PB67-132783
EPA/600/2-86/111
December 1986
BOREHOLE SENSING METHODS FOR GROUND-WATER
INVESTIGATIONS AT HAZARDOUS WASTE SITES
by
Stephen W. Wheatcraft
Kendrick C Taylor
Water Resources Center
Desert Researcii Institute
University of Nevada System
P O. Box 60220 '
Reno, Nv. 83506
John W. Hess
Thomas M. Morris
Water Resources Center
Desert Research Institute
University of Nevada System
2505 Chandler Way, Suite 1
Las Vegas, Nv 89120
Cooperative Agreement No. CR 810052
Project Officer
Leslie G. McMilhon
Advanced Monitoring Systems Divwion
Environmental Monitoring Sjstems Laboratory
Las Vegas, Nevada 39114
ENVIRONMENTAL MONITORING SYSTEMS LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
LAS VEGAS, NEVADA S01I4
-------
TECHNIC(.L R FO T ‘DATA
(Please reaa Insouction: on :r.e Ff1 CF f bef Ode coispleflng)
1. REPORT NO 12
EPA/600/2—86/l1l
3
I ENT1A E I71 iti S
4. TITLE AND SUOTPTLE
BOREHOLE SENSING METhODS ! OR GROUND-WATER
INVESTIGATIONS AT HAZARDOUS WASTE SITES
5 REPORT DATE
December 1986
S. PERFORMING ORGANIZATION CODE
7 AUTHOR(S)
S.W. ‘4neatcraft, K.C. Taylor, LW. Hess: TM. Morris
I PERFORMING ORGANIZATION REPORT NO
9 PLRFORMING ORGANIZATION NAME AND ADDRESS
Desert Research Institute
University of Nevada System
P.O. Box 60220, Reno, NV 89506
10 PROGRAM ELEMENT NO
C104
11. CON1 RACTI RANT NO
CR_810052
(2. SPONSORING AGENCY NAME ANO AODRESS
Environmental Monitoring Systems Laboratory, LV. NV
Office of Research & Development
U.S. Environmental Protection Agency
Las Vegas, NV 89114
13 TYPE OF REPORT AND PERIOD COVERED
Project Report/Summary
14.SPONSOR INGAGEPJCYCOOE
EPA 600/07
15 SUPPLEMENTARY NOTES
6 ABSTRACT
The complex nature of the ground-water contamination problem requires the collection
of extensive amounts of data in order to understand the problem well enough to
recommend and execute the appropriate remedial action. As the complexity and
consequences of ground—water contamination increase, geophysical methods are
becoming a cost effective approach to providing answers to hydrogeologic cuestions
associated with ground—water contamination.
Geophysical methods applicable to hazardous waste site investigations can be broken
into two categories: surface and subsurface methods. Surface methods offer the
advantages of relatively little capital investment at the site and rapid colletion
of data over a horizontal area. However, the interpretation is often ambiguous and
limited in vertical resolution. Subsurface methods can be used c,nly to investigate
an area immediately around the borehole. However, subsurface methods provide
excellent information and resolution for vertical changes in measured parameters.
Also, a synergistic effect is achieved when certain logs are run together, potentially
providing unambiguous interpretation of hydrogeologic parameters, especially in the
vertical dimension.
This report covers borehole geophysical methods and addresses problems of site
characterization, contaminant plume detection and monitoring of contaminant plumes.
17 KEY WORDS AND DOCUMENT ANALYSIS
1 DESCRIPTORS
b IDENTIFIERSIOPEN ENDEG TERMS
C COSATI I-IeIdfCrOup
19 DISTRIBUTION STATEMENT
RELEASE TO ueuc
(9 ECURITYCLAS5fThisRepos’1)
UNCLASSIFIED
21 NO OF PAGES
80
20 SECURITY CI.ASS (Thisplge
UNCLASSIFIED
22 PRICE
EPA Fow, 2220— I ( .v —77) PREVIOU$ EDITION IS OOSOLEIC j
-------
NOTICE
The information in this document has been funded wholly or in part by Ike United
States Environmental Protection Agency under Cooperative Agreement #CR810052 to the
Desert Research Institute, University of Nevada System. It has been subject to the Agency’s
peer and administrative review, and it has been approved for publication as an EPA docu-
ment. Mention of trade names or commerical products does not constitute endorsement or
recommendation for use.
ii -
-------
ABSTRACT
The complex nature of the ground-water contamination problem requires the collection
of extensive amounts of data in order to understand the problem well enough to recommend
and execute the appropriate remedial action. As the complexity and consequences of ground-
water contamination increase, geopysical methods are becoming a cost effective approach to
providing answers to hydrogeologic questions associated with ground-water contamination
Geophysical methods applicable to hazardous waste site i.ivestigations can be broken
into two categories surface and subsurface methods Surface ,iuetliods offcr the advantages of
relatively little capital investment at the site and rapid collection of data over a horizontal
area. However, the interpretation is often ambiguous and limited in vertic ! resolution. Sub-
surface methods can be used only to investigate an area immediately around the borehole
However, subsurface methods provide excellent information and resolution for vertical changes
in measured parameters. Also, a synerg.stic effect is achieved when certain logs are run
together, potentially providing unambiguous interpretation of hydrogeologic parameters, espe-
cially in the vertical dimension.
This report covers borehole geophysical methods and addresses problems of site charac-
terization, contaminant plume (letection and monitoring of contaminant plumes.
Borehole tools and interpretation methodology have been (levelopc(l primarily for the
petroleum ina istry. The environment of the typical hiazardo is waste site is very different
from the environment encountered in petroleum exploration, and the parameters of interest
are very different for the typical hazardous waste site investigation As a result, most boreliohe
tools and virtually all interpretation algorithms for borehole methods are not directly applica-
ble to hazardous waste investigations. The methods developed by the petroleum industry
require modification, and in most cases, new interpretation schemes must be developed
We have developed an interpretation strategy designed specifically to extract data of
hydrogeologic interest from the subsurface. The borehole tools chosen for this strategy include
natural gamma, gamma density, inductioi, (apparent electrical conductivity of the formation),
televiewer, and thermal horizontal flow meter. The data collected using these five tools are
combined so that hydraulic conductivity, ground-water pore velocity, pore fluid conductivity
(a measure of total dissolved solids), and anisotropy can be obtained These data are
obtained as a function of depth, thus providing valuable detailed information as input to con-
taminant transport models
Laboratory experlrne.its conducted to determine the arcuracy of the thermal ground-
water flow meter show that the flow meter can be used to measure ground-water velocities
down to 0 5 meters per day. The results also show poor correlation bet cen e, perimeuits con-
ducted in the t-tube calibration and in the large sand box As a result of this poor correla-
tion, theoretical work was conducted to allow the ground—water 110w meter to l)e umsc(l without
iii
-------
laboratory calibration and to show h w the flow meter can theoretically be used to obtain
aquifer hydraulic conductivity, as well as flow rate and direction.
iv
-------
CONTENTS
Abstract . iii
Figures . viii
Tables ix
Acknowledgments x
1. Introduction I
Monitoring of Ground-water Quality at Hazardous Waste Sites 1
The Problem of Contaminant Transport and Hazardous Waste 1
Site Characterization 2
Contaminant Plume Detcction 3
Monitoring of Contaminant Plumes 3
Purpose and Scope 3
Approach 4
2. Borehole Geophysical Methods S
Traditional Use and Purpose of Borehole Geophysical Methods 5
Effects of the Borehole Environment 6
Introduction 6
Effects of drilling fluids 6
Casing type 8
Summary S
Traditional Borehole Logging Methods S
Electrical-magnetic logs 10
spontaneous potential 10
single point resistance 12
induced current logging 12
Nuclear logs 12
gamma ray logging 16
active gamma (gamma-gamma) logging 16
neutron logs 16
Acoustic logs 18
V
-------
Temperature logs • 19
Televiewer 19
Multiple hole techniques 19
Nontraditional subsurface sensing 10
3. Limitations of Borehole Methods for Hydrogeologic Investigations 20
Introduction 20
Equipment Limitations 20
Limitations of Interpretation Schemes 21
4. Borehole Logging Interpretation Strategy for l- 1 ydrogeologist 22
Introduction 22
Objectives 23
Hazardous Waste Site Investigations 23
Site Characterization 24
Interpretation Strategy 24
Porosity 24
Hydraulic conductivity and regional ground-water velocity 26
Hydraulic anisotropy 27
Lithology 29
Cation exchange capacity 29
Pore fluid electrical conductivity 29
Computer Implementation 32
Further Refinements 32
Effects of drilling disturbances 32
Tool response 32
Improved electrical conductivity model 33
Effects of contaminants 33
Conclusions 33
5. Use of Borehole Thermal Flow Meter for Determination of Ground-water
Velocity and Hydraulic Conductivity 35
Introduction 35
Description and Theory of Operation 36
Potential Problems 40
Goals and Objectives 40
Design and Construction of the Sand Box Flow Chamber 42
Sum*nary of Experiments 42
Experimental Results 45
vi
-------
Experiments using the 5 -cm end cap in 5-cm well casing . 45
Experiments using the pneumatic parker 47
Experiments using the fuzzy packer 51
Instrument accuracy and repeatability 51
Theoretical Development 55
Introduction 55
Direct calculation of aquiler fluid velocity 55
Determination of hydraulic conductivity using the thermal flow meter
60
Potential problems and limitations 61
Summary and Conclusic.ns 62
Experimental work: velocity magnitude calibration 62
Experimental results: directional accuracy 62
Theoretical work 62
6. Summary and Conclusions 64
Introduction 64
Borehole Geophysical Methods 65
Limitations of Borehole Methods for Hydrogeologic Hazardous Waste Inves-
tigations 65
Borehole Logging Strategy for Flydrogeologists 66
Use of a Borehole Thermal Flow Meter for Determination ci Ground-Water
Velocity and Hydraulic Conductivity 66
References 68
vii
-------
FIGURES
2.1 Change of pore fluid in vicinity of borehole. 7
2 2 Electrode placement for spontaneous potential log- 13
ging.
2 3 Simulated geophysical logs. 14
2 4 Schematic of induction tool operating. 15
2.S Schematic of formation density-compensated sonde. 17
4 1 Interpretation strategy. 25
4.2 The relationship between the velocity and the gra- 28
dient for an anisotropic aquifer.
4.3 Lithology cross plot. 30
5 1 Operating principle of the borehole thermal flow 37
meter.
5.2 Diagram of the 5-cm end cap and probe. 38
5 3 Diagram of the packers for the 10-cm borehole. 30
5 4 Diagram of the T-tube calibration chamber 41
5 5 Construction details of the sand box 43
5 6 Placement of the well casings in the sand box. 44
5.7 T-tube calibration of the 5-cm end cap. 48
5.8 Sand box calibration of the 5-cm end cap. 49
5.0 T-tube calibration of the pneumatic packer. 50
5.10 Sand box calibration of the pneumatic packer. 52
5.11 T-tube calibration of the fuzzy packer. 53
5.12 Sand box calibration of the fuzzy packer 54
5 13 Streamlines moving around and through a borehole 58
packer for different hydraulic conductivity ratios
5.14 Tangent streamline for determination of aquifer 50
fluid velocity and hydraulic conductivity
v i i i
-------
TABLES
2.1 Borehole sensing techniques applical)le to various 0
borehole environments
2.2 Borehole sensing methods. 11
5.1 Physical and hydraulic parameters of the sand box. 45
5 2 Summary of experiments. 46
ix
-------
ACKNOWLEDGMENTS
The authors wish to thank the U S Environmental Protection Agency, Environmental
Monitoring Systems Laboratory for funding this project Mr. Leslie McMillion provided valu-
able advice and guidance as Project Officer. Mr. Charles Russell was very helpful with the
laboratory experiments. Finally, we wish to thank the reviewers, E)r Eileen Poeter of Wash-
ington State University, Dr Mark Stewart of the University of South florida and Dr. Aldo
Mazzella of U S EPA Environmental Monitoring Sytems Laboratory for their valuable com-
ments and suggestions which improved the manuscript
‘C
-------
SECTION 1
INTRODUCTION
MONITORING GROUND-WATER QUALITY AT HAZARDOUS WASTE SITES
Monitoring of ground-water quality at hazardous waste sites is a subject of intense
interest in the U S Environmental Protection Agency (EPA) This interest has been sparked
by results of several national surveys that conclude that ground-water monitoring at both
Superfund and Resource Conservation and Recovery Act (RCRA) waste disposal sites falls
short of acceptability. Problems identified during a EPA study ol 22 RCRA sites (R. Murphy,
EPA-Washington DC, personal communicat on, 1985) include:
o Wells at about 50 percent of the sites are screened incorrectly. Consequently, contam-
inant plumes that intersect the wells are not i itercep ed by the screens
• \Vells at 30 percent of the sites are incorrectly placed, .so Lhat the contaminant plume
does not intersect the well at any depth.
• For 10 percent of the sites, placement of monitoring well& occurred before the direction
of ground-water flow had been determined.
A more complete understanding of the site hydrogeologic setting would minimize such
problems. Subsurface information is required to determine hazardous waste site hydrogeolo mc
conditions, locate contaminant plumes, and monitor sites for leaks from disposal facilities.
Borehole geophysics, in conjunction with surface geophysics and other geologic and hydrologic
techniques, can aid in determining site hydrogeologic parameters, particularly in the third
dimension. Borehole geophysics can contribute significantly to our understanding of lithology,
porosity, and structure; zones of saturation; physical and chemical characteristics of fluids;
and ground-water flow speed and direction.
TIlE PROBLEM OF CONTAMINANT TRANSPORT AND HAZARDOUS WASTE
Until the early 1070’s, the science of hydrogeology was primarily concerned with water
supply, which is often referred to as the “ground-water quantity problem”. This problem is
solved by determining the piezometric head or water table elevation distribution in time and
space in response o stresses on the aquifer system. Aquifer tests to determine hydraulic
parameters have been in routine use for more than 20 years Numerical models that provide
predictive models for changes in head di. tributiomi are also used routinely.
1
-------
Concern over the contamnation of ground-water supplies has changed the primary
emphasis of hydiogeology from roblems of ground-water quality problem to problems of
ground-water quality. The ground-water quality problem concerns the transporL of solutes
(contaminants) in ground water To (Ictermine ruLes of contaniiuiant. transport, ground-water
velocity must be known. l!owever, velocity data are normally obtained by indirect methods.
This is true for measurement of velocity in the field and e.ilcui1at oii of velocity in models.
Variations in ground-water velocity owing to the heterogeneous nature of the aqdifer cause
dispesion of the contaminint Thus the contaminant will have a velocity different from the
average velocity of the ground water. The difficulty of obtaining accurate velocity measure-
ments, as well as other difficulties, makes solving the ground-water quality problem much
more difficult than solving the quantity problem In addition, contaminated ground water at
hazardous waste sites is likely to contain numerous contaminants, making the problem consid-
erably more complex.
