p-wave and LLW on the concrete surface (Zhu and Popovics 2007). Digital signal processing techniques such as windowing the wave arrivals with exponential decay or Hanning windows are often performed during data analysis to eliminate any effects of unwanted wave arrivals. Because of the need for separation of the LLW and direct air wave arrivals it is advantageous for the microphone receiver to be located as near the concrete test surface as possible. Multiple studies have also been conducted in which non-contacting microphones were used to perform impact echo testing. Non-contact impact echo testing has proved to be more difficult than non-contact surface wave testing (Zhu and Popovics 2007) because the separation of the impact source and microphone receiver is much less than in surface wave testing, which leads to interference from the direct air wave. The spacing between the receiver and impact source is critical in impact echo testing because the excitability of the S1 Lamb wave mode, which is the impact echo resonance in a slab type structure (Gibson and Popovics 2005), decreases drastically as the source-receiver spacing increases (Gibson 2005). An additional complication is the need for a longer time signal in the impact echo test to determine the resonance, whereas often times in surface wave testing only the first arrival (1 wave cycle) is considered, thus enabling sharp windowing functions to remove unwanted direct air wave arrivals. Zhu and Popovics (2007) demonstrated that sound insulation material can be used for shielding purposes to encapsulate (open on one end) the microphone receiver and reduce the direct air wave energy detected by it Laser Vibrometers Another important emerging technology in the field of non-contacting vibration measurements is the laser vibrometer. Laser vibrometers are used extensively in the automotive, aerospace and other manufacturing fields. The laser vibrometer measures vibration using the Doppler shift effect. Laser vibrometers generally have a wide frequency range, excellent vibration resolution and are well suited to indoor laboratory testing. The ability of the laser vibrometer to 9

16 measure high frequencies (> 100 MHz) has made it the ideal receiver to measure surface waves in thin ceramic and metal materials (Somekh et al 1995). This testing is often conducted to determine material strength and locate defects or anomalies within the material, very similar to testing at lower frequencies on concrete specimens. Some models have been ruggedized and made more portable to allow for field testing situations. The primary drawback of laser vibrometers is the cost, which typically ranges from $10,000 $50,000 for a single receiver. Due to the cost, few studies have been conducted to date in which a laser vibrometer was implemented for surface wave testing on concrete. Abraham et al 2009 performed a successful study in which an extreme number of repetitive surface wave tests were performed on a variety of concrete samples using a laser vibrometer receiver mounted to a semiautonomous robot. The laser vibrometer has also been successfully implemented as a receiver for impact echo testing on concrete structures (Abe et al 2001; Algernon et al 2008). The laser vibrometer has been shown to produce high quality impact echo data and is fairly easy to implement. Because the device relies on the Doppler shift effect of the vibrating surface and not an air coupled wave, proximity to the impact source and shielding of direct air waves are not of concern Microwave Sensors The microwave transducer has also been pursued as a possible non-contacting receiver for structural vibrations. Based upon our understanding of the sensor as well as discussions with other researchers, the sensor is not applicable to the relatively high frequency vibrations found in Impact Echo and Surface Wave testing. However, it is possible that the microwave transducer may be implemented in Slab Impulse Response (SIR) testing in which the frequency range of interest is primarily less than 500 Hz. The current SIR method involves holding a geophone in contact with the concrete structure while the structure is impacted with an instrumented hammer. The geophone measures the transient vibration induced in the concrete slab. Recently multiple research studies have used microwave interferometers to measure movements of large scale structures such as bridges and buildings (Bernardini et al 2007; Farrar et al). These systems, which are commercially available, have typical maximum sampling frequencies from static to 100 Hz (Bernardini et al 2007) to 200 Hz. A separate research study also showed that microwave transducers can be used to measure transient seismic vibrations of the ground (Wijk et al 2005). In the Wijk et al (2005) study, a sledge hammer impacting a steel plate was used to excite the seismic vibrations in the ground 10

17 while a microwave transducer was suspended nearby to receive the vibration signals. The research study involved averaging 32 separate impacts at a single test location to improve the signal to noise ratio. 4.2 Background of Nondestructive Evaluation Methods Applicable for Bridge Decks Sounding Chain dragging or hammer sounding, where either a heavy chain(s) is literally dragged across a bare concrete deck, or a rock-hammer or similarly designed hammer is used to repeatedly strike its surface, are two common acoustic sounding methods widely used to determine areas with shallow surface delamination in bare concrete bridge decks. Common chain configurations consist of four or five segments of 1 in. links of chain that are approximately 18 in. long (ASTM D ). Distinctive hollow sounds produced by the chain drags or hammer impacts are indicative of shallow delaminations. Other investigators have connected the chain drag apparatus to a microphone in an attempt to standardize and automate the evaluation (Henderson et al, 1999). Although chain drags or hammer sounding are simple to perform, most of the damage mapping is at the discretion of the operator due to different levels of experience and hearing among operators. In addition, delamination located deeper than 3 to 4 inches from the surface is hard to determine by acoustic sounds (hollow and drummy due to flexural resonant vibrations of the shallow, horizontally cracked concrete due to steel rebar expansion as a result of corrosion) Impact Echo The IE method involves hitting the concrete surface with a small impactor (or impulse hammer) and identifying the reflected wave energy with a displacement (or accelerometer) receiver mounted on the surface near the impact point (ASTM C ). A simplified diagram of the method is presented in Figure 1. 11

18 Olson Instruments, Inc. handheld test head for Impact Echo tests Receiver Impact Flaw Reflection from concrete/flaw interface Reflection from backside of test member *Reflection from backside occurs at a lower frequency than that from the shallower concrete/flaw interface Figure 1 Schematic of Impact Echo (IE) method. Following the impact, the resulting displacement or acceleration response of the receiver is recorded. The resonant echoes are usually not apparent in the time domain. The resonant echoes are more easily identified in the frequency domain (linear displacement spectrum). Consequently, the time domain test data are processed with a Fast Fourier Transform (FFT) which allows identification of frequency peaks (echoes). The displacement spectrum of the receiver or the transfer function (receiver displacement output/hammer force input vs. frequency) are used to determine the resonant peaks. If the thickness of a slab is known, the compression wave velocity (V p ) can be determined by the following equation: V p = 2*d*f/β (1) where d = slab thickness, f = resonant frequency peak. The above equation is modified by a β (Beta) factor of 0.96 for concrete walls and slabs (Sansalone et al 1997 and per the ASTM standard). 12

19 4.2.3 Spectral Analysis of Surface Waves The SASW method uses the dispersive characteristics of surface waves to determine the variation of the surface wave velocity (stiffness) of layered systems with depth (M.F. Aouad 1993). The SASW testing is applied from the surface which makes the method nondestructive and nonintrusive. Shear wave velocity profiles can be determined from the experimental dispersion curves (surface wave velocity versus wavelength) obtained from SASW measurements through a process called forward modeling (an iterative inversion process to match experimental and theoretical results). The SASW method can be performed on any material provided an accessible surface is available for receiver mounting and impacting. Materials that can be tested with the SASW method include concrete, asphalt, soil, rock, masonry, and wood. Applications of the SASW method include, but are not limited to: 1) determination of pavement system profiles including the surface layer, base and subgrade materials, 2) determination of seismic velocity profiles needed for dynamic loading analysis, 3) determination of abutment depths of bridge substructure, and 4) condition assessment of structural concrete. For bridge decks, the SASW method can be used to check for deteriorated zones in concrete such as cracking from freeze-thaw, alkali-silica/aggregate reactions (ASR/AAR) and fire damage. SASW can also measure crack depths (for cracks perpendicular to the surface) in bridge decks. The SASW method uses the dispersive characteristics of surface waves to evaluate concrete integrity with increasing wavelength (depth). High frequency or short wavelength waves penetrate through shallow depths, and low frequency or long wavelength waves penetrate through deeper depths. Open, unfilled cracks will result in slower surface wave velocities. Weak, fire damaged and poor quality concrete also produce slower surface wave velocities. It should be understood that if a crack is in tight grain-to-grain contact then the SASW dispersion curve will show minimal effect from the crack. This is because the surface wave energy will propagate across a tight crack that is under stress. 13

