|USFDC Home | USF Electronic Theses and Dissertations||| RSS|
This item is only available as the following downloads:
Practical Vibration Evaluation and Early Warning of Damage in Post-Tensioned Tendons by Jaime Lopez-Sabando A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Mechanical Engineering College of Engineering University of South Florida Major Professor: Alberto A. Sag s, Ph.D. Daniel P. Hess, Ph.D. Autar K. Kaw, Ph.D. Date of Approval: December 14, 2007 Keywords: Tendon anchorage, Non destructive methods, Corrosion monitoring, Probes, Electrochemical Impedance Spec troscopy, Linear Polarization. Copyright 2008, Jaime Lopez-Sabando
ii Dedication To my wife, Mayme Hayes, for her unconditional support.
iii Acknowledgements My sincere thanks to Dr. Alberto A. Sags for his guidance and encouragement. I am also grateful to Dr. Autar Kaw and Dr Daniel Hess for serving on the committee. I would also like to thank Kingsley Lau and Dr. Luciano Taveira for their valuable help, and the personnel of the State Structures Laboratory of Florida Department of Transportation for us e of the Tendon Test Facility. This research was supported by the Florida Department of Transportation. The opinions, findings and conclusions ex pressed here are those of the author and not necessarily those of the supporting agency.
i Table of Contents List of Tables iii List of Figures iv Abstract viii Chapter 1 Introduction 1 Chapter 2 Practical Vibration Evaluation Methodol ogy 7 2.1 Tension Spreadsheet 8 2.2 Mass per Unit Length (mu) 11 2.3 Stiffness (S) 12 2.4 Tendon Frequencies 14 2.4.1 Build-in Sound Card Method 15 2.4.2 Data Acquisition Board Method 22 2.4.3 Digital Recorder Microphone Method 23 Chapter 3 Practical Vibration Evaluation Validation 25 Chapter 4 Early Warning Corrosion Probes Methodology 30 4.1 Test Environments 30 4.2 Electrochemical Probes 31 4.3 Electrical Resistance Probes 32 4.4 Probe Material Characterization 38 4.5 Gravimetric Measurements 38 Chapter 5 Early Warning Corrosion Probes Results and Discussion 40 5.1 Electrochemical Probes 40 5.2 Electrical Resistance Probes 52 5.3 Gravimetric Measurement s 54 5.4 Early Warning Probes Discussion 55 Chapter 7 Conclusions 57 References 59 Appendices 61 Appendix 1: Level 1 Block Diagram of Analyzer-M 62
ii Appendix 2: Level 2 Block Diagram of Analyzer-M 63 Appendix 3: Level 3 Subdiagram 0 of Analyzer-M 64 Appendix 4: Level 3 Subdiagram 1 of Analyzer-M 65 Appendix 5: WavePlayer-Mono Block Diagram 67 Appendix 6: Analyzer-M instructions 68 Appendix 7: ANALYZER-DAB Level 3 Block Diagram 82 Appendix 8: ANALYZER-DAB Installation Guide 85 Appendix 9: Block Diagram of WavePlayer-DAB 87 Appendix 10: Wave Player Micro Block Diagram 88 Appendix 11: Derivation of Equation to Obtain P 89 Appendix 12: Derivation of Equation to Obtain Rcorr 90 Appendix 13: Block Diagram P-Measurements 91
iii List of Tables Table 1. Tendon (0A, 0B) Input Paramete rs 28 Table 2. Tendon (0A, 0B) Results 29 Table 3. Eight-Character Tendon Segment Desi gnation 79
iv List of Figures Figure 1. Failed Tendon at Niles Channel Bridge 2 Figure 2. Failed Tendon at Sunshine Skyway Bridge 2 Figure 3. Typical Tendon Configuration 4 Figure 4. Details of a Typical Anchorage Sy stem (Dywidag International) 5 Figure 5. Input Worksheet 9 Figure 6. Results Worksheet 10 Figure 7. Output Field Graphic 10 Figure 8. Strands Configuration, T endon 13A Sloping Section 14 Figure 9. Strands Configuration; T endon 13A, Horizontal Section 14 Figure 10. Level 1 Front Panel 16 Figure 11. Level 2 Front Panel 16 Figure 12. Level 3, Continuous Signal Display 18 Figure 13. Level 3, File Name Prompt 19 Figure 14. Level 3, Ready to Record Prompt 19 Figure 15. Level 3, Test in Progress Prompt 20 Figure 16. Time and Frequency Domain Graphs 21 Figure 17. WavePlayer-Mono Front Panel 22 Figure 18. Front Panel Analyzer-DAB 23 Figure 19. Wave Player Micro Front Panel 24
v Figure 20. Tendon Test Facility 26 Figure 21. Tendon Test Set Up 27 Figure 22. Electrochemical Probe Before Grout ing 32 Figure 23. Grouted Electrochemical Probe 32 Figure 24. ER Probe Interior Design 33 Figure 25. Schematic of ER Probe 36 Figure 26. Front Panel of The P-Measurements Program 37 Figure 27. Probe Materials Characterizati on 39 Figure 28. Schematic of Electrochemical Probe Cross-Section 40 Figure 29. Electrochemical Probe Equi valent Circuit 40 Figure 30. Simplified Equivalent Ci rcuit for the Electrochemical Probe 41 Figure 31. EIS Behavior of (a) 1mm Gap and (b) 0.6mm Gap Probes 43 Figure 32. EIS Interpretation 44 Figure 33. RS and RP Trends for Duplicate (No.3 and 4) 1mm Gap 46 Figure 34. RS and RP Trends for Duplicate (No.3 and 4) 0.6mm Gap 46 Figure 35. I Vcomp Curve of 1mm Gap Electrochem ical Probe 48 Figure 36. RP Values Estimated by LPR and EIS Methods for Duplicate 49 Figure 37. RP Values Estimated by LPR and EIS Methods fo r Duplicate 49 Figure 38. Icorr Trends of Electrochemical Probes with 1.0mm Gap 51 Figure 39. Icorr Trends of Electrochemical Probes with 0.6mm Gap 51 Figure 40. ER Probes Instantaneous Corrosion Rate 53 Figure 41. ER Probes Cumulative Corrosion Rate 53 Figure 42. Corrosion Rate of Grouted and Bare Steel Strands Exposed 54
vi Figure 43. Analyzer-M Level 1 Block Diagram A 62 Figure 44. Analyzer-M Level 1 Block Diagram B 62 Figure 45. Analyzer-M Level 2 Block Diagram A 63 Figure 46. Analyzer-M Level 2 Block Diagram B 63 Figure 47. Analyzer-M Level 3 Subdiagram 0 Block Di agram A 64 Figure 48. Analyzer-M Level 3 Subdiagram 0 Block Di agram B 64 Figure 49. Analyzer-M Level 3 Subdiagram 1 Block Di agram A 65 Figure 50. Analyzer-M Level 3 Subdiagram 1 Block Di agram B 65 Figure 51. Analyzer-M Level 3 Subdiagram 1 Block Di agram C 66 Figure 52. Wave Player-Mono Block Diagram A 67 Figure 53. Wave Player-Mono Block Diagram B 67 Figure 54. Accelerometer Orientation 74 Figure 55. Accelerometer Position 81 Figure 56. ANALYZER-DAB Level 3 Subdiagr am 0-A 82 Figure 57. ANALYZER-DAB Level 3 Subdiagr am 0-B 82 Figure 58. ANALYZER-DAB Level 3 Subdiagr am 0-C 83 Figure 59. ANALYZER-DAB Level 3 Subdiagr am 1-A 83 Figure 60. ANALYZER-DAB Level 3 Subdiagr am 1-B 84 Figure 61. ANALYZER-DAB Level 3 Subdiagr am 1-C 84 Figure 62. WavePlayer-DAB Block Diagram A 87 Figure 63. WavePlayer-DAB Block Diagram B 87 Figure 64. Wave-Player Micro Block Diagram A 88 Figure 65. Wave-Player Micro Block Diagram B 88
vii Figure 66 Schematic ER Probe Initial Conditions 89 Figure 67. Block Diagram P-Measurements A 91 Figure 68. Block Diagram P-Measurements B 92
viii Practical Vibration Evaluation and Early Warning of Damage in Post-Tensioned Tendons Jaime Lopez-Sabando ABSTRACT Severe corrosion damage and even complete failure was recently discovered in external post-tensioned (PT) tendons of three Floridas pre-cast, segmental bridges over seawater. A key de terioration factor was the formation of large bleed water grout voids at or near the anchorages. Steel corrosion may occur at the grout-void interface or in t he air space of the vo id itself. Since the tendons are critical to the structural integrity of the br idges, reliable and nonintrusive damage detection methods are desirable to manage or prevent future occurrences. In recent years several indirect non-destructive methods have been developed or improved to ev aluate the conditions of the tendons. One of those methods is vibration-based tension measurements, consisting of detecting tendon tension loss by analyzing the tendons natural frequencies. Until recently, vibration-bas ed tension measurements were costly and laborious since they required severa l operators to conduct the tests and complicated analysis through different progr ams. The first objective of this research is to provide a practical, si mplified, user-friendly testing and analysis method for screening tendons by vibration measurements.
ix Electrochemical Impedance Spectroscopy, Linear Polarization, and Electrical Resistance are alternative met hods that could nondestructively detect or monitor corrosion before strand failures occur. The reliability and sensitivity of these conventional monitoring methods in solid or liquid media are well proven. However, few investigations exist on applying these methods to air-space corrosion as it may occur in tendon anchors The second objective of this research is to establish the feasibility of using the above conventional monitoring methods for detecting air-s pace corrosion. In this investigation, two different types of Electrical Resistance probes were designed and evaluated. Also, elec trochemical probes were constructed simulating strands conditions in the grout-void interface. Electrochemical Impedance Spectroscopy and Linear Po larization measurements were conducted in the electrochemical pro bes to calculate their instantaneous corrosion rates. Electrical Resistanc e and Electrochemical probes results indicate that both methods provide suffici ent sensibility to determine the ongoing damage.