Understanding the complex nature of the ground-water contamination problem well
enough to recommend and execute an appropriate remedial action requires that extensive
amounts of data be collected. ft is nearly impossible to collect adequate amounts of data
using traditional hydrogeologic methods, consequently, new technology is needed.
Geophysical methods have been widely used in oil and mineral exploration since the
1920’s However due to their cost and the relativc. simplicity of most previous ground-water
problems, geophysical methods have not commonly been used As the complexity and conse-
quences of ground-water contamination increase, geophysics is becoming a more cost effective
approach to answering the hydrologic questions associated with ground-water contamination.
A possible goal of a contaminant study is to recommend remedtal action that will either
clean up the contaminated ground water or prevent it froni entering the biosphere. However,
before remedial action can be taken, it is necessary to understand the e,ctent. of the problem
and the hydrogeologic conditions present at the contaminated site There are many ways to
approach this problem, but from the standpoint of RCRA, the problem can be divided into
three categories: site characterization; contaminant plume detection; and monitoring of con-
taminant plumes
SITE CHARACTERIZATION
The characterzatton of a contaminated (or potentially contaminated) site consists pri-
marily of determining local hydrogeology and contaminant l)lIinlC (histr,l)ntion. Most sites are
small compared to the regional hydrogeologic system, and the regional gradient can be super-
imp ised on the local system to obtain a general flow direction. Structural geologic information
can provide a picture of the geometry of the contaminated aquifer. Aquifer tests provide
information on the horizontal hydraulic conductivity and storativity of the aquifer The
storativity is unnecissary from the standpoint of contaminant transport, because the fluid
mass conservation equation is usually solved as a steady-state equation. The hydraulic con-
ductivity data obtained from aquifer tests are of limited value because of the spatial variabil-
ity of aquifer properties, which greatly affect the transport of contaminants Of particular
importance is the hydraulic conductivity variation in the vertical direction Aquifer tests are
of limited use for vertical variations because they produce hydraulic conductivity values that
are integrated over the vertical domain. Testing of individual strata can be done if the strata
2
-------
are horizontally extersive, but this is most often not the case in shallow, unconsolidated sedi-
ments In an anisotropic aquifer, the solute can be transported in a direction different from
the gradient, thus anisotropy is a very important, parameter for contaminant studies. hleas-
urement .i the hydraulic conductivity tensor requires a minimum of three observation wells
The aquifer must be homogeneous over the area of these observation wells, which is usually a
pcor assumption for shallow, unconsolidated sediments There is a clear need to develop
methods that vill be useful in determining spatial variability and anisotropy for hydraulic
conductivity.
CONTAMiNANT PLUME DETECTION
Mapping the location and distribution of contamnimiants in ground water (“eontaniinant.
plume cletectiomi”) is a difficult. problem l)ecaimse of tIme inaccessibility to time affected environ-
ment. Direct sampling methods are limited to monitoring vells ‘l’o adequately (letermmne con-
taminant concentration and plume migration, a large number of costly monitoring veils must
be drilled.
MONITORING CONTAMINANT PLUMES
Ground-water monitoring is required under RORA regulations at hazardous waste facili-
ties. A successful monitoring network will detect a leak from a permitted hazardous waste
faclity or detect changes in ground-water quality in an aquifer near a hazardous waste site. A
well-designed monitoring network should conform to a number of important criteria. Zones
of high permeability should be located and specifically monitored, because they are most
likely to contain the first arrival of a contaminant. The directional component of the
ground-water velocity will be the primary consideration in locating up- and down-gradient
monitoring wells Often, monitoring of contaminant plumes is thought of ia terms of looking
for certain parameters. In this sense, a parameter is a property or characteristic of the ground
water or the aquifer itself which has been selected for monitoring because it is indicative of a
condition or state of contamination. The chemical, physical, and electrical properties of the
ground water are important in determining what. parameters should be monitored
PIJT1POSE AND SCOPE
Geophysical methods applicable to hazardous waste site iiivcstigatioiis can be broken
into two categories: surface and subsurface methods. Surface methods offer the advantage of
relatively little capital investment at the site (no borehole is required), and rapid cdllection of
data over a horizontal area. however, the interpretation is often ambiguous and limited iii
vertical resolution Subsurface methods require a borehole and can only be used to investigate
an area immediately around the borehole. however, they provide excellent information on
vertical changes in measured parameters Also, a buite of complementary logs can potentially
provide unambiguous interpretation of hydrogeologic data.
The two approaches complement each other very well The subsurface methods provide
the necessary vertical detail for a small area, and the surface methods are used to extend this
detail horizontally between boreholes
In this report, the problems of site characterization, contaminant plume (letection and
monitoring of contaminant plumes are addressed using borehole geophysics ‘The overall goals
3
-------
of this project are to:
1. Determine which traditional and iv ’ntraditional borehole methods are best for addressing
the problems of site characterization and plume detection and monitoring.
2. Develop a borehole logging interpretation strategy that can be used at hazardous waste
sites and that is designed specifically to address the determination of hy”rogeologic
parameters.
3. Develop procedures for using the borehole thermal flow meter to obtain ground-water
velocity and hydraulic conductivity data
Originally, this report was to include procedures for ill borehole tools recommended in
the borehole logging interpretation strategy. Because of problems beyond the control of EPA
and the Desert Research Institute, University of Nevada System (DR !), the only logging tool
available to fully test was the borehole thermal flow meter. The procedures for using the rest
of the recommended suite of borehole logs with the interpretation strategy will be covered in
future reports, after th logging tools have become available and the strategy has been proven
to work, or has been refined as necessary.
APPROACH
Section 2 provides a description of the traditional and nontraditional borehole logging
methods, including principles of operation, discussion of the properties that the tools measure,
and the parameters that the log analyst normally obtains from interpreting the logs.
Emphasis is placed on the objectivcs of traditional borehole logging. More information is pro-
vided for the tools that will be singled out as useful for the interpretation strategy outlined in
Section 4.
Section 3 discusses the limiations of traditional borehole logging techniques The discus-
sion provides the reader with the reasons why lb is necessary to develop an entirely new and
different borehole logging strategy for hydrogeologic applications.
Section 4 outlines the interpretation strategy. The strategy that is presented is only one
alternative. Actual use of the strategy has not yet occurred, due to lack of equipment. The
strategy will be refined and improved based on the experiences of actual use.
Section 5 presents the results of a detailed set of calibration procedures that were con-
ducted with the borehole thermal flow meter. Recommendations for its use are presented.
Procedures for the use of the flow meter to determine aquifer hydraulic conductivity are
developed. The hydraulic conductivity procedures have not been tested, and their verification
will be the subject of a future report.
Section 6 presents the project summary, conclusions and recommendations.
4
-------
SECTION 2
BOREHOLE GEOPHYSICAL METHODS
TRADITIONAL USE AND PURPOSE OF I3OREIIOLE CEOP’IYSICAL METHODS
A widely accepted definition of borehole geophysics can be found in Keys and MacCary
(1071):
Geophysical well logging, also called borehole geophysics, includes all techniques of
lowering sensing devices in a borehole and recording some physical parameter that
may be interpreted in terms of the characteristics of the rocks, the fluids contained
in the rocks, and the construction of the vell
This defini ion reflects the historical development and use of borehole geophysical methods.
The hist known use of a borehole method was to measure the borehole fluid temperature
as a function of depth (llallock, 1807). The first major borehole method to be developed was
in the area of borehole electrical resistivity by Schlurnberger (1920) The work was done in an
oil field in France, and most borehole methods were subsequently developed in response to
needs of the oil industry
The use of borehole geophysics by the water well (or ground-water supply) industry has
been limited by a number of factors. The cost/benefit ratio of using borehole methods in pure
ground-water supply investigations has traditionally been unfavorable compared to using
borehole logging in the field of petroleum exploration Most water wells tend to be shallow
relative to oil wells, and borehole methods have been developed for deep well application In
the oil industry, borehole methods have been used to help locate new wells to maximize pro-
duct recovery The criteria for locating water wells are more often related to cultural neces-
sity, such as locating the well near population centers, and locating shallow water sources to
minimize pumping costs.
In recent years, considerable attention has been focused on the problems of ground-water
contamination, especially at hazardous waste sites. The amount of money being spent on the
clean up of hazardous waste sites is very large, especially compared to traditional ground-
watts supply investigations As a result, borehole geophysical methods are no longer con-
sidered to be relatively costly This has created a need to develop borehole tools and interpre-
tation strategies that can be used in ground-water contamination studies.
Some of the basic reasons for doing geophysical well logging include
5
-------
I. To allow visual interpretation and comparison at the well site.
2. Immediate decisions can be made regardng screening intervals.
3. Well logging is superior to and less expensive than continuous coring (an alternative way
to obtain subsurface information).
4. Well logging provides information on lithologic boundaries.
A well log itself is of limited use without applying sonic sort of interpretation to the raw
data. Nearly all log interpretation in the petroleum industry is designed for open hole wells,
the logging done prior to installation of casing
The objectives of open hole log interpretation include:
1. Remove drilling, mud and borehole effects from logging data.
2. Determine rock properties, such as porosity, hithologic traps, and absolute, relative and
effective permeability.
3. Determine formation fluid properties such as oil and gas specific gravity, hydrostatic
pressure, fluid saturation distribution and temperature.
4. Obtain information on the formation factor, which leads to determination of porosity,
permeability and pore fluid conductivity.
EFFECTS OF THE BOREHOLE ENVIRONMENT
Introduction
The methods of drilling and completion of a borehole will affect the types of geophysical
logs that can be run and the interpretations of the data from them. If the hole is in unconsoli-
dated aadiments, the typical case in the vicinity of many hazardous waste sites, the matrix is
friable and collapses easily. One way of coping with this condition is by adding various
materials to the drilling fluid; usually the weight of the column of drilling fluid is greater than
the pressure of fluids in the matrix rock. This achieves mechanical stability of the hole by
applying backpressure to the formation; that is, stability is achieved at the cost of causing the
drilling fluid to enter the pores of the formation This immediately changes the physical pro-
perties of the formation near the borehole, which is highly undesirable if the purpose of mak-
ing the hole is to ascertain the physical properties of the formation and/or its fluid contents.
The type of material used to case a hok ‘s important. Typical mateiials include steel,
aluminum, PVC or other plastics, and Teflon. Whether the borehole is air or liquid filled will
also affect the geophysical methods available for usc in them
Effects of drilling fluids
After the hole is drilled, the logging tools are lowered into the drill ‘ ole to make nicas-
urements that are hopefully representative of the formation. Such observations must be
adjusted for the changes which have been created in the vicinity of the hole by the drilling
operation. As an illustration, consider Figure 2 1. The electrical resistivity (a very significant
indicator of the type and amount of liquids in the pore space) changes considerably as a func-
tion of radial distance from the axis of the drill hole The drilling fluid has entered into the
formation since it is, for safety reasons, at a slightly higher pressure than the ambient forma-
tion pressure. Some of the material in suspension has been filtered out at the surface of the
sidewall to form a mudcake. The rest of the drilling fluid, which is called the filtrate,
6
-------
FIGURE 2.1. Change of Pore Fluid in Vicinity of Borehole.
BORE HOLE
•0
IMPERMEABLE BED
I
I
0•
/
IMPERMEABLE BED
7
-------
continues penetratiag into the formation; the extent f the formattor. so penetrated is called
the invaded zone The filtrate varies throughout the invaded zone to the percentage of pore
space that is occupied. If the filtrate occupies all of the pore space, then that portion of the
invaded zone is termed the flushed zone; the remaining portion of the invaded zone is called
the transition zone Note that the electrical resistivity of the mud is intermediate n value
between the resistivity of the mudeake and the resistivity of the filtrate
The size of tIre flushed zone or the invaded zone will change over time. Therefore, it is
important to correct the geophysical and geochemical observations for the conditions which
existed at the time of data collection This evaluation is especially important in the monitor-
ing phase of hazardous waste site operations because temporal changes in borehole logs might
be mistakenly attributed to contamination instead of borehole environment changes.
Casing type
The usual material used for casing a hole is steel due to its strength; however, for shah.
low holes, those less than a few hundred meters, other tubing such as PVC, Teflon, or other
plastics can be used This is especially appropriate to monitoring operations in which the hole
is to bc used repeatedly over time Many of the logging techniques cannot be applied in a hole
cased with metal.
Summary -
Table 2.1 shows the various geophysical methods which are appropriate to the wet or
dry tad cased or uncased hole. This is not meant to be an exhaustive presentation, simply
illustrative of the constraints of the hole conditions on the geophysical information that can
be obtained from the hole. Note that the radiation techniques are the most versatile, being
meaningful in an open hole or a cased hole. When used in a cased hole, interpretation is easier
and better if the same log has been run in the same hole before it was cased The induction
techniques can be used in place of contact resistivity techniques for the typical plastic-eased
holes found in hazardous waste site investigations. Geochemical methods are sometimes used
in liquid-filled holes to sample the ambient liquids hi information on depth dependence is
sought, then much more complicated instrumentation is usually required.
TRADITIONAL BOREHOLE LOGGING METHODS
Descriptions of the various borehole techniques and equipment are given below. The
descriptions are rather short because the equipment and interpretation techniques are open
hole, and their application in hazardous waste investigations is limited Enough detail is pro-
vided to set the stage for the discussion of limitations i cr Section 3 Numerous references are
available that provide detailed instructions on the use of borehole logging tools and the
interpretat’on of opei hole logging data. See for example Schlumberger (1972), Asquith
(1982), llilchie (1982) Dresser Atlas (1982) and Keys and MacCary (1071). In Section .1, an
interpretation scheme is developed specifically for application to ground-water investigatnns
at hazardous waste sites, and the borehole methods useful for this application are discu&’ d
individually.
Borehole methods fall into five major categoriew acoustical, electro- magnetic, nuclear,
flow and dimension and thermal Major applications of these techniques include hithologie
correlation, lithology, rock density, fractures, porosity, permeability, flow, water level, water
8
-------
BOREHOLE SENSING
TABLE 2.].
TECHNIQUES APPLICABLE TO VARIOUS BOREHOLE ENVIRONMENTS
Logging Techniques
Single Well
Cross Borehole
In Situ
PVC •
Cased
Stoel
Cased
Uncased
PVC •
Cased
Steel
Csoed
Uncaeed
PVC •
Cased
Stool
Cased
Uncased
WET
DRY
WET
DRY
WET
DRY
WET
DRY
WET
DRY
WET
DRY
WET
DRY
WET DRV WET
DRY
ACOUSTIC
0
0
ELECTRIC
Q
0
0
‘
•
0
0
INDUCTION
0
0
0
0
0
0
0
0
NUCLEAR
0
0
0
0
0
0
FLOW
e
0
0
0
0
0
0
0
0
TEMPERATURE
0
0
0
CHEMICAL
0
0
0
0
0
0
Note, PVC Is shown but other plastic casings (le, teflon) behave similarly.
