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 2200385Ka 4500
controlfield tag 001 002029072
007 cr bnu|||uuuuu
008 090916s2009 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002871
Berke, Brandon Seth.
Long term corrosion of reinforcing strips in mechanically stabilized earth walls
h [electronic resource] /
by Brandon Seth Berke.
[Tampa, Fla] :
b University of South Florida,
Title from PDF of title page.
Document formatted into pages; contains X pages.
Thesis (M.S.C.E.)--University of South Florida, 2009.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
ABSTRACT: Mechanically stabilized earth (MSE) walls are a more advanced form of a retaining wall, often larger and able to hold back more backfill. This is achieved by reinforcing strips or meshes (most often galvanized steel) placed into the soil, which are held in place by friction. The strips mechanically stabilize the earth while undergoing tension. The wall is covered with concrete medallions that connect to the reinforcements. The medallions have only a secondary structural role in holding up the wall but provide cover that protects the soil from washing away. MSE walls are structures expected to have very long service lives (e.g. 100 years). Confirmation is needed that such durability can be achieved, especially to show that the progression of corrosion of the reinforcement is slow enough.Ten MSE walls around Florida were instrumented (electrical connections were made through the concrete covers to the buried elements) between 1996- 1998 and used to survey corrosion rates of galvanized strip or mesh soil reinforcements. Initial estimates of corrosion-related durability were obtained at that time, indicating a good prognosis for long term durability. The objective of the research in this thesis was to obtain additional indications of the durability of reinforcements in MSE walls in Florida so as to perform a more reliable projection of future performance. Corrosion behavior was measured at the same locations as the initial survey by electrochemical nondestructive tests and by destructive tests. The nondestructive testing consisted of half-cell potentials, polarization resistance measurements, and electrochemical impedance spectroscopy. Corrosion rates reported in this thesis are based upon polarization resistance measurements.The destructive testing consisted of soil extraction and hardware extraction. Hardware extraction enabled independent verification of estimates of electrochemical corrosion rate. Analysis of extracted soil verified that soil composition was within construction specifications. The data from the current survey were also used to further improve prediction of corrosion. The present series of evaluations confirm that the structures are performing as desired based upon the updated model projection of future corrosion.
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
Advisor: Alberto A. Sags
x Civil and Environmental Engineering
t USF Electronic Theses and Dissertations.
Long Term Corrosion of Reinforcing Strips in Mechanically Stabilized Earth Walls by Brandon Seth Berke A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Civil Engineering Department of Civil and Environmental Engineering College of Engineering University of South Florida Major Professor: Alberto A. SagÂŸÂŽs, Ph.D. Jeffrey Cunningham, Ph.D. A. Gray Mullins, Ph.D. Date of Approval: March 16, 2009 Keywords: Electrochemical Impedance, Buri ed Metal, Galvanized Steel, Polarization Resistance, Durability Copyright 2009, Brandon Seth Berke
Dedication To my father, Neal S. Berke, Ph.D., who inspired me to choose this career and set an example of excellence for me to follow.
Acknowledgements I am grateful that Dr. Alberto A. SagÂŸÂŽs is my advisor and advocate in the University. Additionally I thank Dr. Jeffrey Cunningham and Dr. A. Grey Mullins for serving on my committee with thei r helpful perspectives. Further I would like to extend gratitude to my fellow colleagues, Kingsley Lau, Luciano Taveira, and Keller Charles at USF who helped me with many aspects of my project. Lastly, I am indebted to Concorr Florida for helping me out with much of my fieldwork.
i TABLE OF CONTENTS LIST OF TABLES iv LIST OF FIGURES v i LIST OF SYMBOLS x i ABSTRACT xi i CHAPTER 1: INTRODUCTION AND OBJECTIVES 1 1.1 Introduction 1 1.2 Objectives and Approach 8 1.2.1 First Approach 8 1.2.2 Second App roach 8 1.2.3 Third Approach 8 1.2.4 Consequences 8 CHAPTER 2: EXPERIMENTAL 8 2.1 Field Sites and Instrumentation Details 9 2.1.1 Brickell Ave., Miami, BR Site 12 2.1.2 Howard Frankland Bridge, Tampa Bay, HFB Site 13 2.1.3 Pensacola Ave ., Tallahassee, PAV Site 14 2.1.4 Palm City Bridge, Stuart/Palm City, PC Site 14 2.1.5 Port St. Lucie Blvd., Port St. Lucie, PSL Site 15 2.1.6 State Rd. 200 Bridge, Ocala, OCA Site 15 2.1.7 Acosta Bridge, Jacksonville, ABJ Site 16 2.1.8 Veterans Expre ssway Overpass, Tampa, VET Site 16 2.2 Field Evaluation Procedure 16 2.2.1 Half cell Potential 17 2.2.2 Macrocell Current 17 2.2.3 Mutual Resistance 1 8 2.2.4 Solution Resistance 18 2.2.5 Linear Polarization Resistance 18 2.2.6 Electrochemical Im pedance Spectroscopy 21 2.2.7 Other Tests 22
2.3 Laboratory Evaluations 23 2.3.1 Soil Tests 23 2.3.2 Metallography 23 CHAPTER 3: RESULTS 2 4 3.1 Field Data 2 4 3.1.1 Visual Appearance of Wall and Extracted Reinforcement 2 4 3.1.2 Solution Resis tance, Linear Polarization, Apparent Corrosion Rates from LPR and Half cell Potential Values 2 6 3.1.3 Electrochemical Impedance Spectroscopy 30 3.1.4 Macrocell Current Values 3 0 3.2 Laboratory Data 3 1 3.2.1 Soil Tests Data 3 1 3.2.2 Metallography and Magnetic Gauge Measurements 3 2 CHAPTER 4: DISCUSSION 3 3 4.1 Direct Assessment Results 3 3 4.2 Electrochemical Estimates of Corrosion Wastage 3 5 4.3 Accuracy and Consistency of Electrochemical Corrosion Measurements 4 2 4.4 Predictive Model 47 CHAPTER 5: CONCLUSIONS 5 2 5 .1 First Conclusion 52 5.2 Second Conclusion 52 5.3 Third Conclusion 52 5.4 Fourth Conclusion 53 5.5 Fifth Conclusion 53 5.6 Sixth Conclusion 53
5.7 Seventh Conclusion 54 5.8 Eighth Conclusion 54 5.9 Final Conclusion 54 REFERENCES 5 5 A PPENDICES 5 8 Appendix 1: Site Instrumentation Diagrams 5 9 Appendix 2: Detailed ACR Data 10 2 Appendix 3: Metallography 11 8
iv LIST OF TABLES Table 1.1 AASHTO backfill guidelines for MSE walls  2 Table 1.2 AA SHTO corrosion rate guidelines for galvanized steel in MSE walls  2 Table 2.1 Structure details for each site 10 Table 2.2 Extracted samples 21 Table 3.1 Visual appearance of wall and extracted reinforcement 23 Table 3.2 Summary of electrochemic al field observations for instrumented structures 2 6 Table 3.3 Macrocell currents for the Brickell Ave. Bridge site from June 2008 3 1 Table 3.4 Averaged soil properties measured in the '94 '98 and '06 '09 surveys 3 1 Table 3.5 Summary of metallogr aphic and magnetic gauge thickness measurements ( m) 3 2 Table 4.1 Analysis of wall averaged ACR results 3 9 Table 4.2 Time dependence parameter n from sources reported in References  and  4 2 Table 4.3 Average galvanized steel half cell poten tials (V) vs. a CSE in contact with soil for the present and previous surveys 4 7 Table A1.1 Dimensions of test elements 100 Table A2.1 LPR data for BR 102 Table A2.2 LPR data for HFB 103
v Table A2.3 LPR data for PC 104 Table A2.4 LPR data for PSL 105 Table A2.5 LPR data for OCA 105 Table A2.6 LPR data for ABJ 106 Table A2.7 LPR data for VET 106 Table A2.8 EIS LPR data for BR 107 Table A2.9 EIS LPR data for HFB 108 Table A2.10 EIS LPR data for PC 109 Table A2.11 EIS LPR data for PSL 11 0 Table A2.12 EIS LPR data for OCA 110 Table A2.13 EIS LPR data for ABJ 111 Table A2.14 EIS LPR data for VET 111 Table A3.1 Detailed direct observation of metal coupons 118 Table A3.2 Listing of all mounted samples and respective measurement in formation 119
v LIST OF FIGURES Figure 1 Microstructure of galvanized steel from an unexposed archival MSE strip, and identification of layers and composition  3 Figure 2 Locations of MSE walls chosen for instrumentation throughout the stat e of Florida . 11 Figure 3 An LPR test using the Gamry TM potentiostat shows the machine configured with the added rebar as a working electrode, the activated titanium as the reference electrode, and the bottom mesh as the counter electrode. 19 Figure 4 Examples of metallographic cross sections showing low (A) and high corrosion wastage (B). 2 5 Figure 5 Comparison of Rp values obtained by LPR and EIS for the same galvanized steel elements, and during the same field visit for the indicat ed walls. 2 9 Figure 6 Comparison of Rp values obtained by LPR and EIS for the same plain steel elements, and during the same field visit for the indicated walls 30 Figure 7 Comparison of galvanized layer plus corrosion product thickness from metallogr aphic (MET Total Thickness) with total film thickness determined with a magnetic thickness gauge (MAG), averaged for coupons extracted in the present survey from each Wall. 34 Figure 8 Wall averaged galvanized steel ACR values for the present and previous surveys. 35 Figure 9 Cumulative distributions of ACR averaged for each wall from the 1994 98 and 2006 09 surveys, fitted with cumulative lognormal distributions. 37
vi Figure 10 Comparison of integrated corrosion rates from metallographic me asurements (MET) of coupons extracted from five walls in the present survey, averaged for coupons for coupons from each wall, with corresponding average ACR values from LPR measurements. 42 Figure 11 Analog circuit used to analyze the EIS data. 44 Figure 12 Model projections of percentage of damaged elements in a generic MSE wall as a function of wall age. 49 Figure 13 Site diagram of the Brickell Ave. Site. 58 Figure 14 Elevation view showing structure, wall, and location information for BRN. 59 Figure 15 Elevation view showing structure, wall, and location information for BRS. 59 Figure 16 Dimensions of typical concrete panels and tie strip locations for the BR site. 60 Figure 17 Reinforcement placement in BR site MSE walls. 61 Figure 18 Clu ster diagram showing layout of elements in BRN. 62 Figure 19 Cluster diagram showing layout of elements in BRS. 63 Figure 20 Test points BRN, bottom layer. 64 Figure 21 Test points BRN, top layer. 64 Figure 22 Test points BRS top (A) and bottom (B) lay ers. 65 Figure 23 Elevation view of HFB MSE wall showing panel nomenclature . 66 Figure 24 Dimensions of typical concrete panels and tie strip locations for the HFB site. 67 Figure 25 Top view of a mesh and panel at HFB. 68 Figure 26 Panel layout for HFB panels R7, R9, R15, and R21. 69 Figure 27 Panel layout for HFB panel R11. 70
vii Figure 28 Panel layout for HFB panel R17. 71 Figure 29 Core hole locations in HFB panel R9. 72 Figure 30 Core hole locations in HFB panel R21. 73 Figure 31 Plan vi ew of PCE showing panels with instrument clusters. Dimensions are not to scale . 74 Figure 32 Dimensions of typical concrete panels and tie strip locations for PCE . 75 Figure 33 Panel layout for panel R1 at PCE. 76 Figure 34 Panel layout for p anel R5 at PCE. 77 Figure 35 Panel layout for panel R14 at PCE. 78 Figure 36 Panel layout for panel R28 at the Stuart NE site. 79 Figure 37 Diagram indicating the location of the panels R3W and R5W PCW . 80 Figure 38 Detailed panel information for panel 3W at PCW. 81 Figure 39 Detailed panel information for panel 5W at PCW. 82 Figure 40 Location of additional core holes at PCW. 83 Figure 41 View of relative panel locations at PSL revised from the previous report . 84 Figure 42 Detailed panel information for panel 3 at PSL. 85 Figure 43 Detailed panel information for panel 7 at PSL. 86 Figure 44 Additional core hole at PSL panel R4, from which a soil sample was extracted. 87 Figure 45 Additional core hole at PSL panel R8 87 Figu re 46 Diagram indicating the location of the panels R6 and R25 at OCA . 88
viii Figure 47 Panel layout for Panel 6 at OCA. 89 Figure 48 Panel layout for Panel 25 at OCA. 90 Figure 49 Approximate location of an additional core hole at OCA panel R5. 9 1 Figure 50 Approximate location of an additional core hole at OCA panel R24. 92 Figure 51 Diagram indicating the location of the panels R9 and R21 of ABJ. 93 Figure 52 Panel layout for Panel 9 at ABJ. 94 Figure 53 Panel layout for Panel 21 at ABJ. 95 Figure 54 Core hole at ABJ panel R20. 96 Figure 55 Diagram indicating the location of the panels R16 and R23 of VET . 97 Figure 56 Panel layout for Panel 16 at VET. 98 Figure 57 Panel layout for Panel 23 at VET. 99 Figure 58 Cumulati ve distribution of ACR from BRN and BRS. 112 Figure 59 Cumulative distribution of ACR from HFB. 112 Figure 60 Cumulative distribution of ACR from PCE. 113 Figure 61 Cumulative distribution of ACR from PCW. 113 Figure 62 Cumulative distribution of ACR from PSL. 114 Figure 63 Cumulative distribution of ACR from OCA. 114 Figure 64 Cumulative distribution of ACR from ABJ. 115 Figure 65 Cumulative distribution of ACR from VET. 115 Figure 66 Cumulative distribution of ACR from all plain steel elements i n all walls grouped by years since insertion. 116
ix Figure 67 Plot of percentage of red rust observed by age by site. 120 Figure 68 Cumulative lognormal distribution of coating thickness measurements from coupons collected from the field. 120 Figure 69 Th e top side of the coupon from PCW panel 2W. 121 Figure 70 The top side of the coupon from PCW panel 4W. 121 Figure 71 The top side of the coupon from PSL panel 8. 122 Figure 72 The top side of coupon A' from OCA panel 5. 122 Figure 73 The top side of coupon B' from OCA panel 5. 123 Figure 74 The top side of the coupon from OCA panel 24. 123 Figure 75 The top side of the coupon from ABJ panel 20. 124 Figure 76 The side views of the coupon from HFB panel 9 top mesh. 124 Figure 77 The bottom side of the coupon from HFB panel 9 top mesh. 124 Figure 78 The side views of the coupon from HFB panel 22 top mesh. 125 Figure 79 The top (A) and bottom (B) sides of part of the coupon from HFB panel 22 top mesh. 126 Figure 80 The top (A), side (B), and botto m (C) sides of part of the coupon from HFB panel 22 bottom mesh. 127 Figure 81 Pictures of segments cut from a hook in HFB panel 9 in the top layer mesh. 128
