Iowa State University Capstones, Theses and Retrospective Theses and Dissertations Dissertations

1994 An ultrasonic approach for nondestructive testing of deteriorating infrastructure: use of Direct Sequence Spread Spectrum Ultrasonic Evaluation to detect embedded steel deterioration Kevin Lee Rens Iowa State University

Follow this and additional works at: https://lib.dr.iastate.edu/rtd Part of the Civil Engineering Commons, and the Electrical and Computer Engineering Commons

Recommended Citation Rens, Kevin Lee, "An ultrasonic approach for nondestructive testing of deteriorating infrastructure: use of Direct Sequence Spread Spectrum Ultrasonic Evaluation to detect embedded steel deterioration " (1994). Retrospective Theses and Dissertations. 10642. https://lib.dr.iastate.edu/rtd/10642

This Dissertation is brought to you for free and open access by the Iowa State University Capstones, Theses and Dissertations at Iowa State University Digital Repository. It has been accepted for inclusion in Retrospective Theses and Dissertations by an authorized administrator of Iowa State University Digital Repository. For more information, please contact [email protected]. U-M-I MICROFILMED 1994 INFORMATION TO USERS

This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer.

The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction.

In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.

Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand corner and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book.

Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.

UMI University Microfilms International A Bell & Howell Information Company 300 North Zeeb Road, Ann Arbor, IVII48106-1346 USA 313/761-4700 800/521-0600

Order Number 9424256

An ultrasonic approach for nondestructive testing of deteriorating infrastructure: Use of direct sequence spread spectrum ultrasonic evaluation to detect embedded steel deterioration

Rens, Kevin Lee, Ph.D.

Iowa State University, 1994

Copyright ©1994 by Rens, Kevin Lee. All rights reserved.

UMI 300 N. Zeeb Rd. Ann Arbor, MI 48106 i

An ultrasonic approach for nondestructive testing of deteriorating infrastructure:

Use of Direct Sequence Spread Spectrum Ultrasonic Evaluation

to detect embedded steel deterioration

by

Kevin Lee Rens

A Dissertation Submitted to the

Graduate Faculty in Partial Fulfillment of the

Requirements for the Degree of

DOCTOR OF PHILOSOPHY

Department: Civil and Construction Engineering Major: Structural Engineering

Approved Members of the Committee

Signature was redacted for privacy.

In Charge of Major Work

Signature was redacted for privacy. For the Major Department Signature was redacted for privacy. Signature was redacted for privacy. For M^^raduate College

Iowa State University Ames, Iowa

1994

Copyright ® Kevin Lee Rens, 1994. All rights reserved. ii

In Memory of

My Grandfather, Bert J. Rens

Who was once convinced that he didn*t need a two bottom plow and later believed that in my life time, if I saw as much technology change as he saw, I would see farmers picking and planting using computerized robotics. iii

TABLE OF CONTENTS Page

LIST OF FIGURES vii

LIST OF TABLES ix

ABSTRACT x

CHAPTER 1. INTRODUCTION 1

Background 1 Objective of Approach 4 Overview 4

CHAPTER 2. AVAILABLE NDE TECHNIQUES 6

Major NDE Techniques 6 Acoustic Emission 6 Thermal Methods 6 Ultrasound 6 Magnetic Methods 7

Minor NDE Techniques 7 X-rays 7 Electrical Methods 8 Vibration Signature Testing 8 New Techniques 8

Acoustic Emission 9 Introduction 9 Theory 10 General 10 Location Models 10 Kaiser Effect 12 Applications 13 Summary 15

Thermal Methods 16 Introduction 16 Theory 17 General 17 iv

Linear System Model . . 17 Applications 18 Summary 20

Ultrasound 20 Introduction 20 Theory 21 General 21 Pulse Echo 21 DSSSUE 21 Applications 22 Summary 24

Magnetic Methods 24 Introduction 24 Theory 24 General 24 Stress Detection Methods 24 Flaw Detection Methods 25 Applications 25 Summary 25

Minor NDE Techniques 26 Introduction 26 X-radiography 26 Theory 26 Applications 27 Electrical Methods 28 Theory 28 Applications 28 Vibration Analysis 29 Introduction 29 Theoiy 29 Applications 30 Sununary 31 New Techniques 32 Theoiy 32 Applications 32 Summary 32

CHAPTER 3. QUESTIONNAIRE RESULTS 34 V

Overview 34 U.S. Summary 34 International Summary 35 Summary 40

CHAPTER 4. STRUCTURAL VIBRATION THEORY 41

Overview 41 Damped Single Degree of Freedom 41 Response Due to Arbitrary Input 43 DSSSUE 47 Random Binary Process 56

CHAPTER 5. COMPONENT IDENTinCATION AND FUNCTION 60

Overview 60 Description 60 Personal Computer 60 , Arbitrary Function Generator 65 Power Amplifier 65 Transmitting Transducer 68 Receiving Transducer 68 Low-noise Pre amplifier 68 Oscilloscope 68 Post Processing 68

CHAPTER 6. EXPERIMENTAL RESULTS 70

Laboratory Experiment 1 71 Laboratory Experiment 2 74 Typical Cross-correlation Signatures (Bar 1) 77 Bar 1 Analysis 80 Bar 2 Analysis 84 Bar 3 Analysis 84 Experimental Results Summary 89

CHAPTER 7. SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS . . 92

Summary 92 Conclusions 94 Reconmiendations 94 r

vi

BraUOGRAPHY 95

ACKNOWLEDGMENTS 109

APPENDIX A. CASE STUDIES Ill

APPENDIX B. THE FOURIER TRANSFORM 122

APPENDIX C. PULSE ECHO SURFACE WAVES 129 vii

LIST OF FIGURES

Page

Figure 2.1 Acoustic Emission Linear Location Model 11 Figure 2.2 DSSSUE Signatures 23 Figure 4.1 Single Degree of Freedom System 42 Figure 4.2 Arbitrary Loading 45 Figure 4.3 Properties of White Noise 51 Figure 4.4 Example of Single Degree of Freedom System 52 Figure 4.5 Graphical Representation of Equation (4.37) 53 Figure 4.6 Graphical Representation of Equation (4.39) 54 Figure 4.7 Graphical Representation of Equation (4.13) 55 Figure 4.8 Random Binary Process 58 Figure 4.9 Pseudo Random Binary Process 59 Figure 5.1 Basic DSSSUE Flow Diagram 61 Figure 5.2 DSSSUE Equipment 62 Figure 5.3 Pseudo Binary Process 63 Figure 5.4 Continuous Sinusiodal Wave 64 Figure 5.5 Continuously Transmitted DSSSUE Signal 66 Figure 5.6 DSSSUE Spectral Density 67 Figure 6.1 DSSSUE Calibration Block 72 Figure 6.2 Water Drop Cross-correlation Signature 73 Figure 6.3 Diffemetial Signatures of Water Drop Experiment 75 Figure 6.4 Bar Specimen Which Simulates a Round Anchorage Bar 76 Figure 6.5 Entire Cross-correlation Signatures for Bar 1 78 Figure 6.6 Bar 1 Cross-correlation Signature from 7,000 to 15,000 81 Figure 6.7 Bar 1 Differential Cross-correlation Signature from 7,000 to 15,000 (baseline - baseline) 82 Figure 6.8 Bar 1 Differential Cross-correlation Signature from 7.000 to 15,000 ... 83 f

viii

Figure 6.9 Bar 2 Cross-correlation Signature from 7,000 to 15,000 85 Figure 6.10 Bar 2 Differential Cross-correlation Signature from 7,000 to 15,000 (baseline - baseline) 86 Figure 6.11 Bar 2 Differential Cross-correlation Signature from 7.000 to 15,000 . . 87 Figure 6.12 Bar 3 Cross-correlation Signature from 7,000 to 15,000 88 Figure 6.13 Bar 3 Differential Cross-correlation Signature from 7,000 to 15,000 (baseline - baseline) 90 Figure 6.14 Bar 3 Differential Cross-correlation Signature from 7.000 to 15,000 . . 91 Figure A.l Interhabor Lock and Dam 114 Figure A.2 Murray Lock and Dam 115 Figure A.3 Bayou Sorell Lock and Dam 116 Figure A.4 Cheatham Lock and Dam 117 Figure A.5 Lock and Dam 24 on the Mississippi River 118 Figure A.6 Dial Gages to Measure Movement at Concrete Interfaces 119 Figure A.7 Anchorage Bar Failure at Lock and Dam 14 on Mississippi River .... 120 Figure A.8 Anchorage Failure at LaGrange Lock and Dam 121 Figure C.l Bonded Concrete Testing With Surface Waves 130 Figure C.2 Unbonded Concrete Testing with Surface Waves 131 ix

LIST OF TABLES Page Table 3.1 U.S. Responses by Testing Method 36 Table 3.2 International Responses by Testing Method 37 Table 3.3 Ultrasound Testing Applications in U.S 38 Table 3.4 Ultrasound Testing Training in U.S 38 Table 3.5 Magnetic Testing Applications in U.S 39 Table 3.6 Magnetic Testing Training in U.S 39 Table 6.1 Bar Flaw Depth (d) Data 77 X

ABSTRACT Better inspection techniques are needed for our deteriorating infrastructure. Recently, many reports have been published concerning the state of deterioration of public works systems which include bridges, navigation structures, pavements, etc. Statistics recently revealed that nearly half of the nation's bridges are either structurally or functionally inadequate. Another example of deteriorating infrastructure occurs in our nation's waterways. Approximately one-half of the Corps' 269 lock chambers along inland waterways will reach or exceed their 50-year design life by the turn of the century. Lock gates are unique and complicated structures consisting of several primary and secondary structural members. The anchorage system is a primary component which connects the gate to the concrete wall. Failure of the anchorage system can be catastrophic and without warning. For example, recently, a failed anchorage system at Lock and Dam 14 on the Mississippi River caused unscheduled maintenance to get the lock chamber operational. A similar failure has occurred on the Illinois Waterway at LaGrange Lock and Dam. Many times the embedded anchorage is badly corroded and a sudden failure could develop. Current inspection techniques consist of removing the surrounding concrete and inspecting the resulting hidden steel. Available nondestructive techniques can be used to determine the strength and serviceability of this embedded connection. Without on-going inspection programs to detect maintenance problems, unwanted down time will be required to solve several neglected problems. A two-week maintenance shut down has enormous effects on the whole transportation network, especially river transportation. To keep structures fully functional and structurally safe, inspectors must determine the overall condition of the total system. A suitable inspection program should (1) measure the structural characteristics in situ, (2) accurately assess the current operating or serviceability condition, (3) reduce current inspection costs, and (4) require a minimum of specialized training. Nondestructive evaluation (NDE) methods can help address these problems. This thesis investigates NDE techniques applicable to civil works structures. xi

Although some sections contain infonnation on NDE of concrete structures, the thrust of this work is devoted to steel structures and steel components of structures. Nondestructive testing can offer assurances, in varying degrees, of the soundness of a structure or a mechanical part. Many nondestructive techniques are used in industry and research laboratories today, but, with the current state-of-the-art, only a few of them can be applied to civil engineering. Specifically, the use of Direct-Sequence, Spread-Spectrum, Ultrasonic Evaluation (DSSSUE) is proposed. This new method uses a continuous transmission technique that increases the detection sensitivity and "illuminates" the structural part as compared to conventional pulse-echo methods where scanning is required. The improved sensitivity can be used to overcome the signal attenuation in large structures and to detect small changes in the structures. It is anticipated that this method will eventually be used in a global lock and dam inspection and rating program and that the transducers and techniques developed in this research will have application in the field of continuous monitoring. There is still a need to continue to bridge the technical gap between Civil Engineers and NDE Engineers. When this is accomplished, the development and use of nondestructive evaluation techniques for assessment of various components of this country's infrastructure will significantly increase. Hopefully, the DSSSUE technique will be eventually incorporated into a global inspection program for civil works structures such as bridges, buildings, and locks and dams. The experimental results indicate that the DSSSUE technique may be feasible to inspect the embedded lock anchorage system nondestructively. 1

CHAPTER 1. INTRODUCTION Background Better inspection techniques are needed for our deteriorating infrastructure. Recently, many reports have been published concerning the state of deterioration of public works systems. Statistics released in the fall of 1989 revealed that 238, 357 (41%) of the nation's 577,710 bridges are either structurally deficient or functionally obsolete [119]. In Iowa, 12,733 bridges (50%) are inadequate. A recent proposal announcement by National Science Foundation [91] states: The infrastructure deteriorates with time, due to aging of the materials, excessive use, overloading, climatic conditions, lack of sufficient maintenance, and difficulties encountered in proper inspection methods. All of these factors contribute to the obsolescence of the structural system as a whole. As a result, repair, retrofit, rehabilitation, and replacement become necessary actions to be taken to insure the safety of the public.

Another example of deteriorating infrastructure occurs in our nation's waterways. The U.S. Army Corps of Engineers operates more than 600 hydraulic structures (lock chambers, flood control dams, power houses, etc.). About 70% of these hydraulic structures are over 20 years of age; 49% are more than 30 years old; and 26% were constructed prior to 1940. Approximately one-half of the Corps' 269 lock chambers along inland waterways will have reached or exceeded their 50-year design life by the turn of the century [57]. The primary navigation structure in the inland U.S. waterways are miter lock gates. The anchorage system for the gates is a primary component which connects them to the concrete wall. Failure of the anchorage system can be catastrophic and without warning. For example, in the late 1980's, a failed anchorage system at Lock and Dam 14 on the Mississippi River caused unscheduled maintenance to make the lock chamber operational. A similar failure has occurred on the Illinois Waterway at LaGrange Lock and Dam. Many times the embedded anchorage is badly corroded and a sudden failure could be developing. Current inspection techniques consist of removing the surrounding concrete and inspecting the previously hidden steel. Appendix A describes the problems, failures. 2 and current inspection techniques associated with the anchorage system on lock and dam structures. Nondestructive techniques are needed to determine the strength and serviceability of this embedded connection. Without on-going inspection programs to detect maintenance problems when they are small, unwanted down time will be required to solve several neglected problems. An unscheduled two-week maintenance shut down has enormous effects on the whole transportation network especially river transportation. Civil engineers are only beginning to embrace the new role of "maintainers" - traditionally, they have not put their energy and money into this yet; however, it is starting to shift. While the trend is going toward maintenance, many civil engineers are just starting to accept this role. For years the emphasis has been on the more glamorous designing and building of new highways and structures. The new emphasis on maintenance in state highway agencies can be seen in the current reorganization of the Iowa Department of Transportation (IDOT). The maintenance office, which had been under the old highway division, has been elevated to a division level equal to the highway project development office which includes design and construction for new facilities. The Intermodal Surface Transportation EfGciency Act of 1991 (ISTEA) of the Federal Highway Administration was created to fund long-term research projects which address problems of the next century. The trend toward maintenance has been growing since the interstate system was nearing completion in the late 1970s and 1980s. The last section in Glenwood Colorado was completed in 1993. The passing of the ISTEA act changed the federal funding habits such that there is no longer any interstate funding. The federal maintenance trend has been sparked by a number of bridge collapses in the late 1970s and 1980s. The bridge inventory system was established in the late 1970s to put into place an inspection and evaluation system to maintain our nations bridges. A similar program was developed in 1984 by the U. S. Army Corps of Engineers (USA- COE). The USA-COE established the Repair, Evaluation, Maintenance, and Rehabilitation (REMR) program to focus more attention on the deterioration rates of 3

hydraulic and navigation structures. A recent statement by Richard Livingston, a member of the ISTEA committee, stated that NDE is needed to monitor the condition of infrastructure [72]. Livingston also stated that NDE priorities should be focused in the area of fractures, cracks, and corrosion as well as the causes and rates of these distresses [72]. To keep structures fully functional and structurally safe, inspectors must determine the overall condition of the total system. A suitable inspection program should (1) measure the structural characteristics in situ, (2) accurately assess the current operating or serviceability condition, (3) reduce current inspection costs, and (4) require a minimum of specialized training. Nondestructive evaluation (NDE) methods can help address these problems. The use of NDE in the assessment of the various components of the United States' infrastructure has significantly increased in the past few years. At a recent annual NDE conference, there were two full sessions of infrastructure presentations [122]. Just a few years ago, this same conference had only a few papers related to the NDE evaluation of steel and concrete structures. Similar trends are occurring at the Center for NDE at Iowa State University where researchers are investigating the application of NDE techniques to civil works structures. A recent article reported that in many cases the appearance of a bridge can be misleading in terms of its load-carrying capabilities [8]. In such cases, inspections and field testing are required to reliably evaluate the bridge. Civil engineering structures are unique and complicated. Typically they have massive size, interact with soil, are partially or completely submerged, are difficult to access, and contain architectural obstructions. Bahkt and Jaeger recently reported that nearly every bridge has some aspect of behavior that can escape the attention of even experienced analysts [8]. In addition, field inspection conditions are usually adverse. Livingston of ISTEA states "NDE researchers many times do not realize the 'field conditions' that are involved in many civil works inspections" [72] All of these problems complicate any global inspection program. 4

Objective The overall goals of this thesis are to; 1) Bridge the technical gap between civil engineers and NDE engineers to apply NDE techniques to civil structures. 2) Study the feasibility of an NDE technique to inspect deteriorated embedded steel, similar to the anchorage system for lock gates.

Overview of Approach To meet the objectives of this thesis, it was necessary to thoroughly review the techniques that are being utilized in the NDE field and evaluate their effectiveness for civil engineering structures. It was also necessary to review the general theory of several NDE techniques. Nondestructive testing can offer assurances, in varying degrees, of the soundness of a structure or a mechanical part. Many nondestructive techniques are used in industry and research laboratories today, but, with the current state-of-the-art, only a few of them can be applied to civil engineering. These methods are detailed later in this paper. After analyzing the available techniques, the direction to focus further study was selected by considering the theoretical, experimental, and analytical aspects of the NDE techniques. This thesis is sub divided into seven main chapters. Chapter 2 is a comprehensive literary review of NDE techniques and their applicability to civil engineering applications. It also provides an overview of the theory involved in each technique. Chapter 3 contains information about a questionnaire which was developed and sent to state highway agencies, industry, and international agencies. Chapter 4 outlines the theory for the selected technique Direct Sequence Spread Spectrum Ultrasonic Evaluation (DSSSUE) and its relationship to structural dynamics vibration theory. Chapter 5 illustrates the function of the equipment associated with the DSSSUE technique. Chapter 6 describes several structural parts that were tested with the DSSSUE technique before and after distressing. Chapter 7 summarizes the thesis, discusses the conclusions reached, and reconunends some areas of future work. A list 5 of bibliographic references appears at the end along with supporting appendices. Although some sections contain information on NDE of concrete structures, the thrust of this work is devoted to steel structures and steel components of structures. 6

CHAPTER 2. AVAILABLE NDE TECHNIQUES M^jor NDE Techniques During the course of this study, some methods were found to be better suited for civil engineering applications. These methods, called major NDE techniques, include acoustic emission, thermal, ultrasonic, and magnetic methods which are either proven techniques in the laboratory and field or appear very promising for application to civil works structures. In addition, these methods can be used for global or overall monitor­ ing of large or complicated structures.