-------
quality, temperature gradient and hole diameter. l’able 2.2 is a summary of borehole objec-
tives and the methods used to achieve them
Hardware for borehole geophysical logging consists of similar basic components for all
the different tools, consisting of sensor, signal conditioners, and a recorder. The sensor or
sonde receives power and transniits the signal to the surface through a conducting cable,
which also serves to position the tool in the hole by means of a winch.
Electronic controls at the surface regulate logging speed and direction, power to the downhole
electronics, signal conditioning, and recorder responses.
Two types of signal conditioning and recorder responses are available: analog and digital.
Most systems that are equipped with digital processing and storage are also equipped with
analog recording for purposes of backup aiid examination of the data during logging. The
return signal from the probe is a function of lithiologic, fluid, and borehole parameters and is
recorded and later analyzed with a computer.
Electrical-magnetic logs
Electric and magnetic logs provide informRtion about lithology, saturation, fracture loca-
tion, fluid movement in the borehole and fluid conductivity. There are many types of resis-
tivity and resistance tools in use.
Electric logs are a record of electrical potentials and rcsistivities These logs can gen-
erally be run only in uncased holes that are filled with a conducting fluid. There are many
types of electric logs which can be run in a borehole, the choice of which is dependent upon
the purpose of the survey. Some electric logs make use of the natural electrical currents that
exist under certain conditions. Most common is the spontaneous-potential (SP), which is
described below.
Other electric logs induce a current to flow in the formation for the purpose of recording
resistance and/or resistivity. There are numerous types of these logs, including single point
and difT ntial resistance, short normal, long normal, lateral, microlog, microfocus log, and
the guard or lateral resistivity log. Each of these logs has a specific application, depending
upon the hithology, depth of drilling mud invasion, and area of interest within the borehole It
is not necessary to run all of these electric logs to determine variations in lithology, aquifer
locations, and relative conductivities of porous media fluids.
spontaneous potential—
The SF logs measure the small differences in natural potentials that develop in the
borehole as a result of fluid movement and electrical potentials due to borehole fluid migra-
tion at hithologic contacts. SF logs can be used to delineate fault zones and water producing
units in the borehole.
Keys and MacCary(197 1) point out that the most important source of SF is due to the
electrochemical electromotive force (emi) produced at the interface between dissimilar materi-
als in the borehole. SP may also be generated in zones of gaining or losing water. ‘l’lic Sl log
can also be used to delineate relative hydraulic conductivities based on the knowledge of
borehole fluid potentials. When the water within an aquifer unit is more conductive than the
borehole water, the emf generated from the streaming potential or from the Junction of dis-
similar materials causes a current to flow from the borehole into that ui it As the SP elec-
trode moves upward, it senses a decreasing potential because the current flows parallel to the
10
-------
Table 2 2
BOIIEFIOLE SENSING METhODS
OBJECTIVE 1 I3OREUOLE METUODS
Electric Log (fully saturated)
Location of Zones Temperature log
of Saturation Neutron log
Gamma-gamma log
Electric log
Physical and Chemical ‘lemperature log
Characteristics of Fluid conductivity log
Fluids Spontaneous potential log
Specific ion electrodes
Fiber optics
D.O., Eh, ph probes
Formation resistivity log
Stratigraphy !nduce polarization log
and Porosity Natural gamma log
Spectral gamma log
Thermal neutrcri log
Cross borehole radar
Cross borehole shear
Resistance log
Acodstic -
Transit time log
Acoustic -
Waveform log
Neutron log
Induction log
Spontaneous potential log
Flow meter
Flow and Direction Tracer
D,fferentia temperature log
Water level
11
-------
well bore. This relationship would be reversed in the case of good quality water within an
aquifer unit and one would see a positive SF response opposite the aquifer unit. Thus a nega-
tive or positive response on the SF log can be used to infer relative fomiat’on water qualities
if the borehole fluid conductivities are known.
The electrode placement for SF logging is shown in Figure 2 2; one electrode is embed-
ded in the surface material and the other is lowered down the hole Figure 2 3 is a suite of
simulated geophysical logs including SF indicating their response to different hydrostrati-
graphic units A more detailed discussion of SF logging methods can he found in llil-
chie(1982)
single point resistance--
Single point resistance logs measure the resistance to current flow of the borehole fluid
and material The logs are useful in geologic correlations because of their unique response to
changes of lithology Fractures or zones of weakness can he mapped with resistance logs, in
addition to changes in borehole fluid conductivitues. The resistance log can be used as a sensi-
tive caliper log and as a water quality indicator The SF and resistance logs are usually inter-
preted as a unit in any one borehole.
induced-current logging--
The methods previously described have required a conductive fluid because the current
flows out. into the formation directly from the electrodes If it is necessary to survey a hole
which contains a non-conductive medium such as air or hydrocarbon, electrical investigations
can be carried out by inducing currents in the formations surrounding the borehole. This
technique is called induction logging, and the theory is analogous to mduction techniques used
on the surface of the earth. This method can also he used in non-metallic cnscl holes.
Figure 2.4 shows a diagram of a conceptual induction tool One coil is used for transmit-.
ting, the other for receiving. An alternating current of constant intensity is passed through
the transmitter coil at a very high frequency This electric field has an associated magnetic
field which, in turn, induces electric currents n the formations. The phenomenon is linear for
liractical ranges of current levels, so the indiicc,l currents will lie of the same frcqucncy as the
source current, but different in intensity and phase Associated with the ground currents
flowing around the hole is a magnetic field, which induces current to flow in the receivei coil.
Of coursc, the magnetic field associated with the transmitter coil also induces current to flow
in the receiver coil, but this can be filtered out by cahihration procedures in an ideal known
environment, e.g., air. In actual practice, the incliictioii tool will have more than two coils,
usually six or even more Additional coils provide selective focusing of the transmitted signal
so that the measurement is weighted away from the borehole environment and its effects.
Figure 2 3 includes a hypothetical inverse induction log
Nuclear logs
Nuclear logs measure radiation emitted from the nuclciis of an atom and have a distinct
advantage over most other geophysical logs in that they can derive inforrn.ition from cased or
uncased holes that are filled with any type of fluid. The most common nuclear logs are
natural gamma, gamma gamma, and neutron-neutron The primary function of these logs is
to determine variations in lithology, density, and moisture content or porosity within the
borehole materials.
12
-------
Ø%Ø
00
FIGURE 2.2. Electrode Placement for Spontaneous Potential
Logging.
13
-------
SIMULATED GEOPHYSICAL LOGS
FIGURE 2.3. Simulated (‘ .cophyslcal Logs.
14
-------
Borehole
Line of magnetic field
of Induced current
Line of magnetic
field of transmitter
FICUR 2.4.
Scher. atic of Induction Tool Operating. The Transmitter
anc Receiver Coils are in a Borehole whereas the In-
duced Current is in the arth.
(e)
/
/
I
(d)
I
Induced
Current
(b)
,-
\
(a)
/
15
-------
gamma-ray logging--
This is a pas .sive technique which uses a scintillation counter to measure the natural
gamma radiation emitted by the matcrials in the vicinity of the so.ide Because the gamma
ray emission is statistical in nature, it. is appropriate to average the counts over time in order
to obtain a stable value ‘I’he duration for averaging depends upon the level of radioactivity,
lower levels requiring longer times A few seconds arc normally adequate in practice. l)ue to
this averaging, any anomaly will be shifted in the d re tion in which the tool is moving
Gamma-ray logg ig can be performed i i a cased hole, If absolute values are desired, ‘hen
corrections should be made for the density of the mud, the diameter of the hole, casing Pro-
perties, etc.
Because radioactive elements tend to concentrate in shales and clays, the gamma-ray log
is usually interpreted to indicate the shale content in sedimentary formations A typical paz-
sive gamma-ray log is shown in Figure 2 3.
active gamma (gamma-gamma) logging—
In gamma-gamma logging the scintillation counter is packaged into a sonde with a
radioactive source, as shown diagrammatically in Figure 2 5(a) The sonde is pushed against
the sidewall to reduce or eliminate the effects of the mud The source emits medium-energy
g iinma-rays which pass into the formation where they collide with electrons and are scat-
tered. Some of these scattered gamma-rays are detected by the scintillation counter lccated at.
a fixed distance from the source. Thus, the active gamma-ray technique depends on the elec-
tron density of the formation. Since this is directly related to the bulk density of the forma-
tion, this technique is usually called the formation density log.
To circumvent the prob’ m of correcting for the mudcake, a sonde with one radioactive
source and two detectors can be used. Assuming that the mudcake conditions are the same at
both the detectors, which are at different distances from the source, then the dili’erence
between the readings will be directly related to the formation electron density and, hence, the
bulk density. The difference is independent of both the mudcake density and its thickness if
both detectors encounter the same conditions Such a sonde, having two detectors and one
source, is depicted in Figure 2 5(b). This is called the Formation Density Coripensated (FDC)
tool.
neutron logs—
The configuration of the probe in neutron logging is oriented so that the tool responds
primarily to a function of hydrogen content in the borehole environment. Neutron energy is
modified by elastic collisions with borehole elements and t ie effectiveness of these energy
moderators is a primary factor in neutron logging The mass of the nucleus of a hydrogen
atom has approximately the same mass as the neutron, so a coilisbn of a neutron with a.
hydrogen atom causes the neutron maximum energy loss
Consequently, the energy of the neutron response is interpreted for moisture content iii the
unsaturated zone of the borehole and the total porosity below the water table Neutron logs
can shcw the location of porous aquifer units, perched ater tables, and , round-water
confining zones, such as clay layers
16
-------
FIGURE 2.5 Schematic of Formation Density Compensated Sonde (b); Original Design
shown in (a).
Mudcake
(a) (b)
-------
Acoustic logs
The ability of a rock to transmit acoustical waves is a measurable rock property that is
useful in determining related properties of interest such as porosity and fractur;ng. The most
commonly used acoustical wave for these studies is a pressure wave that is generated by a
source in the borehole. In the simplest of tool designs, this wave is detected at a receiver
located in the borehole which is a short but constant distance from the source. Although
many travel paths occur, the instrument is set to record only the wave which is refracted
through the formation. Currently, instrumentation is available to measure both seismic velo-
city and the full seismic waveform.
Seismic velocity, as measured by borehole methods, is the speed at which a vertical trav-
eling pressure wave propagates and is dependent on the elastic properties of the formation
Since rock material and fluids have such vastly different elastic properties, the presence of
fluids will greatly effect the velocity. Wyllie(1956) derived an empirical relationship between
porosity and velocity:
I — a 1 n
V. — V 1 + V,,
where
V 1 = measured acoustic velocity
V,, = matrix acoustic velocity
V 1 = pore fluid acoustic velocity
— porosity
The matrix velocity (Vm) is the acoustic velocity within the matrix of the formation.
Since this is nearly a constant for most common earth materials, its value can be assumed
Pore fluid velocity (V 1 ) can also be assumed because it is nearly cor’stant for water. Hence
using this equation, the porosity can be determined. The advantage of using an acoustic tool
to measure porosity iastead of a tool with a radioactive source is the logistical simplification
of eliminating the radioactive source.
Acoustic amplitude can also be recoided by some equipment This equipment normally
records the full pressure waveform. Discontinuities in the formation such as fractures, will
influence the waveform amplitude. Although it. is not currently possible to quantify fracture
patterns using the waveform amplitude, at least an indication of fracture occurrence can be
obtained.
Acoustic logging is commonly done in an open hole. The presence of a casing creates
another travel path for the seismic ener and the instrumentation must be able to distin-
guish the arrival through the casing from the arrival through the formation.
Variations in borehole diameter will significantly effect both velocity and amplitude
measurements. Hence most tools employ two transmitters, one located above and one below a
pair of receivers This geometry allows processing of the signals so as to minimize errors due
to variations in the borehole diameter.
18
-------
Temperature logs
The temperature in a shallow well can eas ly be measured using an appropriate therrnis-
tor. This temperature may be used to correct other observed data, such as the conductivity
hog and the resistivity log. The temperature log can also be used to indicate borehole intervals
that are producing or accepting fluids This may be helpful in locating zones For more detailed
flow measurements with a borehole thermal flow meter (discussed in Section 5).
Televiewer
This tool is simply a television camera which is used to look at the interior of the
borehole and/or well casing. It can delineate and correlate fractures if used prior to casing
installation. For wells in which there is no casing information, the televiewer is useful for
inspecting perforations or slotting and determining their frequency and orientation. Slotting
information is necessary to properly use the thermal horizontal flow meter discussed in a Later
section.
Multiple hole techniques
Multiple hole techniques require several holes for implementation. Seismic methods are
commonly used betweer’ boreholes. A eismic source is placed in one hole, and seismic signals
are received in one or more other boreholes (Oal’perin, 1974). By placing source and receivers
at different depths in di(ierent holes, complex three dimensional interpretations are possible
Borehole-to-borehole radar is another commonly used multiple hole method. Operation-
ally, this method is similar to the multiple hole seismic techniques, with the seismic source
and receiver being replaced by an electromagnetic transmitter and receiver
Nontraditional subsurface sensing
In addition to the traditional logging methods developed by the petroleum and mineral
industries, there are a number of sensing techniques that. can potentially provide valuable
information for hazardous waste site investigations. Non-traditional techniques include geo-
chemical, thermal, moisture and vapor sensing methods Of particular interest is the ground-
water flow meter which senses ground-water flow by induced thermal pulses This instrtime t
is discussed in detail in Section 5. Sensors for many of these methods can be left in situ for
continuous monitoring with time. Table 2 2 presents a list of methods applicable to measuring
different properties of interest in hazardous waste site investigations and monitoring.
In situ methods may prove to be particularly useful for the problem of contaminant
plume monitoring Long term monitoring of borehole parameters will provide information on
the variance of the monitored parameters (from a stochastic point of view) and long term
trends of increase or decrease of the parameters. For many of the in situ methods, little addi-
tional cost is incurred to collect the data, once the instrumentation is in place.
19
-------
SECTION 3
LIMITATIONS OF BOREHOLE METHODS FOR HYDROGEOLOCIC
INVESTIGATIONS
INTRODUCTION
From the information provided in Section 2, it should be evident that considerable prob-
lems are encountered when trying to apply existing borehole methods to the hydrogeologic
investigations at hazardous waste s tes. The problems arise primarily from two considera-
tions•
1. The typicai environment in which borehole methods are commonly used in the
petroleum industry is very different than the environment that commonly exists at
hazardous waste sites.
2. The petroleum industry obtains information from the borehole logs that is quite different
from the information sought by the hydrogeologist for hazardous waste sites.
EQUIPMENT LIMITATIONS
The design considerations for borehole logging equipment have primarily been set by the
petroleum and mineral industry and are a direct result of the environment that the equipment
must function in. It is useful to contrast and compare the “typical environment” for a
petroleum well and a well installed for monitoring a hazardous waste site.
