USF Libraries
USF Digital Collections

Accuracy and bias of TDR measurements in compacted sands

MISSING IMAGE

Material Information

Title:
Accuracy and bias of TDR measurements in compacted sands
Physical Description:
Book
Language:
English
Creator:
White, Newel K
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla.
Publication Date:

Subjects

Subjects / Keywords:
electrical conductivity
water content
dry density
dielectric constant
nuclear
Dissertations, Academic -- Civil Engineering -- Masters -- USF   ( lcsh )
Genre:
government publication (state, provincial, terriorial, dependent)   ( marcgt )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Summary:
ABSTRACT: It is essential to properly monitor in-situ soil compaction properties during most earthwork construction projects. Traditional in-situ soil compaction monitoring methods are often limited in their application. As a result, new methods are being developed to more accurately measure in-situ compaction parameters. Time domain reflectometry (TDR) is one such method. Relying on the propagation of an electromagnetic wave through the soil sample, TDR can be used to measure both in-situ moisture content as well as soil dry density. Although TDR is relatively new to the field of geotechnical engineering, it has previously been implemented in other fields with success. Researchers at Purdue University have made several advances to further incorporate the use of TDR technology into the field of geotechnical engineering and as a result an innovative TDR measurement system has been developed for compaction control monitoring. The method was standardized in the form of ASTM D 6780 in 2002. Further advancements led to an improved method referred to as the Purdue one-step TDR method. Research has indicated that the ASTM TDR method is sufficiently accurate for application in compaction monitoring applications. A comparison between the ASTM TDR method and traditional methods was carried out to evaluate the accuracy of the TDR method to traditional methods. To further expand the application of the TDR method, a correlation was developed between the TDR spike driving process with the in-situ CBR test. A comprehensive review of previous research was conducted to examine recent advancements leading to the improved Purdue one-step method. A study was also performed to evaluate the effect of variable pore fluid conductivity on the calibration of the Purdue one-step method.
Thesis:
Thesis (M.S.C.E.)--University of South Florida, 2004.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Newel Kimball White.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 116 pages.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001478737
oclc - 56389342
notis - AJS2426
usfldc doi - E14-SFE0000378
usfldc handle - e14.378
System ID:
SFS0025071:00001


This item is only available as the following downloads:


Full Text
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001478737
003 fts
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 040811s2004 flua sbm s000|0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0000378
035
(OCoLC)56389342
9
AJS2426
b SE
SFE0000378
040
FHM
c FHM
090
TA145 (ONLINE)
1 100
White, Newel K.
0 245
Accuracy and bias of TDR measurements in compacted sands
h [electronic resource] /
by Newel Kimball White.
260
[Tampa, Fla.] :
University of South Florida,
2004.
502
Thesis (M.S.C.E.)--University of South Florida, 2004.
504
Includes bibliographical references.
516
Text (Electronic thesis) in PDF format.
538
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
500
Title from PDF of title page.
Document formatted into pages; contains 116 pages.
520
ABSTRACT: It is essential to properly monitor in-situ soil compaction properties during most earthwork construction projects. Traditional in-situ soil compaction monitoring methods are often limited in their application. As a result, new methods are being developed to more accurately measure in-situ compaction parameters. Time domain reflectometry (TDR) is one such method. Relying on the propagation of an electromagnetic wave through the soil sample, TDR can be used to measure both in-situ moisture content as well as soil dry density. Although TDR is relatively new to the field of geotechnical engineering, it has previously been implemented in other fields with success. Researchers at Purdue University have made several advances to further incorporate the use of TDR technology into the field of geotechnical engineering and as a result an innovative TDR measurement system has been developed for compaction control monitoring. The method was standardized in the form of ASTM D 6780 in 2002. Further advancements led to an improved method referred to as the Purdue one-step TDR method. Research has indicated that the ASTM TDR method is sufficiently accurate for application in compaction monitoring applications. A comparison between the ASTM TDR method and traditional methods was carried out to evaluate the accuracy of the TDR method to traditional methods. To further expand the application of the TDR method, a correlation was developed between the TDR spike driving process with the in-situ CBR test. A comprehensive review of previous research was conducted to examine recent advancements leading to the improved Purdue one-step method. A study was also performed to evaluate the effect of variable pore fluid conductivity on the calibration of the Purdue one-step method.
590
Adviser: Ashmawy, Alaa K.
653
electrical conductivity.
water content.
dry density.
dielectric constant.
nuclear.
690
Dissertations, Academic
z USF
x Civil Engineering
Masters.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.378



PAGE 1

Accuracy and Bias of TDR Measurements in Compacted Sands by Newel Kimball White 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: Alaa K. Ashmawy, Ph.D. Manjriker Gunaratne, Ph.D. Ram M. Pendyala, Ph.D. Date of Approval: June 25, 2004 Keywords: water content, dry density, dielectri c constant, electrical conductivity, nuclear Copyright 2004, Newel Kimball White

PAGE 2

DEDICATION This work is dedicated to my wife Karen. She has been the motivating force behind my effort.

PAGE 3

ACKNOWLEDGEMENTS I would like to thank Dr. Ashmawy for th e opportunity to further my education as well as for his insight and guidance as a mentor. Also, the Florida Department of Transportation should be thanked for supporting my efforts through a research grant. Rory M. McDonald should be recognized for his assistance while performing several in-situ CBR tests. Rory is also noted for his expert skill as a traffic directions specialist. His guidance, es pecially in a tough spot, was irreplaceable. Hebron B. White “Bob Jr.” is to be acknowledged for his unpa ralleled food service i ndustry experience, which proved to be invaluable during my defense. Nephi So lorzano is to be thanked for his assistance duri ng testing as well. Finally I would like to thank my famil y. Specifically I woul d like to thank my Grandpa and Grandma White for setting a prece dent of the importance of education, my mother, Eileen White, for instilling an apprecia tion for dirt and rock sciences as well as a love and encouragement beyond belief.

PAGE 4

i TABLE OF CONTENTS Background..................................................................................................................... 1 Organization of Thesis....................................................................................................2 Introduction................................................................................................................... ..3 Methods for Determining Moisture Content..................................................................3 Laboratory Determination of Water Cont ent of Soil and Rock (ASTM D2216).......3 Microwave Oven Method (ASTM D 4643)................................................................4 Direct Heating Method (ASTM D4959).....................................................................4 Calcium Carbide Gas Pressure Tester Method (ASTM D4944)................................4 Nuclear Method (Shallow Depth) (ASTM D3017)....................................................4 Methods for Determining In-place Density....................................................................5 Nuclear Method (ASTM D5195)................................................................................5 Nuclear Method (shallow depth) (ASTM D3017)......................................................5 Sleeve Method (ASTM D4564)..................................................................................6 Drive-Cylinder Method (ASTM D2937)....................................................................6 Sand-Cone Method (ASTM D1556)...........................................................................6 Rubber Balloon Method (ASTM D2167)...................................................................7 Summary of Existing Methods.......................................................................................7 Historical Summary of Time Domain Reflectometry.....................................................8 TDR Basics..................................................................................................................... 9 LIST OF TABLES..............................................................................................................v LIST OF FIGURES...........................................................................................................vi ABSTRACT....................................................................................................................... ix CHAPTER 1 INTRODUCTION......................................................................................1 CHAPTER 2 – LITERATURE REVIEW..........................................................................3

PAGE 5

ii Development of the ASTM TDR Measurement System..............................................12 TDR Electromagnetic Wave.........................................................................................12 Complex Dielectric Permittivity...................................................................................14 Soil Dielectric Constant from TDR Waveforms...........................................................15 Soil Apparent Dielectric Constant Relationships.........................................................18 Bulk Electrical Conductivity........................................................................................21 Bulk Electrical Conductivity from TDR Waveforms...................................................21 Bulk Soil Electrical Conductivity Calibrations............................................................23 Dielectric Constant and Bulk Electri cal Conductivity Relationship for Soil...............25 Purdue TDR One-step Methodology............................................................................26 Limitations of TDR Measurement................................................................................29 Introduction................................................................................................................... 30 TDR Measurement System...........................................................................................31 Data Acquisition...........................................................................................................34 TDR Software (PMTDR-SM)......................................................................................35 Testing Procedure.........................................................................................................37 Calibration.................................................................................................................37 In-situ Testing...........................................................................................................44 Introduction................................................................................................................... 47 Calibration Constants “a” and “b”................................................................................47 Calibration Constants “c” and “d”................................................................................52 Calibration Constants “f” and “g”................................................................................53 The Effect of Pore Fluid Conductivity on Calibration Constants.................................55 Effect on Constants “a” and “b”...............................................................................55 Effect on Constants “c” and “d”...............................................................................56 Effect on Constants “f” and “g”................................................................................58 Summary...................................................................................................................60 Effect of Initial Salt Content on Calibration Constants................................................60 CHAPTER 3 EQUIPMENT AND PROCEDURE........................................................30 CHAPTER 4 EVALUATION OF PURD UE TDR SOIL PARAMETERS...................47

PAGE 6

iii Initial Salt Content Testing.......................................................................................60 Results and Discussion for “a” and “b”....................................................................63 Results and Discussion for “c” and “d”....................................................................65 Results and Discussion for “f” and “g”.....................................................................67 Conclusions...............................................................................................................70 Summary.......................................................................................................................7 1 Introduction................................................................................................................... 72 Theoretical Background................................................................................................72 Existing Empirical Correlations to the CBR Test.........................................................73 TDR and In-situ CBR Correlation................................................................................77 Equipment.................................................................................................................77 Procedure..................................................................................................................79 Calculation................................................................................................................81 Test Results................................................................................................................... 83 Discussion/Analysis......................................................................................................83 Proposed Model............................................................................................................84 Conclusion....................................................................................................................8 7 Introduction................................................................................................................... 88 Testing Program............................................................................................................88 Nuclear vs. ASTM TDR Results..................................................................................89 Sand Cone vs. ASTM TDR Results..............................................................................93 Drive Sleeve vs. ASTM TDR Results..........................................................................94 Water Content Measurement Discussion......................................................................96 Dry Density Discussion................................................................................................96 Conclusions...................................................................................................................9 7 Summary.......................................................................................................................9 8 Conclusions...................................................................................................................9 8 CHAPTER 5 IN-SITU CBR CORRELATI ON TO TDR SPIKE DRIVING................72 CHAPTER 6 ASTM METHOD COMPARED TO TRADITIONAL METHODS.......88 CHAPTER 7 – SUMMARY, CONCLU SIONS AND RECOMMENDATIONS...........98

PAGE 7

iv Recommendations for Future Research......................................................................100 REFERENCES...............................................................................................................101

PAGE 8

v LIST OF TABLES Table 2-1 Comparison of Current Methods.................................................................8 Table 4-1 Values of Constants “a” and “b” for Various Sands..................................49 Table 4-2 Error Resulting from Variation of “b” on Predicted Moisture Content.....51 Table 4-3 Error Resulting from Varia tion of “a” on Predicted Dry Density.............51 Table 4-4 Error Resulting from Varia tion of “b” on Predicted Dry Density.............52 Table 4-5 Purdue TDR Constants for Florida Sands..................................................54 Table 4-6 Testing Material Summary........................................................................61 Table 5-1 Summary of CBR Field Testing................................................................84 Table 6-1 Nuclear Testing Locations and Information..............................................89 Table 6-2 Nuclear Water Content Comparison Results.............................................90 Table 6-3 Nuclear Dry Density Comparison Results.................................................92 Table 6-4 Sand Cone Dry Density Comparison Results............................................94 Table 6-5 Drive Sleeve Dry Density Comparison Results.........................................95

PAGE 9

vi LIST OF FIGURES Figure 2-1 Basic TDR Setup..........................................................................................9 Figure 2-2 Electromagnetic Field in Coaxial Line......................................................10 Figure 2-3 TDR Soil Moisture System Configuration................................................11 Figure 2-4 Typical TDR Output Voltage.....................................................................11 Figure 2-5 Purdue TDR Measurement System............................................................13 Figure 2-6 Typical TDR Wave Reflection..................................................................15 Figure 2-7 Electrical Conductivity Wave Analysis.....................................................22 Figure 2-8 Adjusting the Field Samp le to the Laboratory Calibration........................27 Figure 3-1 TDR System Configuration.......................................................................31 Figure 3-2 Configuration of Coaxial Head..................................................................32 Figure 3-3 The Coaxial Cylinde r (CC) Transmission Line.........................................33 Figure 3-4 The Multiple Rod Probe (MRP)................................................................34 Figure 3-5 TDR Field Measurement Case...................................................................35 Figure 3-6 In-situ MRP Input Screen..........................................................................36 Figure 3-7 CC Mold Test Input Screen.......................................................................37 Figure 3-8 Compaction by Tamping with Aluminum Rod.........................................38 Figure 3-9 Mold and Wet Sample Being Weighed.....................................................39 Figure 3-10 Central Spike Being Driven Through the Guide and into the Sample.......40 Figure 3-11 Taking the TDR Measurement...................................................................40 Figure 3-12 Example of “a” and “b” Calibration for Ottawa Sand...............................41 Figure 3-13 Example of “c” and “d” Calibration for Ottawa Sand...............................42 Figure 3-14 Example of “f” and “g” Calibration for Ottawa Sand................................43 Figure 3-15 PMTDR-SM Calibration Tool...................................................................43 Figure 3-16 Driving Spikes through Template into Soil...............................................45 Figure 3-17 Removal of the Template...........................................................................45

PAGE 10

vii Figure 3-18 Placement of Coaxial Head (CH) on Spikes..............................................46 Figure 4-1 Final Calibration Results for “a” and “b”..................................................48 Figure 4-2 Effect of Compaction Energy on Constants “a” and “b”...........................50 Figure 4-3 Ottawa Sand Variation of “b” with Pore Fluid Conductivity....................55 Figure 4-4 Sample #6978 Variation of “b” with Pore Fluid Conductivity..................55 Figure 4-5 Fl Sand #2 Variation of “b” with Pore Fluid Conductivity.......................56 Figure 4-6 Ottawa Sand Variation of “d” with Pore Fluid Conductivity....................56 Figure 4-7 Sample #6978 Variation of “d” with Pore Fluid Conductivity..................57 Figure 4-8 Fl Sand #2 Variation of “d” with Pore Fluid Conductivity.......................57 Figure 4-9 Relationship between ECb and Gravimetric Water Content......................58 Figure 4-10 Ottawa Sand Variation of “g” with Pore Fluid Conductivity....................59 Figure 4-11 Sample #6978 Variation of “g” with Pore Fluid Conductivity..................59 Figure 4-12 Fl Sand #2 Variation of “g” with Pore Fluid Conductivity.......................59 Figure 4-13 Mixing Soil with Deionized Water............................................................62 Figure 4-14 Draining Water from Sample ....................................................................62 Figure 4-15 Ottawa Sand Calibration for “a” and “b”...................................................63 Figure 4-16 Sample #6978 Calibration for “a” and “b”................................................64 Figure 4-17 Fl Sand #2 Calibration for “a” and “b”......................................................64 Figure 4-18 Ottawa Sand Calibration for “c” and “d”...................................................65 Figure 4-19 Sample #6978 Calibration for “c” and “d”................................................66 Figure 4-20 Fl Sand #2 Calibration for “c” and “d”......................................................66 Figure 4-21 Ottawa Sand Calibration for “f” and “g”...................................................68 Figure 4-22 Sample #6978 Calibration for “f” and “g”.................................................68 Figure 4-23 Fl Sand #2 Calibration for “f” and “g”......................................................69 Figure 5-1 DCP and CBR Relationship.......................................................................75 Figure 5-2 SPT and CBR Relationship........................................................................76 Figure 5-3 Equipment for In-Place CBR Test.............................................................78 Figure 5-4 CBR In-place Equipment Setup.................................................................78 Figure 5-5 The Standard Proctor Hammer with the Attached Head...........................79 Figure 5-6 TDR Spike Driving Technique..................................................................81

PAGE 11

viii Figure 5-7 Correction of Penetration Curve................................................................82 Figure 5-8 Proposed CBR and TDR Relationship.......................................................85 Figure 5-9 Proposed Range of In-situ CBR Values....................................................86 Figure 6-1 Nuclear versus ASTM TDR Water Content..............................................91 Figure 6-2 Nuclear versus ASTM TDR Dry Density..................................................93 Figure 6-3 Sand Cone versus ASTM TDR Dry Density.............................................94 Figure 6-4 Drive Sleeve versus TDR Dry Density......................................................95

PAGE 12

ix ACCURACY AND BIAS OF TDR MEAS UREMENTS IN COMPACTED SANDS Newel Kimball White ABSTRACT It is essential to properly monitor in-s itu soil compaction properties during most earthwork construction projec ts. Traditional in-situ soil compaction monitoring methods are often limited in their application. As a result, new methods are being developed to more accurately measure in-situ compaction parameters. Time domain reflectometry (TDR) is one such method. Relying on the propagation of an electromagnetic wave through the soil sample, TDR can be used to measure both in-situ moisture content as well as soil dry density. Although TDR is relatively new to the field of geotechnical engineering, it has previously been impl emented in other fields with success. Researchers at Purdue University have made several advances to fu rther incorporate the use of TDR technology into the field of geotechnical engi neering and as a result an innovative TDR measurement system has been developed for compaction control monitoring. The method was standardized in the form of ASTM D 6780 in 2002. Further advancements led to an improved me thod referred to as the Purdue one-step TDR method. Research has indicated that the AS TM TDR method is sufficiently accurate for application in compaction monitoring app lications. A comparison between the ASTM TDR method and traditional methods was carried out to evaluate the accuracy of the TDR method to traditional methods. To further e xpand the application of the TDR method, a correlation was developed between the TDR sp ike driving process with the in-situ CBR test. A comprehensive review of previous research was conducte d to examine recent advancements leading to the improved Pu rdue one-step method. A study was also performed to evaluate the effect of variab le pore fluid conductivity on the calibration of the Purdue one-step method.