20 4.2.4 Slab Impulse Response Slab Impulse Response (Slab IR) investigations are performed primarily to identify subgrade voids below slabs-on-grade. The method is applicable for evaluating the repair of slab subgrade support conditions by comparing the support conditions before and after repairs. The elements that can be tested include concrete slabs, pavements, runways, spillways, pond and pool bottoms, and tunnel liners. The Slab IR method is often used in conjunction with Ground Penetrating Radar for subgrade void detection and mapping. In addition, the Slab IR test method can be used on other concrete structures to quickly locate areas of delamination or void in the concrete, if the damage is relatively shallow. Slab IR can be performed on reinforced and non-reinforced concrete slabs as well as asphalt or asphalt-overlaid slabs. 4.3 Rolling Contact Transducers Used In Nondestructive Evaluation The only rolling contact transducers used commercially in non-destructive evaluation is the rolling displacement transducer for Impact Echo Scanning. The rolling Impact Echo Scanner (IES) was first conceived by Mr. Larry Olson and researched and developed as a part of a US Bureau of Reclamation prestressed concrete cylinder pipe integrity research project (Sack and Olson, 1995). This technique is based on the impact-echo method (Sansalone and Streett, 1997; ASTM C1338(2004)). In general, the purpose of the impact-echo test is usually to either locate delaminations, honeycombing or cracks parallel to the surface or to measure the thickness of the structures (concrete beams, floors or walls). To expedite the impact-echo testing process, an impact-echo scanning device has been developed with a rolling transducer assembly incorporating multiple sensors, attached underneath the test unit. When the test unit is rolled across the testing surface, an optocoupler on the central wheel keeps track of the distance traveled. This unit is calibrated to impact and record data at intervals of nominally 25 mm (1 in.). If the concrete surface is smooth, a coupling agent between the rolling transducer and test specimen is not required. However, if the concrete surface is rough, water can be used as a liquid couplant. A comparison of the impact-echo scanner and the point by point impact-echo unit is shown in Figure 2. Typical scanning time for a line of 157 in (4 m), approximately 150 points, is 60 s. In an impact-echo scanning line, the resolution of the scanning is about 1 inch (25.4 mm) between impact points. Data analysis and visualization is achieved using impact-echo scanning software 14

21 developed by Dr. Yajai Tinkey for a National Cooperative Highway Research Program Innovations Deserving Exploratory Analysis (NCHRP-IDEA) grant for stress wave scanning of post-tensioned bridges (Tinkey and Olson, 2007). Raw data in the frequency domain were first filtered using a Butterworth filter with a high-pass frequency range of 1-5 khz and a low-pass frequency of typically 20 khz depending on the range of frequencies (inversely related to thickness echo depth) of interest. Automatic and manual picks of dominant frequency are performed on each data spectrum and an impact-echo thickness is calculated based on the selected dominant frequency. A thickness surface plot (skewed 3-D view of X-Y distance and thickness echo depths) of the condition of the scanned element is then generated by combining the calculated impact-echo thicknesses from each scanning line. Figure 2 Impact Echo Scanning Unit and Point by Point Impact Echo Unit 15

22 5.0 INVESTIGATION APPROACH 5.1 Introduction The objective of the research was to develop an effective and reliable system using nondestructive evaluation methods (stress waves and acoustic) to quickly and simultaneously determine the concrete bridge deck thickness profile, stiffness, and internal condition of the deck including top and bottom delamination, crack locations, crack depths and deterioration of concrete bridge decks. The product(s) from this research is to be used as a tool for inspection and non-destructive evaluation (NDE) of concrete bridge decks. The first stage of the research project included a comprehensive study of potential non-contact transducers and rolling contact transducers. The second stage of the research project entailed research and development of the BDS prototype hardware and software. Field experiments using the prototype BDS were conducted on two bridge decks in Wyoming. The tested bridges are referred to herein as the Douglas Bridge in Douglas, WY and the 1 st Street Bridge in Casper, WY. Feedback from the first BDS field experiments on the Douglas Bridge were used to improve the hardware and software. Then the BDS prototype was used on the second tested bridge, the 1 st Street Bridge, for a thorough inspection of the concrete bridge deck. Note that other traditional NDE tests were also conducted on the 1 st Street Bridge. These NDE tests were conducted as part of a Wyoming DOT bridge deck NDE assessment conducted by Dr. Jennifer Tanner of the University of Wyoming and included the following organizations and methods: 1. ground penetrating radar (GPR) with contact and airborne horn antenna performed by the Olson Engineering research team, 2. traditional chain drag by the Wyoming DOT, and 3. Infrared Thermography (IR) and point-by-point Impact Echo tests (3 ft x 3 ft grid) performed by a research team from the University of Wyoming under the direction of Dr. Jennifer Tanner. The results from the traditional NDE tests and the results from the BDS prototype are presented and compared in Section 7.0 herein. 5.2 Preliminary Investigation of Non-Contact Transducers The first stage of the research began with studies of different types of non-contact transducers with potential applications for acoustic sounding (AS), IE, SASW and SIR tests. These 16

23 transducers include microphone, laser vibrometer and microwave transducers. The results and summary of the findings from the non-contact transducers are presented in this section Preliminary Investigation of Non-Contact Microphones The initial part of this study consisted in part of the exploration of non-contact Directional Microphones to be used as receivers for the AS, IE, SASW and SIR tests. This task extends the previous work of the research team at Olson Engineering in the development of the Impact Echo Scanner with a non-contact directional microphone and also followed on the recent research work from many researchers [Holland et al 2003, Gibson 2005, and Ryden et al 2006]. Between , as part of in-house research and development, the research team at Olson Engineering added a non-contact directional microphone in addition to a rolling displacement transducer for Impact Echo Scanning tests. The bottom view of the scanner (in 2003) with non-contact microphone and rolling displacement transducer for Impact Echo tests is shown in Figure 3. Rolling ground contacted displacement transducer Impactor Non-contact directional microphone Figure 3 - Bottom View of the Impact Echo Scanner with non-contact microphone and ground contact rolling displacement transducer and automated solenoid impactor for Impact Echo Scanning Tests Microphone for IE Tests Detailed studies were performed of non-contact microphones as compared with contacting displacement and accelerometer transducers in Impact Echo tests. The studies included looking at the effects of the separation distance between the source and receiver so that the direct interference airborne wave can be excluded, applying a shielding mechanism to protect the microphone 17

24 receivers from acoustic airborne noise, and assessing the best filters to be applied to minimize the effects of ambient or traffic noise [Gibson 2005 and Zhu et al 2007]. One typical laboratory setup of the preliminary experiments is shown in Figure 4. Non-contact microphone (non-directional) Impactor Accelerometer Figure 4 Test Setup to Compare the Non-contacted and Ground-contacted Sensors for Impact Echo Tests A non-contact microphone (ADK SC-1 Small Capsule Condenser Microphone with an external 48V Phantom Power supply) and a small, high frequency accelerometer were used in the comparison study. The tests were performed on a nominally 4 thick concrete slab. The microphone was mounted at various heights above the slab surface directly above the accelerometer. The studies included looking at the effects of the separation distance between the source and receiver so that the direct interference airborne wave could be excluded [after Gibson 2005]. An automatic solenoid impactor was applied on the slab in line with both sensors starting at 4 and performed every 1 away until it was located 24 away from the sensors. Time domain Impact Echo (IE) data and the spectrum (converted to depth scales) from the accelerometer and microphone (mounted 3 inches above the concrete slab) with the impactors located 4 and 12 away from the transducers are presented in Figures 5 and 6. Note that the time domain data presented in Figures 5 and 6 are filtered with a digital bandpass Butterworth filter with a range of 3 20 KHz. 18

25 Time (us) Time (us) Depth Depth Main peak at Multiple peaks due to several wave modes Figure 5a IE Data from an Accelerometer Depth (in) Figure 5b IE Data from the Microphone Figure 5 - Comparison of IE Data from Accelerometer and Microphone with the Source 4 away Depth (in) Review of Figure 5a indicates that the spectrum of the time domain IE data taken from the groundcontact accelerometer had a dominant resonant echo peak corresponding to a slab thickness of 4. However, the spectrum of the time domain IE data taken from a non-contacted microphone showed multiple peaks in Figure 5b. This is because two wave modes (actual Lamb waves and airborne waves) blended together Time (us) Depth Time (us) Depth 15 Multiple peaks 15 Multiple peaks Depth (ft) Depth (in) Depth(ft) Depth (in) Figure 6a IE Data from an Accelerometer Figure 6b IE Data from the Microphone Figure 6 - Comparison of IE Data from Accelerometer and Microphone with the Source 12 away 19