x LIST OF SYMBOLS A Area AC Alternating Current ACR Apparent Corrosion Rate CPE Constant Phase Angle Element CSE Copper/Copper Sulfate Electrode DC Direct Current EIS Electrochemical Impedance Spectroscopy LPR Linear Polarization Resistance Rp Polarization Resistance Rs Solution Resistance S Siemens Delta Layer V Pot ential Difference # Gamma Layer $ Zeta Layer % Eta Layer & Resistivity Ohm
xi Long Term Corrosion of Reinforcing Strips in Mechanically Stabilized Earth Walls Brandon Seth Berke ABSTRACT Mechanically stabilized earth (MSE) walls are a more adva nced form of a retaining wall, often larger and able to hold back more backfill. This is achieved by reinforcing strips or meshes (most often galvanized steel) placed into the soil, which are held in place by friction. The strips mechanically stabilize t he earth while undergoing tension. The wall is covered with concrete medallions that connect to the reinforcements. The medallions have only a secondary structural role in holding up the wall but provide cover that protects the soil from washing away. MS E walls are structures expected to have very long service lives (e.g. 100 years). Confirmation is needed that such durability can be achieved, especially to show that the progression of corrosion of the reinforcement is slow enough. Ten MSE walls around Florida were instrumented (electrical connections were made through the concrete covers to the buried elements) between 1996 1998 and used to survey corrosion rates of galvanized strip or mesh soil reinforcements. Initial estimates of corrosion related du rability were obtained at that time, indicating a good prognosis for long term durability.
xii The objective of the research in this thesis was to obtain additional indications of the durability of reinforcements in MSE walls in Florida so as to perform a m ore reliable projection of future performance. Corrosion behavior was measured at the same locations as the initial survey by electrochemical nondestructive tests and by destructive tests. The nondestructive testing consisted of half cell potentials, pol arization resistance measurements, and electrochemical impedance spectroscopy. Corrosion rates reported in this thesis are based upon polarization resistance measurements. The destructive testing consisted of soil extraction and hardware extraction. Har dware extraction enabled independent verification of estimates of electrochemical corrosion rate. Analysis of extracted soil verified that soil composition was within construction specifications. The data from the current survey were also used to further improve prediction of corrosion. The present series of evaluations confirm that the structures are performing as desired based upon the updated model projection of future corrosion.
1 CHAPTER 1: INTRODUCTION AND OB JECTIVES 1.1 Introduction As widespread use of mechanically stabilized earth (MSE) walls has grown to greater than 40,000 across the United States since 1971 [1, 2], investigations were conducted to determine durability of the structures. Major durabi lity studies for MSE walls are currently ongoing nationally by the Federal Highway Administration, and individual state studies in New York, Kentucky, Georgia, North Carolina, South Dakota, California, and Florida. MSE is an old technology that became fur ther developed in the past century. The process was an ancient practice in China in which branches were inserted into dirt mounds as a strengthening method. The practice was found to be used in 18 th century France as well. Currently the process is execu ted by taking metallic strips or meshes anchored by concrete medallions. Layers of backfill soil are compacted while placing layers of reinforcement like a sandwich. The concrete medallions are typically made out of tessellating patterns . Corrosion in an MSE wall happens due to the oxidizing environment in soils. Various groups have evaluated buried steel and galvanized steel to gather empirical data of corrosion wastage in various soil environments [4,5]. _______________________________________ P arts of the work in the following chapters have appeared in B. Berke and A. SagÂŸÂŽs, "Update on Condition of Reinforced Earthwall Straps", Project No. BD544 32, 97 pages, Draft Final Report to Florida Department of Transportation, University of South Flori da, Tampa, FL, February 2, 2009.
2 Corrosion of zinc is of particular interest since galvanizing the reinforcement strips is a standard method for corrosion protection of MSE wall reinforcements. Wastage information gathered from Stuttgart University's ana lysis of National Bureau of Standards (NBS) data linear wastage approximations, and backfill material in MSE walls (Table 1.1) was used to make design guidelines . American Association of State Highway and Transportation Officials (AASHTO) created desi gn standards based upon the aforementioned data regarding what type of corrosion wastage rat es to expect in MSE structures (Table 1.2). Note that Table 1.1 only applies to soils with a resistivity of less than 5,000 cm [4,5]. Table 1.1 AASHTO backfill guidelines for MSE walls.  Parameter Limit or Range pH 5 10 Chlorides <100 ppm Sulfates <200 ppm Resistivity >3,000 cm Organic Content <1% Table 1.2 AASHTO corrosion rate guidelines for galvanized steel in MSE walls.  Material Layer Age (years) Corrosion Rate ( m/y) 0 2 15 Zinc 2 time of depletion 4 zinc depletion 75 years 12 Base Steel >75 years after zinc depletion 7 Galvanizing provides corrosion resistance first by the intrinsically low corrosion rate of zinc in most natural soil environments. When the zinc layer wastage eventually exposes some of the base steel, it is protected from rapid corrosion by galvanic coupling with the remaining zinc layer. The latter has
3 typically a highly negative corrosion potential that polarizes the steel towards the immune regime. Galvanizing of MSE strips is achieved by hot dippi ng steel into a molten zinc bath, mutually fusing the zinc and iron and creating a series of intermetallics there as noted in Figure 1.  In MSE reinforcements, the hot dipping is rather robust (strong in adhesion, and toughness) as coatings are on the o rder of 100 m thick.  Figure 1 Microstructure of galvanized steel from an unexposed archival MSE strip, and identificat ion of layers and compositions  In initial investigations during the 1980s, the Federal Highway Administration (FHWA) eva luated the durability of MSE walls by extracting strips from field sites and measuring metal loss by comparison with assumed initial
4 dimensions and mechanical tests of yield strength of strips . Currently nondestructive testing (NDT) is often performed involving at least half cell potential measurements typically with a Cu/CuSO 4 electrode (CSE). Linear polarization resistance (LPR) is also frequently measured to evaluate quantitative corrosion rates. LPR measurements yield polarization resistances, wh ich are inversely proportional to corrosion currents and thus apparent corrosion rates (ACR) . Corrosion performance of MSE reinforcement has been the subject of various investigations, highlights of which are noted in the following. A recent FHWA st udy reviewed some national and international practices that lead to severe corrosion and in some instances failures of the walls . Failures were found to be mainly the result of corrosive agents in the backfill materials. Aggressive backfill condition s included high chloride concentrations in the soil (~5000 ppm), low soil pH (less than pH 5), and high concentrations of organic compounds. One notable failure resulted from an accident in which a tanker in Spain crashed into a wall, spilling corrosive c hemicals into the backfill . The New York State Department of Transportation (NYSDOT) established a yearly monitoring reporting program for their MSE walls, starting in 1999 and still in practice. NYSDOT co developed computerized equipment to measure LPR of the metal reinforcements in MSE walls [7,8]. The Kentucky Transportation Center (KTC) only instrumented 4 of the 129 MSE walls in the State in 2003. KTC also inserted corrosion coupons and found the galvanizing of the coupons to still be present a fter two years. KTC's
5 evaluation on backfill from Kentucky MSE walls led to an addition to the AASHTO standard, requiring organic backfill content to be less than 1%. The KTC created a statewide database of all of the MSE walls in Kentucky . The Georg ia Department of Transportation (GDOT) tested 13 walls of which three experienced high corrosion rates. In one location, aggressive conditions resulted when contaminated water from a nearby polluted creek and clay clumping in the backfill . The most notable failure, in which strips were corroded through in places, was observed in a wall built with reinforcements made of an aluminum magnesium alloy. The aluminum alloy failed to passivate in the soil, which was inundated with chlorides and iron, and th us corroded much faster than expected . The North Carolina Department of Transportation (NCDOT) was cited nationally in 2006 by Gladstone et al.  for its good practice in monitoring North Carolina's MSE walls. NCDOT inserted coupons of galvanized and plain steel, which were instrumented for making half cell measurements. This enabled an NCDOT MSE wall inspector to find out when the zinc became depleted on the coupons and presumably the MSE wall strips . The NCDOT found corrosion rates to be v ery low (average of 1.3 m/y) in the five MSE walls. Installation of more monitoring stations at MSE walls throughout North Carolina was considered . The South Dakota Department of Transportation (SDDOT) evaluated an MSE wall while it was being replac ed. This replacement enabled visual inspection of the conditions of the mesh reinforcement grids in the wall. Though
6 deformation due to settling was observed, severe corrosion didn't occur because the backfill didn't allow strong deicing chemicals to pen etrate . Additionally, SDDOT inserted 1 m reinforcement strips of different types (galvanized, epoxy coated, and black steel). Evaluations showed that in areas of elevated sulfate concentrations and lower soil resistivities severe corrosion cover ed the plain steel strips; zinc reaction products were observed on the galvanized strips, and the epoxy remained intact. The study demonstrated that epoxy coating reinforcing strips was a good protection method against aggressive environments whenever bac kfill conditions can't be controlled . The California Department of Transportation (Caltrans) makes evaluations differently than the other states, in that direct examination is emphasized. The method involves pulling out entire reinforcement strips a nd evaluating these elements. Evaluations of the strips entail measuring amounts of pitting observed and residual tensile strength of the strips. Caltrans experienced problems with corrosion at some locations chiefly because AASHTO standards for backfill were not followed . Additionally, in some of these sites, the reinforcement strips were not galvanized . Furthermore there were suspicions that the water used to stop the dusting during the construction was contaminated with corrosion inducing ch emicals . A Florida Department of Transportation (FDOT) investigation completed in 1998  determined from testing conducted from 1994 to 1998 that corrosion rates were, in all structures examined, as low as expected for soils meeting AASHTO specifi cations. Only minor deterioration was observed at one location
7 with partial chloride contamination. The results were used as baseline data to formulate a quantitative durability model [14, 15] that projected that, in the absence of disrupting events, corr osion performance predicted a period of ~ 50 years with negligible reinforcement failure, and only ~5% failure after 100 years. However, laboratory experiments indicated that severe contamination, as may occur during a hurricane induced saltwater flood, could dramatically reduce corrosion related service life. For a wall with a saltwater flood at year zero, the model projected failure development 10 times earlier than without flooding. During the 1994 98 investigation 10 MSE structures were instrument ed at 8 different Florida sites for corrosion measurements; soil and metal samples were retrieved from several of the sites to evaluate the electrochemical properties of the backfill and to assess the condition of the galvanized coating after several years of exposure. The test location connections remained in place for future nondestructive monitoring. Most of those structures were relatively young at the time of testing but are still in service and have now accumulated another decade of service. Theref ore the present age is a significant fraction of a typical (e.g. 75 year) design service life. Assessment of present condition, together with the detailed information available for the same structures one decade earlier, can provide a highly useful indicat ion of corrosion related aging of MSE walls in FDOT service. That information can then be used to improve the accuracy of the durability prediction model to benefit future design and maintenance planning for these structures. The present investigation wa s conducted accordingly.
