Laboratory Simulation of Subsurface Airflow Beneath a

Building

by

Joseph William Corsello

Bachelor of Science

University of New Hampshire, 2011

Submitted to the Department of Civil and Environmental Engineering in Partial Fulfillment of the Requirements for the Degree of

Master of Engineering in Civil and Environmental Engineering at the

Massachusetts Institute of Technology MASSACHUSETTS iNSTITUTE

OF TECHNOLOGY

June 2014 JUN 1 3 2014

LLBRARIS

@2014 Joseph William Corsello. All Rights Reserved.

The author hereby grants to MIT permission to reproduce and distribute publicly paper and electronic copies of this thesis document in whole or in part in any medium now known or hereafter created.

Signature redacted

Signature of Autf or:

Department of Civil and Environmental Engineering

May 9, 2014

Signature redacted

Certified by:

David E. Langseth

Senior Lecturer of Civil and Environmental Engineering

,

t

Signature

hejif Supervisor

redacted

Accepted by:

Heidi Wflepf

Chair, Departmental Committee for Graduate Students

Laboratory Simulation of Subsurface Airflow Beneath a

Building

by

Joseph William Corsello

Submitted to the Department of Civil and Environmental Engineering on May 9, 2014 in Partial

Fulfillment of the Requirements for the Degree of

Master of Engineering in Civil and Environmental Engineering at the

Massachusetts Institute of Technology

ABSTRACT

Vapor intrusion is the vapor-phase migration of volatile organic compounds (VOCs) into buildings due to subsurface soil or groundwater contamination. Oxygen replenishment rates beneath a building are significant for quantifying potential contaminant degradation rates within the vadose zone. Additionally, the migration of VOC soil gas vapors into buildings is partly due to pressure differences between the building and the subsurface.

This study addresses these issues through two laboratory scale experiments. The Wind

Experiment quantifies oxygen replenishment rates as a function of above ground wind speed, while the Depressurization Experiment examines the flow rate of air into a model building as a function of decreased building pressure. For the Wind Experiment, tests were run for basement and slab-on-grade building configurations, as well as with and without a simulated sidewalk. Results show that increased above ground wind speed increases the oxygen replenishment rate and that the presence of a simulated sidewalk inhibits the oxygen replenishment rate. For the Depressurization Experiment, tests were again run for basement and slab-on-grade building configurations, as well as for two different foundation crack percentages. Results of the experiment indicate that increased building vacuum increases the flow rate of air into the building. In addition, basement configuration, increased foundation crack percentage, or some combination of the two results in increased airflow into the building. Additional research is needed for both experiments in order to obtain statistically significant results and resolve remaining uncertainties. Specific research needs include an improved wind source, additional monitoring locations, various sidewalk sizes and shapes, and different foundation crack configurations.

Thesis Supervisor: David E. Langseth

Title: Senior Lecturer of Civil and Environmental Engineering

ACKNOWLEDGEMENTS

I would like to thank Dr. David Langseth, Dr. Atul Salhotra, Dr. Eric Adams and the rest of the MEng staff, Steven Rudolph, Jessica Friscia and the rest of the Course 1 MEng Class of

2014, Dr. Nancy Kinner, Dr. James Malley, Dr. Robin Collins, Kara Connolly, and my parents, without whom none of this would have been possible.

Joseph William Corsello

Cambridge, Massachusetts

May 2014

5

6

TABLE OF CONTENTS

1 In tro d u ctio n ......................................................................................................................................................

1 .1 B a ck g ro u n d ............................................................................................................................................

1.2 Literatu re R eview ...............................................................................................................................

1.3 Stu dy A p p roach ...................................................................................................................................

2 M ethods .......... ........... ............... . ..... ..................................

2.1 M aterials ........................ ........... ............ .

.......

..........

...................

.

.................. 17

............................

17

1 1

1 1

13

15

2.1.1 Sand ........................................ ...... ...............................................

2.1.2 Sand Box and Building................. .....................

2.1.3 Differential Pressure Transducers ........... ...

17

19

....................

20

2.1.4 Data Acquisition System............... ................

2.1.5 Propeller Fan ............................. .............

........

20

..... 21

2.2

2.1.6 Peristaltic Pumps .............................................

Procedures.............. .... .

...................

2.2.1 Building Configuration Naming Scheme ....................

21

.. ......................................

22

....... 22

2.2.2 Barometric Pressure Readings ................................... 22

2.2.3 Differential Pressure Transducer Calibration. ................

2.2.4 Wind Experiment................ .............

........23

................ 23

2.2.4.1 Building Configurations ..........

2.2.4.2 Wind Speeds ............................................

.

... .................................

24

24

2.2.4.3 Data Collection.................................. .........

29

2.2.5 Depressurization Experiment.... .

.. ........

2.2.5.1 Building Configurations ........... ........ ........ ........... 32

2.2.5.2 Data Collection.................... ....... ................ 34

2.2.5.3 Building Depressurization .................................. 37

2.2.6 Apparatus Disassembly.... .

.

.

...

3 Results .... ..................

3.1 Wind Experiment .

.......................

.

.... .... .. ..... .........................................

42

.........-...........42

3.1.1 Data Reduction ................... .............................

42

3.1.2 Total Hydraulic Head .. ...... .

.

....... ................-..... 42

3.1.3 Isopotential Contours ........... ...... .........................

55

3.1.4 Oxygen Replenishment Rate . ......................... ............

55

3.2 Depressurization Experiment........................... ..............

57

7

3.2.1 D ata R ed u ction .............................................................................................................................

3.2.2 Total Hydraulic Head .................................................................................................................

3.2.3 Isopotential Contours ................................................................................................................

3.2.4 Airflow Into Building..................................................................................................................71

4 D iscu ssio n ..........................................................................................................................................................

4.1 W ind Experiment................................................................................................................................73

4.1.1 W in d Sp eed s ..................................................................................................................................

4.1.2 Building Configuration..............................................................................................................74

4.1.2.1 Basement versus Slab-On-Grade.............................................................................

4 .1 .2 .2 Sid ew alk ..................................................................................................................................

4.2 Depressurization Experiment ...................................................................................................

4.2.1 Building Depressurization...................................................................................................75

4.2.2 Building Configuration........................................................................................................

4.2.2.1 Basement versus Slab-On-Grade.............................................................................

4.2.2.2 Foundation Crack ................................................................................................................

5 Additional Research Considerations................................................................................................

6 C o n clu sio n s........................................................................................................................................................8

7 R e fe re n ce s..........................................................................................................................................................8

5 7

57

71

7 3

7 3

74

7 4

75

79

1

2

76

76

76

LIST OF FIGURES

Figure 1 Summary of Grain Size Analysis ............................................................................................. 18

Figure 2 - Boston W ind Rose ............................................................................................................................

Figure 3 Measured Wind Speed for BAS and BAS_SW at 10 ft/s...................

25

26

Figure 4 Measured Wind Speed for BAS and BAS_SW at 15 ft/s................... 26

Figure 5 Measured Wind Speed for BAS and BAS_SW at 20 ft/s................... 27

Figure 6 Measured Wind Speed for SOG and SOGSW at 10 ft/s ................... 27

Figure 7 Measured Wind Speed for SOG and SOGSW at 15 ft/s ...................

Figure 8 Measured Wind Speed for SOG and SOGSW at 20 ft/s ...................

28

28

Figure 9 W ind Experiment: Basement Side View............................................................................. 29

Figure 10 W ind Experiment: Basement Back View.......................................................................... 29

Figure 11 W ind Experiment: Slab-On-Grade Side View .............................................................. 30

Figure 12 W ind Experiment: Slab-On-Grade Back View .............................................................. 30

8

Figure 13 Wind Experiment: Basement Monitoring Locations................................................. 31

Figure 14 Wind Experiment: Slab-On-Grade Monitoring Locations ........................................ 31

Figure 15 Depressurization Experiment: 0.5% Foundation Crack Configuration....... 33

Figure 16 Depressurization Experiment: 0.08% Foundation Crack Configuration...... 33

Figure 17 Depressurization Experiment: Basement Side View.................................................. 34

Figure 18 Depressurization Experiment: Basement Back View ................................................ 34

Figure 19 Depressurization Experiment: Slab-On-Grade Side View........................................ 35

Figure 20 Depressurization Experiment: Slab-On-Grade Back View....................................... 35

Figure 21 Depressurization Experiment: Basement Monitoring Locations ......................... 36

Figure 22 Depressurization Experiment: Slab-On-Grade Monitoring Locations................ 36

Figure 23 -

Figure 24 -

Oxygen Replenishm ent Rates................................................................................................

A irflow Into B u ilding....................................................................................................................

57

7 2

Figure 25 Pressure Transducer 1 Calibration.................................................................................. 97

Figure 26 Pressure Transducer 2 Calibration..................................................................................

Figure 27 Untransform ed Data Exam ple................................................................................................100

98

Figure 28 Log-Transform ed Data Exam ple ........................................................................................... 100

LIST OF TABLES

T able 1 G rain Size A nalysis.............................................................................................................................

Table 2 Building Configuration Naming Scheme ............................

Table 3 Barometric Pressures........... ..... .... .............

...... Table 4 Peristaltic Pump Flow rates...................

Table 5 Average Building Pressures............ ........

Table 6 BAS_10 ft/s Total Hydraulic Head ..................

..........

........

......................

...........

22

......

38

40

................. 43

Table 7 BAS_15 ft/s Total Hydraulic Head ............................

Table 8 BAS_20 ft/s Total Hydraulic Head ...................................