26 Review of Figure 6a indicates that multiple peaks are present in the spectrum of the time domain IE data from the accelerometer. This is mainly because the source was too far from the accelerometer [Sansalone et al 1997]. Multiple peaks were also observed in the spectrum from Figure 6b due to the fact that the spectrum was calculated from both Lamb waves and airborne waves. To eliminate the erroneous response, the airborne wave should be excluded from the analyzed data. The speed of the sound or airborne compressional wave is approximately 1,100 ft/sec and is significantly slower than the speed of Lamb waves in concrete. Consequently, further distances between the microphone and the source can separate the two wave modes. Figure 7 presents an unfiltered time domain data record which shows the time separation of the two wave modes. Therefore, if the airborne waves are excluded from the calculation of the spectrum, the erroneous response can be eliminated. The time domain IE data with the airborne waves excluded and its spectrum (in depth scales) are presented in Figure 8. Reviews of Figure 8 show a dominant response corresponding to a slab thickness resonant echo of 4. Figure 7 Unfiltered Time Domain IE Data from the Microphone with the Source 12 Away Figure 8 Microphone Time Domain IE Data with the Airborne Waves Excluded and the Spectrum 20

27 Microphone for SASW Tests For non-contact SASW tests, two non-contact microphones (ADK SC-1 Small Capsule Condenser Microphones with external 48V Phantom Power) were used in this study. The distance between the two microphones is 4 inches and a solenoid impactor was used as an impact source. The tests were performed on a 4 thick concrete slab. The microphones were mounted 3 inches above the concrete slab and the source was located between 8 and 18 inches away from the closest microphone. Figure 9 shows the un-filtered and un-windowed time domain data from the two microphones when the source was located 8 inches away. The two traces of Figure 10 present the time domain data from the two microphones with an exponential window (decay of 1000), the middle trace of Figure 10 presents the coherence of the data and the last trace of Figure 10 is a plot of the phase difference for the passage of the surface (Rayleigh) wave by the two receivers versus frequency SASW data. The surface wave velocity is calculated from the phase plot as a function of wavelength (velocity = frequency x wavelength). Figure 11 shows a uniform surface wave velocity of approximately 7,000 ft/sec from wavelengths of 0.2 to 0.4 ft and this plot is referred to as a dispersion curve. Amplitude (Volt) Figure 9 Time Domain SASW Data from a pair of Microphones 4 inches apart 21

28 Figure 10 SASW Data Processing of Figure 9 Microphone data 22

29 Figure 11 SASW Surface Wave Velocity vs. Wavelength Plot (Dispersion Curve) 23

30 5.2.2 Preliminary Investigation of Laser Vibrometer Laser Vibrometer for Stationary Impact Echo Tests A Laser Vibrometer continuously transmits and receives the signal and uses a Doppler shift of the laser to measure surface displacement vibrations. In this study, the unit was rented from Polytec, Inc. The maximum Doppler frequency that the unit can acquire is 22 khz. In this experiment, both a Laser Vibrometer and an accelerometer transducer were used as receivers. The Laser Vibrometer was attached to a tripod 40 inches above the tested concrete slab. A small Allen wrench was used as an impact source. In this case, a normal concrete velocity of 12,000 ft/sec was used to calculate the IE thickness. Figure 12 shows the Impact Echo data from the Laser Vibrometer on a 4.5 inch thick concrete slab. Figure 13 shows the Impact Echo data from the accelerometer on a nearby location. The top trace of Figures 12 and 13 is the time domain IE data and the bottom trace is the linear displacement spectrum of the time domain data. Review of Figures 12 and 13 shows that the results from both Laser Vibrometer and accelerometer are of very good quality. Figure 12 IE Results from a Non-Contact Laser Displacement Vibrometer 24

31 Figure 13 IE Results from a Ground Contact Accelerometer at the same locations as the Figure 12 test 25

32 Laser Vibrometer for Moving Impact Echo Tests Next, a test configuration was set up for movable IE tests (scanning fashion) which included the non-contact Laser Vibrometer mounted on a moving tripod (a tripod with wheels). In this setup, a Laser Vibrometer and the automated solenoid impactor from the handheld Impact Echo Scanner (see Figure 2) were used on a smooth four inch thick concrete slab. Figure 14 shows the Laser Vibrometer attached to a movable tripod (with 3 wheels) and the automated impactor (in the IE scanner) attached to the bottom frame of the tripod for the IE test. The IE scanner was attached to the frame of the tripod, therefore the IE scanner rolled at the same speed as the tripod moved. As it was rolled, the automatic solenoid impactor tapped the concrete slab ~ every inch along the scan line distance and the Laser Vibrometer constantly measured the Doppler shift that corresponded to vibration induced displacements in the slab. The IE results from the slowly and very smoothly moving Laser Vibrometer (over 2.3 ft in distance) are presented in Figure 15. Review of Figure 15 shows good quality IE data with the corresponding IE thickness of approximately 4.3 inches. Laser Vibrometer Movable Tripod An automatic solenoid impactor (within the Impact Echo Scanner) attached to the frame of the tripod Figure 14 Test Setup for Impact Echo Scanning Test using Moving Laser Vibrometer 26

33 Time Domain IE Data from Moving Laser Vibrometer Spectrum Figure 15 IE Results from a Slowly Moving Displacement Laser Vibrometer on a smooth concrete slab 27

34 Laser Vibrometer for Stationary Slab Impulse Response (SIR) Test In this study, the Laser Vibrometer was used in the SIR tests. Figure 16 shows the Slab IR test setup using a non-contact Laser Vibrometer mounted 40 inches from the slab and slab contact velocity transducer (vertical 4.5 Hz geophone) for comparison purposes. A 3 lb instrumented impulse hammer was used as a source and calibrated to measure the impact force. The Slab IR results from the Laser Displacement Vibrometer and the velocity transducer are presented in Figures 17 and 18. A comparison of data between the non-contact Laser Vibrometer in Figure 17 and the velocity transducer in Figure 18 shows good coherence of the data from the Laser Vibrometer from near zero frequency to a frequency of approximately 500 Hz. However, low frequencies from ground motion (from the impact) had an influence on the Laser Vibrometer attached to a tripod. The high amplitude of the low frequency showed that the tripod was not able to shield the vibrometer from the ground motion generated by the 3 lb impulse hammer with a hard plastic tip. Laser Vibrometer (mounted on a Tripod) Focused Red dot from Laser Vibrometer 3 lb instrumented Hammer Velocity Transducer 3 lb Instrumented Hammer Velocity Transducer Figure 16 Slab IR Test Setup Using Non-Contact Laser Vibrometer and Ground Contact Velocity Transducer 28

35 Good coherence up to ~500 Hz Due to ground moving Figure 17 Slab IR Test Result Using Non-Contact Laser Vibrometer High coherence up to ~1000 Hz Figure 18 Slab IR Test Result Using Ground-Contact Velocity Transducer 29

36 Laser Vibrometer for Moving Slab Impulse Response (SIR) Test A moving tripod is not practical for the SIR preliminary tests using a Laser Vibrometer as illustrated in Section Thus, a pulley system was attached to roof concrete twin-tee girders to provide an even smoother moving mechanism (see Figure 20). The Laser Vibrometer was attached to an aluminum rod hanging from a roof frame. While the Laser Vibrometer was slowly moved along the frame, hammer impacts were performed manually on the ground along the test line (along the roof frame). An example result from one of the SIR tests from the scan line is presented in Figure 19. Review of Figure 19 shows that the time domain data is noisy with the low frequency moving noise and some spike noises from the small jerking effect of the relatively smooth pulling. Note that the coherence of the data is 1 because there is only one SIR record at each location (scanning fashion). A better moving mechanism was thus found to be required to carry the Laser Vibrometer as the low frequency moving noise has significant impact of the SIR data quality. Figure 19 - Slab IR Test Result Using Laser Vibrometer Moving using a Pulley System 30

37 5.2.3 Preliminary Investigation of Microwave Transducer Microwave Transducer for Stationary Slab Impulse Response (SIR) Test A microwave transceiver continuously transmits and receives the signal. It uses a Doppler shift concept to measure surface vibration in velocity units. The Ka band microwave transceiver used in this study has a rectangle waveguide of 28 and a frequency range between GHz. Figure 20 shows the Slab Impulse Response (Slab IR) test setup using the non-contact microwave transceiver. The microwave transceiver was attached to an aluminum rod connecting to the wooden frame from the ceiling to minimize the effect of the slab movement due to the impulse hammer impact on the microwave transceiver as was similarly done for the laser vibrometer. The study included variation of the height of the non-contact microwave transceiver above the testing surface. Figure 21 presents the data from Slab IR tests using the microwave transceiver attached to the frame with a height of 0.25 inches above ground. The top trace of Figure 21 presents time domain Slab IR data of the transceiver vibration response to the 3 lb instrumented impulse hammer impact. The middle trace of Figure 21 presents a coherence plot (related to signal to noise ratio, a coherence value near 1 indicates good quality data and that the response is due to the impact). The bottom trace of Figure 21 presents a plot of mobility (vibration velocity amplitude per pound force) as a function of frequency measured in cycles per second or Hertz (Hz). Figure 22 shows the data from the Slab IR test using the traditional ground contact velocity transducer (vertical 4.5 Hz geophone). The comparison of data between the non-contact microwave transducer in Figure 21 and the velocity transducer in Figure 22 reveals poor coherence of the data for the microwave transceiver. The Doppler shifts from the microwave transceiver were low frequency and not adequate to acquire good quality Slab IR data. Figure 23 presents the data from a Slab IR test using the microwave transceiver attached to the frame with a height of 2 inches above ground. Review of Figure 23 shows that the quality of the time domain data and coherence drop drastically with a 1.5 inch increase of the height above ground for the transceiver. 31