1 Chapter 1 Introduction Severe corrosion damage and even complete failure was recently discovered in external post-tensioned (PT) tendons of three Floridas pre-cast, PT segmental bridges over seawater. The damage consisted, in each of the three bridges, of a completely separat ed tendon plus one to several partially detensioned tendons (Figure 1,2). A key deter ioration factor was the formation of large bleed water grout voids at or near the anchorages. Steel corrosion may occur at the grout-void interface or in the air space of the void itself. Atmosphericlike air space corrosion may be induced on the bare steel by the high humidity environment inside the grout voids. Since t he tendons are critical to the structural integrity of the bridges, reliable and non-intrusive damage detection methods are desirable to manage or prevent future occurrences , . The following main characteristics of post-tensioned tendons are noted in the Federal Highway Administration Po st-Tensioning Tendon Installation and Grouting Manual . A completely assemble d, post-tensioning tendon consists of anchorages, prestressing strands, duct and cementitious grout. The anchorages are embedded in the concrete pier diaphragm. In many applications including the bridges that experienced corrosion as noted above the duct segments of the tendon are external to the concrete, allo wing them to freely vibrate between the
2 end-span diaphragm and deviation block or between deviation blocks (Figure 3). External segments typically range from 5 to 20 m in length, and their fixity approximates clamped end conditions. Figure 1. Failed Tendon at Niles Channel Bridge  Figure 2. Failed Tendon at Sunshine Skyway Bridge 
3 The grout provides corrosion prot ection to the strand and bonds the internal tendon to the concrete structure surrounding the duct. The primary constituent of grout is ordi nary Portland cement (Type I or II). Other cementitious material may be added such as fly ash to improve corrosion resistance in aggressive environments or a high range water-reducer (HRWR) to enhance fluidity There are several commercial grouts approved by the Florida Department of Transportation (FDOT). The re lative humidity ( RH) inside the void depends on the type of grout being used, and can range from ~75% to ~90% . The duct containing the grout and strands is made of high density polyethylene. The duct size depends on t he number of strands inside the tendon The nominal internal cross sectional area of circular duct should be at least 2.25 times the net area of the post-tensioni ng strands. For example, a 11.45cm outer diameter duct with a wall thickness of 0.475cm can encase up to 38 strands of 98.7mm nominal cross secti on area, or a 9cm outer diameter duct with a wall thickness of 0.432cm can encase up to 23 strands of 98.7mm nominal cross section area). A typical tendon contains from 1 to many strands (e.g. 19,27,etc) made from 7 individual high tensile strength steel wires (A 416), arranged as 6 helically wound outer wires and one ce nter king wire. ASTM A416 is a special steel alloy that it has been heat treaded to obtain a guaranteed ultimate tensile strength (GUTS) of 1860Mpa (270ksi) . St rands used in PT bridges in Florida are mainly of two nominal sizes, 12.7 mm (0.5in) and 15.24mm (0.6in) diameter, with nominal cross sectional areas of 98.7mm2 and 140mm2 (0.153 and 0.217
4 square inches), respectively. After wedge set and relaxation, the terminal stress in post-tensioned strands is on the order of 70% of their GUTS . Therefore, a strand would likely fail if corrosion decreased its cross section area decreased by more than 30%, or even earlier because of stress concentration effects as corrosion is rarely uniform A typical anchorage assembly consists of a wedge plate, anchor, trumpet, and wedges (Figure 4). The wedge plate carries all the strands and bears on the steel anchor. The anchor is typically made of ductile iron (ASTM A27) bearing directly against the concrete. Plastic or galvanized sheet metal trumpets are used to transition from the anchor to the duct. Wedges are of case-hardened, low carbon or alloy steel, and their length is at least 2.5 times the strand diameter . Deviation Blocks Expansion Joint CL Pier Deviation Blocks Expansion Joint CL Pier Figure 3. Typical Ten don Configuration 
5 Figure 4. Details of a Typical Anchor age System (Dywidag International) Direct detection of strand corrosion in the external section is difficult without damaging the tendon since the st rands are encased within polyethylene ducts filled with hardened grout. Observat ion of corrosion in the anchorages is even more difficult unless a grout voids is present, in which case a boroscope may be introduced through a vent hole or unused wedge hole. In recent years several indirect non-destructive met hods have been developed or improved to evaluate the conditions of the tendons such as magnetic flux leakage or pulsed eddy current . One of those me thods is vibration-based tension measurements, consisting of detecting tendon tension loss by analyzing the natural frequencies of the vibrat ing external tendon length. Until recently, vibration-based tens ion measurements required several operators to conduct the tests and complicated analysis through different programs. Frequent implement ation can be costly since Florida has more than
6 80 major, post-tensioned bridges, which would require a co mmensurately large need of specialists work hours and funding. The first objective of this research is to provide a practical, simplified, us er-friendly testing and analysis method for screening tendons by vibrat ion measurements. The research addresses different options of acquiring the tendon frequencies, required the employment of only one operator. Although the vibration te chnique can be easy to im plement, a drawback is that it would only detect a damaged tendon after at least one of its strands has snapped, since tension loss may result only if a strand has failed and the grout cannot support the resulting transferr ed load. Electrochemical Impedance Spectroscopy (EIS), Linear Polarization (LP), and Electrical Resistance (ER) are alternative methods that could nondestruc tively detect or monitor corrosion before strand failures occur. The reliabili ty and sensitivity of these conventional monitoring methods in solid or liquid media are well proven. However, few investigations exist on app lying these methods to air-space corrosion as it may occur in tendon anchors. The second objective of this research is to establish the feasibility of using the above conv entional monitoring methods for detecting air-space corrosion. If these methods pr ove to be sensitive enough, then they could be used as a supplement to vibratio n testing for early warning of tendon deterioration.
7 Chapter 2 Practical Vibration Evaluation Methodology* Vibration-Based tension measur ements consist of measuring the vibrational response of tendons to mec hanical excitation, and using the results along other tendon parameters to esti mate the tendon tension. A damaged tendon can be detected by comparing its ac tual estimated tension against prior tension measurements, peer tendons, or by comparing segments tension at each end of the tendon. Conditions for a dam aged tendon to be detected are that at least one of its strands has snapped, sinc e tension loss may result only if a strand has failed, and that the grout cannot support the resulti ng transferred load. As derived by Morse [2, 11] the vibration frequency (fn) of modes n = 1,2, of a stiff bean of length L, mass per unit length mu, flexural stiffness S, tensioned by force T, and clamp ed at both ends are given by: () ()] ) 4 2 ( 2 1 [2 2 2 1L T S n L T S mu T L n fn + + + =  If S, L, mu, and fn are known then T can be found by solving the above equation for T. An independent estimate of T is obtained for each fn. Parts of the work in this chapter have appeared in A. S ags, T. Eason, C. Cotrim and J. Lopez-Sabando, Validation and Practical Procedure for Vibrational Evaluation of Tendons, Project No. BC 353#44, 158 pages, Draft Final Report to Florida Department of Transportati on, University of South Florida, Tampa, Fl, December, 2007 .
8 2 2 2 1 2 2 4 2 4 2 24 6 2 2 1 L mu mu S mu n S f L S mu S n mu S n S mu S mu L f n mu S n f L Tn n n + =  2.1 Tension Spreadsheet As it was explained in the introduction, the first objective of this research is to provide a practical, simplified, us er-friendly testing and analysis method for screening tendons by vibration measuremen ts with the employment of only one operator. A spreadsheet called Tension-Spreadsheet was prepared in Excel to estimate the tension per strand in a set of 6 tendons, three on each side of a symmetric bridge span, each tendon having ex ternal segments terminating at the diaphragm at each end of the span. The Tension-Spreadsheet consisted of four worksheets called: Inputs Calculations Results and Chart The Inputs worksheet (Figure 5) asks for the following parameters corresponding to each of the before m entioned tendon segments in a span: mass per unit length (mu), stiffness (S), num ber of strands, length of the tendon segment (L), and the first two vibrat ion mode frequencies, corresponding to 4 vibration tests (two straight impact and two side impact). The length of the tendon segment is obtained from direct measurements, and the number and type of strands is obtained from construction data. Other parameters estimations are explained in the following sections. The Calculations worksheet estimates the tension per strand for each tendon using equations (1,2) explained in the previous section. The frequency used for each mode is the average of the four vibration tests. The final estimated
9 tension is the average of the estimated tension for each mode. This worksheet is hidden to make sure the calculatio ns stay uncorrupted by the user. The Results worksheet displays the estima ted average tens ion per strand for each tendon (kN/strand) and the Quality %, or percent difference between the tension estimated from the first and second mode frequencies f1 and f2 (Figure 6). The Chart worksheet (Figure 7) graphically displays peer tendon tensions to facilitate flagging potentially deficient tendons. SegmentL metersStrandsmuSTestMode 1Mode 2 SWL18.651919.26812727518.817.6 28.717.4 38.817.5 48.817.6 SWM13.242426.99616570011224.1 212.124.1 312.124.2 412.124.3 SWS7.8522829.641246875120.742.3 220.742.3 320.842.5 420.842.5 SEL18.6411919.26812727518.817.5 28.817.4 38.817.5 48.817.5 SEM13.2582426.996165700112.124.3 212.124.3 31224.5 412.124.5 SES7.8792829.641246875120.642.2 220.642.2 320.943 420.943 NWL18.5811919.26812727518.717.4 28.617.4 38.717.4 48.717.5 NWM13.2152426.99616570011224 21224 31224 41224 NWS7.9112829.64124687512040.9 220.240.9 320.241.4 420.241.4 NEL18.5541919.26812727518.917.7 28.917.7 38.917.7 48.917.7 NEM13.232426.99616570011224 21224 312.124.2 412.124.2 NES7.9012829.641246875120.341.6 220.341.6 320.541.9 420.541.9 SegmentL metersStrandsmuSTestMode 1Mode 2 Bridge's Name Figure 5. Input Worksheet
10 Results Segment Avg Tension (kN/strand) Quality% SWL 102.15 0.85 SWM 105.47 0.89 SWS 94.49 0.26 SEL 102.05 2.03 SEM 106.83 1.07 SES 95.83 1.41 NWL 99.56 0.31 NWM 103.55 1.14 NWS 90.04 0.02 NEL 103.58 1.72 NEM 104.72 1.12 NES 92.53 0.57 Figure 6. Results Worksheet 50 60 70 80 90 100 110 120 130 140 150 TENDONkN/STRANDFIRST ESTIMATE SPAN # SWLSES SEM SWMSEL SWS Figure 7. Output Field Graphic
11 2.2 Mass per Unit Length (mu) Prior to testing the operator can ca lculate the tendon mass per unit length (mu) from design or construction dat a by using the following equation. )] ( ] 4 ) 2 ( ) [( [ 1 0 ) (2 2 g s s g p m kgA N d D d D d mu + + =  mu = mass/length (kg/m) D = outer tendon diameter (cm) D = 8.92 (cm) (3.51 in diameter duct) D = 11.45 (cm) (4.51 in diameter duct) d = polyethylene duct wall thickness (cm) d = 0.43 (cm) (3.51 in diameter duct) d = 0.48 (cm) (4.51 in diameter duct) N = # of strands As = Area of one strand As = 0.99 (cm) ( in strand diameter) As = 1.44 (cm) (0.6 in strand diameter) p = polyethylene density = 1.0 (g/cm) g = hardened grout density = 1.84 (g/cm) s = steel density = 7.8 (g/cm) Example: Tendon of 10cm di ameter, 0.6cm duct wall th ickness, and 17 strands of type in diameter, m = 23.04 kg/m.
12 This value of mu is permanently input to the spreadsheet before conducting the tests. If different types of tendons are tested in the same bridge various values of mu are entered in designated cells. 2.3 Stiffness (S) Geometric strand arrangem ent within the tendon cr oss section is nonuniform. The strands tend to crowd again st the inside curvature of the tendon path as it is altered at end span and deviation blo cks. As the steel strands contribute the most to the composite fl exural stiffness, their non-isotropic distribution provides greater flexure stiffne ss in the horizontal than in the vertical direction. Thus, different sets of vibration frequencies may be expected for vibration deflection along those two directio ns [2, 12]. Averaging the two sets of frequencies (peak doublets) for each mode in the tension spreadsheet corrects to some extent the stiffness difference along the tendon. For an ideally bonded, close-packed arrangement of strands and grout the sti ffness can be estimated by the following equation (4). Th is is a rough estimation, since strands distribution along the tendon are not always arr anged as bonded and close-packed (Figure 8,9), in which case stiffness can be signif icantly larger as the strands moment of inertia increase. )] ( 4 ) ( ) 2 ( [ 4 6 ) (2 4 4 2 g s s p g pE E A N E E d D E D m N S + + =  S = Tendon stiffness (N.m) D = outer tendon diameter (cm) D = 8.92 (cm) (3.51 in diameter duct)
13 D = 11.45 (cm) (4.51 in diameter duct) d = polyethylene duct wall thickness (cm) d = 0.43 (cm) (3.51 in diameter duct) d = 0.48 (cm) (4.51 in diameter duct) N = # of strands As = Area of one strand As = 0.99 (cm) ( in strand diameter) As = 1.44 (cm) (0.6 in strand diameter) Ep = polyethylene modulus of elas ticity = 1.276 (GPa)  Eg = hardened grout modulus of elasticity = 40 (GPa)  Es = steel modulus of elasticity = 206.8 (GPa)  Example: A tendon with a 3.51in diameter duct, and 19 strands of type in diameter; S = 120kN-m. The average of 19-strand tendon S values observed at Niles Channel Bridge shows order-ofmagnitude values of 125kN-m and 140kNm , in reasonable agreem ent with the above estimate. This value of S is per manently input to the sp readsheet before conducting the tests. If different types of tendons are tested in the same bridge various values of S are entered in designated cells.
14 Figure 8. Strands Configuration, Tendon 13A Sloping Section  Figure 9. Strands Configuration; Tendon 13A, Horizontal Section  2.4 Tendon Frequencies This section addresses different methods of acquiring the first two vibration mode frequencies, requiring only one operator One of these methods uses the built-in sound card of a computer, another method uses a dataacquisition-board (DAB), and a third method uses a microphone recorder.