Oil wells are nearly always deep (hundreds to thousands of meters) Logging is done
immediately upon hole completion and often before completion, the hole is logged to aid in
determining how much further to drill. Once the hole is logged upon completion, it is seldom
logged again. Most oil wells are at least 30 cm in diameter The hole is filled with mud during
logging, and the mud has infiltrated into the formation.
There are a number of common aspects regarding the design of borehole equipment for
these environments. The downhole sondes must be able to withstand high pressure and tem-
perature because of the depth. Several logging tools are commonly installed in one long sonde.
This is done for two reasons: the large depths coupled with the desire to obtain a multi-log
suite; and the need to obtain multi-log suites that can be accurately correlated for depth with
respect to each other. Multi-log sondes can be as much as 5 m in length. F’or a 5000-rn
borehole, a 5-rn long sonde can be considered to be essentially a point Many of the logging
tools will only work in an open hole. Table 2 1 summarizes the kinds of environments that
20
-------
various types of logging tools can be ised in.
The typical borehole that exis s at (or around) hazardous wa. te sites is shallow (prob-
ably less that 100 m and typically about 30 m), narrow diameter (5 cm) and cased, usually
with PVC, teflon, or some other plastic. None of the borehole tools that have been designed
for petroleum well logging are usable in such an environment. Although a 5 m sonde would
fit in a 30 m borehole, its upper most tools vould miss logging the bottom portion of the
borehole. Unless the bole has not been cased, none of the open hole tools such as the electric
logging tools, can bt used. If the tools could be used, they would be over-designed for tem-
perature and pressure requirements, however the hazardous waste site ground water may con-
tain chemicals r r which the sondc was not designed to withstand.
LIMITATIONS OP INTERPRETATION SCHEMES
Most of the interpretation schemes that have been developed have been for ope’ hole
logging. The emphasis is on eliminating borehole mud and mud filtrate effects from the data,
and trying to determine formation contacts and other parameters that are useful to the
petroleum reservoLr engineer. The parameters that are of interest to the petroleum industry
are discussed in Section 2. The parameters that are of interest t ’ the hydrogeologist are del-
ineated in Section 4, along with an interpretation strategy to obtain them
21
-------
SECTiON 4
BOREHOLE LOGGING INTERPRETATION STRATEGY FOR
HYDROGEOLOGISTS
INTRODUCTION
The oil and mineral industries have been collecting and interpreting borehole logging
data for many years. The interpretation strategies that have been developed are naturally
oriented towards solving problems and answering questions that are of interest to these dis-
ciplines. Borehole logging data are potentially of great value to hydrogeologists for evaluating
hazardous waste sites. However, new interpretation strategies must be developed that are
designed to provide quantitative information on the formation parameters that are of hydro-
geologic interest
There is some overlap in the interests of the hydrogeologist and those of the oil or
mineral geologist. however, the collection and interpretation of logging data requires a com-
plete re-thinking in order to maximize the information that is useful for the study and predic-
tion of contaminant transport in aquifers.
The vertical variation in hydraulic parameters within an aquifer is recognized to be of
primary importance in determining the fate and transport of coiitamiuiants in ground-water
systems Traditionally, the process of hydrodynamic dispersion has been thought to be the
dominant process causing contaminant mixing Macro-scale heterogeneity and vertical
stratification induce large variations in the advective how rate of the ground water. This pro-
cess has been termed macroscopic dispersion (Schwartz, 1077, Smith and Schwartz, 1080), and
it is the dominant mechanism controlling contaminant mixing and transport in many aquifers
Largely because of macroscopic dispersion, traditional techniques of vertical ir ration
(Bear, 1079, Chapter 5) to der vc two-dimensional vertically-averaged ground-wate it equa-
tions are inadequate to describe contaminant transport in aquifers. Although it is important
to account for vertical variation in hydraulic parameters, there has been little effort to
develop adequate borehole methods that would provide such paiameters Aquifer parameters
that are important for contaminant transport studies include litliology, effective and total
porosity, tensor hydraulic conductivity, fluid velocity, cation exchange capacity, aquifer
dispersivity and total organic carbon
In a general sense, geophysical methods applicable to hazardous waste site investigations
can be divided into two categories, subsurface methods and surface methods. Surface
22
-------
methods have the advantage of not requiring a borehole and in general a large horizontal area
can be quickly covered. Unfortunately, the interpretation of surface methods is often ambigu-
ous and additional inrormation is usually required
Subsurface methods require a borehole and can only investigate an area immediately
around the borehole. They do however provide excellent informati’ n on vertical changes and
because of the wide variety of complementary measurements winch can be made, they can
provide a great amount of information. Obviously, the two approaches complement each
other nicely. The subsurface methods provide the necessary detail but only in a small area,
and surface methods extend this detail in the horizontal directions. This is not to say borehole
methods do not have pitfaLls but that an integrated approach of surface and borehole geophy-
sics along with more conventional techniques is desirable.
OBJECTIVES
The purpose of this section is to outline a borehole logging interpretation strategy for
hydrogeologists for use in the saturated zone. The development of a similar strategy for the
unsaturated zone may be part of a proposed future cooperative agreement. The objectives of
the strategy are:
I. Select a suite of logging tools that can be used to obtain the hydraulic parameters neces-
sary for contaminant transport studies
2. Develop a technique to quantitatively combine the information from the selected logging
tools so that the hydraulic parameters and their vertical variations can be obtained for a
specific hazardous waste site.
Originally, the strategy was to be tested in the field, but due to factors heyond the con-
trol of Dltl and EPA, the logging equipment to (10 tlic testing was not available The field
testing, and refinement of the strategy may be an important part of a proposed future
cooperative agreement between EPA and DRI.
hAZARDOUS WASTE SITE INVEST IGATIONS
Before a strategy can he developed, it is important to understand the environment in
which it will be used. Typically wells are shallow and details of perforations and mud logs are
poor or unreliable. Unlike traditional logging which is done in uncased holes just after hole
completion, logging at hazardous waste sites will normally have to be done in PVC-cased
holes with short perforated intervals Usually there are several wells in a relatively small area
which were initially drilled for monitoring but which ran be used for borehole geophysics.
Typically the site has an abundance of cultural activity which restricts the use of some sur-
face geophysical methods.
If borehole methods are to be of use for hydrogeologists, it is essential that they answer
questions of hydrologic significance. In particular this strategy strives to describe how the fol-
lowing parameters vary with depth: effective and total porosity; tensor hydraulic conduc-
tivity; lithology; ground-viater velocity; cation exchange capti.city of the formation, and elcctri-
oal conductivity of the pore fluid.
It is necessary to have a clear understanding of how to get from the measurements taken
in the field to our desired goals. The exploration ind ustries have developed sophisticated
23
-------
strategies to do this, but because they operate in different environments and have difTerent
goals, many of their well-developed and tested strategies cannot be applied to hazardous
waste site investigations. While relying on the same principles it is necessary to develop
modifications of equipment, field techniques and interpretation st.ategies The following sec-
tions explain a strategy which can be applied at a hazardous waste site.
SITE ChARACTERIZATION
Hazardous waste sites are located in every conceivable geologic setting Each one is
unique and relationships developed for one site cannot he considered valid elsewhere It is
essential that relationships used in interpretations be based on data collected at the site under
study. To do this, it is necessary to drill characterization holes at each site.
The characterization holes should be drilled with a technique than allows good core sam-
ples to be taken These cores will be analyzed for li hology, hydraulic conductivity, effect 1 ve
porosity, and cation exchange capacity This information will be combined with the well logs
of the hole to provide the necessary site speciII relationships for interpretation of the other
wells from which cores are not available. Although the characterization wells do not provide
an absolute calibration of the logging tools, it permits the tool response to be related to the
local conditions.
Obviously, the characterization holes must be representative of the site It may be neces-
sary to have several char cterization holes and different relationships must be developed in
different formations if th’ geologic environments were significantly different during their for-
mation. Because most hazardous waste sites are small and the geologic processes that formed
them are usually consistent across them, the extrapolation of relationships across the site is
usually valid. By comparing the relationships developed for a given unit with data from
different wells the validity of this assumption can be checi ed.
INTERPRETATION STRATEGY
This strategy combines the geophysical information from the well logs and the geologic
information from the characterization well to answer the hydrologic questions of interest. Fig-
ure 4.1 shows a block diagram of the strategy and the reader is urged to refer to it frequently.
The strategy is only developed for use in the saturated zone. Additional strategies for use in
the unsaturated zone will be developed as part of a future cooperative agreement
Poros ity
The gamma nsity log which measures the bulk (lensity of the formation can be used to
determine porosity if matrix density E . known. The relationship is derived from the fact that
the bulk density is a weighted average of the density of the formation’s constitue its:
pg “ pj +(I—n)p,,, (4.1)
and is
PBP, .
4.2
P1 — P ,ji
24
-------
MATRIX
DENSITY
F i
___________ LITHOLOGY
CROSS
__________ SITE SPECiFIC CATION
“1 NATURAL GAMMA EX I1ANGE
VS CEC RELATIONSHiP CAPACITY
___________ ____
rCONDUCT,V,TY MODEL’ PORE FLUID1
___________ (v1AxMAN/sf iTS) J CONDUCTIVITY
________ —
ACCEPTABLE
[ PERFORATED ZONE
‘ IHORIZON T Atl’I —
FLOW METER
PT7 F I I I I I
Dosired Information
— — — — Empirical Rolatlon
From Litoroturo
GROUNDWATER
VELOCITY
( FLOW METER )
- Site Specific Data
— —. — Hydrologic Judgament
1
I
REGIONAL
GRADIENT
‘ ‘‘‘ ‘ F IdData
-JGROUNDWATER1
I VELOCITY I
INDICATION OFI
ANISOTROPyJ
SiTE SPECIFIC
HYDRAULIC CONDUCTIVITY
VS POROSITY RELAT1ONSH 1P
‘ ‘ Ii , , ! , ,
I NATURAL
1•, i , it ,
_____
1 I L _i. ’ F F
1 TELEVIEWER
FIGURE 4.1. InterpretatIon Strategy.
-------
where
n = total porosity
PB = bulk density (from density log)
= matrix density 2 65 gm/ m 3
= fluid density 1.0 gm/cm 3
The choice of a matrix density, which is the (lensity of the formation material excluding the
pore space and fluid, is determined by noting the mineralogy in the core sainpks. Once the
mineralogy of a particular formation is known its matrix density can be calculated from a
weighted average of the density of the matrix minerals
Effective porosity differs from total porosity in that it is the percentage of total pore
space available for fluid transport. Fluid in pore spaces that are dead ended (vuggs) or
extremely small (clays) is not mobile and hence this pore space is not included in effective
porosity. In coarse sediments with inter-granular porosity, effective porosity will be the same
as total porosit.y but as the percentage of dead end pores increases or pore size decreases,
effective porosity will become less than total porosity
The core samples from the characterization wells can be evaluated to determine if there
is a significant difference between effective and total porosity. This can easily be done with a
tracer test when the cores hydraulic conductivity is determined, II there is a significant
difference, it is necessary to attempt to develop a site and formation specific relation between
the two If this can be done then an effective porosity log can be obtained by applying the
relationship to the total porosity log
Hydraulic conductivity and regional ground-water velocity
There are two possible ways to obtain hydraulic conductivity and ground-water velocity.
Both metlio(ls have advantages and (lisadvantages and when ii ,ed together, they can yield
information on the hydraulic anisotropy of the formation
The most direct way of obtaining hydraulic conductivity and velocity is with a borehole
thermal ground-water flow meter. The use of this device for measuring fluid velocity and
hydraulic conductivity, along with its limitations, is discussed in detail in Section 5.
An alternate method of determining hydraulic conductivity is based on the more conven-
tional logging tools. The first step is to determine the hydraulic co:i(liictivity of many short
sections in several of the boreholes This can be done by analysis of core samples or hydraulic
tests of short borehole intervals The intervals investigated must be representative of the
range of conductivities in the formation To investigate the possibility of a relationship
between hydraulic conductivity and porosity the conductivity of each interval is plotted
against the log derived porosity that corresponds to that interval. Riclo (1962) has demon-
strated such relationships for shallow sediments with intergranular porosity. It is important to
keep in mind the relationship, if it exists, will only be valid for the formation and locality
from which the data was obtained. The relationship tends to occur because with in a given
depositional system any change in sorting, the controlling factor for porosity, will also be
26
-------
accompanied by a change in grain size distribution, the controlling factor for hydraulic con-
ductivity. Various statistical approaches can be used to quantify the relationship and used to
predict the hydraulic conductivity that corresponds to a given porosity. This relationship is
then applied to the porosity logs and a corresponding hydraulic conductivity log can be
obtained.
Once the hydraulic conductivity logs have been obtained, this information akng with
the effective porosity log and the regional gradient can be used to find the pore velocity from
Darcy’s Law.
It is important to understand the limitations and assumptions used in these two
methods for obtaining ground-water velocity information. The flow meter technique requires a
well with acceptable perforations. A televiewer is commonly ised to determine if this condi-
tion exists. There are also restrictions on the minimum velocity that can be measured The
porosity relationship approach is dependent on the development of a porosity versus hydraulic
conductivity relationship which may not be as unique as desired.
Hydraulic anisotropy
In addition to direct measurement of hydraulic conductivity, the thermal how meter can
be modified to directly measure the velocity direction. The result is that the flow meter can
be used in conjunction with the regional and local piezometric head gradient to obtain the
true hydraulic conductiviy for an anisotropic aquifer. For an anisotropic aqd fer, the
hydraulic conductivity is a second rank tensor, and the direction of flow is (in general)
different from the direction of t ie gradient.
The direction of flow (8,) is a function of the gradient direction (6,) and the principle
values of the hydraulic conductivity tensor, K , and K,,, where x and y are the coordinate
axes of the principle directions. Figure 4 2 shows the relationships and definitions of the vari-
ables. For more information on anisotropy, see Bear (1979, chap. 4). The relationship between
8, and 8 is determined by using trigonometry and Darcy’s Law:
8, = tant(.! _) (1.3)
where
q, K,, J, ; q, K,, .1, (Darcy’s Law) (4 4a,b)
and
7 = i, + ,‘ .r,
Substituting eq (4) into (3) and using J, = Jcos6, and J, = Jsin6,, we have:
e, = tan (K. tan8, ) , (4.5)
where K, K,, 1K,,.
The importance of determining anisotropy cannot be overemphasized. As an example,
assume that the direction of the gradient is 30 degrees from the x-axis and K, =2, then using
eq. (4), the velocity direction would be nearl; 50 degrees In other words, changing the con-
ductivity ratio by a factor of two can change the velocity d:rection by as much as 20 degrees
27
-------
Y Ks.)
x
q -K,
ax
q -K 5cD
óy
eq tar (K rdn e )
Where ( K
FIGURE 4.2 The Relationship Between the Velocity and the Gradient
for an Anfootropic Aquifer.