38 Frame Pulley System Aluminum Rod Microwave Transceiver Figure 20 Slab IR Test Setup Using Non-Contact Microwave Transceiver with Roof Frame 32

39 Figure 21 Slab IR Test Result from the Microwave Transceiver Positioned 0.25 inch above the Slab 33

40 Figure 22 Slab IR Test Result from the Slab Contact Velocity Transducer 34

41 Figure 23 Slab IR Test Result from the Microwave Transceiver Positioned 2 inches above the Slab 35

42 5.3 Development of the Bridge Deck Scanner Prototype The design and development of the Bridge Deck Scanner prototype involved the fusion of knowledge gained from our literature review, discussions with other researchers, our extensive prior experience with the test methods and equipment, preliminary investigations with non-contact transducers as well as significant mechanical and electrical research and development. Because of our mixed results with the non-contacting transducers, it was considered critical that our early prototype incorporate both contacting transducers as well as non-contacting transducers where applicable. Olson Instruments has had excellent success with the Impact Echo Scanner, which was designed to perform impact echo testing at 1 inch intervals while rolling across a formed or smooth concrete surface. The IE Scanner was designed for high resolution testing on finished concrete surfaces such as concrete floors, walls, girders, etc. The major limitations of the IE scanner are the scan rate (maximum of 1 ft/sec) and the poor results on rough surfaces due to poor contact of the transducer, impactor, or both. Therefore the central idea at the beginning of development was to create a large scale IE scanner that could achieve greater scan rates, perform well on relatively rough surfaces (typical of concrete bridge decks), incorporate additional test methods such as SASW (by synchronizing multiple rolling transducer wheels) and SIR (by automating a 3-lb instrumented hammer impact and measuring the induced lower frequency vibration) and be easily towed and maneuvered by a van or truck. The Bridge Deck Scanner (BDS) wheel is shown in Figure 24 and was designed to include six impact echo piezocermaic displacement transducers at 6 inch spacings, resulting in a wheel circumference of 3 feet or a diameter of approximately 11.5 inches. The 6 inch transducer spacing was considered to provide relatively close measurement intervals consistent with a high data resolution bridge deck survey. Six transducer elements from the Olson Instruments IE-1 head were incorporated into the wheel. The 6 transducers were spring mounted with rubber isolators and captured with a thin (1/16 ) urethane tire approximately 2.5 wide that is replaceable. The thin urethane tire was added as a dust cover to prevent dirt from entering the sensor housing and more importantly to increase sensor contact area and coupling. The Bridge Deck Scanner wheel design uses a slightly larger solenoid impactor than is typical in our other IE products. The larger solenoid imparts more energy into the concrete creating higher amplitude signals which are more easily measured. The larger solenoid also performs better on rough surfaces than a smaller solenoid because it is less affected by the immediate surface condition such as loose material, roughness, 36

43 paint coatings, etc. The urethane tire, larger impacting solenoids, and overall sensor weight (approximately 25 lbs), which effects contact pressure, are the primary changes that improved the rough surface performance over the handheld Impact Echo Scanner. Six solenoids per wheel were used in the design. The solenoids were mounted to the side of the rolling transducer wheel in line with the sensor element, instead of suspending a single solenoid from the Bridge Deck Scanner frame, thus ensuring the solenoid height (distance between bridge surface and solenoid) remained constant to improve test consistency. This style mounting also reduced the wear and tear on the solenoids by avoiding slippage and spreading the impacts out among six solenoids rather than relying on a single solenoid. A similar approach was taken with the electronics to power and acquire data from the sensors; instead of having a single very complex system housed independent of the rolling wheel, 6 small circuits were designed and incorporated into the wheel itself (Figure 26). This system has many advantages: first it reduces the number of wires which must be passed through the spinning hub assembly; second it makes the system more modular and robust where a single small component can easily be replaced if broken or damaged; and, third it makes the system more economical and simpler to produce six identical circuit boards than one large complex board. Slip-Ring Hub Assembly IE and SW Impact Solenoids Figure 24 - Bridge Deck Scanner Transducer Wheel, hub assembly side (outside). Embedded IE Test Head Sensors 37

44 IE Sensor Retaining Screw Thin Urethane Tire On-board Electronics Figure 25 - Bridge Deck Scanner Transducer Wheel, axle side (inside) with dust cover removed. To incorporate SASW we chose to use multiple Bridge Deck Scanner transducer wheels, described above, oriented, synchronized and timed in a transverse (across the bridge lane) line. As can be seen in Figure 26, the transducer wheels were mechanically connected using two u-joint slip couplers that would allow the wheels to move up and down independently and remain rotationally aligned such that one transducer from each wheel was in contact with the bridge deck surface at the same time. A mechanical adjustment was designed into the system so that either wheel could be delayed slightly if this was later deemed necessary due to the speed of travel in the forward direction. For SASW testing, the 2 nd wheel s solenoids would be turned off so that only one solenoid was firing at a time. The 2 nd wheel would become a SASW measurement only wheel. The wheels could also be offset 30 degrees apart in rotation and the solenoids on both wheels turned on to allow IE only testing on both wheels simultaneously. 38

45 Inside Dust Cover and Axle Mount Microphone U-joints Couplers Microphone Figure 26: Bridge Deck Scanner Sister Transducer Wheels with two u-joint slip couplers for rotational synchronization in SASW tests or offset 30 degrees for IE tests. To incorporate Slab Impulse Response (SIR), a rolling or sliding geophone receiver was designed as well as an automated 3-lb instrumented impulse hammer. The rolling SIR system incorporated a geophone receiver into the axle of the Bridge Deck Scanner wheel. Therefore the geophone itself would not be rotating with the wheel but it would be continuously coupled to the concrete surface through the wheel. This type of contact has the potential to be able to transfer the relatively high amplitude and low frequency signals typical of SIR testing. Several designs of an automated impulse hammer were considered which included the following approaches: hydraulic driven, gravity driven, pneumatic driven, electrically driven, and coil-spring driven. Ultimately it was decided to purchase and adapt a pneumatic framing nailer to drive the automated instrumented impulse hammer. The nail magazine, contact mechanism, and other unnecessary parts were removed from the nailer. Several new parts were designed and machined to support the added weight of an impulse head load cell (Dytran Model 1060V) and rubber/plastic impact tip, including: a stronger piston rod and bolt assembly/piston retainer. Two springs were added to the exterior of the piston to help return the piston and hold it in the neutral position. A large solenoid was installed to trigger the framing nailer once per revolution of the Bridge Deck Scanner instrument wheel or every 3 ft. The nailer was then mounted to a frame as shown in Figure 27 which had two rubber 39

46 wheels for stability and was positioned next to the bridge deck scanner transducer wheel. The pneumatic framing nailer was air driven from a small gas powered air compressor mounted in the back of the vehicle. The hammer system was independently mounted to the towing frame to travel alongside the bridge deck scanner instrument wheel which housed the geophone sensing element at its axle as shown in Figure 28 Nailer plus Solenoid for Triggering Impact Hammer Air Hose Fitting Figure 27 - Bridge Deck Scanner SIR Impulse hammer System, side view. 40

47 Return and Hold Springs New Bolt Assembly / Piston Retainer Transducer Wheel Axle with Embedded Geophone Load Cell Plastic Impact Tip Figure 28: Bridge Deck Scanner Impulse Hammer System, rear view. Microphone transducers were incorporated into the Bridge Deck Scanner prototype design in order to perform real world field testing of their applicability to Acoustic Sounding (AS), Impact Echo and Surface Wave testing. The microphones were shielded by inserting them into a short section of rubber tubing. This helped block unwanted direct air wave arrivals and exterior noise due to the wind, vehicle or rolling apparatus. The original design included two microphones, one mounted on the outsides of each of the two mirrored instrument wheels. The microphones were vertically oriented near the solenoid impact points to perform AS and IE testing (see Figure 29). Two additional microphones were added to allow SW testing in later iterations of the prototype design. 41