PAGE 13

1 CHAPTER 1 INTRODUCTION Background Most earthwork construction projects re quire some form of quality control to ensure that design conditions are actually met in the field. For an adequate compaction control monitoring system, it is essential to measure the in-situ moisture content and density. Currently there ar e several methods available for determination of these necessary compaction parameters. The majori ty of these methods are limited in their application. As a result, improvements over current methods could be invaluable in reduction of operation time and cost as well as improved measurement accuracy. A new method for in-situ soil moisture content and density measurement has recently been developed using time domain reflectometry in an effort to achieve the aforementioned improvements. Although time domain reflectometry (TDR) ha s been widely used in other fields, it is relatively new to the fiel d of geotechnical measurement and is an altogether different approach to measuring in-situ soil propertie s when compared to traditional geotechnical monitoring methods. Recent studies have indica ted that TDR is a legitimate tool for insitu geotechnical measurement and may be desi rable over traditional methods in several compaction control applications. Current me thods used to measur e in-situ soil density and moisture content often rely on separate and independent tests that are generally run on different soil samples. The TDR method allows for both density and water content measurement at the same time using the same soil sample by evaluating an electromagnetic wave that is sent into the soil sample. The method was standardized in the form of ASTM D6780 in 2002, due in la rge part to research done at Purdue University. Further developments have since been made reducing the time and equipment required for testing. This improve d method is referred to as the Purdue onestep TDR method.

PAGE 14

2 Results from previous researchers i ndicate that the ASTM D6780 method is sufficiently accurate for geotechnical measuremen t. However, little research has been carried out to evaluate the recent developmen ts leading to the Purdue one-step method. An effort has been made to further explore the application of the ASTM TDR method as well as to evaluate recent developments lead ing to the improved Purdue one-step method. Organization of Thesis A comprehensive review of time domain re flectometry and its relation to the field of geotechnical engineering is presented in Ch apter 2. Also included in the chapter is a review of current geotechnical measuremen t methods as well as the basic concepts associated with the soil parameters extracte d from TDR waveforms. Chapter 3 presents both the equipment and procedure used in c onjunction with the ASTM and Purdue onestep TDR methods. Calibration procedures ar e also outlined with in the chapter. A detailed review of the calibration constants us ed with the Purdue one -step TDR method is presented in Chapter 4. The results from a study on the effe cts of pore fluid conductivity on TDR calibration constants are also included in the chapter. Chapter 5 discusses the results obtained from a series of tests that we re carried out in an effort to establish a relationship between the TDR spike driving process and the California bearing ratio (CBR) test. Chapter 6 includes results obtain ed from a testing pr ogram carried out in conjunction with the Florid a Department of Transportation (FDOT) to evaluate the accuracy of the ASTM TDR method. Included within the chapter are comparative results with traditional geot echnical testing methods. Finall y, Chapter 7 provides a summary and conclusions from the work presented within this thesis. Also recommendations are made as to the need of future research.

PAGE 15

3 CHAPTER 2 – LITERATURE REVIEW Introduction Time domain reflectometry has recently been introduced to the field of geotechnical engineering as a reliable tool fo r in-situ measurement. Several traditional methods are currently being used to measure both in-situ soil densit y and water content; however, many of these methods are limited in their application. An understanding of traditional methods and their limitations is presented to demonstrate the need for an alternative comprehensive method. As stat ed previously, the use of time domain reflectometry is relatively new to the fiel d of geotechnical engineering and is a new approach all together for in-s itu soil moisture and densit y measurement. The theory behind the use of time domain reflectometry as it relates to geotechnical measurement is discussed in this chapter as well as its im plementation in to the geotechnical field. A review of recent work relating to th e ASTM TDR method is also presented. Methods for Determining Moisture Content Several methods are used for determining soil moisture content in both the field and the laboratory. The following is a brief summary of these tests with commentary on the limitations of each. Laboratory Determination of Water C ontent of Soil and Rock (ASTM D2216) Equivalent Method: AASHTO T 265 This method is widely known in geotec hnical practice as the “Oven Dry Method”. The underlying principle behind this test is to determine both the weight of solids and weight of water contained in the given soil sample. The sample is placed in a

PAGE 16

4 conventional oven at 110 C for a time period of 24 hours. The wet and dry weights of the sample are determined before and after dryi ng. The apparatus consis ts of: a drying oven, balances and specimen containers. The oven dry method has traditionally been accepted as the baseline standard for geotechnical appl ications. The methods main limitation is the amount of time required to perform the test. Microwave Oven Method (ASTM D 4643) The microwave oven method is similar to the oven dry procedure, except a microwave oven is used in pla ce of a conventional oven. Soils that contain organics may ignite upon drying. When compared to the oven dry method, the microwave oven method yields less accurate results. Direct Heating Method (ASTM D4959) Again, the idea behind the di rect heating method is si milar to the previously mentioned oven methods. The only difference being that a di rect heat source is used; such as a hotplate, stove, or blowtorch. Th e direct heating method yields faster results than the oven dry method, but again, is less accurate. Calcium Carbide Gas Pressure Tester Method (ASTM D4944) Equivalent Method: AASHTO T 217 The calcium carbide method is commonly referred to as the “speedy moisture content” method. The method relies on the us e of a chemical reaction using calcium carbide as a reagent to react with the soil pore water. Th e method is not accurate for highly plastic clays and soils containing minerals that deh ydrate with heat. The test method is limited to soils with particles less than No. 4 sieve size. Since flammable and explosive acetylene gas is i nvolved, appropriate guidelines a nd rules should be followed by the operator. Nuclear Method (Shallow Depth) (ASTM D3017) Equivalent Method: AASHTO T 310 The shallow depth nuclear method calls for a fast neutron source to be applied to the surface of the soil. Using a surface slow ne utron detector, the slowing ratio of the fast

PAGE 17

5 neutron is measured. Using this ratio and the calibration data the moisture content of the soil is calculated. The hydrogen present in wate r is the main factor in this test. The apparatus is highly sensitive to water contained in the top 2 to 3 inches of soil. Hydrogen in forms other than water will cause readings to be in excess of the true value. Some chemical elements such as boron, chlori ne, and minute quantities of cadmium cause measurements to be lower than the true value. Methods for Determining In-place Density Several methods are used for determining the in-place density of soil. The following is a summary of commonly used tests and discussion of their limitations. Nuclear Method (ASTM D5195) The nuclear method requires that a radiatio n tube be inserted into the soil to the desired depth. The tube contains a source a nd a detector of gamma radiation; these are used to measure the attenuation of gamma radi ation through soil. The soil density is then determined by comparing the detected rate of gamma radiation with previously established calibration data. If the dry unit weight is required, the measurement of the inplace water content is needed. Measurements will be higher than the actual values if some elements with greater atomic number s than 20 are encounter ed. Voids around the access tube can greatly affect the measuremen ts. The equipment utilizes radioactive materials that may cause hazards, so proper precautions have to be taken by the operator. Nuclear Method (shallow depth) (ASTM D3017) Equivalent Method: AASHTO T 310 The shallow depth nuclear method is the same as the regular nuclear method, but either the source and detector remains on the surface (Backscatter Method) or one of them is at the surface while the other is at a known depth up to 300mm (Direct Transmission Method). The same limitations apply as with the other nuclear methods.

PAGE 18

6 Sleeve Method (ASTM D4564) The sleeve method requires the insertion of a metal sleeve into the soil. The soil within the sleeve is then removed, and a de termination of the dr y mass of soil removed per linear inch of the depth of the excavation is obtained. The sleeve method is used for soils that are predominantly fine gravel size, with a maximum of 5% fines, and a maximum grain size of ” (19 mm). The test is applicable for c ohesionless soils in a confined or limited space since the test met hod requires less working area compared to the other methods. Consistency in the grada tion and particle angular ity of the soil being tested is critical to the test. The test is ope rator sensitive. The sleeve should be examined periodically for wear. Drive-Cylinder Method (ASTM D2937) Equivalent Method: AASHTO T 204 The drive-cylinder method requires that a th in-wall steel cylinder be driven into a smoothed soil surface using a fall hammer. Th e soil is then excavated from around the cylinder to allow for removal of the steel cyli nder. Using a straightedge, the ends of the cylinder are then trimmed. The weight and the volume of the empt y cylinder are known and the weight of the removed soil can be de termined. Using this information the unit weight of the soil can be determined. To get the dry unit weight of the soil, the moisture content has to be determined using a standa rd method. The test is not applicable for organic soils, very hard natural soils, heav ily compacted soils, and soils which contain appreciable amount of sand. The cutting edge of the cylinder should be examined after each test to ensure that it is still sharp. If any damage o ccurs to the cylinder edge or body, the test results should be discarded. Sand-Cone Method (ASTM D1556) Equivalent Method: AASHTO T 191 In the Sand Cone Method a hole is excavat ed in the ground and the excavated soil is weighed. The volume of the hole is then de termined using standard sand replacement. The standard sand should be dry, clean, uniform, uncemented, durable, and free flowing. Knowing the weight of the standard sand fill and its density, the volume of the hole can

PAGE 19

7 be calculated. The density of the excavated so il can be computed ac cordingly. To obtain the dry density of the soil, the water content of the extracted portion is determined using a standard method. The method is not suitable fo r saturated, soft, orga nic, deformable or highly compressible soils. It is also not suitable for soils that contain appreciable amount of rock or coarse materials (larger than 38mm). Rubber Balloon Method (ASTM D2167) The rubber balloon method relies on the sa me concept as the sand-cone method, but instead of replacing the soil with standard sand, wate r and a balloon are used. A flexible membrane filled with water and conne cted to a water-filled calibrated vessel is used to measure the volume of the hole after extracting the soil. Prior to first use, the apparatus should be calibrated. The suitabi lity of this method is the same as the sandcone method. Summary of Existing Methods The nuclear method is the most commonl y used method in current practice for soil moisture and density measurement. The nuclear method requires training and special licensing to operate and field measurements are only as good as th e calibration of the device. The drive sleeve me thod is also commonly used, but is a destructive method. Both the Sand Cone and Rubber Balloon methods are less frequently used and rely heavily on the skill of the test operator and as a result measurements vary considerably. Further discussion on the comparison of these tests with the ASTM TDR method will be addressed in Chapter 6. A summary of key points of the more commonly run tests is displayed in Table 2-1.

PAGE 20

8 Historical Summary of Time Domain Reflectometry The power and telecommunication industrie s first implemented the use of time domain reflectometry in the 1950s to locate transmission line disc ontinuities (O’Conner and Dowding, 1999). Hugo Freller-Feldegg (1969) expanded the use of time domain reflectometry to the measurement of the diel ectric permittivity of liquids. Topp et al. (1980) furthered the applications of TDR by de monstrating that the apparent dielectric constant of a soil is strongly dependent on the amount of water contained within the soil. Further developments by Topp and Davis (1985) led to the transmission of TDR signals into in-situ soil using metallic rods. Resear chers then began to incorporate the use of bulk electrical conductivity TD R measurements in an effort to estimate soil salinity (Dasberg and Dalton, 1985). As a result, TDR ha s been widely utilized in agricultural applications, where the soil water content of crops can be monitored to optimize irrigation procedures. The us e of TDR has since been expa nded to a wide variety of applications including: soil/rock deforma tion, structural deformation and air-liquid interface monitoring (O’Conner and Dowding, 1999). Recent improvements to soil moisture monitoring systems by researchers at Purdue University have led to the development of the ASTM TDR method for measuring in-situ soil gravimetric water content and density (Siddiqui and Drnevich, 1995; Yu and Drnevich, 2004). These more recent developments from Purdue University are of great interest to the field of geotechnical measurement. Table 2-1. Comparison of Current Methods. TestApplicationRequired TimeMajor Source of Error Oven DryWater Content24 hoursConsidered as baseline measurement Speedy MoistureWater Content15-20 min + calibration Operator's ability to perform test correctly Nuclear MethodWc and Density30 min + calibrationHighly dependant on proficient calibration Sand Cone Density30 min + calibrationOperator dependant

PAGE 21

9 TDR Basics Time domain reflectometry is derived fr om the same technology used in radars, which have been in use since the 1930s. TD R is similar to radar in that a short electromagnetic pulse is first emitted, and then a reflection is measured. TDR can be defined as a measurement device that relies on the use of remote electrical sampling to determine the location and nature of objects. A TDR system essentially consists of a pulser, a sampler, an oscilloscope and a coaxial cable (Figure 2-1). The TDR pulser generates an electronic st ep pulse that travels into the coaxial line. As the pulse travels down the coaxial li ne, a potential differenc e is created between the inner and outer conductors of the coax ial line creating an electromagnetic field between the conductors (Figure 2-2). Pulser Oscilloscope Coaxial cable Figure 2-1. Basic TDR Setup. Sour ce: O’Conner and Dowding (1999)

PAGE 22

10 As this electromagnetic field travels thr ough the coaxial line it creates a wave. If the medium between the inner and outer c onductors is uniform both geometrically and physically, its measured reflection is also uni form. However, if a discontinuity in the coaxial line is encountered (i.e. a change in cable geom etry or the medium between conductors) a distinct reflecti on jump is observed. The time it takes for a reflection to occur along with its sign, length and amplitude are useful in determining both the location of the fault as well as its nature For soil moisture monitoring purposes, the coaxial line is transmitted into the soil using metallic spikes (See Figure 2-3). Figure 2-2. Electromagnetic Field in Coax ial Line. Source: O’Conner and Dowding

PAGE 23

11 Soil properties can be extracted from TDR reflections measured by the configuration displayed in Fi gure 2-3. A typical TDR wave reflection is shown in Figure 2-4 (O’Conner and Dowding, 1999). Figure 2-3. TDR Soil Moisture System Confi guration. Source: Drnevich et al. 2000. Figure 2-4. Typical TDR Output Voltage.

PAGE 24

12 Development of the ASTM TDR Measurement System Topp et al. (1980) suggested th at research be carried ou t to develop transmission line components that would be sufficiently accurate for water content measurement purposes. Several research projects were carri ed out to evaluate different transmission line configurations (Ledieu et al., 1986; Topp et al., 1982; and Dasberg and Dalton, 1985). Although results from these tests indicat ed a reliable relationship existed between the dielectric constant and volumetric water co ntent, the need for a reliable and routine field technique to measure gravimetric water content was still evident. Zeglin et al. (1989) studied several coaxial soil probe configurations an d found that three and four wire configurations were superior to a tw o wire system. Studies investigating cable length, quality and type of pr obe and cable dimensions were carried out by Heimovaara (1993) to determine their in fluence on the accuracy of TDR measurements. These improvements led researchers at Purdue Univer sity (Drnevich et al., 2000) to develop the ASTM TDR measurement system. Their work warranted acceptance of the method in the form of ASTM standard D 6780 in 2002. The basic TDR measurement system used for this research was obtained from Purdue Un iversity and is displayed in Figure 2-5. It includes a Campbell Scientific TDR100 tester which is then connected to a Multiple Rod Probe (MRP) Head. The MRP sits on a series of spikes that are driven into the soil. TDR Electromagnetic Wave The propagation of an electromagnetic fiel d through a transmission line is governed by the wave equation derived from Maxwell’s equations. Drnevich et al. (2000) stated that “there are two important components of the wave equation solution; the characteristic impedance (Z) and the propagation constant ( ).

PAGE 25

13 The characteristic impedance is the rati o of voltage to current propagating along the line. It is a function of the geometry of the transmission line and the dielectric permittivity of the insulating material.” It can be derived for a coaxial line as: * 0 01 2r p rZ a b Ln Z (2-1) Where “b” is the inner diameter of the out er conductor, “a” is the outer diameter of the inner conductor, 0 is the vacuum permittivity (8.854x 10-12), 0 is the vacuum permeability (4 x10-7 H/m), r is the equivalent dielectric permittivity, and Zp is defined as the impedance of the same line filled with air as the medium (Krauss, 1984). Figure 2-5. Purdue TDR Measurement Sy stem. Source: Yu and Drnevich (2004).

PAGE 26

14 The propagation constant is the other intrin sic property of a transmission line. It is only a function of the dielectric permittivit y of the insulating material. It can be derived, for a coaxial line, as: j c f jr *2 (2-2) In which c is the velocity of the electromagnetic waves in free space, and and are the real and imaginary parts of the propa gation constant, respectively. The real part represents the attenuation of the wave, wh ereas the imaginary part is the spatial frequency, which gives the ve locity of wave propagation when divided by temporal frequency (2 f). The TDR waveform recorded by sa mpling oscilloscope is a result of multiple reflections and dispersions. As the water content, conductiv ity, mineral content, density, and chemical composition of the soil vary, different wave reflections are expected. Wave reflection analysis can th en be used to determine soil properties (O’Conner and Dowding, 1999). The relevant pa rameters derived from wave reflection analysis used in conjunction with the AS TM TDR method are th e complex dielectric permittivity and the bulk electrical conductivity. Complex Dielectric Permittivity The complex dielectric permittivity of a material consists of a real and an imaginary portion. The imaginary portion is a ttributed to electric al loss and the real portion is a measure of the amount of energy stored in the materi al (Drnevich et al., 2000). In low loss materials the imaginary portion of the complex permittivity is not significant enough to alter the propagation ve locity and thus the complex dielectric permittivity can be estimated as the real po rtion. Additionally, Da vis and Annan (1977) demonstrated that the real portion of a soils complex permittivity was much more prominent than the imaginary portion over the frequency range of 1 MHz to 1 GHz. As a

PAGE 27

15 result, the real portion of the dielectric permittiv ity is referred to as the apparent dielectric constant of the soil (T opp et al., 1980). Soil Dielectric Constant from TDR Waveforms There are two general approaches to meas uring the permittivity of a material: 1) placing the material between two plates of a ca pacitor and 2) placing the material into a coaxial line and measuring its complex impeda nce. The second approa ch is used in TDR technology. For a complete char acterization of the material several measurements are to be taken over a wide range of frequencies. However, the same information can be obtained in the time domain by using an electroni c pulse that is sent down the coaxial line (Fellner-Feldegg, 1969). After Fellner-Feldegg (1969), TDR ha s been used extensively to measure the complex dielectric permittivity of polar and non-polar liquids (Giese and Tiemann, 1975; Clarkson et al ., 1977). A typical measured TDR wave reflection from a TDR soil measurement system is displayed in Figure 2-6. Figure 2-6. Typical TDR Wave Reflecti on. Source: Yu and Drnevich (2004). ScaledDistance(m) Relative Voltage (V)