(is g.w. gradient)
q g.w. velocity
28
-------
Therefore, the degree of anisotropy has significant implications for the placement of down-
gradient monitoring wells and the prediction of contaminant plume migration
Lithology
Borehole methods are of great help in determining lithology If a plot is made with
porosity on one axis and the corresponding natural gamma response on the other axis it may
be found that formations with similar litliology will form groups on the plot. Figure 4 3 is an
example of this concept The core and geophysical data from the characterization holes can be
used to make such a crossplot.
This is done by plotting the lithology of numerous samples from the characterization
wells on the crossplot If enough samples representative of the different lith&ogies are avail-
able, it will be possible to define the lahologic boundaries on the crossplot Once the crossplot
is defined, the unknown lithologies in wells from which core data is unavailable can be deter-
mined by plotting the natural gamma and porosity response on the crossplot and noting in
which hthiology field it plots.
Cation exchange capacity
The cation exchange capacity (CEO) of the formation is important because it
significantly affects the mobility of some contaminates and affects the geochemistry and
electrical properties of the formation and pore water. CEO is primarily a function of clay
content and type. It is closely related to the natural gamma response because formations with
high CEC preferentially absorb high valence cations includiflg U” and Th 4 . Naturally occur-
ring radioactive isotopes of these cations frequently account for the majority of the radiation
associated with clays. In some environments the Potassium in the clay matrix will be the
dominant source of radiation (Donovan and llilchie, 1079, Johnson and Linke, 1977) The
cores from the calibration well can be anelyzed for CEC and a statistical correlation can be
made to the natural gamma log. From this relationship, the CEC corresponding to a given
natural gamma response can be determined. This relationship, if it exists, can then be
applied to all the holes and CEO logs can be obtained.
Pore fluid electrical conductivity
Electrical conductivity of a fluid is closely related to the ionic strength of the solution
and in many, but not all, instances can be an indicator of contamination Unfortunately, it is
not possible to directly measure the electrical conductivity of the pore fluid The fluid in the
well is a mixture of fluids and has been exposed to the atmosphere and hence is not neces-
sarily representative of the actual pore fluid. It is desirable to have a method of measuring the
in 811u pore fluid conductivity.
The induction log measures the electrical conductivity of the formation, but this is
influenced by many things. Current in the formation is conducted by two major pathways,
through the pore fluid and by ion movement along surface exchange sites. In general the
matrix dGes not contribute significantly to the electrical conductivity.
To determine the pore fluid conductivity, it is necessary to ( 1 uiantify these effccis The
path through the pore fluid is affected by pore fluid conductivity and porosity. The surface
path is controlled by the number of exchange sites that are available for ion movement and
hence is related to CEO.
29
-------
POROSITY
• Data from a
particular depth
FIGURE 4.3 Lithology rosa ‘ lot.
z
30
-------
In clay free formations, the number of exchange sites is negligible and the surface path
can be ignored. Under this situation, Arcluie’s Law (Keller and Prischkneclit, 1965), expressed
here in conücttvities can be used to express the re lationship between a, c 1 , and a ,
= 8001 (4.6)
where
= bulk conductivity (from induction tool)
= fluid conductivity (unknown)
a = porosity (from density tool)
o = tortucoity (0 62-1 0)
at = saturatinn exponent a42A)—2 Is)
The values for a and at have been empirically obtained and are dependent on litbology.
The porosity can be obtained from the density log as explained in Section LI, and the
induction log measures Oj Archie’s Law tan then be applied and an s ri situ pore fluid conduc-
tivity log obtained.
If clays are present, the surface conductance along the surface of the clays cannot he
neglected and it is necessary to use a model such as that proposed by Waxrnan and Smits
(1968):
e =C (B Q,+aj) (47)
where
B = 0.046 11—0.6 exp-{u 1 /0.013)1
— CEC(t—n)g pg
loon
CEC = cation exchange capacity, ineciJlOognu (from CEC log)
p 9 = bulk formation density (from density log)
In this equation, u is the unknown, 04 is obtained from the induction log, and n is obtained
as in Sect on 5 , 1. Density is measured with the density tog, and knowing the lithotogy, at can
be obtained from the literature (Keller and Frischknecht, 1966) CEC is obtained from the
naturai gamma log as explained in Section £5. The term D Q, is the additional conductivity
31
-------
from surface conductance. Using this equation to relate the log data, pore fluid electrical con-
ductivity can be determined
Ii is important to emphasize that this method allows the measurement of an in silu
profile of pore fluid conductivity that uses the CEC log to account for the influence of chys
It is not necessary to determine the clay type or amount because only the CCC which is
obtained from Lhe natural gamma log is necessary.
COMPUTER IMPLEMENTATION
Large data sets and nonlinear relationships make computer processing a necessity It. is
desirable to record he data digitally, but if only a limited number of sites are investigated
and turnaround time is not critical, the analog data could be recorded and digitized later
The processing software would only require a microcomputer and a plotter. The software
can he written to guide the interpreter through the strategy, and perform the important task
of keeping track of the propagation of uncertainties. It is not desirerable to reduce the
software to a standardized package as this would encourage users not to take the time to
fully understand the assumptions in the processing sequence This software package could be
readily used by small firms. The actual field work could be contracted out, thi s eliminating
the need of a high capital equipment cost.
FURThER REFINEMENTS
Effects of drilling disturbances
All drilling techniques will disturb the formation in the immediate vicinity of the well,
The effect this will have on the interpretation is dependent on the degree and depth of distur-
bance, and the effective penetration of the logging measurements If two gamma density logs
are made with two different source detector sp ’cings, thereby allowing measurement of forma-
tion density weighted over two different horizontal distances from the borehole, an indication
of drilling disturbances can be obtained. Under favorable conditions, the drilling effects can be
removed from the density measurement or at least intervals of significant disturbances can be
identified and treated with an alternate approach than is suggested here In theory the same
concept. could be used for the induction tool, but a small diameter multiple spacing induction
tool is not currently available.
Tool response
Logging instrumentation does not make a measurement at a point, but because of prac-
tical design restrictions, the value is a weighted average over a short section of (he borehole
This spatial average varies depending on the tool design. It is important that when logs are
compared from different tools that the measurements are weighted averages of the same sec-
tion of the well. This is particularly important near thin bec 1 s and abrupt contacts.
The effects of different spatial averaging can be removed by digital filtering of the data so
that the two measurements are averaged over the same iiiterval, or in some instances by
deconvolution, a digital process which removes the effects of spatial averaging.
32
-------
Improved electrical conductivity model
The calculased pore fluid electrical conductivity is strongly dependent on the way the
matrix properties are accounted for. Because clay-bearing formations vdl have a frequency
dependent electrical conductivity (Ward and l?raser, 1967) it is important to have a model
that. is accurate at the measurement frequency. Field tests of this strategy will b ne with a
Geonics EM-39 which operates at a frequency of 50 NIla Unfortunately, conductivity starts
becoming strongly frequency dependent. at around 50 kI I,. and above. ‘l’liereforc, the
Waxman/Smits model is not wholly acceptable much above this frequency, it would be useful
to develop a better model.
There are major instrumental difficulties in measuring conductivities of reasonable sized
lab samples at 50 l(Hz. The Waxmari/Smit.s model can be used but it is important to be
aware of its possible pitfalls in high clay content rorn ations Additional thought should he
given to developing a practical set of experiments from which an improved model could be
developed.
Effects of contaminants
This strategy assumes that the site specific relationships obtained from the calibration
well hold throughout the site. Although different relationships could be developed for different
formations it is assumed that these rek.tionships are valid throughout. the formation for which
they were developed In unusual cases, it is possible that the presence of the contaminant.
could alter these relationships and invalidate the interpretation. Because these relationships
are based on fairly simple physical and chemical principles, a review of the literature along
with an understanding of he mechanisms involved may make it possible to identify coridi-
tions where the contaminant might be altering the relationships used in the interpretation
CONCLUSIONS
A strategy has been dev loped that combines selected tradit onal borehole logging tech.
niques with hydrogeologic borehole equipment and methodology. When the strategy is
applied as outlined in this section, the parameters that are important to contaminant plume
detection, monitoring and transport prediction can be calculated for every borehole located in
and around the hazardous waste site Because logging gives information with reapect to depth,
the strategy provides these parameters as a function of depth in the aquifer. The parameters
that. can be obtained from this strategy include litliology, total porosity, effective porosity,
ground-water velocity, hydraulic conductivity (including anisotropy), cation exchange capacity
and pore fluid conductivity.
Considering the wide range of conditions present at hazardous wast sites it is not possi-
ble to de elop a universal technique that will york everywhere The interpretation strategy
presented here relies heavily on the use of core data to determine site and formation specific
relations between measured parameters and parameters of interest Although the relations
used here have been noted in some geologic environments this does not imply that they will
always he present. However when they are it desirerable to take advantage of them, when
they are not an alternate approach must be employed Although the strategy can be reduced
to a simple set of computer programs the processing and interpretation must. always by done
by an individual who understands the significance of both the hydrologic and geophysical
assumptions and relations that are made along the processing path and can decided if they
33
-------
are appropriate.
Macroscopic variations in hydraulic parameters are the most important contribution to
contaminant mixing and dispersion in aquifers. This strategy is designed to provide sufficient
detailed information so that the mixing and dispersion process can be modeled by strictly
advecti”e transport processes. This is an attractive alternative to using advective-dispersion
models which require complex field experiments to determine dispersion coefficients and still
often require trial and error calibration.
Under a proposed cooperative agreement with EPA, D I II will evaluate this strategy at a
number of field sites. Refinements will be made to it based on the field experience.
34
-------
SECTION 5
USE OF A BOREHOLE THERMAL FLOW METER FOR DETERMINATION OF
GROUND-WATER VELOCITY AND HYDRAULIC CONDUCTIVITY
INTRODUCTION
Subsurface data .e required to determine hydrogeologic conditions of hazardous waste
sites, to locate contaminant plumes and to monitor for leaks from disposal and storage facili-
ties. Ground-water velocity is the most important information to help determine direction of
contaminant plume migration. Ground-water velocity is a vector, possessing a speed (magni-
tude) and a direction. It is needed to locate the up and down gradient monitoring wells and
to estimate potential contaminant pathways and travel times.
The traditional way of determining ground-water velocity is to calculate it using Darcy’s
Law and regional or local piezometric bead gradient information. This is an indirect measure-
ment, and does iiot take into account velocity variations in the vertical dimension. A more
desirable method to obtain velocity information in principle would be to direc tly measure it in
a borehole. One way of doing this is with a thermal ground-water flow meter.
The ground-water flow meter evaluated in this study is the KV Associates Model 30L
Geo-Flo. The instrument operates by inducing a heat pulse and monitoring the pulse move-
ment away from the source. The following section will describe in detail the theory of instru-
ment operation.
It is necessary to understand the instrument before discussing the specific objectives of
this study, but the general goals are to determine the instrument limitations and accuracy;
and evaluate the manufacturer’s suggested calibration procedures. In addition, theoretical
work will be presented to show that there is an alternative to the empirical calibration and
that it is possible in theory to use the instrument to determine the hydraulic conductivity of
the aquifer, in addition to the ground-water velocity.
‘It would be more appropriate to call the instrument a ground-water velocity meter, but the term
flow meter is iu widespread use, we use the same terminology
35
-------
DESCRIPTION AND THEORY OF OPERATION
The probe itself consists of a central heat source surrounded by five pairs of thermistors
The basic principle of operation is that the central heat source generates a pulse of heat
energy. This pulse diffuses ra’liahly from the center of - prohe by heat con(IuctIoIl and is
advected by the ambient ground water The direction and relative magnitude of the advective
ground-water velocity can be determined by measuring the temperature difference between
opposite pairs of thermistors (see Figure 5.1) Various (actors affect the performance of the
instrument, including thermal conductance of the solid and liquid ihaacs, surface area of the
solid phase, thermal transfer coefficient of the liquid phase and the advective transport rate of
the fluid phase (Kerfoot, 19 2)
The flow meter is 4.4 cm in diameter and can be used with a simnie end cap packed itli
glass beads in a 5-cm well The glass beads are packed around the thermisto&s and heat
source in order to eliminate heat convection and to insure a continuous porous medium from
the aquifer into the borehole for more accurate velocity measurements. A diagram of the 5
cm end cap and flow meter probe is shown in Figure 5.2.
This particular model of the flow meter is designed to be used primarily with 5-cm well
casing, but the manufacturer provides two different packer configurations to allow the flow
meter to be used in 10-cm well casings. These two packers are shown in Figure 5 3. The first
is a pneumatic packer, consisting of inflatable tubes above and below the thermistor array. A
nylon mesh sock is installed between the tubes which contains the glass beads and thermistor
array. This sock expands outward to grip the inner sides of the well casing when the tubes are
inflated, in theory providing a continuous porous medium in the borehole, similar to the 5 cm
end cap arrangement.
The second packer shown in Figure 5.3 is somewhat simpler in design than the pneu-
matic packer It is referred to as the “fuzzy packer”, and consists of a simple cylinder with an
outside diameter equal to the inside diameter (or slightly less) of a 10-cm well casing. The
fuzzy packer is filled with glass beads and the probe is screwed into it by means of an
adapter. The fuzzy packer must fit into the well casing very lightly in order to achieve the
continuous porous medium arrangement of the 5 cm end cap and the pneumatic packer.
In practice, the probe is lowered down the borehole to the level at which the ground-
water velocity is to be measured opposite a screened or slotted section. The submerged probe
creates a short duration point source of heat. After a period of time, the relative thermal
differences between each of the five pairs of thermistors are displayed using a rotary switch
which selects the pairs to be read. The information is then used to calculate the ground-water
speed and direction.
Prior to initiation of the heat pulse, the thermistors do not necessarily show zero tem-
perature difference between opposite pairs of thermistors even under isothermal conditions.
This condition is referred to as the natural background response The operator records the
temperature differences between the five pairs of thermistors prior to inducing the heat pulse
to correct for this.
The probe is provided with consistent orientation towards north by means of a set of
rods that snap together in only one direction.
36
-------
NO FLOW CONDITION
FLOW CONDITION
DIRECTION
FIGURE 5.]. Operating Principle of the Borehole Thermal Flow Meter
(from KV—Associates).
-J
TMAX
THERMISTORS
HEAT SOURCE
-------
FIGURE 5.2
IN SAND
Diagram of the 5—cm End Cap and Probe.
PROBE
END CAP
38
-------
PNEUMATIC PACKER
THERMAL
FLOW
METER
INTERiOR FINE MESH
BAG FILLED WITH
GLASS BEADS
4” WELL SCREEN
FUZZY PACKER
WIRE MESH SCREEN
FUZZY MATERIAL
(SIMILAR TO THAT
ON A PAINT ROLLER)
TUBE TO AIR PUMP
INFLATION UNIT
FIGURE 5.3
Diagram of the Packers for the 10—cm Borehole.
-------
The probe is on the downhole end, and a compass is attached to the uphole snap pole.
Corrertions for magnetic bearing are made with a l3runton compass, and the standard prac-
tice is to orient the number 1 thermistor towards true north.