48 Microphone Hung from Frame Next to Impactor Figure 29: Bridge Deck Scanner System showing Microphone Placement. Concerning the overall prototype system, multiple mechanical and electrical adjustments were incorporated into the design to facilitate solenoid/sensor timing, wheel #1/wheel #2 timing, trigger/acquisition timing, and multiple test method timing. The original prototype with one pair of transducer wheels and a instrumented impulse hammer could theoretically perform IE and SASW at 6 inch spacings with the contacting transducers, IE and AS at 6 inch spacings with non-contacting microphone transducers and SIR with the impulse hammer and axle mounted geophone at 3 foot spacings. The transducer wheels and impulse hammer system were attached to a towing apparatus as shown in Figure 30. The apparatus consisted of a triangular frame with a ball hitch coupler at the apex. The corners of the frame were designed to be supported on the concrete surface with small rubber dolly wheels. This design allowed the axle-mounting-bar, attached to the transducer wheel, to maintain a consistent angle regardless of variation of the height of the truck hitch, which is 42

49 critical in the solenoid firing and acquisition timing of the system. The impact hammer system was also attached to this towing apparatus for simplicity. The apparatus was mounted to the vehicle via a standard ball hitch. Because the two transducer wheels were rotationally synchronized for SW testing, the system cannot make sharp corners without one of the wheels skidding on the concrete surface. The prototype system also did not easily allow for traveling in the reverse direction. Transducer Wheels Pneumatic Impulse Hammer Dolly Wheels To Put on Ball Hitch on Vehicle Figure 30: Bridge Deck Scanner System Original Design. 5.4 Description of Test Structures and Test Procedures Douglas Bridge in Douglas, WY The Douglas Bridge located near Douglas, WY is composed of two sister bridges, each supporting two lanes of traffic on Interstate 25. Only the south-bound bridge was evaluated during the investigation. The bridge consists of four spans and is supported by wide flange concrete girders. The bridge was 38 feet wide (curb to curb) and approximately 179 feet long. The bridge 43

50 was mostly straight however both ends were skewed. The bridge deck consisted of silica fume overlay concrete with a nominal thickness of 8 ¼ inches and it was reinforced in both directions. Figure 31: Douglas Bridge, Douglas, WY, Bridge Deck Scanner Testing 8/6/2009. The testing on the Douglas Bridge was the first field testing performed with the bridge deck scanner (see Figure 31). It was not the objective to perform a full investigation of the concrete bridge deck, but rather to test the bridge deck scanner system. Therefore all testing was performed on approximately the same test line in the right hand lane of the bridge. Test runs were conducted 44

51 the full length of the bridge deck. Multiple test runs were conducted with different test methods (e.g. IE, SW, SIR, AS) active for each test run (see Figure 32). Once a run was completed, the Bridge Deck Scanner was disconnected from the towing vehicle and manually rolled back to the beginning of the bridge. The vehicle was also returned to the north end of the bridge and the Bridge Deck Scanner (BDS) was reconnected and another test run was performed. Gas Powered Air Compressor Freedom Data PC - Data Acquisition System Figure 32: Douglas Bridge, Douglas, WY, Bridge Deck Scanner (BDS) Test Run 8/6/ st Street Bridge in Casper, WY The 1 st Street Bridge in Casper, WY is a four lane concrete structure over the North Platte River on 1 st Street. Only the two east-bound lanes were evaluated during our field investigation. The bridge is curved and skewed at both ends, with a centerline distance of approximately 357 feet and a deck width of ~ 36 ft (curb to curb). The deck is bare concrete with a nominal thickness of 7 inches. Note that the areas on top of girders are a couple of inches thicker than the nominal thickness since the slab was thickened to bear on the steel girders. Figure 33 shows the BDS on the concrete deck of the 1 st Street Bridge. Figure 34 shows the steel girders underneath the deck. A plan drawing of the Casper Bridge is included in Appendix A. 45

52 Figure 33: 1 st Street Bridge, Casper, WY, Bridge Deck Scanner Testing 8/19/2009. Figure 34: The Underneath View of the 1 st Street Bridge, Casper, WY which crosses over the North Platte River 46

53 The testing on the 1 st Street Bridge was performed as a full investigation of the concrete bridge deck conditions with the BDS. Testing was performed in test runs the full length of the bridge deck with approximately 1 foot transverse spacings. Improvements to the towing apparatus to permit moving the scanner in 1 foot increments across the width of a 12 foot lane were made after initial testing (see Section 6.1.2) which allowed test runs to be performed near the edges of the bridge deck as shown in Figure 35 below. Once a run was completed, the Bridge Deck Scanner was disconnected from the towing vehicle and manually rolled back to the beginning of the bridge. The vehicle was also returned to the west end of the bridge, then the Bridge Deck Scanner was reconnected and another test run was performed. In some areas of the bridge, significant gravel was present on the roadway and brooms were used to sweep the surface so that it was free of debris. Figure 35: 1 st Street Bridge, Casper, WY, Bridge Deck Scanner Test Run 8/19/

54 6.0 BRIDGE DECK SCANNER HARDWARE AND SOFTWARE IMPROVEMENTS 6.1 Hardware Original Hardware Design The original hardware design is described in Section 5.3. The original prototype of the Bridge Deck Scanner (BDS) was used for all testing on the Douglas Bridge in Douglas, WY as described in Section First Iteration BDS Improvements After initial testing on the Douglas Bridge in Douglas, WY several changes were made to the Bridge Deck Scanner to address issues with the system. In general, the IE and AS testing worked extremely well with good reliability and excellent data quality. The SW testing resulted in some locations having good data and some with poor data. The SIR testing provided only poor quality data. One significant limitation of the original prototype was the fact that it attached directly to the ball hitch on the towing vehicle; therefore the Bridge Deck Scanner system was always directly behind the center of the truck, making it impossible to perform test runs near the edges of the bridge deck. To address this problem, a 10 foot steel beam was attached to the towing hitch of the vehicle in the transverse direction as shown in Figure 36. The beam had trailer ball hitches at 1 foot spacings and would allow test runs to be performed at any location within a lane width while driving in the center of that lane. 48

55 Bridge Deck Scanner offset from Vehicle Center Figure 36: Bridge Deck Scanner 10 foot steel beam addition. It was determined that the major issue in the collection of SW testing data was the synchronization of the two transducer wheels. It was discovered that the two slip u-joints connecting the two transducer wheels had sufficient play to allow the wheels to become unsynchronized. To address this issue, it was decided to replace the slip u-joints with a solid axle between the transducer wheels. The original design employed slip u-joints to allow the two transducer wheels to independently move up and down following the contour of the road. After testing on the Douglas Bridge, it was determined that the contour differences within a one foot transverse spacing were minimal and would not effect the data acquisition, thus a solid axle was deemed appropriate. 10 Foot Steel Beam Holes for Ball Hitch Mounting Due to the promising results of other researchers in performing surface wave testing with audio microphones, two additional microphones (resulting in a total of 4) were added to the frame 49

56 of the transducer wheels as shown in Figure 37. This allowed several configurations of microphones with regards to spacing between transducers as well as the spacing from the point of impact to the transducers for experimentation purposes. Rigid Microphone Figure 37: Bridge Deck Scanner Additional Microphones and Rigid Axle Updates. There were several apparent issues when employing the SIR testing. The geophone (28 Hz resonant frequency) that was originally designed to attach to the transducer wheel axle did not have adequate response at low frequencies; therefore the original geophone was replaced with a 4.5 Hz resonant frequency geophone. The geophone also showed that the vibration from the firing of the impulse hammer was traveling through the frame and affecting the measured vibration readings, thus more isolation was required. Taking advantage of the new towing apparatus, which consisted of the 10 foot long beam with ball hitches at 1 foot spacings, the impulse hammer was reconfigured to have an independent frame and connect to a separate ball hitch, thus providing more isolation. The final obvious issue with SIR testing was the instability of the impulse hammer. Although the impulse hammer system weighed approximately 25 lbs, it still bounced significantly from the force of the impact on the bridge deck. To quickly address this issue in the short-term, three additional 25 lbs bags of lead shot were attached to the impulse hammer system (see Figure 38). 50