Acoustic Emissions Acoustic emission testing can be thought of as listening for structural flaws to develop. As a structural system deforms, high-pitched energy is released. By separating normal and abnormal sound for a particular structure, a condition assessment can be made.

Thermal Methods Thermal methods use principles applied in monitoring residential housing insulation systems as well as assessing masonry wall conditions. By taking an infra-red picture of a building or home, areas of heat loss can be observed. The same principle can be applied to structural systems. Stress-concentrations in joints or areas of high fatigue show up as "hot spots" in infra-red scans.

Ultrasound Two types of ultrasonic testing are available. The more traditional method is called pulse- echo, while the other recently developed method is called Direct-Sequence, Spread-Spectrum, Ultrasonic Evaluation (DSSSUE). This new technique, which was developed at the Center for Nondestructive Evaluation (CNDE) at Iowa State University (ISU), uses a continuous-transmission technique that increases the detection sensitivity by a factor of 100,000 over conventional pulse-echo methods. In the DSSSUE system. 7

a mechanical part or structural member is "flooded" with ultrasound. Changes in aggregate properties such as volume, shape, dimension, composition, acoustic velocity, etc. are measured in one single test. In pulse-echo systems, a wideband pulse of acoustic energy is introduced into the test object. The pulse propagates in the material and is scattered or reflected by various surfaces and inhomogeneities within the object.

Magnetic Methods Magnetic methods can be used in testing ferromagnetic materials such as steel bridges to determine structural parameters such as stress, strain, and microstructure, and to detect flaws. Hysteresis properties such as permeability, coercivity, and remanence, are known to be sensitive to stress, strain, grain size, and thermal properties. Other magnetic methods such as magnetic particle, magnetic flux, and eddy currents can be used to determine flaws such as cracks, voids, corrosion, and section loss. These major techniques are described in more detail in the literature review which follows.

Minor NDE Techniques Methods that were determined to not be well suited for civil engineering field applications included X-ray and electrical methods. These techniques, termed minor NDE techniques, were too complicated for field use, not proven in the field, or not directly applicable to steel bridge structures. New techniques and methods used for local inspections were also included in the minor NDE section. Although not officially called an NDE technique by many individuals, vibration signature testing was also located in this section.

X-Rays The process of radiography requires exposing film to X-rays or gamma radiation that has penetrated a specimen, processing the exposed film, and interpreting the resulting radiograph. The resulting picture is then studied and analyzed to determine 8

corrosion, cracks, weld defects, etc. Another way is to pass gamma rays through the specimen and then measure the intensity of the emerging radiation by using Geiger or Scintillation counters.

Electrical Methods Monitoring steel deterioration by electrical methods is based on the measurement of the potential or resistivity of the specimen. Potential measurements are made by completing a half-cell circuit between the structural member. The potential method is a particularly useful method when used in conjunction with other nondestructive techniques for the interpretation of results. Resistivity measurements of steel and concrete members are made primarily to determine the rate of corrosion rates of steel because the resistivity changes as corrosion reduces the cross-sectional area of a steel rod.

Vibration Signature Testing All structural systems have dynamic properties which are unique to itself. For example, a pulse load on a structural member creates a response which is, by theory, repeatable for a given pulse. If a flaw is introduced in this member, the response is different. A vibration signature analysis can be an indicator of structural imperfections or deterioration.

New Techniques Several new techniques to inspect bridges nondestructively have been developed. Shearography is a technique based on laser interferometry. The technique provides fringe patterns corresponding to the derivative of the displacement field. Dynamic testing methods provide information about bridge capacities on the basis of the natural frequencies or dynamic response of the bridges. Sonar techniques use differences in travel time of sound waves to measure displacements. The impulse radar techniques measure differences in dielectric constants of materials to detect flaws. 9

Acoustic Emission Introduction Acoustic emission (AE), which utilizes high-frequency sound waves, is a nondestructive testing technique that has been applied to civil engineering structures. The basic principle behind AE is that a developing flaw emits bursts of energy in the form of high-frequency sound waves. By separating background noise from AE, the ongoing condition of a structure can be monitored. Several other terms that have been used to describe this phenomena include: stress wave emission, microseismic activity, and microseismic emission [83]. AE is a highly sensitive technique which detects microscopic events in a material. Because the events consist of elastic waves that propagate into the material, it is not necessary to focus on the exact location of the source event to detect it; one only needs to surround the event. Thus, AE is a passive monitoring technique relying on remote sensors. This contrasts other NDE techniques such as radiography, pulse-echo ultrasonics, or eddy current, which require 100% volummetric scanning for flaw/defect isolation. AE is therefore more efficient than these other techniques with respect to inspection time and preparation, which results in cost savings. In many applications, AE is far more sensitive than other NDE tech­ niques. Many times this sensitivity is used only to locate flaw/defect areas and the detailed sizing and classification is left to other NDE techniques. AE testing has been used for several years. The concept was first applied to structural monitoring of a bridge in 1939. Watchmen located in anchor houses of a suspension bridge reported that on quiet nights they could detect the sound of cable strands fracturing. Soon after this event, a decision was made to recable the bridge [46]. In the early 1940s, the U.S. Bureau of Mines prepared a document entitled "Use of Subaudible Noise for Prediction of Rock Bursts," which noted that the noise rate of stressed rock pillars increased as the structure became more highly loaded. These noises were later termed "rock talk" [86, 113]. Kaiser published a paper in the early 1950s entitled "Results and Conclusions of 10

Sound in Metallic Materials Under Tensile Stress" [113]. This research is generally accepted to be the beginning of acoustic emission as it is known today. Kaiser discovered an effect which bears his name and is discussed later in this review.

Theory General AE can be defined as a transient elastic wave generated by the rapid release of energy within a material. These emissions can come from plastic deformation such as grain boundary slip, phase transformations, and crack growth. The energy released by a single dislocation is normally too small to detect, but many dislocations are detectable by acoustic emission equipment [61, 88]. AE signals are defined by two categories: burst or distinct pulses and continuous emissions. Burst emissions are categorized as crack propagation and continuous emissions as movements or dislocations [61].

Location Models Several location models exist in current AE technology. These include: point location, zone isolation, and order of arrival. Only the point location method will be discussed in detail. In point location, the position of the AE source can be calculated, given the location of the sensors on the structure and the sequence and arrival times of signals from various sensors. In the linear location model (point location) for flaw location [88], the distance the flaw is from a particular transducer can be calculated from a burst signal. Fig. 2.1 illustrates the linear location model. Assume two transducers i and j are separated by a distance L and that a flaw is a distance D from transducer i. An acoustic emission event at the flaw location arrives at transducer i in time t|. The same event arrives at transducer j in time tj. The relative arrival time between the two transducers is

A/jy = tj - t. (2.1) r

11

Transducer i Traiuductr I

I •i

Figure 2.1 Acoustic Emission Linear Location Model 12

A calibration procedure establishes an average time value T between the two transducers by using an acoustic emission analyzer and a puiser receiver placed next to transducer i. The location of the flaw from transducer i is

D . (2.2) IT

Zone isolation is used when acoustic emission activity is known to likely come from a well-defined region, for example, near a connection, a bolt hole, or a stiffener. A zone encompassing this region can be defmed, and all AE activity is assumed to come from the known source. The theory states that a locus of points corresponding to differences in AE arrival time is hyperbolic. Two intersecting hyperbolas define an area on the specimen or structure where the flaw is located. The order of arrival method is based on arranging AE transducers in a triangular configuration. AE activity can be isolated into specific areas on the basis of the order of arrival of the AE event to the transducer. For example, assume there are three transducers (A, B, and C) placed in a triangular array. This triangular array can be divided into six areas. If transducer A first receives the signal followed by B then C, the particular AE event can be isolated in one of the six areas. These models are discussed in more detail in Ref. 113.

Kaiser Effect Generating acoustic emissions usually requires that a stress be applied to the structure tested. A unique property of acoustic emissions is that it has a memory known as the Kaiser effect. If a material is loaded, unloaded, and then reloaded, acoustic emissions will not be produced until the previous highest load is surpassed [88]. This feature has very practical consequences- it can be used to detect crack propagation, such as fatigue crack growth and stress-corrosion cracking. Additional theory can be found in several other references: [49,83,86,113]. 13

Applications Acoustic emission has been used to monitor and examine the behavior of metals, ceramics, composites, rocks, and concrete for rupture, yielding, fatigue, corrosion, stress concentration, and creep [61]. Additional AE applications to concrete structures include crack detection [14,31,41,44,66,67,98] expansion joint evaluation [131], in-situ stress [77], and fracture toughness [90]. AE has also been used to monitor behavior of fiber reinforced plastic [110], asphalts [126], and mortars [129]. Power and light companies have used AE on aerial devices and associated equipment to determine the safety of equipment used by their personnel [11]. Other applications include evaluating the structural integrity of cables; acoustic emission techniques have been used to determine the number of wires that break when a cable is loaded [68]. The Federal Republic of Germany has used acoustic emission to determine the safety of earth embankments [33]. Similar studies on earth structures and soil have been performed in the United States [90]. Different standards and personnel certification and qualifications are briefly discussed in Refs. 55 and 124. There are several readily available acoustic emission devices that were developed in the mid- 1980s. Stresswave Technology has developed a sensor that converts acoustic emission stress waves to electrical signals [76]. This device has applications to gears, valves, pumps, etc., and could have applications in structural engineering. AVT Engineering Services of Stockport, United Kingdom, has developed a fully automatic AE crack detection system applicable to all kinds of steel structures ranging from pressure vessels to bridges [26]. Gard Inc., a Chicago-based subsidiary of GATX Inc., has developed a device that detects growing cracks in steel bridges [18]. Several Ph.D dissertations have been published on the subject of acoustic emission monitoring of bridges. In 1985, Ghorbanpoor published a study on steel bridge component cracking [34]. Ghorbanpoor concluded that AE monitoring could be used to monitor and detect small cracks in highway bridge structures. In 1987, Mathieson published a study on AE of fatigue cracks in steel [81]. He concluded that crack growth rate is only possible from AE signals that occur at peak load. The 14

presence and location of cracks can be determined from secondary AE sources that result from crack closure, friction, and crushing of corrosion products. This phenom­ ena was also reported in Ref. 17. In 1989, Mohammad published a study on AE monitoring of bridge components [6] and developed a model that predicts the stages of crack growth based on AE signal strength. Several recent studies have been conducted related to acoustic emission monitoring of steel bridges [18,28,35,38,46]. An older study on acoustic emission for flaw detection of bridges was first published in the 1970s by the Federal Highway Administration [29,30]. Gong, Nyborg, and Oommen of Monac International, Incorporated have reported results from acoustic emission monitoring of 36 steel railroad bridges [38]. The number of active cracks and the crack safety index is reported for each of the bridges. Ghorbanpoor has published recent work on AE of bridges [35]. Similar to his findings in his Ph.D dissertation, he concluded AE to be an effective tool in determining fatigue crack activities in steel bridges. Similar studies were reported at a recent conference on nondestructive testing [102,135]. Generally, all of these studies have the same scope of determining crack conditions in structural steel bridge members. A traditional field application of acoustic emission is being performed at the University of Kentucky (U.K.) [18,28,46]. The project team has determined acoustic emission to be a promising method for large-scale nondestructive testing of structures. The research team placed several sensors on critical structural bridge members, as well as in the region of known cracks and dents. Those areas were monitored for a given period of time while the structure was stressed by normal service loads. Improved acoustic emission instrumentation helped separate mechanical noise from flaw-related noise by pattern recognition. In general, the instrumentation evaluates the data and compares three criteria to determine if they meet pattern recognition related to crack activity. The three criteria are (1) cracks produce many acoustic emission events over a short period of time (an event rate criteria), (2) acoustic emission events have a discrete energy level (an energy criteria), and (3) crack acoustic emission tends to come from a very localized area (location criteria). Data meeting these three 15

parameters are designated as crack related. The U. K. research team has field tested these techniques on nine bridges containing varying degrees of acoustic emission activity. Potential crack prone areas were identified and visible cracks were classified as propagating or non propagating.

Summary AE is applicable to civil engineering structures and bridge inspections and can be used in conjunction with other nondestructive techniques. Inspection methods are often complimentary. For example, visual techniques can be used to identify a problem area that can be further analyzed by AE. A quote from a recent engineering manual states "acoustic emission is presently being used to monitor crack propagation in bridges. This technique can be easily adapted to steel lock and dam structures" [61]. Generating acoustic emissions usually requires that a stress be applied to the structure. Bridge structures can be loaded manually or naturally by traffic and wind. Materials undergoing welding are stressed thermally. Miter lock gate structures are stressed by emptying and filling the lock chamber. The Kaiser effect has very practical applications in monitoring crack growth or overstress and can compliment a structural inspection by detecting areas of overstress as well as stress concentrations around dents and cracks. Electronic data processing greatly simplifies interpretation of the results. AE equipment can detect active and propagating flaws; however, data interpretation can be difficult for complex geometrical structures. Probably the most desirable attribute of acoustic emissions with respect to bridge or lock inspections is the ease of inservice monitoring of structural components. The overall goal of any type of bridge inspection procedure should be to avoid interrupting the flow of traffic. The AE inservice inspection capability offers this advantage. One other major advantage, especially for bridge inspections, is that the equipment is not affected by extensive rough surfaces since the transducer contact surface is small and requires minimal preparation of the structural component surface. 16

Thermal Methods Introduction A wide range of materials has been evaluated using infrared (IR) emissions. A recent U.S. Army Corps of Engineers technical report states, "Large structures can be loaded cyclically and stresses can be determined in-situ [61]." Areas of localized weakness or deterioration caused by a crack or corrosion can be detected. The basic principal behind IR emissions is that equipment measures temperature changes caused by tension or compression on the surface of a part or structural member. Since this equipment can measure temperature changes within a few tenths of a degree Celsius, it can be used in a variety of industrial situations and is being utilized in the field. Thermal analysis techniques have been used for several years. The phenomenon of an elastomer material (rubber) changing temperature upon tensile stretching was first discovered in 1805 by Gough, who performed simple tests on strands of India rubber. In metallic materials, the first genuine observation of the thermoelastic effect was made by Weber who noted the frequency of a vibrating wire did not change as quickly as expected when a tensile force was applied. He reasoned that this gradual frequency change was due to a temporary change in temperature of the wire as higher stress was applied. In the 1930s, vibratory, thermoelastic effects were studied by Tamman and Warrentrup (1937) and Zener (1938) [43]. In 1967, Belgen compiled all the existing theories and became the first scientist to use IR radiometry to estimate the amplitudes of dynamic stresses [118]. In 1978, after four years of research, the British Admiralty Research Establish­ ment developed a laboratory prototype called SPATE (Stress Pattern Analysis by Thermal Emission), which further developed the relationship between stress and temperature changes. Current SPATE versions use powerful computers that allow large structural areas to be scanned [43]. 17

Theory General Stress pattern analysis using thermal emission is based on measuring the thermoelastic effect in structural materials. Compression and tension cycles (fatigue) produce heating or cooling of the solids [61]. Pressures in an area of a solid show up as a stressed region where heat is generated or absorbed. This is known as the thermoelastic effect, i.e., the change of temperature that accompanies elastic deforma­ tion of a body. Thermal effects can be monitored in two ways; however, both ways involve heating the structure or specimen and measuring the flux change. Thermal effects can be obtained by physically heating the structure with lamps (until thermal equilibrium is reached) or applying stress reversals to fatigue the structure or part. In each of these methods, the change in heat can be related to stress or areas of deterioration. Thermal properties can be measured during the day or night as long as heat transfer is taking place on the structure, i.e., the surrounding environment is a different temperature than the part or structural system.

Linear Svstem Model The following linear system model for IR emissions is from an article in Optical Engineering [118]. Assuming that the time scale of the stress change is such that heat losses in a specimen are negligible, a change in the stress state of a solid produces a temperature change given by

AT ^T(a, + a, + a,) (2.3) p*Cf

where Q is the coefficient of thermal expansion; p is the mass density; C is the specific heat constant; T is absolute temperature; and a,, a;, and are the changes in principal stress of the material. A relationship for radiant emission change (flux change) from a surface that experiences a change of temperature is given by 18

A = 4eBT^AT (2.4)

where e is the surface emittance and B is the Stefan-Boltzmann constant. If a linear system is used to detect and measure flux change, the system signal S is proportional to the flux change or

S = /?A0 (2.5)

where R is the detector response factor (the signal per unit incident flux). Combining all three equations and solving for stress gives

a, + tFj + aj = AS (2.6)

where A is a constant associated with material properties. This last equation is the basic equation relating IR emissions to stress change for a structural component. The stress state over the surface of a structural component will vary and be dependent on the load characteristics, component geometry, and possibly the type of material. If an accurate flux can be detected and if the constant A is assumed to be calibrated, the stress changes can be accurately determined. The emitted thermal spectrum is part of the IR portion that requires the use of SPATE equipment to accurately detect change. The advantages and disadvantages of SPATE are listed in the following section. In addition, one-dimensional and two-dimensional defect reconstruction (inverse problem) from data recorded from the thermal equipment is under study [63,64]. Further theory on thermal methods can be found in Refs. 19 and 49.