17

....... 44

45

46 Table 9 BASSW_10 ft/s Total Hydraulic Head................................

Table 10 BASSW_15 ft/s Total Hydraulic Head.............................. 47

Table 11

Table 12

-

-

BASSW_20 ft/s Total Hydraulic Head...............................

SOG_10 ft/s Total Hydraulic Head...................... .............

48

49

50 Table 13 SOG_15 ft/s Total Hydraulic Head..................................

Table 14 SOG_20 ft/s Total Hydraulic Head........................... ......... 51

9

Table 15 SOG_SW _10 ft/s Total Hydraulic Head...............................................................................

Table 16 SOG_SW_15 ft/s Total Hydraulic Head...............................................................................

Table 17 SOG_SW _20 ft/s Total Hydraulic Head...............................................................................

Table 18 Oxygen Replenishment Rates................................................................................................

Table 19 BAS_0.08%_10.9 Pa Total Hydraulic Head .......................................................................

Table 20 BAS_0.08% _19.4 Pa Total Hydraulic Head.......................................................................

Table 21 BAS_0.08%_27.2 Pa Total Hydraulic Head.......................................................................

Table 22 BAS_0.5%_5.84 Pa Total Hydraulic Head..........................................................................

Table 23 BAS_0.5% _10.1 Pa Total Hydraulic Head..........................................................................

Table 24 BAS_0.5% _13.8 Pa Total Hydraulic Head..........................................................................

Table 25 SOG_0.08%_18.5 Pa Total Hydraulic Head.......................................................................

Table 26 SOG_0.08%_35.6 Pa Total Hydraulic Head.......................................................................

Table 27 SOG_0.08%_49.3 Pa Total Hydraulic Head.......................................................................

Table 28 SOG_0.5% _8.45 Pa Total Hydraulic Head..........................................................................

Table 29 SOG_0.5%_14.8 Pa Total Hydraulic Head..........................................................................

Table 30 SOG_0.5% _20.6 Pa Total Hydraulic Head..........................................................................

Table 31 Airflow Into Building ......................................................................................................................

Table 32 Calculated Hydraulic Conductivities..................................................................................

61

62

63

64

65

56

59

60

52

53

54

66

67

68

69

70

72

92

10

1 Introduction

1.1 Background

Vapor intrusion is the vapor-phase migration of volatile organic compounds (VOCs) into buildings due to subsurface soil or groundwater contamination. The vapor intrusion pathway can cause potential human exposure to VOCs and result in-adverse health effects.

The contaminants of interest related to vapor intrusion are typically petroleum hydrocarbons (PHCs) or chlorinated hydrocarbons (CHCs), though the pathway is applicable to any volatile chemical. The vapor intrusion pathway can potentially result in concentrations of VOCs in indoor air that exceed the human health-based indoor air regulatory concentrations, set by the United States Environmental Protection Agency

(USEPA). Additionally, methane and certain other volatile chemicals can pose explosion hazards when they accumulate in confined spaces in the presence of oxygen at levels between the upper and lower explosion limit (UEL and LEL) (EPA 2013a, b).

The USEPA issued the first set of draft guidance documents (EPA 2001, 2002) in 2001 and

2002 to address the issue of vapor intrusion and offer guidance on how to manage a vapor intrusion site. These draft documents provided guidance for the investigation and management of vapor intrusion at Resource Recovery and Conservation Act (RCRA),

Superfund, and Brownfield sites, but did not address vapor intrusion for petroleum releases at underground storage tank sites (Kalmuss-Katz 2013). Several proposed revisions (EPA 2013a, b) to the draft guidance documents have been released for comment with the intention to have them finalized in 2013. To date these have not been finalized. In addition, beginning in the 2000s, several states began to develop their own guidance

11

documents, in many cases consistent with the federal guidance. Today, a majority of states have adopted some form of a vapor intrusion regulatory program (Levy 2013).

If there is reason for suspicion of the vapor intrusion pathway at a site, an investigation is required per federal guidance. Practical experience with assessing vapor intrusion has demonstrated that this pathway can be extremely challenging to assess (Swartjes et al.

2011). Typically, indoor air regulatory levels are low (on the order of micrograms per cubic meter), so extra care must be taken to avoid contamination during sampling. Certain contaminants may also be detected in air inside buildings due to emissions from the use of consumer products, building materials, and outdoor air sources, thus the contribution from the subsurface may be difficult to determine. Furthermore, indoor air concentrations of

VOCs vary with season, location, weather, life-style, and building ventilation rate. These issues contribute to a high level of uncertainty and variability when assessing the potential and degree of vapor intrusion (Swartjes et al. 2011).

This variability in indoor air concentrations and the large number of controllable and uncontrollable factors that affect indoor air concentrations can lead to unnecessary remediation costs. The uncertain nature of vapor intrusion investigations and analyses can predict a greater contribution to indoor air from vapor intrusion than actually exists.

Mitigation strategies and technologies (e.g., sub-slab depressurization, soil vapor extraction, vapor barriers) are relatively inexpensive when compared to performing extensive investigations (Swartjes et al. 2011). Depending on site conditions, mitigating vapor intrusion in the absence of sufficient data may be more cost effective than completely

12

evaluating the pathway through extensive data collection and negotiating with the various stakeholders, including the regulatory agencies.

To obtain an estimate of the vapor intrusion pathway, scientists and engineers often measure subsurface soil, soil vapor, or groundwater concentrations and use a computational model to predict an attenuation factor (a). The attenuation factor, a dimensionless factor, is the ratio of the indoor air concentration to the sub slab concentration. The models used to predict the attenuation factor heavily rely on several assumptions and often result in a relatively inaccurate analysis (Swartjes et al. 2011).

Additional research on the vapor intrusion pathway is needed to create a more reliable and accurate model.

1.2 Literature Review

The physical transport of soil gas through the vadose zone and into basements has been studied in both the field and laboratory, but is still not fully understood. It is commonly assumed that under natural conditions, soil vapor diffusion dominates the migration of

VOCs through unsaturated soils (Johnson et al. 1998). However, the major contributor to soil vapor entry into basements is often thought to be dominated by advection caused by the depressurization of buildings (Swartjes et al. 2011). There are many ways in which depressurization of buildings occurs including the "stack effect," temperature differences between the building and outdoor air, barometric pressure variations, pressure differentials created by wind loads against a building, and the operation of heating ventilation air condition (HVAC) systems (Sakaki et al. 2011).

13

Studies have been conducted to quantify these methods of depressurization and the factors that affect these processes. A subset of the relevant literature includes studies of the effects of soil moisture and temperature (Sakaki et al. 2011), atmospheric pressure fluctuations

(Robinson at al. 1997), soil type and stratigraphy (Escobar et al. 2010), and basement floor integrity (Fischer et al. 1996). Based on these studies, the most important factors affecting the migration of soil vapor are soil properties such as moisture content and permeability, as well as pressure fluctuations within the building. Using the critical soil properties in combination with physical environmental conditions, general analytical transport models describing soil vapor transport in the subsurface have been quantified, as described in the this literature.

In addition to the physical transport of soil vapor in the vadose zone, biodegradation rates of subsurface contaminants are an important factor in determining the attenuation factor.

Especially in the case of PHC contamination where aerobic conditions enhance the biodegradation process, the rate at which oxygen is replenished in the subsurface is an important factor that has not been extensively studied or quantified. To gain a better understanding of the oxygen replenishment rate beneath a full-scale house, Lundegard et al. (2008) performed field studies and modeled the migration of nitrogen and oxygen in unsaturated soil. This study concluded that strong winds (>3 meters per second) increased the oxygen replenishment rate, while an increase in soil moisture due to the infiltration of rainwater decreased this replenishment rate. In addition, the spatial and temporal oxygen data suggest rapid replenishment immediately below the building slab followed by downward diffusion of oxygen.

14

1.3 Study Approach

This study examines subsurface flow of air beneath a model building through two laboratory experiments. The Wind Experiment is intended to quantify the oxygen replenishment rate beneath a building due to variations in above ground wind speed. The

Depressurization Experiment models how air flows into a building as a result of negative pressures within the building.

The Wind Experiment was intended to address the transport of oxygen below a building on a VOC-impacted property. Properties contaminated with VOCs as a result of leaking fuel storage tanks, or other sources can cause indoor air contamination as a result of vapor intrusion. Depending on the type of VOC of concern, oxygen may promote (in the case of

PHCs) or inhibit (CHCs) the degradation process of these compounds in the subsurface.

The rate at which atmospheric oxygen is transported beneath a building is significant for the biodegradation rate of VOCs and thus the potential for vapor intrusion. The Wind

Experiment further evaluates the findings of Lundegard et al. (2008) that strong winds experienced above ground increase the rate at which oxygen is transported in the subsurface.

In addition, depressurization within a building also increases the potential for vapor intrusion by influencing the magnitude and direction of advective soil gas intrusion into a basement (Fischer et al. 1996). There are many parameters that cause depressurization within a building including HVAC systems, temperature gradients, and barometric pressure changes. However, the relative contributions of advection by these parameters are not well

15

understood (Lundegard et al. 2008). The Depressurization Experiment provides additional research on subsurface airflow into a model building due to building depressurization.

16

2 Methods

The following sections discuss the materials used and procedures of the study.

2.1 Materials

2.1.1 Sand

The experiments utilized clean, kiln-dried Quikrete All-Purpose sand.

A grain size distribution analysis was performed on the sand using a RO-TAP mechanical sieve machine

(Photo 1 in Appendix D) in accordance with ASTM C136 (ASTM Standard C136). The results of the grain size analysis are shown in Table 1 and Figure 1.