57 Separate Mounting for Impact Hammer and Geophone Receiver Additional Weight for Impact Hammer 4.5 Hz Geophone Figure 38: Bridge Deck Scanner Design After First Iteration of Modifications, Highlighting SIR Improvements Second Iteration BDS Improvements (Current Design) The Bridge Deck Scanner with the first iteration of improvements (as described above in Section 6.1.2) was used to perform testing on the 1 st Street Bridge in Casper, WY (described above in Section 5.4.2). The 1 st Street Bridge Investigation was conducted as a full investigation of the 2 east-bound lanes of the bridge. Test runs were performed the full length of the bridge at 1 foot transverse spacings. This extensive day of real world field testing again led to several hardware improvements and a better understanding of the system. 51

58 One of the evident differences in the data quality after the first iteration improvements was notably more vibration noise in the sister transducer during surface wave testing. It is believed that the major contributing change was the addition of a fixed axle between the transducer wheels to provided exact and fixed rotational alignment. The original design with the two u-joint slip couplers did not transfer notable vibrations, but it also did not provide reliable rotational alignment needed for surface waves (SW) testing. The second design provided excellent alignment but also transferred vibrations through the fixed axle and distorted the surface wave data. To address this problem, a third design was implemented using a solid axle with a rubber high frequency isolator inserted in the middle of the axle. The Slab Impulse Response (SIR) testing again proved to be problematic on the 1 st Street Bridge. The added weight to the impulse hammer system significantly improved the consistency of the hammer impulse force applied to the deck. The adjustments to the frame which isolated the impulse hammer system, by mounting it to a separate ball hitch, eliminated most of the direct vibration noise traveling through the frame. However, the 4.5 Hz geophone, which is much more sensitive and linear in its response to low frequency vibrations than the original 28 Hz geophone, was sensitive to the so-called rolling noise. This vibration noise is generated by the rolling wheel following the contours of the roadway and is at the frequencies important for SIR data analysis. Dr. Kenneth H. Stokoe, II and his students at the University of Texas at Austin have done similar testing with Rolling Dynamic Deflectometers (RDD). The RDD s have overcome rolling noise issues with geophone measurements by using extremely large input forces (10,000 pounds peakpeak is typical and therefore the vibration of interest is much greater than the rolling noise), forced frequency vibrations (the vibration of interest is at a single frequency between Hz instead of a wide frequency range), and have coupled the rolling geophone transducer mounted on a 2-wheel platform with an air piston spring to hold it down (Lee et al 2009). Based upon our results thus far from testing in our research lab and on both the Douglas and 1 st Street bridges, we believe a rolling geophone approach may be unsuitable for the SIR vibration measurement if implementing some of the RDD approaches do not resolve the problem in the future. The research team is currently exploring other possibilities that include a walking geophone design in which the geophone would be placed on a discrete location while the testing is performed and picked-up and moved ahead to the next test location as the vehicle proceeds forward. 52

59 The primary changes made in the second iteration were to the towing apparatus. To simplify the design, the dolly wheels were moved from the intermediate frame to the ends of the 10 foot long steel beam (square tube). The beam was then outfitted with a rotating slider system that allowed it to move up and down and twist in order to follow the contour of the road yet still be attached to the truck. The dolly wheels were also changed from small rubber wheels to larger, airfilled rubber wheels. The 10 foot long beam was spliced in two locations to allow a more compact shipping package. The two transducer wheels were attached to one another with a u-shaped yoke to make them easier to pick up together. The yoke was designed with rubber isolation joints to dampen any vibration between the two wheels. A handle was also attached to the yoke that enables the transducer wheels to be easily lifted and provides support to the rubber isolation joints. Previous versions of the system also had multiple cable connections to the power source and data acquisition system. In this iteration, significant re-wiring and design refinement was performed to concentrate all cables from the sister transducer wheel system into a single connection. This makes a much more user friendly and less complex system that is also quicker and easier to set up. Near the end of testing, one of the transducer wheel hub bearings seized and all transducer and solenoid cables were broken (due to twisting). Upon disassembly, it was discovered that a granular particle (either gravel or metal) had become embedded in the smooth plastic bushing causing the bushing to wear and eventually seize. The design was altered to utilize a more durable brass bushing and to increase clearances within this portion of the hub assembly so that particles will not wear on surfaces. It is believed that this particle was a metallic shaving from our manufacturing shop and was not picked up in the field during deck testing. 6.2 Software The Bridge Deck Scanner prototype was developed to run under Microsoft Windows XP on the Olson Instruments Freedom Data PC data acquisition system (1.1 GHz Intel Pentium M with 1 GB of RAM) which utilizes a 16 channel, 16-bit A/D data acquisition card by National Instruments. To support the new hardware prototype, software improvements were added to the original Impact Echo Scanner software. Multi-channel data acquisition capability was added to acquire data from the second rolling transducer, and two additional microphones. Relevant data analysis concerning microphone and SASW analysis was also added to the existing software. 53

60 7.0 TEST SETUP AND RESULTS FROM 1 st STREET BRIDGE (CASPER, WY) The internal condition study of the bridge deck of the 1 st Street Bridge was a collaboration effort between the research team at Olson Engineering, Inc and the University of Wyoming under the supervision of Dr. Jennifer E. Tanner of the Department of Civil Engineering along with the support of the Wyoming DOT. The title of the research project conducted by the University of Wyoming is Bridge Deck Evaluation using Non-destructive Test Methods and their project is funded by the Wyoming Department of Transportation (WYDOT). The scope of work of the University of Wyoming research included the studies of traditional Impact Echo method (point by point testing with an Olson Instruments Concrete Thickness Gauge) and Infrared Thermography to delineate the areas with top delamination. In addition, personnel from WYDOT performed a traditional chain drag on the bridge deck to locate areas with hollow sounds indicative of shallow delamination on the bridge deck. The scope of work of Olson Engineering, Inc. included the studies of the newly developed Bridge Deck Scanner prototype as part of this research, and radar surveys with ground-coupled and non-contact air horn antennae for Ground Penetrating Radar (GPR) based deck condition assessments in support of the University of Wyoming research. 7.1 Test Setups and Results from Traditional NDE Test Methods This section includes test setups and results from traditional nondestructive evaluation (NDE) test methods including Ground Penetrating Radar (GPR), Impact Echo (point by point), Infrared Thermography and chain drag acoustic sounding (AS) methods Test Setup and Results from Sounding Using Chain Drags This section is a summary of the test setup and results using traditional chain dragging for acoustic sounding (AS) to locate areas with hollow, drummy sounds indicative of shallow delamination. The chain drag testing was performed by WYDOT personal. The test setup and results presented herein were summarized from the quarterly report written by Tanner and Robinson submitted to WYDOT in August 09 [Tanner et al 2009]. 54

61 The chain dragging was performed using a row of chains that is attached to a handle and is brushed back and forth across the bridge deck (Figure 39). Common chain configurations will consist of four or five segments of 1 in. links of chain that are approximately 18 in. long (ASTM D standard). A 3x3 ft grid was previously laid out on the deck to assist in documenting delaminations. The operator must have a trained ear to hear the lower frequency, hollow, drummy tones that correspond to delaminated sections of the deck which flexurally resonate when excited by the dragging of the chains and are typically audible for the top 3-4 inches of a deck. Sound concrete has a sharper, higher frequency ringing sound by comparison. The hollow, drummy sounds denote a delamination and are marked directly on the bridge deck using paint. After the entire deck has been sounded, the operator then marks the delamination locations and develops a map of the bridge deck indicating the location of the delaminations. However, most of the damage mapping is at the discretion of the operator due to different levels of experience and hearing among operators. The results from the chain drag tests, which were performed by the WYDOT bridge crew, are presented as shaded areas in Figure 40 on a 3 ft square grid. of Wyoming). Figure 39: Chain Dragging Evaluation by Wyoming DOT (photo courtesy of the University 55

62 N Figure 40: Top Delamination Map from Traditional Sounding Using Chain Drags (courtesy of University of Wyoming) 56

63 7.1.2 Test Setup and Results from Ground Penetrating Radar (GPR) Tests The GPR tests were performed by the Olson Engineering research team using a Geophysical Survey Systems, Inc. (GSSI), 1500MHz ground coupled antenna as well as a 1GHz (1000MHz) air horn antenna along the length of the concrete bridge deck using a cart as shown in Figure 41. The tests were performed on the top of the deck per drawings provided by the Wyoming Department of Transportation (WYDOT). Traffic control for the testing was provided by WYDOT. The deck was scanned using a grid spacing of 1.5 feet along the N-S direction (width of the bridge) and 0.25 inch along the W-E direction (along a scan line). GPR data files were recorded in the eastbound direction, in one and a half foot transverse intervals from the centerline of the bridge to the south curb edge. The 1GHz air horn antenna data was collected from 4.5 feet inside the centerline to 4.5 feet from the curb due to the width of the truck the radar was mounted on. The objective of the GPR tests was to determine areas of the bridge deck with potential corrosion or delamination (cracks) at the top layer of steel reinforcement. Figure 41 - GPR testing with the 1500MHz ground-coupled antenna over the North Platte River in Casper, Wyoming. Data collected with the 1500MHz antenna contained clear reflections from each individual rebar in the deck. The raw data even shows evidence of some variance in signal attenuation within the concrete. The areas undergoing corrosion show up as weaker, attenuated signals than areas in 57