Applications IR emission has been used in a variety of applications, including the inspection of air-frame structures, automotive and farm components, turbine blades, and chains; it has also been used to detect stress concentrations in welds [21,112,127]. Stress analysts have been the major users of this equipment for designing and modifying structures to determine areas of stress and to reveal areas of overstress [61]. Other applications 19

include the scanning of offshore oil-rig joints [96] and electrical control panels to identify potential electrical problems [32]. An extensive field application of IR emissions on concrete structures is being performed by Greiner Consultants [48]. The project team has used IR scanning to inspect concrete bridge decks for debonding and delamination conditions and for determining the corrosion levels of reinforcing steel. The IR method required five days to inspect one million square ft. of bridge deck. On the basis of the results of this global inspection procedure, an engineering economic analysis was performed and an optimum repair and maintenance solution was determined. This study showed that structural problems can be identified and located quickly causing minimal disruption to traffic since the system is remote. A similar inspection technique is used in Canada [80] and at the University of Wisconsin-Milwaukee [136]. Thermal techniques have been used to analyze steel corrosion levels. Louisiana State University has used infrared spectroscopy to identify developing rust phases and weathering characteristics of steel coupons taken from deteriorating bridge spans. Recently, the NASA Langley Research Center has evaluated aircraft structure corrosion and debonding by thermal techniques [130]. In this study, portable thermographic equipment was used to quantify the extent of material loss due to corrosion as well as to identify regions of debonding. As contrasted with point source NDE techniques (eddy current. X-ray, pulse-echo ultrasound) thermal methods can be used to scan large areas. Harwood and Cummings [43] cite the following advantages and disadvantages for thermal stress monitoring. The advantages include; data acquisition is fast, minimal surface preparation, works good on complex geometries, highly sensitive, no contact needed, portable, and it can be used during the day or evening. The disadvantages include: a high first cost, structure must be fatigued or heated for thermal activity, and only surface stress obtainable (no internal stress). 20

Summary The application of thermal stress monitoring can be applied to the inspection of bridge structures. The output of a complete thermal scan usually consists of stress maps or contours which can be used to determine structural integrity. IR scanning can help monitor fatigue connections and as indicated in Refs. 101 and 128, it can monitor differences in corrosion levels on structural steel. Thermal stress monitoring requires no contact with the surface of the structure thus eliminating the tedious task of mounting strain gages for measuring stress concentrations. IR emission has been used significantly in the laboratory, but because of the size of testing equipment, field applications have been rare. Like all nondestruc­ tive methods, microprocessing increases the cost significantly.

Ultrasound Introduction Ultrasound refers to sound which is too high-pitched for the human ear to hear. Under normal circumstances, the ear can perceive sound up to 20,000 cycles per sec. Because sound travels at particular speeds in any one material, the distance it has traveled can be determined by measuring the elapsed time. This distance can be quantified by taking the sound velocity times the elapsed time. Like other NDE techniques, ultrasonic concepts have been in existence for many years. The discovery of piezoelectricity, which is the fundamental principle behind the relationship between electricity and mechanical vibrations in transducers, was made by Curie in 1880. In spite of this discovery, no important use of it was made for more than 35 years [36]. In 1917, Langevin developed a method for detecting submarines using a piezoelectric substance mounted to the bottom of a ship [36]. Submarines in the area could be detected by means of echoes called echo-ranging. Even greater advances in echo ranging sonar were made during the Second World War. The modem era of ultrasound is based on the technological developments of electronic circuitry, transducer design, and computer advancements. 21

Theory General Two types of ultrasonic testing are available. The first and more traditional method called pulse echo consists of a number of short pulses of inaudible sound. The second recently developed method is called Direct-Sequence, Spread- Spectrum, Ultrasonic Evaluation (DSSSUE). This new technique, which was developed at the Center for Nondestructive Evaluation (CNDE) at Iowa State University, uses a continuous-transmission technique that increases the detection sensitivity over conven­ tional pulse-echo methods. This improved sensitivity can be used to overcome the signal attenuation in large structures and to detect small changes in the structure.

Pulse Echo In pulse-echo systems, a wideband pulse of acoustic energy is introduced into the test object. The pulse propagates in the material and is scattered or reflected from the various object surfaces and from inhomogeneities within the object. Because flaws such as cracks, corrosion, inclusions, and voids represent major material inhomogeneities, considerable acoustic energy is scattered or reflected back to the receiving transducer where the corresponding return signal is recorded as amplitude versus time (an A-scan). The position of a flaw can be determined by scanning the test object and recording the round-trip transit time of the pulse; i.e, transit time and flaw location are directly related. Pulse returns from various locations can be isolated according to their respective transit times by bracketing the received signal with a time window.

DSSSUE In the DSSSUE system [1,7,59,108], a continuously transmitted signal of wideband acoustic energy is used to flood the test object with ultrasound. As with pulse-echo systems, this signal also propagates in the material and is scattered or reflected from the various object surfaces and from inhomogeneities (such as flaws) within the object. However, unlike pulse-echo systems, there is no equivalent "time-of- arrival" or transit time concept inherent in the resulting correlation function. What the DSSSUE system really does is to measure the composite characteristic of the entire acoustic system (structure and transducers) and measure the cross-correlation functions between the various input and output transducers of the composite system. The advantage of this approach is that the cross-correlation functions represent signatures for the test object that can be used to detect changes in object characteristics, such as volume, shape, dimension, composition, density, homogeneity, and acoustic velocity. The ability to measure the changes of so many of the properties of the structure gives the DSSSUE system a significant advantage over pulse-echo systems. For example, a fatigue crack will show up as a spike in the cross-correlation waveform that makes up the characteristic signature for the structure. When a baseline cross-correlation signature is established, a condition assessment can be made by comparing the baseline signature with one taken at a later point in time. For example. Fig. 2.2a represents a baseline signature of a structure. The same structure with a flaw has the signature in Fig.2.2b. The difference in the two signatures is a direct indication of the flaw. Because the DSSSUE technique has the potential to measure large or complicated structures, this technique will be further pursued in later chapters of this thesis.

Applications Ultrasound has been used on all metals and alloys, welds, structural members, forgings and castings [61]. Ultrasound has been applied to crack detection of concrete structures [22], thickness measurements and corrosion detection [20,60,116,117], and has been used to determine cracks and inclusions in welds [20]. Recently, ultrasound has been used in the field to determine the quality of bridge welds [87], bridge pin integrity [121], and the measurement of corrosion in steel bridges [94]. Ultrasonic techniques have also been used to detect corrosion damage in aircraft structures [70,101]. 23

Back-Surfac<

a

0 100 200 300 400 500 600 700 800 900 1000

Code Delay a) Without Flaw

Surface

I «

0 100 200 300 400 500 600 700 800 900 1000

b) With Flaw Code Delay Figure 2.2 DSSSUE Signatures 24

Summary Pulse-echo ultrasonic methods tend to work well for local flaw detection in both laboratory and field applications and is being extensively used by state departments of transportation to determine the extent of fatigue cracks. The DSSSUE technique looks very promising for global evaluation applications of large, complicated, structural systems.

Magnetic Methods Introduction Although there are several different magnetic methods, all of them fall into one of two categories: (1) flaw detection, and (2) stress detection. Flaw detection methods include magnetic particle inspection (MPI), eddy current (EC), and magnetic flux leakage (MFL) techniques. Stress detection methods include magnetic Barkhausen effect (MBE), magnetic acoustic emission (MAE), hysterisis, residual field, and magneto elastic methods (MIVC).

Theory General Flaw detection techniques rely on flaws to disrupt the magnetic field, while stress detection techniques rely on the fact that magnetic properties such as permeability, coercivity, etc., can be related to stress. The general descriptions and theory are discussed fully in Refs. 50 and 32 and will be briefly defined in the next sections.

Stress Detection Methods Stress detection methods include MBE, MAE, hysteresis, residual field, and MIVC. MBE causes discontinuous changes in flux density which is related to residual stress. MAE is dependent on the magnetostriction coefficient, which, in turn, is dependent on the applied stress. The hysteresis properties are sensitive to stress. Residual field techniques detect changes in strain while the MIVC method measures acoustic velocity which can be related to stress. 25

Flaw Detection Methods Flaw detection methods include MPI, MFL, and EC. MPI uses powder to detect leaks of magnetic flux. The MFL method uses a magnetometer to detect flux change, while EC locates flaws by detecting disruptions in magnetic fields.

Applications Several papers on magnetic methods discussed a variety of applications. For example, MPI techniques have been used to identify fatigue cracks in structural members, gears, pumps and shafts, heat cracks, and weld defects such as undercuttings and lack of fusion [61]. Eddy current techniques have been used to monitor the condition of nuclear reactors [104]; improved transducer design has helped aid in image restoration [42]. Eddy current responses due to rectangular flaws are also presented in Ref. 4. Magnetic methods have been used to determine the condition of several types of ferromagnetic parts [39,100,111], including the detection of corrosion in aircraft parts [16,89,123]. Many applications involve relating the magnetic properties to stress [27,45,51,54,56,58,109,125,132]. Magnetic flux leakage has been used to determine the steel deterioration in concrete and to determine the condition of main cables in suspension bridges [85,120]. A portable magnetic inspection system has been developed that is adaptable to many different environments to determine fatigue and creep damage [53]. This device, called the magnescope, has been used in the field to determine stress gradients in steel railroad bridge girders [25].

Summary Most magnetic methods have been fully laboratory tested and a few have been used in the field. The magnescope seems to be the most promising magnetic technique for global testing in the field. 26

Minor NDE Techniques Introduction Techniques that were determined to not be well suited for civil engineering field applications are classified as minor NDE techniques in this thesis. These include X- radiography, electrical techniques (e.g., electrochemical and A.C. impedance methods), and newer techniques which include shearography, sonar, and impulse radar. The techniques have been used in a variety of applications with varying degrees of success. Vibration signature testing is also included in this section. The following sections briefly explain the theory and applications of these methods.

X-radiography Theorv 11241 Radiography is a well accepted technique for NDE of materials. For example, it is extensively used to examine the internal structure of weldments on bridge structures. Radiography relies on the penetration and absorption characteristics of X-radiation to test materials. When radiation passes through a material, some of the radiation is absorbed. The amount absorbed is dependant on the thickness and density of the material being tested. The emerging beam contains information about flaws in the specimen. This is usually recorded on special radiographic film. Non-film techniques like fluoroscopy and real time radiography are used for special applications. The three essential components for producing a radiograph are: (1) the radiation source, (2) the object (specimen) to be radiographed, and (3) the cassette containing the film. The usual source of X-rays is an X-ray Tube which has a source of electrons (a tungsten filament), a means of accelerating the electrons (a high-voltage electric field), and a target for the electrons (the tungsten anode). When current is passed through the tungsten filament, electrons are produced. These electrons are accelerated by the electric field towards the anode. They strike the anode to release X-rays. Beam intensity can be controlled by the current in the filament, while the penetration power can be controlled by the field voltage. Usually the cassette is made of cardboard, metal, plastic, or a combination of these materials. The cassette has an upper, and a 27

lower lead screen, between which the film is placed. The screens help to intensify the image and reduce the amount of scatter. The film has to be processed by suitable darkroom techniques to get the final image. Darker areas in the image usually imply a crack or an undercut, whereas lighter areas imply a weld build up. The severity of the defect is indicated by the intensity of the image. While the defect is visible, it is difficult to get information about the position of the flaw inside the specimen. The technique is excellent to detect flaws in structural members nondestructively. Radiographic examination of steel structures is typically guided by the requirements of AWS Code D1.1. For examination methodology, AWS references ASTM standards E- 94, "Standard Recommended Practice for Radiographic Technique, " and ASTM E-142 "Standard Method for Controlling Quality of Radiographic Testing." AASHTO specifications also contain some provisions for the above purpose. The technique can be used with most materials to detect internal flaws and produces permanent records (when film is used). However, there are disadvantages due to its inability to accurately determine the depth of a defect and to locate defects perpendicular to the beam. Inaccessible places and complex geometries are difficult to radiograph. Also, the technique is a safety hazard, and care should be taken to avoid excessive exposure of personnel to radiation. In addition, the technique is not global and, thus, can be used for defect detection only in a localized zone. Even given these limitations, the technique is excellent for flaw detection in bridge structures.

Applications The technique has been extensively used for NDE of bridge weldments [124] and in the aerospace industry for aircraft inspection. Recent advances in dry processing techniques have simplified field applications [133]. Radiography has been used to conduct nondestructive depth profiling analysis of composite specimens [135]. In addition, radiation techniques have been used recently to determine areas of corrosion in layered structures such as aircraft [12,69]. 28

Electrical Methods Theory This section includes techniques like the electrochemical method, A C. impedance spectroscopy, galvanostatic pulse techniques, and electrochemical potential monitoring (half-cell method). The electrochemical method is based on polarization experiments to obtain anodic polarization curves [128]. Certain parameters like peak current density may then be evaluated from the curves. These parameters are used to detect deterioration. Alternating current impedance spectroscopy relies on the change in the impedance of a conductor to detect corrosion in steel rebars [75]. Impedance of a reinforcing bar is significantly influenced by the extent of corrosion. An alternating current is applied to the reinforcing bar and the voltages measured. The measured voltages are indicative of the impedance, and hence locate the corroded section in the reinforcing bar. The galvanostatic pulse technique is based on the analysis of the transient response of a steel-concrete system to an applied pulse [93]. The electrochemical potential monitoring technique (half cell method) has been the most commonly used technique to detect the corrosion of reinforcing bars in concrete [73]. It uses a half-cell reference electrode (usually a copper/copper sulfate electrode (CSE)) to measure the electrochemical potential. The potential is representa­ tive of the probability of corrosion in reinforcing bars. The electrode is placed at various points on a wet concrete surface, and the potential is measured. Equipotential contours are then drawn to give an indication of areas with probable corroded reinforcement.

Applications The electrochemical method has been used for evaluating degradation of in-service materials made of low alloy steels [128]. Both A C. impedance spectroscopy and the galvanostatic pulse technique have been investigated as possible methods for detecting the corrosion of reinforcement [75,93]. The electro­ chemical potential monitoring method, which has been used for reinforcement corrosion 29

detection for several years, has also been used for monitoring corrosion and coating protection [73]. A reference moving wheel electrode, which has recently been developed, allows for quick corrosion monitoring of large areas of reinforced concrete [74].

Vibration analysis Introduction Mechanical equipment and structural systems very rarely break down without warning. Mechanical or structural problems are almost always accompa­ nied by an increase in vibrational levels which can be measured. The basic principal behind a vibration signature analysis lies in structural vibrations and dynamics. Collectively, these dynamic characteristics form a signature or fingerprint which will not change unless the dynamic parameters are altered. This fingerprint is fundamentally the same as the the signature in the ultrasonic section. However, the frequency of vibration signature analysis is on the order of one megahertz slower. Again, like in all NDE techniques already presented earlier, signal interpretation can be difficult.

Theorv The subject of structural vibrations and dynamics is well established and documented in numerous textbooks and papers. Elements that are used in describing a vibrational signature are as follows: 1) natural frequency, 2) structural damping, and 3) mode shapes. These system parameters, called the dynamic response, can be determined from frequency response plots, which can be measured directly from a vibrating structure using accelerometers [84]. Modal analysis is a tool which has been developed over many years as a method of determining the dynamic response of a structure. In order to obtain the parameters, a series of tests are carried out on the structure. Accelerometers are attached to a fixed location on the stmcture and response data is gathered after the structure is excited. Signal processing then creates frequency response plots which is an indicator of frequency and damping. Several different response plots taken at different locations on the structure can be used to obtain mode shapes. Changes in the signatures (frequen­ 30

cies, damping, or mode shapes) indicate that changes have occurred in the structure such as cracks, dents, corrosion, loose bolts, or other forms of deterioration. In this respect, it is similar to the DSSSUE method (Fig. 2.2). The wave length associated with vibration analysis is, usually, considerably longer than for the DSSSUE method.

Applications Vibration analysis has been used to monitor and examine bearings, pumps, motors, sumps, hydraulic cylinders, shafts, and almost any moving part that will produce vibrations. For example, in gears, tooth deflection under load and tooth wear will give rise to tooth meshing frequency [61]. This application is of interest to the author of this paper because many sector lock gate structures (costal structures) are driven by a bull gear and roller rack. Other applications include condition monitoring of rotating machinery where vibrational trends are measured at random times to help schedule appropriate maintenance activity [106]. European practice has used dynamic response to help determine void areas under concrete slabs [5]. Vibrational signature analysis has been applied in offshore platforms by observing changes in structural dynamic parameters [107]. The most extensive laboratory/field application of vibration signature analysis is being performed at the University of Connecticut [84]. The project team has concluded that an ongoing vibration signature monitoring program is effective in determining structural degradation. The U. Conn, research team used a small scale laboratory bridge model and placed accelerometers along the bridge spans. Mock vehicles were driven over the bridge and a series of tests were performed. First of all, verification tests were performed to compare modal analysis techniques to other vibration analysis techniques. These tests correlated well with modal analysis. Second, before vibration signature analysis to defect structural degradation could be performed, the effects of normal operating conditions had to be accounted for. Tests were performed to see how vehicle velocity, roadway roughness, and vehicle mass affects structural dynamics parameters. The results of these tests show modal parameters to be consistent with the baseline verification study case with the exception of modal amplitudes with respect to 31

vehicle mass. The mass of the vehicle affected the vibration level more than expected. The most valuable application of the U. Conn, laboratory bridge model was the evaluation of structural degradation and how it affects the vibration signature. Two structural failures were considered, support failure and crack propagation. The bridge model consisted of a double girder two span bridge with three support locations. Each support location had two pins, on one each side of the road. One of the pins was loosened from the center support, the structure was excited, and modal analysis performed. The response spectra showed a significant mode shape change near the released support location. A crack was introduced in one of the two girders at a point of maximum bending moment. The crack was developed in three stages, reducing the bending moment of inertia to amounts of 81%, 68%, and 67% of the original cross- section. The structure was excited for each different cross-section and modal analysis was performed. Like in the support removal experiment, the response spectra showed a significant mode shape change near the crack location. The deviation increased for each crack increment. The U. Conn, research team has conducted lull scale experiments with similar results to the laboratory results. Collectively, the results of this research indicate the concept of an automated vibration monitoring system is feasible.

Sunmiarv Vibration signature analysis has applications in civil engineering and specifically the miter lock gate inspection and rating program. A wide range of equipment is available to measure vibrations. A single frequency band meter yields a simple vibrational reading. Narrow band meters can provide a hard copy of the measured spectra in any parameter (acceleration, velocity, or displacement) [61]. An inspector can make maintenance decisions and schedule replacement or repair based on the response spectra. 32

New Techniques Theory Several new techniques for inspecting bridges nondestructively have been developed. Shearography is a technique based on laser interferometry. The technique provides fringe patterns corresponding to the derivative of the displacement field. Dynamic testing methods provide information about bridge capacities on the basis of the natural frequencies or dynamic response of the bridge. Sonar techniques use differences in travel time of sound waves to measure displacements. The impulse radar technique measures differences in dielectric constants of materials to detect flaws.

Applications Shearography has been used to detect cracks and strain concen­ trations [78]. Instantaneous frequency concepts have been used to evaluate the actual response of a bridge to loading [24]. Sonar techniques were applied to measure the motion of several points along a vibrating beam [65]. Impulse radar has been used for various purposes including detection of ungrouted masonry units in a masonry wall, pavement voids, misaligned dowel bars, and high chloride concentrations in concrete decks [71]. Information regarding on going deformation and cracking has been acquired by permanently instrumenting the structure with various sensors such as accelerometers and tilt meters during constructed so that the structure can be routinely monitored [2]. Methods to identify faults in beams on the basis of differences in natural frequency have been developed [15]. Television cameras and image digitizers have been used to estimate the dynamic response of a bridge [95]. Diagnostic load testing methods have been effective in monitoring the condition of old bridges [23]. Telemetry devices have been used to help transport data from inaccessible locations [40]. In short all NDE, both new and old, requires some sort of signature analysis.