Table 1 Grain Size Analysis

Measurements Overload ing Check Results

Sieve Sieve Mass of Mass of Mass of Sieve Accumulated Percent

Number Opening Sieve Soil + Sieve Retained Limit

(mm) (g m) (gim) Soil (gm) (9m)

% Retained Finer

(%)

(%)

8

20

40

2.36

0.85

0.43

528.6

410.6

574.6

532.20

426.60

600.90

3.60

16.00

26.30

211.0

76.0

38.0

4.2

23.1

54.0

95.8

76.9

46.0

60

80

100

Pan

0.25

0.18

0.15

0.00

529.6

430.7

540.3

377.3

550.10

439.70

543.70

383.50

20.50

9.00

3.40

6.20

22.4

15.8

13.4

0.0

78.1

88.7

92.7

100.0

21.9

11.3

7.3

0.0

17

100

90

80

70

60

S50

40

30

2 20

10

0

10.000

Summary of Grain Size Analysis

-

1.000

Particle Diameter (mm)

-

0.100

Figure 1 Summary of Grain Size Analysis

Based on the grain size analysis, the sand used in this study is classified as poorly graded sand (SP) according to the Uniform Soil Classification System (ASTM Standard D2487). The sand had a porosity of approximately 0.29, which was calculated by adding a known amount of water to a representative sample of the sand. This process is further described in Appendix A.2 and is shown in Photo 2. The hydraulic conductivity of the sand (1.7x10

4 meters per second [m/s]) was obtained via a constant head test in accordance with ASTM -

D2434 (ASTM Standard D2434). A photo of the apparatus is shown in Photo 3 in Appendix

D. An in-depth description of this process is included in Appendix A.3 along with the conversion to conductivity of air (1.05x105 m/s) and calculation of intrinsic permeability

(1.7x10-" M 2

).

18

2.1.2 Sand Box and Building

The experiments were conducted in a three-quarter inch thick plywood box (4 feet long, 2 feet wide, 1 foot tall) filled with approximately eight cubic feet of the sand. A one cubic foot simulated building structure (building) was constructed of three-quarter inch thick plywood that was glued and screwed together to create a near-air tight seal. The size of the building was intended to simulate a small, two-story building with approximate footprint area of 625 square feet (1:25 scale). See Appendix A.1 for an in-depth description of scaling the model building. The building was placed four inches below the top of the sand to simulate a basement condition, and placed on top of the sand for slab-on-grade configuration. A small piece of wood was also placed on the sand in front of the building to simulate a sidewalk. For more information regarding the various experiments and building configurations, refer to Sections 2.2.4 and 2.2.5. For the depressurization experiment,

3/16" diameter holes were drilled in the bottom of the building to simulate cracks that are found within the concrete slab of a building. The holes promoted airflow from the porous media to the building so differential pressure readings could be recorded beneath the building. This study utilized two different percentages (0.08% and 0.5%) of "cracks" in the foundation. The orientations of the holes are described in Section 2.2.5.1. The calculation of crack percentage is available in Appendix A.4. It should be noted that round holes were used rather than straight cracks because scaling down straight cracks to laboratory scale would have created considerable edge effects (Fischer et al. 1996). Although round holes cause different patterns of air flow than thin cracks, this difference is small compared to the large edge effects of a thin crack on a laboratory scale (Fisher et. al. 1996).

19

2.1.3 Differential Pressure Transducers

The specifications of the differential pressure transducers used in this experiment

(Freescale Semiconductor) are available in Appendix G. On one end, the transducers were connected to wiring harnesses, which connected to the data acquisition module. Two pieces of 7/32" outer diameter flexible poly tubing (ClearFlex 60, Flex Tubing Products) were connected to the opposite end of the transducers. One piece was connected to a 14inch long piece of 3/16-inch stainless steel welded tubing (McMaster Carr) installed at various locations in the box, while the other piece was left open to the atmosphere in an undisturbed location to reduce the chance of pressure interference. The stainless steel tubing was installed to the desired locations described in Section 2.2.4.1 and 2.2.5.1 prior to filling the tank with sand in order to help minimize the chance of void spaces and preferential pathways to occur. In total, ten transducers were available for use. However, after initial testing, some significant variability was discovered in the various transducers.

For this reason, only one or two transducers were used during the experiments to help ensure instrument consistency.

2.1.4 Data Acquisition System

The data acquisition system (DAS) used in this study consisted of an IOTech Personal Daq

56 data acquisition module connected to a desktop computer equipped with Acqlipse Data

Acquisition software. The data acquisition module was comprised of an analog-to-digital

(A/D) converter connected to a sequential sampling multiplexer. A total of ten ports were wired to the multiplexer, allowing a maximum of ten devices to be used at one time. The differential pressure transducers were provided with approximately five volts of electricity, as recommended by the transducer specifications (see Appendix G). As the

20

transducers experienced pressure fluctuations, output voltages were sent from the transducers to the A/D converter via the multiplexer, which were then displayed real-time on the desktop computer software. The input voltage was also recorded on a separate port and then sent to the computer. The software was set to achieve optimum accuracy and precision in voltage readings for the expected output voltage readings and was set to collect a reading once per second during each test. Photo 4 shows the components of the

DAS, while Photo 5 shows a screenshot of the software interface.

2.1.5 Propeller Fan

Wind was created using a Dayton propeller fan (model 4UX61J) placed at one end of the experimental tank. A variable frequency drive (Variac, General Radio Company) was connected to the fan to allow consistent control of the wind speed. Further discussion of how wind speeds were obtained is provided in Section 2.2.4.2. At higher wind speeds, a large amount of turbulence created by the fan was experienced during the experiment. A discussion of this air turbulence is presented in Section 5.

2.1.6 Peristaltic Pumps

Two Cole Palmer Instruments Co. peristaltic pumps (models 7015-02 and 7015-72) were used to depressurize the building. Masterflex controllers connected to the pumps allowed for the airflow rate to be controlled manually. These flow rates are discussed further in

Section 2.2.5.3 and 3.2.4. Photo 7 shows the peristaltic pumps and how they were connected to the building. The process of varying the building vacuum is described in

Section 2.2.5.3.

21

2.2 Procedures

The experimental procedures are described in the following sections.

2.2.1 Building Configuration Naming Scheme

A total of eight different building configurations were used during this study. Table 2 provides brief descriptions of each configuration and its corresponding name used throughout the study.

Table 2 Building Configuration Naming Scheme

Description Test Name

Basement

Basement with Sidewalk

Slab-On-Grade

Slab-On-Grade with Sidewalk

Basement with 0.08% of Foundation Crack

Slab-On-Grade with 0.5% of Foundation Crack

BAS

BASSW

SOG

SOGSW

BAS 0.08%

Basement with 0.5% of Foundation Crack BAS_0.5%

Slab-On-Grade with 0.08% of Foundation Crack SOG 0.08%

SOG_0.5%

2.2.2 Barometric Pressure Readings

Throughout the duration of this study, barometric pressure readings were recorded at the beginning and end of each test using a wireless advanced weather station (Ambient

Weather WS-1171A). Although the transducers used in this study measured differential pressure, and the barometric pressure should not theoretically have an effect on the results, barometric pressure was recorded in order to obtain a complete set of data for the study. The barometric pressure readings are provided in Table 3.

22

Test

BAS

BAS SW

SOG

SOG SW

BAS 0.08

BAS 0.5

SOG_0.08

SOG_0.5

Table 3 Barometric Pressures

Barometric Pressure (kPa)

Start End Change

102.17

100.58

99.97

99.73

100.58

100.34

99.73

99.59

1.59

0.24

0.24

0.14

102.78

101.96

103.01

102.68

101.83

102.98

0.1

0.13

0.03

102.98 102.84 0.14

Overall, barometric pressure remained fairly constant throughout the study. Note that these barometric pressures were not used in any calculations performed as part of this study.

2.2.3 Differential Pressure Transducer Calibration

Prior to beginning the experiments, the differential pressure transducers were calibrated using the apparatus shown in Photo 6. The pressure reading from the transducer was recorded simultaneously with the measured difference in the water column. A linear data fit was applied to the transducer and water column readings to develop an equation to convert the transducer output of volts to kilopascals (kPa) for each transducer. Appendix

A.5 further describes the transducer calibration process.

2.2.4 Wind Experiment

The Wind Experiment examines the flow of air beneath a small building induced by simulated wind conditions. To monitor the flow of air within the porous media, differential

23

pressure readings were collected at various locations beneath the building at different wind speeds.

2.2.4.1 Building Configurations

Four different building configurations were used in this experiment. For basement conditions (BAS), the building was placed four inches below the top of the sand. A 1-foot

by 3.5 inch piece of wood three-quarter inch thick wood was placed upwind of the building to simulate a sidewalk attached to a building (BASSW). These same configurations were applied to the slab-on-grade conditions (SOG; SOGSW), where the building was placed at the surface of the sand. Photos 8 through 12 show the various building configurations.

2.2.4.2 Wind Speeds

A variable frequency drive (VFD) was connected to the propeller fan to allow for various wind speeds to be achieved consistently. Three different average wind speeds were tested during the wind experiment (approximately 10 feet per second [ft/s], 15 ft/s, and 20 ft/s).

These values were chosen based on the average wind speeds experienced in Boston,

Massachusetts (Windfinder.com). A wind rose for Boston is shown in Figure 2.

24

NW

NNW

10

N

NNE

NE

WNW ENE

W E

WSW ESE

SW SE

SSW SSE

S

Source: Windfinder.com

Figure 2 Boston Wind Rose

Higher wind speeds would have been ideal; however the propeller peaked at approximately 20 ft/s. Due to the turbulent and inconsistent flow of air created by the fan, wind speeds were recorded at various locations on the building using a handheld anemometer (LaCrosse, EA-30104). The average speeds and locations are shown on Figure

3 through Figure 8.