8 1.2 Objectives and Approach The objectives of this investigation are to extend the baseline of FDOT MSE corrosion performance measurements to reveal long term trends, and to improve the durability projections by using updated mo deling input. 1.2.1 First Approach A ssess the present condition of existing sites evaluated in the 1994 98 surveys by nondestructive measurements and by extracting soil and reinforcement samples. 1.2.2 Second Approach Evaluate field samples in the laboratory including experiments with simulated systems for comparison as needed. 1.2.3 Third Approach Operate and expand as needed forecasting models to predict future evolution of corrosion damage in existing and future FDOT MSE sites. 1.2.4 Consequ ences The activities and findings conducted toward achieving the objectives are detailed in the following sections.
9 CHAPTER 2: METHODOLOGY 2.1 Field Sites and Instrumentation Details The following terminology applies to these d escriptions: Site: An overall locale (e.g. Howard Frankland Bridge) Structure: One or more structural components associated with the Site (e.g. Tampa end causeway o f the Howard Frankland Bridge) Wall: One of the MSE walls in the Structure Locatio n: A place at the Wall where one or more test clusters have been implemented Test Cluster: A group of neighboring buried metallic components that have been instrumented for testing and/or exposed and sampled for direct metallic component and/or soil asse ssment. A cluster may include carbon steel rods embedded in the soil at the time of an earlier field visit. Metallic components in a cluster may be all associated with a single wall concrete panel (sometimes referred to as a medallion) or involve componen ts of two medallions i mmediately above each other Test Points: Permanent external electric contacts to instrumented metallic components and openings for reference electrode placement In the past survey from 1994 98, ten walls were instrumented to allow for electrochemical measurements on the buried elements in the MSE walls [14,15]. The sites/structures were chosen to represent the diversity of MSE walls found across Florida. The site list and rationale for each site is compiled in Table 2.1 and geog raphic locations in Florida are shown in Figure 2. The same sites as available were revisited in the present study.
10 Table 2.1 Structure details for each site. Site # and Code Structure and Wall Regime and Rationale for Testing Year Built Age (Years)Â # o f Test Clusters* 1A BRN Brickell Ave. Bridge NW Wall, Miami Coastal, Possible inundation 1995 10 2 1B BRS Brickell Ave. Bridge SE Wall, Miami Coastal, Possible inundation 1995 10 2 2 HFB Howard Frankland Bridge, Tampa Coastal, Possible inundation 1992 13 6 3 PAV Pensacola Ave., Tallahassee Land, oldest in FL 10 years ago 1979 N/A** 4 4A PCE Palm City Bridge NE Wall, Stuart Coastal, Possible inundation 1991 15 4 4B PCW Palm City Bridge NW Wall, Palm City Coastal, Tidal Saltwater Aggressiv e Regime 1991 15 2 5 PSL Port St. Lucie Blvd., Port St. Lucie Coastal, Tidal Saltwater Aggressive Regime 1992 14 2 6 OCA State Rd. 200, Ocala Land, Old, Long Term Baseline 1984 23 2 7 ABJ Acosta Bridge, Jacksonville Coastal, Non Spec. Backfill 199 0 17 2 8 VET Veteran's Expressway, Tampa Land, Representative of Present Practice 1995 12 2 *Set of reinforcements instrumented for electrical contacts Â Age of the structure when visited during the current survey. 4 Original clusters and two new ones from 2006 ** Demolished before second survey. Site codes are same as structure codes except for BR designating BRN/BRS and PC designating PCE/PCW.
11 Figure 2 Locations of MSE walls chosen for instrumentation thr oughout the state of Florida . Details of each site, structure, wall, location, test cluster and test points were presented by SagÂŸÂŽs et al . Those details are reproduced in Appendix 1 amended and updated to account for the actual condition of th e sites, new buried components, accurate test point cluster information, and the presence of any damaged test points. All geographic coordinates given for a site (or structure/wall if differentiated within a site) in the following are simplified to one s econd resolution. The coordinates correspond to a point located centrally to the instrumented wall locations. Additional directions and descriptions are given to facilitate accessing a site. Specific details on sites and subcategories including test po ints and reinforcing element surface areas are found also in Appendix 1. Details on how
12 permanent external connections to buried metallic elements were made are given by SagÂŸÂŽs et al. . During field visits alligator clips attached to the external con tact enabled interfacing the field equipment. 2.1.1 Brickell Ave., Miami, BR Site The BR site includes two structures on opposite sides (BRN (North) and BRS (South)) of the drawbridge that crosses the Miami River in downtown Miami. Test locations were implemented in walls at the Northwest and Southeast portions of the corresponding structures. Each location consisted of two test clusters, one near ground level and the other elevated. The BRN wall location and its test clusters and points are accessible through a gate on 64 SE 4 th St. for a car to drive to the site. There are City of Miami offices (with which it may be required to coordinate access operations) at the Knight Center, which is nearby on SE 4 th Street. The BRS wall location and cluster t est points were at the time of the visits adjacent to a construction site, yet a DOT access road exists which is the first left turn available southbound after crossing the bridge, extending to the South side of the bridge and running along the MSE wall. The coordinates of are 25Â¡46'13"N x 80Â¡11'25"W and 25Â¡46'10"N x 80Â¡11'23"W for the BRN and BRS wall locations respectively. The electric test points for each cluster are fitted in either one or two 4" diameter capped PVC ports.
13 2.1.2 Howard Frankland Bridge, Tampa Bay, HFB Site The HFB site has one Wall located on the Tampa end causeway of the Howard Frankland Bridge on southbound I 275. The Wall can be reached by a DOT service road, which emerges beyond the breakdown lane. The coordinates are 27Â¡5 6'25"N x 82Â¡33'10"W. There were 4 locations (each at one panel, R7, R11, R15 and R17) instrumented in the 1994 98 survey. In August 2006, two additional panels were instrumented (R9 and R21). The instrumentation of the two panels utilized similar metho ds to those used in the first survey, by attaching stainless steel rods to the galvanized meshes and other added electrodes. While instrumenting the new panels, coupons in the form of ~4" segments were cut out from the galvanized mesh wire for examination See Table 2.2 and Appendix 3 for additional information on extraction points and coupons. Furthermore, for each panel that didn't previously have a plain steel rebar (R7, R9 R15, and R21), a 2.4 m long No. 4 (4/8 in. 1.77 cm diameter) rebar was inserte d at positions indicated in Appendix 1. Moreover, extra 1 in diameter holes were drilled into the panels and covered with PVC fittings to allow future insertion of other electrodes. Electrical connections to panels R7 and R15 were no longer working as ev idenced in resistivity measurements between the meshes to other buried elements. SagÂŸÂŽs et al  provide details.
14 2.1.3 Pensacola Ave., Tallahassee, PAV Site The site was demolished before the present project started in 2006. The location of the site was on the Florida State University campus at the coordinates of 30Â¡26'24"N x 84Â¡18'24"W. 2.1.4 Palm City Bridge, Stuart/Palm City, PC Site The PC site consists of one structure containing two walls on the Northeast (PCE) and Northwest (PCW) on the bridge on Florida State Road 714 that crosses the South fork of the St. Lucie River. The PCE wall's coordinates are 27Â¡10'32"N x 80Â¡15'49"W, and is currently (2009) across from a Marine Max yacht dealership. The PCE wall contains 4 locations with test clusters at panel rows, R1, R5, R14, and R28. The PCW wall was the most complicated wall to reach in this investigation as the access point is in a tidal zone. To get to PCW, one makes the first left turn on the Southwest side of the bridge to park u nder the bridge's West side just past its causeway. The PCW contains 2 locations with test clusters at rows R3W and R5W. The coordinates for PCW are 27Â¡10'24"N x 80Â¡15'30"W. The reference electrode connection was damaged at the PCW Panel R3W and was i nitially repaired by inserting a stainless steel screw into the wire stump. The repair was not effective, so a CSE temporarily inserted in the soil hole of the panel was used instead as a reference electrode for the polarization measurements. At PCE, the Panel R1 connection stainless steel rod to the
15 bottom strip was broken were it emerged at the concrete surface, making electrical connections with alligator clips difficult but still feasible. 2.1.5 Port St. Lucie Blvd., Port St. Lucie, PSL Site The P SL site is on Florida State Road 716 on the Southeast corner of the bridge that crosses the Northern fork of the St. Lucie River. The coordinates are 27Â¡16'21"N x 80Â¡19'5"W. The Wall is located in the flood plain of the river so over the course of time f looding brought dirt to the site. This added dirt caused panel R7's bottom strip connection to be buried, and the area became covered in overgrowth. Additionally, when the site was instrumented, the stainless steel rods for test points were not thoroughly cleaned when the concrete patch was made to fill in the access hole made for inserting the rebar and reference, so further filing or grinding may be necessary to ensure good electrical contacts. The wall has two locations with test clusters R3 and R7. T he area adjacent to the wall was cleared to remove the overgrown brush. Additionally a shovel was needed to uncover the bottom galvanized strip connection in panel R7. 2.1.6 State Rd. 200 Bridge, Ocala, OCA Site The OCA site is on Florida State Road 2 00 where it crosses the CSX railroad tracks. The wall is adjacent to the newly created Thompson Bowl Park of Ocala. The OCA site coordinates are 29Â¡10'45"N x 82Â¡ 8'42"W. Access to the site is achieved by turning North from State Rd 200 onto Southwest 10 th Ave. and turning east onto Southwest 9 th St. The wall has two locations with test clusters, R6 and R25.