Summary A review of current nondestructive evaluation (NDE) techniques applicable to large- scale, civil engineering structures has been presented. The literature reveals that acoustic emission testing is the most widely used state-of-the-art technique for 33 monitoring conditions of bridges and other large or complicated structures. This method was developed from extensive laboratory and field testing in the mid-1970s to mid-1980s and is currently in the application stage. Thermal techniques have been applied to several civil engineering projects-primarily asphalt and concrete pavement condition assessments. Similar to acoustic emission testing, thermal methods are making the transition from the research phase to the application phase. Although magnetic methods have been extensively tested in the laboratory, the field testing of these methods is only beginning. More than likely, the application of magnetic methods for field monitoring of civil works structures is years away. A new ultrasonic technique, called Direct-Sequence, Spread-Spectrum, Ultrasonic Evaluation (DSSSUE), has recently been developed. This technique has the potential to become an effective global monitoring technique for civil engineering applications; however, it is still in the developmental stage. Other techniques, such as pulse-echo ultrasonic, electronic, and radiographic, are effective for local investigations. 34

CHAPTERS. QUESTIONNAIRE RESULTS Overview This chapter presents the results of an investigation which was undertaken to determine the current NDE trends in civil and construction engineering [105]. The investigation involved a survey of various transportation agencies of NDE applications in the field of civil engineering. This chapter contains a portion of the information obtained from questionnaires sent to agencies in the United States as well as international agencies. The questionnaire form can be found in Ref. 105. Of the 58 questionnaires sent to U.S. agencies, 50 were sent to state departments of transportation (DOT'S) and 8 to industry. The DOT's returned 94 percent of the questionnaires while the industry returned 63 percent. The overall return rate was approximately 90%. The results of the survey by NDE technique are summarized in Table 3.1. All 38 international questionnaires were sent to public works organizations. Four agencies (10%) responded and the limited results by NDE technique are summarized in Table 3.2.

U.S. Summary Of the 36 responses pertaining to ultrasonic testing (UT), local investigation of pin and hanger assemblies in bridges, castings, welds, bolts, and corrosion thickness are being performed. In addition, some agencies use portable ultrasonic testers operated by NDE engineers to determine the extent of corrosion or cracking where potential problems were noted during visual inspections by technicians. One response also indicated use of ultrasound to find flaws in concrete. Table 3.3 sununarizes the ultrasonic applications. According to Table 3.4, all 37 ultrasonic responses include some sort of formal training. Of the 21 responses related to magnetic testing (MT), a significant amount of magnetic particle testing is used as an aid to visual inspections. Applications include checking for internal or surface flaws either in structural steel members or welds. Table 3.5 shows the applications associated with magnetic testing. Magnaflux and yoke 35

probes appear to be the most widely used magnetic particle devices. Note that eddy current testing is also a magnetic technique; however, it was a separate category on the questionnaire. According to Table 3.6, approximately 50 percent of the responses have some sort of formal training. Five responses indicate that they contract out large-scale NDE testing. Several agencies use local NDE techniques (UT, MT, DP), but contract out large or complicated jobs. It is suspected that those who indicate that they "currently do not use NDE techniques" actually contract out their NDE work (most likely AE testing). For example, Wisconsin DOT indicated they don't use NDE techniques; however the literature showed that significant AE testing had been performed there [100]. Two DOT responses indicated that X-ray NDE is being used. One was used during in-house fabrication of steel beams for weld condition determination. The other used radiography during construction to also determine weld condition. One DOT had recently stopped radiography testing because of the strict regulations that were adopted in 1991. One industrial response indicated that X-rays were used to assist in any questionable NDE evaluations made at the plant. Three other industrial firms use X-ray NDE techniques. Die penetrants (13 responses) are used widely for enhancing visual inspections. Apparently, liquid die penetrant testing is also a widely used NDT technique.

International Summary Because only four agencies responded to the questionnaire, it is difficult to judge which methods are preferred. However, given that masonry and concrete structures seem to be the predominant building materials, techniques suited to concrete and masonry will probably prevail. The reported methods include: rebar locators, pin-hole permeability devices, fiber-optic scopes, eddy current, Schmidt hammers, and optical/camera devices. 36

Table 3.1. U.S. Responses by Testing Method Method Number

Ultrasonic Testing (UT) 37 Magnetic Testing (MT) 21 Voltmeter (VM) 4 Rebar Locator (RL) 6 Schmidt hammer (SH) 6 Dye Penetrant (DP) 13 Radiographic Testing (XR) 6 Eddy current Testing (ET) 6 Contract out NDE work (C) 6 Do not use NDE techniques 5 (N) Other (0) 7 37

Table 3.2. International Responses by Testing Method

Method Number

Ultrasonic Testing (UT) 0 Magnetic Testing (MT) 0 Voltmeter (VM) 0 Rebar Locator (RL) 3 Schmidt hammer (SH) 1 Dye Penetrant (DP) 0 Radiographic Testing (XR) 0 Eddy current Testing (ET) 1 Contract out NDE work (C) 0 Do not use NDE techniques 0 (N) Other (0) 1 Fiber scope 1 Pin hole (permeability) 1 Table 3.3. UT Applications in U.S. Application Number

Thickness 9 Internal flaws 16 Pin 4 Anchor bolts 4 Welds 15 Pin and 5 Hanger

Table 3.4. UT Training in U.S. •BBr^^^aBBssasBaB^B^aaaBnBBs^BBS Type Number

ASNT Level I 6 ASNT Level II 15 Formal Training 14 Req'd BS & Level II 1 Technical School 1 39

Table 3.5. MT Applications in U.S. Application Number

Surface Flaw 8 Sub-surface flaw 10 Welds

Table 3.6. MT Training in U.S. Type Number

Level II 3 None or on the Job 7 Formal Training 7 Req'd 40

Summary An NDE questionnaire was developed and sent to 58 organizations within the United States and 38 organizations within the international community; 90% of the United States organizations responded. Results from this questionnaire indicate that pulse-echo ultrasound and magnetic particle testing methods are widely used by public agencies. Liquid die-penetrant testing is also a popular method. Most large-scale NDE testing is contracted out; based on responses to the questionnaire, acoustic emission testing of state bridges is being done mostly by consultants. The literature suggests acoustic emission testing to be a application oriented technique and suitable to field investigations. The questionnaire was distributed mostly to state agencies - those responsible for inspecting and maintaining public infrastructure. Because of the practitioners fondness toward ultrasound, the experimental and theoretical aspects of this thesis are devoted toward ultrasound. Specifically, the use of DSSSUE is proposed. 41

CHAPTER 4. STRUCTURAL VBRATION THEORY Overview The area of defect identification in which a response to an input is used to determine the dynamic characteristics of a system has been discussed in Chapter 2 in the vibration analysis section. It is this area which supplies the link from a structural engineer's point of view to the area of ultrasonic testing using spread spectrum techniques. This chapter will introduce deterministic structural dynamics theory, extend it to a arbitiary situation, and relate it to the general DSSSUE theory.

Damped Single Degree of Freedom System Consider the case of a damped single degree of freedom (SDOF) subjected to a harmonic excitation as shown in Fig. 4.1. The differential equation of motion can be obtained by using D'Alembert's principle and sunmiing the forces on the free body diagram.

+ q}(t) + ky(t) = x(t) (4.1)

where m is the mass, c is the damping coefficient, k is the stiffness, x is the input, and y is the resulting output. The solution to this differential equation will be of the form

y(0 = >'«(0 + yp(t) (4.2) where yH(t) is the homogeneous or transient solution and yp(t) is the particular or steady state solution. The homogeneous solution to Eq. (4.1) is of the form

yff (/) = e [/4C0S Wg f + 5sin o)^ f] (4.3) where

(4.4) 2\/km 42

Figure 4.1 Single Degree of Freedom System 43

(ùjj = (i)\/l - (4'5)

w = ,f (4.6) \ m

The particular solution for a sinusoidal driving function of frequency w,

x(t) = e'^' (4.7)

is an out-of-phase harmonic output. The particular solution will be of the form

yp(t) = H((ù) e'"' (4.8)

where the transfer function is

. «) ""

and

r„ = - (4.10) b)

The general solution is the sum of the homogeneous solution Eq. (4.3) and the particular solution Eq. (4.8). The constants A and B are determined from initial conditions using the general solutions in Eq. (4.2). The presence of the exponential (damping) in Eq. (4.3) will cause the transient component of the response to vanish after some time, leaving only the steady state solution given by Eq. (4.8).

Response Due to Arbitrary Input The previous section was concerned with the analysis of a linear structural system where the excitations are periodic. Many times the excitation function will have an irregular shape. An approximate method can be utilized to calculate relationships 44

between the irregular input x(t) and the irregular output y(t). The impulse response function h(t) gives the response at time t due to a unit response at t = 0, that is h(t) is the response when

xit) = S(t) (4.11)

where 6(t), the Dirac delta function, is defined as

S(t) = 0 (t^O)

+ 00 (4.12) f S(t)dt = 1

-00

For example, for the single degree of freedom system in Eq. (4.1)

0 /SO (4.13) h(t) = e ^"'sinWgf f & 0 ma.

Similarly, h(t - T) is a response at time t due to an impulse at r. Fig. 4.2 shows that a general input force x(t) can be divided into several small pulses at time r of area [92]

Pulse Magnitude = x{T)dr (4.14)

The response at time t due to this small pulse at r is given by

h(t - T)x(T)dr (4.15)

Now, because superposition applies for a linear system, the total response at t is the sum of all the individual pulses as the limit of the pulse width goes to zero I y(t) = - T)K(T)dT (4.16)

which is commonly known as the convolution integral. An alternate form of the 45

X

X x+dx

Figure 4.2 Arbitrary Loading 46

convolution integral can be seen by noting there is no response for t less than r. In other words there is no response because there has been no impulse applied or

h(t - T) = 0 (t < T) (4.17)

The upper limit of the integral can be extended to infinity without changing the result. Introducing a change of variable

0 = t ~ T (4.18)

Eq (4.16) can be rewritten as

•fOQ >(0= fh(6Xt-B)de (4.19)

-00

which is just another form of the convolution integral. Taking the Fourier Transform (FT) of the response in the time domain produces the response in the frequency domain, as shown in Appendix B,

4-00 +00 y(w) = ^/ [ - 0)dd]e-'^'dt (4.20)

-00 -00

Reintroducing Eq. (4.18) as

t = T + 0 , dr = dt (4.21)

Keeping 6 constant and rearranging

+00 *00 y(w) = fh(e)dee-"^<'[—fx(T)e-'^'dT] (4.22)

-00 -00 or, the solution in the frequency domain can be written as

y(w) = H((ù)X((ù) (4.23) in which X(w) is the FT of the input x(t) and H(w) is the FT of the impulse response 47

function h(t). Eq. (4.23) is a very important relationship in the frequency domain between the input X(w), output Y(w), and the frequency response function H(w). Eq. (4.23) brings out the very important property of convolution. A time domain convolution is equivalent to multiplication in the frequency domain. Using similar thinking, it can be shown that convolution in the frequency domain is equivalent to multiplication in the time domain.

DSSSUE Defect identification, where the response to an intentionally applied random input is recorded, is an effective way to predict the dynamic characteristics of a system. The need for measuring the dynamic characteristics of a system arises from the idea that if the system Anger print or signature is known, changes in this fmger print can indicate distress to the system (see Fig. 2.2). Spread spectrum ultrasound operates on this concept and uses a random-like continuous transmitted signal. The general theory to this technique is outlined below. The description of the random-like signal is discussed towards the end of this chapter. The cross-correlation between the random input x(t) and random output y(t) to a linear system is defmed by [92]

1 ^ /(^(T) = -fx(t - T)y(t)dt (4.24) •' 0 where T is time over which data was recorded. The convolution integral (Eq. (4.19)) can again be written as

+00 y(l) = - W C M)

-00 where ^ is again a time shift variable. Substituting Eq. (4.25) into Eq. (4.24) yields

T +00 w = !>(( - T)[/A(CW' - (XW (4.26) •' 0 -00 48

Switching the order of integration and rearranging

• 00 7" RJj) = I/ /''(fW - CW - (4.27) -mo

Introducing another change of variable

9 = t - r , dd = dt (4.28)

+00 T «^(T) = ^/ Jh((X»*'r-C>(0)ded( (4.29)

•' -X 0

The autocorrelation function is defined as [92]

1 ^ RJt) = ^J.K(0+T)*(0)rfd (4.30)

which, by replacing T with r - F, can be rewritten as

1 ^ - 0 = ^fx(0^'r-C)x(e)de (4.31) 0

which, with Eq. (4.29) leads to the final version of the input to output cross correlation

+ 00 - CMC (4-32)

-00

In the special case where the auto-correlation function of the input is a delta function, i.e.,

RJO = 5(0 (4.33) 49 and using Eq. (4.12) it is seen that

= H(T) (4.34)

in other words, the cross correlation between the input and the output will give the impulse response Ainction of the system if the autocorrelation function of the input, Rxx(r), is a delta function 6(r). In the frequency domain, the cross spectral density can be expressed as a complex conjugate multiplication which is given by [92]

5^(0>) = %(w) y(w) (4.35)

where X(«r) and Y(w) are the FT of x(t) and y(t) respectfully. An equivalent cross spectral density also exists which is given by [92]

5^(<ô) = //(c5)5«(«) (4.36)

where the power spectral density of the input, S^^(w), is the FT of the auto correlation function; and S,y(tiT) is the FT of Rxy(T). For the special case of Eq. (4.33), the power spectral density of the input is

(4.37) that is, it is constant which is one definition of white noise. The cross spectral density for the special case is

(4.38) where H(tDr) is from Eq. (4.9). 50

White noise is a random process whose autocorrelation function is a delta function [92]. A sample excitation and autocorrelation function for such a process is shown in Fig. 4.3. The power spectral density S** (or), which is constant at So, is also shown in Fig. 4.3. A simple example utilizing the white noise input is illustrated next. Suppose the SDOF system in Fig. 4.4 is subjected to a normally distributed white noise input shown in Fig. 4.3. The absolute value of the transfer function (Eq. (4.9) is

1 |//(w)| = (4.39)

The absolute value of the cross spectral density, as found from Eq. (4.38), is plotted in Fig. 4.5 as the W/O flaw line. The phase angle, 6, for cross spectral density is

2r £ tan(g) = —^ (4.40)

which is plotted in Fig. 4.6 as the W/O flaw line. As the system deteriorates, the stiffness of the system will change as mentioned in the vibration signature analysis section of the literature review in Chapter 2. In this example, if the stiffness is decreased by 10 %, the corresponding cross spectral density and 0 plot are shown overlaid as the W/ flaw line in Fig. 4.5 and Fig. 4.6, respectfully. The amplitudes and corresponding natural frequencies of the system have changed. This is the fundamental principal of the DSSSUE signature analysis. In the time domain, taking the inverse transform of Sxy(w) gives the approximate impulse response function h(t) of the no flaw and flawed structure. This is shown in Fig. 4.7. 51

x(t)

co^

Figure 4.3 Properties of White Noise 52

m B8 100 Ib-sec^inch c p 632 Ib-sec/inch x(t) = white noise, (ib.) I $0^271 Ib.-in. - sec. le s 100,000 ib/incii y(t) = displacement output (inch)

Figure 4.4 Example of Single Degree of Freedom System 53

xy h.- lb - stc 5E-6 w/o flaw

4E-6 w/ lliw

3E-6

2E-6

IE-6

OE+0 (80) (60) (40) (20) 0 _20 40 60 80 Fraqaency, ^

Figure 4.5 Graphical Representation of Eq. (4.38) 54

e Radians 2

f.5 w/flaw

1

0.5

0

(0.5)

(0

(2) 1 ' 1 ' ^ (80) (60) (40) (20) 20 40 60 80 FraquMKy, nd/sac

Figure 4.6 Graphical Representation of Eq. (4.40) 55

3E-4 w/o flaw w/flaw 2E-4

IE 4

OE+0

•fE-4

-2E-4 0.3 0.4 ThM (saconds)

Figure 4.7 Graphical Representation of Eq. (4.13) 56

Random Binary Process Since the process of white noise is only a theoretical concept an alternate technique to determine a broad band power spectral density must be found. A random binary process is such a process. Consider a stationary ergodic process [92] where by an event (coin flip) occurs at a regular interval At. A stationary process is a process in which the probabilistic properties do not depend on time. An ergodic system is a property where by any sample average will be representative of the entire process. When a "head" event is achieved in a process, x(t) takes on the value +1 for the interval At; if it is a "tails" it is -1 for the interval. If this event would go on for a "long" time, the autocorrelation process for this process x(t) would be as shown in Fig. 4.8 [92]. The corresponding spectral density can be found by taking the transform (see Appendix B)

(4.41) which is shown in Fig. 4.8. As the clock frequency (At) gets shorter, the broader the random binary process spectrum band and the closer approximation to white noise is approached. The concept of "long" coin flipping is not a practical means of producing a white noise-like signal. An alternative method of determining this signal is preparing a finite length binary sequence in advance and, subsequently, repeating it. The signal becomes periodic and repeats itself with period (length of signal) NAt where N is the number of coin flips. Such a generation of a process is termed a pseudo-random binary process. The pseudo-noise is a good means of producing an approximate "white noise" process. The autocorrelation function of such a process will have the same shape as Fig. 4.8; however, it will repeat itself with period NAt as shown in Fig. 4.9. The pseudo­ random power spectrum S'„ , called a line spectrum, is given by [92] 57

where S„(tJ7) is the random binary power spectrum shown in Fig. 4.8. In other words, it has the same shape and it differs only by the constant given in Eq. (4.42). The pseudo random binary sequence is the fundamental input to the DSSSUE system. If the input is a pseudo-noise and its autocorrelation function approaches a delta function, the input-output cross coirelation will be representative of the impulse response (Eq. (4.34)). This cross correlation signature is the basis for condition assessment of a structural member or mechanical part. 58

$xx i

CO Ù)

Figure 4.8 Random Binary Process 59

*xx

X

Figure 4.9 Pseudo Random Binary Process 60

CHAPTER 5. COMPONENT IDENTIFICATION AND FUNCTION Overview The previous chapter gave a basic idea of structural dynamics theory and the relationship to the spread spectrum method. This chapter describes the equipment and function of the equipment needed to make the DSSSUE signal. There are two approaches to the DSSSUE signal generation [59]. The first and more compact approach is called the maximal hardware approach (MHA). In this technique, the signal generation, amplification, and cross correlation signature are performed in one unit. The MHA will be the more portable technique and, thus, will be more useful in the field application stage. The second approach, called the maximal software approach (MSA) is more suitable for laboratory work. In this technique, separate units perform each of the stages of DSSSUE transmission, reception, and post processing. This chapter describes the equipment associated with MSA DSSSUE.