PAGE 28

16 Two distinct reflections can be noted from the measured TDR reflection. The first reflection occurs at the air and soil interface and the se cond occurs at the end of the TDR probe (Yu and Drnevich, 2004 ). Topp et al. (1980) previ ously defined the apparent dielectric constant (Ka) as being related to the velocity of the electromagnetic wave traveling through the transmission line. Th e apparent propagation velocity of an electromagnetic wave can be related to the dielectric constant by the following: (2-3) Where ( v ) is the apparent propagation velocity and ( c ) is the velocity of an electromagnetic wave in free space (2.998 x 108 m/s). The apparent propagation velocity can also be related to the travel time between reflection points by the following: (2-4) Where t is the travel time and L is the length of the probe in the soil. Combining Equations 2-3 and 2-4, the apparent dielectric can be expressed as: (2-5) The term 2 ct can be expressed as the apparent length (La) (Baker and Allmaras, 1990) and Equation 2-5 can be reduced to a simplified form as: (2-6) aK c v t L v 2 22 L ct Ka 2 p a aL L K

PAGE 29

17 This relationship is then used to dete rmine the apparent dielectric from the measured TDR reflection. Where La is the apparent length, wh ich is a scaled horizontal distance between the two reflection points and Lp is the length of th e soil probe (Yu and Drnevich, 2004). Several methods have been proposed to select the reflection points displayed in Figure 2-6 (Topp et al., 1982; Baker and Allm aras, 1990). Researchers at Purdue University have adopted an algorith m developed by Drnevich and Yu (2001) for use with the Purdue TDR method (Yu and Dr nevich, 2004). The re flection points are used to determine the apparent length (La) and then the dielectric constant (Ka) can be computed using Equation (2-6). The soil apparent dielectric constant from TDR measurements is affected by soil temperature. Although temper ature effects on the apparent dielectric constant of soil solids are almost negligible, that of wate r experiences a decrease with increasing temperature (Drnevich et al., 2001 ). In an effort to improve the accuracy of the ASTM TDR method, Drnevich et al (2001) studied the effects of soil temperature on soil dielectric constant by TDR. Th eir results indicate that appa rent dielectric constants in soils are somewhat dependant on soil temperatur e. For sands a decrease in dielectric constant was observed with increasing te mperature, whereas in clays the opposite behavior was observed. It was also determined that within 5C of 20C temperature effects could be neglected. The following lin ear temperature correction functions (TCF) were proposed for temperatures ranging between 4C and 40C: (2-7) (2-8) The TCF function is then used to make the temperature correction to the measured value of Ka in the following form: soil cohesive for T TCF soil ss cohesionle for T TCFtest test 005 0 10 1 0015 0 97 0

PAGE 30

18 (2-9) Although the linear corrections derived do not account for all factors relating to the soil apparent dielectric c onstant, the method has yielded satisfactory results (Drnevich et al., 2001). Soil Apparent Dielectric Constant Relationships For most soil solids the dielectric consta nt is between three and four, whereas the dielectric constant of water is near eighty. Due to the di fference between the dielectric constants of water and soil solids, it has been determined that the real part of permittivity of wet soil is dominated by the volumetric water content (O’Conner and Dowding 1999). Topp et al. (1980) incorporated these ideas to develop an empirical relationship between the apparent dielectric permittivity (Ka) and volumetric water content ( ) of soil (Equation 2-10), this relationship has since been used by several researchers with good results in a wide variety of soils (Dasberg and Dalton, 1985; Heimovaara, 1994; Roth et al., 1992; Topp et al., 1984; Zeglin et al., 1992). (2-10) Where the volumetric water content ( ) is defined as: (2-11) Although Topp’s equation has been used with fairly accurate results using a wide variety of soils, several resear chers have indicated that the calibration is not suitable for organics soil, fine-texture d soils and clays (Dobson et al., 1985; Roth et al., 1992; Dirksen and Dasberg, 1993). Variation in Topp’s dielectric rela tionship with soil 2 2 2 4 3 610 3 5 10 92 2 10 5 5 10 3 4 x K x K x K xa a aTCF K KC T a C a 20 solids waterV V

PAGE 31

19 volumetric water content is attributed mainly to soil density and texture effects (Abdula et al., 1988). Several other researchers (Ledieu et al ., 1986; Alharthi and Lange, 1987) assumed a linear relationship between the square root of the apparent dielectric constant (Ka 0.5) and the volumetric water content ( ): (2-12) Where “a” and “b” are calibration constants: “a” = 1.545 and “b” = 8.787 in Ledieu et al. (1986); “a” =1.59 and “b” = 7.83 in Alharthi and Lange (1987). Ledieu et al. (1986) demonstrated that an improved relationship could be obtained by adding bulk dry density: (2-13) Where “a,” “b” and “c” are calibration constants: “a” = 0.297, “b” = 8.79, “c” = 1.344 and d is the bulk dry density in grams/cm3. Ferre et al. (1996), Malicki et al. (1996) and Yu et al. (1997) also assumed a linear calibration similar to Eqn. (2-12): (2-14) Where a and b are soil constants obtained by regression analysis. Malicki et al. (1996) suggested the idea of accounting for density effects of the soil: (2-15) 'a K ba b b b aK 18 1 17 7 159 0 618 0 819 02 5 0 b a Ka c b a Kd a

PAGE 32

20 Where b is measured in units of grams/cm3. Yu and Drnevich (2004) point out that th ese calibration equations are difficult to apply to the field of geot echnical engineering for two reasons: 1) the calibration equations are expressed in terms of volumetri c water content, whereas gravimetric water content is used in the field of geotechnica l engineering and 2) calibrations which have incorporated the density effect such as Malicki et al. (1996 ) are complex and difficult to apply (Yu and Drnevich, 2004). As a result, Si ddiqui and Drnevich (1995) developed the following expression: (2-16) Where “a” and “b” are soil specific calibration constants, d is the dry density of the soil, w is the density of water and w is the gravimetric water content. The SiddiquiDrnevich equation has been used with sati sfactory results using a wide variety soils (Drnevich et al., 2002; Sallam et al., 2004). Further treatmen t of this equation and its constants will be addressed in chapter 4. The gravimetric water content (w) is defined as: (2-17) The gravimetric water content (w) is related to the volumetric water content ( ) by the following: (2-18) If Eqn. (2-18) is combined with E qn. (2-16) the follo wing is obtained: bw a Kd w a solids waterW W w w dw

PAGE 33

21 (2-19) Eqn. (2-19) is the Siddiqui-Drnevich (1995) equation expressed in terms of volumetric water content and is comparable to the linear calibrations derived in Eqns. (212), (2-13) and (2-14). Bulk Electrical Conductivity When an electromagnetic field travels thr ough soil energy dissipat ion into the soil is expected. There are two major causes for TDR signal attenuation: 1) Losses due to the complex nature of a materials dielectric pe rmittivity which causes the wave to have an out of phase portion and 2) the electrica l conductivity of the medium. Electrical conductivity of a medium is caused by surfac e conduction that results from electrically charged particles on the surface of the solids and from ionic conductance that is a result of dissolved electrolytes in the po re water (White et al., 1994). Bulk Electrical Conducti vity from TDR Waveforms Dalton et al. (1984) proposed the simultaneous measur ement of the apparent dielectric constant as well as the bulk elec trical conductivity of soil in an attempt to estimate pore fluid conductivity from TDR measurements. This work led to the following expression for bulk electrical conductivity (ECb): (2-20) Where Ka and Lp are previously defined, and V1 and V2 are defined in Figure 2-7. b a Kw d a 2 1ln 120 V V L K ECp a b

PAGE 34

22 Dalton (1992) later modified his appr oach by accounting for the intervening coaxial cable and the impedance matching transf ormer. Also Zeglin (1989) made further improvements by adding the multiple reflection idea. Yu and Drnevich (2004) argue that the aforementioned methods have two major shortcomings: 1) ECb is coupled with Ka, which may be a source of error and 2) selecting accurate values of V1 and V2 may be problematic. As a result they propose that a long term response analysis be used to determine the bulk el ectrical conductivity: (2-21) Figure 2-7. Electrical Conductivity Wave Analysis. Yu and Drnevich (2004). Scaled Distance (m) Relative Volta g e ( V ) 1 1f bV Vs C EC

PAGE 35

23 Where, Vs is the voltage source which is twice the step pulse and Vf is the long term voltage level (Figure 2-7) (Yu and Drnevich, 2004). C is a constant related to the probe configuration that is obtained th rough calibration. For coaxial probes: (2-22) Where Lp is the probe length in the soil, Rs is the internal resistance of the pulse generator and d0 and d1 are the inner and outer conductor diameters (Giese and Tiemann, 1975). Results obtained by Topp et al. (1988) and Yu and Drnevich (2004) indicate that these relationships yield satis factory results for electrical conductivity measurement. The bulk electrical conductivity measuremen t is affected by soil temperature. Unlike temperature effects for soil apparent dielectric constant, those for bulk electrical conductivity measurement are consistent for both cohesive and cohesionless soils (Yu and Drnevich, 2004). Conductivity increases in a linear fashion with temperature for temperature ranges generally encountered in construction (Abu-Hassanein et al., 1996; Rinaldi and Cuestas, 2002). While a temper ature correction functi on could be developed similar to that for apparent dielectric consta nt, a simplified approach is taken by Yu and Drnevich (2004). The approach is based on a developed relation between apparent dielectric constant and bul k electrical conductivity of the soil, where the field measurement is adjusted to the calibration temperature. This relationship will be discussed later in this chapter. Bulk Soil Electrical Co nductivity Calibrations Soil electrical conductiv ity is influenced by several f actors including: pore water, mineralogy, soil structure, degree of satu ration and surface conductance. Due to the complex nature of soil electrical conductivity, theoretical equations are not available for use (Yu and Drnevich, 2004). However, seve ral empirical correlations have been 1 0ln 2 d d R L Cs p

PAGE 36

24 developed that have yielded satisfactory results (Rhoades et al., 1990; Mitchell, 1993; Malicki et al., 1994). Rhoades et al. (1976) attempted to relate bulk electrical conduc tivity to pore fluid conductivity, volumetric water content and the soil surface conductance in the following form: s w bEC EC T EC (2-23) Where ECw is the pore fluid conductivity, ECs is the soil surface conductance, T is a geometric factor (T=a '+b where a and b are soil specific constants) and is volumetric water content. The bulk electrica l conductivity of soil can then be expressed as: s w w bEC EC b EC a EC '2 (2-24) The above relationship has been used with fairly accurate results (Kalinski and Kelly, 1993). However, Yu and Drnevich (200 4) argue that the e quation is inadequate for geotechnical engineering a pplications for the following reasons: 1) it does not account for soil skeleton density and 2) electrical c onductivity is expressed in terms of volumetric water content (Yu and Drnevich, 2004). Yu and Drnevich (2004) further argue th at the electrical conductivity from the pore fluid is typically the dominating factor in the determination of the bulk electrical conductivity of soil. As a re sult, the amount of pore fluid present in the soil generally dominates the bulk electrical c onductivity of the soil. This phenomenon is also noted in the measurement of the apparent dielectric cons tant. As previously mentioned, due to the large difference in dielectric constants be tween water and soil solids the apparent dielectric constant of wet soil is domi nated by the volumetric water content (Dowding and O’Conner, 1999). Due to the dominance of the pore fluid in th e apparent dielectric constant and bulk electrical conductivity meas urement, a relationship is derived for bulk electrical conductivity and gravimetric water content th at is similar to the Siddiqui-

PAGE 37

25 Drnevich equation for apparent dielectric constant and gravimetric water content (Yu and Drnevich, 2004). The proposed re lationship can be expressed as: (2-25) Where “c” and “d” are soil specific calib ration constants. Yu and Drnevich (2004) state that Eqn. (2-25) has several advantages incl uding: 1) gravimetric water content is used, thus it is more suitable to the geotechnical field; 2) conduction from the pore water and soil surf ace is accounted for; 3) soil skelet on density is accounted for; and 4) the equation is simple in format and ea sy to apply. The relationship has been investigated by Feng (1999) an d Lin (1999) and their findi ngs are consistent with previous research done by White et al. (1994) and Hilhorst (2000). Also, data from Amenta et al. (2000) indicates that a good linear relation ex ists when applying Eqn. (225). Further treatment of this equation and its constants will be addressed in chapter 4. Dielectric Constant and Bulk Electric al Conductivity Relationship for Soil Although apparent dielectric constant and bulk electrical conductivity are typically viewed as independent measurements obtained from the TDR waveform, the two parameters can be related and used to simplify TDR measurements and make them more accurate (Yu and Drnevich, 2004). Both Malicki et al. (1994) and Hilhorst (2000) found that a good linear relationship existed betw een the apparent dielectric constant and bulk soil electrical conductivity. Also, White et al. (1994) noted a linear relationship between the apparent dielectric constant of the water phase and the square root of the bulk electrical conductivity of the water phase. Yu and Drne vich (2004) point out that Equations (2-16) and (2-25) are two indepe ndent equations that both are functions of water content and dry density of the soil. As a result, Yu and Dr nevich (2004) suggest that the apparent dielectric constant (Ka) and bulk electrical conductivity (ECb) must be dw c ECd w b

PAGE 38

26 related to one another. If Eqns. (2-16) and (2-25) ar e combined, the following is obtained: (2-26) Eqn. (2-26) can be expressed as: (2-27) Where “f” and “g” are soil specific calibration constants. Applicati on of Equation (2-27) as well as the calibration of “f” and “g” will be treated more in depth in chapter 4. Purdue One-step TDR Methodology After the TDR calibration constants are dete rmined for a particular soil the dry density ( d) and the water content can be computed in the field by solving for Eqns. (216) and (2-25). The dielectric constant (Ka) and the bulk electrical conductivity (ECb) are measured and the following equations can be used to determine field dry density and water content: (2-28) (2-29) a bK g f EC a w d bK b d b d a c b EC cb ad EC b K db a d a b b aK d EC b EC a K c w

PAGE 39

27 However, Yu and Drnevich (2004) point out that when Eqns. (2-28) and (2-29) are applied satisfactory results are typically not obtained. This is due to the dominance of pore fluid conductivity on Eqn. (2 -25). When calibration is pe rformed in the laboratory, the pore fluid conductivity measured does not generally correspond to the field condition and as a result, accurate m easurements are not obtained. Yu and Drnevich (2004) propose to “adjust ” the field sample to the laboratory calibration. Figure 2-8 is a graphic representation of the adjustment procedure. Figure 2-8. Adjusting the Field Sample to the Laboratory Calibration. Source: Yu and Drnevich (2004).

PAGE 40

28 The dotted lines represent the calibration lines obtained for different values of pore fluid conductivities. If a TDR measurement is taken in the field and lies somewhere between an ECw of 0.08 and 0.10 S/m the sample can be adjusted to any laboratory calibration line (in this case the ECw = 0.08 S/m line). This is done by replacing the same amount of pore fluid of the field sample with that of the laboratory calibration (see Figure 2-8). In this manner the adjusted sample ha s the same water content and dry density of the field sample, the only difference bei ng the pore fluid conduc tivity. The bulk electrical conductivity of the adjusted sample changes due to the replacement of the field pore fluid with the laboratory pore fluid. Howeve r, the dielectric constant of the adjusted sample is the same as the original field sample. This is due to insensitivity of the Siddiqui-Drnevich equation (Eqn. 2-16) to pore fluid conductivity. The adjustment applied to the field condition is thus a ver tical projection to the laboratory calibration line. The water content and dry density of the fi eld sample can then be calculated using Eqns. (2-28) and (2-29) with the adjusted values of Ka, adj and ECb, adj. field a adj aK K, (2-30) 2 , f a adj bK g f EC (2-31) As mentioned previously and as displa yed in Eqn. (2-30) no adjustment is necessary to the measured dielectric consta nt. Calibration values of “f” and “g” are required to determine the adjusted bulk el ectrical conductivity and as a result no measurement of pore fluid conductivity need be taken except that it remains constant during calibration. Tap water is recommende d for accurate laboratory calibration (Yu and Drnevich, 2004).

PAGE 41

29Limitations of TDR Measurement Significant research has been carried out to determine possible s ources of error in TDR measurements. Particularly problem s with “lossy” materials and multiple reflections have received attention. Mo jid et al. (2003) studied problems with geotechnical TDR measurement in lossy materi als. In highly conductive materials, a dispersive or lossy phenomenon is observed. Th ese lossy or dispersive materials are finegrained materials or conductive materials wher e large amounts of ions are present in the pore water (Mojid et al. 2003). Yanuka et al (1988) also discusse d the significant error that arises when taking TDR measurements in conductive materials. These same observations were made by Topp et al. (2000). Without equipment or method modifications, the TDR procedure discussed here in is not applicable to highly dispersive materials such as fat clays at high water co ntents and concentrated ionic solutions (Yu and Drnevich, 2004).

PAGE 42

30 CHAPTER 3 EQUIPMENT AND PROCEDURE Introduction Siddiqui and Drnevich (1995) studied th e factors which influence the wave transmission and as a result, transmissi on line components were designed and built. These components were designed to be robust easy to use, and pr ovide superior wave transmission for field measurement of the soil apparent dielectric constant. The TDR system was standardized in the form of ASTM D6780. The method outlined in ASTM D6780 is referred to as the two-step Purdue TDR method. Two steps in that the method calls for two tests to be run (one in-situ a nd one in a compaction mold) to determine the in-situ water content and density. Several lab and field investigations have been carried out by researchers at Purdue University as well as through a beta testing program designed to evaluate the accuracy of the me thod. Results indicate that the method is sufficiently accurate for geotechnical applic ations (Lin, 1999; Siddiqui et al., 2000; Drnevich et al., 2002; Sallam et al., 2004). However, the two-step method was determ ined to be somewhat limited in its application. It only made use of the soil apparent dielectric c onstant, the method was destructive (requiring the excavation of th e soil to be tested) and time consuming compared with other methods. In an effort to streamline the two-step method, Yu and Drnevich (2004) proposed the use of a bul k electrical conductivity measurement to improve the accuracy of the two-step method as well as to eliminate the need to perform a second test in a compaction mold. This im proved method is referred to as the one-step Purdue TDR method. The equipment used is essentially the same but, only one in-situ measurement is required to obtain both wate r content and density. The test is nondestructive and the need for excavation is eliminated (Yu and Drnevich, 2004).