15 Programs were developed in Lab-VIEWTM for each of those methods to analyze and display the processed acceleromete r output. Other components used in the data acquisition process are coaxial ca bles, accelerometer Model PCB 338B34, and signal conditioning amplif ier ICP-Model 480E09; all of which are common for the three acquisition methods. 2.4.1 Built-in Sound Card Method The method using the built-in sound card is based on a Dell Latitude 840 computer operating Windows XP. A Lab VIEWTM-based program Analyzer-M was developed to manage data acquisition th rough the Line In port of the 16-bit resolution computer sound card. The program Analyzer-M has three Levels (1,2, and 3). Levels 1 and 2 are both graphica l interfaces that display screens information. Level 3 is involved in data acquisition, data processing, and graphics. On Level 1 the first interface screen appears (Figure 10), which displays the University South Flori da logo and requests the user to press F2. When the user presses the F2 key, the screen front panel for Level 1 closes and the program opens Level 2. The block diagr am of Level 1 is in Appendix 1. The Level 2 interface displays a screen with copyright information (Figure 11). When the user accept s the conditions of use, the Level 2 modulus opens Level 3, otherwise the screen closes and the program stops The block diagram of Level 2 is in Appendix 2.
16 Figure 10. Level 1 Front Panel Figure 11. Level 2 Front Panel
17 Level 3 is the main program which controls data acquisition, creates a .wav file of the acquired data, converts the processed accelerometer signal from time domain to frequency domain, and gra phically displays both domains. The Level 3 block diagram consists of a sequence structure with two main subdiagrams (0,1). Subdiagram 0 (Appendix 3) has two sequence st ructures and one case structure. The first sequence structure specif ies the file size in bytes, without the header, to be recorded (218) and the path to create the file. The number of samples to collect (217) is half the file size without the header since each sample requires 2 bytes. The time for acquiri ng the data is 11.8886s for a sampling frequency of 11,025Hz, and the frequency resolution is 0.084Hz. The second sequence structure contains a while loop The while loop terminates if Start or Record is activated, otherwise the while loop continues to iterate. The Start key F1 activates t he case structure. The case structure contains two while loops, the first of which is used to configure the sound input device (computer sound card) with the chosen options of 11,025Hz sampling, monaural sound quality, 16 bits per sample, and 8192 buffer size The second while loop reads data from the buffer and di splays it in a chart continuously (Figure 12) until the stop ke y (F2) is pressed. This function is used to adjust hammer impact.
18 Figure 12. Level 3, C ontinuous Signal Display Subdiagram 1 (Appendix 4) is activated when the st op key F2 or the start key F3 is pressed, prompting t he user to enter a name for the .wav file to be recorded (Figure 13). Subdiagram 1 ta sks are acquiring, recording, analyzing, and displaying the data. The first s equence structured deals with initial preparations of acquiring and recording t he data such naming the file to be recorded and sound input configuration as s ubdiagram 0. Once the file is named and accepted the next while loop is activated displaying on the front panel in the info box Hit F3 or the push button to begin recording (Figure 14). The push button can be actuated by using a wireless presentation remote control such as TargusTM Model PAUM30, which has a distanc e range of up to 50 feet and thus permits the operator to hit the tendon at the required time without the need of an assistant. This while loop is followed by the module SI Start which starts the data acquisition once the start button is actuated.
19 Figure 13. Level 3, File Name Prompt Figure 14. Level 3, Read y to Record Prompt
20 The next frame is a sequence structure in which the data from the buffer is read, producing a one-dimens ional array of 16-bit inte rgers. A string TEST IN PROGRESS is assigned to info to be displayed on t he front panel (Figure 15) while the data is being collected for the time assigned in subdiagram 0. Figure 15. Level 3, Test in Progress Prompt In the next sequence structure the one dimensional array containing the time domain signal is pa ssed as input to module F(x) which computes the real Fast Fourier Transform. The resulting complex number is normalized with respect to the array size and separated into its polar components. The frequency magnitude is displayed on t he front panel at a frequency spacing or resolution of ~0.084Hz (Figure 16). The time domain signal is also displayed beside the frequency domain graph for the purpose of quality control. If the data are
21 acceptable then the one-dimensional array containing the time domain signal is saved as a 257kB 16bit wave file. Figure 16. Time and Frequency Domain Graphs The Analyzer-M program was prepared as an executable file. Also, an executable program called WavePlayer-Mono (Figure 17) was made to retrieve the data from the wave files that were sa ved during the vibratio n tests; its block diagram is in appendix 5. Appendix 6 incl udes the program in stallation guide, the Step by Step Procedure to operate equipment and Analyzer-M software, and tendon preparation.
22 Figure 17. WavePlayer-Mono Front Panel 2.4.2 Data Acquisition Board Method The data acquisition board (DAB) method uses a LabVIEW-based program ( ANALYZER-DAB ) to acquire and analyze data from a 12-bit resolution NI USB-6008 data acquisi tion board. The program ANALYZER-DAB is similar to the program ANALYZER-M but several key differences exist in level 3 of the program (Appendix 7). Data acquisition in ANALYZER-DAB is controlled by a NIDAQmx Base 2.0 driver instead of the Sound Input VI used in the ANALYZER-M The number of samples collected by ANALYZER-DAB is 213 and the sampling frequency is 800Hz, therefore it has a frequency resolution of ~0.0977Hz. Also, the collected data ar e saved as a binary Single Precision Floats (SGL) file which requires 4 bytes per data sample, which is 32kb total file
23 size. An example of ANALYZER-DAB data collection quality is in Figure 18. The installation guide is in Appendix 8. An executable program called WavePlayerDAB was made like the WavePlayer-MON to retrieve the data from the SGL files that were saved during the vibrat ion tests; the block diagram of WavePlayer-DAB is in Appendix 9. Figure 18. Front Panel Analyzer-DAB 2.4.3 Digital Recorder Microphone Method Another method of acquiring the fi rst fundamental tendon frequencies is by using a digital recorder microphone. Accelerometer, cables, and signal conditioner set up is like that of the pr evious methods but with the adapter cable coming from the signal conditioner connec ted to the microphone line input (mic). The digital microphone recorder to be used is a SONYTM ICD-P210. The SONYTM ICD-P210 allows you to save the record er files in a PC as a 16bit 11000kHz
24 monaural wave file. The record ing files are also easy to identify since their initial assigned names correspond to the micr ophone folder used, their recording position with respect to the other files, and the date and time of the recording. The recording files can be analyzed by using a Lab VIEW program Wave Player Micro which works in a similar way as WavePlayer-MON The block diagram of Wave Player Micro is in Appendix 10. A sample of data quality recording and analysis of the microphone method is in Fi gure19 with microphone sensitivity set to low and recording mode set to HQ. Figure 19. Wave Player Micro Front Panel
25 Chapter 3 Practical Vibration Evaluation Validation* The practical vibration evaluation ex plained in chapter 2 to detect tendon tension was validated on nearly full-sca le tendons constructed at the FDOT Structures Laboratory as part of an ongoing parallel investigation . A Tendon Test Facility (TTF) was c onstructed at the FDOT Structures Laboratory (FSL) in Tallahassee, FL. for large-scale model validation tests. The TTF had one fixed reinforced concrete anchor block (South end) and one movable anchor block (North end) with horizontal anchors approximately 9m away. The anchor assemblies were Ty pe E manufactured by VSL. The fixed block had provisions for horizontal anchors at the same elevation as those in the movable block, allowing for straight hor izontal tendons with a free length of ~9m (Figure 20). Tension was applied by disp lacing the movable block by the required amount with hydraulic jacks and than placing stops between the block and the end of the frame. Load cells monitored the tensioning force at the movable block end allowing for precise computation of the force. Two dupli cate full length horizontal tendons (0A and 0B) were c onstructed. Both tendons had twelve one* Parts of the work in this chapter have appeared in A. S ags, T. Eason, C. Cotrim and J. Lopez-Sabando, Validation and Practical Procedure for Vibrational Ev aluation of Tendons, Project No. BC 353#44, 158 pages, Draft Final Report to Florida Department of Transportati on, University of South Florida, Tampa, Fl, December, 2007 .
26 half in low-relaxation seven wire str ands per ASTM A416 grade 1860 supplied by VSL. Figure 20. Tendon Test Facility The design stretching stress capability was 1800kN, corresponding to 80 % of the Guaranteed Ultimate Tensile Str ength (GUTS) but actual stretching stresses were typically 1500kN, (67% GUTS). The strands were contained in a high density polyethylene (HDPE) duct type DR17 3 in NPS (8.9cm outer diameter with a wall thickness of 0.4cm) Galvanized steel pipes 7.62cm internal diameter and 0.48cm wall thickness emerged from the end and deviation blocks and served as attachment points for t he polymer duct by means of an 8.9cm inner diameter 15.24cm long Neoprene duct coupler. After stretching the
27 tendons, they were grouted with QPL938 grout manufactured by Masters Builders using a colloidal pump. There wa s no indication of gr out voids in any of the tendons constructed. Further details are given in . After a grout setting period of 7 da ys minimum, vibration testing was conducted on the free length(s) of the tendon. The vibrat ion tests consisted of basic tests as explained in chapter 2, in which the tendon was impacted at a point 1/6 of the free length away from one of the blocks at either end of the free length of the tendon, and the accelerometer was pl aced at 1/3 of the distance from the same or opposite end of the tendon (Figure 21). Figure 21. Tendon Test Set Up Tendon impacts were conducted with a rubber hammer with a total mass of 611 grams. The accelerometer, Model 338B34 by PCB Piezotronics with a
28 sensitivity of 10.00 mV/g, was placed wi th its sensing axis attached to 45 from horizontal side of the tendon. A flexib le coaxial wire ~0.25cm diameter connected the accelerometer to a Model 480E09 signal conditio ning unit by PCB Piezotronics, with a voltage gain set to 10x, resulting in a signal amplitude upon impact typically < 0.6V The signal was acquired using the Sound Card line input of a Model Latitude C840 computer by Dell and controlled by a LabVIEWTM program similar to Analyzer-M which was described in Chapter 2. Input tendon parameters (Table 1) we re obtained as explained in chapter 2. Tendon tension ap proximation results (Table 2) indicates a small difference with the load cells (less than 4%) for both tendons (0A, 0B) for the experiments chosen for analysis. This result is consis tent with the general level of agreement between vibrational and load cell tens ion estimates obtained in a broader investigation in progress , although results from particu lar test sequences may differ Table 1. Tendon (0A, 0B) Input Parameters SegmentL metersStrandsmuSTestMode 1Mode 2 Test # 0B-South9.2791217.5759102477115.6532.050BBASSC 215.6532.050BBASSD 315.6531.880BBATSA 415.6531.880BBATSB 0B-North9.2791217.5759102477115.6532.050BBASNA 215.6532.050BBASNB 315.6531.880BBATNA 415.6531.880BBATNB 0A-South9.3191217.5759102477116.1632.80ABASSA 216.1632.80ABASSB 316.3532.80ABATSA 416.4132.80ABATSB 0A-North9.3191217.5759102477116.2432.80ABNASA 216.2432.80ABNASB 316.2432.80ABNATA 416.2432.80ABNATA
29 Table 2. Tendon (0A, 0B) Results Results Segment Avg Tension (kN/strand) Quality% Load Cell (kN/strand) % Difference 0B-South 109.17 1.86 113.14 3.6 0B-North 109.17 1.86 113.2 3.6 0A-South 118.30 0.77 122.28 3.3 0A-North 118.07 0.37 121.8 3.1
30 Chapter 4 Early Warning Corrosion Probes Methodology* 4.1 Test Environments Controlled humidity chambers were constructed using lidded glass fish tanks. The dimensions of the glass cham ber were 51x31x26cm, total volume of 39000cc. A 95% RH environment was implemen ted by adding 266gr of NaCl (common salt) to 3 liters of distilled water at the bottom of the chamber. The concentration of NaCl to water was calculated using equation (5) derived by Cinkotai (1971) . Where Cs=ms/(m s+mw), and ms, mw are the mass of solute and water, respectively. 255 1 4867 0 1 Cs Cs RH = (5) This steady state RH should be stabl e over the experiments length even if the lid of the chamber is removed for short periods of time. The mass of water vapor present in the air inside the chamber at 95% RH and 23 Celsius was calculated to be 0.81g . The latter is an insign ificant amount compar ed with the 3000g total mass of water in the chamber so lost water can be replenished by Parts of the work in this chapter have appeared in L. Ta veira, A. Sags, J.Lopez-Sabando, and B. Joseph, Detection of Corrosion of Post-Tensioned Strands in Grouted Assemb lies, Project No. BD544-08, 71 pages, Final Report to Florida Department of Transportation, University of Sout h Florida, Tampa, Fl, October 31, 2007 .