Because the probe measures the relative temperature difference between thermistor pairs
at a set time after the heat pulse is emitted, the measurement of ground-water speed is a rela-
tive number which must be calibrated against some set of standards. The manufacturer pro-
vides a small calibration chamber in the shape of a 15-cm v-tube. The calibration process
involves filling the horizontal portion of the v-tube (see Figure 5.4) with porous material,
embedding the flow meter in the porous material, and inducing a series of known flow rates
through the t-tube. The relative measurements of the t-tube are plotted against the
corresponding known flow rates, thus providing a purely •mpirical calibration curve.
POTENTIAL PROBLEMS
There are a number of significant problems associated with the calibration technique and
instrument usage. If the porosity and permeability of the porous material are the same as that
of the formation to be measured in the field, then the calibration is exact. The manufacturer
offers no way of correcting the velocities for different aquifers other than conducting v-tube
calibration experiments for each anuifer material In addition, the aquifer material placed in
the t-tube has been disturbed and probably has hydraulic characteristics different from the
in-situ values.
Several potential problems and questions arose concerning the applicability of the ther-
mal flow meter after initial field testing of the equipment (Uess, et at, 1984). They included
the following:
• Effect of the two different packers on the ground-water velocity measurements.
o Low flow velocity measurement limit.
o Suitability of the t-tube calibration chamber to construct calibration curves
In order to solve these problems and answer the questions, a number of specific goals and
objectives were developed.
GOALS AND OBJEOTWES
There are four primary goals of this study:
1. Evaluate calibration proceduies for the flow meter.
2. Determine limits of accuracy and precision under controlled laboratory conditions
3. Develop methods and techniques to extend and improve accuracy of the flow meter.
4 Develop methods to yield hydraulic conductivity information from the flow meter
In order to achieve these goals, five specific ob3ectlves were developed:
1. Construct a sand box flow chamber for calibration experiments to allow comparison with
the I -tube.
2: Conduct experiments with the t-tube calibration chamber and the sand box to determine
I-tube calibration accuracy.
40
-------
6” FLOW TUBE FILLED
WITH MEDIUM SAND
FlOUR! 5.4
Diagram of the T—Tube Calibration Chamber
(from Ky—Associates).
4
FLOW
PROBE
OUTFLOW
SLOTTED PVC SCREEN
INFLOW
FILTER
METERED FLOW
PUMP
41
-------
3. Test the 4.4-cm end cap and the 10-cm packers in the t-tube and the sand box for accu-
racy and precision.
4. Develop solutions for flow around and through a cylinder (packer) of (lifFerent hydraulic
conductivity than the surrounding aquifer.
5. Derive equations to allow determination of hydraulic conductivity using the flow rieter
from he solutions developed in (4).
DESIGN AND CONSTRUCTfON OF THE SAND BOX FLOW CIL\MBER
The sand box was designed to allow precision control and repeatability of the ground-
water velocity and to have uniform flow characteristics throughout. the vertical cross-section.
Numerous 5- and 10-cm well casings had to be installed, thus the dimensions of the sand box
are rather large: 1 2 m by 1.2 m (cross-sectional area in direction of flow) by 2 m (length of
flow path). The sand box is constructed of 3/8 inch plexiglas, arid N held together with angle
iron and braced with 2x4 and 2x6 beams Figure 5 5 is set of engineering drawings, providing
the construction details of the sand box The end chambers provided constant head reservoirs
which could be adjusted in height with a precision of 001 cm, providing excellent reproduci-
bility of flow rates through the sand box. Figure 5 6 is a diagram of the neil casing arrange-
ment within the sand box. Each well casing was placed at least three casing diameters away
from its nearest neighbor. Some of the well casings were deliberately placed flear a wall to
allow study of edge/wall effects.
Uniform size silica sand oF #16 sieve size was used in the sand box (geometric mean size
1.19 mm) This sand size was chosen primarily to prevent sand grains from entering the cas-
ing slots When the sand was emplaced in the box, care was taken to avoid stratification
Using Darcy’s Law, the hydraulic conductivity of the sand was calculated to be 252 rn/day
Determining the porosity is of primary importance. The volumetric flow rate (Q), cross-
sectional area (A), effective porosity (n) and pore velocity (v) are related by the following sim-
ple equation
An
In this equation, Q and A can be measured with great accuracy It is important to have
a value for the effective porosity that is equally accurate, so that the velocity can be calcu-
lated accurately from the flow rate A number of different methods were tried to calculate
porosity, with calculated alues ranging from approximately 35-15% Rather than take an
average, It was decided that the most accurate and appropriate method was a conductive
tracer test in which breaKthrough curves were constructed The effective porosity ‘as nen
calculated based on time-distance relationships for peak arri,al from one well to another The
value calculated from this procedure was 43 7%. Table 5.1 is a summary of the relevant phy-
sical and hydraulic parameters of the sand box
SUMMARY OF EXPERIMENTS
The experiments are divided into four categories which are summarized in th,s section
Table 5 2 summarizes the experiments that are discussed in this section, and their relevant
parameters.
42
-------
FRONT VIOW
FIGURE 5.5 Construction Details of the Sand Box.
poeTs
PLANX
STUD
-------
.——-—- 1 ’
,fF 2
—
,
, I
, , .
t /
I
1c1 1’TT -
‘ r
, -
,
_____._I———— —
I I
I:
3 3I4 ’
FLOW
/
—
/
13/I O u
/
,
/
FICUR.E 5.6 PLacement of the Well Casing5 in the Sand Box.
-------
Table 5.1
[ ‘hysical and Hydraulic 1 arameters of the Sand Box
PARAMETER
VALUE
average grain size
1.19 mm
hydraulic cor.ductivit.y
252 rn/day
effective porosity
43 7’
The initial evaluation experiments included work with the instrument in open water, and
with the instrument inserted in sand without the end cap (naked probe in sand). Experi-
ments were also conducted to determine basic cha acter. iics of the instrument, including
bias and other parameters. This work helped gain familiarity with the instrument, but Lhe
resul .s were relatively insignificant and are not reported here.
A 5-cm well casing was placed inside the t-tube With the flow meter inside the casing,
experiments were conducted for a range of velocity measurements. This set of experiments
was repeated in the send box, providing a direct comparison of the calibration in the t-tube
and in the sand box for the 5-cm casing and cnd cap.
The third set of experiments was conducted similar to the second set, except that the
pneumatic packer was used with 10-cm well casings in both the L-tube and the eand box.
Similarly, the fourth set tested the fuzzy packer calibration in the t-tube and the sand
box. The fuzzy packer experiments were performed using the same 10-cm well casings that
were used for the pneumatic packer.
EXPERIMENTAL RESULTS
The type of packer chosen to be used with the instrument had by far the strongest effect
on instrument accuracy and operational limitations. Therefore, discussion of experimental
results will be grouped into three categories on the basis of the packer tested. The first
category will include all experiments with the 5-em end cap (i.e. all experiments conducted
with the 5-cm well casing) The second category will include all experiments with the pneu-
matic packer and the third category will be all fuzzy packer experiments
Experiments using the 5-cm end cap in 5-cm well caning
This set of experiments follows the caiil,ration procedure for the 5-cm end cap as recom-
mended by the manufacturers The data from experiments 20l-A3.l-.l are plotted in Figure
5.7. The y-axis is the velocity in the t.-tube. The x-axis is the vector sum of the temperature
‘Actually, the flow rate is the measured parameter, and the veloznty is calculated by knowing the
porosity
45
-------
TaI,Ie 5.2
Summary of Experiments
Experiment Place Casing Pucker Velocity
Number Size Type
(cm) (rn/day)
201-A3.1 I
201-A3.2
20 1-A3.3
201-A3.4
201-133.1
201-133.2
201-B3.3
201-133.4
201-33.5
201-133 6
201-133 7
201-133.8
t-tube
t-tube
t.—tube
t-tube
sand box
sand box
sand box
sand box
sand box
sand box
sand box
sand box
5
5
5
5
5
5
5
5
5
5
5
5
end cap
end cap
end cap
end cap
end cap
end cap
end cap
end cap
end cap
end cap
end cap
end cap
5.08
3.03
2 07
0 87
5 26
3.74
2.23
1.02
0.55
0 35
0 195
0080
401-A2.1
.1Q1-A2 2
401-A2 3
40l-A2.4
401-132.1
401-132.2
401-132.3
401-112 4
401-B2.5
401-B2 6
401-112.7
401-B2 8
t-tube
t.-tube
t.-tube
t-tube
sand box
sand box
sand box
sand box
sand box
sand box
sand box
sand box
10
10
10
10
10
10
10
10
10
10
10
10
pneumatic
pneumatic
pneumatic
pneumatic
pneumatic
pneumatic
pneumatic
pneumatic
pneumatic
pneumatic
pneumatic
pneumatic
5 66
3.83
1 00
080
5 26
4 30
2.22
1 03
0 55
0 35
020
0.10
40 1-A3 1
40 1-A3.2
401-A3 3
401-A3 4
401-113.1
401-113.2
401-113.3
401-113 4
401-133 5
401-133 6
401-B3.7
401-B3 8
1.-tube
t-tube
t-tube
t-tube
sand box
sand box
sand box
sand box
sand box
sand box
sand box
sand l)ox
10
10
10
10
10
10
10
10
10
10
10
10
fuzzy
fuzzy
fuzzy
fuzzy
fuzzy
fuzzy
fuzzy
fuzzy
fuzzy
fuzzy
fuzzy
fuzzy
5.48
3 99
2.10
1 05
5.21
3 74
2.21
1.0-I
0 55
0 35
0 22
0 10
46
-------
difference between the five opposing thermistor pairs. The p’otted temperature difference was
calculated according to instructions provided by the manufacturer The arrows placed above
the data points are the indicated directions for each experiment North (or “up”) is arbitrarily
chosen ns the true direction of flow. Hence, an arrow pointing north represents a measurement
with essentially no error in directional indication.
The data show an excellent linear relationship between the temperature dilferenee and
the tr te velocity of the fluid. The flow meter can read up to approximately 500 relative tern.
perature unit.s, so the response covers about 20% of the iiistriinieut. range
Figure 5.8 shows the resuht.s from the series of experiments using the 5-cm end cap in the
sand box. A good calibration is shown for the data in the velocity range from about 0 5
rn/day to the maximum velocity tested (about 5 rn/day). however 1 the directional error gets
large for velocities less than I rn/day. The two diverging arrows near the bottom of the
graph indicate the range of directions measured by repeating the experiment at that particu-
lar velocity.
The velocity magnitude error gets large below 05 rn/day. This error at low velocities is
attributed to the fact that the water is moving very slowly through the end cap and the
diffusion of the heat pulse dominates the advective transport. Note that the average direc-
tional measurement is still towards north, indicating that repet1tiv. measurements in the field
may provide a direction of acceptable accuracy. The resulting temperature difference is too
small for the thermistor pairs to measure. This will be discussed further in the following sec-
tions on the other two packers.
The sand box calibration curve for the 5-cm end cap (Figure 5.8) is somewhat non-linear,
compared to t - linear calibration in the t-tube (Figure 5 7) One reason suggested for the
non-linear calibration curve in the sand ox was related to the method used to con rol flow
rate in the sand box. The flow rate (and t.erefore velocity) was set by establishing constant
heads at different levels in the two end chamb r reservoirs Inaccuracies in setting the con-
stant heads could have caused a non-linear relationship between head differences and flow
rates. This hypothesis was eliminated by plotting Figurc 5.8 as flow rate (which is directly
measured) versus flow meter units for the experiment, series. The same non-linear cirve was
observed.
Another possible reason for the non-linearity is that the sand-box may not be operating
with truly one-dimensional flow. On the basis of some simple nye tra er tests, the observed
flow paths appear to be essentially hior, outal. Therefore, it is not known wl.y the non-
linearity was observed in this set of experiments at this time. Because the same porous
medium was used for the t-tube and the sand box, a much better agreement was expected br
the two calibration curves.
Experiments using the pneumatic packer
A 10-cm well was installed in the t-tube for the pneumatic packer and fuzzy packer
experiments. The data for the pneumatic packer experiments are shown in Figure 5 9. The
results are much different than those for the 5-cm end cap. The total instrument response was
about 5 flow meter units, or only about 2% of the instrument range. As seen in Figure 5 9, a
very small change in thermistor temperature difference indicates a large change in actual velo-
city. The arrows show that the directional accuracy is even worse than the magnitude
47
-------
>-
0
Q
0
LU
>
0
U i
0
10
8
6
4
2
00 40 80 80
FLOW METER UNITS
FIGURE 5.7 T—Tube Calibration of the 5—cm End Cap.
— 1 ’ - ‘ I
t
I
t
t
I I
. I
20
I I
. I
100
120
48
-------
10
0
0
4
2
0
0
>.
0
0
-j
LU
>
LU
0
FIGURE 5.8 Sand Box Calibration of the 5—cm End Cap.
FLOW METER UNITS
49
-------
10
8
6
4
2
0
0
FLOW METER UNITS
FICURE .9 T—Tube Calibration of the Pnewratie Packer.
>-
a
>-
I-
C-,
0
-J
Lu
>
Lu
I-
-j
C)
20 40 60 80 100
120
51)
-------
accuracy, and is useless at velocities less than 3 rn/clay.
Figure 5.10 shows the analogous experiments in the saud box (I e. same parameters as in
the (-tube). The instrument response is about 50 flow meter units, as opposed to only 5 in the
(-tube. A larger response indicates larger temperature diFferences and therefore more advective
transport across the thermistois Since the t-tube and sand box velocities are in the same
range (roughly 05 to 5 rn/day), it can be conduded that the velocity in the packer is higher
when it is in the sand box thaii when it is in the t—tu lie I iie explanation for this is that the
ratio of slotted area to total flow area is smaller in the t-tube tliaii in the sand box It. can be
concluded that the (-tube does not provide an acceptable calibration procedure for (lie pneli-
matic packer because of the edge and wall eflects of the t-tul)e
The directional accuracy in th’ sand box for the pneumatic packer is fairly good from I
rn/day through the highest velocity measured (about 6 rn/day) however, for (lie experiments
with velocities above 1 rn/day, there is an average bias of +7 degrees (with a standard devia-
tion of +/- 3 degrees) Several of the experiments were repeated several times without inov-
ing the probe, and the measured direction was the same to within 0 5 degree k is therefore
believed that this bias or offset is caused by the pneumatic packer rather thaic expeiimcntal
error.
The fan-shaped symbol near the bottom of Fugr 5 10 shows the range of indicated
(:irection for the velocities less than I rn/clay Tue l.irge variation in velocity inagnittm(le auid
direction is attributed to the fact that in this velocity range, (lie amount of water moving
through the packer is very small. At low velocities, the advective transport is dominated by
the diffusive heat transport. As a result, (tie heat pulse is spreading primarily by diffusion,
and the pulse reaches all thermistors at almost exactly the same time When the vector addi-
Lion is performed, the vectors are primarily noise and the resiiltiiig magnitude and (lurection
are essentially meaningless.