PAGE 43

31TDR Measurement System The system configuration of the basic Pur due TDR device is show n in Figure 3-1. The TDR device used in this study was acquire d from Purdue University and its central components include a Campbell Scientific TD R 100 Time Domain Reflectometer, 3 feet of Intcom 50cable as well as soil probe equipment. The soil probe is essentially comprised of three main components: 1) the coaxial cable; 2) the co axial head (CH); and 3) either a coaxial cylinder (CC) (used for calibration purposes) or multiple rod probe (MRP) (used for field measurem ent) (Drnevich et al., 2001). The coaxial cable consists of a cente r conducting wire su rrounded by a cylinder casing, which acts as the outer conductor (Lin et al., 2000). The coaxial head (CH) (Figure 3-2) is a transition from the coaxial cable to either the CC or MRP and consists of three components: 1) a 50mating BNC connector; 2) a metal cylindrical head with an insulating material; and 3) a multiple rod sec tion that contains three perimeter rods and one center rod. The coaxial head (CH) has one center stud and three perimeter studs. The center stud and two of the perimeter studs ar e of the same length (21mm), whereas the third perimeter stud can be adjusted to ensure full contact with either the MRP or CC. Figure 3-1. TDR System Configura tion. Source: Lin et al. (2000).

PAGE 44

32 The coaxial cylinder (CC) transmission line consists of a CC mold, a ring, and a central rod all made of stainless steel (F igure 3-3). The CC mold is essentially a modified compaction mold with an inner di ameter of 102 mm and a length of 203 mm. The CC ring rests on top of the CC mold. It acts as an extension during the compaction process and as a part of the coaxial cylinder (CC) during measurement. A central rod is driven after the soil has been compacted into the mold to complete the coaxial line (the mold acts as the outer conductor). The central rod is made of stainless steel and has a diameter of 8 mm. A template is used to guide the central r od during driving stage (Drnevich et al., 2001). Figure 3-2. Configuration of Coaxia l Head. Source: TDR User’s Manual

PAGE 45

33 Other equipment relevant to the CC test includes: a tamping rod for compacting soil, a guide for installing the center rod, an acrylic hammer for driving the center rod, a metal screed for leveling the surface of the soil in mold, a brush and a digital scale (Drnevich et al., 2001) The multiple rod probe (MRP) consists of one central rod and three perimeter rods (9.5 mm diameter and a length of 236 mm ), which are driven into the soil. The configuration of the MRP rods is made to match the coaxial head (CH) by means of a detachable template. After the spikes have been driven, the template is removed and the coaxial head (CH) is placed on top of the spik es (Figure 3-4). This forms a coaxial line in the soil (Drnevich et al., 2001). Figure 3-3. The Coaxial Cyli nder (CC) Transmission Line. Source: Siddiqui and Drnevich (1995).

PAGE 46

34 Other equipment relevant to the MRP fiel d test includes: a brass hammer to drive the spikes, a tool for loosening the template and a tool for removing the spikes (Drnevich et al., 2001). Data Acquisition A laptop computer is connected to SP232 serial communication module to retrieve data obtained from the Cam pbell Scientific TDR100 Time Domain Reflectometer. Computer software named TDR++ was developed by Feng et al. (1998) to automate TDR data acquisition. A rugge d, oversized briefcase (Figure 3-5) was designed for field measurement. The briefcas e houses the CS TDR 100, a power supply for the TDR100, a 120-volt charger, a laptop computer, a charger for the laptop, a BNC cable to connect the TDR100 to the MRP head and a cable to connect the chargers to a 120 V AC source (Drnevich et al., 2001). Figure 3-4. The Multiple Rod Probe (MRP). Source: Siddi q ui and Drnevich ( 1995 ) .

PAGE 47

35 TDR Software (PMTDR-SM) The Purdue Method TDR Simplified Method (PMTDR-SM) is an updated version of the two-step software developed for ASTM D6780. PMTDR-SM Version 1.2.2 was the most current version availa ble at the time this research was conducted. Researchers continue to update and improve the software The software was designed with a userfriendly interface. It consists of two input sc reens. The first screen is the In-Situ MRP Test which prompts the user to input project name, contract No., operator, test location, test number, temperature, and type of soil (c ohesive or cohesionless) (Figure 3-6). Other input parameters include the MRP probe configuration measurements and the soil specific constants for the soil being tested. The software obtains a TDR waveform from the aforementioned configurati on by using the “Get Wavefo rm” command. A TDR wave analysis can then be performed by using the “S tart” function. The analysis is consistent with that discussed in Chapter 2 and will determine the in-situ apparent dielectric constant (Ka) and the in-situ bulk electrical conductivity (ECb). The gravimetric water content and the dry density can then be co mputed with the inputted soil specific Figure 3-5. TDR Field Measurement Case. Source: TDR User’s Manual.

PAGE 48

36 calibration constants using the “Compute” function. Test resu lts can then be saved using the “Save Test Results” command. The second screen is used for the CC Mold Test in which the user is prompted to input the same parameters mentioned for the In-situ MRP test as well as the mass of empty mold, mass of mold, wet soil and the volume of mold and the mold probe dimensions (Figure 3-7). The waveform is obtained in the same manner mentioned previously and the wave analysis is also performed in the same manner. The mold moisture content and dry dens ity can then be calculated. Figure 3-6. In-situ MRP Input Screen. Source: PMTDR-SM.

PAGE 49

37 Testing Procedure Testing includes two parts: 1) a soil specific calibration is performed in the lab prior to field testing and 2) field tests are carried out with the calibration constants obtained in part (1). Calibration Calibration should be carried out in a sim ilar fashion as the procedure outlined in ASTM D6780 – Annex 2. The laboratory ca libration can be perf ormed in conjunction with ASTM D698 and ASTM D1557 in order to determine the opt imum water content and the maximum dry density for field compacti on control. Typical calibration requires Figure 3-7. CC Mold Test Input Screen. Source: PMTDR-SM.

PAGE 50

38 that at least 3 or 4 tests be performed at different water contents that encompass the expected field conditions. The calibration te sts are performed in the cylindrical mold (CC). Proper calibration will yield the soil spec ific calibration constants (a, b, c, d, f and g). Calibration should be carried out with soil temperatures in the range of 15C to 25C. The calibration phase can be summ arized by the following steps: 1) Determine the length of the central rod (typically 0.263m), the volume of the mold (typically 1888 cm3) and the mass of the empty mold (typically 4,380 g). These values should be consistent with those outlined in ASTM D6780. 2) Obtain a representative soil sample from the testing site. 3) Air-dry the soil sample. 4) Prepare an adequate number of specimens having water contents that encompass the expected range of values anticipated in the field. Water contents should vary by about 2-3%. 5) Place the soil into the cylindrical mold in six uniform lifts applying ten blows per lift with the supplied aluminum tamping rod (Figure 3-8). Figure 3-8. Compaction by Tamping with Aluminum Rod.

PAGE 51

39 6) Remove the ring collar a nd level the soil surface. 7) Remove any excess soil from the mold a nd then weigh the mold with the wet soil (Figure 3-9). 8) Place the driving template over the mold and drive the central spike (Figure 3-10). 9) Replace the ring collar and place the coaxia l head such that all four legs are in contact with the mold or central spike. 10) Obtain a TDR wave using the Purdue TDR measurement system and software (Figure 3-11). Figure 3-9. Mold and Wet Sample Being Weighed.

PAGE 52

40 Figure 3-10. Central Spike Being Driven Through the Guide and into the Sample. Figure 3-11. Taking the TDR Measurement.

PAGE 53

41 11) Remove the soil from the mold and retain a sufficient sample to obtain the water content in accordance with ASTM D 2216. 12) Repeat steps (5) through (11) until th e desired number of measurements are obtained. For each of the specimens the water content (from the oven dry method), dry density (from mold and wet soil measur ement), apparent dielectric constant (Ka), and bulk electrical conductivity (ECb) can be calculated. Specific constants for the soil being calibrated can be obtained through a se ries of linear regression plots. Soil constants “a” and “b ” are found by plotting oven d w aw vs K where Ka is the measured apparent dielectric constant, d is the dry density of the soil in the mold, w is the density of water (1000 kg/m3), and woven is the oven dry water content. A best fit linear line is then fitted to the data and “a” is the zero intercept of the line and “b” is the slope of the line. Figure 3-12 is an example calibration for Ottawa Sand. y = 9.7123x + 0.9531 R2 = 0.99520.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 0.0%1.0%2.0%3.0%4.0%5.0%6.0%7.0%Water Content (%) Ka* w/ d Figure3 12.Exampleof “ a ” and “ b ” CalibrationforOttawaSand.

PAGE 54

42 Soil constants “c” and “d” are found by plotting oven d w bw vs EC where ECb is the measured bulk electrical conductivity, d is the dry density of the soil in the mold, w is the density of water (1000 kg/m3), and woven is the oven dry water content. A best fit straight line is then fitted to the data and “c” is the zero inte rcept of the line and “d” is the slope of the line. Figure 3-13 is an example calibration for Ottawa Sand. Soil constants “f” and “g” are found by plotting a bK vs EC where ECb is the measured bulk electrical conductivity and Ka is the apparent dielectric constant. A best fit straight line is then fitted to the data a nd “f” is the zero intercept of the line and “g” is the slope of the line. Figure 3-14 is an example calibration for Ottawa Sand. In order to facilitate the calibration process PMTDR-SM has incorporated a calibration tool (Figure 3-15). The data points obtained from calibration are inputted into the appropriate fields and the tool will automatically generate the desired calibration plots. The calibration results from bet fit regression line analysis are then displayed. The soil specific constants can then be uploaded in to either the field testing input screen (Figure 3-6) or the mold input screen (3-7) for later use. y = 0.1826x + 0.01 R2 = 0.97230.00 0.01 0.01 0.02 0.02 0.03 0.0%1.0%2.0%3.0%4.0%5.0%6.0%7.0%Water Content (%) ECb* w/ d Figure 3-13. Example of “c” and “d” Calibration for Ottawa Sand.

PAGE 55

43 Figure 3-14. Example of “f” and “g” Calibration for Ottawa Sand. y = 0.0194x 0.0134 R2 = 0.98340.00 0.01 0.01 0.02 0.02 0.03 0.03 0.04 1.401.601.802.002.202.40 Ka ECb Figure 3-15. PMTDR-SM Calibration T ool. Source: PMTDR-SM Software.

PAGE 56

44 The entire calibration process requires at least 24 hours to perform. The calibration constants can then be used for th e same soil encountered in the field under a wide variety of in-situ conditions. It may be possible to catalog commonly encountered soils and their constants to avoid the calibration process (Yu and Drnevich, 2004). In-situ Testing Field testing is similar to the proce dure outlined in ASTM D6780. However, there is no need to remove the soil for a compaction mold measurement. Field testing should be carried out with the soil specif ic calibration constants obtained in the calibration phase. The in-situ testing procedure can be summarized by the following steps: 1) Level the soil surface that is to be tested If the soil surface has been exposed such that it is dried out or wet from recent rain, the top in ch of soil may be removed. The leveled surface should be free of voids and if they are present, they should be filled. 2) Place the driving template on the leveled so il surface, making sure the locking pin is in place. Ensure that the soil is in full contact w ith the template. 3) Drive the three outer spikes through the template and into the soil with the brass hammer (Figure 3-16). The central spike shou ld be driven last. Ensure that all spikes are touching the template. 4) Remove the locking pin from the templa te. Spread the template apart and remove it (Figure 3-17). 5) Place the coaxial head (CH) on top of the driven spikes so that each spike makes contact with the CH. It may be necessary to slide or rotate the CH to ensure good contact with the spikes (Figure 3-18).

PAGE 57

45 Figure 3-16. Driving Spikes th rough Template into Soil. Source: TDR Manual. Figure 3-17. Removal of the Template. Source: TDR Manual.

PAGE 58

46 6) The coaxial head (CH) is then connected to the CS TDR100 by the provided coaxial cable. 7) Obtain a TDR measurement using the TD R measurement system and software. The field testing procedure requires only a few minutes of setup and about 2 to 3 minutes to run the test. This is a significan t reduction in the time required to run the test than the two-step method and is comparable to the nuclear test. After the equipment is setup, several tests can be performed in succession in a relatively short amount of time (Yu and Drnevich, 2004). Figure 3-18. Placement of Coaxial Head (CH) on Spikes. Source: TDR Manual.

PAGE 59

47 CHAPTER 4 EVALUATION OF PURDUE TDR SOIL PARAMETERS Introduction For accurate Purdue one-step TDR field m easurement, a series of soil specific calibration constants must be obtained prev iously. Soil constants “a” and “b” are parameters that relate the gravimetric moistu re content to the soil dielectric constant. Constants “c” and “d” relate the gravimetri c moisture content to the bulk electrical conductivity of the soil. Finally, constants “f” an d “g” relate the dielec tric constant to the bulk electrical conductivity. The soil cons tants described can va ry widely with soil composition and site specific conditions. A critic al factor affecting calibration is the pore fluid conductivity of the soil. This chapter pr esents experimental results from a study on the effect of initial salinity on the calibration constants. Also an effort is made to determine the typical range of TDR constant s for Florida sand by performing a series of TDR tests in the calibration mold for several soils obtained at loca l construction projects in the vicinity of Tampa. Calibration Constants “a” and “b” The Siddiqui and Drnevich (1995) relations hip (Eqn. 2-16) is used to determine soil constants “a” and “b.” The process for obtaining calibration constants “a” and “b” was previously discussed in Chapter 3. Substituting the volumetric water content into Eqn. (2-16), the following relationship is obtained: (4-1) b a Kw d a

PAGE 60

48 When the volumetric water content ( ) is zero: d w s aK a (4-2) Soil constant “a” is thus termed the refraction index of the soil solids that is normalized by the soil dry density. Typical va lues of “a” range from 0.7 to 1.85 (Yu and Drnevich, 2004). When the volumetric water content ( ) is 100 percent: (4-3) Soil constant “b” is defined as the refr action index of the pore fluid. Typical values of Ka,w measured by TDR are close to 81 at 20C. This yields a “b” value of about 9 (Yu and Drnevich, 2004). Sallam et al. (2004) determined the calibration parameters “a” and “b” encountered in common soils in the state of Florida. A fi nal recommendation of “a” = 1 and “b” = 8.5 was made (Figure 4-1). Soils te sted in the study were mainly Florida sands that can be characterized as A3 or A-1-b soils (Table 4-1). Fi g ure 4-1. Final Calibration Results for “a” and “b.” Source: Sallam et al. ( 2004 ) Kax w/ d = 8.48 Wc + 0.99 R2 = 0.98260.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 0.000.050.100.150.200.25Wc Kax w/ d w aK b,

PAGE 61

49 Note – Test 1 was discarded for incorrect testing procedures. Table 4-1. Values of Constants “a” and “b” for Various Sands. Source: Sallam et al., (2004). TestDescriptionOperatorUSCSAASHTOabComment 1Ottawa SandAmrSPA-1-b1.2211.68Discarded 1-aOttawa SandBrianSPA-1-b0.959.00Accepted 1-bOttawa SandBothSPA-1-b0.919.41Accepted Average 0.939.21 2Outside LabAmrSPA-31.008.20Accepted 2-aOutside LabBrianSPA-31.038.35Accepted Average 1.028.28 3MP-1AmrSPA-1-b0.938.78Accepted 3-aMP-1BrianSPA-1-b1.017.48Accepted 3-bMP-1BrianSPA-1-b0.988.21Accepted Average 0.978.16 4Sample # 515BothSPA-31.058.19Accepted 4-aSample # 515BrianSPA-31.018.93Accepted 4-aSample # 515BrianSPA-31.038.73Accepted Average 1.038.62 5Sample # 2AmrSPA-1-b1.107.40Accepted 5-aSample # 2BothSPA-1-b1.048.06Accepted Average 1.077.73 6Sample # 6944AmrSPA-31.088.09Accepted 6-aSample # 6944BrianSPA-30.998.65Accepted Average 1.048.37 7Sample with # 6944BrianSPA-1-b 0.998.80 Accepted 8Sample # 6965BothSWA-1-b0.998.80Study the effect of compaction 8-aSample # 6965BothSWA-1-b1.048.03Study the effect of compaction 8-bSample # 6965BothSWA-1-b0.998.31Study the effect of compaction 8-cSample # 6965BothSWA-1-b1.007.96Study the effect of compaction Average 1.018.28 9Sample With # 6965BrianSPA-1-b 1.028.20 Accepted 10Sample # 6974BrianSPA-1-b 0.9 9 8.2 7 Accepted 11Sample # 6978BrianSPA-1-b 1.027.9 3 Accepted 12Sample # 6926BrianSPA-1-b0.909.24Accepted 13Sample # 6927BrianS P A-1-b0.879.83Accepted

PAGE 62

50 Sallam et al. (2004) performed a series of tests to study the fact ors influencing the calibration of “a” and “b.” It was conclude d that the compaction energy only slightly affects the value of constants “a” and “b” (F igure 4-2). Operator dependency was also found to have an insignificant in fluence on testing results. The effect of variation or inaccuracy in determining “a” and “b” on field measurement of water content and dry density was also studied. The effect of changing constant “b” was more critical than the eff ect of changing constant “a” in water content measurement. The predicted value of the mo isture content changes noticeably as “b” varies, especially when the actual value is su rpassed by 1.0 (Table 4-2). It was also noted that at higher Ka values the error in predicting the moisture content increased. This is expected since a change in the predicted mo isture content resulti ng from a variation in “b” (which is the slope of the straight line) is likely dramatic. Figure 4-2. Effect of Compaction Energy on Constants “a” and “b.” Source: Sallam et al., (2004). Case 1 y = 8.6565x + 1 Case 2 y = 8.3668x + 1 Case 3 y = 8.2424x + 1 Case 4 y = 8.1448x + 11 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 00.020.040.060.080.10.120.140.160.18Wc (Ka)* w/ d

PAGE 63

51 Variability in both constants “a” and “b ” was determined to have a significant effect on dry density measurement (Tables 43 and 4-4). Also, inaccurate water content measurement added significant error in dry density measurement (S allam et al., 2004). -0.2-0.15-0.1-0.0500.050.10.150.2 0.02-2.60-1.96-1.32-0.660.000.671.352.042.74 0.05-2.60-1.96-1.32-0.660.000.671.352.042.74 0.1-2.60-1.96-1.32-0.660.000.671.352.042.74 0.15-2.60-1.96-1.32-0.660.000.671.352.042.74 0.2-2.60-1.96-1.32-0.660.000.671.352.042.74 0.25-2.60-1.96-1.32-0.660.000.671.352.042.74 Wc a d/ d,true %ge Table 4-3. Error Resulting from Varia tion of “a” on Predicted Dry Density. Source: Sallam et al. ( 2004 ) .-1-0.75-0.5-0.2500.250.50.751 0.020.310.230.150.070.00-0.07-0.13-0.19-0.24 0.050.810.590.380.180.00-0.17-0.33-0.48-0.61 0.11.721.240.790.380.00-0.35-0.68-0.99-1.28 0.152.721.951.250.600.00-0.55-1.07-1.55-1.99 0.23.812.731.740.830.00-0.77-1.48-2.14-2.76 0.255.003.572.271.090.00-1.00-1.92-2.78-3.57 Wc b Wc, %ge Table 4-2. Error Resulting from Variati on of “b” on Predicted Moisture Content. Source: Sallam et al., (2004).