31 evaporation without substant ial change in the solution composition. The above estimation is based on the following equations: vaporv V m = v vaporP T R v = g vP RH P = (6) Where m is the mass of water vapor V is the volume of air, vvapor is the saturated vapor specific volume, R is the water gas constant (461.5J/(kg K), T is the temperature in Kelvin, Pv is the partial pressure of water vapor, RH is the relative humidity, and Pg is the saturated water pressu re at 23 Celsius (2.82kPa). The 75% relative humidity was a ccomplished by introducing inside the chamber 1.0 liter of water saturated with sodium chloride (NaCl) . 4.2 Electrochemical Probes An atmospheric corrosion electrochem ical test array was designed and constructed using a methodology inspired by that of Mansfeld and Kenkel . The probes consist of two 5mm diameter steel wires extracted from an ASTM A416M-98 high strength str and. The wires were 0.508cm diameter and 10.5cm long. Both wires were attached parallel to each other, with a gap between them of 0.6mm. Plastic spacers at the end of the probes kept the two wires electronically isolated fr om each other. A thin, stainless steel wire <1mm diameter was spot welded to the end of the each wire probe to permit external measurements. Two probes with the afore mentioned characteristics were made and immersed in fluid, 0.42 water/cement ratio grout (type 1 Portland cement) and then lifted, forming upon curing a thin grout layer on the surface and across the gap. The length of the probes cove red with grout was 9.0cm, with the
32 extremes of the wires uncovered by grout [Figures 22, 23]. Another two probes were made three days later with similar characteristics but with a gap between the wires of 1.0 mm, and a thicker layer of grout covering the wires. The grouted probes were cured in a 100% RH glass chamber for a day and inserted afterwards into the 95% RH chamber at a temper ature of 232C. Holes were made and then caulked in the lid of the chamber for the stainless steel wires attached to the probes to al low external measur ements of the probes without disturbing the corroding conditions inside the chamber. Figure 22. Electrochemical Probe Before Grouting Figure 23. Grouted El ectrochemical Probe 4.3. Electrical Resistance Probes The Electrical Resistance (ER) method is another corrosion monitoring approach. ER probes use the simple prin ciple of an increase in electrical resistance produced by a decrease in the se ction thickness of the metal as it corrodes. Two different ER probes were developed in this study. The first
33 generation ER probe contained two identic al plain low carbon steel rebar tie wires 120cm long and 1.60mm di ameter, in the as-receiv ed condition (dark mill scale on the metal surface). One of the wires, the wo rking element, was exposed to the corrosive atmosphere inside th e chamber. The other, the reference element, was protected by sealing it in side the probe body (pvc pipe) from the corrosive medium. The covered wire pr ovided a reference for evaluating changes in the uncovered wire and also served to compensate for the effects of temperature changes on resistance (Figur e 24). The latter can be an important source of error, as the resistivity of st eel varies roughly by 0.3% for every 1C change near ambient temperature . To signal conditioning amplifiers Exterior Wire Interior Wire Figure 24. ER Probe Interior Design The corrosion rate (Corrrate) of ER probes can be det ermined by the radius change of the corroding wire ( r) over its exposed time in days (t) and multiplied by 365(days/year): t r Corrrate365 = (7)
34 A 60Hz AC, 80mA excitation curr ent was created with a 21V output transformer in series with a 260 resistor and the probe. The resistance of each wire was ~0.8, resulting on only ~10mW to tal probe power dissipation, a negligible amount of heat pr oduction rate considering the dimensions of the probe. Two 100X amplifiers and a 0.1mV A. C resolution multimeter were used to measure the potential drop across wires. The sensitivity of this probe was calculated to be 1/1428 parts (0.6 m) of the corroding wire radius. The above estimation is based on the following equations: Probe sensitivity = rcorr/ro (8a) f o corrr r r = (8b) EIL 1 r f2 1 r o2 (8c) where E is the minimum drop in potential that can be detected (0.05mV), I is the current in the circuit, L is the length of the wires, is the resistivity of steel, ro is the original radius, rf is the final radius after corroding, and rcorr is the change in radius due to corrosion that can be detected. Two ER probes with the above menti oned characteristics were tested only in the low RH (75%) insulated glass ch amber at a temperature of 232C, mainly to check operation at the electr onic signal acquisition system as corrosion rates were very low in that environment.
35 A second generation of ER probes was later designed to simplify measurements and improve s ensitivity. A potentiomete r was added to the earlier design to create a bridge as a way of m easuring the change in the resistance of the corroding wire [Figure 25]. The bridge was initially balanced (Vout = 0) by adjusting the potentiometer, so the resist ance ratio of the probe wires was the same as that of the potentiometer. The initial resistance ratio, Rout / Rin (see Figure 25), of the probe was ne arly 1. From that initia l condition, when the wire corrodes the resistance increases by a factor of (1+P). P is a function of the input voltage (Vin) and output voltage (Vout), according with the relationship below (equation develop ed in appendix 11). out in outV V V P = 2 4 (9) The corroding wire radius ()corrr can be calculated by the following equation: () P r ro corr+ = 1 1 (10) Where or is the original radius of the wire, Equation (10) is derived in Appendix 12. For a constant supply voltage from the transformer, the input voltage shouldnt change over time since the increa se in the resistance of the corroding wire is negligible in comparison with the resistance of the whole assembly. If the input voltage is known a pr iori, P can be calculated by just measuring the bridge output voltage.
36 Figure 25. Schematic of ER Probe. A 0.03 F capacitor was placed across R2 to minimize phase shift. Since the input voltage is not constant, because of fluctuation in the power grid, input and output voltages need to be measured at the same time for an accurate calculation of P. To measur e simultaneously the i nput and output AC voltage across the bridge, a Lab ViewTM program P-Measurements was developed. Other components of the data acquisition system are: a data acquisition board (DAB) and a 100-gain amp lifier to condition the signal between the bridge voltage divider and the acquisition board. The block diagram developed to m easure the bridge input and output voltage is shown in the Appendix 13. The P-Measurements program consists of three parts, the first part configures the DAB and converts binary counts to engineering units, the second part perfo rms voltage measurem ents, and the third 1 R R R R 0 P 0 V @ C I2 1 in out out = = = out in outV 2 V V 4 P = Probe Wires Potentiometer V in = 21v Vout R out (1+P) RinR 1 R2 260 ohms
37 part analyzes the data and calculat es P. The front panel of the P-Measurements program (Figure 26) lets you choose the voltage range and displays the RMS, average voltage peaks and P. Figure 26. Front Panel of The P-Measurements Program The data acquisition board used for this task was an USB-1608FS from Measurement Computing, with a 16-bit pr ecision (0.03mV resolution error). The sampling rate was set to 24000Hz and the number of samples was 6000. Another improvement was placing a ~0.03 F capacitor (value selected by trial and error) across the potentiometer arm R2 until ther e was nearly zero phase shift across the Vout terminals. Without that c apacitor a small phase shift, due to the mutual inductance of the inter nal and external co iled wires and the magnetic properties of steel, was present which prevented obtaining a sharp null during initial adjustment. The probes were made with counter turn coils to
38 minimize induction effects, but the c apacitor was still needed for improved compensation. The resulting system imbalance sensitivity was ~0.005mV, corresponding to a detectable change of corr oding wire radius in the order of 1 part in 4000 (0.2 m). Four of this second generation ER probes were made and placed in the 95% RH glass chamber described earlier fo r the Electochemical probes. Two of the ER probes were dipped in grout as described earlier for the Electrochemical probes. 4.4 Probe Materials Characterization Metallographic examination of t he strand wires (ASTM A416) and the steel tie wires cross sections was conducted to re veal and compare their microstructures. The specimens were mounted metallogr aphically, ground, fi ne polished to a 0.05 m alumina suspension finish, and etc hed with 2% nital solution. The micrographs in Figure 27 reve al the fine eutectoid pearli tic microstructure of the high strength PT wires, and the nearly all ferrite (low carbon) microstructure of the softer steel tie wires These structur es were as expected and the tests served to confirm the identity of the probe materials. 4.5 Gravimetric Measurements Gravimetric measurements were c onducted to compare against the results of the Electrochemical and ER probes to verify their reliability. The weight loss experiments were conducted with ba re and dip-grouted wires. The tests included 20 helically shaped outer wires extracted from actu al 7-wire steel strands from the same stock used for the Electrochemica l probes, and 8 low
39 carbon steel tie wires with mill scale as used for the ER probes. The specimens grouted by dipping were processed as fo r the other tests, and cured for 2 days inside a 100% RH chamber before in troducing in the 95% RH chamber. The helically shaped wires were 0. 508cm in diameter and 35cm long; the steel tie wires were 0.16cm in diameter and 46cm long. Before and after the test, the strand and tie wire specimens were cleaned per ASTM G1 and then weighed to 10-3 and 10-5 grams precision respectively. ASTM A416 Metallography ER probe wire 100 m 100 m 100 m 100 m Cross SectionTransverse SectionFine eutectoid Pearlitic microstructure Microstructure of High-Strength PT Wire (Eutectoid) and Steel Tie Wire ER Probe Wire (Low carbon steel). Nearly all Ferrite Figure 27. Probe Materials Characterization
40 Chapter 5 Early Warning Corrosion Probes Results and Discussion* 5.1 Electrochemical Probes The Electrochemical probe system, illust rated in cross section in Figure 28 can be approximated as behaving as the equivalent circuit in Figure 29. Grout film Grout bridge Wire 1 Wire 2 Figure 28. Schematic of Electro chemical Probe Cross-Section Figure 29. Electrochemical Probe Equivalent Circuit *Parts of the work in this chapter have appeared in L. Ta veira, A. Sags, J.Lopez-Sabando, and B. Joseph, Detection of Corrosion of Post-Tensioned Strands in Grouted Assemblies, Project No. BD544-08, 71 pages, Final Report to Florida Department of Transportation, University of Sout h Florida, Tampa, Fl, October 31, 2007.. Ci Rp Cc Rs CiRp Wire-grout interface Wire-grout interface Grout Bridge CS
41 Ci are constant phase angle elem ents, of admittance Y = Y0i (j2f)n, representing the interfacial capacitance of the me tal-grout interface of each wire (Y0i and n are the constant phase angle elem ent parameters ,), RP is the polarization resistance of that interface, and CS and RS represent, respectively, the dielectric capacitance and ohmic resistance of the grout bridge, the former of which was found to be not negligible for the system and frequency r ange examined. For simplicity the two metal-grout interfaces were assumed to behave similarly. Thus, the measured impedance could be represented by a single Cm-Rm parallel combination (where Rm=2 RP, and Cm is a constant phase element with parameters Y0m= Y0i/2 and n ) in series with the CS RS parallel combination as shown in Figure 30. Figure 30. Simplified Equivalent Circ uit for the Electrochemical Probe EIS measurements of the Electrochemical prob es were carried out at the open circuit potential (OCP) with 10mV RMS amplitude in the frequency range Cm Cs Rm Rs
42 from 100kHz to 10mHz, to determine polarization resistance and corrosion currents of the probes. The experiment s were performed periodically during 264 days. Both experiments were perform ed using Parstat 2263 from Princeton Applied Research, Oak Ridge, U.S.A. and Gamry PCI-4 from Gamry Instruments, Warminster, U.S.A. potentio stats. The electroc hemical parameters were estimated by using the program s Gamry Echem Analyst from Gamry Instruments, Warminster, U.S.A or Zv iew2 from Scribner Associates, Inc., Southern Pines, U.S.A. The reference a nd counter-electrode were connected to one wire of the probe and the working el ectrode to the other one, so the impedance measured corresponded to the wi re-grout-wire series combination. Other records were temperature and re lative humidity inside the chamber. Examples of EIS results for 1mm and 0.6mm gap electrochemical probes exposed to the 95% RH environment ar e shown in Figure 31. Two depressed semicircles can be distinguished. Th e first semicircle corresponds to RS and CS, while the second is related to Rm and Cm as discussed above and illustrated in Figure 32. Because the grout resist ance-capacitance component has a very short time constant, the analysis to deter mine the circuit parameters relevant to the polarization of the corrosion reactions was limited to the frequency interval 10mHz to 1Hz, where the effect of Cs is small. Thus, the eq uivalent circuit used for the actual EIS data analysis had RS, Rm and Cm as the only fit parameters and the results in the following are discussed in terms of the parameter values for one of the interfaces.