16 2.1.7 Acosta Bridge, Jacksonville, ABJ Site The ABJ site's wall is the westernmost MSE wall on the Acosta Bridge with coordinates 30Â¡19'9"N x 81Â¡ 39'46"W. Driving to the MSE wall requires going westbound on Prudential Dr. and turning North at the railroad tracks but not crossing them. After driving parallel to the tracks about 100 m, one can park the vehicle some 20 m away from the MSE wall. The wall has two locations with test clusters R9 and R21. 2.1.8 Veterans Expressway Overpass, Tampa, VET Site This site is located on the NW part of the overpass of the Veterans Expressway (Florida State Toll Road 589) as it crosses Gunn Highway (Hillsbor ough County Road 587) at coordinates 28Â¡ 3'59"N x 82Â¡34'2"W. There are two locations, R16 and R23, with test clusters. The steel rod connection to the buried rebar in R16 was modified in July, 2007 to enable a banana plug wire connector tip to fit direc tly into the steel rod instead of using an alligator clip, creating a more secure electrical contact. 2.2 Field Evaluation Procedure During each field visit, a battery of nondestructive tests was conducted to evaluate corrosion behavior of the reinforc ements and added rebars. In addition, at some sites the panels were cored through so actual coupons of reinforcing strips or meshes could be examined. Soil samples were extracted from selected sites. A summary description is given in the following; further procedure details are given by SagÂŸÂŽs et al [14,15].
17 2.2.1 Half cell Potential At each site a copper/copper sulfate electrode (CSE) was placed in both soil through a hole in the panel and against a freshly chipped sample of panel concrete. The poten tial of each metallic element was measured against the reference electrode using a high impedance voltmeter. To verify measurements, mutual metal to metal potentials were obtained which would correspond to the difference in values obtained from half cells and contrasted with the regular measurements for consistency. 2.2.2 Macrocell Current At the Brickell Ave. MSE walls, additional galvanized strips and rebars were placed in two pieces. Each piece was set up as a front (closest to the external panel ) and back piece. The two pieces would act as a long strip or rebar when they were electrically shorted together. By opening the jumper connectio n between front and back and inserting an ammeter of 5 resistance, a macrocell current was measured. The current direction and magnitude enabled determination of which end of the strip behaved as a net anode/cathode and the extent of corrosion macrocel l action. 2.2.3 Mutual Resistance Using a Nilsson model 400 AC soil resistance meter (Nilsson Electrical Laboratory Inc., Jersey City, NJ) in the two point setting (coupling connectors C1 and P1 and C2 and P2 with jumpers; C and P denote current and potential
18 terminals respectively), resistances were measured between each pair of elements in the same cluster. This method allows determining if broken or shorted connections exist. The Nilsson meter uses a square wave at 97Hz to avoid interference fr om power line stray currents. 2.2.4 Solution Resistance The Nilsson model 400 meter was used in a three point configuration to obtain the solution resistance for the IR compensation of the linear polarization resistance measurements. The current and potent ial terminal at one end (C2 P2) was coupled with a jumper. The current terminal at the other end (C1) was connected to the working electrode (the chosen strip, mesh, or steel). The corresponding potential terminal (P1) was connected to the reference ele ctrode, and C2 P2 were connected to the counter electrodes (the opposite strip or mesh). 2.2.5 Linear Polarization Resistance In the Linear Polarization Resistance (LPR) method, a potentiostat connected to a computer (either a Gamry TM Reference 600 with a laptop computer or Gamry TM PCI4 300 with a built in computer (Gamry, Westchester, PA)) records current response to an applied potential ramp in the cathodic direction. Figure 3 shows the Gamry TM PCI4 300 being used in the field. Details on the princi ples of the method are given by Jones .
19 Figure 3 An LPR test using the Gamry TM potentiostat shows the machine configured with the added rebar as a working electrode, the activated titanium as the reference electrode, and the bottom mesh as the coun ter electrode. The polarization scan started from the open circuit potential (OCP) which is the initial undisturbed potential between the working electrode and the reference electrode. The potential was scanned from the OCP to 10mV below the OCP at a ~ 100 V/s scan rate; potential and current data were acquired typically at ~0.1 mV steps The working, reference and counter electrode configurations were the same as those used for the solution resistance measurement arrangements. The Working and Working Sense tips of the Gamry TM devices were coupled together to connect to the working electrode, the reference electrode tip was connected to the reference electrode, and the counter electrode tip was connected to the counter electrode. When performing the f irst field investigations the standard software script for LPR was used in the Gamry TM Framework software. However it became apparent in many applications that the E I curve showed a current step causing
20 data reliability issues. The step appeared to origi nate from current range switching problems related to the large apparent interfacial capacitance of the buried elements. Gamry engineer, Dr. Bob Rodgers created a customized script named, "_USF Polarization Resistance Ver 4.exp" that limited the lowest cu rrent range to 300nA and increased the stability settings. The script also required sample times greater than one second and minimized the common mode voltage on the I/E converter. This script was then upgraded for use with the Gamry TM 600, which was firs t taken to the field for the Jacksonville Site inspection and then used since (August 2007). The new script was called "Concrete polarization resistance.exp". In addition to adding compatibility to the new Framework software this script also uses 10 point s from the initial voltage scan to calculate a sample period. Some minor anomalies still remained but the updated procedures yielded generally adequate results, confirmed with test measurements with dummy cells using discrete components. The equipment cr eated a .dta file from each field test. The data were imported into a Microsoft TM Excel TM spreadsheet. The columns of Vf' and Im' were copied into a new worksheet and the data was plotted against the open circuit potential, OCP, so a column of Vf Vo (V o=OCP) was made. From the plot a 2 nd order polynomial fit was applied to the graph and the equation was used to find the slope of the E I curve at the terminal potential value 10 mV below the OCP. The trend line polynomial fit generally had very good f it quality, with R 2 >0.95 in most cases. The reported polarization
21 resistance (Rp) is the difference of the terminal slope value and the solution resistance value obtained per Item 2.2.4. An alternative simplified galvanostatic method to evaluate Rp was used in selected reinforcement strips of VET and PCW, where interference of unknown origin introduced intermittent artifacts in the computer controlled tests. In the alternative method a nearly constant current of ~100 A was impressed, by means of a 9V ba ttery and a high value resistor, between counter and working electrodes while monitoring the working electrode reference electrode potential with a 0.1 mV resolution high impedance voltmeter. The current level was adjusted to obtain <10 mV cathodic potenti al excursion after 180 s of current application. The ratio of potential excursion at 180 s to the impressed current, minus the value of Rs obtained per Item 2.2.4, was reported as the value of Rp. Comparison with Rp values obtained with the computerized sy stem under normal operating conditions showed typical better than 20% agreement between both methods. Details of the alternative method will be published elsewhere . 2.2.6 Electrochemical Impedance Spectroscopy The Electrochemical Impedance Spectr oscopy (EIS) tests were conducted for selected elements in the frequency range from 5mHz to 5kHz with 3 points per decade resolution. The Gamry TM potentiostats indicated above were used for these measurements as well. Details on the principles of the meth od are given by Jones .
22 2.2.7 Other Tests Reinforcement coupons were collected at selected locations listed in Table 2.2 at cored holes indicated in Appendix 1. Cutting out samples of the reinforcement was achieved by using a small power saw that could be operated through a 5" diameter cored hole through the panel. Soil samples were obtained at selected locations through core holes used for extracting coupons, as indicated in Appendix 1. Table 2.2 Extracted Samples Wall Sample /Panel Elevation (m)* Soil Extracted Â Top Mesh / R9 1.3 Top Mesh Hook / R9 1.3 Top Mesh/ R21 1.4 X HFB Bottom Mesh/ 21 0.7 X Top Right Strip / R2W 0.4 PCW Top Left Strip / R4W 0.4 PSL Top Left Strip / R8 0.3 Bottom Right Strip / R5 A and B 0.2 OCA T op Right Strip / R24 0.5 ABJ Top Right Strip / R20 0.5 Â Indicates a 10 liter bucket of soil was removed from the panel row location *Elevations are with respect to ground or designated reference level per Appendix 1. 2.3 Laboratory Evaluations 2. 3.1 Soil Tests Soil tests were performed using FDOT method s  for soil resistivity, pH 1 chloride concentration and sulfate concentration. Resistivity and pH determinations were made at times ranging from 1 4 weeks of sample extraction. Chloride and sulfate analyses were conducted after sample storage periods 1 Determinations using pH paper, not available for HFB.
23 ranging from 3 20 months after extraction, possibly affecting the results as noted later on. 2.3.2 Metallograp hy The reinforcement coupons were cleaned of any loose debris or loose corrosion products and the overall coating plus remaining deposit thickness was measured with a magnetic coating thickness gauge (Mikrotest III, ElektroPhysik, Arlington Heights, IL) at multiple sampling points. Small portions of the coupons were cut out with a slow speed diamond saw (Isomet, Buehler, Lake Bluff, IL) with non aqueous lubricant (Isocut Fluid, Buehler, Lake Bluff, IL) and then cold mounted in a metallographic epoxy c ompound which promotes edge retention (Epoxicure, Buehler, Lake Bluff, IL). The metallographic preparation was conducted with water free grinding and polishing to prevent oxidation of the zinc. The polished samples were etched to provide contrast betwee n the base steel and zinc layers with a 1% nitric acid solution in denatured ethyl alcohol. A metallographic microscope was used to measure the thickness of the corrosion wastage and remaining zinc at multiple locations around the sample perimeter.
24 CHAPTE R 3: RESULTS 3.1 Field Data This section presents some results in summary form. A comprehensive listing of the corresponding detailed primary field data appears in Appendix 2. 3.1.1 Visual Appearance of Wall and Extracted Reinforce ment Results are summarized in Table 3.1. Figure 4 shows metallographs demonstrating both high and low corrosion. Table 3.1 Visual Appearance of Wall and Extracted Reinforcement Wall Location Panel Coating Condition Rating 1 Red Rust 1 External Wall Cond ition BRN No apparent distress BRS One concrete spall 9 Top Mesh VG NP 9 Top Hook VG NP 21 Top Mesh VG NP HFB 21 Bot. Mesh VG NP Numerous small concrete spalls reflecting estuary chloride exposure and low rebar cover. PCE No apparent distress 2W VG <5% PCW 4W VG NP Scale on panels from tidal exposure PSL 8 VG <5% No apparent distress 5 G <10% OCA 24 VG <5% Covered in ivy, otherwise no apparent distress ABJ 20 G <10% No apparent distress VET No apparent distres s 1. VG = Very Good, red rust < 5%, G=Good, red rust between 5 20%. Rating reflects percentage of rust on entire specimen surface. NP: No red rust present.