25

242

2'

Figure 3 - Measured Wind Speed for BAS and BASSW at 10 ft/s

45.

Figure 4 - Measured Wind Speed for BAS and BASSW at 15 ft/s

26

2447

T2-JL6'

197

Figure 5 - Measured Wind Speed for BAS and BASSW at 20 ft/s

69.8

6r

8..2 FS .9

Z

Figure 6 - Measured Wind Speed for SOG and SOGSW at 10 ft/s

27

34

*4 .8

3

14.4

3.

13.8112.8-1

3-

Figure 7 - Measured Wind Speed for SOG and SOGSW at 15 ft/s

17.1

1.1901

11 .7

20.0+

Figure 8 - Measured Wind Speed for SOG and SOGSW at 20 ft/s

28

2.2.4.3 Data Collection

Differential pressure data was collected individually at each monitoring location with a single transducer at a rate of one reading per second. Schematics of the apparatus including monitoring locations for basement and slab-on-grade configuration are shown on

Figure 9 through Figure 14.

Propeller Fan

Sand

Differential pressure transducer location

Figure 9 - Wind Experiment Basement Side View

Data acquisition system

Differential pressure transducer

Stainless Steel Tubing

Figure 10 - Wind Experiment: Basement Back View

29

Propeller Fan

29 40 60 $0 1 IM0

It*70 9*0 t

Differential pressure transducer location

Figure 11 Wind Experiment: Slab-On-Grade Side View

Sand

I

Data acquisition system

Differential pressure transducer

I

I

ZZ

Stainless Steel Tubing

Figure 12 Wind Experiment: Slab-On-Grade Back View

30

15

S

I41

6'

3S

'5

15

3.

Z'.O

---------------------------------

3.

2.V

Is ft-

~~ 2S'

3.

3'

2S'

- -----

-

I

Is

S

---------

Figure 13 Wind Experiment: Basement Monitoring Locations

19

-

21

SL35

------------------

ZOO

-----a

3'.

1)IL b S -

-

t

3-

2-S

T,

- - - - - - - - - - - - - - - - - - - - - - - - - - - -

Figure 14 - Wind Experiment: Slab-On-Grade Monitoring Locations

At each location, a baseline of pressure data was recorded for approximately 90 seconds before setting the VFD and propeller fan to approximately 10 ft/s wind speed. The

31

pressure was allowed to reach a steady state prior to increasing the wind speed to 15 ft/s.

After the pressure readings stabilized a second time, the wind speed was increased to approximately 20 ft/s. Once a steady state was reached, the fan was turned off and the final baseline was recorded. This process was repeated at each location and for each building configuration.

2.2.5 Depressurization Experiment

The Depressurization Experiment examines the effect of negative pressure within a small building on the differential pressure beneath the building slab. To monitor these conditions, differential pressure was recorded both in the building and in the porous media beneath the building.

2.2.5.1 Building Configurations

Similar to the wind experiment, basement (BAS) and slab-on-grade (SOG) configurations were tested (see Photo 15 and 16). Within each configuration, two different building foundation integrities were tested. In one condition, 25 3/16-inch diameter holes were drilled in the bottom of the building in a symmetrical fashion (see Figure 15 and Photo 13).

This results in a total percentage of cracks in the foundation floor of approximately 0.5%

(BAS_0.5%; SOG_0.5%).

32

2.5"

2'

2.5"QQ

0

0

0 0 0

0

0

Figure 15 Depressurization Experiment: 0.5% Foundation Crack Configuration

For the second foundation condition, all but four of the holes were duct taped closed (see

Figure 16 and Photo 14) to create a total crack percentage of approximately 0.08%

(BAS_0.08%; SOG_0.08%). The calculation of basement crack percentage is available in

Appendix A.4.

2.5"

2"

2"0

2.5"

Figure 16 - Depressurization Experiment: 0.08% Foundation Crack Configuration

33

2.2.5.2 Data Collection

Differential pressure was recorded simultaneously in the building and at the various subsurface monitoring locations using two transducers at a rate of one reading per second.

Schematics of the apparatus including monitoring locations for basement and slab-ongrade configurations are shown on Figure 17 through Figure 22.

Silicon Tubing

Peristaltic

Pump 10170

20

60 79 so 90

100 110 120 130

160

Peristaltic

Pump

Figure 17 Depressurization Experiment: Basement Side View

Data acquisition system

Figure 18 - Depressurization Experiment: Basement Back View

34

Data acquisition system

Silicon Tubing

Peristaltic

Pump 10

20

9

3.

S. 11

1010 110

140

40

70

130

170

160

Peristaltic

Pump

Figure 19 - Depressurization Experiment: Slab-On-Grade Side View

I --"" I-

Figure 20 - Depressurization Experiment: Slab-On-Grade Back View

35

Z.

0

4-1

2.5"

2.5

2

2.5

2.,S'

2

Figure 21 - Depressurization Experiment: Basement Monitoring Locations

Zr

I" i's

A

I.s

3

Figure 22 - Depressurization Experiment: Slab-On-Grade Monitoring Locations

36

At each location, a baseline was recorded for approximately 60 seconds prior to turning on the peristaltic pumps to create a building vacuum. Following the baseline, one pump was turned on to full speed (17.4 cm 3 /s) and pressure readings were allowed sufficient time to stabilize. Once the pressure in the building and in the subsurface reached a steady state, the second pump was turned on to half speed (27.9 cm 3 /s). Again, once readings stabilized, both pumps were turned up to full speed (37.3 cm 3 /s) to create the final building vacuum.

Section 2.2.5.3 provides a discussion of the pump speeds and resulting building vacuums.

After stabilization, the pumps were turned off and the original baseline pressures were recorded. This process was repeated at each location and for each building configuration.

2.2.5.3 Building Depressurization

During the tests, two peristaltic pumps worked in parallel to depressurize the building to various vacuums, as shown in Photo 7. The silicon tubing of the pumps was connected to two-inch pieces of 3/16-inch outer diameter stainless steel welded tubing (McMaster Carr).

Each of the two pump speeds was set using a Masterflex controller. Three building pressures were tested during this experiment. Building pressures varied depending on the percentage of foundation crack and how flush the building was placed on the sand, making it difficult to set the building pressure at a consistent value for each run of each test.

Therefore, the settings on the peristaltic pump controllers remained consistent for each of the three building pressures for the duration of the experiment.

To calculate the approximate airflow rates at each of the three pump speeds tested, water was used in place of air. To gather this information, a volume of water was collected over a timed interval for each pump speed. The data for these test runs are shown in Table 4.

37

Table 4 Peristaltic Pump Flow rates

Pump Setting

Pump 1100%

Pump 1 100%

Pump 1100%

Plump 1 100%,

Pump 2 50%

Plump 1 100%,

Pump 2 50%

Plump 1 100%,

Pump 2 50%

Pump 1 100%,

Pump 2 100%

Pump 1 100%,

Pump 2 100%

Pump 1 100%,

Pump 2 100%

Pump 1 100%,

Pump 2 100%

Pump 1 100%,

Pump 2 100%

Volume

Collected (cm 3 )

265

172

174

276

Time (s) Flow rate Average Flow

(cm 3 /s) rate (cm 3 /s)

15

10

10

17.7

17.2

17.4

17.4

10 27.6

282 10 28.2 27.9

279 10 27.9

381

378

369

368

370

10

10

10

10

10

38.1

37.8

36.9

36.8

37

37.3

Theoretically, these average values calculated using water should be approximately equal to those of air due to the mechanical function of the peristaltic pumps. As the rotating shoe of the peristaltic pump passes along the tubing, a volume of fluid is isolated within the tubing, creating a strong vacuum towards the pump. Once the shoe completely passes along the tubing, the fluid volume is released. Although this volume within the tube is initially compressed, it is restored to its original volume once it is released. This action suggests that compressibility should not be a factor and that using water as a substitute fluid for air should result in approximately equivalent flow rates. Average building

38

vacuums resulting from these peristaltic pump flow rates for each of the experiments are provided in Table 5.

39

Note that the values listed are in units of Pascals and that the "Pressure Change in Building" row values are the differences in the average baseline and the average values at each peristaltic pump flow rate for each test. Building vacuums observed in real buildings range from 0 to 10 Pa (Hodgson 1992), but are typically 5 Pa or less (Fischer et al. 1996). The building vacuums in this study range from approximately 5.8 to 49 Pa. Two of the 12 vacuums were less than 10 Pa, six were between 10 and 20 Pa, and the remaining four were greater than 20 Pa

2.2.6 Apparatus Disassembly

Once all the necessary data for this study was collected, the experimental apparatus was disassembled. The reusable pieces of wood to make the box and building were unscrewed and returned to the MIT Course 1 undergraduate laboratory for potential future use. The pieces of wood deemed unusable in the future were discarded in accordance with the MIT

Environmental Health and Lab Safety Building 1 Chemical Hygiene Plan. The sand was placed into five-gallon buckets and was stored in the MIT Course 1 laboratories for potential future use.

41

3 Results

The following sections describe the results of the experiments and the procedures to obtain the results.

3.1 Wind Experiment

3.1.1 Data Reduction

The data collected from the data acquisition system was reduced to single representative values at each location for each condition tested. For example, an arithmetic average pressure value was calculated for basement configuration for the initial baseline, at 10 ft/s wind, at 15 ft/s wind, at 20 ft/s wind, and a final baseline at each monitoring location.

Appendix A.6 discusses the rationale for using arithmetic average values. As the wind speed was increased, it took approximately 70 seconds for the recorded pressures in the subsurface monitoring locations to stabilize. Data prior to this stabilization (gray shaded areas in the time series data plots in Appendix B.1) were not used to compute the representative values.