43 Figure 31. EIS Behavior of (a) 1mm Gap and (b) 0.6mm Gap Probes. The solid line indicates the model fitting -5.E+04 -4.E+04 -3.E+04 -2.E+04 -1.E+04 0.E+00 0.E+00 1.E+04 2.E+04 3.E+04 4.E+04 5.E+04 Z' () Z'' () 1mm gap probe 236 days exposed 1 Hz RS = 6.1 10 3 RP = 1.6 10 5Y0 =6.8 105 F sn-1n =0.4710 mHz(a) -2.5E+05 -2.0E+05 -1.5E+05 -1.0E+05 -5.0E+04 0.0E+00 0.0E+00 5.0E+04 1.0E+05 1.5E+05 2.0E+05 2.5E+05 Z' () Z'' () 0.6mm gap probe 39 days exposedto 95% RH RS = 6.8 10 4RP = 2.8 10 5Y0 =1.4 105 Fsn-1n =0.381 Hz 10 mHz(b)
44 Figure 32. EIS Interpretation Treating the reactions in the system as if they were under simple activation polarization, t he corrosion rate can be estimated by the Stearn-Geary relationship (equation 11) between RP and the corrosion current density (icorr). P corrR I = ) ( 303 2 c a c a + = (11) where a and c are the Tafel slopes ~ 0.12V ,. CpeC c Rm R s C c Grout Dielectric Rs Rs+Rp Z Z
45 Another important paramet er is the solution resistance (RS). The solution resistance or grout resistance (equation 12) is on first approximation proportional to the resistivity of the grout ( ), the cross section ar ea of the grout between wires (A), and the distance between wires (d). A d Rs = (12) The actual system is more complica ted (Figure 28), but for a fixed effective distance between wires, the resistivity of the grout can be determined if an effective area is known, and vice versa. For probes with similar grout, distance, and conditions such as RH and temperatur e, their solution resistance can be assumed to be equal to the inverse of thei r effective contact areas multiplied by a constant. The RS and the RP trends for the 1mm gap electrochemical probes exposed to 95% RH environment are shown in Figu re 33. Upon initial exposure to 95% RH the RS and the RP values were small, but t hen increased drastically tending to stabilize after ~50 days. The increase in Rs likely reflects the establishment of a less interconnected pore network in the grout as curing matures. Other factors that can alter the resistivity of the grout are temperature and t he relative humidity. Figure 34 shows that the values for RP in the 0.6mm probes were in the same order as those for the 1mm probes, but not as stable. The values of Rs were about one order of magnit ude greater than those for the 1mm gap probes and less stable as well.
46 Figure 33. RS and RP Trends for Duplicate (No.3 and 4) 1mm Gap Electrochemical Probes Figure 34. RS and RP Trends for Duplicate (No.2 and 0) 1mm Gap Electrochemical Probes 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05 1.E+06 0 50 100 150 200 250 300 t (days) Rs 3 Rs 4 Rp 3 Rp 4 Rs, Rp () 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05 1.E+06 1.E+07 0 50 100 150 200 250 300 t (days) Rs 2 Rs 0 Rp 2 Rp 0 Rs, Rp ()
47 Polarization resistance can also be determined by the linear polarization resistance (LPR) technique. The LPR method is based on the relationship between small amplitude appl ied polarization potentials and the corresponding polarization current of a corroding syst em. The LPR experiments were conducted using the Gamry PCI-4 Potentiostat at 0.1mV/s, starting fr om the OCP to an overpotential of 10mV in the cathodic dire ction. A typical LPR potential-current curve of a 1mm gap electrochemical probe is shown in Figure 35. The results of the LPR measurements are not strai ghtforward since the experimental arrangement to measure the polarization re sistance can only directly sense the values of the total voltage and the total ap plied current without distinction of the current demanded by any elem ent of the system . Therefore, to determine the corrosion currents values a refined RP (LPRRP) value was calculated. The LPRRP values were compensated for RS and for the presence of interfacial CPE behavior using the corresponding para meters obtained from the EIS measurements. The compens ation was made by firs t subtracting an amount equal to I Rs from the potential V at each point of the measured current ( I ) V curve obtained in the LPR test, thus obtaining an ohmic resistance-compensated curve I Vcomp. The correction for the current demanded by the CPE used the following relationship [20, 23]: 1 max)] 1 ( / / 1 [ + = n V S Y R Rn n o p ap (13) where Vmax is the maximum compens ated potential applied, S is the scan rate (0.1mV/s), is the Eulers Gamma function, and Rap is the apparent RP determined by the slope at V = Vmax of the I Vcomp curve, and Y0 and n are the CPE
48 parameters obtained from an impedance experiment perfo rmed shortly before or after the test. It is noted that the corre ction represents only a first approximation as it does not take into consideratio n the convolution resulting from the simultaneous presence of the RS and the CPE ,. Figure 35. I Vcomp Curve of 1mm Gap Electrochemical Probe Figures 36 and 37 show co mparable relative trends for the RP values estimated by LPR and EIS methods for both the 1mm and 0.6mm gap probes, but the RP obtained from LPR tended to be lower than those ones from EIS (EIS-RP) by about a factor of 2. 9.00E-03 1.10E-02 1.30E-02 1.50E-02 1.70E-02 1.90E-02 2.10E-02 -2.00E-07-1.00E-07 0.00E+00 1.00E-07I (A) Vcomp R P = 319 k Ra= 95 k
49 1.E+03 1.E+04 1.E+05 1.E+06 200220240260280 t (days) Rp ( ) Rp-EIS 3 Rp-EIS 4 Rp-LPR 3 Rp-LPR 4 Figure 36. RP Values Estimated by LPR and EIS Methods for Duplicate Probes with 1.0mm Gap 1.E+04 1.E+05 1.E+06 1.E+07 200220240260280 t (days)Rp ( ) Rp EIS2 Rp EIS 0 Rp-LPR 0 Rp-LPR 2 Figure 37. RP Values Estimated by LPR and EIS Methods for Duplicate Probes with 0.6mm Gap t (days) t (days) Rp ( ) Rp ( )
50 Corrosion current Icorr values were calculated by the Stearn-Geary relationship (equation 11), assuming that both wires were corroding equally. The corresponding nominal corrosion rates were estimated per equation (14) assuming an area of 8cm2 for the metal in effective contact with grout on each of the probe wires. That area value was estimated by making the rough assumption that all the excitation current flows th rough the one-half of the wire perimeter facing the other wire. The time evolution of Icorr for both types of electrochemical probes is shown in Figure 38 and 39. The Icorr values were, in general, larger in the first days of exposure but after several days decreased to ~0.3A and ~0.1A for 1mm and 0.6mm gap probes, respectively. The Icorr for 0.6mm gap probes were less stable than those of the 1mm gap probes, reflecting the instability of the estimated RP values noted before. The instantaneous corrosion rate wa s calculated by the Stearn-Geary relationship introduced earlier (Equation 11) and by the Faradaic conversion formula (equation 14). ) /( A F n M I dt dWcorr = (14) where W is the mass lost of the corroding metal in g/cm2, t is the time in seconds, M is the atomic mass of iron 55,845g/mol, A is the effective area, n is the valence of iron (2), and F is the Faradays const ant 96,485C/mol. The result is converted to cm/y (later expressed as m/y) by dividing the ma ss by the density of iron 7.87g/cm3. While there is considerable uncertain ty in the effective area of these
51 specimens and from the other simplifying assumptions used, the results suggest that corrosion rates estimated by this method were in the order of ~0.3m/y. Figure 38. Icorr Trends of Electrochemi cal Probes with 1.0mm Gap Figure 39. Icorr Trends of Electrochemical Probes with 0.6mm Gap I (1.0mm) 1.E-08 1.E-07 1.E-06 1.E-05 1.E-04 0 50 100 150 200 250 300 t (days) Icorr (A) EIS 3 EIS 4 I (0.6mm) 1.E-08 1.E-07 1.E-06 1.E-05 1.E-04 0 50100 150 200 250 300 t (days) Icorr (A) EIS 2 EID 0
52 5.2 Electrical Resistance Probes The cumulative (i.e. averaged from the beginning of exposure until an exposure time t ) corrosion rate of ER probes in m/year was determined by: t r CR 365 = (7) where r is the radius change of the corroding wire in m ( rinitial rcorr) and t is the exposure time in days An instantaneous (actually short interval) corrosion rate can be calculated by evaluating equations (10) and (7) using the short interval t between two measurements, and using the first measur ement as the initial condition. The instantaneous and cumulative corro sion rate trends for grouted and bare ER probes exposed to the 95% RH environm ent are shown in the Figures 40 and 41. Higher corrosion rates were observed in the first days of exposure especially for the bare steel probes but after 50 days the rates for both conditions reached a plateau of ~5 m/y. The fluctuations of the instantaneous corrosion rate may be attributed to the resolution of the individual measurements and/or minor temperature and RH fluctuations. The cumu lative corrosion rates for ER probes had comparable decreasing trends. After 98 days the cumulative corrosion rates were 12 m/y and 24 m/y for grouted and bare ER probes respectively, reaching, after 196 days, 8 m/y and 15 m/y.
53 Figure 40. ER Probes Instantaneous Corrosion Rate Figure 41. ER Probes Cumulative Corrosion Rate 0.1 1 10 100 1000 0 50100150200 250 300t (days) m/y Bare steel (L) Bare steel (S) Grouted 1 Grouted 2 0.1 1 10 100 1000 0 50 100150200250300t (days) m/y Bare steel (L) Bare steel (S) Grouted 1 Grouted 2
54 In the 75% RH chamber none of t he specimens exam ined experienced measurable corrosion rates, so that cond ition served as a baseline control. The result is consistent with the expectat ion that exposure at 75% RH does not meet the conditions necessary to trigger atmosp heric corrosion. For bare metals that condition is typically encountered abov e 85% RH , consistent with the present results. 5.3 Gravimetric Measurements The weight loss measurements yielded re sults comparable to those of the ER probes as illustrated in Figure 42. After 98 days of exposure the average corrosion rate for grouted and bare helically shaped wires were 11 m/y and 12 m/y respectively, and 10 m/y and 11 m/y after 196 days. The average corrosion rate for bare st eel tie wires was ~ 13 m/y after 98 and 196 days of exposure. grouted bare 196 days 98 days 6 7 8 9 10 11 12 m/yGravimetric Measurements of Strand Wires 196 days 98 days Figure 42. Corrosion Rate of Grouted and Bare Steel Strands Exposed to a 95% RH, Estimated by Weight Loss Measurement
555.4 Early Warning Probes Discussion The observation of rust on some electroch emical probe wires clearly indicated that significant corrosion was taking plac e in the 95% RH chamber. However, the electrochemical probe apparent corrosion rates were one order of magnitude lower than those obtained by ER or weight loss. This discrepancy may be attributed in part to uncertainty in estimating the effective probe area in contact with grout. That area may be much less than the nominal assumed value because of cracks in the grout or disbondm ent at the grout-metal interface, thus greatly underestimating the actual rates over the remaining area of contact. Another likely cause of insensitivity would be that the assumption of equal electrochemical behavior at the two metalgrout interfaces in a probe is wrong. If corrosion were to start at only one of the interfaces with the other largely in the passive condition, then the total series impedance would still be very large and the corresponding apparent current density would stay low until both wires are simultaneously in the active condition. The ER probes and weight loss measur ements showed evidence that at 95% RH the corrosion rates were considerably high. That rate was expected from the relatively thin effective electrolyte layer present on the metal surface in the air space case. The grout film was not particu larly protective, as shown by similar corrosion rates in the bare steel and grouted specimens. Tests, after long exposure times, with pH paper and spra yed phenolphthalein on the grout film on probes in the 95% RH chamber indicat ed a near neutral pH, meaning that the thin layer of hardened grout had eventually carbonated in the chamber
56 environment. Thus, the grout no longer h ad passivating properties to the steel and it is not surprising that measurabl e corrosion was taking place. This was further confirmed by direct observation of rusting on ER, weight loss and even some of the electrochemical probe wires. The worst-case surface-averaged corro sion rate values observed in the high humidity chamber (about 10 m/y), if sustained over 10 years would correspond to an average loss of diameter of 200 m, or about 8% reduction in cross-sectional area in a 5 mm diamet er wire. Such a loss may already be considered of concern even if it were uniform, considering that one decade is a relatively short time compared to ty pical design life goals (e.g. 75years). As corrosion is likely to show some degree of localization, critical loss of strength could occur even earlier. Thus, these findi ngs highlight air space corrosion as a potential cause of strand failure.