Description The general flow chart diagram for the MSA is shown in Fig. 5.1. The components consist of a computer, arbitrary function generator, power amplifier, low- noise pre-amplifier, and an oscilloscope. These components provide the mechanisms for DSSSUE signal generation. The general component layout is shown in Fig. 5.2. The next seven sections describe the functions that each of these components perform.

Personal Computer The MSA starts with two computer generated signals. The first signal is an extremely long pseudo random binary signal which is a wide band frequency signal centered at the origin that has an autocorrelation function shown in Fig. 4.9 which approaches a delta function. The second signal is a discrete sinusoid signal at the center frequency of the transmitting and receiving transducer. The frequency of the sinusoid signal is twice that of the pseudo random binary signal. These two signals in time are shown in Fig. 5.3 and Fig. 5.4, respectfully. These two input signals are multiplied 61

Arb Function Gen.

Scope

Figure 5.1 Basic DSSSUE Flow Diagram 62

UcwAjWey PwKdeii

Computer %% Inputt Wwifonii K#wd OM»MOMI I # T • • \ Power Amplifer ] SSS7 / On // / Input- Output -—%' y y/g T (

Figure 5.2 DSSSUE Equipment 63

(0.5)

(1.5)

Figure 5.3 Pseudo Binary Process 1.5

0.5

(0.5)

(1)

(1.5) 65

together to give a wide band signal centered at the frequency of the transducers. The discrete signal in time is shown in Fig. 5.5 ~ this is the continuous signal which is used to drive the transducer. The choice of the length of the long pseudo random binary signal must be such that the operating characteristics of the transducer are not exceeded. The longer the pseudo signal, the closer an approximation to the white noise; however, more computational time is needed to post process the results. In addition, the longer the pseudo signal, the more sensitive the signal is to changes in the member or part that is being inspected. The typical length of pseudo signals that are used in this research include a 11 bit code which has 2,047 coin flips and 8191 coin flips respectfully ' . In digitizing the data, each coin flip of the pseudo signal is sampled 40 times giving the total number of points in the 11 bit and 13 bit codes to be 81,880 and 327,640, respectfully. The main lobe of the spectrum of the combined pseudo signal and continuous sine wave must be smaller than or equal to the maximum operational limitations of the transducer as shown in Fig. 5.6.

Arbitrary Function Generator The LeCroy Model-9101 arbitrary function generator shown in Fig. 5.2 serves to make a continuous time waveform signal out of the discrete time signal created by the computer. This unit converts the digital voltage signals to the analog voltage wave.

Power Amplifier The power amplifier shown in Fig. 5.2 magnifies the continuous wave received from the function generator to an acceptable voltage that the transducer can accept. This concept is illustrated pictorially in Fig. 5.6.

' See Chapter 4 Random Binary Process Section. 1.5

1

0.5

0

(0.5)

(O

(15), 67

Transducer Main Lobe Spectrum

Of) Center Frequency of Transducer Figure 5.6 DSSSUE Spectral Density 68

Transmitting Transducer The transmitting trajisducer converts the continuous amplified DSSSUE voltage signal into high frequency mechanical vibrations. The continuous amplified DSSSUE signal is sent into the piezoelectric crystal inside the transducer which transmits the mechanical energy into the structural member or mechanical part.

Receiving Transducer The receiving transducer accepts the continuous transmitted signal after it has been sent through the structural member or mechanical part. The receiving transducer acts in the reverse order of the transmitting transducer.

Low-noise Fre-amplifier The Panametrics pre-amplifier shown in Fig. 5.2 is a device which is needed to magnify the received signal while adding as little circuit noise as possible to the setup.

Oscilloscope The Lecroy Model 9310L multichannel digitizing oscilloscope shown in Fig. 5.2 is able to digitize and display two channels at one time. The scope also has the capability to perform transforms to display in time or frequency, but this capability is not used. The Lecroy scope accepts the continuous received signal and digitizes it so it can be uploaded to the computer for post processing.

Post Processing The entire process terminates at the personal computer which is able to communicate with both the oscilloscope and function generator via interfaces built into these units (GPIB or IEEE-488-1987). This is where post processing of the data can occur. Specifically, the input signal and the output signal are cross-correlated to obtain the approximate impulse response of the structure or mechanical part. Consider two sets of records x(t) and y(t) each being of 69

length N. Matlab software [82] is used to post process the received and transmitted signal. To produce the cross correlation signature in the time domain, the input record, x(t) is multiplied with the output y(t) to produce each data point in the cross-correlation signature by shifting one record with respect to the other and multiplying which is similar to Eq. (4.24). This shifting process is called code delay which will be used in the plots which are shown throughout this thesis. Each code delay is 1 nano second. As an example, to produce the first point in the cross-correlation signature, x(t) and y(t) would be multiplied together without any shift (zero code delay), i.e. a simple multiplication of two vectors of length N. Next y(t) would be shifted by one record (code delay) and each of the records again would be multiplied to obtain the second point in the cross correlation signature. This would go on for N shifts or N code delays. Because this process involves N^ multiplication steps, an equivalent and more efficient response can be obtained in the frequency domain. First transform each of the records y(t) and x(t) into the frequency domain which will yeild Y(w) and X(w). Next, multiply the two records to obtain the entire cross spectral density signature in one process which is similar to Eq. (4.35). For example, the product of the first records in the frequency domain, Y(W|) and X(W|) will give the first point of the cross spectral density. Next, Y(w2) would be multiplied by Xiui) to produce the second data point of the cross spectral density. Note, multiplication of these records implies complex conjugate because both of the records include both real and imaginary components. This goes on for all N records which will produce the entire cross spectral density in just N multiplication steps. Transforming this back into time will produce the equivalent record in the time domain as previously described. 70

CHAPTER 6. EXPERIMENTAL RESULTS This chapter presents the results of a set experiments which have been performed with the DSSSUE equipment. Chapter S detailed the equipment associated with the MSA DSSSUE technique. In all of the experimental results presented in this chapter, the basic layout of the equipment shown in Fig. 5.1 and Fig. 5.2 were utilized with the exception of the arbitrary function generator shown in Fig. 5.2. During the course of the experimental work, uncertainty of the signal produced by this device led to utilizing the equipment of the MHA technique to generate the DSSSUE continuous signal [59]. Testing of bars embedded into concrete was fu-st attempted with pulse echo ultrasound*. Chapter 2 briefly explains the concept behind pulse echo ultrasound. Appendix C illustrates some of the equipment used and samples that were tested. These initial tests showed that any concrete surrounding the steel had no effect on ultrasound once the bond has deteriorated between the steel and concrete. This is usually the case for embedded anchor bars on lock and dams; therefore, all further testing involved testing bars without the surrounding concrete. The first laboratory experiment consists of a sensitivity study in which a standard calibrated test block with holes was partially filled with water. Baseline signatures were obtained. Next incremental signatures were obtained by adding additional drops of water to each of the holes with an eye dropper to simulate distress or change to the system. The second experiment consists of three 18 in. long round steel rods which were specifically chosen to simulate the deteriorated round anchor bars that are shown in Fig. A.3. Transducers were mounted on each end and baseline signatures were obtained. Next the bars were distressed to varying degrees starting with very small imperfections and comparisons were made to the baseline signatures.

' Assisted by Scientist David Hsu of the CNDE center. 71

Laboratory Experiment Number 1 The first experiment was specifically designed to determine the sensitivity of the DSSSUE technique and was called the water drop experiment^. Earlier experimental work has indicated that the DSSSUE technique is very sensitive to transducer placement. To avoid this, transducers were permanently attached to the DSSSUE test block shown in Fig. 6.1. By permanently attaching the transducers, the cross correlation signature effects of transducer placement (registration) in both the X and Y directions as well as radial (6) directions are eliminated (see Fig. 6.1). In addition, error associated with the mineral oil couplant that is used to transmit the ultrasonic wave into the specimen is eliminated. This means that the water drop experiment should give the best experimental results that the system is capable of producing for a given length of transmitted signal. An example of the first 50,000 data points in the cross correlation signature utilizing a 13 bit transmitted signal with 25 drops of water in each hole and 30 drops in each hole are shown in Fig. 6.2. In other words, this represents a total change in the system of 10 water drops. The ordinate is the normalized amplitude, scaled by the maximum cross-correlation value in the record, and the abscissa is the code delay in nano seconds, which was defined in the Post Processing Section of Chapter 5. At first glance, each of the records shown in Fig. 6.2 appear random. However, they are not. Under closer observation, changes in the amplitudes are present. This can be seen by zooming into a region and by taking differences of the signatures with respect to the baseline condition. Signatures were obtained after adding one, three, and five drops of water to each hole which already contained 25 drops of water. Signature differences were obtained by subtracting the incremental water signatures from the baseline of 25 drops/hole. These differential signatures in the code delay range of 10,000 to 20,000 are shown in Fig. 6.3. Fig. 6.3a represents the current limitations of the measurement - it is the difference between two different baseline

^ Assisted by Ph.D. graduate student Sangmin Bae of the Electrical Engineering Dept. 72

1/4 in.

6 In.

^ \ Transducer Registration

Figure 6.1 DSSSUE Calibration Block 73

i"

I"8-0.5

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 code delay X 10"

a) Baseline cross correlation signature (25 water drops in each hole)

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

code delay .|q4

b) 10 additional drops of water to the calibration block (30 water drops in each hole)

Fig. 6.2 Water drop cross-correlation signature (25 and 30 water drops in each hole). 74

signatures and is a measure of noise in the system when a 13 bit code is utilized. The lower three plots (Fig. 6.3b, Fig. 6.3c, Fig. 6.3d) represent 26 drops/hole minus the baseline (2 drop change), 28 drops/hole minus the baseline (6 drop change), and 30 drops/hole minus the baseline (10 drop change), respectfully. Although not directly applicable, the water drop test shows the sensitivity of the DSSSUE technique. As shown in Fig. 6.3, the technique can easily detect the volume change of an additional drop of water in each of the holes. The signal for one drop in each hole is significantly larger than the noise.

Laboratory Experiment Number 2 The fmal experiment was specifically designed to simulate the deteriorated round anchor bars that are shown in Fig. A.3 \ The specimens were chosen to be of simple geometry to reduce the number of parameters. The bars are shown in Fig. 6.4. To reduce sensitivity due to transducer placement, transducer recess housings were milled into the ends of the bars as shown in Fig. 6.4. The transducer recess housings restricted movement in the X and Y directions, but not the 6 direction. A mark was placed on the bar to ensure, as best as possible, rotational repeatability (6) of the transducer placement. A large C-clamp was used to hold both transducers in place. A mark was made on the threading of the C-clamp to help repeat transducer pressure. Mineral oil couplant was placed on the transducers before placing them in the recess areas. In this experiment, errors due to removing the transducers and getting them precisely to the same location, the same C clamp pressure, and the amount of couplant, will be present in each cross correlation signature. In this sense, the accuracy will be poorer than that obtained in the water drop experiment which had permanently attached transducers. In addition, a shorter transmitted signal was used (11 bit). Three different

^ Assisted by Ph.D. graduate student MAK Afzal of the Electrical Engineering Dept. 75

1 1 1 1 1 1 1 1 1 i d 0 c-o.l _L _L J_ 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 code delay X 10 a) Differential of two baseline signatures

9- 0.1

c-o.l

X 10 b) Differential of 2 drops referenced to the baseline signature

9- 0.1

code delay X 10 c) Differential of 6 drops referenced to the baseline signature

9- 0.1

code delay X 10 d) Differential of 10 drops referenced to the baseline signature

Figure 6.3 Differential signatures of Water Drop Experiment 76

Transducer Recess (each end) Registration

X

18 In.

a) Baseline

Flaw depth (d)

b) Distressed

Figure 6.4 Bar Specimen Which Simulates a Round Anchor Bar (see Table 6.1 for flaw depths) Table 6.1 Bar Flaw Depth (d) Data (see Fig. 6.5)

Bar # Baseline (in.) Flaw 1 (in.) Flaw 2 (in.) 1 0.0125 0.100 0.200 2 0 0.0125 0.025 3 0.025 0.050 0.100

bars were used and each was distressed by milling in a groove at mid-length to different depths which are given in Table 6.1.

Typical Cross-correlation signature (Bar 1) Fig. 6.5 represents the complete cross-correlation signature utilizing a 11 bit transmitted signal for Bar number 1 at two different stages of milling. The baseline condition signature for the bar with a milled groove of 0.0125 in. at the bar mid length is shown in Fig. 6.5a. The signature for a milled groove of 0.100 in., which is approximately a 20% cross-sectional area reduction, is shown in Fig. 6.5b. As in the water drop experiment, the ordinate is normalized amplitude, scaled by the maximum value in the record, and the abscissa is code delay. For the case of simple geometries like a bar with transducers on each end, the code delay can be interpreted as transit time as the continuous signal propagates through the bar. In this sense, the bar can be used to relate signature activity to flaw location. As discussed in Chapter 5, each code delay is 10'^ seconds (one nano-second). For more complicated geometries and transducer placement like the calibration block, it is more difficult to distinguish the code delay in terms of transit time of the continuous signal propagating through the object. Referring to Fig. 6.5a, note that there are specific groups of signature activity occurring around 0 code delay, 16,000 code delay, 32,000 code delay, etc. Since the signatures approximately represent responses due to unit impulses (see chapter 4), these signature groupings can be interpreted as a unit impulse wave ringing back and forth in the bar. 78

1 1 I

à"§ 0.5 "S. .1.. F" 1 rrt ' " 1 rg-0.5 2

-1. 1 .... 4 5 code delay X 10 a) Bar 1 Baseline cross-correlation (d=0.0125")

1 •S 0.5

7"'lu '1It . i . - .

Ig-0.5 8

-1 4 5 code delay X 10

b) Bar 1 flaw 1 cross-correlation (d=0.100")

Figure 6.5 Entire cross-correlation signature for Bar 1 79

At the zero code delay ^ position, the unit impulse has gone from the transmitting transducer, through the bar, and to the receiving transducer. The pulse then reflects off the receiving end, travels back through the bar, reflects off the transmitting end, and travels back to the receiving transducer. This is the second group of signature activity which appears around the 16,000 code delay region. Using the longitudinal speed of sound in steel to be 234,646 in/sec, the equivalent code delay for this unit pulse to travel up and back through the length of the bar is 15,342 nano seconds. Looking at Fig. 6.Sa, one can see that this corresponds to the first spike of activity near the 15,000 to 16,000 code delay mark. Comparing Fig. 6.5b to Fig. 6.5a, one can see that the same series of signature groups occur at the 0 code delay, 16,000 code delay, and 32,000 code delay regions. However, there are differences in the cross-correlation signature as well. Looking in the 7,500 to 15,000 code delay region, obvious activity is present in Fig. 6.5b that is not in Fig. 6.5a. This activity is indicative that a change of some sort has occurred in the bar. Going a step further, referring to Fig. 6.4, the flaw in the bar was placed at the midpoint of the bar which is 9 inches from the end. Using logic similar in the previous paragraph, one would expect the flaw to introduce cross correlation amplitude activity at the mid-point between the initial signal and the first reflection. This would theoretically be at 7,671 code delay or half of the 15,342 code delay which was previously calculated. This peak, which appears at about the 8,000 code delay mark, is indicative of a flaw in the bar at the 9 inch mark. This method of comparing signatures to a baseline condition to detect structural changes is the fundamental concept of signature analysis (see Chapter 2 and Chapter 4). As noted above, the region of interest for comparing the two bar signatures was found to be in the 7,500 to 15,000 code delay region. The next three sections will present the cross-correlation signatures in this region for the three bars in the baseline.

* Zero code delay implies TQ code delay. In all the bar cross correlation plots, the data was shifted to the left to remove the section of no correlation signature activity. 80

flaw 1, and flaw 2 distress conditions as shown in Table 6.1.

Bar 1 Analysis Fig. 6.6 contains the 3 cross-correlation signatures in the ranges of 7,500 to 15,000 code delay for Bar 1. For example. Fig. 6.6a and Fig. 6.6b are taken directly from Fig. 6.5a and Fig. 6.5b, respectfully. Fig. 6.6c represents approximately a 40% cross-sectional area reduction (d=0.200 in.) when compared to the baseline case. In each of the plots shown in Fig. 6.6, an increased level of cross-correlation amplitude appears for each of the increasing flaw sizes. In fact, comparison of Fig. 6.6c to Fig. 6.6b corresponds to roughly a two to one amplitude increase on the cross-correlation signature which corresponds approximately to the flaw depth increase. Further analysis can be performed by subtracting the baseline case from the distressed signatures. Fig. 6.7 shows differences between two baseline signatures. This low activity of difference shows good repeatability of the experiment. It should be noted; however, that in Fig. 6.7, the transducers were not removed. This is why only one plot is shown in Fig. 6.7 because it is not equivalent to the data shown in Fig. 6.8. The effect of the noise due to removing the transducers while the bar was brought to the machinist for milling is illustrated in Fig. 6.8a and Fig. 6.8b. Figure 6.8 shows differences between the baseline and flaw 1 and the baseline and flaw 2 condition. Again, note that the differential amplitude for Flaw 2 (Fig. 6.8b) is about twice that of Flaw 2 (Fig. 6.8a), especially in the 11,(XX) to 14,000 code delay region. However in the regions near 8,0(X) and 9,0(X) code delay the ratio appears slightly larger. Subsequent tests were performed to study the effects of transducer registration^ Six data sets were taken which had four different registrations (three data sets had the same transducer registration). Cross correlation signature analysis was performed on all six records. There was no visual difference in the signatures themselves between the 7,500 to 15,0(X) code delay range. (This is equivalent to looking at Fig. 6.6 and

^ Assisted by MAK Afzal (plots not included in this dissertation) 81

a 0.5 I " g H).7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 code delay X 10 4 a) Bar 1 baseline cross-correlation (d=0.0125")

a. 0.5

Jkl p*' " I" f 0.8 0.9 1.1 1.2 1.3 .1.4 1.5 4 code delay X 10 b) Bar 1 flaw 1 cross-correlation (d=0.100")

a 0.5

I 0 4 —wk- -0, ?7 0.8 0.9 1.1 1.2 1.3 1.4 1.5 code delay X 10 4 c) Bar 1 flaw 2 cross-conelation (d=0.200")

Figure 6.6 Bar 1 cross-correlation signature from 7,500 to 15,000 82

0.5 1 r -i r

J I I L •"•g- 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 4 code delay X 10

Figure 6.7 Bar 1 differential cross correlation signature from 7,500 to 15,000 (baseline - baseline) 83

0.5 — — - - 1 T - I i14 i^Ai fW ' i V f F

1 •"•87 0.8 0.9 1.1 1.2 1.3 1.4 1.5 4 code delay X 10 a) Bar 1 differential (flaw 1 - baseline)

0.5

"—|h—Wk !•E

0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 4 code delay X 10 b) Bar 1 differential (flaw 2 - baseline)

Figure 6.8 Bar 1 differential cross-correlation signature from 7,500 to 15,000 84

visually (subjectively) seeing differences in the signatures themselves). The error comes into play when differences are performed (Like Fig 6.7 and Fig. 6.8). Signatures with the same registration (Fig. 6.7) cancel out much better than the ones which have different registration (Fig. 6.8). This experiment showed that if higher sensitivity is desired and differential tests are performed, one may need to permanently mount the transducers to the specimen. More work is needed to quantify these effects.