PAGE 64

52 Calibration Constants “c” and “d” Equation (2-25) is used in conjunction w ith the procedure outlined in Chapter 3 to determine constants “c” and “d.” If Eqn. (225) is expressed in terms of volumetric water content the following expression is obtained: d c ECw d b (4-4) When the volumetric water content is zero: (4-5) Soil constant “c” is related to surface conduc tance of the soil particles normalized by dry density (Yu and Drnevich, 2004). T ypical values for “c” have not yet been established. However, data taken from Sallam et al., (2004) using the ASTM TDR method was uploaded into the Purdue one-step software. Values of “c” ranged between -1-0.75-0.5-0.2500.250.50.751 0.02-0.31-0.22-0.14-0.070.000.060.120.180.24 0.05-0.77-0.56-0.36-0.170.000.160.310.450.59 0.1-1.54-1.11-0.71-0.340.000.320.630.911.18 0.15-2.31-1.67-1.07-0.520.000.480.941.361.76 0.2-3.08-2.22-1.43-0.690.000.651.251.822.35 0.25-3.85-2.78-1.79-0.860.000.811.562.272.94 Wc b d/ d,true %ge Table 4-4. Error Resulting from Variati on of “b” on Predicted Dry Density. Source: Sallam et al., (2004). s d wEC c

PAGE 65

53 0.0051 and 0.0359 for Florida sands. Further di scussion of the value of “c” is addressed later in this chapter. When the volumetric water content is 100 percent: (4-6) Eqn. (4-6) expressed in terms of Rhoades equatio n (Eqn. 2-24): (4-7) Soil constant “d” accounts for the effect of soil type and por e fluid properties (Yu and Drnevich, 2004). Again, no typical range of values has been established for “d.” This is predominantly due to the dependence of “d” on th e pore fluid conductivity of the soil being tested. This phenomenon is addresse d later in the chapter. Values obtained for Florida sands ranged from 0.146 to 0.671. Calibration Constants “f” and “g” Equation (2-27) is used with the calibra tion procedure outline d in Chapter 3 to determine soil constants “f” and “g.” Eqn. (2-27) can be expressed as: (4-8) Constant “g” is a function of “d” and “b” and is theref ore related to pore fluid properties and is largely dependant on the pore fluid conductivity. Typical values for both “f” and “g” have not been established. For Florida sands, values of “f” range from 0.0873 to 0.0157, and values of “g” ranged from 0.021 to 0.0754. A summary of Purdue one-step TDR constants for Florida sa nds is displayed in Table 4-5. bEC d wEC a d' a w d bK b d b d a c b EC

PAGE 66

54 When the volumetric water cont ent is zero, the value of Ka,s can be extracted from Eqn. (4-2). This value should remain consta nt for a given soil as long as a standardized compaction procedure is followed. Similarly, the value of ECb,s can be obtained from Eqn. (4-5). Again, assuming that the variation of dry density is small, the value of ECb,s should remain constant. If both Ka,s and ECb,s remain constant with a consistent compaction procedure, a unique point may be defined at zero water content through which all calibration plots for constants “f” and “g” must pass for a given soil. This concept is discussed and demonstrat ed in the subsequent section. SampleSoil Typeabcdfg #6974 brSP or A-1-b0.9868.280.01990.326-0.02860.0385 #6926 NKWSP or A-1-b18.830.00560.632-0.08730.0681 #6926 brSP or A-1-b0.9299.130.00640.517-0.05820.0526 #6978 NKWSP or A-1-b0.9698.80.01640.486-0.05320.0549 #6978 brSP or A-1-b1.018.090.02170.223-0.00810.027 #6944 brSP or A-1-b0.9878.820.00910.485-0.05780.0521 #6965 brSP or A-1-b1.028.220.01710.342-0.03720.0402 #6965 Amr/Br 1SP or A-1-b0.9828.850.00730.464-0.04880.0462 #6965 Amr/Br 2SP or A-1-b1.018.430.00860.471-0.05640.0498 #6965 Amr/Br 3SP or A-1-b0.9938.340.01050.437-0.04990.0469 #6965 Amr/Br 4SP or A-1-b0.9938.390.00780.502-0.05770.0505 #6927 brSP or A-1-b0.8759.880.01230.446-0.04480.0462 #515 6-11 brSP or A-31.038.760.03590.671-0.06350.0754 #515 5-28 brSP or A-31.028.920.01910.367-0.03310.0404 #515 5-22-02 Amr/BrSP or A-31.048.370.01790.308-0.02940.0359 #mp-1br 5-21-02SP or A-1-b17.480.01550.217-0.0220.0291 #mp-1br 6-11-02SP or A-1-b0.9838.150.02180.397-0.04280.0494 #mp-1 amr 5-15-2002SP or A-1-b0.9617.940.00510.258-0.04120.0324 Clayey Sand brSP or A-31.058.240.01160.236-0.02530.0281 Clayey Sand AmrSP or A-31.038.270.01170.235-0.02120.0266 Beach Sand BrSP or A-1-b0.969.060.00880.348-0.0490.0406 Beach Sand AmrSP or A-1-b0.9588.510.02210.215-0.00220.0248 Beach Sand Amr/BrSP or A-1-b0.87710.150.0090.386-0.03940.038 #2 Amr 6-11-02SP or A-1-b1.087.490.03340.1460.01570.021 #2 Amr/Br 5-23-02SP or A-1-b1.038.220.02340.391-0.02640.0426 #6944 Amr 6-10-02SP or A-31.078.270.02430.294-0.02030.0354 Max1.0810.150.03590.6710.01570.0754 Min0.8757.480.00510.146-0.08730.021 Averages0.9948.5340.0150.377-0.0380.042 Table 4-5. Purdue TDR C onstants for Florida Sands.

PAGE 67

55The Effect of Pore Fluid Conducti vity on Calibration Constants In an effort to better characterize th e effect of pore fl uid conductivity on the Purdue one-step TDR calibration constants, se veral calibrations were carried out using varying pore fluid conductivities on three different sands. Ottawa Sand and two local materials (Sample #6978 and Florida Sand #2) were selected for testing. Effect on Constants “a” and “b” The values of constant “b” were recorded for every pore fluid conductivity tested and then plotted for each soil. Results ar e displayed in Figures 4-3, 4-4 and 4-5. 8.6 8.8 9.0 9.2 9.4 9.6 024681012ECw (mS/cm)b Figure 4-4. Sample #6978 Variation of “b” with Pore Fluid Conductivity. 8.0 9.0 10.0 11.0 12.0 13.0 0102030405060708090ECw (mS/cm)b Figure 4-3. Ottawa Sand Variation of “b” with Pore Fluid Conductivity.

PAGE 68

56 From Figures 4-3, 4-4 and 4-5 it appears that no defi nite relationship exists between pore fluid conductivity and constant “b.” Also, constant “a” is virtually unaffected by changes in pore fluid conductivity. This observation is consistent with past research that has indicated that pore fluid conductivity has little effect on calibration for the dielectric constant in common soils (D alton, 1982; Topp et al., 1988; White et al., 1994). Effect on Constants “c” and “d” The values of constant “d” were recorded for every pore fluid conductivity tested and then plotted for each soil. Results ar e displayed in Figures 4-6, 4-7 and 4-8. 0.0 0.5 1.0 1.5 2.0 2.5 020406080100ECw (mS/cm)d Figure 4-6. Ottawa Sand Variation of “d” with Pore Fluid Conductivity. 8.8 8.8 8.9 8.9 9.0 9.0 9.1 0123456ECw (mS/cm)b Figure 4-5. Fl Sand #2 Variation of “b” with Pore Fluid Conductivity.

PAGE 69

57 Figures 4-6, 4-7 and 4-8 show that a relationship exists between pore fluid conductivity and constant “d.” Constant “d” increases with pore fluid conductivity. The relationship appears to assume a parabolic form. It has been noted previously that pore fluid conductivity has a large effect on the calib ration of soil bulk elec trical conductivity. Data from Amenta et al. (2000) demonstrates the effect of pore fluid conductivity on bulk electrical conductivity ca libration (Figure 4-9). 0.0 0.2 0.4 0.6 0.8 0123456ECw (mS/cm)d Figure 4-8. Fl Sand #2 Variation of “d” with Pore Fluid Conductivity. 0.0 0.2 0.4 0.6 0.8 1.0 1.2 024681012ECw (mS/cm)d Figure 4-7. Sample #6978 Variation of “d” with Pore Fluid Conductivity.

PAGE 70

58 As displayed in Figure 4-9 the slope of the calibration line (constant “d”) systematically increases with an increase in pore fluid conductivity. This is consistent with the previous discussion of the domina nce of the pore fluid on bulk electrical conductivity measurement. Although pore fluid conductivity largely governs the calibration of “d,” constant “c” should be affect ed to a lesser degree. This is seen in Figure 4-9 where the y-intercept (constant “c”) varies little. It appears that calibration lines pivot around constant “c” at the y-axis. Further discussion as to the significance of constant “c” is addressed later in this chapter. Effect on Constants “f” and “g” The values of constant “g ” were recorded for every po re fluid conductivity tested and then plotted for each soil. Results ar e displayed in Figures 4-10, 4-11 and 4-12. Figure 4-9. Relationship between ECb and Gravimetric Water Content. Source: Amenta et al. ( 2000 ) 0 0.05 0.1 0.15 0.2 0.25 00.050.10.150.20.250.3Gravimetric Water Content ECb* w/ d ECw=0.56 S/m ECw=0.40 S/m ECw=0.37 S/m ECw=0.30 S/m ECw=0.24 S/m ECw=0.21 S/m ECw=0.15 S/m ECw=0.10 S/m

PAGE 71

59 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0123456ECw (mS/cm)g Figure 4-12. Fl Sand #2 Variation of “g” with Pore Fluid Conductivity. 0.00 0.02 0.04 0.06 0.08 0.10 0.12 024681012ECw (mS/cm)g Figure 4-11. Sample #6978 Variation of “g” with Pore Fluid conductivity. 0.00 0.05 0.10 0.15 0.20 020406080100ECw (mS/cm)g Figure 4-10. Ottawa Sand Variation of “g” with Pore Fluid Conductivity.

PAGE 72

60 Figures 4-10, 4-11 and 4-12 disp lay a similar trend to that seen with constant “d.” This is expected, as constant “g” is dependa nt on constant “d.” Constant “f” varies widely as constant “g” changes with pore fluid conductivity. This is due to the nature of the bulk electrical conductivity and dielectric constant calibration. Calibration lines pass through a point that is define d by the dry calibration point, wh ich is not located on the yaxis. As a result a change in slope will significantly affect the y-intercept. This concept will be further developed in the subsequent section. Summary Although there are several possible calib rations for “c,” “d,” “f” and “g” depending on pore fluid conductivity, Yu and Drnevich (2004) point out that only one calibration is needed to adjust the field m easurement to the calibration condition. The assumption is that the pore fluid conductivit y remains constant during calibration. Calibration with tap water is recommende d for accurate parameter evaluation. Effect of Initial Salt Content on Calibration Constants If a soil contains an appreciable amount of salts, it is reasonable to assume that the pore fluid conductivity does not re main constant during the ca libration phase. With the addition of tap water to a salt y soil, the salt will dissolve into the tap water and increase the pore fluid conductivity. As more water is added, the pore fluid conductivity will decrease as the dissolved salt is diluted. Th e system will eventually reach a constant pore fluid conductivity at higher water contents. Under these conditions it is clear that the pore fluid conductivity will not be constant dur ing calibration and will not yield the true calibration constants. Initial Salt Content Testing In an effort to demonstrate th e effect of initial salt cont ent on the calibration of the Purdue one-step method, a series of tests were carried out on three soil types (Table 4-6). Ottawa Sand and two local sand samples were us ed with varying amounts of fines. Each soil was prepared by cleaning the sample w ith deionized water until a low initial pore

PAGE 73

61 fluid conductivity was obtained. For the thre e samples tested, the pore fluid conductivity values after washing are displayed in Table 4-6. After the samples had been washed, they were calibrated with water at several different pore fluid conductiv ities. These calibrations were termed “true” calibrations. The samples were then calibrated with an initial salt concentration by either soaking the sand in a known pore fluid conductiv ity or by testing the soil in its natural condition. The samples were then calibrated with tap water and compared to the true calibrations. The testing procedure can be summarized as follows: 1) Add deionized water to the sample and vi gorously mix the soil with the deionized water (Figure 4-13). 2) Allow the sample to stabilize by allowing the fine-grained particles to settle (time varies depending on the amount of fines contained in the soil). 3) Drain the excess water from the sample. Fo r Ottawa Sand (low fines content) the sample was placed in a geofabric and drained (Figure 4-14) For the other samples (higher fines content) the exce ss water was siphoned from the sample after all fines had settled. 4) Repeat steps 1) through 3) several times until the pore fluid conductivity reaches an acceptable level (washed values are recorded in Table 4-6). SoilClassification% FinesECw Cleaned Ottawa SandSP or A-1-b0< 10 S/cm #6978SP or A-1-b1.36< 35 S/cm Fl Sand #2SP or A-31.82< 45 S/cm Table 4-6. Testing Material Summary.

PAGE 74

62 Figure 4-14. Draining Water from Sample. Figure 4-13 Mixing Soil with Deionized Water.

PAGE 75

63 5) Perform calibration outlined in Chapter 3 to obtain soil. Calibration should be performed at several known pore fluid conductivities to obtain the “true” calibrations. After each calibration the sample should be washed using the procedure outlined in steps 1) through 4). 6) Add a known amount of salt content to the sample and then perform a calibration using tap water. Results and Discussion for “a” and “b” Each of the three sands was calibrate d at several diffe rent pore fluid conductivities ranging fro m deionized water (ECw = 0 mS/cm) to 79 mS/cm. For Ottawa Sand and Sample #6978 a sample was soaked in 2 mS/cm water, dried and calibrated with tap water. For Florida Sand #2, a sample was calibrated in its natural condition with tap water. Figure 4-15 displays calibration results for “a” and “b” for Ottawa Sand, Figure 4-16 displays results for Sample #6978 and Figure 4-17 di splays results for Florida Sand #2. The open diamonds repres ent the calibration that was performed with an initial amount of salt. The solid lines represent the true calib rations obtained after washing the sample. Fi g ure 4-15. Ottawa San d Calibrationfor “a” and “b.” 0.80 0.90 1.00 1.10 1.20 1.30 1.40 1.50 1.60 1.70 1.80 0.0%1.0%2.0%3.0%4.0%5.0%6.0%7.0%8.0%Water Content (%) Ka* w/ d Ecw=0 mS/cm Ecw = 0.103 mS/cm Ecw=0.300 mS/cm Ecw=0.590 mS/cm tap Ecw=51mS/cm Ecw=79mS/cm Ecw=30mS/cm Ecw=10mS/cm 5mS/cm 2mS dried

PAGE 76

64 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 0%2%4%6%8%10%12%14%16%Water Content (%) Ka* w/ d ECw = natural w tap ECw = 0 mS/cm ECw = 5 mS/cm ECw = 1 mS/cm Figure 4-17. Fl Sand #2 Calibration for “a” and “b.” Figure 4-16. Sample #6978 Calibration for “a” and “b.” 0.80 1.00 1.20 1.40 1.60 1.80 2.00 2.20 2.40 0.0%2.0%4.0%6.0%8.0%10.0%12.0%14.0%Water Content (%) Ka* w/ d ECw = 0 mS/cm ECw = 1 mS/cm ECw = 5mS/cm ECw = 10 mS/cm ECw = 2 mS/cm dried

PAGE 77

65 Results from Figures 4-15, 4-16 and 4-17 show that the true calibrations are similar to the calibrations that contained an initial amount of salt, indicating that initial salt content has not effect the calibration of soil constant “a” and “b.” This was expected as it has been shown that pore fluid conductiv ity has little to no effect on dielectric constant calibration for comm on geotechnical materials. Results and Discussion for “c” and “d” Figure 4-18 displays calibr ation results for “c” and “d” for Ottawa Sand. Figure 4-19 displays results for Sample #6978, and Fi gure 4-20 displays results for Florida Sand #2. The open diamonds represent the calibra tion that was performed with an initial amount of salt. The solid lines represent th e true calibrations obtained after washing the sample. Figure 4-18. Ottawa Sand Calibration for “c” and “d.” 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.0%1.0%2.0%3.0%4.0%5.0%6.0%7.0%8.0%9.0%Water Content (%) ECb* w/ d Ecw=0mS/cm Ecw = 0.103 mS/cm Ecw=0.300 mS/cm Ecw=0.590 mS/cm Ecw=51mS/cm Ecw=79 mS/cm Ecw=30mS/cm Ecw=10mS/cm ECw=5mS/cm 2mS/cm dried

PAGE 78

66 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.0%2.0%4.0%6.0%8.0%10.0%12.0%14.0%16.0%Water Content (%) ECb* w/ d ECw = natural w tap ECw = 0 mS/cm ECw = 5 mS/cm ECw = 1 mS/cm Figure 4-20. Fl Sand #2 Calibration for “c” and “d.” Figure 4-19. Sample #6978 Calibration for “c” and “d.” 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.0%2.0%4.0%6.0%8.0%10.0%12.0%14.0%Water Content (%) ECb* w/ d ECw = 0mS/cm ECw = 1mS/cm ECw = 5mS/cm ECw = 10 mS/cm ECw = 2mS/cm dried

PAGE 79

67 In Figures 4-18, 4-19 and 4-20 the slopes of the true calibration lines vary systematically with pore flui d conductivity as demonstrat ed by Amenta et al. (2000) Figure 4-9). The open diamonds indicate the ca libration points that were determined with an initially salty sample. Clearly these po ints demonstrate that a calibration obtained with an initially salty soil will not yield th e true calibration constants; rather the slope (constant “d”) will be less than the true value and the intercept (constant “c”) will be slightly higher than the true value. This can be explained by the initial amount of salt dissolving into the tap water and increasing the pore fluid conductivity. This causes the calibration to follow a higher po re fluid conductivity calibration. As more tap water is added, the pore fluid conductivity will decrea se and the calibration will follow a lower calibration line. This process will continue until the pore fluid conductivity stabilizes or in this case reaches that of the tap wate r being added. The calibration obtained is therefore not a true calibration, but a calibration that is bein g subjected to a constantly changing pore fluid conductivity. Such a ca libration would not be valid under the requirement set forth by Yu and Drnevich ( 2004) that a constant pore fluid conductivity be used during calibration. Constant “c” varies only slightly for each calibration regardless of the pore fluid conductivity of water that is used for calibration. The tr ue calibrations obtained all passed through or near this point All true calibrated values of “c” were close to 0.01 for each of the three sands tested. Constant “c” may be unique for sands similar to those tested and will vary only slightly with dry de nsity according to Eqn. (4-5). This indicates that the electrical condu ctivity of the soil soli d particles for sands is essentially constant and does not depend on an initial amount of sa lt contained within th e sample or the pore fluid conductivity used to calibrate the sample since the soil is dry. Results and Discussion for “f” and “g” Figure 4-21 displays Ottawa Sand calibration results for “c” and “d.” Figure 4-22 displays results for Sample #6978, and Figure 4-23 displays results for Florida Sand #2. The open diamonds repres ent the calibration that was performed with an initial amount of salt. The solid lines represent the true calib rations obtained after washing the sample.