57 Chapter 7 Conclusions A simplified test and analysis procedur e for rapid screening of structures with commonly encountered tendon config urations was developed and validated on nearly full-scale tendons constructed at the FDOT Structures Laboratory. The developed practical vibration tendon tension approximation was validated against nearly full-scale tendon s, showing less than 4% difference between the tension obtained by the simp lified vibration me thod and independent measurements from load cells. Electrical Resistance (ER) probes customized for PT anchor air space conditions were constructed and their oper ation with readily available electronic instrumentation was demonstrated. The probes showed adequate sensitivity to detect the corrosion rates of interest, and the results were validated against direct gravimetric measurements. Electrochemical probes for EIS and LPR measurements in PT anchor air space conditions were constructed and their operation with readily available electronic instrumentation was demonstra ted. However, sensitivity may be low and the interpretation of the electrochemic al probe results needs to be refined to better assess their usefulness.
58 There was good correlation betw een EIS and LPR measurements showing that the latter, simpler me thod has good potential for practical implementation. Simulated air-space corrosion experi ments showed that an aggressive environment may evolve in the grout void even on strand wires covered with a residual hardened grout layer, resulti ng in corrosion rates that may have damaging effects in a relatively short service time.
59 References  H. Wang, A.A. Sags; Corrosion of Post-Tensioning Strands, Final Report to Florida Department of Transportation BC353-33, Florida, November 1, 2005.  A.A. Sags, S.C. Kranc, and T.G. Ea son; Vibrational Tension Measurement of External Tendons in Segmental Pos ttensioned Bridges Journal of Bridge Engineering, Vol. 11, No. 5, September/October 2006, pp. 582-589.  J.Corven, A. Moreton; Post-Tens ioning Tendon Installation and Grouting Manual Federal Highway Administration U.S. Department of Transportation, May 2004.  A.A. Sags, R.G. Powers, and H. Wang; Corrosion/2003, Paper No. 03312, NACE International, Houston, 2003.  H. Wang, A.A. Sags, and R.G. Powe rs; Corrosion of the Strand-Anchorage System in Post-Tensioned Grouted As semblies, Corrosion/2005, Paper No. 05266, NACE International, Houston, 2005.  A.T. Ciolko and H. Tabatabai; Nondestructive Methods for Condition Evaluation of Prestressing Steel Strands in Concrete Bridges Final Report to Transportation Research Board Nationa l Research Council NCHRP Project 1053, 1999.  DYWIDAG Bonded Post-Tensioning Systems (Brochure) DYWIDAG-Systems International, 2005.  B. Pielstick, and G. Peterson; Gr outing of Bridge Post-Tensioning Tendons Training Manual Florida Department of Transportation, Federal Highway Administration, July 2002  A. Naaman, and J. Br en; External Prestressing in Bridges, American Concrete Institute, ACI SP-120, Detroit, 1990.  A. A. Sags,T. Eason,C. Co trim,J. Lopez-Sabando Validation and Practical Procedure for Vibrational Evaluat ion of Tendons, Draft Final Report to Florida Department of transportation BC 353#44, Florida, December, 2007.  P.M Morse; Vibration and Sound Mc.Graw Hill, N.Y., 1948.
60  A.A. Sags, S.C. Kranc, and R.H. Hoehne, Initial Development of Methods for Assessing Condition of Post-Tension ed Tendons of Segmental Bridges, Final Report to Florida Department of Transpo rtation BC374, Florida, September 1999.  A.A. Sags, C. Cotrim, V.Balakrishna, Vibrational Evaluation of Tendons in Segmental Sections of Sunshine Skyw ay Bridge Main Spans, Final Report to Florida Department of Transportation BD544, Florida, August 1, 2005.  L. Taveira, A. Sags, J.Lopez-S abando, and B. Joseph, Detection of Corrosion of Post-Tensioned Strands in Grouted Assemblies, Project No. BD544-08, 71 pages, Final Report to Florida Department of Transportation, University of South Florida, Tampa, Fl, October 31, 2007  F.F. Cinkotai. The Behavior of Sodi um Chloride Particles in Moist Air, J. Aerosol Sci. Vol 2, 1971, pp 325329.  M.J. Moran, and H.N. Shapiro. Fundamentals of Engineering Thermodynamics 5th Edition. John Wiley & Sons, Danvers, 2004.  J.F. Young. Humidity control in the laboratory us ing salt solutions, A. Review, J. Appl. Chem. Vol 17, 1967, pp 241-245.  F. Mansfeld, and J.V. Kenkel. Electrochemical Monitoring of Atmospheric Corrosion Phenomena, Co rrosion Sci., Vol 16, 1976, pp.111-122.  D.C. Giancoli, Physics for Scientists & Engineers Third Edition. Prentice Hall, New Jersey, 2000.  A.A. Sags, Lectures Notes on Electrochemical Impedance Spectroscopy for Corrosion Testing, University of South Florida, 2006.  R. Coelho; Physics of Dielectrics for the Engineer Elsevier, N.Y., 1979.  A. Bentur, S. Diamond, and N.S Berke. Steel Corrosion in Concrete London: E&FN Spon, 1997. pp 77.  A.A. Sags, S.C. Kranc, and E. Moreno. "The Time-Domain Response of a Corroding System with Constant Phase An gle Interfacial Component: Application to Steel in Concrete" Corrosion Science, Vol 37, 1995, pp 1097.  R.B. Griffin, Corrosion in Mari ne Atmospheres, in Corrosion: Environments and Industries, ASM Handbook Vol. 13c, Materials Park: ASM International, 2006, pp. 42-60.
62Appendix 1: Level 1 Block Diagram of Analyzer-M Figure 43. Analyzer-M Level 1Block Diagram A Figure 44. Analyzer-M Level 1Block Diagram B
63Appendix 2: Level 2 Block Diagram of Analyzer-M Figure 45. Analyzer-M Level 2 Block Diagram A Figure 46. Analyzer-M Level 2 Block Diagram B
64Appendix 3: Level 3 Subdiagram 0 of Analyzer-M Figure 47. Analyzer-M Level 3 Subdiagram 0 Block Diagram A Figure 48. Analyzer-M Level 3 Subdiagram 0 Block Diagram B
65Appendix 4: Level 3 Subdiagram 1 of Analyzer-M Figure 49. Analyzer-M Level 3 Subdiagram 1 Block Diagram A Figure 50. Analyzer-M Level 3 Subdiagram 1 Block Diagram B
66Appendix 4:(Continued) Figure 51. Analyzer-M Level 3 Subdiagram 1 Block Diagram C
67Appendix 5: WavePlayer-Mono block Diagram Figure 52. WavePlayer-Mono block Diagram A Figure 53. WavePlayer-Mono block Diagram B
68Appendix 6: Analyzer-M Instructions  INSTRUCTIONS FOR INSTALLATI ON / UN-INSTALLATION OF ANALYZERM.EXE Install: 1. If Windows Explorer is not already configured to show file extensions, configure your Windows Explorer to always show file extensions. 2. Install LabVIEW Run-time Engine 7.1 if LabVIEW 7.1 or higher is not installed in the computer. 3. Go to M-remote folder in the Installation CD. 4. Navigate to the Installer folder inside M-remote folder. 5. Click on install.msi and follow the instructions in the WIZARD. i. During the installation proce ss, the setup will display the location where the program will be installed. The Destination Folder item should display EXACTLY c:\ANALYZER-M (and NOTHING ELSE) as the path for the destination of the program. If it does not, navigate to the folder c:\ANALYZER-M using the Browse button located on the right hand side of the setup window. 6. After installation is complete, locate the file ANALYZER-M.EXE (it should be located in the folder c:\progr ams\ANALYZER-M or go to the start menu and choose programs) and create a shortcut for that file on the desktop. 7. A simplify way to install the progra m would be to copy the file ANALYZERM.EXE from the M-remote fol der and paste on the desktop. Un-Install: 1. In the Installation CD, navigate to M-remote\Installer folder, click on install.msi 2. Follow the instructions in the wizard. Imp. Note: In the un-installation process, the wizard should indicate that the process is indeed un-installation, not an installation. Run-Time Engine installation From the installation CD run LVRunTimeE ng 7.1.exe to instal l the LabVIEW 7.1 Run-Time Engine.
69Appendix 6:(Continued) TENDON VIBRATIONAL TESTING STEP-BY-STEP PROCEDURE Prepared by University of South Florida All Rights Reserved A. EQUIPMENT AND SETUP FOR ANALYZER-M Minimum items required: 1. Laptop DELL LATITUDE 840 computer with Analyzer-M software 2. Memory stick with at least 1GB of empty space for drag and drop operation. 3. 110 V 60Hz AC Power Sour ce adequate for computer 4. Thermometer to record ambient temperature 5. Long (100 ft) and accurate measuring tape 6. Log Form binder/clipboard/ballpoint pens. 7. Accelerometer Kit Box containing 7a. Accelerometer (PCB 338B34) 7b. Accelerometer Extension Cable 7c. Sensor Signal Condit ioner (ICP Model 480E09) 7d. Stereo Adapter Cable 7e. Spare 9V Alkaline Batteries (bag of 3) 8. BNC Black Extension Cable (2 50 -ft sections with 2 Female-Female couplers) 9. Hammer 10. Tuning Fork 11. Wireless presentation remo te (Targus, model PAUM30). 12. Card Table and Stool essential for accurate work. 13. Adequate lighting Note: Items 1,7 and 8 must be on site in duplicate to provide full spare backup. Physical arrangement (see Figure 1): a) Set up Card Table centrally in t he Test Station area chosen, so the computer screen is within easy view from the impact position b) Set up power source outlet next to Card Table. c) Ensure that accelerometer wiring can run unobstructed to each of the accelerometer locations in the Test Station. Ensure that operator movement does not result in tripping over wires or equipment falling down. d) Place DELL Computer and Log Form on Card Table. Leave space also for tuning fork. A comfortabl e working space is essential for reliable operation and record keeping.