25 Figure 4 Examples of metallographic cross sections showing low (A) and high corrosion wastage (B ). The base metal is at the bottom. 3.1.2 Solution Resistance, Polarization Resistance, Apparent Corrosion Rates From LPR and Half cell Potential Values Solution resistance (Rs), polarization resistance (Rp), and apparent corrosion rate (ACR) results for individual tests are given in Tables A2 1 to A2 7 of Appendix 2 for both the 1994 98 and the present survey. Averaged ACR results of multiple tests of each element for each field visit and corresponding half cell potentials for both surveys are present ed in Table 3.2
26 Table 3.2 Summary of electrochemical field observations for instrumented structures. (continued on next page)
27 Table 3.2 (cont.) Summary of electrochemical field observations for instrumented structures.
28 Table 3.2 (cont.) Sum mary of electrochemical field observations for instrumented structures.
29 3.1.3 Electrochemical Impedance Spectroscopy Comparison of polarization resistance values obtained by LPR and EIS are shown in Figure 5 and Figure 6. Detailed listing of EIS pa rameters obtained for each EIS test of the present survey, as well as the corresponding ACR values is presented in Tables A2 8 to A2 14 of Appendix 2. Figure 5 Comparison of Rp values obtained by LPR and EIS for the same galvanized steel elements and during the same field visit for the indicated walls. The diagonal line corresponds to an ideal 1:1 correlation.
30 Figure 6 Comparison of Rp values obtained by LPR and EIS for the same plain steel elements and during the same field visit for th e indicated walls. The diagonal line corresponds to an ideal 1:1 correlation. 3.1.4 Macrocell Current Values These data could only be reliably obtained during the June, 2008 visit to the BR site. Results are shown in Table 3.3.
31 Table 3.3 Macr ocell currents for the Brickell Ave. Bridge site from June 2008. BRN BRS Connection (+) / ( ) Time (s) (mA) 5 0.01 0.07 10 0.01 0.06 Back Galv / Front Galv 60 0.01 0.07 1 0.105 10 0.10 Top Back Steel / Front Steel 60 0.105 5 0.09 0.00* 10 0.09 0.00* Back Galv / Front Galv 60 0.09 0.004** 1 0.07 10 0.075 Bott om Back Steel / Front Steel 60 0.075 *Partial data indicates negligible macrocell activity **Value taken at higher resolution setting Positive values i ndicate the element denoted by (+) is the net cathode. 3.2 Laboratory Data 3.2.1 Soil Tests Data Soil properties obtained in the present survey are displayed in Table 3.4. Table 3.4 Averaged soil properties measured in the 1994 98 and 2006 09 surveys Cl (ppm) SO 4 2 (ppm) Resist ivity (k cm) pH Wall 94 98 06 09 94 98 06 09 94 98 06 09 94 98 06 09 BRN 13 30 42 9 .1 BRS 5 .3 9 34 9 .1 HFB 22 ND 1 5 .0 16 16 8 .3 PCE 2 .5 ND 40 9 .1 PCW 160** ND 67** 11 1 .2 ** 13 8 PSL 20 ND 8 15 7. 5 9.3 8 .3 7.8 OCA 8 .3 ND 3 4 .0 37 24 7.3 6.5 ABJ 4.7 ND 1 3.5 29 16 8.4 7 VET 2 .3 7.3 21 5.5 Resistivity average reported as inverse of average conductivity if multiple samples existed. *ND indicates below detection limit **'94 98 survey showed high vari ability of composition of PCW
32 3.2.2 Metallographic and Magnetic Gage Measurements Except for HFB, all samples extracted came from the lowest elevations possible per core drill placement. Figure 4 shows metallographs illustrating instances of low corrosio n (nearly complete galvanized layer) and high corrosion (including a thick corrosion product layer). Table 3.5 shows averages of thickness measurements for each wall and for a coupon from a control regular production MSE strip retained from experiments per formed during the 1994 7 investigation. Table 3.5 Summary of Metallographic and Magnetic Gage Thickness Measurements ( m) Metallographic Magnetic Gage Wall Age at time of Coupon Extraction (y) Number of Coupons Remaining Galvanized Layer Corrosion Product Layer Total Thickness ACR ( m/y) MET Total Thickness BRN BRS HFB 13.8 3 107 55 161 4.0 210 PCE PCW 17.8 2 56 50 106 2.8 145 PSL 11.9 1 139 22 161 1.8 189 OCA 24.6 3 119 15 135 0.61 150 ABJ 17.7 1 75 15 90 0.85 104 VET Control 151 151 140 (Average of measurements for all coupons of each wall)
33 CHAPTER 4: DISCUSSION 4.1 Direct Assessment Results Visual appearance of the walls examined (Table 3.1) did not reveal any outward signs of distress related to corrosion of the earth reinforcement or any other obvious structu ral distress. The numerous small spalls present in HFB were examined separately from this investigation and found to be consistent with instances of very small (e.g. 6 mm) concrete cover over the medallion's rebar. The small cover likely permitted rapid p enetration of chloride from the concrete surface (the wall is placed ~ 5m from the shore of the causeway facing estuarial waters containing in the order of 10,000 ppm Cl ) with consequent initiation and propagation of corrosion of the rebar. BRS had one spall which was similar in rational to the spalls in HFB, except the BRS wall was also painted. Chloride content of the soil extracted from core holes (Table 3.4) was non detectable. This result may reflect the long storage period before the analyses wer e conducted, possibly promoting conversion of the chloride into evasive species  or forms non detectable by the method used. Thus, the chloride analysis results from the present survey will not be considered to be relevant by themselves. Sulfate conte nts and pH were in the general range of those obtained in the earlier survey, suggesting that, if no artifact from long term storage affected
34 those analyses, no adverse evolution of that parameter in the intervening period. Unlike the soil chemical compos ition analysis, most pH and resistivity measurements were conducted shortly after sample extraction and thus considered to be reliable, an expectation supported by the results being generally close to those obtained in the 1994 98 survey. The resistivity values were amply above the 3,000 ohm cm design minimum, consistent with the interpretation that no recent adverse soil contamination took place in the structures from which soil samples were taken. Visual examination of reinforcement coupons showed lit tle indication of distress with condition ranging from Fair to Very Good, and only small regions of incipient rust. The rating of the galvanizing was based upon the amounts of red rust visible with Very Good and Good ratings corresponding to less than 5% and between 5 20% of the surface area respectively. These observations are consistent with results from examination of strip holding hardware and visual inspection at cored locations conducted in the 1994 98 survey and further indicative of no severe agin g deterioration since. Yet the more exhaustive metallographic examination of coupons conducted in the present survey showed however distinct indications of wastage of the galvanized layer in progress. Corrosion product layers are on average a sizable frac tion of the remaining galvanized layer. As a confirmation of the metallographic measurements Figure 7 shows general consistency between the metallographic total thickness results and the magnetic gage measurements. Quantitative analysis of the results to
35 e stimate integrated corrosion rates and comparison with results from electrochemical measurements is presented later in this thesis. Figure 7 Comparison of galvanized layer plus corrosion product thickness from metallographic measurements (MET Total Thickness) with total film thickness determined with a magnetic thickness gage (MAG), averaged for coupons extracted n the present survey from each Wall. The diagonal line corresponds to an ideal 1:1 correlation. Data from Table 3.5. 4.2 Electrochemical Estimates of Corrosion Wastage The corrosion rates estimated from electrochemical measurements (obtained from the LPR method unless indicated otherwise) are the result of calculations based on numerous assumptions  which can be only be partially
36 ful filled in any given system. Those values are considered therefore as an approximation of the actual corrosion rate at the time of the measurement and will be reported in the following as apparent corrosion rates (ACR), expressed in m/y. The information from Table 3.2 was used to obtain ACR averages for each galvanized reinforcing element over all visits in the present survey. The highest and lowest element ACR averages for each wall (except for BR where the two walls at that site were treated as one) we re noted. Those element averages were in turn averaged for each wall. The results including high and low values are displayed in Figure 8, along with similar results for the 1994 98 survey, indicating the structure age range spanned in each survey. Fi gure 8 Wall averaged galvanized steel ACR values for the present and previous surveys.
37 The averaged ACR results for galvanized steel at each wall were generally of the same order for both surveys, with a very small value of<1 m/y, in most cases. This value is, encouragingly, at the low end of the range commonly anticipated for buried galvanized steel . With one exception (PSL) individual wall survey to survey differences in average ACR were markedly smaller than the range spanned by the results of individual measurements in a given wall, thus obscuring any effect of interim aging on ACR. To assist in revealing overall trends, the results were graphically summarized in Figure 9, which shows cumulative distributions of ACR for both surveys.
38 F igure 9 Cumulative distributions of ACR averaged for each wall from the 1994 98 and 2006 09 surveys, fitted with cumulative lognormal distributions. Results for Galvanized steel are grouped per survey. Results for plain steel rods are grouped per element age group as indicated in the text. The galvanized steel ACR cumulative distributions from the 1994 98 and 2006 09 surveys nearly overlap so a more detailed analysis was implemented to elucidate possible underlying trends. The distributions are marke dly skewed when displayed in a linear plot but become more symmetric in a logarithmic plot as in Figure 7. Following the presentation by SagÂŸÂŽs et al. , the data were fit with ideal lognormal distributions with resulting parameters summarized in Table 4.1.