PAGE 80

68 0.000 0.050 0.100 0.150 0.200 0.250 1.401.902.402.903.403.90 Ka ECb ECw = 0 mS/cm ECw = 1 mS/cm ECw = 5 mS/cm ECw = 10 mS/cm ECw = 2mS/cm dried Figure 4-22. Sample #6978 Calibration for “f” and “g.” Figure 4-21. Ottawa Sand Calibration for “f” and “g.” 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 1.401.601.802.002.202.402.60 Ka ECb Ecw=0 mS/cm Ecw = 0.103 mS/cm Ecw=0.300 mS/cm Ecw=0.590 mS/cm tap Ecw=51mS/cm sea Ecw=79mS/cm Ecw=30mS/cm Ecw=10mS/cm Ecw=5mS/cm 2mS/cm dried

PAGE 81

69 Figures 4-21, 4-22 and 4-23 yield results th at are similar to those obtained for “c” and “d.” Soil constant “g” varies systema tically with pore flui d conductivity as does constant “d.” This was expected as “g” is a function of “d” and “b.” Since “b” is largely unaffected by pore fluid conductiv ity, the variation of “g” wi th pore fluid conductivity is attributed to changes in “d.” The open diamonds indicate th e calibration points that were determined with an initially salty sample. Results indicate that constant “g” will be smaller than the true value and that “f” will be slightly higher than the true value. Since constant “g” varies closely with constant “d,” similar behavior in te rms of the effect of initial salt content is expected. Figures 4-21, 4-22 and 4-23 show that the true calibration values for “f” and “g” pass through virtually the same point at the dr y condition. It appear s that this point will be unique for a given soil that will vary slightly with dry de nsity. This indicates that the bulk electrical conductivity of th e soil solids is not affected by the presence of salt under dry conditions. 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 1.41.92.42.93.43.9 Ka ECb ECw = natural w tap ECw = 0 mS/cm ECw = 5 mS/cm ECw = 1 mS/cm Figure 4-23. Fl Sand #2 Calibration for “f” and “g.”

PAGE 82

70Conclusions Calibrations obtained for the dielectric constant (“a” and “b”) are unaffected by initial salt content. Calibrations obtained for electrical conductivity are affected by the initial salt content according to amount present in the sample. True calibrations that are unaffected by initial salt conten t can be obtained by washing the sample with deionized water and then calibrating the sample. Calibration points obtained at zero wate r content may be true calibration points regardless of salt content. Constant “c” ma y be a true calibration point regardless of initial salt content and por e fluid conductivity. Also the dry condition on the bulk electrical conductivity a nd dielectric constant plot may be a unique point regardless of pore fluid conductivity and initial salt content. If calibration constant “c” is a unique value for a give n soil it may be possible to extract a true calibration from this point. This may be done by obtaining the remainder of the calibration points at high water contents where the pore fluid conductivity can be assumed to be constant as the solution will be dilute. The calibra tion points along with the zero water content point may be used to extract the true calib ration slope (constant “d”). Another option may be to generate a calibration line with only constant “c” being known. Knowing that the value of “d” varies systematically with pore fluid conductivity, an arbitrary slope can be selected within a typical range of values. The pore fluid conductivity of such a calibration would not be known; however it would be a true calibration line as it passes through the zero water content intercept point. If a unique point can be defined at the zero water cont ent point for the “f” and “g” calibration a similar process may be used to extract a true calibration line from the known dry point. This would require a back calculation of the value of “g” from the arbitrarily selected value of “d” and a previously calibrated value of constant “b.” The validity of any such extraction of true calibration lines from know n dry points requires further validation. However, such a process would eliminate th e need to wash a soil to obtain a true calibration that is unaffected by initial salt concentration.

PAGE 83

71Summary Soil constants “a” and “b” are unaffected by the pore fluid c onductivity of soil. Constant “d” varies systemati cally with pore fluid conductiv ity as does related constant “g.” Both constants increase ra pidly at lower pore fluid conductivities and then appear to increase towards an asymptote at a slower ra te. Constant “c” may be a unique point for a given soil that may be obtained from a TDR measurement at zero water content. It appears that “c” may be the same for similar soil s. Constant “c” varies only slightly with the dry density of the soil. A unique point on the “f” and “g” calibration plot located at the zero water content condition may exist where all true calibration lines intersect. This point varies only slightly depending on the dry density of the material. If an initial amount of sa lt is present in the soil, th e calibration constants obtained are different from the true calibration values. Values of “d” and “g” will be lower than the true values and “c” and “f” will be higher than the true values. To obtain the true calibration values a soil must be washed with deionized water before it is calibrated. To avoid washing a soil, it may be possible to use the zero water conten t calibration points to extract true calibration lines.

PAGE 84

72 CHAPTER 5 IN-SITU CBR CORRELATION TO TDR SPIKE DRIVING Introduction The Purdue TDR method requires that four sp ikes be driven into the soil to obtain a field TDR waveform. Using a brass hamme r, the TDR spikes are driven into the ground to transfer the coaxial line into the soil medium. The effort required to drive the spikes into the soil varies with the soil’s resi stance to penetration. This resistance to penetration is related to the soil’s strength. Similar tests, which rely on the measurement of resistance to penetration, have been co mmonly used to estimate soil strength. One such test is the commonly r un California Bearing Ratio (CBR) test. It is therefore proposed that the energy require d to drive the TDR spikes be correlated to the CBR test. Instead of driving the spikes with a brass hammer, a modified standard proctor hammer is used to measure the amount of effo rt required to drive the spikes. A steel cap with a groove having the same dimensions as the spike head is attached to the standard hammer. This promotes uniformity in the m easurement of the energy required to drive the TDR spikes into the soil. Depending on the anticipated soil strength, further modification could be introduced to the stan dard hammer to allow changes in drop height. A correlation between energy required to drive the spikes and the soil resilient modulus or CBR value will be pr esented in this chapter. Theoretical Background The TDR procedure requires that four 12 in ch stakes be driven into the soil to obtain a TDR measurement (the waveform is used for the measurement of soil properties as previously explained in chapter 2). The energy required to drive the TDR stakes into the soil is variable and is related to the soil’s strength. The str onger the materi al is the

PAGE 85

73 more hammer blows are required to drive the stakes. The California Bearing Ratio (CBR) test operates on the same principle. The CBR test involve s comparing a soil’s resistance to penetration to that of a standard material. This test is commonly used for evaluation and design of flexib le pavement components. The piston is pushed into the ground as both force and displacement are meas ured. A measurement of strength is then derived from the relationship of stress to pe netration. The measurement varies according to the strength of the material. The CBR stre ngth parameter is then used to estimate the bearing capacity of the soil for de sign and analysis purposes. Both the CBR test and the driving of the TDR spikes can be used to measure the amount of energy required to displace a soil a given amount. As a result, it is proposed that a correlation exists between the energy required to driv e the TDR spikes into a soil and the CBR number for the same soil. A modi fied proctor hammer is used to introduce consistency in the TDR stake driving process. The number of blows required to drive the 12 inch stakes into the TDR template can then be recorded. The average of the four stakes can then be taken and compared to the in-situ CBR value obtained for the same soil specimen. Existing Empirical Correlations to the CBR Test Correlative research to the CBR test has been carried out using the dynamic cone penetration test (DCP), the sta ndard penetration test (SPT), the vane shear test (VST) and unconfined compression tests. The VST and unconfined compression tests are primarily concerned with silty and clayey subgrades and for that reason will not be dealt with here. The DCP test is used to approximate soil st rength to depths of one meter. This is accomplished by driving a rod with a penetration cone at the end into the soil. The penetration associated with each drop is recorded and a relationship is derived between the number of drops and the penetration of each drop. The SPT test is widely used in site-investigation works. The SPT test is a measure of penetration as a result of an applied force. The measurement of penetra tion to blows has been used to formulate a correlative relationship between the SPT te st and the CBR test (Livneh, 1989).

PAGE 86

74 Laboratory research done by Kleyn (1975) led to the fo rmulation of the following DCP to CBR correlation: DCP CBR log 27 1 62 2 log (5-1) Where CBR is the CBR value for the soil and DCP is the ration of penetration (mm) to the number of blows requi red to achieve such penetration. Smith and Pratt (1983) performed a series of field investig ations and came up with a slightly different co rrelation than Eqn. (5-1): DCP CBR log 15 1 56 2 log (5-2) From analysis of laboratory-based res earch, Harison (1984 and 1986) established the following model: DCP CBR log 32 1 81 2 log (5-3) Livneh and Ishai (1987) took a modified a pproach by adding an exponential to the DCP term: 5 1 log 71 0 20 2 log DCP CBR (5-4) Other similar correlations have been devel oped; all equations have similar values, with the constant ranging from 2.555 to 2. 81 and the log multiplier ranging from 1.12 to 1.32. The only exception is Liveneh’s (1987 ) equation (Eqn. 5-4). The differences between proposed correlations can be attributed to the differences in cone head angle of among other differences in data co llection and procedure.

PAGE 87

75 A final recommendation based on previous research was made by Harison (1989): DCP CBR log 14 1 55 2 log (5-5) Eqn. (5-5) various only slightly when compared to the correlation proposed by Smith and Pratt (1983) (Eqn. 5-2). Figure 5-1 is a graphical representation of the several proposed correlations between the DC P and CBR tests (Harison 1989). Research done on granular materials by Ha rison (1989) suggests that a slightly steeper relationship may exist for granular materials. Figure 5-1. DCP and C BR Relationship. Source: Harison 1989. 0 10 20 30 40 50 60 70 80 01020304050607080DCP blows per 12 inchesIn-situ CBR Numbe r Harison (1984) Livneh (1987) Smith & Pratt (1983) Kleyn (1975)

PAGE 88

76 Due to the wide spread use of the SPT te st in site-investiga tion, Livneh and Ishai (1987) developed an equation for transformi ng SPT values to LBR numbers. Further improvements were made yielding the follo wing equation (Livneh and Ishai, 1988 and Livneh, 1989): 26 0) (log 55 6 13 5 log SPT CBR (5-6) Where SPT is the relationship between de pth of penetration (mm) and number of blows required. Figure 5-2 shows the relationship betw een the SPT values and LBR values derived by Livneh (1989). Note that the data is displayed in SPT blow count per 12 inches for comparative purposes. Figure 5-2. SPT and CBR Relations hip. Source: Livneh (1989). 0 10 20 30 40 50 60 70 80 90 100 110 120 130 010203040506070SPT blows per 12 inchesIn-situ CBR Numbe r

PAGE 89

77TDR and In-situ CBR Correlation The DCP and SPT tests rely on a penetration to resistance relati onship to evaluate the strength of a material. The CBR test reli es on the same principle. The correlations developed for the DCP and SPT tests to the CBR test indicate that an exponential relationship exists between them. The same type of relationship between the energy required to drive the TDR spikes and the CBR test is expected. Equipment ASTM standard D4429-93 (in-place CBR test) was used as a guideline in acquiring necessary equipment. The components of the in-place CBR test according to ASTM standard D 4429-93 are: 1) Mechanical screw jack ca pable of applying 5950 lbf 2) Two proving rings of 1984 lbf and 5070 lbf in capacity 3) Penetration piston 3 i n2 in cross section 4) Piston adaptors 5) Pipe extensions 6) Dial gages 7) Surcharge plate 8) Surcharge weights 9) Reaction truck capable of transmitting a 6970 lbf reaction force 10) Loading frame and wood supports 11) Beam used for dial gage support The truck, proving rings, j ack, and standard piston we re already available at University of South Florida. The reaction b eam had to be designed and machined in the university workshop. The surcharge plates, su rcharge loads, extension rods and materials required for the reaction beam, were purchas ed. Figure 5-3 shows typical in-situ CBR test equipment. Figure 5-4 displays the act ual truck and set up used for testing.

PAGE 90

78 Additional equipment re quired for the test: 1) TDR driving template 2) Four standard TDR spikes 3) Modified standard Proctor hammer Figure 5-4. CBR In-place Equipment Setup. Figure 5-3. Equipment for In-Place CBR Test. Source: ASTM D4429.

PAGE 91

79 The TDR equipment was previously acq uired in the purch ase of standard equipment obtained from Purdue University. The standard Proctor hammer was modified and machined at the University of South Florida (See figure 5-5). Procedure 1) Select a test location that is relatively flat and readily accessible to the truck required for the CBR test. 2) Prepare the testing area by removing a ny loose organic material lying on the surface to be tested. Typically, it is nece ssary to remove the t op one to two inches of soil to expose soil that does not s how inconsistency fro m rain or organic growth. Figure 5-5. The Standard Proctor Hammer with the Attached Head.

PAGE 92

80 3) Attach the loading frame to the hydraulic jacks located at the front of the truck, position the truck over the desired testing s ite, and lift the truck so no weight rests on the front springs. 4) Attach the jack to the loading frame and connect the proving ring along with the loading piston in series. Extension can be used to come within 4.9 inches of the surface that is to be tested. 5) Position the circular surcharge plate directly over the material that is to be tested. Align the plate so that the penetrati on piston passes through the central hole. Place an additional twenty pound surcharge pl ate over the circular plate to achieve the minimum total surcharge of thirty pounds. 6) Place the dial support in such a way that it does not interfere with the penetration piston or disturb the material being tested. Attach the dial gage to the beam and place it over the penetration piston were it can be easily viewed. 7) Zero both the dial gage for the proving ring and for the displacement gage. 8) Apply the load to the penetration piston. The rate of pe netration should be approximately 0.05 in./min. This is done by monitoring the displacement gage while applying load with the jack. Reco rd the deflection of the proving ring at increments of 0.025 inches to a final depth of 0.500 inches. 9) Prepare a location for the TDR driving template that is as close to the penetration piston as possible, but no closer than the ASTM recommended distance of 15 inches. The area should be cleared of any organic material and leveled. 10) Place the first spike into the template. Th e spike may need to be held in place manually. Place the modified standard Pr octor hammer over the spike fitting the grove over the head of the spike. The Proctor hammer is then dropped in the traditional manner and the number of drops is recorded. To ensure that the hammer drops are being applie d perpendicularly to the gr ound and in line with the spike, an observer should view the pr ocess at ground level while holding the spikes in position. Figure 5-6 disp lays the TDR spike driving technique.

PAGE 93

81 11) Repeat the process outlined in the previous step until all four spikes are driven. The number of blows required to drive each stake to the template should be recorded. 12) After testing is completed, obtain a soil sample which will be used to classify the soil. Calculation 1) Compute the average number of blows requi red to drive the TDR spikes from the four recorded values. 2) Calculate the applied stre ss at each displacem ent recorded from the CBR test. Generate a plot of the stress penetrati on curve using the measured stress versus displacement relationship. Figure 5-6. TDR Spik e Driving Technique.

PAGE 94

82 3) If necessary, correct the penetration graph for surf ace irregularities or upward concavity (See ASTM D 4429 for complete explanation). Fi gure 5-7 shows an example of a corrected penetration curve. 4) Record the stress at 0.1 in. and 0.2 in. penetration from the corrected penetration curve. Divide the recorded stress at 0.1 in. by a value of 800 psi and then multiply by 100. Divide the stress at 0. 2 in penetration by a value of 1200 psi and then multiply by 100. Record the values obtained for both the 0.1 in. and 0.2 in. penetrations. If the 0.1 in. value is greater this number is recorded as the CBR number. If however the 0.2 in value is greater, the test should be repeated. Then if the results are consistent with the first test, the 0.2 in. number should be reported as the CBR number. Figure 5-7. Correction of Penetration Curve. Source: ASTM D 4429.

PAGE 95

83Test Results Several CBR field tests were run along with TDR spike driving tests in an attempt to establish a correlative transformation e quation. Tests were mainly carried out at various locations on th e University of South Florida cam pus. A smaller number of tests were performed on roadway projects in the greater Tampa area (S R 54 tests). Test locations were selected to give a wide range of CBR values that would cover the typical values encountered in Florida. All materials th at were tested were A3 or A-1-b type soil. Each test was performed in accordance with the procedure outlined previously. A summary of test locations, soil types and results is displayed in Table 5-1. Table 5-1 also displays the variation between the TDR spikes dr iven in the form of a standard deviation. Significant variability was encountered in driving the TDR stakes; however most standard deviation values were under 3. Table 5-1 also indicates that at higher LBR values the support system used was prone to slipping. This was evident in the wooden support blocks sliding against e ach other at higher loading conditions. Slippage between the wooden blocks was caused by higher loading conditions (above 2400 pounds) and also by frictional force induced by a slight in clination experienced by the truck as it was raised by the hydraulic jack system. Th is limited data collection to CBR values approximately 65. However, based on previo usly discussed DCP correlative research, the TDR/CBR relationship developed is expect ed to yield acceptable results at higher range CBR values. Discussion/Analysis As the results indicate, the correlation obt ained for the TDR spikes to the in-situ CBR number compares favorably with the tran sformation equations pr eviously discussed (Figures 5-1 and 5-2). This was expected, as the driving of the TDR spikes is not much different from the DCP test. Both tests operate on the same principle of using a penetration to resistance relationship to esti mate soil strength. It follows that the correlative transformation equations develope d are of similar shape. Of course, the shape, size, geometry and force used to drive the object va ry between tests.