70Appendix 6:(Continued) e) Make sure that cooling fan in computer is not obstructed Equipment is now ready for operation. Proceed To Part B, System Startup and Pre-Test Steps. B. SYSTEM STARTUP AND PRE-TEST STEPS READ EACH STEP COMPLETELY BEFORE ACTING Step 0. Span and tendon segment ID and preparatory measurements See ADDENDUM 1 for ID proc edures and preparatory work. Step 1. Set up computer: 1.1 Write down Test Station number (for example 07) on Log Form. 1.2 Power up and boot up computer. Record on Log Form designation of computer being used. 1.3 Perform audio input setup check: Perform once at beginning of shift. Perform also if ma chine was operated by others during shift or if abnorma l test results are observed. On desktop, double click Volume Control icon Ensure Mute All is selected. Ensure Line In Volume is all the way up and not muted. Click Options, Properties. Click Recording, then OK. Ensure Stereo Mix Volume is all the way up. Ensure Stereo Mix Select is clicked. Ensure nothing else is selected. Close window. 1.4 Double-click ANALYZER-M icon on desktop. After Logo appears, press F2 and choose OK or Cancel conditions. If OK is chosen, the operating panel shows up on screen. Turn Caps Lock on. Step 2. Wire accelerometer, Sensor Signal Conditioner, and connection to computer. Check/replace Sensor Signal Conditioner batteries:
71Appendix 6:(Continued) Remove from Accelerometer Kit Box #1 the Accelerometer. [Use parts in spare Kit Box #2 only if parts in #1 fail]. Record on Log Form Serial Number (SN) of accelerometer being used. The Accele rometer is a precision instrument. Handle it gently and do not drop Connect Accelerometer to white A ccelerometer Extension Cable Turn floating clamp nut, never the acce lerometer as that may damage connector pin. Do not kink or stress Accelerometer Extension Cable. Connect Accelerometer Extension Cable to BNC Black Extension Cable with provided adaptor. If only one 50-ft l ength is sufficient, store away the other 50-ft length. Do not use cables to pull on or hold equipment! 2.2 Connect other end of BNC Black Extension Cable to XDCR jack on Sensor Signal Conditioner. Verify that Sensor Signal Conditioner controls are as follows: Gain: 10 Red Rocker: Press ri ght side (ON) and let go. 2.3 Connect Stereo Adapter Cable to SCOPE on Sensor Sign al Conditioner #1. 2.4 Connect other end of Stereo Adapter Cable to LINE INPUT of computer. MAKE SURE NOT TO USE THE MICROPHONE INPUT For easier identification, the LINE INPUT has been marked by a white ring. 2.5 Check batteries by momentarily pre ssing right side Red Rocker in Sensor Signal Conditioner all the way down. Meter should point to the BATT OK region. If it doesnt, replace all three batteries (open box by loosening Phillips-head screw on back). 2.6 After verifying the batteries are OK, check that needle in Sensor Signal Conditioner is in green regi on. If it isnt, check cables, accelerometer and connectors and reconnect/replace until condition is remedied. Step 3. Place accelerometer on Tuning Fork 3: Attach accelerometer with wax se curely and precisely between scribe marks on Tuning Fork 3. Route Acce lerometer Extension Cable loosely
72Appendix 6:(Continued) through the back of the tuning fork so cable does not touch vibrating beams or interferes with acce lerometer motion. Step 4. Run Tendon Test program se ction and acquire Tuning Fork vibration data. Examine response: 4.1 Press (CTRL+R). Press F3. A File Save window appears on screen. Type XXT (where XX is the Test Station designation, for example 07 and the letter T indicates Tuning Fork te st). DO NOT ENTER ANY FILE EXTENSIONS OR ALTER THIS PROCEDURE IN ANY WAY. Press ENTER. Note: this does not cause a file to be saved yet. It only prepares the system to save the result of the test under the file name chosen, if the test is completed successfully. 4.2 Gently place 1/2 inch dowel cro sswise just inside Tuning Fork end until dowel touches stop screw. 4.3 Press F3 or use the wireless re mote to press the PUSH BOTTON display on the computer screen (TEST IN PROGRESS appears on screen). Immediately start counting: one-thous and-one, one-thousand-two, so as to have a 2-second wait. Immediately follo wing, pull dowel straight out (along main axis of tuning fork) in one quick motion and without introducing torque. The data acquisition st ops automatically 12 seconds after pressing F3. Do not touch anything while the TEST IN PROGRESS message shows. A short while later the test output will appear on the screen. 4.4 a) The frequency plot should show a clear peak at about 33 Hz (electric noise may also cause another peak near 60 Hz; ignore it). b) Press TAB repeating as needed to select RANGE box. Pressing the UP or DOWN keys causes the spectrum to zoom into a 10 Hz wide window that shifts in 5 Hz steps (w indow cycles to full width after multiple steps). Move the zoom until it incl udes the peak near 33 Hz. Read the peak frequency as shown in the Peak box. Peak frequency for Tuning Fork 3 should be a value from 33.8 Hz to 34.0 Hz. c) Read the peak height, as indicated in the vertical axis Height for Tuning Fork should be between 200 to 500 units.
73Appendix 6:(Continued) Note 1: The Peak box always shows the frequency of the highest point in the window. Read the box only when the near-33 Hz peak is the only one showing in the 10 Hz wide window. Note 2: Since the windows move only in short steps, the peak near 33 Hz will appear in two consecutive windows. The Peak reading in those two windows may differ by 0.1 Hz. That small variation is normal; in such case record only one of those values. Step 5. If Tuning Fork response is adeq uate, save data and proceed to Step 7: If response is as indicated in each of 4.4 (a), (b) and (c), Press F4, press ENTER (to select Yes), and the machi ne saves the results of the Tuning Fork test under the file name sele cted earlier. A confirmation message appears (see note after Step 16). The system is now ready for testing the tendons. Write down Peak frequency and Height in Log Form Supplementary Information section and proceed to Step 7. Otherwise, proceed to Step 6. Step 6. If Tuning Fork response is inadequate, check all settings and connections, and proceed to Step 4: If response fails to result in any one of 4.4 (a), (b) or (c), the test response is inadequate, indicating a problem Press F4 and then TAB (to select No) and then ENTER. This resets the program. Check everything (including that accelerometer is firmly attached to tuning fork, connections, switch positions, etc.) and repeat test starting at Step 4. Step 7. Record ambient temper ature and Span test start time: Enter operator initials, temperat ure and date/time information on Log Form. Always use ball-point pen. If there are any entry errors, strike over and write correct entry on the side do not attempt to write over old entry. Testing system is now ready to operate. Proceed to Part C.
74Appendix 6:(Continued) C. TEST EXECUTION STEPS Step 8. Select tendon to test, measure length, mark positions for accelerometer and hammer impact, and place accelerometer: 8.1. Tendon segment designations segment length measurement and accelerometer placement position. 8.2 Select tendon segments to be test ed in the order indi cated in the Log Form for the appropriate Test Stat ion. Special procedures for tendon segments obstructed from free vibrati on will be provided in a separate document. 8.3 Attaching accelerometer to tendon: Figure 54. Accelerometer Orientation The accelerometer is to be placed, with its axis on a plane approximately 45o from vertical, on the plastic duct as shown in Figure 2. Use mounting wax, cleaning any dust first. Av oid dropping accelerometer. If accelerometer is dropped make a note of it on Log Fo rm Supplementary Information section. Loose accelerometers are major source of rejected data. If necessary, further secure the accelerometer to the duct using adhesive tape or a Velcro strap. Tendon Vertical ~45 o A cceleromete r Accelerometer Axis
75Appendix 6:(Continued) 10.1 Route Accelerometer Extension Ca ble so that it does not rattle against tendon during vibration. Cable may be lightly wrapped around the next tendon to avoid accidental yanking and to restrain accelerometer fall. 8.4 Hammer impact is to be always appli ed on a direction perpendicular to the axis of the tendon, at a point approximately halfway between accelerometer and the deviation saddle end of the tendon. Impact will be applied in two manners: Straight and Side. In Straight impact the direction of the blow is contained in a vertical plane. In Side impact the direction of the blow is in a horizontal plane. Step 9. Run Hammer Practice program section and practice to deliver adequate hammer blow strength. SKIP STEP IF ALREADY TRAINED: 9.1 Ensure that Steps 1 through 8 are completed. 9.2 Using the designated hammer, impact (Straight) the tendon. See ADDENDUM 2 for important Notice and Disclaimer. For this operation screen must be within easy view from impact position. Hammer hitting: Adjust impact to obtain desired amp litude as detailed in instructions below. If duct is not tightly filled with grou t at impact point (as indicated by unusual sound), change impac t position to a point a few inches to the right or left of initial position) 9.3 Press (CTRL+R). Press F1 and hit tendon repeatedly waiting about 3 seconds between hits. Watch signal di splay. Signal trace at impact should go well beyond inner lines but should not cross the outer lines. Train yourself to adjust Straight impac t strength until signal stays within limits. With display still running, swit ch impact direction to Side and train for it similarly. Press F2 when operation within limits is achieved in both directions and stop hitting tendon. The Save Wave File window wi ll appear; do not attempt to close it. Training is complete. Wait about 20 seconds before next action.
76Appendix 6:(Continued) Step 10. Run Tendon Test program section and acquire vibration data: Ensure that Steps 1 through 8 ar e completed and that operator has already been trained to deliver adequat e impact strength. Each tendon is tested 2 times with Straight impact (tests 1 and 2) and 2 times with Side impact (tests 3 and 4). 10.2 Press F3 if the Save Wave File window is not already on screen. Type the file name (ALL CAPITALS) for the tendon segment to be tested and press ENTER. The file name is the same as tendon segment designation (see ADDENDUM 1) but with the number 1 appended for the first test performed for the tendon segmen t, 2 for the second test, etc. Example: File for 1st test on tendon segment 116209NE is named 116209NE1 File for 3rd test is nam ed 116209NE3. If a test is a repeat of a test that was not acceptable (due to implementation of Step 11), repeat the same file name used in the fa iled test (the failed test file will be written over). DO NOT ENTER ANY FILE EXTENSIONS OR ALTER THIS PROCEDURE IN ANY WAY. Note: this does not cause a file to be saved yet. It only prepares the system to save the result of the test under the file name chosen, if the test is completed successfully. Have operator standby with hammer ready to hit (Straight for tests 1 and 2, Side for tests 3 and 4) when directed. 10.3 Press F3 or the PUSH BOTTON (TEST IN PROGRESS appears on screen). Immediately start coun ting: one-thousand-one, one-thousandtwo, so as to have a 2-second wait. I mmediately following, direct operator to hit tendon only once. The data ac quisition stops automatically 12 seconds after pressing F3. Do not touch the tendon, accelerometer or anything else in the equipment while the TEST IN PROGRESS message shows. About 20 seconds later the test output appears on the screen. Signal analysis by the computer is now complete. Step 11.Check to see If data are adequate or not:
77Appendix 6:(Continued) a) Examine strip record of top channel. Is signal in top chart within limits described in step 9.3 ? b) Examine the spectrum record in t he lower chart. A dist inct peak should appear near the left end of the chart. That is the Mode 1 peak An overtone peak (Mode 2) should be visi ble at about twice the frequency of Mode 1. Higher overt ones may be visible at about three or four times the frequency of Mode 1. Also, the line between the peaks should be relatively smooth with few jagged regions. The signal in the top chart should be relatively symmetric and showing a gentle decay. See Figure 3 for ex amples of good and bad signals and spectra. Do signal and spectrum have the good' appearance shown in Figure 3? c) If answers to both (a) and (b) are YES, go to Step 12. d) If answer to (a) is NO, too low or too high impact has been applied. Press F4 and then TAB (to select No) and then ENTER. Check equipment and go to Step 9 for hammer practice. e) If answer to (a) is YES but answer to (b) is NO, test needs to be rerun. Bad signals and spectra are oft en due to a loose accelerometer, obstructions in the tendon motion, or abnormal hammer impact. Check for those problems as well as equipment and connections. Correct deficiencies. Press F4 and then TAB (to select No) and then ENTER. Press (CTRL+R). This resets the program. Go to Step 10 to repeat test. Note: If a bad spectrum or signal persists after a few tries, complete the tests anyway, make a note of the problem, and proceed to the next tendon. Repeat ed difficulties in subsequent tendons may indicate equipment malfunction. Step 12. Identify and record peak frequencies: 12.1 Press TAB and select RANGE box. Afterwards, pressing the UP or DOWN keys causes the spectrum to zoom into a 10 Hz wide window that shifts in 5 Hz steps. Shift the windo w until it includes the Mode 1 peak. Read the peak frequency as shown in the Peak box and enter in Log Form. Repeat for Mode 2 peak. Note 1: The Peak box alwa ys shows the frequency of the highest point in the window. Read the box only when the desired Mode is the