Experiments using the fuzzy packer
The calibration curves for the fuzzy packer in the (-tube and the sand box are shown in
F’igiires 5.11 and 5.12 respectively. \Vitli the Iiiziy picker iii the I-tube, the Ilow iiictcr pr
duces a marginally useful calibration curve, although LIce directions become unreliable below
about 3 rn/day. In Figure 5.12, (lie data show that the flow meter cannot be calibrated with
the fuzzy packer iii (lie sc id box, there is just iio correlation cst.chlishu d l)etwcen (he how
meter units and the actual velocity
At first, it see ins 0(1 d LI cat (lie p ii en iii atic p.ick cr e,cl ii ira irs I ccli er in Lice sa 1111 I )OX I Ii .uii
in the 1.—tube, while the reverse is (rile for the fuzzy IEH ker. The reason is (lint the fuzzy
packer fits iiito the 10 cm well casing very loosely, leaving a 0.5 cm anniiliis betweeii the
packer and the inside of the casing wall The result is that water entering the casing preferen-
tially flows through the annulus rather than through (lie p.tckcr ‘I’bc wall ellects of the I-i nbc
cause a slightly l)etter calibration than is possible in the sand box The sand box approxi-
mates the con(litions in an actual aquifer more closely than the 1.-Lube, so the conclusion is
that the fuzzy packer is not recommended for use in actual ground-water investigations.
Instrument accuracy and repeatability
Using the 5-cm end cap and flow meter in the sand box, several experiments crc
repeated to get information on the accuracy and repedi Ll)Ility of the flow meter ‘rile
51
-------
10
8
6
4
2
0
FLOW METER UNITS
FIGURE 5.10 Sand Box Calibration of the Pneumatic Packer.
>-
>-
I-
0
0
-J
w
>
w
I-
1
C.)
0 20 40 60 80
100
120
rn
-------
>-
I-
0
-J
L i i
>
LU
I-
C)
10
8
6
4
2
0
FLOW METER UNITS
FIGURE 5.11 T—Tube Calibration of the Fuzzy Packer.
0 20 40 60 80 100 120
53
-------
10
S
6
4
2
0
I C
0
‘a
0
0
-a
w
>
0
U i
‘a-
IC
a - i
C
0
f iGURE 5.12 Sand Box Calibration of the Fuzzy Packer.
FLOW METER UNITS
0 20 40 60 60 100 120
54
-------
following results arc Valid for 5-un end cap packer only, and for velocities greater than 05
rn/day. Several measurements were taken successively without moving time probe or cliamigung
flow conditions The standard deviations were all less Luau 2% of the averages if the probe
was removed and replaced to its original cricntatmoui and vertical position (to the be L of the
operator’s ability to do this manually), and under time same flow coiidut,ons, time standard
(levi .ltuons were still less thai. 2% TIme averages of tue temnpermiture diFferences (i e velocity
magnitudes) were about the same also however, time averages of the mmm(hicate(l velocity direc-
tions were usually different by about 5 degrees Time differences varied, and eeined to iimdic .ttc
that it is not possible to reliably place the instrument in the borehole with bet tar luau 5
degrees of directional accuracy. For real life problems, this kind of a ciiracy is more than ade-
quate considering the normal uncertainty and spatial variability in aquifer hydraulic parame-
ters
ThEORETICAL DEVEI.OPMENT
Introduction
Time calibration techniques provided by time manufacturer are purely empirical, relating
measured velocity to instrument readout Iii the previous section, we have shown thi. t. the
calibration technique suffers from numerous problems Eveii for tIme 5-cm end cap in a 5-cmii
well, the calibration curves for time sand box are different than for the t-tiihe, although exactly
the same type of sand was used for both sets of experiments Bringing disturbed aquifer
material in from the field and using it in tIme t—tuhc to develop a site-speculur calibration curve
won lii be very question able base(l on time COmumparison expem iiui ii ts (lone iii this study
On the basis of the poor caiuh)ratlon comparisons in the l.ihor .mtory experiments, it would
be very desirable to develop a way to directly muisure veloeity ni:mgiiilude ‘Flue tlueorelitaf
vork in this section addre es this goal In a(lditioim, it goe a major step further amid outlines
a technique to calculate formation hydraulic conductivity from the measumrcineiits taken by
the flow meter Thus technique represents a potentially very uuiportsnt appli(atioum of the Flow
meter that has not previously been considered
Direct calculation of aquifer fluid velocity
From a theoretical standpoint, the problem can he thought of as flow around nd
through a cylinder oh hydraulic conductivity dullercmmt from the sumrrountliimg aquifer The flow
meter with the 5-em end cap in a 5-cm well can he thought of as I lie pcriimcable C) lmimdcr. It us
assumed that flow through the end cap packed with glass heads obeys i) trcy’s Law an(l us
steady-state. Another assumption is that the hydraulic io scs ii . the casing slots arc negligible
Thus us a good assim mptlou under ii atii ral flow con(hmtuon , wh mcli are norumi ally well heloit’ Lime
transition from laminar to tuirhaiteumt ilow.
The sohui Lion is oi)t.amned thu rough application of coimu plex pot euutuuh theory Bouuumd .ury
conditions are applied at the cylinder (packer) boundary The boundary COfl(hitionS state that
flow must, be the same in the aquifer and time packer as the bouiuidary is appronclicul ‘l’hie
potent ml must alSO be he same as the boundary :s alproa lied ‘l’he complete uleruvatmon us
quite lengthy and can be found in \Vheatcraft and ViimLerherg (1985) The solution us in the
form of two complex potentials, one for (lie aquifer ( a’. ), mmmd omic br Flue packer ( me
2
= U z + 0 (5 I)
-------
and
w,=UK ” z, (52)
where
A 1
I +K,
K
K” =2
1- 1-K,
K
K,=- -,
It’
z = + I y , (complex variable).
= specific discharge = V , a,
V , = aquifer pore velocity,
a, = aquifer porosity,
K, = hydraulic conductivity of the aquifer,
K ,, = hydraulic conductivity of the packer.
56
-------
These complex potential equations can be separated into real and imaginary parts. The
real part is related to the potential and the imaginaly part is related to the stream function
Figure 5.13 shows the streamlines in the vicinity of the packed borehole for several values of
K,. For a packed borehole that is more pei meablc than the aquifer, fluid preferentially moves
into the packer, causing greater velocity in the packer than in the aquifer. Vhen the packed
borehole is less permeable thaii the aquifer, fluid tends to move around the borehole, causing
the packer velocity to be less than the fluid velocity in the borehole.
To obtain the pore velocity in the packed borehole, the stream function equation is writ.-
ten on the borehole boundary. Equations (5.1) and (5 2) are both valid on the borehole boun-
dary. Equation (5 2) will be used because it is algebraically simpler. The amount . of flow
through the borehole packer can be determined by finding the particular streamline which is
tangent to the borehol.-. The value of the tangent streamline, ‘1’s represents the flow that
moves through half aI the borehole (the half that. is on the positive side of the x.axis), as seen
in Figure 5.14. The efore, the total flow through the cylinder, Q , is twice the value of ‘I’s.
On the cylinder boundary, the stream function can be written by obtaining the ima-
ginary part of equation (5.2):
(5.3)
The tangent streamline, ‘ 4 ’s , occurs at the point x =0, p =r,, and its value is (see Figure 5.13):
—K’ q r . (5.4)
The total flow through the borehole packer, Q,, is:
= 2’I’ = 2 r, K’ q . (5.5)
But Q can also be expressed as velocity in the borehole times borehole diameter:
Q, = V (2r,)n, . (5.6)
By equating the two expressions for Q, (eqs. 5.5 and 5 6).
V n ,
K’
Substituting in the expression for K’
v, n (1+K )
2K (57)
Equation 5.7 is an extremely important development. If all of the variables on the right
hand side are known, one can directly calculate the average specific discharge for the aquifer.
In theory, equation 5.7 eliminates the need for empirical calibration of the flow meter with the
t-tube. Based on the questionable results of the calibration experiments, this would be a very
desirable thing to be able to do.
Equation 5.7 requires the instrument to actually measure the borehole packer velocity,
(vp) However, the thermal flow meter reads out in “flow meter units,” which are actually
temperature differences between thermistor pairs The flow meter needs to be modified to
allow calculation of the packer velocity directly. Such a modification would require that the
57
-------
-
H
2• .
K iWi
a
3o
FICURE 5.fl Streamlines Moving Around and Through a l3orehole Packer
for Different Hydraulic Conductivity Ratios.
IC,..t,tO
I
2
—2 —I 1 2
I c r — I l 5
-2
1
iO
—
2
I
ii
—3—2 —1 0 1
2
I
1 —2 —1 0 1
I —2 —l
0 1
a
—3 —2 —1
I
g 1 2 1 3 2 —1
0
I
58
-------
k12 k21
FIGURE 5.14
Tangent Streamline for Determination of Aquifer
Fluid Velocity and Hydraulic Conductivity.
3 2
I
0
UNIFORM FLOW FIELD
kl=1 k2=5
—1 —2 —3
3
2
I
3 2 1 0
—I —2
0
59
-------
output from each thermistor be recorded on a chart recorder as a function of time (starting
when the temperature pulse is initiated). Using the peak arrival times of the heat pulse for
each thermistor, the advective portion of the heat transport could be obtained. This in turn
would allow direct calcjlation of the fluid velocity in the packer.
The other parameters on the right hand side of equation 5.7 are easier to obtain in some
cases. The packer porosity and hydraulic conductivity (n, and Kr) can be measured in the
laboratory, arid the aquifer hydraulic conductivity can be obtained undep ndently from aquifer
pumping tests, or other means.
It should be noted that the aquifer hydraulic conductivity (K,) measured by a pump
test may be different from the K, immediately adjacent to the vertical position where meas-
urement with the flow meter is being taken. In such a case, the K, would be incorrect. This
problem will be directly addressed in the following section.
Equation 5.7 has been written in terms of q , but if the aquifer velocity varies vertically
only as a function of porosity, (e g. due to variations in litholgogic stratification) then the
gamma density log will allow conversion of the specific discharge into aquifer fluid velocity in
situations where the relationship between effective and total porosity is known.
Determination of hydraulic conductivity using the thermal flow meter
If the aquifer hydraulic conductivity is unknown, equation 5.7 can be used to determine
it. The process consists of conducting two experiments with the flow meter, using two
different packers that have different, but known hydraulic conductivities, K, 1 and K, 2 . Thus
we h*ve two equations, and two unknowns. En the first experiment, we use a packer with
hydraulic conductivity K, 1 , obtaining a packer velocity V, 1 :
V 1 n, (i+K 1 /K 1 ) 58
— 2 K, i/K.) ( . )
In the second experiment, we change the packer so that we have K, 2 and V, 2 :
— V 2 n, 2 (1+K , 2 /K )
‘)1I/ ,f/ ( )
The porosities, and may be different, but they can be determined in the laboratory
along with the hydraulic conductivities, K, 1 and K, 2 .
The two unknowns are q and K . Thus by conducting two experiments, we can actu-
ally determine the aquifer hydraulic conductivity as well i ’s the specific discharge. This is a
very significant finding, because until this Lheoretical development, there has been no reason
to suspect that the thermal flow meter could be used to obtain the aquifer hydraulic conduc-
tivity.
60
-------
Equations (5.8) and (5.9) can easily be solved for the two unknown8 K 1 and q 1 :
K 1 =K,z1;r1 ‘1 (5.10)
,,-q ,- ,
au d
= .! .( KF—l ) , (5.11)
p,-q,i•
qp 2 Vp2 92
where qp, = — =
q,i ‘p1
and IC ,, =
This “two-IC 9 method” theoretically offers an alternative to laboratory calibration with
the t-tube, which has been shown to be of questionable value. This method would r. quire a
fairly involved field procedure, and may be cumbersome for multiple measurements, but con-
sidering the problems of direct calibration, it may be the only way to obtain any data from
the flow meter other than directional data.
Potential problems and limitations
It should be strongly emphasized that the method outlined in this section, while theoreti-
cally correct 1 has not been tested in the laboratory or in the field. Plans for future research
include using the sand box laboratory constructed for this project to evaluate this method for
obtaining aquifer fluid velocity and hydraulic conductivity using the thermal flow metP.r.
Some of the problems that may be encountered include:
1. The packer hydraulic conductivity may be subject to considerable variability from one
measurement to the next, due to rearrangement of the glass beads in the packer. Sensi-
tivity of K, to such rearrangement is an important area of further research.
2. The accuracy of the method may be sensitive to the choice of packer hydraulic conduc-
tivities, and/or sensitive to the aquifer hydraulic conductivity relative to the choice of
packer hydraulic conductivity.
These and other problems that could develop may limit the use of the method with the
thermal flow meter as it is currently designed. Understanding the problems and limitations
may lead to ideas for a better design of the thermal flow meter, taking into consideration the
design criteria that would make the method work in practice, as well as in theory.
61
-------
SUMMARY AND CONCLUSIONS
Experimental work: velocity magnitude calibration
The results of the calibration comparisons show that there are significant wall or
geometric effects when using the t-tube to proiide an empirical calibration curve for the
thermal flow meter when used with the 5-cm end cap in a 5-cm well casing The calibra-
tion curve was linear iii the t-tube and non-linear in the sand box. ‘liiiis calibrations
done in the t-tube do not apply to actual Field situat,or.s. The t-tube calibration is even
more questionable when considering the fact that the t-tiibe must be filled with dis-
turbed aquifer material, which may have different hydraulic properties than the original
aquifer.
The pneumatic packer had a reasonably good calibration in the sand ho; but not
in the t-tube. Again, the discrepancy appears to be due to the wall effects in the t-tube.
The difference in the calibration curves between the t-tube and the sand box indicate
that the pneumatic packer should not be used for velocity magnitude detetminations in
the field.
The fuzzy packer has a linear calibration in the t-tube, but very small changes in
temperature difference are indicative of large changes in velocity. This condition makes
the calibration curve too sensitive to provide good correlation between temperature
difference and velocity magnitude. In the sand box, there is almost no correlation
between temperature and velocity, indicating that most of the fluid is moving around
the fuzzy packer. The fuzzy packer is the least useful of the three packer configurations
tested.
Experi:nental results: directional accuracy
The directional accuracy is good for velocities down to about 0 5 rn/day for the 5-
cm end cap For repeated experiments in which the flow meter is not moved, accuracy is
at least ± 0 5 degrees. If the flow meter is moved, the repeatability is about ± 3 degrees.
The directional accuracy for the pneumatic packer was not quite as good in the
sand box as the 5-cm end cap. The pneumatic packer showed a consistent bias offset of 7
degrees to the east. The results show that the pneumatic packer can be used to deter-
mine direction for velocitites down to 0.5 rn/day, even though it is not capable of being
calibrated for velocity magnitude. Correction for the bias can be done by subtracting 7
degrees from the indicated direction.