PAGE 96

84 Proposed Model The data from Table (5-1) was plotted a nd a relationship was derived between the TDR spike driving and the in-s itu CBR test (Figure 5-8). Table 5-1. Summary of CBR Field Testing. Test LocationOperatorSoil TypeAvg. # BlowsStnd DevCBR #Comment Water ChaseNewel & RorySP / A-311.750.53OK Shriners 1Newel SP / A-1-b35.753.9525OK Shriners 2Newel SP / A-1-b29.252.520OK Shriners 3Newel SP / A-1-b29.51.7319OK Shriners 4Newel SP / A-1-b36.752.535OK Shriners 5Newel SP / A-1-b332.5823OK Shriners 6Newel SP / A-1-b29.52.0818OK Moffit 1Newel SP / A-1-b42.55.3237OK Moffit 2Newel SP / A-1-b63.752.2254OK Moffit 3Newel SP / A-1-b597.3960OK Moffit 4Newel SP / A-1-b597.3968OK Botanical 2Newel & RorySP / A-317.751.59OK Botanical 3Newel & RorySP / A-325.252.2224OK Botanical 4Newel SP / A-323.51.2915OK Botanical 5Newel SP / A-327.251.7120OK Botanical 7Newel SP / A-322.51.2924OK Botanical 8Newel SP / A-3172.3116OK Botanical 9Newel SP / A-319.52.8718OK Botanical 10Newel SP / A-311.50.59OK SR 54-2Newel & RorySP / A-3671.1565OK Botanical 6Newel SP / A-317.252.9719OK Botanical 1Newel & RorySP / A-315.250.95717OK Moffit 5Newel SP / A-1-b54.513.7271Support slipped SR 54-1Newel & RorySP / A-370.256.857Support slipped Moffit 6Newel SP / A-1-b41.53.768Support slipped

PAGE 97

85 The following relationship was obtained to estimate the in-situ CBR number from the average number of blows re quired to drive the TDR stakes: 896 2 438 0 007 0 2 TDR TDR CBR (5-7) Where TDR is the average number of blows required to drive the 12 inch TDR spikes into the soil. However, it may be more valuable to obtain a range of possible in-situ CBR values due to the inherent variability of soil pr operties within a given site. The variability of in-situ CBR measurements is evident in the data collected as well as previous attempts to establish in-situ CBR correlations (Figures 5-1 and 5-2). Figur e 5-9 displays the collected data points along with a low and a high value based on one and a half standard deviations from Eqn. 5-7. Re porting a range of possible valu es based on one and a half Figure 5-8. Proposed CBR and TDR Relationship. y = 0.0072x2 + 0.4832x + 2.8964 R2 = 0.92140 10 20 30 40 50 60 70 80 01020304050607080Average BlowsCBR Numbe r

PAGE 98

86 standard deviations gives a confidence inte rval of 0.8664 th at the actual in-situ CBR lies within the prescribed range of values. Th e following range is recommended for in-situ CBR number estimation: 94 7 CBR (5-8) Where both CBR is defined previously in Eqn. (5-7). Figure 5-9. Proposed Range of In-situ CBR Values. y = 0.0072x2 + 0.4832x + 10.831 y = 0.0072x2 + 0.4832x 5.03860 10 20 30 40 50 60 70 80 01020304050607080Average TDR BlowsCBR Numbe r

PAGE 99

87Conclusion The collection of data from a variety of fi eld tests can greatly aid in the design of earthwork structures. Moisture content a nd density are two parameters that can be obtained with the Purdue TDR method. An a dditional in-situ CBR range of values can be obtained through the correlation to the spikes driven when TDR measurements are taken. This additional measurement can readily provide an estimation of an in place CBR number that otherwise would re quire the mobilization of speci al equipment that in most cases is costly to operate and maintain. This additional feature allows the TDR operator to obtain three important soil parameters that are important in design and construction: in place moisture content, in place de nsity and an in place CBR number.

PAGE 100

88 CHAPTER 6 ASTM METHOD COMP ARED TO TRADITIONAL METHODS Introduction Previous discussion has been centered on advances in TDR technology that led to the development of the Purdue one-step TDR met hod. Due to the recent nature of these developments no testing program has been implemented as of yet to compare the Purdue one-step method with traditional methods. Howe ver, a study was recently carried out in conjunction with the Florid a Department of Transportation (FDOT) to evaluate the relative accuracy of the ASTM TDR two-step method compared to traditional methods. The testing program included a series of side by side testing with the ASTM TDR method and the nuclear, sand cone and driv e sleeve methods. The nuclear method has become increasingly popular in recent years, due to its ability to measure both density and water content, whereas the sand cone and drive sleeve method are more traditional approaches that are limited to the measuremen t of density and rely on a separate method to determine water content. As a result, the primary comparison made here is with the nuclear method. Testing Program The widespread use of the nuclear met hod by the Florida DOT allows for ready access to a large amount of data. The nuclear method is commonly implanted at a variety of job sites across the state of Florida. TDR measurements can easily be taken simultaneously with routine nuclear gauge testing using the ASTM TDR method. For purposes of this study, an effort was also ma de to collect comparative data for both the sand cone and drive sleeve methods for additio nal insight. Samples were collect ed from all testing locations and were taken to the laboratory to obtain a baseline oven dry water

PAGE 101

89 content. Due to the lack of a baseline method for measuring in-situ soil dry density, the nuclear moist density was used as a baseline, with the dry density back-calculated from the oven dry moisture content. This, of c ourse, is not entirely accurate, however it was the method most readily availabl e. A brief review of the nuclear, sand cone and drive sleeve methods are contained in Chapter 2 and will not be dealt with here. Nuclear vs. ASTM TDR Results A series of side-by-side tests were car ried out at severa l locations throughout Florida using the nuclear gauge method and the ASTM TDR method. Blanket calibration values of “a” = 1 and “b” = 9 were used for field TDR measurements. The blanket values used for testing were made based on preliminar y results and it should be noted that a final recommendation for Florida sands was made by Sallam et al. (2004) that was slightly different than the value used during the testi ng program. The effects of this discrepancy will be discussed later in the chapter. A summary of test locations and soil types for the nuclear to ASTM TDR comparison is displaye d in Table 6-1. Testing was carried out across the state at a variety of highway projects. Several tests were run at each location. All samples tested were common construction soils encountered in Florida (A-3 and A-24 sands). Table 6-1. Nuclear Testing Locations and Information. Location County/CityNo. of SamplesSoil Type(s) I-295 / I-95Duval Co.2A-3 SR44Sumter Co.5A-3 SR54 E 41Pasco6A-2-4 SR 207St Johns8A-3/A-2-4 Santa FeUnion7A-3/A-2-4 SR 54 E 19Pasco6A-3 SR 98Walton4A-3 SR 291Pensacola6A-2-4 I-4Volusia6A-3/A-2-4

PAGE 102

90 Table 6-2. Nuclear Water Content Comparison Results. LocationTestOven wcTDR wcNuclear wc% error TDR % error Nuclear Absolute TDR Absolute Nuclear Duval Co.19.0%13.0%10.2%44.4%13.3%4.00%1.20% Duval Co.29.7%12.5%9.2%28.9%-5.2%2.80%-0.50% Sumter Co.37.6%7.2%7.9%-5.3%3.9%-0.40%0.30% Sumter Co.48.0%7.6%7.9%-5.0%-1.3%-0.40%-0.10% Sumter Co.57.9%7.1%7.6%-10.1%-3.8%-0.80%-0.30% Sumter Co.612.9%12.2%13.4%-5.4%3.9%-0.70%0.50% Sumter Co.714.0%13.6%16.4%-2.9%17.1%-0.40%2.40% Sumter Co.813.2%12.3%14.8%-6.8%12.1%-0.90%1.60% Pasco16.3%5.9%6.2%-6.3%-1.6%-0.40%-0.10% Pasco26.2%5.7%5.3%-8.1%-14.5%-0.50%-0.90% Pasco37.2%6.0%7.2%-16.7%0.0%-1.20%0.00% Pasco45.5%4.7%5.0%-14.5%-9.1%-0.80%-0.50% Pasco56.6%5.7%6.9%-13.6%4.5%-0.90%0.30% Pasco68.6%8.4%8.7%-2.3%1.2%-0.20%0.10% St Johns113.8%13.4%14.1%-2.9%2.2%-0.40%0.30% St Johns29.7%9.3%9.0%-4.1%-7.2%-0.40%-0.70% St Johns3a11.4%10.8%12.2%-5.3%7.0%-0.60%0.80% St Johns3b11.0%10.8%12.4%-1.8%12.7%-0.20%1.40% St Johns412.2%11.4%13.4%-6.6%9.8%-0.80%1.20% St Johns58.2%11.9%13.8%45.1%68.3%3.70%5.60% St Johns68.7%6.9%8.7%-20.7%0.0%-1.80%0.00% St Johns712.8%8.8%8.6%-31.3%-32.8%-4.00%-4.20% Union110.2%9.4%12.3%-7.8%20.6%-0.80%2.10% Union210.3%9.2%11.4%-10.7%10.7%-1.10%1.10% Union39.4%8.4%10.1%-10.6%7.4%-1.00%0.70% Union47.5%6.6%11.6%-12.0%54.7%-0.90%4.10% Union510.5%8.9%9.8%-15.2%-6.7%-1.60%-0.70% Union79.1%8.0%13.4%-12.1%47.3%-1.10%4.30% Pasco14.0%3.5%4.5%-12.5%12.5%-0.50%0.50% Pasco26.9%6.3%9.1%-8.7%31.9%-0.60%2.20% Pasco36.6%6.1%9.6%-7.6%45.5%-0.50%3.00% Pasco45.8%5.3%7.2%-8.6%24.1%-0.50%1.40% Pasco56.6%6.0%7.7%-9.1%16.7%-0.60%1.10% Pasco65.2%4.5%6.3%-13.5%21.2%-0.70%1.10% Walton114.0%9.6%15.1%-31.4%7.9%-4.40%1.10% Walton213.2%10.2%13.1%-22.7%-0.8%-3.00%-0.10% Walton312.8%9.9%12.9%-22.7%0.8%-2.90%0.10% Walton413.5%10.5%13.6%-22.2%0.7%-3.00%0.10% Pensacola19.2%6.6%7.2%-28.3%-21.7%-2.60%-2.00% Pensacola28.7%5.5%6.5%-36.8%-25.3%-3.20%-2.20% Pensacola38.4%5.5%6.7%-34.5%-20.2%-2.90%-1.70% Pensacola49.7%6.6%7.7%-32.0%-20.6%-3.10%-2.00% Pensacola59.7%6.6%8.4%-32.0%-13.4%-3.10%-1.30% Pensacola69.2%6.3%7.4%-31.5%-19.6%-2.90%-1.80% Volusia110.0%9.2%12.3%-8.0%23.0%-0.80%2.30% Volusia211.3%9.8%13.1%-13.3%15.9%-1.50%1.80% Volusia312.0%10.2%12.3%-15.0%2.5%-1.80%0.30% Volusia48.9%7.5%8.7%-15.7%-2.2%-1.40%-0.20% Volusia510.3%8.8%10.0%-14.6%-2.9%-1.50%-0.30% Volusia67.5%6.6%8.9%-12.0%18.7%-0.90%1.40%

PAGE 103

91 Table 6-2 displays the ASTM TDR and nuclear water content measurements recorded at each test site along with the oven dry baseline water content measured in the laboratory. The percent error and absolute error was then calculated by comparing the field water content measurements to the baseline water content values. The data from Table 62 is displayed graphically in Figure 6-1. Data points for both nuclear and ASTM TDR testing were plotted alon g with a 1:1 line. Table 6-3 summarizes the correspondi ng nuclear and ASTM TDR dry density measurements along with the dry density back calculated using th e wet nuclear density and the oven dry water content. The values were compared and percent and absolute error were calculated. All data points fo r both the nuclear and ASTM TDR tests were plotted in Figure 6-2 along with a 1:1 line. Figure 6-1. Nuclear versus ASTM TDR Water Content. 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.000.020.040.060.080.100.120.140.16 Oven Dry MoistureTDR and Nuclear Moisture TDR Nuclear 1:1

PAGE 104

92 Table 6-3. Nuclear Dry Density Comparison Results. LocationTestOven dTDR dNuclear d% error TDR % error Nuclear Absolute TDR Absolute Nuclear Duval Co.1102.894.1101.6-8.5%-1.2%-8.70-1.20 Duval Co.2102.688.0103.1-14.2%0.5%-14.600.50 Sumter Co.3112.0105.5111.7-5.8%-0.3%-6.50-0.30 Sumter Co.4112.5108.3112.6-3.7%0.1%-4.200.10 Sumter Co.5112.9108.3113.2-4.1%0.3%-4.600.30 Sumter Co.6108.8107.1108.3-1.6%-0.5%-1.70-0.50 Sumter Co.7107.3102.2105.1-4.8%-2.1%-5.10-2.20 Sumter Co.8111.2127.7109.714.8%-1.3%16.50-1.50 Pasco1108.9103.8109.0-4.7%0.1%-5.100.10 Pasco2107.8100.1108.7-7.1%0.8%-7.700.90 Pasco3107.3106.7107.3-0.6%0.0%-0.600.00 Pasco4107.2101.4107.7-5.4%0.5%-5.800.50 Pasco5112.2110.4111.9-1.6%-0.3%-1.80-0.30 Pasco6107.496.1107.3-10.5%-0.1%-11.30-0.10 St Johns1101.398.0101.1-3.3%-0.2%-3.30-0.20 St Johns2101.594.6102.2-6.8%0.7%-6.900.70 St Johns3a103.399.1102.6-4.1%-0.7%-4.20-0.70 St Johns3b107.0100.0105.7-6.5%-1.2%-7.00-1.30 St Johns4103.0106.0101.92.9%-1.1%3.00-1.10 St Johns5108.7100.8103.3-7.3%-5.0%-7.90-5.40 St Johns6100.8100.1100.8-0.7%0.0%-0.700.00 St Johns799.098.1102.9-0.9%3.9%-0.903.90 Union198.187.896.2-10.5%-2.0%-10.28-1.92 Union2100.1102.499.12.3%-1.0%2.30-0.99 Union3105.799.7105.0-5.6%-0.6%-5.92-0.67 Union4107.2108.1103.20.9%-3.7%0.93-3.94 Union5102.9107.7103.64.7%0.6%4.810.66 Union7109.6103.8105.5-5.3%-3.7%-5.80-4.10 Pasco1102.698.0102.1-4.5%-0.5%-4.63-0.49 Pasco2106.0100.9103.8-4.8%-2.0%-5.04-2.14 Pasco3107.698.6104.7-8.4%-2.7%-9.00-2.95 Pasco4102.896.1101.5-6.5%-1.3%-6.69-1.34 Pasco5105.196.8104.0-7.8%-1.0%-8.24-1.07 Pasco6105.197.4104.0-7.4%-1.0%-7.75-1.09 Walton1109.4111.9108.32.3%-1.0%2.50-1.10 Walton2110.1108.9110.2-1.1%0.1%-1.200.10 Walton3111.0107.3110.9-3.3%-0.1%-3.70-0.10 Walton4108.9105.9108.8-2.8%-0.1%-3.00-0.10 Pensacola1115.9112.9118.1-2.6%1.9%-3.052.16 Pensacola2116.4118.1118.81.5%2.1%1.752.40 Pensacola3114.7116.6116.51.7%1.6%1.911.83 Pensacola4114.7117.9116.82.8%1.9%3.242.13 Pensacola5116.1117.7117.51.3%1.2%1.551.39 Pensacola6116.1112.7118.1-3.0%1.7%-3.461.95 Volusia1115.5109.7113.2-5.0%-2.0%-5.82-2.37 Volusia2114.1113.0112.3-1.0%-1.6%-1.13-1.82 Volusia3108.2106.3107.9-1.7%-0.3%-1.88-0.29 Volusia4112.1106.6112.3-4.9%0.2%-5.480.21 Volusia5112.5107.8112.8-4.2%0.3%-4.760.31 Volusia6109.9113.4108.43.2%-1.3%3.49-1.41

PAGE 105

93 Sand Cone vs. ASTM TDR Results A series of side by side tests were ca rried out between the sand cone method and the ASTM TDR method. Again, the blanket calibration values of “a” = 1 and “b” = 9 were used. Table 6-4 displays test informa tion for the sand cone tests. The baseline dry density was calculated by using the nuclear wet density and th e oven dried water content. The percent error was then computed for each test. Figure 6-3 displays the sand cone vs. ASTM TDR testing results. Figure 6-2. Nuclear versus ASTM TDR Dry Density. 80 85 90 95 100 105 110 115 120 95100105110115120 Nuclear/Oven Dry Density (pcf)TDR and Nuclear Dry Density (pcf ) TDR Nuclear 1:1

PAGE 106

94 Drive Sleeve vs. ASTM TDR Results A series of side by side tests were car ried out between the drive sleeve method and the ASTM TDR method. Blanket calibra tion values of “a” = 1 and “b” = 9 were used. Table 6-5 displays test information fo r the drive sleeve tests. The baseline dry density was calculated by using the nuclear wet density and th e oven dried water content. Table 6-4. Sand Cone Dry Density Comparison Results. LocationTestOven dTDR dSand Cone d% error TDR % Error Sand Cone I-295 / I-951102.894.195-8.46%-7.59% I-295 / I-952102.68897.4-14.23%-5.07% Sumter Co.3112105.5112.4-5.80%0.36% Sumter Co.4112.5108.3111.9-3.73%-0.53% Sumter Co.5112.9108.3114.3-4.07%1.24% Sumter Co.6108.8107.1106.8-1.56%-1.84% Sumter Co.7107.3102.2114.3-4.75%6.52% Sumter Co.8111.2127.7110.514.84%-0.63% Figure 6-3. Sand Cone vers us ASTM TDR Dry Density. 80 85 90 95 100 105 110 115 120 9698100102104106108110112114116118 Nuclear/Oven Dry Density (pcf)TDR and Sand Cone Dry Density (pcf ) TDR Sand Cone 1:1

PAGE 107

95 The percent error was then computed for each test. Figure 6-4 displays the drive sleeve vs. ASTM TDR testing results. Table 6-5. Drive Sleeve Dr y Density Comparison Results. LocationTestOven dTDR dDrive Sleeve d% error TDR % Error Drive Sleeve I-295 / I-951102.894.1102.8-8.46%0.00% I-295 / I-952102.688101.3-14.23%-1.27% Sumter Co.3112105.5112.9-5.80%0.80% Sumter Co.4112.5108.3115.5-3.73%2.67% Sumter Co.5112.9108.3114.8-4.07%1.68% Sumter Co.6108.8107.1109.6-1.56%0.74% Sumter Co.7107.3102.2121.8-4.75%13.51% Sumter Co.8111.2127.7112.914.84%1.53% Figure 6-4. Drive Sleeve ve rsus ASTM TDR Dry Density. 80 85 90 95 100 105 110 115 120 9698100102104106108110112114116118 Nuclear/Oven Dry Density (pcf)TDR and Drive Sleeve Dry Density (pcf ) TDR Drive Sleeve 1:1

PAGE 108

96Water Content Measurement Discussion The water content measurement comparis on displayed in Table 6-2 shows the absolute error for both the ASTM TDR and nuclear methods varied similarly (ASTM TDR between -4.4 and 4.0 and while the nuclear varied between -4.2 and 5.6). Figure 61 indicates that the ASTM TDR method consis tently under predicted the baseline water content value. This observati on is consistent with a study performed by Sallam et al. (2004) which indicated that the ASTM TDR method is likely to underpredict water content. The ASTM TDR measurements have a higher correlation coefficient with the oven dry measurements than the nuclear measurements (0.899 for the ASTM TDR method compared to 0.791 for the nuclear me thod). As mentioned previously, blanket values of “a” = 1 and “b” = 9 were used to determine the TDR field water contents. In their final report prepared for the FDOT, Sallam et al. (2004) suggest that a more representative “b” value for Florida sand is 8. 5 (Sallam et al., 2004). The reduction in the “b” value will predict slightly higher wate r content values in the field ASTM TDR measurements displayed in Figure 6-1. This would mean that the TDR measurements would be shifted upwards towards the 1:1 line. If the recommended value for constant “b” would have been used, it is reasonable to suggest that the ASTM TDR water content measurement is more accurate than the nucle ar method for water content measurement. Dry Density Discussion Due to the fact that the baseline dry de nsity was calculated from a combination of the nuclear moist density and the oven dry mo isture content results should be viewed with caution. Obviously, it is to be expect ed that the nuclear measurements be more accurate than the TDR measurements. The bias towards the accuracy of the nuclear method is apparent in Figure 6-2 where the co rrelation coefficient is much higher for the nuclear method than the ASTM TDR method (0 .931 compared with 0.707). In order to evaluate the true accuracy of both the nuclear and the ASTM TDR method, the use of an objective and independent point of reference fo r dry density is needed. The development of such independent measurements was, howev er, beyond the scope of the current study.