3.1.2 Total Hydraulic Head

These average representative pressure values were then used to calculate pressure head values at each location. Appendix A.7 provides an in-depth analysis of converting the pressure readings and the elevations from each location to total hydraulic head values.

Table 6 through Table 17 shows the pressure readings, elevations, and corresponding total hydraulic head values for each of the building configurations and wind speeds tested.

42

3.1.3 Isopotential Contours

In order to visualize and quantify the subsurface flow conditions, the computer software program SEEP/W 2012 by GEO-SLOPE International Ltd. was used. SEEP/W is a software product used for analyzing groundwater flow within porous media. In order to modify the model to match this specific study, the air conductivity calculated in Appendix A.3 was entered as the hydraulic conductivity of the fluid. Once this value was entered, the subsurface region was drawn to scale and total hydraulic head values listed in bold in Table

6 through Table 17 were entered at their corresponding monitoring locations. The model then drew isopotential contour lines based on the specified head values. These isopotential contours are provided in Appendix C.1, and their corresponding SEEP/W data reports are available in Appendix E. Note that SEEP/W extrapolates isopotential contours, so contours drawn in locations with no hydraulic head data should be considered highly uncertain.

3.1.4 Oxygen Replenishment Rate

In order to quantify the rate at which oxygen is replenished beneath a building, the Flux

Section tool in SEEP/W was used to calculate the total flow of air across a specified area.

To calculate the flow rate of air beneath the building, the Flux Section line was drawn directly downward from the upwind end of the house down to approximately three inches below the lowest monitoring locations. This accounts for all air transported into the 12 inch long by 12 inch wide by 8 inch deep target area, which was consistent for both basement and slab-on-grade building configurations. For this study, it is assumed that the area at which potential subsurface contamination exists is located directly below the outside edges of the building. Oxygen flux rates and oxygen replenishment rates were then calculated from these airflow rates into this simulated zone of contamination, as described

55

in Appendix A.8. Note that the airflow rates calculated from the Flux Section tool do not take into account the width of the experimental tank. To account for this, the given flow rates calculated by SEEP/W were multiplied by the tank width (two feet or 61 centimeters). Table 18 provides these airflow rates, the oxygen flux rates, and the oxygen replenishment rates. To further analyze these results, oxygen replenishment rate versus wind speeds scatter plots were formulated and are present in Figure 23. Section 4.1

provides a discussion of these results. Note that 1% per minute is equivalent to 1,440% per day.

Table 18 - Oxygen Replenishment Rates

Building

Configuration

BAS

BASSW

SOG

SOGSW

Wind Speed Flow rate Oxygen

3

(ft/s) (cm /s) Rate

(g/min)

10

15

20

10

15

20

10

15

20

10

15

20

3.02

5.34

6.58

1.84

3.38

4.40

3.01

5.16

6.90

1.75

2.74

3.49

0.05

0.09

0.11

0.03

0.06

0.07

0.05

0.09

0.12

0.03

0.05

0.06

% of Oxygen

Replenished per Minute

0.96%

1.70%

2.09%

0.59%

1.07%

1.40%

0.96%

1.64%

2.19%

0.56%

0.87%

1.11%

56

Oxygen Replenishment Rate

S0012,

L302t,

A 0U

2UB

HA

d L

Figure 23 -

Oxygen Replenishment Rates

3.2 Depressurization Experiment

3.2.1 Data Reduction

Similar to the Wind Experiment, arithmetic average representative pressure values were calculated for each test condition. Time series plots of the pressure readings collected in the building and in the subsurface are available in Appendix B.2. Again, gray areas on these plots indicate data that was excluded from the average data calculation due to equilibration time.

3.2.2 Total Hydraulic Head

The same process used in the Wind Experiment was used to calculate head values for the

Depressurization Experiment. Appendix A.7 describes the process by which the average pressure values and the elevations of each location were converted to total hydraulic head

57

values. Table 19 through Table 30 display the pressure data collected during the experiment, monitoring location elevations, and resulting total hydraulic head values for each test of the experiment.

58

3.2.3 Isopotential Contours

Using the same process described in Section 3.1.3, the total hydraulic head values listed in

Table 19 through Table 30 were entered into SEEP/W where isopotential contour diagrams were created for each test condition. The isopotential contour diagrams are available in Appendix C.2 and subsequent SEEP/W data reports are shown in Appendix E.

Similar to the Wind Experiment, isopotential contour extrapolations should be considered

highly uncertain due to lack of data in those regions.

3.2.4 Airflow Into Building

The total flow of air from the subsurface into the building was quantified by drawing

SEEP/W Flux Sections below the sand surface on both sides of the building. These locations theoretically capture the entire flow of air from the subsurface into the building assuming there is no leakage within the experimental tank. These airflow values

(multiplied by the two feet wide dimension not included from the Flux Section tool, as well as a factor of two to include airflow entering the building from all four sides) are tabulated in Table 31 and are graphically displayed with scatter plots in Figure 24. Table 31 also includes the measured peristaltic pump flow rates and the corresponding factor of difference between these and the SEEP/W flow rates. A discussion of these results is provided in Section 4.2.

71

BAS_0.5%

BAS_0.08%

SOG_0.5%

SOG_0.08%

Building

Pressure (Pa

5.84

10.1

13.8

10.9

19.4

27.2

8.45

14.8

20.6

18.5

35.6

49.3

Table 31 Airflow Into Building

Calculated Flow Peristaltic rate (cm

3

/s)

Flow rate (cm

3 /s)

Factor

Difference

14.29 17.4 1.2

23.27

32.46

13.98

20.77

27.07

11.39

17.46

23.53

10.77

27.9

37.3

17.4

27.9

37.3

17.4

27.9

37.3

17.4

1.4

1.5

1.6

1.2

1.1

1.2

1.3

1.6

1.6

15.81

21.42

27.9

37.3

1.8

1.7

LKJo

H

PI/

Flowrate Into Building

""

0.998

4 i.2 -, 86Ui. v t -d 3 , " " ,

E 20 0 x.D

i :, W

1000

5 00

/

4

/ix

+ 4 *114

09 91 1

*Bk~ b 0~

~(X U ~

30

Su~dng Vacm IP.)

Figure 24 Airflow Into Building

72

4 Discussion

4.1 Wind Experiment

4.1.1 Wind Speeds

Figure 23 depicts the relationship between the percent of oxygen replenished in the target area per minute and above ground wind speed. In general, it is clear that increased wind speed increases the rate at which oxygen is replenished in the target area, based on the positive slope of the best-fit lines for each experimental setup (BAS, BAS_SW, SOG,

SOGSW). The limited amount of data in this study does not allow for a statistically sound quantifiable relationship to be established between oxygen replenishment rate and above ground wind speeds, however the available data points are fairly linear based on their R 2 values, which range from 0.97 to 0.99. Examination of the isopotential contour diagrams suggests that there may be stagnant zones below the downwind edge of the building at lower wind speeds for basement building configuration, possibly resulting in decreased oxygen replenishment rates at these locations. For example, tests BAS and BASSW at 10 ft/s resulted in stagnant zones, while these zones did not appear for tests run at 15 and 20 ft/s. The extent of these stagnant zones should be considered approximate due to lack of data within these regions. In addition, the airflow rates and oxygen replenishment rates calculated in this experiment are most likely much larger than those that would be observed in the field. Factors that could decrease these rates include soil moisture content and soil stratigraphy (Sakaki et all. 2011, Escobar et al. 2010), both of which were idealized in this study. Section 5 addresses these data gaps and the need for further research.

73

4.1.2 Building Configuration

4.1.2.1 Basement versus Slab-On-Grade

Further examination of Figure 23 shows that there is not a large difference in the best-fit lines between the basement and slab-on-grade building configurations. The slope for BAS is 0.0011, while the slope for SOG is only slightly greater at 0.0012. This comparison is also somewhat true for the BASSW and SOGSW experiments (slopes of 0.0008 and 0.0006, respectively). There did not appear to be any significant qualitative trends between basement and slab-on-grade configurations upon examination of the isopotential contour diagrams.

4.1.2.2 Sidewalk

Based on the slopes of the best-fit lines in Figure 23 between BAS or SOG and their corresponding BASSW or SOGSW tests, it is fairly clear that the addition of a simulated sidewalk decreased the overall oxygen replenishment rate beneath the building to the target area. Both the slopes of the BAS (0.0011) and SOG (0.0012) tests are larger than that of the BASSW (0.0008) and SOGSW (0.0006) tests, respectively. Comparison of the isopotential contour diagrams for experiments with and without the simulated sidewalk further confirm this data as there appears to be a smaller hydraulic head gradient on the upwind side of the sidewalk experiments. It should be noted that this is a subjective qualitative analysis and no additional calculations were performed to confirm this hypothesis. A discussion of suggested methods to further quantify these results is presented in Section 5.

74

4.2 Depressurization Experiment

4.2.1 Building Depressurization

The data plotted in Figure 24 clearly indicates that increased building vacuum results in increased airflow into the building. This is evident based on the positive slopes of the best-

fit lines for the data. There is insufficient data to suggest statistical correlations, however this data suggests a linear relationship based on R 2 values all greater than 0.99. These calculated airflow rates should be considered idealized and real flow rates observed in the field would most likely be less due to soil moisture and soil stratigraphy. There does not appear to be any qualitative trends from examining the isopotential contour diagrams other than higher building pressures result in greater hydraulic head gradients beneath the building.