The fuzzy packer is inaccurate for the direction as well as the velocity magnitude,
so it should not be used at all.
In summary, the flow meter as presently marketed is only reliable for directional
measurements using the 5-cm end cap and with additional caution, the pneumatic
packer.
Theor tlcal work
Equations have been developed that expre aquifer fluid velocity as a function of
the fluid velocity in the borehole packer, hydraulic conductivity and porosity of the
borehole packer and the hydraulic conductivity of the aquifer. With some modification
to the thermal flow meter, it is theoretically possible to measure these parameters.
62
-------
Therefore, the aquifer fluid velodty can be directly calculated, thus eliminating the need
for the questionable t-tuhe ealibra.tio: procedure Although these equations are theoreti-
cally correct, they are untested. Considerable experimental work will be nece& ary to
determine how well they work with the thermal flow meter.
The most significant result of the theoretical work is that the hydraulic conduc-
tivity of the aquifer can be calculated with the thermal flow meter. This method requires
two measurements to he taken with the flow meter, with two different packers of
different hydraulic conductivity. These two measurements provide two equations and
two unknowns, the aquifer hydraulic conductivity and the aquifer specific discharge. If
the aquifer porosity is known from other borehole logs, then the aquifer fluid velocity
can be calculated as a function of aquifer depth. This “two-K, method” theoretically
provides the only known way of simultaneously obtaining aquifer hydraulic conductivity
and specific discharge vector data. If it can be shown to be of practical use in the field,
it will be a valuable addition to the suite of tools available to the hydrogeologist.
63
-------
SECTION
SUMMARY AND CONCLUSIONS
INTRODUCTION
The complex nature of the ground-water contamination problem requires the collection
of extensive amounts of data in order to undrstand the problem well enough to recommend
and execute the appropriate remedial action, It is nearly impossible to collect adequate
amounts of data with traditional hydrogeologic methods, thus creating a need for technology
that can help answer the complex questions that arise when dealing with ground-water con-
tamination,
Geophysical methods have been widely used in oil and mineral exploration since the
1920’s. However, due to their cost and the relative simplicity of most previous ground-water
problems, geophysical methods have not commonly been used for ground-water investigations.
As the complexity and consequences of ground-water contamination increase, geophysics is
becoming a more cost effective approach to answer the hydrologic questions associated with
ground-water contamination
Geophysical methods applicable to hazardous waste site investigations can be broken
into two categories: surface and subsurface methods. Surface methods oiler the advanti gc of
relatively little capital investment at the site (no borehole is required), and rapid collection of
data over a horizontal area. However, the interpretation is often ambiguous and limited in
vertical resolution Subsurface methods require a borehole and can only investigate an area
inmediately around the borehole. However, they provide excellent information on vertical
changes in measured parameters. Also, a suite of complementary logs can potentially provide
unambiguous interpretation of hydrogeologic parameters, especially in the vertical dimension.
The two approaches complement each other very well. The subsurface methods provide
the necessary vertical detail for a small area and the surface methods can then extend this
detail horizon tally between borehotes
This report has examined only subsurface geophysical methods. Research in surface
methods is being performed by other EPA contractors l roblems of site characterization, con-
taminant plume detection and monitoring of contaminant plumes have been addressed using
borehole geophysics.
64
-------
BOREHOLE GEOPIIYSIOAL METHODS
Borehole methods fall into five major categories, acoustical, electromagnetic, nuclear,
flow and dimension, and thermal. Major applications of these techniques include: litliologic
correlation, lithology, rock density, fractures, porosity, permeability, flow, water level, water
quality, temperature gradient and hole diameter.
Hardware for borehole geophysical logging consists of similar basic components for all
the different tools. consisting of sensor, signal conditioners, and a recorder. The sensor or
sonde receives power and transmits the signal to the surface through a conducting cable,
which also serves to position the tool in the hole by means of a. winch Electronic controls at
the surface regulate logging speed and direction, power to the downhole electronics, signal
conditioning, and recorder responses. The return signal from the probe is a function of litholo-
gic, fluid, and borehole parameters and is recoided and later analyzed with a computer.
LIMiTATIONS OF BOREHOLE METHODS FOR HYDROGEOLOCIC HAZARDOUS WASTE
INVESTIGATIONS
Borehole logging methods have been developed primarily by and for the petroleum
industry. Logging tools are designed to be used in open, large diameter, deep holes. Several
logging tools are usually attached to one downhole sonde that can be as .nuch as 5 m in
length. Interpretation schemes are designed to remove effects of drilling and to determine
parameters that are of interest in petroleum reservoir engineering
The typical borehole that exists at (or around) hazardous waste sites is shallow (prob-
ably less than 100 m), narrow diameter (5 cm) and cased, usually with PVC, Teflon, or some
other plastic. None of the borehole tools designed for the petroleum industry are usable in
such an environment. Many downhole sondes used in the petroleum industry have 4 or S
tools attached and are 5 m long Even if a sonde such as this could fit in a typical 30-rn hole
at a hazardous waste site, the bottom 15% of the hole could not be logged due to the length
of the sonde None of the open hole logging tools (such as electric logging) can be used in the
PVC cased holes. Since most downhole tools are designed for high temperature, high pressure
environments, they would be over-designed for the typical shallow monitoring well around
hazardous waste sites. Moreover, the same tools may be subjected to hazardous chemical
borehole environments in monitoring wells around hazardous waste sites for which they are
not designed to withstand.
The interpretation schemes developed for the petroleum industry are designed to remove
effects of drilling fluid from the data. Logging is normally done before, or just alter hole com-
pletion and holes are almost never re-logged, especially after casing has been set. For hazar-
dous waste site investigations, borehole logging will commonly be (lone after PVC casing has
been set, and it. will be desirable to re-log holes on a regular basis to monitor for changes in
formation fluid chemistry and ground-water velocity.
The borehole logging parameters that are of interest to the hydrogeologist investigating
ground-water contamination are quite different from the parameters commonly sought l)y the
petroleum reservoir engineer. As a result of the above eoiisi’hcrations, it is of primary linpor-
Lance to develop a new borehole logging strategy that is designed to provide the iuiformation
sought by the hydrogeologist for hazardous waste site investigations.
65
-------
BOREHOLE LOGGING STRATEGY FOR UYDROCEOLOGISTS
The vertical variation in hydraulic parameters . ‘ithin an aquifer is recognized to be of
primary importance in determining the fate and transport of contaminants in ground-water
systems Traditionally, the process of !iydrodynamic dispersion has been thought . to be the
dominant process causing contaminant mixing Macro-scale heterogeneity and vertical
stratification induce large variations in the advective flow rate of the ground water. This pro-
cess has been termed macroscopic dispersion, and is the dominant mechanism controlling con-
taminant mixing and transport in many aquifers
Largely because of macroscopic dispersion, traditional ground-water flow equations are
inadequate to describe contaminant transport in aquifers Although it is important to account
for vertical variation in hydraulic parameters, there has been lit.tle effort to develop adequate
borehole methods that would provide such parameters.
If borehole methods are to be of use for hydrogeologists, it. is essential that they answer
questions of hydrologic significance. In particular, the strategy outlined in this report
describes how the following parameters vary with depth: porosity; hydraulic conductivity;
lithology; gro .ind-water velocity; cation exchange capacity of the formation; and electrical
conductivity of the pore fluid.
Hazardous waste sites are located in every roncei i5le geologic setting Each on’ is
unique and relationships developed for one site cannot be considered valid elsewhere It is
essential that relationships used in interpretations be based on data collected at the s e under
study. To do this, it is necessary to drill a characterization hole at each site. The interpreta-
tion strategy combines geophysical information from the well logs and geologic information
from the characterization well to answer the hydrologic questions we are interested in.
The assumption in this strategy i i that. the site specific relationships oi,tained from the
c. librat.ion well hold throughout the site Shallow, unconsolidated sedime.ts are among the
most heterogeneous deposits to be found It is not expected that the formations will be
laterally continous for the site specific relationships to hold, only that the correlations
between lithologic characteristics and geophysical parameters remain the same. Even this
assumption may be risky in some instances and the practicing liydrogeologist that employs
this strategy must be aware of the possible variations that may exist for specific problems.
Although different relationships could be developed for different formations, it is assumeo that
these relationships are valid throughout the formation for which they were developed In
unusual cases, it is possible that the presence of the contaminant could alter these relation-
ships are based on fairly simple physical and chemical principles, a review of the literature
along with an understanding of the mechanii ms involved may make it possible to identify
conditions where the contaminant might be altering the relationships used in the interpreta-
tion.
USE OF A BOREHOLE THERMAL FLOW METER FOR DETERMINATION OF OROLND-
WATER VELOCITY AND HYDRAULIC CONDUCTIVITY
The traditional way of determining ground-water velocity is to calculate it using Darcy’s
Law, regional or ocal piezometric head gradient information, and effective porosity This is an
indirect measurement, and does not take into account velocity variations in the vertical
dimension. A more desirable method to obtain velocity information in principle would be to
66
-------
directly measure it in a borehole. One way of doing this is with a thermal ground-water flow
meter, as described in Section 5.
Laboratory attempts to calibrate the instrument for veiocity, including a specially
designed sandbox, were not entirely successful. The data generated clearly indicated that the
thermal flow meter, when used with the fuzzy packer, is inaccurate for velocity as well as
direction. It is recommended that the fuzzy packer not he .ised at all.
Because of poor calibration comparisons in the laboratory experiments, study was ini-
tiated to develop a way to directly measure velocity magnitude. Equations were developed
that express aquifer fluid velocity as a function of the fluid velocity in the borehole packer,
hydraulic conductivity and porosity of the borehole packer and the hydraulic conductivity of
the aquifer. With some modiiication to the thermal flow meter, it is theoretically possible to
directly measure these parameters. Therefore, the aquifer fluid velocity can be directly calcu-
lated, thus eliminating the need for the questionable calibration procedure. Although these
equations are theoretically correct, they are untested. Considerable experimental work will be
necessary to determine how well they work with the thermal flow meter.
The most significant result of the theoretical work is that the hydraulic conductivity of
the aquifer can be calculated with the thermal flow meter. This method requires two measure-
ments to be taken with the flow meter, with two different packers of different hydraulic con-
ductivity. These two measurements provide two equations and two unknowns, the aquifer
hydraulic conductivity and the aquifer specific discharge. If the aquifer porosity is known
from other borehole logs, then the aquifer fluid velocity can be calculated as a function of
aquifer depth.
67
-------
REFERENCES
Asquith, C. and C. Gibson, Basic Well Log Analysis for Geologists, American Association of
Petroleum Geologists, lulsa, Ok., 1982. 216 pp.
Bear, J., Hydraulics of Groundwater, McGraw-Hill, New York, 1979. 567 pp.
Dobrin, M. B., Introduction to Geophysical Prospecting, McGraw-hill, New York, 1076 630 pp.
Donovan, W. S. and D. W. Hilchie, Seventh Formation Evaluation Symposium of the Canadian
Well Logging Society, Calgary, Canada, 1079.
Dresser Atlas,, Well Logging and Interpretation Techniques, The Course for Home Study,
Dresser Industries, Inc., Houston, Texas, 1982, 350 pp.
GaI’perin, E.l., Vertical Seismic Profiling, Society of Exploration Ceophysicist ’i. Houston, Texas,
1974, 27Opp..
Hallock, W. B., “Subterranean temperatures at Wheeling, W. Va., aad Pittsburgh, Pa.,” Colum-
bia School of Mines Quarterly, vol. 19, Pp. 148-154, 1897.
Hess, J. W., SW. Wheatcraft, D. D. Spencer, and W. M. Adams, “Evaluation of the applicabili-
ty of some existing borehole instruments to hazardous waste site characterizaticn and mon-
itoring,” in Proceedings of the NWWA Conference on Surface and Borehole Geophysical
Methods in Groundwater Investigations, pp. 762-772, National Water Well Association,
1 .S4.
Hilchie, D., Applied Open Hole Log Interpretation, Douglas llilchic, Inc., Golden, Co, 1082, 220
pp..
Johnson, W.L. and W.B Linke, “Some practical applications to improve formation evaluation of
sandstone in the MacKenzie delta,” in Transactions, Fifth Formation Evaluation Symposi-
um, pp. R.1-R.!9, Canadian Vell Logging Society, Calgary, Canada, 1977.
Keller, C. V. and F. C. Frischknecht, Electrical Methods in Geophysical Prospecting, Pergam-
mon Press, New York, 1066, 523 pp..
Kerfoot, W., Groundwater Flowmeter System, Operalion and Maintenance Manual; Model 30L,
K-V Associates, Inc., Falmouth, Massachusetts, 1982, 27 pp..
Keys, W. S. and L. M. MacCary, Techniques of Water-Resources Investigations of the United
States Geological Survey; Application of Borehole Geophysics to Water-Resource8 Investi-
gations, U.S Geological Survey, Washington, D.C, 1971 126 pp.
68
-------
Mlay, M., “Coordinating protection efforts,” i. ..PA Journal, vol. 10, no. 5, pp. 50-60, 1084.
Pirson, S. 3., Rr.ndlook of Well Log Analysis, Prentice.iTa , Inc., Englewood Cliffs, New Jersey,
1963, 32Gpp..
Riclo, 3., ‘Poiouity and pe mcahiIity studies of sonie unconsolidated sediments,” Compass, vol.
39, pp. 51-63, 1( 52.
Schiumberger, Ltd.,, Log interp7etalion, Vol. 1-Principles; Vol. 2-Applications, Schiumberger,
Ltd., New York, 1072, 116 pp. and 112 pp.
Schlumberger, C., “Electrical logs and correlations in drill holes,” Mining and Itfe1alh:r jy, vol.
10, pp. 515-518, 1929.
Schwartz, F. W., “Macroscopic dispersion in porous media,” lVater Resour. 1?es, vol. 13, no. 4,
pp. 743-752, 1977.
Smith, L. and F. W. Schwartz, “A stot’hastic analysis of macroscopic dispersion,” Water Reaour.
Re&, vol. 16, no. 2, pp. 303-313, 1980.
Ward, S. H. and D. C. Fraser, “Conduction of electricity in rocks,” in Mining Geophysics, vol.
11, Theory, pp. 197-223, Society of Exploration Geophysicists, Tulsa, Oklahoma, 1967, 300
pp..
Waxman, M. H. and L. J. M. Smit.s, “Electrical conductivities in oil-bearing shaly sands,” Aime
Trans., vol. 243, pp. 107.122, 1968.
Vheatcraft, S. \V. and F. Winterberg, “Steady-state flow passing through a cylinder of permea-
bilil .y different from the surrounding medium,” IVal r 1?esour. Res., vol. 21, pp. 1925-1929,
1985.
Wyllie, M.R.J., “Elastic wave properties in heterogeneous and porous media,” Geophysics. vol.
21, pp. 41-70, 1956.
69
------- |