PAGE 109

97 Results indicate that ASTM TDR dry density was consistently below the nuclear baseline value. A possible explanation may be from th e use of a slightly different “b” value from the recommended value. Table 6-4 and Figure 6-3 should only be us ed for comparative purposes. Again, the dry density baseline is ta ken as the nuclear moist densit y back calculated to the dry density using the oven dry water content. As displayed in Figure 6-3 the scatter between the two methods appears to be comparable. Ho wever, more data needs to be collected to make a realistic determination of the relati ve accuracy of the ASTM TDR method to the sand cone method. Again in Table 6-5 and Fi gure 6-4 the TDR and drive sleeve dry density results are compared to the measured nuclear dens ity and oven dry water content. The drive sleeve method appears to compare favorably with the nuclear dry density baseline. More testing would be needed to make a reasona ble statement of the relative accuracy of ASTM TDR method compared to the drive sleeve method. Conclusions The nuclear method has been accepted as a reliable method for both water content and density for base course materials, due to its ability to measure both water content and dry density, and its perceived reliability and accuracy. The sand cone method and the drive sleeve method are other methods that have been commonly used to estimate dry density in the field. Side -by side measurements compar ing the ASTM TDR method to the nuclear method for water content measur ement on Florida construction soils indicate that the ASTM TDR method displa ys less scatter than the nucl ear gauge and as a result is likely more accurate than the nuclear gauge w ith the proper selection of constants “a” and “b.” It thus appears that the TDR method is more reliable than the nuclear method, at least in terms of water content measurement. Density results were inconclusive, due to the lack of a comparative baseline. However, a comparison of the data scatter between methods is similar. Further research is requir ed to fully evaluate the absolute accuracy of the methods for field density measurement.

PAGE 110

98 CHAPTER 7 – SUMMARY, CONCLU SIONS AND RECOMMENDATIONS Summary A comprehensive review of time domain reflectometry and its course to the field of geotechnical engineering wa s performed. Also a detaile d review of work done at Purdue University in the development of both the ASTM TDR and Purdue one-step methods was completed. Testing was carried out to evaluate the effect of pore fluid conductivity and initial sa lt content on the soil specific calibration of TDR constants. A greater understanding of the soil specific TD R constants was obtaine d. A relationship was developed to obtain an in-situ CBR va lue from the field TDR procedure. The accuracy of the ASTM TDR method was compared to that of current geotechnical testing methods. Research has lead to a greater understanding of the Purdue one-step TDR method that will aid in its further devel opment and implementation to the field of geotechnical measurement. Conclusions Investigation into the soil specific TDR calibration constants proved to be valuable. Soil constants “a” and “b” were fo und to behave consistently with previous studies. It appears that soil constant “c” may be a unique point for a given soil and it may be possible to catalog values based on soil type Soil constant “d” was demonstrated to change systematically with pore fluid conducti vity as previous research had indicated. Soil constants “f” and “g” change with the por e fluid conductivity of the soil. Constant “g” behaves similar to constant “d.” It is also noted that for a calibration plot of the dielectric constant and the bul k electrical conduc tivity there may exist a unique point at the dry condition for a particular soil at which true calibration lines intersect.

PAGE 111

99 Further investigation into the effect of pore fluid conductivity on the calibration to obtain the soil specific TDR cons tants revealed that the ini tial salt content may effect calibration. A soil containi ng an appreciable amount of salt will alte r calibration constants “c,” “d,” “f” and “g.” This is due to the constant change of pore fluid conductivity as the soil is calibrated. It appe ars that values of “c” will be higher than the true calibration value for initially salty soils. Constant “d” will most likely be lower than the true calibration value. Constant “f” will most lik ely be higher than the true calibration value and constant “g” will be lower than the true calibration value. For a true calibration to be obtained a soil must be washed from ions. It was noted that the true calibration of constant “c” may be determined at a dry soil condition. It appears that this point may be obtained regardless of salt content as long as the soil is tested at zero water content. To avoid washing a soil to obtain a true calibration line, it may be possible to obtain true calibration parameters under dry co nditions which can then be used to extract the true calibration lines. Several California bearing ratio (CBR) te sts were carried out adjacent to the driving of the TDR spikes. This allowed fo r a relationship to be developed between the CBR number and the number of blow required to drive the TDR stakes. The developed relationship showed a similar trend to simila r correlations developed in the past. The derived transformation equation should be used to give an approximate value for the insitu CBR number. Several tests were performed side-by-si de using the ASTM TDR method and the nuclear, sand cone and drive sleeve methods. Data was compared for both moisture content and dry density measurement. Resu lts indicate that the ASTM TDR method displays less scatter than the nuclear method for moisture content measurement and may be more accurate with the proper selection of calibration constants. Dry density results were inconclusive due to the lack of a ba seline method for comparison. However, it was observed that ASTM TDR measurements disp layed larger scatte r than the nuclear method for the comparative method adopted. Sand cone and drive sleeve comparisons were inconclusive due to insufficient testing results.

PAGE 112

100 Research carried out further validates the ASTM TDR method as a viable tool for geotechnical measurement. A greater understand ing of the soil specific constants used in conjunction with the Purdue one-step TDR me thod was also achieved. The results of studies carried out on to evalua te the effects of pore fluid conductivity on calibration will be valuable to establishing the Purdue one -step TDR method as a reliable geotechnical measurement system. Recommendations for Future Research It is clear that further testing is need ed to more fully compare the ASTM TDR method to traditional methods. An effort shoul d be made to evaluate the accuracy of dry density measurement to that of the nuclear method. Future te sting may also be useful in the further evaluation of consta nts “a” and “b.” Developing constants that could be used for common soils would be of value. Research relating to the Purdue one-step method should be focused on a procedure to accurately obtain the soil specific calibration c onstants without the need of washing the sample of ions. Further investig ation into the range of each constant and typical values for each would be of great valu e. As only sands were tested in research presented within this thesis, testing on soils with higher fines content is needed to validate the conclusions made herein for sands. Evaluation of the accuracy of the Purdue one-step method is needed. A testing program involving the Purdue one-step method and traditional methods would be of benefit in validating the accur acy of the method. A comp arison between dry density measurements would be of particular value; this would require the use of a baseline method for unbiased comparative results.

PAGE 113

101 REFERENCES Alharthi, A., and Lange, J. (1987), “Soil Wate r Saturation: Dielectric Determination,” Water Resources Research Vol. 23, pp. 591-595. Amente, G., Baker, John M. and Reece, Cliv e F. (2000), “Estimation of Soil Solution Electrical Conductivity from Bulk Soil Electrical C onductivity in Sandy Soils,” Soil Science Society of America Journal, Vol. 64, pp. 1931-1939. Birchak, J.R., Gardner, C.G., Hipp, J.E ., and Victor, J.M., ( 1974), “High Dielectric Constant Microwave Probes for Sensing Soil Moisture,” Proceedings IEEE Vol. 62, pp. 93-98. Campbell, Jeffrey E. (1990), “Dielectric Propert ies and Influence of Conductivity in Soils at One to Fifty Megahertz,” Soil Science Society of America Journal, Vol. 54, pp. 332341. Clarkson, T. S., Glasser, L., Tuxworth, R. W., and Williams, G. (1977), “An appreciation of experimental factors in Time-Domain Sp ectroscopy,” Adv. Mol. Relax. Processes, Vol. 10, pp. 173-202. Dalton, F. N., Herkelrath, W. N., Rawlins, D. S. and Rhoades, J. D. (1984), “Time Domain Reflectometry: Simultaneous Measurem ent of Soil Water Content and Electrical Conductivity with a Single Probe,” Science New Series, Vol. 224, No. 4652, pp. 989990. Dasberg, S. and Dalton, F. N. (1985), “Tim e Domain Reflectometry Field Measurement of Soil Water Content and Electrical Conductivity,” Soil Science Society of America Journal, Vol. 49, pp. 293-297. Dirksen, C, and Dasberg, S., (1993), “I mproved Calibration of Time Domain Reflectometry Soil Water Content Measurements,” Soil Science Society of America Journal Vol. 57, pp. 660-667. Drnevich, V. P., Lin, C., Quanghee, Yi, Yu, X., Lovell, Janet, (2000), “Real-Time Determination of Soil Type, Water Conten t, and Density Using Electromagnetics,” Final report, FHWA/IN/JTRP-2000/20.

PAGE 114

102 Drnevich, Vincent P., Yu, Xiong, Lovell, Janet and Tishmack, Jody, (2001), “Temperature Effects on Dielectric Constant Determined by Time Domain Reflectometry,” Proceedings of the Second Interna tional Symposium and Workshop on Time Domain Reflectometry for I nnovative Geotechnical Applications, Northwestern University, Evanston, Illinois, 2001. Fellner-Feldegg, J., (1969), “The Measurement of Dielectrics in Time Domain”. Journal of Physical Chemistry Vol. 73, pp. 616-623. Feng, W., Lin, C. P., Deschamps, R. J. a nd Drnevich, V. P. (1999), “Theoretical Model of a Multisection Time Domain Reflectometry Measurement System,” Water Resources Research, Vol. 35, No. 8, pp. 2321-2331. Giese, K. and Tiemann, R. (1975), “Determina tion of the complex permittivity from thinsample Time Domain Reflectometry: Improve d analysis of the syep response Wave form,” Adv. Mol. Relax. Processes, Vol. 7, pp. 45-59. Harison, J. A. (1989), “In -situ CBR Determ ination by DCP Testing Using a Laboratorybased Correlation,” Australian Road Research Vol. 19, pp. 313-317. Heimovaara, T. J. (1993), “Design of Triple -Wire Time Domain Reflectometry Probes in Practice and Theory,” Soil Science Society of America Journal, Vol. 57, pp. 1410-1417. Heimovaara, T. J. and Bouten, W. (1990), “A Computer-Controlled 36-Channel Time Domain Reflectometry System for Monitoring Soil Water Contents,” Water Resources Research, Vol. 26, No.10, pp. 2311-2316. Herkelrath, W. N., Hamburg, S. P. and Mu rphy, Fred (1991), “ Automatic, Real Time Monitoring of Soil Moisture in a remote Fi eld Area with Time Do main Reflectometry,” Water Resources Research, Vol. 27, No. 5, pp. 857-864. Hilhorst, M. A. (2000), “A Po re Water Conductivity Sensor,” Soil Science Society of America Journal, Vol. 64, pp. 1922-1925. Kleyn, E. G. (1975), “The Use of th e Dynamic Cone Penetrometer (DCP),” Transvaal Roads Department Report, No. L2/74. Krauss, J. D., (1984), “Electromagnetics ,” McGraw-Hill, New York Ledieu, J.P., Ridder, De., and Dautrebande A., (1986), “A Method for Measuring Soil Moisture Content by Time Domain Reflectometry ,” Journal of Hydrology Vol. 88, pp. 319-328.

PAGE 115

103 Lin, C., Siddiqui, S.I., Feng, W., Drnevich, V., and Desc hamp, R., (2000), “Quality Control of Earth Fills Using Time Domain Reflectometry (TDR)”, Constructing and Controlling Compaction of Earth Fills, ASTM STP 1384 D. W. Shankin, K. R. Rademacher, and J. R. Talbot. Eds., American Society for Testing and Materials, West Conshohocken, PA,2000. Livneh, M. (1989) “In-situ CBR testing by indirect methods,” Proceedings of the 12th International Conference on Soil Mechanics and Foundation Engineering Rio de Janero, Vol. 1, pp. 267-270. Livneh, M. and Ishai, I. (1987), “The Use of Between Dynamic Cone Penetrometer in Determining the Strength of Existing Pacements and Subgrades,” Proceedings of the 9th Southeast Asian Geotechnical Conference, Bangkok, Thailand. Livneh, M. and Ishai, I. (1988), “The Re lationship Between In-Situ CBR Test and Various Penetration Tests,” Proceedings of the Firs t International Synapse on Penetration Testing, Orlando, Florida, ISOPT-1. Malicki, M. A., Plagge, R. and Roth, C. H. (1996), “Improving the Calibration of Dielectric TDR Soil Moisture Determina tion Taking into Account the Solid Soil,” European Journal of Soil Science, Vol. 47, pp. 357-366. Mojid, M. A., Wyseure, G. C. L. and Rose D. A. (2003), “Electrical Conductivity Problems Associated With Time-Domaon Reflectometry (TDR) Measurement in Geotechnical Engineering,” Geotechnical and Geol ogical Engineering, Vol. 21, Is. 3, pp. 243-258. Nadler, A., Dasberg S. and Lapid, I. (1991) “Time Domain Reflectometry Measurements of Water Content and Electrical C onductivity of Layered Soil Columns,” Soil Science Society of America Journal, Vol.55, pp. 938-943. Noborio, K. (2001), “Measurement of Soil Wa ter Content and Electrical Conductivity by Time Domain Reflectometry: A Review,” Computers and Electroni cs in Agriculture, Vol. 31, pp. 213-237. O’Connor, K., and Dowding, C., (1999), “GeoMeasurements by Pulsing TDR Cables and Probes” A book published by CRC Press LLC. Rhoades, J. D. and J. van Schilfgaarde. (1976), “An Electrical Conductivity Probe for Determining Soil Salinity,” Soil Science Society of America Journal, Vol. 40, pp. 647651. Roth, C. H., Malicki, M. A. and Plagge, R. (1992), “Empirical Evaluation of the Relationship Between Soil Dielectric Constant and Volumetric Water Content as the Basis for Calibrating Soil Moisture Measurements by TDR,” Journal of Soil Science, Vol. 43, pp. 1-13.

PAGE 116

104 Sallam, Amr M., White, Newel K. and Ashmawy, Alaa K. (2004), “Evalauation of the Purdue TDR Method for Soil Water Content and Density Measurement,” Final Report to the Florida Department of Transportation, Contract No. BC-353-30. April 2004. Siddiqui, S.I. and Drnevich, V.P., (1995), “Use of Time Domain Reflectometry for the Determination of Water Content and Density of Soil,” FHWA/IN/JHRP-95/9, Purdue University. Smith, R.B. and Pratt, D. N. (1983), “A fi eld Study of In-Situ Ca lifornia Bearing Ratio and Dynamic Penetrometer Testing for Road Subgrade Investigation,” Australian Road Research Vol.13, No.4, pp. 285-294. Topp, G.C., Davis, J.L., and Annan, A.P. ( 1980), “Electromagnetic Determination of Soil Water Content and Electrical Conductivity Measurement Using Time Domain Reflectometry,” Water Resources Research Vol. 16, pp. 574-582. Topp, G. C., Zeglin, S., and White, I. ( 2000), “Impacts of the Real and Imaginary Components of Relative Permittivity on Time Domain Reflectometry Measurements in Soils,” Soil Science Society of America Journal, Vol. 64, No. 4, pp. 1244-1252. White, Ian, Zegelin, Steven J. and Topp, G. C. (1994), “Effect of Bulk Electrical Conductivity on TDR Measurement of Wa ter Content in Porous Media,” USBM special publication SP 19-94, Washington, D.C.: U.S. Bureau of Mines, pp. 294-308. Yanuka, M., Topp, G. C., Zegelin, S. and Zebc huk, W. D. (1988), “Multiple Reflection and Attenuation of Time Domain Reflectometry Pulses: Theoretical Considerations for Applications to Soil and Water,” Water Resources Research, Vol. 24, No.7, pp. 939-944. Yu, X. and Drnevich, V. P. (2004), “Soi l Water Content and Dry Density by Time Domain Reflectometry,” The Journal of Geotechni cal and Geoenvironmental Engineering, ASCE, accepted for publica tion, to appear in Vol. 130, Issue 9, September 2004. Zegelin, S. J., White, I. and Jenkins, D. R. (1989), “Improved Field Probes for Soil Water and Electrical Conductivity Measurement Using Time Domain Reflectometry,” Water Resources Research, Vol. 25, No. 11, pp. 2367-2376.