To perform a quality check on the airflow data produced by SEEP/W, these airflow rates were compared to the peristaltic pump flow rates measured using water. As shown in

Table 31, these two values are different by a factor ranging from 1.1 to 1.8. The following are possible reasons for these differences:

" The method of converting water flow rates to air flow rates through the peristaltic pumps may have included some error. There was no correction factor applied to either value, and since water and air have different physical properties, notability differences in compressibility, this may not be an entirely accurate assumption.

" As air was pumped from the building, a vacuum was created. The actual volumetric air flow rate may have been decreased due to the pressure difference between the

75

building and the atmosphere. The vacuum created in the building reduces the density of air, thus decreasing the total volume of air pumped.

* The analysis performed by SEEP/W was modified in order to match the laboratory setup. This modeling program is designed to characterize and model groundwater for full-scale sites. It is possible that in converting a full-scale groundwater site scenario to a laboratory-scale airflow scenario there is some potential for error.

4.2.2 Building Configuration

4.2.2.1 Basement versus Slab-On-Grade

Examination of Figure 24 suggests a possible relationship between basement and slab-ongrade building configuration. The slopes of the best-fit lines for BAS_0.5% (2.28) and

BAS_0.08% (0.803) tests are over two times greater than those of the SOG_0.5% (0.998) and SOG_0.08% (0.344) tests. One possible explanation for this difference is that the more sand surrounding the basement of the building created a better seal around the foundation holes, allowing for a higher vacuum to be achieved inside the building. When comparing the building vacuums produced at equal peristaltic pump speed settings, the basement configuration does appear to result in higher building vacuums. As discussed in the previous section, a higher building vacuum results in a greater airflow rate into the building. Examination of the isopotential contour diagrams does not provide any additional significant trends between basement and slab-on-grade configurations.

4.2.2.2 Foundation Crack

Figure 24 provides information that possibly suggests a decrease in percentage of foundation cracks results in decreased airflow into a building. Similar to basement versus

76

slab-on-grade, there appears to be a factor of approximately two difference between in the slopes of the airflow rate equations between basement crack percentages of 0.5% and

0.08% for both basement and slab-on-grade configurations. For example, the slope of

BAS_0.5% (2.28) is approximately double that of BAS_0.08% (0.803). A similar trend appears with the slab-on-grade tests (SOG_0.5% of 0.998 and SOG_0.08% of 0.344).

However, due to the similarities between this analysis of foundation crack percentage and the basement versus slab-on-grade analysis described in the previous section, it is unclear whether this difference in airflow into the building is a result of building configuration or foundation crack percentage.

Although there appears to be a fairly strong trend between foundation crack percentages, the data may not be entirely accurate based on the wide range of building vacuums experienced during the experiment. The results of test SOG_0.08% (slope of 0.344) are much lower than those of the other three tests. One possible reason why the results between SOG_0.08% and the rest of the experiments vary significantly could be because the building pressures for SOG_0.08% were much higher than the other experiments

(SOG_0.08% ran at building pressures of 18.5 to 49.3 Pa, while the rest ran from 5.84 to

27.2 Pa). A possible reason why the building pressures for SOG_0.08% were higher could be because there was less void space beneath the building as it was placed on the sand, creating a more airtight seal around the bottom of the building, thus increasing the vacuum potential inside the building. Examination of the isopotential contour diagrams for

BAS_0.08% suggests that this may have been true, as there appears to be one side of the building with higher head gradients. There may have been a void space on the opposite side of the building floor, which could result in a smaller gradient and possibly less flow of

77

air, assuming constant hydraulic conductivities. If this hypothesis is true, it may be seen that a smaller percentage of foundation crack could result in higher building pressures and a higher flow rate of air into the building.

78

5

Additional Research Considerations

In order to obtain a more complete and accurate set of data for future research on the topics discussed in this study, the following issues should be addressed:

" Use of a more powerful fan with perhaps a different blade configuration that is able to provide a more laminar flow of air. Based on the pressure data in the Wind

Experiment time series plots provided in Appendix B, as well as the wind speed figures in Figure 3 through Figure 8, it is clear that the propeller fan provided a turbulent flow of air during the study. As shown in the in the Wind Experiment time series plots, there is a high level of variability in the pressure readings, especially at higher wind speeds. For future research, it would be ideal to use a wind source that provides a more laminar flow, especially at higher wind speeds. This would decrease variability in the data, and perhaps provide more accurate results.

Additionally, a more powerful wind source could allow for a larger number of wind speeds to be analyzed, which could result in a more significant trend in the wind speed versus oxygen flux rate data.

* A larger apparatus with additional differential pressure monitoring locations. As a number of the isopotential contour diagrams were calculated, it became evident that additional data would have been useful to obtain a more complete understanding of subsurface conditions. A larger apparatus would also allow for the entire extent of the area of influence for both the Wind and Depressurization Experiments to be quantified. For the isopotential contour diagrams in this study, there is a large

79

amount of extrapolation calculated by SEEP/W. It would be ideal to have data that captures the entire zone of influence so no extrapolation would be necessary.

* For the Wind Experiment, it would be beneficial to vary the size and shape of the object used as the sidewalk. From the data collected in this study, there may be a trend that the simulated sidewalk reduces the amount of airflow beneath the building. Additional information that correlates the size and shape of the sidewalk and the amount of airflow in the subsurface could yield additional information that may be directly useful to sites impacted with subsurface contamination.

* During the Depressurization Experiment in this study, there may have been an inconsistency in how the building was placed on the porous media. As described in

Section 4.2.2.2, this inconsistency may be the reason why the airflow into the building results for SOG_0.08% does not correlate well with the rest of the tests.

Additional experimentation on this topic could potentially yield significant correlations between foundation crack and subsurface airflow.

" It would be beneficial to test different foundation crack configurations. For this experiment, only symmetrical configurations resulting in 0.5% and 0.08% of foundation crack were tested. Subsurface airflow patterns as a result of different foundation crack percentages, as well as different placement of the cracks (e.g., around the perimeter of the building footprint) could provide a more complete understanding of the effect foundation integrity has on the vapor intrusion pathway.

80

6 Conclusions

This study quantified the flow of air in porous media beneath a building due to above ground wind speed, building depressurization, and various forms of building configuration or construction through laboratory simulations. Relationships between these factors and the amount of air or oxygen flowing beneath the building were obtained and provide basic insight and understanding of airflow through porous media.

The primary findings from the Wind Experiment show that increased wind speed increases the rate at which oxygen is transported beneath a model building. Although there appears to be no significant quantifiable difference in basement versus slab-on-grade construction, the presence of a simulated sidewalk may decrease this oxygen replenishment rate.

Results from the Depressurization Experiment indicate increased subsurface airflow into the building due to increased building depressurization. There are still questions left unanswered regarding whether building configuration, foundation crack percentage, or some combination of the two is a major contributor to increased airflow in the model building.

The purpose of this study was to quantify the flow of air beneath a building susceptible to subsurface contamination. Although this study provides a basic understanding of these concepts, there are data gaps and additional research is needed to gain a complete understanding of the factors studied here.

81

7 References

ASTM Standard C136. 2001. "Standard Test Method for Sieve Analysis of Fine and Coarse

Aggregates". ASTM International, West Conshohocken, PA. www.astm.org

ASTM Standard D2434. 2006. "Standard Test Method for Permeability of Granular Soils

(Constant Head)". ASTM International, West Conshohocken, PA. www.astm.org

ASTM Standard D2487. 2006. "Standard Practice for Classification of Soils for Engineering

Purposes (Unified Soil Classification System)". ASTM International, West

Conshohocken, PA. www.astm.org

Crowe, Clayton, Donald Elger, Barbara Williams, and John Roberson. 2009.

Engineering Fluid Mechanics. Ninth. John Wiley & Sons.

EPA, a. 2001. "RCRA Draft Supplemental Guidance for Evaluating the Vapor

Intrusion to Indoor Air Pathway (Vapor Intrusion Guidance)."

EPA, b. 2002. "OSWER Draft Guidance for Evaluating the Vapor Intrusion to Indoor

Air Pathway from Groundwater and Soils (Subsurface Vapor Intrusion Guidance)".

EPA 530-D-02-004.

EPA, a. 2013. "Guidance for Addressing Petroleum Vapor Intrusion at Leaking

Underground Storage Tank Sites". EPA 510-R-13-xxx.

EPA, b. 2013. "OSWER Final Guidance for Assessing and Mitigating the Vapor

Intrusion Pathway from Subsurface Sources to Indoor Air (External Review Draft)".

EPA xxx-xx-xxx.

Escobar, Elsy, Paul Dahlen, and Paul Johnson. 2010. "Transport of Petroleum

Hydrocarbon Vapor Components in the Subsurface; a Laboratory Soil Column

Study". Ira A. Fulton School of Engineering, Arizona State University.

82

Fischer, David, and Christopher Uchrin. 1996. "Laboratory Simulation of VOC Entry into Residence Basements from Soil Gas." Environmental Science and Technology 30

(8): 25998-2603.

Hodgson, Alfred, Karina Garbesi, Richard Sextro, and Joan Daisey. 1992. "Soil-Gas

Contamination and Entry of Volatile Organic Compounds into a House Near a

Landfill." Air& Waste Management Association 42 (3): 277-83.

Johnson, Paul, Cristin Bruce, Richard Johnson, and Mariush Kemblowski. 1998. "In

Situ Measurement of Effective Vapor-Phase Porous Media Diffusion Coefficients."

Environmental Science and Technology 32 (21): 3405-9.

Kalmuss-Katz, Jonathan. 2013. "EPA Finalizes Vapor Intrusion Regulations." ENN:

Environmental News Network, January 23.