64 good condition (Figure 42). The data from the 1500MHz antenna was of good quality as shown in the figures below. Data from the 1GHz air horn was accurate but lacked the resolution (due to the wavelength of the signal) to pick out individual rebar. Figures 43 and 44 show data collected over the same location with the 1,500MHz antenna and the 1GHz air horn, respectively. Both plots show the depth and amplitude of the signal, but only the 1500MHz data allows for precise location of the reinforcement. Good Weak Figure 42: 1500MHz GPR Scan 6 feet offset from the bridge centerline. Note the variance of the signal strength as the radar passed areas of suspected corrosion. 58

65 Rebar denoted by hyperbolic reflectors Figure 43: 1500MHz ground coupled antenna GPR Scan 4.5 feet offset from the South curb. The West joint is located at the far left of the plot. Surface Reflecti Rebar Figure 44: 1GHz air horn antenna GPR Scan note the rebar reflections are not distinct feet offset from South curb. The West joint is located at the far left of the plot. 59

66 The GPR data from the deck was processed using GSSI RADAN 6.5 software to measure the reflection amplitudes (db) in each GPR data file of the individual transverse reinforcing bars within the top reinforcement mat as well as the depth of concrete cover over the rebar. Signal losses in the reinforcing bar reflection amplitudes vary according to the bar size and the relative abundance of moisture and chloride in the concrete cover and concrete above the top reinforcing bar mat. The signal losses have been correlated in previous studies (Gucunski et al 2008) with the location and extent of corrosion and corrosion-induced damage of the surface cover layer. The reflection amplitude data was corrected for geometric losses due to reinforcing bar depth using a statistical regression approach fit to the 90th percentile amplitude (db) versus the twoway travel time of the GPR signal. Predictions of the location and quantities of probable delamination and probable active corrosion were evaluated using proprietary thresholds calibrated for use on exposed-surface reinforced concrete bridge decks developed in research by Dr. Christopher Barnes at Dalhousie University, Halifax, Nova Scotia, Canada (Barnes et al 2008). This approach assumes that the 90th percentile strongest reflection amplitudes correspond to undamaged regions of the deck containing low quantities of moisture and chlorides. Areas with significantly more attenuated data below the thresholds correspond to upper reinforcement mat corrosion and/or corrosion induced-cracking of the concrete cover layer. Please note that the GPR investigation for delamination survey is most accurate for bridge deck areas with no previous repairs. The GPR results presented in Figure 45 show the deck surface in plan view and indicate probable delaminations in red and probable active corrosion areas in red and yellow. The quantity of probable delaminations was estimated to be 1,167 sq ft, or 10.8 percent of the deck surface area. The quantity of probable active corrosion was estimated to be 1,798 sq ft, or 16.7 percent of the deck surface area. Depth-corrected GPR amplitudes that were outside the damage thresholds are shown in grayscale to indicate the predicted relative variation in moisture and chlorides over the undamaged deck surface. Darker regions may indicate areas where moisture and chloride ingress is approaching levels sufficient to initiate corrosion. The chain drag AS results are presented in the top of Figure 45 for comparison purposes. 60

67 N Figure 45: GPR Evaluation of Delaminated (red), Corroded (yellow) and Darker Gray (possibly beginning to corrode) Areas on 1 st Street Bridge Deck with Chain Drag Acoustic Sounding Results at top for comparison 61

68 7.1.3 Test Setup and Results from Point by Point Impact Echo Tests The traditional point by point Impact Echo (IE) tests were performed by graduate students from the University of Wyoming (Dr. Jennifer Tanner s team). The IE tests were performed using a Concrete Thickness Gauge (CTG-1TF) manufactured by Olson Instruments. The tests were performed on a 3 ft x 3 ft grid fashion. The test results from the point by point IE tests are presented in Figure 46. Note that an interpolation technique was used to estimate the data between the grid lines. The test results summarized in this section were obtained from the quarterly report written by Tanner and Robinson submitted to WYDOT in August 09 [Tanner et al 2009]. In Figure 46, the darker blue areas represent shallow readings and darker red areas represent thicker readings from the CTG. Shallow regions represent potential areas of delamination and thicker regions correspond to sound concrete. The dark blue regions on either side of the contour map represent the skewed ends of the deck. Figure 47 is a simplified version of Figure 46 and only presents outlined damaged and delaminated zones. In all grid figures, the top section is the north portion of the bridge and the bottom section is the south section. The chain drag AS results are presented at the top of Figure 47 for comparison purposes with the point by point IE results. 62

69 Figure 46: Test Results from the Point by Point Impact Echo Tests (3 ft x 3 ft Grid) (courtesy of University of Wyoming) Figure 47: Shallow Delamination Map from the Point by Point Impact Echo Tests - 3 ft x 3 ft Grid (courtesy of University of Wyoming) with Chain Drag Acoustic Sounding Results at top for comparison N 63

70 7.1.4 Test Setup and Results from Infrared Thermography The Infrared Thermography tests were performed by the researchers from the University of Wyoming. The test results in this section are a summary from the quarterly report written by Tanner and Robinson submitted to WYDOT in August 09 [Tanner et al 2009]. Bridge deck delaminations are indicated by hotter temperatures as a deck warms up and comparatively cooler temperatures as a deck cools down from solar radiation. Approximately 900 images were overlaid to produce the thermal image of the bridge deck as presented in Figure 48. Figure 49 presents the outline of the shallow delamination damages from the results in Figure 48, Figure 48: Temperature Images of the Bridge Deck from Infrared Thermography Tests (courtesy of University of Wyoming) Figure 49: Shallow Delamination Map of the Bridge Deck from Infrared Thermography (courtesy of University of Wyoming) 64

71 7.2 Test Setups and Results from the Bridge Deck Scanner Prototype This section presents test results from all tests performed using the BDS prototype and discussions of current limitations from each test and future modifications planned for the BDS prototype Test Setup Using the BDS Prototype The BDS prototype was used on the 1 st Street Bridge to determine the damage conditions and damage locations of the concrete bridge deck. The BDS unit was mounted to a hitch behind a truck and the data acquisition system (controller) was placed on the tailgate of the truck. A maximum speed of 1 to 1.5 mph was achieved for the testing in order to maintain good data quality which degenerated at higher speeds. The BDS prototype performed Impact Echo tests using one transducer/impactor wheel in a line and Spectral Analysis of Surface Waves (SASW) tests were conducted using both transducer/impactor wheels simultaneously with the IE tests. Automated Sounding (AS) using microphones was also done simultaneously in the same test line as the IE test line. The BDS test setup for the IE, SASW and AS is presented in Figure 50. The BDS unit was then driven again on the same test line using the automated pneumatic nail gun impulse hammer and a geophone attached to the axle to perform the Slab Impulse Response (SIR) tests. The BDS test setup for the SIR tests is presented in Figure 51. IE, SASW and AS tests were performed every 1 ft along the entire width of the deck and 0.5 ft along each scan line over the length of the deck. SIR tests were performed on a separate run and only performed on one line. The SIR tests were performed every 3 ft along the scan line. 65

72 Figure 50: BDS Test Setup for IE, SASW and Automated Sounding on the 1 st Street Bridge Figure 51: Bridge Deck Scanner Test Setup for SIR Tests on the 1 st Street Bridge 66