Bar 2 Analysis The previous section had distress ranges from baseline to a 20% cross-sectional area reduction, and to a 40% cross-sectional area reduction. Bar 2 has a much smaller flaw differential: baseline (d=0.0 in.), a 3% cross-sectional area reduction (d=0.0125 in.), and a 6% cross-sectional area reduction (d=0.025 in.). Again the flaws differ by a factor of two. Fig. 6.9 contains the three cross-correlation signatures in the ranges of 7,500 to 15,000 for Bar 2. In each of the plots, an increased level of signature activity can be seen - especially in the regions from 11,000 to 14,000 code delay. Fig. 6.10 shows difference between baseline signatures. Again low activity was present which shows good repeatability when transducers are not removed. As in the Bar 1 example, differences between the distressed signatures and the baseline case were performed which is shown in Fig. 6.11. Again note, that in both Fig. 6.9 and Fig. 6.11, the differential amplitude has increased by a factor of two.

Bar 3 Analysis The set of tests on Bar 3 had flaws that were in between the flaws introduced in Bar 1 and 2. Bar 3 had flaws of: baseline (d=0.025 in.), a 6% cross-sectional area reduction (d=0.025 in ), and a 17% cross-sectional area reduction (d=0.050 in.). As in the previous experiments, the flaws differ by a factor of two. Fig. 6.12 contains the 3 cross-correlation signatures in the ranges of 7,500 to 15,000 for bar 3. As in the previous cases, larger amplitudes in this region correspond 85

Q. 0.1 %

1.1 1.2 code delay X 10 a) Bar 2 baseline cross-correlation (d=0.0")

Q. 0.1 %

1.1 1.2 code delay X 10 b) Bar 2 flaw 1 cross-correlation (d=0.0125")

a 0.1 1 1 t 1 I N|N/|/^yy>^V>rVV^^ °-0. i i i I

code delay X 10 c) Bar 2 flaw 2 cross-correlation (d=0.0250")

Figure 6.9 Bar 2 cross-correlation signature from 7,500 to 15,000 86

0.1 -i r 1 r -i r (D Î 0.05 "8 0 .M J i •0.05 g

j L j L -H 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 4 code delay X 10

Figure 6.10 Bar 2 differential cross correlation signature from 7,500 to 15,000 (baseline - baseline) 87

— . 0.1 r 1 • " • I r 1 I •S i 0.05 Q. I ' § -0.05

-0, i 1 1.7 0.6 0.9 1.1 1.2 1.3 1.4 1.5 code delay X 10 4 a) Bar 2 differential (flaw 1 - baseline)

0.1 —r 1 ••• 1 ' ' ' •S I 0.05

- wvMvyvY.

Ig -0.05" 8 -0. i 1 I).7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 code delay X 10 4 b) Bar 2 differential (flaw 2 - baseline)

Figure 6.11 Bar 2 differential cross-correlation signature from 7,500 to 15,000 88

Ia 0.1= 1 1 1 1 c-O i i i 4.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 4 code delay X 10 a) Bar 3 baseline cross-correlation (d=0.025")

a 0.1 I « °-0 •H? 0.8 0.9 1.1 1.2 1.3 1.4 1.5 code delay X 10 4 b) Bar 3 flaw 1 cross-correlation (d=0.050")

a 0.1 %

1.1 code delay c) Bar 3 flaw 2 cross-correlation (d=0.100")

Figure 6,12 Bar 3 cross-correlation signature from 7,500 to 15,000 89

to a increased level of distress. Fig. 6.13 shows difference between baseline signatures. Again low activity was present which shows good repeatability when transducers are not removed. Differences between the distressed signatures and the baseline case are shown in Fig. 6.14. As in the Bar 1 and 2 analysis, the differential amplitude on the cross- correlation has increased in the approximate proportions of the flaw increase, that is, by a factor of 2, especially in the regions from 11,000 to 14,000 code delay.

Experimental Results Summary The experimental portion of this thesis demonstrates the sensitivity and flaw delectability characteristics of the DSSSUE technique. The experiments indicate that the DSSSUE technique may be feasible to inspect the embedded lock anchorage system nondestructively. It also appears that the inverse problem of flaw sizing may also be possible by comparing relative amplitudes of subsequent signatures. Both experiments produced acceptable results. That is, in all cases, changes due to the specimen were detectable by the DSSSUE technique. Differential signatures also produced evidence that the specimen has changed. There are many variables that play very important roles in the signature noise. These include: couplant, transducer pressure, registration (x, y, and 6), equipment noise, transducer wear, analog to digital (A to D) data conversion, local magnetic fields (computer, lights, etc.), human, etc. There are many ways of producing better results by using ingenuity in the signal processing. For example, there are many alignment techniques that can help improve differential experiments. In each of these tests, each signature was normalized by its own maximum value. No doubt, some information is lost by doing this. However, better differential results are obtained this way. In other words, there is a trade off between getting good differential or good stand alone signatures. Other amplitude scaling techniques may produce better results in signature comparison, including: maximum ensemble scaling, cross correlate the cross correlation signatures, and variance scaling. 90

0.1 @ 0.05 I i -0.05 g

'H 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 code delay X 104

Figure 6.13 Bar 3 differential cross correlation signature from 7,500 to 15,000 (baseline - baseline) 91

0.1 — - - r •• • r — • 1 " •S .•i 0.05 n L,,,, , . - I " It' • t- - ip.. u , ,ÉànMt t -0.05

-0. i.7 0.8 0.9 1.1 1.2 1.3 1.4 1.5 code delay X 10 4 a) Bar 3 differential (flaw 1 - baseline)

0.1

i -0.05

0.8 0.9 1.2 1.3 1.4 1.5 code delay ,4 X 10 b) Bar 3 differential (flaw 2 - baseline)

Figure 6.14 Bar 3 differential cross-correlation signature from 7,500 to 15,000 92

7. SUMMARY, CONCLUSIONS AND RECOMMENDATIONS Summary Improved inspection techniques are needed for our deteriorating infrastructure. Recently, many reports have been published concerning the state of deterioration of public works systems. Statistics recently revealed that nearly half of the nation's bridges are either structurally or functionally inadequate. Another example of deteriorating infrastructure occurs in our nation's waterways. Approximately one-half of the Corps' 269 lock chambers along inland waterways will reach or exceed their 50- year design life by the turn of the century. Without on-going inspection programs to detect maintenance problems, unwanted down time will be required to solve several neglected problems. Nondestructive evaluation (NDE) methods can help address these problems. A literature review of current nondestructive evaluation (NDE) techniques applicable to large- scale, civil engineering structures is presented. The literature reveals that acoustic emission testing is the most reliable state-of-the-art technique for monitoring conditions of bridges and other large or complicated structures. This method was developed from extensive laboratory and field testing in the mid-1970s to mid-1980s and is currently in the application stage. Thermal techniques have been applied to several civil engineering projects-primarily asphalt and concrete pavement condition assessments. Similar to acoustic emission testing, thermal methods are making the transition from the research phase to the application phase. Although magnetic methods have been extensively tested in the laboratory, the field testing of these methods is only beginning. More than likely, the application of magnetic methods for monitoring civil works is years away and may never occur. A new ultrasonic technique, called Direct-Sequence, Spread-Spectrum, Ultrasonic Evaluation (DSSSUE), has recently been developed. This technique has the potential to become an effective global monitoring technique for civil engineering applications; however, it is still in the developmental stage. Other techniques, such as pulse-echo ultrasonic, electronic, and radiographic, are effective for local investigations. 93

An NDE questionnaire was developed and sent to 58 organizations within the United States and 38 organizations within the international community; 88% of the United States organizations responded. Results from this questionnaire indicate that pulse-echo ultrasound and magnetic particle testing methods are widely used by public agencies. Liquid die-penetrant testing is also a popular method. Most large-scale NDE testing is contracted out; based on responses to the questionnaire, acoustic emission testing of state bridges is being done mostly by consultants. The literature suggests acoustic emission testing to be a application oriented technique and suitable to field investigations. The questionnaire was distributed mostly to state agencies - those responsible for inspecting and maintaining public infrastructure. Because of the practitioners fondness toward ultrasound, the experimental and theoretical aspects of this thesis are devoted toward ultrasound. Specifically, the use of DSSSUE is proposed. This new method uses a continuous transmission technique that increases the detection sensitivity and "illuminates" the structural part as compared to conventional pulse-echo methods where scanning is required. The improved sensitivity can be used to overcome the signal attenuation in large structures and to detect small changes in the structures. It is anticipated that this method could eventually be used in a global lock and dam inspection and rating program and that the transducers and techniques developed in this research will have application in the field of continuous monitoring. Several tests were performed in two different experiments. First, A water drop experiment showed the sensitivity of the technique. Results indicated that the difference of one drop of water could be detected in the cross correlation signature. The second laboratory experiment consisted of three 18 in. long round steel rods. Transducers were mounted on each end and baseline signatures were obtained. Next the bars were distressed starting with very small imperfections and comparisons were made to the baseline signatures. Results indicated that even with the smallest of flaws, the DSSSUE technique was able to detect differences comparing to the baseline signature. 94

Conclusions There is still a need to bridge the technical gap between Civil Engineers and NDE engineers. When this is accomplished, the development and use of nondestructive evaluation techniques for assessment of various components of this country's infrastructure will significantly increase. At Iowa State University, significant progress towards achieving this goal has been initiated. HopeAilly, the DSSSUE technique will be eventually incoiporated into a global inspection program for civil works structures such as bridges, buildings, and locks and dams. The results of the experiments indicate that the DSSSUE technique may be feasible to inspect the embedded lock anchorage system nondestructively. This research was performed during on going system and equipment development process. It is expected that improvements in the postprocessing and equipment will bring forth even better data. The current flaw delectability presented in this thesis are certainly acceptable. Finally, the question still arises as to how good will this system perform under adverse field conditions.

Recommendations Field testing of the DSSSUE technique is the ultimate goal of this work. Field testing the DSSSUE technique on an actual lock and dam anchorage system that is scheduled for repair is recommended. Bridge testing of the technique could also be performed. But first, the equipment should be tested on structural sections of W-beams in the Town Engineering Structures Laboratory Further, rigorous analytical modeling of the DSSSUE technique utilizing fmite element techniques and a super computer is recommended. Also, ray tracing software that has been developed at the CNDE may provide a computer simulation tool. Finally, research between the Electrical Engineering Department and the Civil and Construction Engineering department should be continued. 95

BIBLIOGRAPHY

1. Afzal, M. A., and S. F. Russell. "Signal Processing Simulations for a Direct- Sequence Spread-Spectrum Ultrasonic Testing System. " Presented at 20th Annual Review of Progress: Quantitative Nondestructive Evaluation, August 1- 6, 1993, Bowdoin College, Brunswick, Maine.

2. Aggour, M. S., and H. M. Fouad. "Use of Nondestructive Testing Techniques for Bridge Inspection. " Proceedings of Nondestructive Evaluation of Civil Structures and Materials. University of Colorado, Boulder, Oct. 1990, pp. 315- 326.

3. Anisimov, V. K. "Possibility of Continuous Acoustic Inspection of Plate Material Structures." Soviet Journal of Nondestructive Testing. Vol. 27, No. 1, 1991, pp. 42-45.

4. Antimirov, M. Ya, A. A. Kolyshkin, and R. Vaillancourt. "Eddy Current Nondestructive Testing by a Perturbation Method. " Journal of Nondestructive Evalution. Vol. 10, No. 1, 1991, pp. 31-37.

5. Armstrong, D. M., A. Sibbald, G. M. Tomlinson, and M. C. Forde. "Modal Analysis of Model Concrete Road Slab. " Yet to be published. Civil Engineering Seminar, 1990.

6. Azmi, Mohammad. "Acoustic Emission Detection and Monitoring of Highway Bridge Components." Ph.D. Dissertation, University of Maryland, College Park, 1989.

7. Bae, S., and S. F. Russell. "Partial-Length Correlation Theory for Spread- Spectrum Ultrasonic Evaluation Applications." Presented at 20th Annual Review of Progress: Quantitative Nondestructive Evaluation, August 1-6, 1993, Bowdoin College, Brunswick, Maine.

8. Bakht, B., and L. G. Jaeger. "Bridge Testing-A Surprise Every Time." Journal of Structural Engineering. ASCE Vol. 116, No. 5, 19%, pp. 1370-1383.

9. Baston-Pitt, J. "Chlorides, Concrete and Radar-Bridging the Problem." Proceedings of Fifth International Conference on Structural Faults and Repair. June 29, 1993, Univ. of Edinburgh, pp. 83-88.

10. Bein, B. K., J. Gibkes, J. H. Gu, R. Huttner, and J. Pelzl. "Thermal Wave Characterization of Plasma-Facing Materials by IR Radiometry. " Journal of 96

Nuclear Materials. Vol. 191, No. 194, 1992, pp. 315-319.

11. Bingham, Allen H., C. W. Ek, and J. R. Tanner. "Acoustic Emission Testing of Aerial Devices and Associated Equipment Used in the Utility Industries. " ASTM, July 1992, pp. 1-82.

12. Birt, E. A. "Quantification of Corrosion Damage in Aircraft Skin Using a Novel X-Ray Radiography System. " Presented at 20th Annual Review of Progress: Quantitative Nondestructive Evaluation, August 1-6, 1993, Bowdoin College, Brunswick, Maine.

13. "Bolt Stress Measured by Calculating Wave Speed." Machine Design. May 24, 1990, p. 64.

14. Chen, L., and C. T. Cheng. "Study of Three Dimensional Crack Tip Location of Mortar by Acoustic Emission. " Proceedings of Nondestructive Evaluation of Concrete Elements and Structures. ASCE, 1992, pp. 25-37.

15. Choy, F. K., R. Liang, and P. Xu. "Fault Identification in Beam Structures." Journal of the Franklin Institute. Vol. 329, No. 4, 1992, pp. 697-713.

16. Chung, R., and A. Hibbs. "Corrosion Measurement with a High Resolution Scanning Magnetometer." Presented at 20th Annual Review of Progress; Quantitative Nondestructive Evaluation, August 1-6, 1993, Bowdoin College, Brunswick, Maine.

17. Clark, G. "Commentary on "QAE" Source Characterization of Microcracking in Steel." Research in Nondestructive Evaluation. A Journal of the ASNT, Vol. 4, No. 3, 1992, pp. 183-190.

18. "Computers Find Bridge Flaws." ENR. Feb. 2, 1984, pp. 14-16.

19. Connolly, M. P. "Review of Factors Influencing Defect Detection in Infrared Thermography." Journal of Nondestructive Evaluation. Vol. 10, No. 3, 1991, pp. 89-96.

20. Cook, J. F., D. A. Aldrich, B. C Anderson, and V. S. Scown. "Development of an Automated Ultrasonic Inspection System for Verifying Waste Water Container Integrity." Materials Evaluation. Vol. 42, Dec. 1984, pp. 1626-1630.

21. Copley, D. W. "Considerations in the Application of Thermoelastic Stress Analysis to the Vibration Testing of Aero-Engine Structures." Experimental 97

Techniques. Mar. 1988, pp. 11-14.

22. Daponte, P., F. Maceri, and R. S. Olivito. "Frequency-Domain Analysis of Ultrasonic Pulses for the Measure of Damage Growth in Structural Materials." Ultrasonics Symposium Proceedings. Vol. 2, Dec. 4-7, 1990, Piscataway, N.J., pp. 1113-1118.

23. Dauenheimer, E., and J. Schuring. "Diagnostic Load Tests Help Aging Bridges Cope." Proceedings of Nondestructive Evaluation of Civil Structures and Materials. University of Colorado, Boulder, Oct. 1990, pp. 245-256.

24. Davis, A. G., and J. Paquet. "Monitoring Bridge Performance Using NDE Techniques." Proceedings of Nondestructive Evaluation of Civil Structures and Materials. University of Colorado, Boulder, May 1992, pp. 227-240.

25. Devine, M. K., A. Strom, and D. C. Jiles. "Evaluation of Steel Bridges by Magnetic Hysteresis Measurements. " Presented at 20th Annual Review of Progress: Quantitative Nondestructive Evaluation, August 1-6, 1993, Bowdoin College, Brunswick, Maine.

26. Dunn, J. "Flaw Detector Sits and Waits for Cracks to Squeal." The Engineer. Jan., 1986, pp. 32.

27. Eichmann, A. R., D. C. Jiles, and M. K. Devine. "In Site Determination of the Magnetic Properties of Soft Magnetic Materials Using an Automated Magnetic Measuring System. " Journal of Magnetism and Magnetic Materials. Vol. 104, No. 107, 1992, pp. 375-376.

28. "Engineers 'Listen' to Structural Flaws in Bridges." Metal Progress News. Sept., 1986.

29. Federal Highway Administration, "Acoustic Emission Methods for Flaw Detection in Steel in Highway Bridges, Phase II. " Federal Highway Administration, Washington, D C., March 1978, Final Report.

30. Federal Highway Administration, "Acoustic Emission Methods for Flaw Detection in Steel in Highway Bridges, Phase I." Federal Highway Administration, Washington, D C., March 1975, Final Report.

31. Fokwa, D., Y. Berthaud, and D. Breysse. "Experimental Determination of the Relation Between the Damaged Zone and the Aggregate Size in Concrete Through Acoustic and Mechanical Techniques. " Engineering Mechanics. 1992, 98

pp. 131-135.

32. Freund, A. "Infrared Scanning Added to Routine Maintenance. " EC&M. Oct. 1987, pp. 67-70.

33. "Germans Detect Dam Defects with AE." Materials Evaluation. Vol. 45, Mar. 1987, pp. 325.