Kearl, Peter, Nic Korte, T.A. Gleason, and John Beale. 1991. "Vapor Extraction

Experiments with Laboratory Soil Columns: Implications for Field Programs." Waste

Management 11 (June): 231-39.

Laurent, Levy. 2013. "A Review of Vapor Intrusion Guidance by State." Northeast

Waste Management Official's Association, November 28.

Lundegard, Paul, Paul Johnson, and Paul Dahlen. 2008. "Oxygen Transport From the

Atmosphere to Soil Gas Beneath S Slab-On-Grade Foundation Overlying Petroleum-

Impacted Soil." Environmental Science and Technology 42 (15): 5534-40.

Nazaroff, William, Steven Lewis, Suzanne Doyle, Barbara Moed, and Anthony Nero.

1987. "Experiments on Pollutant Transport from Soil into Residential Basements by

Pressure-Driven Airflow." Environmental Science and Technology 21 (5): 459-66.

Neff, David, and Robert Meroney. 1996. "Reynolds Number Independence of the

83

Wind-Tunnel Simulation of Transport and Dispersion About Buildings". Fluid

Mechanics and Wind Engineering Program, Department of Civil Engineering,

Colorado State University.

Plummer, Mitchell, Larry Hull, and Don Fox. 2004. "Transport of Carbon-14 in a

Large Unsaturated Soil Column." Vadose Zone

Journal

3: 109-21.

Robinson, Allen, and Richard Sextro. 1995. "Direct Measurement of Soil-Gas Entry into an Experimental Basement Driven by Atmospheric Pressure Fluctuations."

Geophysical Research Letters 22 (14): 1929-3 2.

-- . 1997. "Radon Entry into Buildings Driven by Atmospheric Pressure

Fluctuations." Environmental Science and Technology 31 (6): 1742-48.

Sakaki, Toshihiro, Paul Schulte, Abdullah Cihan, John Christ, and Tissa Illangasekare.

2011. "Airflow Pathway Development as Affected by Soil Moisture Variability in

Heterogeneous Soil." Vadose Zone

Journal,

September. doi:10.2136/vzj2011.0118.

Swartjes, Frank. 2011. Dealing with Contaminated Sites. Springer Science & Business

Media.

Warrick, A.W. 2001. Soil Physics Companion. CRC Press.

"Windfinder: Wind Statistics." 2014. Windfinder.com. Accessed March 26. http://www.windfinder.com/windstatistics/boston logan airport.

84

Appendix A

-

Calculations

85

A.1 Scaling

In order to obtain data from the experiment that is representative of airflow beneath a full- scale building, the appropriate conditions must be replicated in the laboratory. Here, the concept of similitude must be taken into account to scale up the experimental conditions.

Similitude is the theory and art of predicting prototype performance from model observations and is often quantified by various dimensionless numbers (Crowe et al.

2009). Examples of these dimensionless numbers include the Reynolds number, Mach number, and Froude number. For this experiment, the important physical parameters that will be present include the kinetic and viscous forces around the model building caused by the fan. Thus, the Reynolds number is a key criterion for evaluating the correspondence between experimental and full-scale conditions. The Reynolds number is defined as follows:

Where Re is the Reynolds number (dimensionless), V a representative velocity, L is the length of the object, and v is the kinematic viscosity of air.

Ideally, the Reynolds number at full scale and model scale should be equal. However, this is often difficult to achieve. For example, assuming a ten-mile per hour wind against a 25 foot

by 25 foot building, the wind speed required on a one foot by one-foot model in the lab would need to be 250 miles per hour (see Appendix A.1.1). Not only is this infeasible, but additional effects such as air compression would be introduced to the system that would be absent at full scale. According to Crowe et al. (2009), Reynolds number effects often become insignificant at sufficiently high Reynolds numbers or become independent of the

86

Reynolds number. In order to determine the point at which the Reynolds number is sufficiently high or has become independent, the relationship between the pressure coefficient, a dimensionless number that describes the relative pressures throughout a flow field, versus Reynolds number must be examined. According to various wind tunnel experiments for buildings, flow conditions on a model building should be set such that the

Reynolds number is greater than 15,000 to assure Reynolds number independence (Neff et al. 1996). As shown in Appendix A.1.2, in order to keep a Reynolds number greater than

15,000 on the one-foot model building, a wind speed of at least 2.45 feet per second, or approximately 1.7 miles per hour is required.

87

A.1.1 Scaled Wind Speed

To scale wind effects on structures, the Reynolds number is commonly used:

Equation 1

Where Re is the Reynolds number (dimensionless), V is the velocity, L is the length, and v is the kinematic viscosity of air.

In order for the model to simulate realistic conditions, the Reynolds numbers must be approximately equal:

Equation 2

Assuming the kinematic viscosity of air will be the same, a combination of Equation 1 and

Equation 2 reduces to,

Equation 3

For an average wind velocity of ten miles per hour against a 25 foot by 25 foot house, the wind velocity required on the one foot by one foot model is expressed as:

10 25 1

This yields a wind velocity on the model of 250 miles per hour.

88

A.1.2 Minimum Wind Speed

In order to obtain a minimum Reynolds number of 15,000, the minimum velocity on a onefoot by one-foot house can be expressed by:

Equation 4

Inserting the given values for the model building yields:

15,000 1.63 10

-

1

2.45

-

Note the kinematic viscosity value was obtained from Crowe et al. (2009). Therefore, the wind velocity on the model house will need to be greater than 2.45 feet per second.

89

A.2 Porosity

The porosity of the sand used for this study was calculated using the following relationship:

Equation 5

Where 0 is the porosity, Vv is the volume of void space, and VT is the total volume. The void space volume was measured by adding water to a known amount of soil. Once the soil became fully saturated, the volume of water added to the soil was equal to the void volume.

Note that this process assumes the sand does not expand due to the addition of moisture.

For this experiment, 29 mL of water fully saturated 100 mL of soil. This results in a porosity of 0.29.

29

100

0 .29

90

A.3 Sand Hydraulic Conductivity and Permeability

The following section describes the process by which the hydraulic conductivity and intrinsic permeability of the sand was calculated.

The hydraulic conductivity of a soil material can be estimated empirically by use of the

Hazen equation:

Equation 6

Where K is hydraulic conductivity in water in centimeters per second, C is the dimensionless shape factor (varies widely depending on literature), and D

1 0 is the grain diameter at which 10% of the particles by weight are finer. Using a shape factor of 1

(Warrick 2001), and an approximate Dio based on the sieve analysis from Table 1 and

Figure 1, a hydraulic conductivity was calculated as follows:

1.0 0.18 3 .2 10 /

From the hydraulic conductivity test performed in the lab in accordance with ASTM -

D2434 (ASTM Standard D2434), a hydraulic conductivity can be calculated using the following equation:

Equation 7

Where V is the volume of water collected, L is the length of the soil sample, A is the cross sectional area of the soil sample, h is the head difference, and t is the elapsed time to collect

91

the volume of water, V. Multiple measurements were collected over a three-day period in which the test was run. The resulting hydraulic conductivity values are shown in Table 32.

1

2

3

Table 32 Calculated Hydraulic Conductivities

Time of Calculated Hydraulic

Day Conductivity (m/s)

Morning

Afternoon

4.70E-05

5.50E-05

Morning

Afternoon

Morning

1.70E-04

1.65E-04

1.50E-04

The maximum hydraulic conductivity recorded during this test was used as the value for this study. The maximum represents the most accurate value because there was some lag during the first day of the test in which the sand was yet to become fully saturated, resulting in an underestimate. On the third day of the test, a biological film began accumulating on the sand sample, which decreased the flow of water, resulting in an underestimate. Therefore, the maximum value (collected on the second day) represents the most accurate approximation of hydraulic conductivity. The resulting maximum hydraulic conductivity was calculated as follows:

8.6 10

0.002

0.432

0.103

60

1 .7 10

The hydraulic conductivity calculated from the Hazen equation and the constant head test were within a factor of two, suggesting representative values for the porous media used in this study.

To convert this water conductivity to air conductivity, the following relationships were used (Warrick 2001):

92

Equation 8

Where Ka is the air conductivity, k is the intrinsic permeability Pa is the density of air, pa is the dynamic viscosity of air, and g is the acceleration due to gravity. A similar relationship exists for water conductivity:

Equation 9

Where Kw is the water conductivity, pw is the density of water, and [w is the dynamic viscosity of water. Solving for the air conductivity using Equation 8 and Equation 9 yields the following relationship:

Equation 10

Using the calculated water conductivity, the properties of water and air at a temperature of

294 K (Crowe et al. 2009), and one atmosphere of pressure results in an air conductivity as follows:

1 .7 10 -

1.2

-

1._982

997.95 -

9.8210

1.9210

*

* 1.0510

/

Intrinsic permeability is related to hydraulic conductivity through the following equation

(Warrick 2001):

93

Equation 11

Inserting the values calculated and used above yields the following:

1.0510

-

1.9210

1.2 -

*

9.81 -

1 .7 10

Permeability values for clean sand typically range from 1O-9 to

10-12 m 2 (Warrick 2001).

The permeability calculated for the sand in this experiment is within this range of values.

94

A.4 Percentage of Fully Penetrating Foundation Crack

The percentage of fully penetrating foundation crack is calculated by the area of crack divided by the total foundation area:

% 100

The area occupied by each 3/16" diameter hole drilled into the foundation is calculated using the area of a circle:

3

-

32

0.0276

The building foundation has dimensions of one foot by one foot, resulting in an area of 144 in 2 .