78Appendix 6:(Continued) main feature showing in the 10 Hz wide window. Some peaks may be split into two closely spaced peaks; record only the tallest. Note 2: Since the windows move only in short steps, each desired mode may appear in two consecutive windows. The Peak reading in those two windows may differ by 0.1 Hz. That sma ll variation is normal; in such case record only one of those values for the Mode. The approximate Mode 1 frequencies expected are listed in T able 1 (rough estimates actual behavior may be substantially different). Step 13. Save file: Press F4, press ENTER. The machine sa ves the results. A confirmation message appears (see note after Step 16). Press CTRL+R. Step 14. Conduct second, third or f ourth test of the tendon segment. (Steps 10 13): If the previous test was not the fou rth successful test for this tendon segment, repeat Steps (10) through ( 13). Otherwise, go to Step (15). Step 15. Proceed to next tendon segment in span starting at step (8): After the 4th successful test for this tendon segment is concluded, go to Step (8). Continue until all tendon segment s in the Test Station are tested. Then go to Step (16). Step 16. After the last segment in the span is tested, record temperature, copy data to CDRW drive, and prepare for next Span: Press left side of Red Rocker switch of Sensor Signal Conditioner to the OFF position. Exit Analyzer-M program by pressing (ALT+F4). Read temperature and record temperature and time in Log File. The files from all tendon te sts plus the Tuning Fork test file for this Test Station have been stored in the Folder named ANALYZER FILES in the C drive (folder ANALYZER FILES accumulates all the data from all the Test Stations). Copy all the files for th is Test Station to a folder named STATION## (where ## is the Test St ation designation) onto the formatted
79Appendix 6:(Continued) CDRW disk that is in the CD driv e. That CDRW disk has been formatted to act as if it were a hard drive. Copy into the same folder also any other files that may have originated for this Test Station (for example, from extra tests). Go through Windows Shut Down sequence and turn off computer. Preparations Procedure for identifying tendon segment s, measuring and recording lengths, and marking for accelerometer and impact location. 1. Tendon segment designations Use following order: Span number, dire ction, position along span, position across span, length of the sloping segment, creating an eight-character name FFFGGHIJ. The values that each of the characters can take are per Table 1. FFFGGHIJ Table 3. Eight-Character Tendon Segment Designation Span Designation Span Number FFF Direction GG Position Along Span H Position Across Span I Length of Sloping Segment J 088 to 105 or 117 to 134 SB: Southbound NB: Northbound S: South End N: North End W: West Side E: East Side L: Longest M: Medium S: Shortest
80Appendix 6:(Continued) 2. Measuring and recording length. a. For each tendon segment to be tested, clear any obstructions or debris that may prevent tendon from vibrati ng freely. Note: If the tendon is obstructed by unremovable obstacles (e.g, contact with walls or with other tendons), frets will need to be installe d. Procedures for fret placement and associated testing will be indica ted in a supplementary guide. b. On each tendon segment, measure and record in Log Form the clear concrete-to-concrete distance. Make a note of any unusual details such as uneven concrete surface. If available, use a metric tape and writ e result with 1 mm precision (if only English-units tape is available, writ e result in inches with 1/8 inch precision). If any other device [e.g. Las er unit] is used instead of tape, ensure first that device is accu rate by making independent tape and device measurements in at least 12 di fferent segments in actual field conditions. Send table of results fo r USF where statistical analysis of results will be conducted for verification. If using tape, ensure that an accurate, stretch-free tape is used. Do not pull on tape excessively. Replace kinked or damaged tape. Make sure that any folding tabs at end of tape are pr operly positioned. If concrete face is irregul ar, refer distance to main plane representing surface. 3. Marking accelerometer and impact positions a. Mark with tape or bright marker position where accelerometer is to be placed. See Figure A1 (if end point s are not against a bulkhead or a deviation saddle, measure distances from lowest point). Position is at distance LA from low end, where LA is ~1/3 of the tendon segment length. The value of this distance is not critical but once chosen it must be recorded. b. Mark with tape or bright marker position where impact is to be made. See Figure A1. Position is at distance LI from low end, where LI is ~1/6 of the tendon segment length. The value of th is distance is not critical but once chosen it must be recorded.
81Appendix 6:(Continued) Figure 55. Accelerometer Position SPREADSHEET FOR TENSION COMPUTATION. Tension.xls Spreadsheet Instructions 1 Open the Excel Workbook called Tension.xls. 2 In the Worksheet called Inputs enter the following data in their corresponding cells: -Bridges name. -Length L (meters) and Mode Frequencies f (Hz) for each of the tendon segments tested, from the Log Form. Number of strands Strands from bridge construction data. Mass mu per unit length (kg/meter), and stiffness S of the tendon (N-m2), from the Estimation Tables for each tendon segment of being analyzed. The Estimation Tables use as input the number of strands and the tendon diameter, the latter to be measured in situ for each tendon. 3 The calculated Tension per strand (k N/strand), in each of the bridge segments, appears on the Results Worksheet. BULKHEAD DEVIATION SADDLE TENDON SEGMENT LA ACCELEROMETER BULKHEAD DEVIATION SADDLE TENDON SEGMENT LA ACCELEROMETER BULKHEAD DEVIATION SADDLE TENDON SEGMENT LA ACCELEROMETER L I LI
82Appendix 7: ANALYZER-DAB Level 3 Block Diagram Figure 56. ANALYZER-DAB Level 3 Subdiagram 0-A Figure 57. ANALYZER-DAB Level 3 Subdiagram 0-B
83Appendix 7:(Continued) Figure 58. ANALYZER-DAB Level 3 Subdiagram 0-C Figure 59. ANALYZER-DAB Level 3 Subdiagram 1-A
84Appendix 7:(Continued) Figure 60. ANALYZER-DAB Level 3 Subdiagram 1-B Figure 61. ANALYZER-DAB Level 3 Subdiagram 1-C
85Appendix 8: ANALYZER-DAB Installation Guide INSTRUCTIONS FOR INSTALLATION / UN-INSTALLATION OF ANALYZERDAB.EXE Install: 1. If Windows Explorer is not already configured to show file extensions, configure your Windows Explorer to always show file extensions. 2. Install LabVIEW Run-time Engine 7.1 if LabVIEW 7.1 or higher is not installed in the computer. 3. Install NI-DAQmx Base Version 2.0 for Windows 2000/XP. 4. Go to DAQ-remote folder in the Installation CD. 5. Navigate to the Installer folder inside DAQ-remote folder. 6. Click on install.msi and follow the instructions in the WIZARD. During the installation process, the setup will display the location where the program will be installed. The Destination Folder item should display EXACTLY c:\ ANALYZER-DAQ (and NOTHING ELSE) as the path for the destinati on of the program. If it does not, navigate to the folder c:\ ANALYZER-DAQ using the Browse button located on the right hand si de of the setup window. 7. After installation is complete, locate the file ANALYZER-M.EXE (it should be located in the folder c:\programs\ ANALYZER-DAQ or go to the start menu and choose programs) and create a shortcut for that file on the desktop. 8. A simplify way to install the pr ogram would be to copy the file ANALYZERDAQ.EXE from the DAQ-remote folder and paste on the desktop. Un-Install: 3. In the Installation CD, navigate to DAQ-remote\Installer folder, click on install.msi 4. Follow the instructions in the wizard. Imp. Note: In the un-installation process, the wizard should indicate that the process is indeed un-installation, not an installation. Wave-Player DAQ.EXE installation Follow same instructions as for ANALYZER-DAQ.EXE but taking into consideration that the files shoul d be under wave-player DAQ folder. Run-Time Engine installation From the installation CD run LVRunTimeEng 7.1.exe to install the LabVIEW 7.1 Run-Time Engine
86Appendix 8:(Continued) NI-DAQmx Base Version 2.0 installation 1. Create a temporary folder on your local hard drive. 2. Extract the NIDAQmxBase200.zip file into the fol der created in Step 1. This will create the installation files necessary for installing NI-DAQmx Base. 3. To launch the NI-DAQmx Base installer, run setup.exe from the folder created in Step 1.
87Appendix 9: Block Diagram of WavePlayer-DAB Figure 62. WavePlayer-DAB Block Diagram A Figure 63. WavePlayer-DAB Block Diagram B
88Appendix 10: Wave Player Micro Block Diagram Figure 64. Wave Player Micro Block Diagram A Figure 65. Wave Player Micro Block Diagram B
89Appendix 11: Derivation of Equation to obtain "P" Figure 66. Schematic ER Probe Initial Conditions R P R I I R I P R I E E E Eo o2 ) 2 ( 0 ) 2 ( ) 2 ( 01 2 2 1 4 3 2 1+ = = + = + + + ) 2 ( 2 ) 2 4 ( ] 2 1 ) 2 ( 1 [1P R R P R R R R P R Ro o T o T+ + + = + + = ) 2 ( 2 ) 2 ( 1 1 ) 2 ( 2 ) 2 4 ( 2 ) 2 (1 1 1 1 2 1P R E I R P R P R R P R R E I R P R I I R E I I Io in o o o in o T in+ = + + + + + = + + = + = () 2 ) 2 ) 2 ( ( 1 ) 1 (1 1 1 2 1 1 4P R I E R R P R I P R I E R I P R I E E E Eo out o o out o out out = + + = + = = out in out o o in out o outE E E P P R P R E E P R I E 2 4 2 ) 2 ( 21 = + = = EinEout Rout(1+P) RinR1R2E4E3E1E2I1I2Rin=Rout=R0R1=R2=R E1+E2+E3+E4=0 I=I1+I2
90Appendix 12: Derivation of Equation to Obtain "Rcorr" A L R = where R is the resistance of the wi re in ohms, L is the length of the wire, is the resistivity of the steel, and A is the cross section area of the wire. 2r A =, where r is the radius of the wire. () () 2 2 corr o corr or L r L R R = where Ro is the original resistance of the wire, Rcorr is the Resistance of the wire after it corroded, ro is the original radius of the wire, and rcorr is the radius of the wire after it has corroded. Simplifying the above equation. corr o o corrR R r r = Since ) 1 ( P R Ro corr+ = ) 1 ( P R R r ro o o corr+ = ) 1 ( 1 P r ro corr+ =
91Appendix 13: Block Diagram P-Measurements Figure 67. Block Diagram P-Measurements A
92Appendix 13:(Continued) Figure 68. Block Diagram P-Measurements B
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001970345
007 cr mnu|||uuuuu
008 081125s2007 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002311
Practical vibration evaluation and early warning of damage in post-tensioned tendons
h [electronic resource] /
by Jaime Lopez-Sabando.
[Tampa, Fla.] :
b University of South Florida,
ABSTRACT: Severe corrosion damage and even complete failure was recently discovered in external post-tensioned (PT) tendons of three Florida's pre-cast, segmental bridges over seawater. A key deterioration factor was the formation of large bleed water grout voids at or near the anchorages. Steel corrosion may occur at the grout-void interface or in the air space of the void itself. Since the tendons are critical to the structural integrity of the bridges, reliable and non-intrusive damage detection methods are desirable to manage or prevent future occurrences. In recent years several indirect non-destructive methods have been developed or improved to evaluate the conditions of the tendons. One of those methods is vibration-based tension measurements, consisting of detecting tendon tension loss by analyzing the tendon's natural frequencies.^ Until recently, vibration-based tension measurements were costly and laborious since they required several operators to conduct the tests and complicated analysis through different programs. The first objective of this research is to provide a practical, simplified, user-friendly testing and analysis method for screening tendons by vibration measurements. Electrochemical Impedance Spectroscopy, Linear Polarization, and Electrical Resistance are alternative methods that could nondestructively detect or monitor corrosion before strand failures occur. The reliability and sensitivity of these conventional monitoring methods in solid or liquid media are well proven. However, few investigations exist on applying these methods to air-space corrosion as it may occur in tendon anchors. The second objective of this research is to establish the feasibility of using the above conventional monitoring methods for detecting air-space corrosion.^ In this investigation, two different types of Electrical Resistance probes were designed and evaluated. Also, electrochemical probes were constructed simulating strands conditions in the grout-void interface. Electrochemical Impedance Spectroscopy and Linear Polarization measurements were conducted in the electrochemical probes to calculate their instantaneous corrosion rates. Electrical Resistance and Electrochemical probes results indicate that both methods provide sufficient sensibility to determine the ongoing damage.
Thesis (M.S.M.E.)--University of South Florida, 2007.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 92 pages.
Adviser: Alberto A. Sags, Ph.D.
Non destructive methods.
Electrochemical Impedance Spectroscopy.
x Mechanical Engineering
t USF Electronic Theses and Dissertations.