34. Ghorbanpoor, A. "Acoustic Emission Characterization of Fatigue Crack Growth in Highway Bridge Components. " Dissertation Abstracts International. Jan., 1987.

35. Ghorbanpoor, A. and A. T. Rentmeester. "NDE of Steel Bridges by Acoustic Emission." Proceedings Structural Engineering in Natural Hazards Mitigation. Irvine, Calif. : ASCE Structural Division, Vol. 2, 1993, pp. 1008-1013.

36. Glickstein, C. Basic Ultrasonics. New York: John F. Rider Publisher Inc., 1960.

37. Gomez-Rivas, A. "Nondestructive Evaluation of Bridge Design Assumptions. " Proceedings of Fifth International Conference on Structural Faults and Repair. June 29, 1993, Univ. of Edinburgh, pp. 53-58.

38. Gong, Z., E. O. Nyborg, and G Oonunen. "Acoustic Emission Monitoring of Steel Railroad Bridges. " Materials Evaluation. July 1992, pp. 883-887.

39. Gorkunov, E. S. "Magnetic Methods and Instruments for the Quality Control of the Case-Hardening of Ferromagnetic Steel Objects (Review)." Soviet Journal of Nondestructive Testing. Vol. 27, No. 1, 1991, pp. 1-18.

40. Green, R. E., Jr. "Telemetry Systems for Self-Monitoring of Highway Bridges." Presented at 20th Annual Review of Progress: Quantitative Nondestructive Evaluation, August 1-6, 1993, Bowdoin College, Brunswick, Maine.

41. Grobe, C. "Detection of Cracks in Reinforced Concrete Cans." Proceedings of 9th Conference on Engineering Mechanics. ASCE, pp. 413-416.

42. Groshong, B. R., G. L. Bilbro, and W. E. Snyder. "Eddy Current Transducer Model for Image Restoration." Journal of Nondestructive Evaluation. Vol. 10, No. 2, 1991, pp. 55-61.

43. Harwood, N., and W. M. Cummings. Thermoelastic Stress Analvsis. New 99

York: Adam Hilger, 1991.

44. Henkel, D. P., and J. D. Wood. "Monitoring Concrete Reinforced with Bonded Surface Plates by the Acoustic Emission Method. " NDT and E International. Vol. 24, No. 5, 1991, pp. 259-264.

45. Hognestad, H. "New Method for Determining Absolute Stress in Ferromagnetic Steel via Potential Drop Measurements. " Materials Science and Technolopv. Vol. 8, No. 7, 1992, pp. 649-651.

46. Hopwood, T., II. "Acoustic Emission Inspection of Steel Bridges. " Public Works. May 1988, pp. 66-69.

47. Hrinko, W., R. Miller, A. Aktan, B. Shahrooz, and C. Young. "Instrumentation Considerations for Nondestructive and Destructive Testing of Reinforced Concrete Slab Bridges. " Proceedings Structural Engineering in Natural Hazards Mitigation. Irvine, Calif. : ASCE Structural Division Vol. 2, 1993, pp. 1557-1562.

48. "Infrared Scanning Speeds Bridge Deck Inspection." Public Works. June 1988, pp. 77-79.

49. Jacobs, L. J., W. R. Scott, D. M Granata, and M. J. Ryan. "Experimental and Analytical Characterization of Acoustic Emission Signals. " Journal of Nondestructive Testing. Vol. 10, No. 2, 1991, pp. 63-70.

50. Jiles, D. C. "Review of Magnetic Methods for Nondestructive Evaluation. " NDT International. Vol. 23, No. 2, Apr. 1990, pp. 83-92.

51. Jiles, D. C. "The Effect of Stress on Magnetic Barkhausen Activity in Ferromagnetic Steels." IEEE Transactions of Magnetics. Vol. 25, No. 5, Sept. 1989, pp. 3455-3457.

52. Jiles, D. C. "Review of Magnetic Methods for Nondestructive Evaluation." NDT International. Vol. 21, No. 5, Oct. 1988, pp. 311-319.

53. Jiles, D. C., S. Hariharan, and M. K. Devine. "Magnescope: A Portable Magnetic Inspection System for Evaluation of Steel Structures and Components. " IEEE Transactions of Magnetics. Vol. 26, No. 5, Sept. 1990, pp. 2577-2579.

54. Jiles, D. C., and D. Utrata. "Effects of Tensile Plastic Defonnation on the Magnetic Properties of AISI 4140 Steel. " Journal of Nondestructive Evaluation. 100

Vol. 6, No. 3, 1987, pp. 129-134.

55. Jolly, W. D. "Status and Future Directions for Acoustic Emission Standards." Nondestructive Testing Standards. Present and Future. ASTM STP 1151, 1992, pp. 56-62.

56. Kaminski, D. A., D. C. Jiles, S. B. Biner, and M. J. Sablik. "Angular Dependence of the Magnetic Properties of Polycrystalline Iron Under the Action of Uniaxial Stress. " Journal of Magnetism and Magnetic Materials. Vol. 104, No. 107, 1992, pp. 382-384.

57. Kao, A. "An Overview of the Repair, Evaluation, Maintenance, and Rehabilitation (REMR) Research Program 1984-1989. " U.S. Army Construction Engineering Research Laboratories (USA-CERL), Champaign, IL.

58. Kashiwaya, K. "Fundamentals of Nondestructive Measurement of Biaxial Stress in Steel Utilizing Magnetoelastic Effect Under Low Magnetic Field." Japanese Journal of Applied Phvsics. Part 1: Regular Papers and Short Notes. Vol. 30, No. 11A, 1991, pp. 2932-2942.

59. Kayani, J. K., S. F. Russell, and K. Hoech. "Hardware/Software Designs for Spread-Spectrum Ultrasonic Evaluation Systems. " Presented at 20th Annual Review of Progress; Quantitative Nondestructive Evaluation, August 1-6, 1993, Bowdoin College, Brunswick, Maine.

60. Kayser, J. R., and E. D. Miller. "Accuracy in the Thickness Measurement of Corroded Bridge Steel Using an Ultrasonic Thickness Meter. " Proceedings of Nondestructive Evaluation of Civil Structures and Materials. University of Colorado, Boulder, May 1992, pp. 241-246.

61. Kisters, F. H., and F. W. Kearney. "Evaluation of Civil Works Metal Structures." Technical Report REMR-CS-31, January 1991.

62. Koemer, R. M., and A. E. Lord, Jr. "Spill Alert Device for Earth Dam Failure Warning." Journal of Hazardous Materials. Vol. 9, 1984, 373-380.

63. Krapez, J. C., and P. Cielo. "Thermographic Nondestructive Evaluation: Data Inversion Procedures: Parti: 1-D Analysis." Research in Nondestructive Evaluation. A Journal of the ASNT, Vol. 3, No. 2, 1991, pp. 81-100.

64. Krapez, J. C., and P. Cielo. "Thermographic Nondestructive Evaluation; Data Inversion Procedures, Part II; 2-D Analysis and Experimental Results." 101

Research in Nondestructive Evaluation. A Journal of the ASNT, Vol. 3, No. 2, 1991, pp. 101-124.

65. Kuc, R. "Measuring Structural Motion and Location Remotely with Sonar." Proceedings Nondestructive Evaluation of Civil Structures and Materials. University of Colorado, Boulder, Oct. 1990, pp. 141-154.

66. Landis, E. N., and S. P. Shah. "Applications of Quantitative NDE to Basic Fracture Research of Concrete." Presented at 20th Annual Review of Progress: Quantitative Nondestructive Evaluation, August 1-6, 1993, Bowdoin College, Brunswick, Maine.

67. Landis, E. N. and S. P. Shah. "Signal Analysis for Quantitative AE Testing." Proceedings Structural Engineering in Natural Hazards Mitigation. Irvine, Calif.; ASCE Structural Division Vol. 2, 1993, pp. 1563-1568.

68. Laura, P. A. A. "Evaluation of Structural Integrity of Cables Using Acoustic Emission." Materials Evaluation. Vol. 44, Jan. 1986, pp. 38.

69. Lawson, L. and N. Kim. "Deconvolution and Detectability in Compton Backscatter Profilometry." Presented at 20th Annual Review of Progress: Quantitative Nondestructive Evaluation, August 1-6, 1993, Bowdoin College, Brunswick, Maine.

70. Lhota, J. R , M. C. Gregory, G. C. Panos, and E. C. Johnson. "Novel Ultrasonic Backscatter Technique for Corrosion Detection. " Presented at 20th Annual Review of Progress: Quantitative Nondestructive Evaluation, August 1- 6, 1993, Bowdoin College, Brunswick, Maine.

71. Lim, M K., and C. A. Olson. "Use of Non-Destructive Impulse Radar in Evaluating Civil Engineering Structures. " Proceedings of Nondestructive Evaluation of Civil Structures and Materials. University of Colorado, Boulder, Oct. 1990, pp. 167-176.

72. Livingston, R. A. "Future Directions in NDE Research for Highways." Presented at 20th Annual Review of Progress: Quantitative Nondestructive Evaluation, August 1-6, 1993, Bowdoin College, Brunswick, Maine.

73. Loto, C. A. and E. T. Odumbo. "Electrochemical Potential Monitoring of Corrosion and Coating Protection of Mild Steel Reinforcement in Concrete. " Corrosion. Vol. 45, No. 7, 1989, pp. 553-557. 102

74. Lozen, K. M. "Detecting Invisible Corrosion in Concrete. " Concrete Construction. Aug. 1988, pp. 757-760.

75. MacDonald, D. D., M. C. H. McKubre, and M. Urquidi-MacDonald. "Theoretical Assessment of AC Impedance Spectroscopy for Detecting Corrosion of Rebar in Reinforced Concrete." Corrosion-NACE. Vol. 44, No. 1, 1988, pp. 4-6.

76. Macilwain, C. "Wave Sensor Feels Stress in Machines." The Engineer. Vol. 1, June 1989, p. 36.

77. Maji, A. K. "Determination of In-Site Stresses From Acoustic Emissions. " Proceedings of 9fli Conference on Engineering Mechanics. ASCE, 1992, pp. 405-409.

78. Maji, A. K., and J. L. Wang. "Inspection of Model Steel Bridge Components with Electronic Shearography. " Proceedings of Nondestructive Evaluation of Civil Structures and Materials. University of Colorado, Boulder, May 1992, pp. 263-278.

79. Maldague, X., J.-C. Krapez, and P. Cielo. "Temperature Recovery and Contrast Computations in NDE Thermographic Imaging Systems." Journal of Nondestructive Evaluation. Vol. 10, No. 1, 1991, pp. 19-30.

80. Manning, D. G., and T. Masliwec. "Operational Experience Using Radar and Thermography for Bridge Deck Condition Survey." Proceedings of Nondestructive Evaluation of Civil Structures and Materials. University of Colorado, Boulder, Oct. 1990, pp. 233-244.

81. Mathieson, P. A. R. "Acoustic Emissions from Fatigue Cracks in Steels." Ph.D. Dissertation, Cranfield Institute of Technology, United Kingdom, 1987.

82. Matlab Software User's Guide. The Math Works, Inc. Englewood Cliffs, New Jersey: Prentice-Hall, Inc., IWl.

83. Matthews, J. R. Acoustic Emission. New York: Gordon and Breach Science Publishers, 1983.

84. Mazurek, D. and J. DeWolf. "Experimental Study of Bridge Monitoring Technique," Journal of Structural Engineering. Vol. 116, No. 9, Sept. 1990, pp. 2532-2549 103

85. McGogney, C. H. "Magnetic Flux Leakage for Bridge Inspection." Proceedings Structural Engineering in Natural Hazards Mitigation. Irvine, Calif.: ASCE Structural Division, Vol. 2, 1993, pp. 1576-1583.

86. Mclntire, P. Nondestructive Testing Handbook. Second Edition. Columbus, Ohio: American Society for Nondestructive Testing, Vol. 5, 1987.

87. Miki, C. "Applications of NDI in Fracture Control of Honshu Shikoku Bridge Projects. " Presented at 20th Annual Review of Progress: Quantitative Nondestructive Evaluation, August 1-6, 1993, Bowdoin College, Brunswick, Maine.

88. Moslehy, F. A. "Application of Acoustic Emission to Flaw Detection in Engineering Materials." Experimental Techniques. Feb. 1990, pp. 31-33.

89. Moulder, J. C., E. Uzal, P. Sancheti, and R. Patankar. "Quantitative Characterization of Corrosion in Aluminum Lap Joints with Sweep-Frequency Eddy Current Method. " Presented at 20th Annual Review of Progress: Quantitative Nondestructive Evaluation. August 1-6, 1993, Bowdoin College, Brunswick, Maine.

90. Muravin, G. B., A. I. Merman, and L. M Lezvinskaya. "Acoustic-Emission Method for Estimating the Fracture Toughness of Concrete in Large-Scale Structures and Buildings." Soviet Journal of Nondestructive Testing. Vol. 27, No. 3, 1991, pp. 166-171.

91. National Science Foundation, "Quantitative Nondestructive Evaluation for Constructed Facilities." Announcement Fiscal Year 1992. Directorate for Engineering. Division of Mechanical and Structural Systems.

92. Newland, D. E. An Introduction to Random Vibrations. Spectral, and Wave Analysis. New York: John Wiley & Sons, 1993.

93. Newton, C. J., and J. M. Sykes. "A Galvanostatic Pulse Technique for Investigation of Steel Corrosion in Concrete." Corrosion Science. Vol. 28, No. 11, 1988, pp. 1051-1074.

94. Nishikawa, K. "An Attempt on Evaluation of Corroded Steel Bridge Members. " Presented at 20th Annual Review of Progress: Quantitative Nondestructive Evaluation, August 1-6, 1993, Bowdoin College, Brunswick, Maine.

95. Olaszek, P. "The Photogrammetric Method for Dynamic Bridge Investigations." 104

Proceedings of Fifth International Conference on Structural Faults and Repair. June 29, 1993, Univ. of Edinburgh, pp. 89-94.

96. Oliver, D. E. "Stress Pattern Analysis by Thermal Emission (Spate). " Experimental Techniques. Mar. 1988, pp. 3-7.

97. Oppenheim, A., V. and R. W. Schafer, Discrete-time Signal Processing. Engelwood Cliffs, New Jersey: Prentice-Hall Inc., 1989.

98. Ouyang, C., E. Landis, and S. P. Shah. "Damage Assessment in Concrete Using Acoustic Emission. " Proceedings of Nondestructive Evaluation of Concrete Elements and Structures. ASCE, 1992, pp. 13-24.

99. Paz, M. Structural Dvnamics: Theorv and Computation. Third Edition. New York: Van Nostrand Reinhold, 1991.

100. Popov, G M. "Methods of Checking the Mechanical Properties of Ferromagnetic Parts Based on Using the Oscillating Envelopes of the EMF Spectra." Soviet Journal of Nondestructive Testing. Vol. 27, No. 1, 1991, pp. 27-33.

101. Prabhu, D. R., M. N. Abedin, and P. H. Johnston. "Simultaneous Detection of Disbonds and Top-Layer Corrosion Through Ultrasonic Contact Scanning." Presented at 20th Annual Review of Progress: Quantitative Nondestructive Evaluation, August 1-6, 1993, Bowdoin College, Brunswick, Maine.

102. Prine, D. W. "Crack Characterization in Steel Highway Bridges Using Acoustic Emission and Strain Gage Monitoring. " Presented at 20th Annual Review of Progress: Quantitative Nondestructive Evaluation, August 1-6, 1993, Bowdoin College, Brunswick, Maine.

103. Raman, A., and B. Kuban. "Infrared Spectroscopic Analysis of Phase Transformation Processes in Rust Layers Formed on Weathering Steels in Bridge Spans." Corrosion Engineering. Vol. 44, No. 7, 1988, pp. 483-488.

104. Rao, B. P. C., M. T. Shyamsunder, P. Kalyanasundaram, D. K. Bhattacharya, and B. Raj. "Development of Eddy Current Test Techniques for Condition Monitoring of Pressure Tube/Calandria Tube Assemblies of Indian Pressurized Heavy Water Reactors." British Journal of Non-Destructive Testing. Vol. 33, No. 9, 1991, pp. 437-440.

105. Rens, K. L., T. Wipf and F. W. Klaiber. "Literature Review: Nondestructive 105

Evaluation of Civil Engineering Structures. " Report R-853 Association of American Railroads, Chicago, IL. Auguust 1993.

106. Renwick, J. "Condition of Machinery Using Computerized Vibration Signature Analysis," IEEE Transactions on Industry Applications, Vol. lla-20, No. 3, May/June 1984, pp. 519-527.

107. Rubin, S. "Ambient Vibration Survey of Offshore Platform." Journal of Engineering Mechanics Division. June 1980, pp. 425-441.

108. Russell, S. F. "Theory of Spread-Spectrum Ultrasonic Evaluation. " Presented at 20th Annual Review of Progress: Quantitative Nondestructive Evaluation, August 1-6, 1993, Bowdoin College, Brunswick, Maine.

109. Sablik, M. J. "Modeling Stress Dependence of Magnetic Properties for NDE of Steels." Nondestructive Testing Evaluation. Vol. 5, 1989, pp. 49-65.

110. Sami, Z., H. L. Chen and H. V. GangaRao. "Acoustic Emission Monitoring of Stressed Composite Bars. " Proceedings Structural Engineering in Natural Hazards Mitigation. Irvine, Calif.: publisher ASCE Structural Division Vol. 2, 1993, pp. 1569-1575.

111. Sandomirskii, S. G. "Analysis of a Method of Inspection of Moving Ferromagnetic Parts Using Coercive Force. " Soviet Journal of Nondestructive Testing. Vol. 27, No. 6, 1992, pp. 392-398.

112. Sarihan, V., D. Oliver, and S. Russell. "Thermoelastic Stress Analysis of an Automobile Engine Connecting Rod." Experimental Techniques. Mar. 1988, pp. 7-10.

113. Scott, I. G. Basic Acoustic Emission. Montreux, Switzerland: Gordon and Breach Science Publishers, 1991.

114. Shablov, S. V., V. I. Kapustin, and V. B. Nazaruk. "Electrographic Inspection of the Root Part of Welded Joints in Steam Turbines in the Hot State. " Soviet Journal of Nondestructive Testing. Vol. 27, No. 6, 1992, pp. 413-419.

115. Silk, M. G. "Flaw Size Distributions in Pressure Vessels and Flaw Detection Probabilities in NDT. " British Journal of Non-Destructive Testing. Vol. 33, No. 10, 1991, pp. 491-494.

116. Singh, A., and R. McClintock. "Computer-Controlled System for Nondestructive 106

Thickness Measurement of Corroded Steel Structures. " Journal of Energy Resources Technology. Vol. 105, Dec. 1983, pp. 499-502.