39 Table 4.1 Analysis of wall averaged ACR results. Elements Galvanized Plain Steel Survey 1994 98 2006 09 1994 98* 2006 09* Average Age (y) 5.9 16.0 0.5 10.6 Average ln(ACR) ( m/y) 0.30 0.45 1.72 1.54 Std. Dev. ln(ACR) ( m/y) 0.53 0.29 0. 77 0.70 Lognormal Std. Dev. expressed as ratio 1.7 1.3 2.1 2.0 Median per Lognormal Dist. ( m/y) 0.74 0.64 5.60 4.65 Average ( m/y) 0.85 0.67 7.65 5.68 ACR Early/Aged Ratio Median 1.15 1.20 ACR Early/Aged Ratio Avge. 1.27 1.35 n based on median ACR 0 .59 0.94 n based on average ACR 0.76 0.90 *Steel elements newly placed in HFB grouped with those of the 1994 98 survey. As shown in Table 4.1 and consistent with visual appearance in Figure 7, the median and average ACR values for the earlier survey are somewhat higher (by 15% and 27% respectively) than those found in the current survey. Such change would be in the expected direction (corrosion rates decreasing with time of burial [3,4]) but statistical significance of these figures is limited in vi ew of the large variability of results. The lognormal standard deviations, when expressed as ratios from the value one lognormal standard deviation above median to that
40 of median, are high (1.7 and 1.3 for the earlier and present surveys respectively, or 70% and 30% variations). In an ideal lognormal distribution the standard deviation of the median of the 8 value sample considered would be ~3 times smaller than the overall standard deviation. If that were to apply to the current case, then the calculated decrease in median (and average) ACR values from the previous to the current surveys may be considered to be only marginally significant. A similar treatment was applied to ACR data in Table 3.2 for the plain steel bars, with the results seen in Figure 7 and Table 4.1. With the exception of BR the steel bars were buried after the walls had been in place for some time, and in the case of HFB half of the steel bars were buried during the first survey and half during the second. Therefore, the plain steel AC R data are grouped by age at the time of testing, with one group for average age ~11y (buried during the first survey and tested during the second) and the other group for average age ~0.5y (for all bars buried and tested during the first survey plus the 4 bars buried at HFB during the second survey). The plain steel ACR values in both surveys were, with little statistical uncertainty, much higher (average ~ 6 m/y) than those for the galvanized elements. On the other hand, as in the galvanized steel cas e the distributions for the early and late age data are quite close to each other. Other parallels with the galvanized steel case are noted in the following. The plain steel ACR data are better approximated by a lognormal distribution than by a linear one. The lognormal fit values are shown in Table 3.2. The median and average plain steel
41 ACR values for the earlier ~0.5y tests are higher (by 20% and 35% respectively) than those in the ~11 year tests, but the wall to wall variability indicated by the standa rd deviation (corresponding to ~100% in the upward direction, ~50% downward) is even greater that in the case of the galvanized steel elements. Thus, even accounting for lesser uncertainty in the value of the median (and average) on account of the multipl e sample size the calculated overall decrease with time in the ACR of plain steel is also marginally statistically significant. Regardless of the time dependence question, it is noted that the ~6 m/y average corrosion rates found for steel in the relatively long term ~11y tests are well within the range of those reported in the literature for similar buried conditions [4,5]. While recognizing the uncertainty in the time dependence indicated above it is instructive to determine how those trends would compare with general observation of corrosion wastage in buried metals. As found in the investigations by Romanoff and others [4,5] corrosion metal loss x tends to follow a dependence with time t give n by x= k t n Eq. (1) where k is a proportionality constant and n a parameter with value between 0 and 1. The corresponding corrosion rate time dependence is therefore dx/dt = k n t n 1 Eq. (2) Rearranging Eq.(2) to solve for n and usin g the median and average ACR values from Table 4.1 yields nominal n values displayed further below on the same
42 Table. For galvanized steel the nominal n values computed using either the median or the average ACR values were near 0.7, which is in agreement with values often reported in the literature for buried galvanized steel as summarized in Table 4.2 [4,5]. In contrast, the nominal n values computed for plain steel were closer to unity, denoting a corrosion rate that decreases relatively slowly with time That trend also approximates results with reported values of n for plain steel that are somewhat higher than those for galvanized steel. Due to the variability and associated uncertainty in time trends noted above, later measurements over a wider time b aseline may be needed to better resolve time dependence of corrosion rate in these structures. Table 4.2 Time dependence parameter n from sources reported in References  and . Study Galvanized Plain Steel NBS* Avg.  0.65 0.80 NBS* Max.  0.6 5 0.80 France** Low  0.60 0.65 France** High  0.60 1.00 National Bureau of Standards ** French soil box investigations. 4.3 Accuracy and Consistency of Electrochemical Corrosion Measurements A validation check of the accuracy of the ACR d eterminations was conducted for galvanized steel elements by comparison with integrated wastage estimates from the metallographic examinations. The corrosion products were assumed to have a zinc content (66 wt%) and density (3.05 g/cm 3 ) similar to those of solid Zn(OH) 2 a common composition for zinc corrosion products .
43 The average corrosion product thickness for each structure per Table 3.5 was converted accordingly to the equivalent thickness of solid zinc as an estimate of galvanized layer wastage The wastage thickness was divided by the age of the wall at the time of coupon extraction to obtain a metallographically estimated corrosion rate (MET CR) reflecting the average corrosion rate experienced during the entire exposure period. Results are sh own in Figure 10, where the diagonal line represents ideal agreement between LPR ACR and MET CR results. Figure 10 Comparison of integrated corrosion rates evaluated from metallographic measurements (MET) of coupons extracted from five walls in the pre sent survey, averaged for coupons from each wall, with corresponding average ACR values from LPR measurements. The calculations assumed that the corrosion products behaved as solid Zn(OH) 2 Triangles: 1994 98 survey. Circles: 2006 09 survey. Open symbols: results from individual walls. Filled symbols: average of all values in each survey. The diagonal line corresponds to an ideal 1:1 correlation.
44 There is order of magnitude agreement between metallographic and electrochemical corrosion assessments for ind ividual walls based on either survey, and significantly better correlation when the average of all walls is considered (filled symbols). It is emphasized that the metallographic method result reflects metal wastage rate averaged over the entire exposure p eriod, while the electrochemical measurement determine instantaneous corrosion rate. Thus the value of a direct comparison is limited by variability in corrosion rates both long term (as in the expected gradual decrease of rate with time while the galvani zed layer is in place) and short term reflecting seasonal and tidal influences. Further limitation is due to the small amount and size of reinforcement coupons available, and that the coupons were always from one region of the wall, closest to the outer su rface thus introducing a sample bias that may not reflect conditions further in. In contrast, the electrochemical measurements involve the entire buried element length. Keeping in mind these factors and the typical method uncertainty and actual variabil ity in corrosion distribution, the metallographic results generally support the validity of the electrochemical LPR measurements. Although the approximate validity of the LPR estimates of corrosion rate for galvanized steel was supported by the direct met allographic observations, it is important to examine to what extent the ACR values may vary when alternative electrochemical techniques are used. Consequently, the internal consistency of the electrochemical ACR determinations was examined by contrasting LPR
45 results with those from independent EIS measurements performed from time to time at selected elements, as detailed in Appendix 2. Analysis of the results was performed by fitting the EIS data with the analog circuit shown in Figure 11, restricting t he analysis to the frequency range 0.01 Hz to 0.8Hz. That procedure yields values for the polarization resistance, Rp EIS, which can be compared with those obtained by the LPR method. EIS analysis reliability was limited by uncertainty inherent to the gen erally low values of the frequency dispersion coefficient observed in soil systems , especially for cases where the value of Rp was high . Figure 11 Analog circuit used to analyze the EIS data. The constant phase angle element (CPE) has param eters Yo (Ss n ) and n. Consequently only analyses for which Rp EIS <1k were contrasted with the LPR polarization resistance (Rp LPR) determinations. The comparison results are shown in Figures 5 and 6 for tests with galvanized and plain steel elements respectively. With the exception of one Rp EIS value at the high end CPE Rp Rs
46 of the range in both cases, there is an approximate correlation between both methods over a wide range of values. Superimposed on a moderate amount of scatter there is also an overall o ffset of ~1.5:1 for the observed Rp EIS / Rp LPR, which is not surprising given the many working assumptions and consequent model uncertainty involved in the interpretation of these types of data [19,21]. This comparison supports concluding that a reasonab le degree of internal consistency exists for the electrochemical ACR determinations. The systematic offset between results of alternative test methods underscores the importance of using consistent electrochemical measurement and analysis methodology from survey to survey. Another independent indication of corrosion activity is provided by the macrocell current measurements, for which the BR walls have been instrumented. Table 3.3 shows that significant macrocell action was taking place in five of the six instrumented elements, with macrocell currents in the 10 to 130 A range. In three of those the cases (one galvanized and two steel divided elements) the cathode was the half of the element further away from the external surface indicating that corrosion was greater at the front. The elements with the opposite polarity were both part of divided galvanized strips. When translated into a current density and corresponding average enhanced corrosion rate at the anode, the effect is on the order of a fraction of 1 m/y. Although small in absolute terms, it must be recalled that the typical ACR values are also of the same order, so corrosion macrocells could easily double the local corrosion rate at the net anode. Any further localization of the macrocell current co uld likewise
47 multiply the metal loss in small regions, possibly leading to a substantial decrease of cross section there. The half cell potentials observed in the present survey (Table 3.2) had values comparable to those in the 1994 98 survey, as shown i n the summary of average values in Table 4.3. As observed in the previous survey, these potentials are only roughly informative of corrosion condition. More detailed analysis of the results did not reveal a clear correlation between ACR and the half cel l potential. It is possible, however, that a correlation may be observed in the future as consumption of the galvanized layer begins to expose some of the more Fe rich lower layers of the film and eventually the base steel itself. This consumption would results in potentials between those of galvanized steel and plain steel. Table 4.3 Average galvanized steel half cell potentials (V) vs a CSE in contact with soil for the present and previous surveys. Wall 2006 09 1994 98 BRN 0.750 0.480 BRS 0.586 0.532 HFB 0.569 0.744 PCE 0.603 0.643 PCW 0.662 0.742 PSL 0.733 0.781 OCA 0.577 0.587 ABJ 0.481 0.507 VET 0.565 0.530 4.4 Predictive Model The predictive model used is the same as described previously [14,15], operated to reflect th e corrosion distributions obtained in the present as well as
48 the previous surveys. A generic Florida condition is considered, so as a working estimate the lognormal distribution parameters based on those given in Table 4.1 are used and assumed to reflect t he distribution of corrosion rates over the strips in a given structure. The corrosion rates are considered to be time invariant as opposed to time dependent as discussed previously for simplicity and conservativeness. Separate calculations are conducted using the 1994 98 and the 2006 09 distributions to reveal sensitivity to the parameter choices and to examine the implications of the added data developed in the present investigations. As the model is detailed in those previous publications, only salient points are addressed, following the treatment by SagÂŸÂŽs et al. . The following modeling assumptions apply: The corrosion rates actually used for the calculations are those per the Table 4.1 distributions but multiplied by 2 to account for corrosion l ocalization per the discussion in the previous section. Except for that multiplier, corrosion is treated as uniform along the strips. Corrosion at the strip edges is ignored. Element failure is declared upon the base steel reaching one half of its origina l thickness (one quarter loss of thickness on each side), since when that condition is reached stresses on the strip are likely to have grossly exceed the original design value. The average strip is considered to have a steel thickness s = 4mm and a galvan ized layer thickness g = 150 m.
4 9 Per the above assumptions and as indicated in  the time to failure of a galvanized strip is given by Eq. (3) where v g and v s are the corrosion rates of the galvanized and the plain steel (af ter it is exposed) for a given element. Each element has its own galvanized and plain steel corrosion rate values assigned per the assumed distributions. Calling C g (v g ) and C s (v s ) the cumulative distribution of ACR values for the galvanized layer and base steel respectively and calling P g (v g ) the probability distribution for the galvanized layer ACR: Eq. (4) the derivation in  shows that Ff ( t ) = P g v g ( ) 1 C s s 4 t g v g ( ) [ ] ( ) d v g = g / t # $ v g Eq. (5) where Ff(t) is the fraction of elements in the wall that failed by time t. Though the co rrosion rates are conservatively assumed to be time invariant; this simplification may be relaxed in future model implementations as more reliable time dependence data are developed over a longer period of time.
50 Figure 12 shows the projections based on Eq. (10) and the parameters abstracted from the previous and the present surveys. Consistent with the very small ACR values obtained for the galvanized layers, both inputs result in projections of minimum damage (<5% elements failed) at age 100 years, and for reaching one half of elements damaged after 200 years. The results from the present survey yield a moderately more optimistic outlook due to the fractional decrease in ACR values with respect to the first survey discussed earlier. Given the sustained character of the trends confirmed by the present survey, no further model expansion was deemed necessary at this time. Figure 12 Model projections of percentage of damage elements in a generic MSE wall as function of wall age. The projections are bas ed on the lognormal distribution parameters in Table 4.1. The dashed and solid lines correspond to the parameters abstracted from the 1994 98 and the 2006 09 survey data respectively.
51 It is emphasized that the model projections are based on sweeping assu mptions as well as on apparent corrosion rate values only approximately validated by direct observation. Consequently, the projections are subject to considerable uncertainty and should be treated with caution. Nevertheless, the overall observations indica te that corrosion deterioration so far has been mild in the structures investigated, and that there is a good prognosis for adequate corrosion performance in future decades barring unusual circumstances such as extensive backfill contamination. Periodic c ontinuation surveys should nevertheless be conducted for verification. The low apparent corrosion rates observed appear to reflect successful control of backfill composition to avoid corrosive agents. Events such as saltwater inundation (addressed by SagÂŸ ÂŽs et al ) or aggressive chemical spills could dramatically degrade corrosion performance and in such circumstances the corrosion condition of the affected structure should be promptly assessed in detail.