73 7.2.2 Findings from Impact Echo Scanning Tests from the BDS Prototype The graphical IES test results from the Bridge Deck Scanner are presented in Figure 52. The plot is a surface thickness tomogram presented in a 3D thickness tomogram to elaborate the general condition of the tested concrete deck. The color thickness/echo depth scales are all in inches in Figure 52. The majority of the indicated anomalies are predominantly top delaminations based on the IES results. The green color represents areas where the thickness results ranged from 7.5 to 9 inches indicative of sound concrete, normal thickness deck areas. Dark green and light blue represent areas with greater thickness echo results of approximately 9-10 inches or areas with thickened slabs over the steel girders underneath the deck. Purple, Gray, and black colors represent areas with top delaminations. Yellow and red colors represent areas with thinner thickness results or more likely areas with either bottom delamination or internal cracks. Figure 53 presents a shallow delamination map of the bridge deck by the BDS IE system and the delamination map from chain drag AS in the top of Figure 53 for comparison purposes. The quantity of probable delaminations detected from the BDS was estimated to be 1,004 sq ft, or 11.1 percent of the tested deck surface area which compares well with the GPR results. There is a decent correlation of the Bridge Deck Scanner IE top delamination results with the chain drag AS results shown in Figure 53. However, review of Figure 52 shows a much more precise delineation of deck damage conditions with both top and bottom delamination and other deck integrity information from the BDS IE tests. The IE echoes indicative of the thickened slab over girder areas are evident as the 5 linear features in Figures 52 and 53 along the length of the deck. This further validates the accuracy of the Impact Echo scanning data obtained by the BDS prototype. 67

74 68 Figure 52: IE Test Results from the BDS Prototype from the 1 st Street Bridge Deck

75 69 Figure 53: IE Test Results from the BDS Prototype from the 1 st Street Bridge Deck Showing Top Delamination Mapping with Chain Drag AS results shown at top for comparison purposes (Probable Delamination Area = 1,004 sq ft or 11.1%)

76 7.2.3 Findings from Spectral Analysis of Surface Waves Tests from the BDS Prototype Full analysis of the SASW data was not performed for the 1 st Street Bridge as the bridge deck had not suffered extensive freeze-thaw damage where the cracking damage depths (from the top surface) and extent are of interest. This section presents example BDS SASW data from sound and delaminated concrete in Figures 54 and 55, respectively, with the following information: 1) Windowed data in time domain from the transducer near the impact (see Trace 1) 2) Windowed data in time domain from the transducer located 1 foot from the impact (see Trace 2) 3) Frequency spectrum representing thickness (or condition) of concrete deck from the transducer near the impact (see Trace 3) 4) Surface wave velocity between the two transducers (see Trace 4) 5) Phase plot calculated from data from both transducers (see Trace 5). Review of Figure 54 reveals an average surface wave velocity of 7,000 ft/sec which is indicative of normal, good quality concrete. This surface wave velocity predicts a compressional wave velocity of 12,500 ft/sec which is indicative of sound concrete. Review of Figure 55 reveals an average surface wave velocity of 3,000 ft/sec. This surface wave velocity predicts a compressional wave velocity of 5,357 ft/sec which is indicative of deteriorated concrete, in this case a delamination. 70

77 Trace 1 Trace 2 Trace 3 Trace 4 Trace 5 Figure 54: BDS SASW Data Obtained from Sound Concrete Trace 1 Trace 2 Trace 3 Trace 4 Trace 5 Figure 55: SASW Data Obtained from Concrete with Surface Delamination 71

78 7.2.4 Findings from Automated Acoustic Sounding with the BDS Prototype This section presents example data from sound concrete and delaminated concrete. In this case, a microphone was placed 1 inch away from the impact and 0.7 inch off the ground. A simplified diagram in Figure 56 shows the location of ground contacted displacement transducer (on the transducer wheel), impactor and microphone. Displacement Transducer (inside the transducer wheel) Microphone Impactor Figure 56: Locations of Microphone, Impact and Displacement Transducer on the BDS Wheel Figure 57a shows the time domain data from the displacement transducer and Figure 57b shows the time domain data from the adjacent microphone. The first arrival time of the data from the displacement transducer is 3,560 us and the first arrival time of the data from the adjacent microphone is 3,620 us with a phase change at 3,680 us. The following paragraph shows calculations for the impact time. An average compressional wave velocity of concrete is 12,000 ft/sec. Therefore the speed of the Rayleigh wave is 6,720 ft/sec from elastic wave equations. The impact time can be calculated in Eq. 2 as follows: Ddisp t 0 = t 1..(2) V where t 0 is the impact time, D disp is the distance between the impact and displacement transducer, t 1 is the first arrival time of the displacement transducer and V r is the Rayleigh wave velocity. In this case, t 1 is 3,560 us, D disp is inches and V r is 6,720 ft/sec. Therefore t 0 is calculated to be 3551 us. The paragraph below shows the calculation for the first arrival of the airborne wave (direct acoustic wave). r 72

79 The speed of air (V air ) is ~1,100 ft/sec and the distance between the microphone and the impact is 1.7 inches (1 inch inch). The travel time for the airborne wave from the impact to the microphone is calculated to be 129 us. The impact time (from the above paragraph) is calculated to be 3551 us. Therefore, based on the speed of air of 1,100 ft/sec, the first arrival time of the airborne wave is 3680 us ( us). This agrees well with the change in phase at 3680 us shown in Figure 57b us Figure 57a: Time Domain Data from Displacement Transducer 3620 us Change of phase at 3680 us Figure 57b: Time Domain Data from Adjacent Microphone Figure 57: Time Domain Data from Displacement Transducer and Microphone 73

80 Figure 57b shows that the time domain data obtained from the microphone adjacent to the impact is a combination of energy from the leaky Lamb wave and direct airborne wave. However, findings from the automatic sounding using microphone adjacent to the impact also showed that the microphone can be used to determine severe surface delamination when the leaky Lamb wave is a dominant portion within the time domain data. The top trace of Figure 58 shows the unfiltered time domain data from the displacement transducer located on areas with severe surface delamination and the bottom trace is the frequency spectrum of the top time domain data which has a high amplitude resonance indicative of flexure of a near-surface delamination. The top trace of Figure 59 shows the time domain data from the adjacent microphone and the bottom trace is the frequency spectrum of the top time domain data and has a similar high amplitude resonant frequency peak around 2000 Hz as identified with the IE displacement transducer in Figure 58. High amplitude of low frequency (typically an indication of surface delamination) Figure 58: Time Domain and Spectrum of Data from Displacement Transducer from an Area with Severe Top Delamination 74

81 . This is from the IE data Figure 59: Time Domain and Spectrum of Data from Non-contact Microphone from an Area with Severe Top Delamination Findings from Slab Impulse Response Tests from the BDS Prototype The Slab Impulse Response (SIR) component in the BDS prototype unfortunately did not result in a fully successful field experiment in this research. An example of typical Slab IR time domain data is shown in Figure 60. The left trace in Figure 60 shows the time domain data from the geophone attached to the axle and the right trace in Figure 60 shows the time domain data of force from the automated nail gun. Review of Figure 60 Reveals that the geophone on the axle was unable to sense the movement of the concrete deck due to rolling noise. In addition, interference from rolling results in low frequency rolling noise also adversely affected the SIR data. This section also presents the results of a laboratory experiment with the SIR test using the geophone attached to the axle of the BDS prototype and the automated nail gun. 75

82 Figure 60 Time Domain SIR Data from the Geophone and Automated Nail Gun Laboratory SIR Testing with the Bridge Deck Scanner Prototype Once the Bridge Deck Scanner original prototype design was complete, extensive testing was performed with the system in the laboratory on the shop floor. Of particular interest was the performance of the axle mounted geophone (at the time of laboratory testing, the 28 Hz natural frequency geophone was installed in the prototype) for SIR testing. One of the primary concerns was if the resonant frequency of the transducer wheel itself would interfere with the SIR data. This issue is not a problem in IE and SASW testing because the frequency ranges of interest are much higher. SIR testing is typically performed by impacting the concrete slab with an instrumented 3-lb impulse hammer while holding a geophone (4.5 Hz resonant frequency) in contact with the floor near the impact (within 4-6 inches) in order to measure the resulting vibration. In order to test the response of the geophone mounted to the axle of the bridge deck scanner, stationary tests were performed by using a 3-lb instrumented hammer to impact a 5 inch thick concrete slab within 6 inches of both the transducer wheel and a 4.5 Hz geophone held in contact with the floor. This allowed the results of both geophones to be directly compared for the same test location and the same impact. Figures 61 (a-d) and 62(a-d) below show the test results from a sound and voided test location (the voided location shows signs of significant loss of subgrade support beneath the slab). In both Figures 61 and 62, plot (a) shows the time domain vibration signal from the axle mounted geophone, plot (b) shows the transfer function (mobility = velocity/force vs. frequency) between the input force and axle mounted geophone measured vibration, plot (c) shows the time domain vibration signal from the geophone held in contact with the concrete slab, and plot (d) shows the corresponding mobility transfer function for the geophone on the slab. 76