117. Singh, G. P., Ed. "Inspection of Corroded Material." Materials Evaluation. Vol. 43, Aug. 1985, pp. 1054-1055.

118. Stanley, P. "Appraisal of a New Infrared-Based Stress Analysis Technique." Optical Engineering. Vol. 26, No. 1, Jan. 1987, pp. 75-80.

119. Tarricone, P. "Bridges Under Surveillance." Civil Engineering. May 1990, pp. 48-51.

120. Teller, C. M., and H. Kwun. "New Methods for Inspecting and Monitoring Bridge Structures Using Magnetostrictive Sensors. " Presented at 20th Annual Review of Progress: Quantitative Nondestructive Evaluation, August 1-6, 1993, Bowdoin College, Brunswick, Maine.

121. Thomas, G. H., S. Benson, P. Durbin, N. Del Grande, and D. Schneberg. "Nondestructive Evaluation Techniques for Enhanced Bridge Inspection. " Presented at 20th Amiual Review of Progress: Quantitative Nondestructive Evaluation, August 1-6, 1993, Bowdoin College, Brunswick, Maine.

122. Thompson, D. O. and D. E. Chimenti, Proceedings. "Review of Progress in Quantitative NDE," Plenum, New York. Vol. 13.

123. Thompson, J. G. "Second Layer Corrosion Detection in Aircraft Structures." Presented at 20th Annual Review of Progress: Quantitative Nondestructive Evaluation. August 1-6, 1993, Bowdoin College, Brunswick, Maine.

124. U.S. Department of Transportation, "Nondestructive Testing Methods for Steel Bridges." Participants Training Manual, Federal Highway Administration, U.S. DOT, June 1986.

125. Utsunomiya, T., H. Nishizawa, and K. Kaneta. "Biaxial Stress Measurement Using a Magnetic Probe Based on the Law of Approach to Saturation Magnetization." NDT and E International. Vol. 24, No. 2, 1991, pp. 91-94.

126. Valkering, C. P., and D. J. Jongeneel. "Acoustic Emission for Evaluating the Relative Performance of Asphalt Mixes Under Thermal Loading Conditions. " Acoustic Emission. 1991, pp. 160-187.

127. Vavilov, V. "A Review of Thermal Nondestructive Testing Methods for 107

Aerospace Structures in the Former USSR." Journal of Recherche Aerospatiale (English Edition). No. 6, 1991, pp. 1-16.

128. Watanabe, Y., and T. Shoji. "The Evaluation of In-Service Materials Degradation of Low-Alloy Steels by the Electrochemical Method. " Metallurgical Transactions à, Sept. 1991, pp. 2097-2106.

129. Weng, C. C., M. T. Tarn, and G. C. Lin. "Acoustic Emission Characteristics of Mortar Under Compression. " Pavement and Concrete Research. Vol. 22, 1992, pp. 641-652.

130. Winfree, W. P., K. E. Cramer, P. R. Emeric, P. A. Howell, and H. L Syed. "Thermographic Detection of Corrosion in Aircraft Structures. " Presented at 20th Annual Review of Progress: Quantitative Nondestructive Evaluation, August 1-6, 1993, Bowdoin College, Brunswick, Maine.

131. Woodard, M. J., and S. S. Kuo. "Evaluating Polymer Concrete Bridge Expansion Joints Using Acoustic Emission. " Proceedings of 9th Conference on Engineering Mechanics. ASCE, pp. 409-412.

132. Woodham, D "Magnetically Induced Velocity Change to Measure Residual Stress." Proceedings of Nondestructive Evaluation of Civil Structures and Materials, University of Colorado, Boulder, May 1992, pp. 297-312.

133. "X-Rays and NDT-An Updated Dry Processing/Silver Imaging X-Ray System is an Effective Non-Destructive Test System. " Aerospace Engineering. July 1992, pp. 20-21.

134. Yamauchi, S., M. Hirai, M. Kusaka, M Iwami, H. Nakamura, S. Minomura, H. Watabe, M Kawi, and H. Zoezima. "Nondestructive Depth Profiling Using Soft X-Ray Emission Spectroscopy by Incident Angle Variation Method. " Japanese Journal of Applied Phvsics. Part 1 : Regular Papers and Short Notes. Vol. 31, No. 2A, 1992, pp. 395-400.

135. Yen, B. T., S. P. Pessiki, and J. W. Fisher. "Nondestructive Evaluation of Bridge Fatigue Crack Growth by Incorporating Acoustic Emission Monitoring." Presented at 20th Annual Review of Progress: Quantitative Nondestructive Evaluation, August 1-6, 1993, Bowdoin College, Brunswick, Maine. 108

136. Zachar, J., and T. R. Naik. "Principles of Infrared Thermography and Application for Assessment of the Deterioration of the Bridge Deck at the 'Zoo Interchange'." Materials: Performance and Prevention of Deficiencies and Failures 92. New York, ASCE, Aug. 1992, pp. 107-115. 109

ACKNOWLEDGEMENTS First, all thanks goes to God, creator of all things. I would like to thank my major professor, Lowell Greimann, who is my mentor and friend. One paragraph is too short to name all of places we have been and discussions we have had. If granted the chance to be a faculty member some day, I will try to model my day and events as we have for the past six and a half years. I'm veiy thankful for how I have been treated throughout my graduate studies - one would have to look far to fmd a student who has been treated better. I would like to give special thanks to Jim Stecker, who like Lowell, knows the defmition of hard work and knows how to get things accomplished. We have been through many long discussions which eventually always lead to the solution of a problem. I will never forget the field trips - especially the "egged out" gudgeon pin at Lock 19 and the many anchorage measurement devices that we developed, constructed, and used. Our four tool boxes contain many specialized tools that work only on a particular lock and dam in the United States. I would like to mention thanks to our project manager, Dr. Anthony M. Kao. I'm glad I was a part of your research team. As you end your career, you can remember being involved with the "ACME" lock and dam inspection team. Thank you, Tony, for funding my salary during my graduate career. Thanks is given to graduate students Mr. Joel Veenstra and Mr. Mike Nop. I enjoyed working with you both and wish both of you the best of luck as you go off on your own careers. Joel, I will never forget our Nashville golfing trip and the Mazda 929 - what a machine. I can not begin to mention all of the Corps people who influenced my thinking in both my Masters degree and Ph D degree. Just to name a few: Lynn Midgett, Tom Hood, D. Wayne Hickman, Mike Kn*ebutg, Jim Fischer, Dick Atkinson, Fred Jores, Jerry Dean, Harold Trahan, Charlie Bryan, Gene Ardine, Jack Syrac, and many more. Thanks to my committee members Terry Wipf, Fouad Fanous, Loren Zachary, Tom Maze, and Steve Russell. Special thanks to Steve for supplying all of the 110 equipment, transducers, students, and DSSSUE knowledge. It took over four years to find a person to go to bat for me in the nondestructive testing field of study. Without you, Steve, I could not have finished my Ph.D this year. Also thanks to Wayne Klaiber who hired me to work on the NDE literature review project for the American Association of Railroads of which John Charos was project monitor. Thanks to the DSSSUE graduate students MAK Afzal, Sangmin Bae, J. K. Kayani, and Karl Hoech. All of you spent time in helping teach me a new way of thinking. Thanks to my family and friends for all of your support. This thesis is dedicated in memory of my Grandfather, Bert J. Rens, who passed away February 22, 1994 at the age of 90 years, 8 months, and 1 day. Finally thanks to my wife. Amy, who doubles as my best friend. The last 5 months have not been easy with your studying and taking your professional engineering exam and my working ridiculous hours trying to finish up this thesis. Thank you for all of your help and support. I hope we have many years together to make up for the lost time. Referring back to my Masters degree page ISO - obviously she said yes to the proposal. Ill

APPENDKÂ. CASE STUDIES An example of deteriorating infrastructure occurs in our nation's waterways. Approximately one-half of the Corps' 269 lock chambers along inland waterways will reach or exceed their 50-year design life by the turn of the century [84]. Lock gates are unique and complicated structures consisting of several primary and secondary structural members. The anchorage system is a primary component which connects the gate to the concrete wall. Failure of the anchorage system can be catastrophic and without warning. For example, in the late 1980's, an anchorage system at Lock and Dam 14 on the Mississippi River failed and caused unscheduled maintenance to get the lock chamber operational. A similar failure has occurred on the Illinois Waterway at LaGrange Lock and Dam. Many times the embedded anchorage is badly corroded and a sudden failure could be developing. Current inspection techniques consist of removing the surrounding concrete and inspecting the previously hidden steel. This appendix gives a pictorial illustration of the problem, failures, and current inspection techniques. In other words, this appendix presents the purpose for this study.

Problem Figures A.l, A.2, and A.3 illustrate the problem that occurs at the embedded steel connection on a miter lock gate structure. Fig. A.l Interharbor Lock and Dam , New Orleans, Louisiana is a lock where obvious concrete damage has occurred. Although the lock is currently operational, the condition of the concrete and exposed steel illustrate both the field conditions and deterioration state of many of the lock and dam structures. It is probably safe to assume that the embedded steel in this area has deteriorated also. Fig. A.l is an example where one would expect a structural problem - many times the problem is not evident by the deterioration of the surrounding concrete. Fig A.2a, Murray Lock and Dam, Little Rock, Arkansas shows a typical anchor bar that is in relatively good shape. When one removes the small concrete piece at the interface between the concrete and steel it becomes evident that corrosion of the embedded steel is occurring. This is illustrated in Fig. A.2b with a flat piece of steel 112

and pocket knife. It should be noted that some times the concrete appears new and can mislead one to believe that the steel is in good condition. Fig. A.3, Bayou Sorell Lock and Dam illustrates another example where by the steel appears good however just beyond the interface a problem is developing. In this situation, lock personal had not looked closely at the anchor for years. When brought to their attention, this problem was corrected inunediately. Finally, Fig. A.4, Cheatham Lock and Dam shows a temporary solution to a problem that occurred in the Nashville District when excessive movement was visually detected in the anchorage system. A temporary anchorage reinforcement was fitted over the anchorage while inspectors had to determine whether the movement was deterioration related. Nashville District personal suspect the anchorage was constructed based on a flexible anchorage design whereas plans called for the more traditional rigid system. Recently, the author of this thesis was visiting several miter lock gate structures in the New Orleans District of the U. S. Army Corps of Engineers. On all two of the structures visited, evidence of deteriorated steel was present. In addition, the bond between the concrete and steel was not present.

Current Inspection Techniques Many times the current inspection techniques are rather primitive but, nonetheless, effective. Fig. A.5, Lock and Dam 24 on the Mississippi River, illustrates what the inspectors feel may be a problem. Destructive testing is performed where concrete is cut and removed to examine the condition of the steel. If the steel is deteriorated, reinforcement plates will be added. After the embedded steel is inspected, concrete is placed back in the anchor recess. A significant amount of labor and lock down time is required for this routine inspection technique performed by the St. Louis District of the USA-COE. Fig. A.6 shows current Repair Evaluation Maintenance Rehabilitation (REMR) inspection techniques. Inspectors are require to place dial gauges on the interface between the concrete and steel. Any movement at this location over an acceptable amount is considered a potential problem. Although this technique is i 113

rather elementary and the fact that most failures are immediate with little warning, the program does force Corps personnel to inspect the anchorage more closely. Continual monitoring of the interface movement could flag to the inspector that a potential problem is developing. In fact, the deteriorated anchorage on Bayou Sorell lock (Fig. A.3) was discovered during a visit by ISU experts during REMR project development for sector gates.

Failures Fig. A.7, Lock and Dam 14 on the Mississippi River, and Fig. A.8, LaGrange Lock and Dam on the Illinois Waterway, show failures that do occur when the anchorage system deteriorates. Both of these failures were catastrophic and immediate. It is possible that ongoing inspections may have alleviated these failures. Although not shown another failed anchorage has occurred at Emsworth Lock and Dam in the Pittsburgh District. Nondestructive techniques could have noted the presence of a crack forming. In addition, there are several failures that have been reported but not documented with photos in this appendix. The St. Paul District and the Tulsa District have reported problems with deteriorated anchorages. a) Overall View of Anchorage Area

b) Deterioration at Concrete Interface Figure A. 1 Interharbor Lock and Dam 115

a) Embedded Anchorage System (Channel Type)

b) Steel Deterioration at Concrete Interface Figure A.2 Murray Lock and Dam 116

a) Embedded Anchorage System (Round Type)

b) Steel Deterioration at Concrete Interface Figure A.3 Bayou Sorell Lock and Dam 117

a) Temporary Bracing at Suspected Anchorage System Failure

b) Close-up of Temporary Anchorage Reinforcement Figure A.4 Cheatham Lock and Dam 118

a) Cut Away 6" of Concrete at Anchorage Interface

'\''L b) Removal of Concrete at Anchorage Interface Figure A.5 Lock and Dam 24 on Mississippi River 119

a) Valve Anchorage

b) Miter Gate Anchorage Figure A. 6 Dial Gages to Measure Movement at Concrete Interfaces 120

Figure A.7 Anchorage Bar Failure at Lock and Dam 14 on Mississippi River 121

Figure A.8 Anchorage Failure at LaGrange Lock and Dam 122

APPENDIX B. FOURIER THEORY Fourier theory, in some way, shape, or form, has found its way into ahnost every branch of physical science. Vibrations theory is no exception. The Fourier series and integral were discovered by J. B. Fourier in the early 1800's. Often the integral in the continuous form is not available. Therefore, the discrete form is utilized. This appendix shows the general theory for the discrete transform, the exponential transform, and the relationship between them. The last section shows the development of the Fourier Integral, which is used throughout the body of this thesis.

Exponential Form of the Fourier Series The exponential form of the Fourier series is given by [97] [99]

rt=+oo F(t) = S (B.l) where the complex expansion coefficients are

1 ^ C„==dt (B.2) and where F(t) is a periodic function with Fourier period T, n is the harmonic number index, i the imaginary number. The limits can be arbitrary; that is, the lower limit can be t and the upper limit t + r. The frequency of the forcing function is given by

(Ù = ~ (B.3)

We can define the complex expansion coefficients as

C„ = - ib„ (B.4)

Using the property of the exponential (euler's identity) 123

e'" = cos(a) + isin(a) (B.5)

the positive terms of Eq. (B.2) using Eq. (B.4) can be written as

= jF(t)[œs(-n(ùt) + ;sin(-/iwf)Xf (B.6)

where, by comparison to Eq. (B.4),

= —jF(t)cos(-nûtt) dt = —jF(t)cos(n<ùt)dt (B.7) ^0 ^0 and

= - —jF(t)sin(-n(ùt) dt = —jF(t)sin(n(ot)dt (B.8) ' 0 ' 0

The negative terms of Eq. (B.2) can be written as

1 ^ C.„ = j;fF(t)e^^'dt (B.9) ' 0

T C.„ = jf(f)[cos(mwr) + ism(inûtt)]dt (B.IO)

where, by comparison with Eq. (B.3), 124

1 ^ = — jf(f)cos(MWf) dt (B.ll) and

1 ^ f>.„ =-—fF(t)sm(n(ùt) dt (B.12)

Therefore,

a.„ = (B.13)

i>.„ = -K (B.14)

or

C„ . C*

Discrete Fourier Series When the periodic forcing function F(t) is only supplied at N time intervals, the discrete form of the Fourier Series is appropriate and the integrals in Eq. (B. 1) can be replaced with sunmiations [97]. Defme

A/ = - /, = jàt (B.16) N '

The discrete form is then 125

N - 1 F(tj) = Ss (C„ e2mWA0 n = 0

n = 0,l,2,3,...,(Ar-l)

where

C = — /-(/ V-Z^nW/AO " N j-o ;=0,1,2,3,...,(^-1)

The n"" term of Eq. (B. 18) can be written as

The N-n term as

or

Now e-2mj ^ 2

So that ... 1 126

therefore

(B.24)

In summary, the relationship between the discrete and exponential frequency €„ terms are as follows

Q.. = c * = c „ (B.25)

The N/2 folding of the discrete is similar to the n=0 folding in the exponential case. This folding frequency phenomena is termed Nyquist frequency.

Fourier Transform In presenting the theory for vibrations and DSSSUE, the Fourier Transform is needed. Substituting Eq. (B.2) int Eq. (B.l)

(B.26)

where the limits of the integrand have been changed to - T/2 to + T/2 and the integrated variable in Eq. (B.2) was changed from t to T. The frequency nor in Eq. (B.26) consists of discrete equally spaced frequency values separated by the interval Aw

(B.28)

Substituting 127

T = ^ (B.29) Ab) into Eq. (B.26) and as T becomes large, the sunmiations become integrals and the discrete intervals become infmitesimal

2 —./ (B.30) Aw -*-* d<ù

Therefore

+ 00 400 ^(0 = ^ f [ /f(T)e-'"WT (B.31)

which is called the Fourier transform of F(t) where the term in brackets including the constant 2T is given by

C((ù) = f F(t)e"^'dt (B.32) and

+ 00 F(t) = fC((ù)e"^'dùi (B.33)

which are called a Fourier Transform pair. 128

Fast Fourier Transform Analysis of data utilizing Fourier theory was discouraged because of the numerous mathematical operations. In the late 1940's, recognition of the symmetries and periodicities made the DT more efficient. In 1965, an algorithm known as the Fast Fourier Transform (FFT) was developed which revolutionized the analysis of time series data [99]. In reality all transforms utilized in software today utilize FFT techniques which is a very efficient form of the discrete case mentioned in the previous section. 129

APPENDK C. PULSE ECHO SURFACE WAVES Fig. C.l and Fig. C.2 show attempts to investigate the inspection of the steel embedded in concrete. Fig C.l shows a small steel piece embedded in a small concrete cylinder. This experiment showed that, when a perfect bond exists between the steel and concrete, ultrasonic surface waves are unable to propagate beyond the concrete interface. In reality, in the field, a perfect bond is rarely achieved. In fact, most of the time there is no bond. Fig. C.2 illustrates a box of concrete with slots for steel specimens. This more closely represents the condition in the field. Bars with flaws were inserted into the box and ultrasonic surface waves were utilized. Results indicated that surface waves in this case were able to penetrate the concrete and determine the location of the flaw. Based on these initial results, it was determined that ultrasonic tests could be performed without concrete surrounding the steel. 130

a) Surface Wave Equipment and Bonded Concrete Sample

b) Surface Wave Testing of Bonded Concrete Figure C.l Bonded Concrete Testing with Surface Waves 131

a) Concrete With Slots

b) Surface Wave Testing of Unbonded Concrete Figure C.2 Unbonded Concrete Testing with Surface Waves