52 CHAPTER 5: CONCLUSION S 5.1 First Conclusion Nine reinforced mechanically stabilized earth (MSE) walls (average age 16 years) in seven Florida sites representing a variety of service conditions were evaluated as a continuation of a previous survey conducted a decade earlier. Surface appearance of coupons extracted from actual reinforcement elements in five of the walls showed in general little evidence of distress. There was no external evidence of any earth reinforcement corrosion in any of the walls. Chemical analysis and r esistivity measurements of extracted backfill revealed no unusual contamination. 5.2 Second Conclusion Metallographic examination of the reinforcement coupons showed only moderate wastage of the galvanized layer, corresponding to low corrosion rates, est imated to range from ~0.2 to ~1.1 m/y. 5.3 Third Conclusion An extended series of linear polarization resistance (LPR) measurements yielded galvanized reinforcement apparent corrosion rates (ACR) that were also low (average ~0.7 m/y) and in general a greement with those estimated
53 metallographically. Similar analysis of results from the earlier survey yielded ~0.9 m/y. These values are at the low end of the range commonly anticipated for galvanized steel reinforcement in MSE walls. 5.4 Fourth Conclu sion LPR tests produced average ACR ~6 m/y values for ~11 year old embedded plain steel elements in the same walls. The average ACR for measurements conducted for the same or similar elements at average age ~0.5 year was ~8 m/y. These values are within the anticipated range for buried steel in similar conditions. 5.5 Fifth Conclusion The fractional drop in average ACR values between both surveys, although subject to uncertainty due to variability in the data and subject to future confirmation, was on the order of that expected for buried components. 5.6 Sixth Conclusion The approximate accuracy and the electrochemical ACR estimates for galvanized steel was supported by agreement with direct metallographic examination, while internal consistency wa s established for both galvanized and plain steel ACR measurements by comparison with the results of independent electrochemical impedance spectroscopy tests in the field.
54 5.7 Seventh Conclusion Corrosion macrocell current measurements at one of the tes t sites showed interactions between the front and back portions of the reinforcement that corresponded to an appreciable fraction of the overall corrosion rate. This information was used to apply a localized corrosion multiplier in the damage prediction mo del. 5.8 Eighth Conclusion A statistical model that takes into account the estimated galvanized steel and plain steel corrosion rates and their variability was applied to project the evolution of corrosion related damage in a generic Florida MSE wall, us ing the data developed in the present and previous surveys. Consistent with the very small ACR values obtained for the galvanized layers, both inputs result in projections of minimum damage (<5% and <1% elements failed for the earlier and present survey) a t age 100 years, and for about one half of elements experiencing damage after 200 years. 5.9 Final Conclusion The overall observations indicate that corrosion deterioration so far has been mild in the structures investigated, and that there is a good pr ognosis for adequate corrosion performance in future decades barring unusual circumstances such as extensive backfill contamination. Periodic continuation surveys should be conducted for verification.
55 REFERENCES 1. Gladstone, R., An derson, P., Fishman, K., Withiam, J., "Durability of Galvanized Soil Reinforcement: 30+ Years of Experience with MSE," presented at the 85 th Annual Meeting of TRB, Washington, D.C., 2006. 2. "The performance of buried galvanized steel earth reinforceme nts after years or service", P. Anderson and J. Sankey, in Landmarks in Earth Reinforcement, Ochlai et al. (eds), Swets & Zeitlinger, ISBN 90 2651 863 3, 2001. 3. Elias, V., Christopher, B., Berg, R., "Mechanically Stabilized Earth Walls and Reinforced S oil Slope Design and Construction Guidelines," Rep. No. FHWA NH1 00 043, National Highway Institute, Federal Highway Administration, Washington D.C., 2001. 4. Romanoff, M., "Underground Corrosion," NBS Circular 579, U.S. Dept. of Commerce, Washington, 1957. 5. Elias, V., "Durability/Corrosion of Soil Reinforced Structures," Report No. FHWA RD 89 186, NTIS, Springfield, VA, December 1990. 6. "Coating Properties", Galvanized Rebar Resource Center, 22 Aug. 2008 < http://www.galvanizedrebar.com/coating_properties.htm > 7. Wheeler, J., "MSES Corrosion Evaluation Program, End of the Year Report for 1998," New York State Department of Transportation Geotechnical Engineering Bureau Albany, 1999. 8. Wheeler, J., "MSES Corrosion Evaluation Program, End of the Year Report for 2001," New York State Department of Transportation Geotechnical Engineering Bureau, Albany, 2001. 9. Beckham, T., Sun, L., Hopkins, T., "Corrosion Evaluation of Mechanically Stabilized Earth Walls," Rep. No. KTC 05 28/SPR 239 02 1F, Kentucky Transportation Center, Lexington, KY, 2005. 10. McGee, P., "Reinforced Earth Wall Strip Serviceability Study," Final Report: Study No. 8405, Georgia Department of Transpo rtation, Atlanta, GA,1985.
56 11. Medford, W.M., "Monitoring the Corrosion of Galvanized Earth Wall Reinforcement North Carolina Department of Transportation Field Investigation, presented at 78 th Annual Meeting of TRB, Washington, D.C., January 13, 1999 (preprint). 12. Johnston, D., "Corrosion Monitoring of Hot Springs VSL Mechanically Stabilized Earth Wall," Rep. No. SD2004 02 F, South Dakota Department of Transportation Office of Research, 2005. 13. Coats, D., Castanon, D., Parks, D., "Needs Assess ment for a Maintenance Monitoring Program For Mechanically Stabilized Embankment Structures," State of California Department of Transportation, Sacramento, CA, 2003. 14. SagÂŸÂŽs, A., Scott, R., Rossi, J., PeÂ–a, J., Powers, R., Corrosion of Galvanized Strip s in Florida, Reinforced Earth Walls, Journal of Materials in Civil Engineering, August 2000, pp. 220 227, 2000. 15. SagÂŸÂŽs, A., Rossi, J., Scott, R., PeÂ–a, J.A., Simmons, T. "Influence of Corrosive Inundation on the Corrosion Rates of Galvanized Tie Stri ps In Mechanically Stabilized Earth Walls", Rep. No. WPI0510686, Florida Department of Transportation Research Center, Tallahassee, FL, 1998. 16. Jones, D, "Principles and Prevention of Corrosion", 2 nd ed., Prentice Hall, Upper Saddle River, NJ, 1996 17. SagÂŸÂŽs, A. and Berke, B, To Be Published, 2009. 18. Florida Department of Transportation, Florida Test Methods, FM 5 550, 551, 552, and 553, FDOT State Materials Office, www.dot.state.fl.us 19. Oeberg, G., "Chloride and Organic Chlorine in Soil", Acta Hydrochim, Hydrobiol. 26, pp. 137 144, 1998. 20. SagÂŸÂŽs, A., Kranc, S. and Moreno, E., ""An Improved Method for Polarization Resistance from Small Amplitude Potentiodynamic Scans in Concrete", Corrosion Vol 54, p.20, 1998. 21. Zhang, X.G., "Corrosion and electrochemistry of Zinc", Plenum Press, New York, 1996 22. SagÂŸÂŽs, A., Kranc, S. and Moreno, E. "Time Domain Response of a Corroding System with Constant Phase Angle Interfacial Component: Application to Steel in Concrete", Corrosion Science, Vol.37, p.1097, 1995
58 Appendix 1 : Site Instrumentation Diagrams Site: BR Figure 13 Site diagram of the Brickell Ave. Site.
59 Appendix 1: Site Instrumentation Diagrams (Continued) Fig ure 14 Elevation view showing structure, wall, and location information for BRN . Figure 15 Elevation view showing structure, wall, and location information for BRS .
60 Appendix 1: Site Instrumentation Diagrams (Continued) Figure 16 Dimensions of typical concrete panels and tie strip locations for the BR. site. All dimensions are in meters .
61 Appendix 1 : Site Instrumentation Diagrams (Continued) Figure 17 Reinforcement placement in B R site MSE walls. All dimensions are in meters .
62 Appendix 1 : Site Instrumentation Diagrams (Continued) Figure 18 Cluster diagram showing layout of elements in BRN. All dimensions are in meters .
63 Appendix 1 : Site Instrumentation Diagrams (Continued) Figure 19 Cluster diagram showing layout of elements in BRS. All dimensions are in meters .
64 Appendix 1 : Site Instrumentation Dia grams (Continued) Figure 20 Test points BRN, bottom layer. Normally connected jumpers shown. Figure 21 Test points BRN, top layer. Normally connected jumpers shown.
65 Appendix 1 : Site Instrumentation Diagrams (Continu ed) A B Figure 22 Test points BRS top (A) and bottom (B) layers. Normally connected jumpers shown.
66 Appendix 1 : Site Instrumentation Diagrams (Continued) Site: HFB Figure 23 Elevation view of HFB MSE wall showin g panel nomenclature .
67 Appendix 1 : Site Instrumentation Diagrams (Continued) Figure 24 Dimensions of typical concrete panels and tie strip locations for the HFB site. All dimensions are in meters .
68 Appendix 1 : Site Instrumentation Diagrams (Continued) Figure 25 Top view of a mesh and panel at HFB. All dimensions are in meters and not to scale .
69 Appendix 1 : Site Instrumentation Diagrams (Co ntinued) Figure 26 Panel layout for HFB panels R7, R9, R15, and R21. Panel R9 is shown in this figure. Panels R7 and R15 do not have PVC caps over their reference electrode covers. WT and WB refer to the top and bottom mesh connectors respectivel y.
70 Appendix 1 : Site Instrumentation Diagrams (Continued) Figure 27 Panel layout for HFB panel R11. WT and WB refer to the top and bottom mesh connectors respectively.
71 Appendix 1 : Site In strumentation Diagrams (Continued) Figure 28 Panel layout for HFB panel R17. WT and WB refer to the top and bottom mesh connectors respectively.
72 Appendix 1 : Site Instrumentation Diagrams (Continued) Figure 2 9 Core hole locations in HFB panel R9. Soil samples were extracted from each hole. Metal reinforcement coupons (a piece of mesh and the connector hook) were only removed from the top hole.
73 Appendix 1 : Site Instrumentatio n Diagrams (Continued) Figure 30 Core hole locations in HFB panel R21. Metal reinforcement coupons were removed from the top and bottom holes. Small spalls, caused by panel concrete reinforcing steel corrosion at points of low concrete cover, are vi sible.
74 Appendix 1 : Site Instrumentation Diagrams (Continued) Site: PC PCE Wall Figure 31 Plan view of PCE showing panels with instrument clusters. D imensions are not to scale .
75 Appendix 1 : Site Instrumentation Diagrams (Continued) Figure 32 Dimensions of typical concrete panels and tie strip locations for PCE. All dimensions are in meters .
76 Appendix 1 : Site Instrumentation Diagrams (Continued) Figure 33 Panel layout for panel R1 at PCE. WT and WB refer to the top and bottom strip connectors respectively.
77 Appendix 1 : Site Instrumentation Diagrams (Contin ued) Figure 34 Panel layout for panel R5 at PCE. WT and WB refer to the top and bottom strip connectors respectively.
78 Appendix 1 : Site Instrumentation Diagrams (Continued) Figure 35 Panel layout for panel R14 at PCE. WT and WB ref er to the top and bottom strip connectors respectively.
79 Appendix 1 : Site Instrumentation Diagrams (Continued) Figure 36 Panel layout for panel R28 at the Stuart NE site. WT and WB refer to the top and bottom strip connectors respect ively.
80 Appendix 1 : Site Instrumentation Diagrams (Continued) PCW Wall Figure 37 Diagram indicating the location of the panels R3W and R5W PCW .
81 Appendix 1 : Site Instrumentation Diagrams (Conti nued) Figure 38 Detailed panel information for panel R3W at PCW.
82 Appendix 1 : Site Instrumentation Diagrams (Continued) Figure 39 Detailed panel information for panel R5W at PCW.
83 Appendix 1 : Site Instrumentation Diagrams (Continued) Figure 40 Location of additional core holes at PCW. Soil samples and metal reinforcement coupons were removed from each hole in panels R2W and R4W.
84 Appendix 1 : Site Instrumentation Diagrams (Continued) Site: PSL Figure 41 View of relative panel locations at PSL. This figure is revised from the report for the 1994 98 survey to show correct cluster locations .
85 Appendix 1 : Site Instrumentation Diagr ams (Continued) Figure 42 Detailed panel information for panel 3 at PSL.
86 Appendix 1 : Site Instrumentation Diagrams (Continued) Figure 43 Detailed panel information for panel 7 at PSL.
87 Appendix 1 : Site Instrumentation Diagrams (Continued) Figure 44 Additional core hole at PSL panel R4, from which a soil sample was extracted. Figure 45 Additional core hole at PSL panel R8. A soil sample a nd metal reinforcement coupon were removed. Core Hole R8 Core Hole R4
88 Appendix 1 : Site Instrumentation Diagrams (Continued) Site: OCA Figure 46 Diagram indicating the location of the panels R6 and R25 at OCA .
89 Appendix 1 : Site Instrumenta tion Diagrams (Continued) Figure 47 Panel layout for Panel 6 at OCA. D corresponds to the top galvanized strip, E to the steel rebar and reference electrode (the reference electrode is identified on site by a green cable) and F corresponds to the b ottom galvanized strip .
90 Appendix 1 : Site Instrumentation Diagrams (Continued) Figure 48 Panel layout for Panel 25 at OCA. A corresponds to the top galvanized strip, B to the rebar and reference electrode (the reference electrode is identified on site by a green cable) and C corresponds to the bottom galvanized strip .
91 Appendix 1 : Site Instrumentation Diagrams (Continued) Figure 49 Approximate location of an add itional core hole at OCA panel R5. A soil sample and two metal reinforcement coupons were extracted from that hole. Approximate Core Hole R5
92 Appendix 1 : Site Instrumentation Diagrams (Continued) Figure 50 Approximate location of an additional core hole at OCA panel R24. A soil sample and one metal reinforcement coupon were extracted from that hole.
93 Appendix 1 : Site Instrumentation Diagrams (Continued) Site: ABJ Figure 51 Diagram indicating the location of the panels R9 and R21 of ABJ .
94 Appendix 1 : Site Instrumentation Diagrams (Continued) Figure 52 Panel layout for Panel 9 at ABJ. A corresponds to the top galvanized strip, B to the bottom galvanized strip, C to an added Zn Al strip, an d D to the rebar and reference electrode (the reference electrode is identified on site by a green cable) .
95 Appendix 1 : Site Instrumentation Diagrams (Continued) Figure 53 Panel layout for Panel 21 at ABJ. E corresponds to the top galvanized strip, F to the bottom galvanized strip, and G to the rebar and reference electrode (the reference electrode is identified on site by a green cable) .
96 Appendix 1 : Site Instrumentation Diagrams (Continued) Figure 54 Additional core hole at ABJ panel R20. A soil sample and metal reinforcement coupon were removed from that hole.
97 Appendix 1 : Site Instrumentation Diagrams (Continued) Site: VET Figure 55 Diagram indicating the loc ation of the panels R16 and R23 of VET .
98 Appendix 1 : Site Instrumentation Diagrams (Continued) Figure 56 Panel layout for Panel 16 at VET. A corresponds to the top galvanized strip, B to the rebar and reference electrode (the reference ele ctrode is identified on site by a green cable), C to the bottom galvanized strip, and D to the added Zn Al strip .
99 Appendix 1 : Site Instrumentation Diagrams (Continued) Figure 57 Panel layout for Panel 23 at VET. E corres ponds to the top galvanized strip, F to the reference G to the bottom galvanized strip, and H to the added Zn Al strip. Note: This location does not have rebar steel inserted .
1 00 Appendix 1 : Site Instrumentation Diagrams (Continued) Tab le A1.1 Dimensions of test elements. Length was estimated based on the height of the reinforcement compared to other structures Â New rebars were inserted in to panels R7, R9, R15, and R21
101 A ppendix 2: Detailed ACR Data This Append ix contains detailed electrochemical data. Tables A2.1 to A2.7 tabulate Nilsson meter solution resistance measurements, number ("Disc") identifying the Gamry TM data file for LPR tests, calculated LPR Rp value and corresponding ACR values for each test run in the field in both the first and second surveys. Tables A2.8 to A2.14 list, for the second survey only, all the electrochemical impedance spectroscopy (EIS) test results together with the corresponding LPR results of tests conducted for the same eleme nt during the same visit. In all tables the first column refers to the element tested and its paired counter electrode. The abbreviations are as follows: GT: galvanized top element GB: galvanized bottom element, S: buried plain steel rebar Z: b uried Zn Al strips Comb.: (Combination) test arrangement measured In BR various combinations are not repeated as there are 6 reference electrodes used at BRN and 4 reference electrodes used at BRS for each elevation. Values of X' denote discarded te sts where the calculated Rp values numbers were near 0 or over 1,000 except for Zn Al Rp > 1,000 Figures 58 to 65 are cumulative distributions of ACR for galvanized steel elements, grouped by site. Figure 66 shows the corrosion rate behavior of the inserted plain steel rebars grouped by age.
102 Appendix 2: Detailed ACR Data (Continued) Table A2.1 LPR data for BR.
103 Appendix 2: Detailed ACR Data (Continued) Table A2.2 LPR data for HFB.
104 Appe ndix 2: Detailed ACR Data (Continued) Table A2.3 LPR data for PC.
105 Appendix 2: Detailed ACR Data (Continued) Table A2.4 LPR data for PSL. Table A2.5 LPR data for OCA.
106 Appendix 2: Detailed ACR Data (Continued ) Table A2.6 LPR data for ABJ. Table A2.7 LPR data for VET.
107 Appendix 2: Detailed ACR Data (Continued) Table A2.8 EIS LPR data for BR.
108 Appendix 2: Detailed ACR Data (Continued) Table A2.9 EIS LPR data for HFB.
109 Appendix 2: Detailed ACR Data (Continued) Table A2.10 EIS LPR data for PC.
110 Appendix 2: Detailed ACR Data (Continued) Table A2.11 EIS LPR data for PSL. Table A2.12 EIS LPR data for OCA.
111 Appendix 2: Detailed ACR Data (Continued) Table A2.13 EIS LPR data for ABJ. Table A2.14 EIS LPR data for VET.
112 Appendix 2: Detailed ACR Data (Continued) Figure 58 Cumulative distribution of ACR from BRN and BRS. Figure 59 Cumulative distribution of ACR from HFB.
113 Appendix 2: Detailed ACR Data (Continued) Figure 60 Cumulative distribution of ACR from PCE. Figure 61 Cumulative distribution of ACR from PCW.
114 Appendix 2: Detailed ACR Data (Continued) Figure 62 Cumulative distribution of ACR from PSL. Figure 63 Cumulative distribution of ACR from OCA.
115 Appendix 2: Detailed ACR Data (Continued) Figure 64 Cumulative distribu tion of ACR from ABJ. Figure 65 Cumulative distribution of ACR from VET.
116 Appendix 2: Detailed ACR Data (Continued) Figure 66 Cumulative distribution of ACR from all plain steel elements in all walls grouped by years since insertion.
117 Appendix 3: Metallography This appendix contains detailed information of the Metallographic examinations conducted on the structural element coupons collected from the field sites. Table A3.1 is a more detail ed version of the concise Table 3.1 presented earlier. Table A3.2 lists all of the details regarding the extracted metal coupons including how many metallographic mounted samples were made. Figure 67 presents a summary of average percentage of coupon surf ace showing rust for all coupons from each site, as function of age of the wall at the time of coupon extraction. Figure 68 shows the distribution of galvanized and corrosion product thicknesses from metallographic examinations, averaged per site. Figur es 69 to 75 show views of the as extracted metal coupons after light cleaning to remove loosely adhering soil. Dashed lines indicate where the specimens where cut to prepare the metallographic cross sections (only post sectioning pictures are available fo r HFB). In all cases, "Top Side" indicates the face of the strip that was facing upwards in the structure. Figures 76 to 81 contain pictures of post sectioned HFB coupons.
118 Appendix 3: Metallography (Continued) Table A3.1 Detail ed direct observation of metal coupons. Site Location Panel Side 1 Coating Condition 3 Red Rust BRN BRS 9 Hook All 2 VG NP 9 Top All 2 VG NP 21 Top All 2 VG NP HFB 21 Bottom All 2 VG NP PCE Top <5% Bottom <5% 2W Si de VG <5% Top NP Bottom NP PCW 4W Side VG NP Top <5% Bottom <5% PSL 8 Side VG <5% Top <10% Bottom <10% 5A Side <50% Top <10% Bottom <5% 5B Side G <50% Top <5% Bottom <5% OCA 24 Side VG NP Top <20% Bo ttom NP ABJ 20 Side VG <5% VET 1. Refers to face of the strip in contact with the soil. 2. In HFB the mesh are cylindrical, so the entire surface was examined uniformly. 3. VG=very good (<5% red rust on entire surface), G=good (5 20% red rust on ent ire surface)
119 Appendix 3: Metallography (Continued) Table A3.2 Listing of all mounted samples and respective measurement information.
120 Appendix 3: Metallography (Continued) Figure 67 Plot of percentage of red rust observed by age by site. Figure 68 Cumulative lognormal distribution of coating thickness measurements from coupons collected from the field.
121 Appendix 3: Metallography (Continued) Note: in Figures 69 to 75 the dashed lines represent the cros s section examined using the microscope. Figure 69 The top side of the coupon from PCW panel 2W. Figure 70 The top side of the coupon from PCW panel 4W.
122 Appendix 3: Metallography (Continued) Figure 71 The top side of the cou pon from PSL panel 8. Figure 72 The top side of coupon A' from OCA panel 5.
123 Appendix 3: Metallography (Continued) Figure 73 The top side of coupon B' from OCA panel 5. Figure 74 The top side of the coupon from OCA panel 24.
124 Appendix 3: Metallography (Continued) Figure 75 The top side of the coupon from ABJ panel 20. Figure 76 Elevation views of the coupon from HFB panel 9 top mesh. A: Top is up. B: Opposite side, top is down. The wall panel connection is to the left in both pictures. Figure 77 The bottom side of the coupon from HFB panel 9 top mesh. The wall panel connection is to the left.
125 Appendix 3: Metallography (Continued) Figure 78 Elevation views of the coupon from HFB panel 22 top mesh. A: Top is up. B: Opposite side, top is down. The wall panel connection is to the left in both pictures.
126 Appendix 3: Metallography (Continued) Figu re 79 View from above (A) and below (B) of part of the coupon from HFB panel 22 top mesh.
127 Appendix 3: Metallography (Continued) Figure 80 View from top (A), side (B), and bottom (C) of part of the coupon from HFB panel 22 bottom mesh.
128 Appendix 3: Metallography (Continued) Figure 81 Pictures of segments cut from a hook in HFB panel 9 in the top layer mesh. The hook was heavily deformed during the coring of the hole to make an electrical contact to the mesh causing much of the disbanding of the galvanizing and later the rusting (B).