Using all 25 holes drilled in the foundation yields the following foundation crack percentage:

/

25 0.0276

144100

After covering up all but four holes yields the following:

/

4 0.0276

144100

0.5%

0.08%

95

A.5 Differential Pressure Transducer Calibration

In order to convert the output voltage value produced from the pressure transducer, a calibration had to be performed that related output voltage to pressure. To do this, an output voltage reading was recorded for a known pressure. The apparatus shown in Photo

3 was used to create a known pressure value. By forcing air into the tubing, a pressure value resulting from the difference in height of the water column was obtained

(centimeters of water). This value of centimeters of water was then converted to kilopascals (kPa) using the following conversion:

0.09806648

Equation 12

Once this value of centimeters of water was measured, a data point was collected from the output voltage of the pressure transducer, which was connected to the apparatus. The resulting data provides both an input and an output voltage from the transducer for the pressure corresponding to the water column height. The ratio of output volts/input volts was then plotted with the kPa values for each height of water column, as shown in Figure

25.

96

8

.M

6

4

12 in

2

0

0

Seriesl

Linear (Series 1)

1 y = 11.198921x-

R

2

=

0.332572

0.999967

0.2 0.4 0.6

Output/Input (volts/volts)

0.8

Figure 25 Pressure Transducer 1 Calibration

From the equation of the linear fit line to the data, the input and output volts can be converted to Pascals for each transducer using the following equation:

11.198921

0.332572

Equation 13

Where 11.198921 and -0.332572 are constants derived from the line of best fit. Note that the pressure value calculated from the second transducer used in the Depressurization

Experiment included different constants based on a separate calibration analysis. The calibration plot and resulting equation for the second transducer are provided in Figure 26 and Equation 14.

97

4

0 0' 9J

-

-L

~c~es 1 fleJ~ ~Serf~1

,~

11 :9b8?JA U 3$7/1~

0T~1Y4

1 0 04 u' Ok

Output/input (vofts/vokts)

07

Figure 26 Pressure Transducer 2 Calibration

11.195877 0.357714

Equation 14

98

A.6 Frequency Distribution Plot

To determine if the data collected from the Wind Experiment is normally or log-normally distributed, a frequency distribution plot for the data was created. For the data used at each wind speed, the pressure readings were sorted in order from smallest to largest. Each of these values was assigned a rank (1 being the smallest, 2 being the next smallest, etc.).

From this rank, a plotting position was calculated using the following equation:

1

Equation 15

Where n is the number of pressure readings in the data set. The plotting position was then plotted with the following calculated value:

Equation 16

An example of the graph created using untransformed data is provided in Figure 27:

99

Untransformed Data

a

25 281d

Sta Wad Deviate tue 2!

Figure 27 Untransformed Data Example

The same procedure was followed, except the log of each pressure reading was taken prior to calculating the standard deviate. An example of the graph created using log-transformed data is provided in Figure 28:

Log-Transformed Data

4 0 28'2W 4

2 S

Standard Devtate

Figure 28 - Log-Transformed Data Example

For some data sets there is a more significant difference between the two plots, which leads one to choose either an arithmetic or geometric average to properly characterize the data

100

set. In this case, since the two graphs are so similar, there is no significant difference in using an arithmetic or geometric average. One possible reason for the small variability in this analysis is that the range (e.g., 27.1 to 30.2 Pa) of the pressure readings used in the data set is so small.

For this study, the arithmetic average of the pressure data was used. Note that multiple frequency distribution analyses were performed, and each resulted in similar conclusions.

101

A.7 Total Hydraulic Head

Total hydraulic head equation:

2

Equation 17

Where z is the elevation head, referenced to a common datum; is the pressure head, where P is the absolute pressure, p is the density of air, and g is the acceleration due to gravity; and is the velocity head, where V is the air velocity, generally taken as the average linear velocity over a control cross-section.

102

A.7.1 Elevation Head

The elevation head was measured from a common datum, specified in Figure 13, Figure 14,

Figure 21, and Figure 22. Elevation head values range from 0 to 0.27 meters.

103

A.7.2 Absolute Pressure Conversion

To determine the pressure head, differential pressure readings were added to the absolute pressure at each monitoring location. To calculate the absolute pressure at a given elevation, the ideal gas law can be used to relate pressure and elevation for relatively small elevation changes using the following equation (Warrick 2001):

Equation 18

Where the initial pressure, Po, is equal to the pressure, P, at z = zo, where zo is equal to an elevation of 0 meters; M is the molecular weight of air (28.96 g/mol); g is the acceleration due to gravity (981 cm/s 2 ); R is the universal gas constant [8.314x10

7 erg/(mol*K)]; and T is temperature (degrees K). For this study, it was assumed that the initial pressure at zo was atmospheric (101,325 Pa) for each of the experiments. This initial pressure value was just used as a reference point and was intended to allow each test to be compared at an equal starting value. The absolute pressure used in this study represents the initial pressure at a given monitoring location elevation before any wind or vacuum was applied to the apparatus, referenced to a common pressure at zo. An example calculation using an elevation of 26.67 cm, and T of 294 K:

101,325

* 101,324.88

This calculation was performed for each monitoring location elevation and is displayed in

Table 6 through Table 17 and Table 19 through Table 30.

104

A.7.3 Total Pressure

To determine the total pressure at each monitoring location, the absolute pressure was added to the change in gage pressure observed from the impact of the wind or building vacuum. As described in Appendix A.6, the change in gage pressure was calculated by taking the difference of the arithmetic average initial baseline and the arithmetic average pressure readings recorded for each wind speed or building vacuum:

Equation 19

An example calculation using the previously calculated absolute pressure is as follows:

101,324.88 32.59 32.69 101,324.78

Total pressure calculations for each location are available in Table 6 through Table 17 and

Table 19 through Table 30.

105

A.7.4 Pressure Head

Using the pressure head portion of Equation 17, a total pressure head was calculated:

Equation 20

Where p is the density of air and g is the acceleration due to gravity (9.81 m/s 2 ). To account for the fact that air is a compressible fluid and its density changes with elevation, the density of air at each calculated absolute pressure was calculated using the ideal gas law:

Equation 21

Where P is the calculated absolute pressure and R is the universal gas constant (8.314

m

3 *Pa/mol/K). Using the absolute pressure, the air density at this example monitoring location is calculated:

28.96-

8.314

101,324.88

294

Inserting this density into Equation 20 yields:

1000-

101,324.78

1.2-- 9.81--

8,604.30

1.20049-

106

A.7.5 Velocity Head

To calculate velocity head, the velocity of the air was approximated by the following equation (Warrick 2001):

Equation 22

Where

Va is the average linear velocity of the air,

Ja is the volumetric flux density, and

Oa is the saturated air content. The sand used in this study was dry; therefore the saturated air content is equal to the porosity. By using the total pressures, the volumetric flux density for horizontal airflow is described as a simplified form of Darcy's Law:

Equation 23

Where k is the intrinsic permeability (1.7x10-11 M

2

, as calculated in Section Appendix A.3), ta is the dynamic viscosity of air (1.92x10-s kg/m/s at 294 K), and P is the pressure gradient. To estimate the pressure gradient, total pressure values were divided by the distance between two locations at equal elevations for the wind experiment, assuming completely horizontal airflow within the subsurface. Although this assumption may not be entirely accurate, it is a reasonable in order to obtain an order-of-magnitude estimate of velocity head. An example calculation of Ja using Locations 1 and 14 of the slab-on-grade test:

1.7 10

1.9210

,

81.9 Pa

0.36 m' 2 .0

10

107

Inserting this Ja into Equation 22:

2.010

0.29

6 .9 10 -

Inserting this velocity into Equation 20 yields the following:

6.910 -

2 9.81-

5 .9 10

Compared to the prior pressure head calculation (8,608.48 m), this value is negligible and is therefore ignored in the total hydraulic head equation. For additional comparison, differences in total head at different locations are generally larger than 10-2 m (see Table 6 through Table 17 and Table 19 through Table 30), further suggesting that velocity head is negligible. Note that the largest pressure gradient was observed in the wind experiment, so a separate calculation for the depressurization experiment was not performed.

Overall, these calculations show that elevation and pressure are the major contributors to total head for this study. These total head values are the values used in the SEEP/W analysis to create the isopotential contours described in Sections 3.1.3 and 3.2.3.

108

A.8 Oxygen Replenishment Rate

In order to calculate the oxygen replenishment rate, the ratio of volume of air/moles of air must be calculated. Assuming ideal gas conditions, this relationship can be calculated from the ideal gas law re-arranged:

Equation 24

Where V is volume of air, n is the number of moles, R is the universal gas constant (0.082

L*atm/K/mol), T is the temperature in Kelvin, and P is pressure. Assuming a temperature of 294 K and a standard pressure of 1 atm yields the following:

0.82 * 294

24.1

1

To convert the flow rate of air to an oxygen flux rate, the following equation was used:

0.21 32

24.1 -

Equation 25

Where Qair is the flow of air through the subsurface beneath the building; 0.21 moles of oxygen/moles of air corresponds to the approximate fraction of oxygen in air under ideal gas conditions; 32 grams/mol of oxygen is the molecular weight of oxygen (02); and 24.1

L/mol is the value calculated using Equation 24. Inserting an example Qair of 3.02 cm 3 /s into Equation 25 yields an oxygen flux rate of approximately 8.42x10

4

g02/ second or 0.05 g02/minute.

109

The total amount of oxygen present beneath the building in the specified target area can be calculated by replacing the flow of air with the volume of air in Equation 24, as shown here:

18,878 0.21

32

5.26

24.1 - 1000

Where 18,878 cm 3 is the volume of air in the target area beneath the building. The oxygen replenishment rate was calculated as follows:

100

Using the example values calculated above yields following value:

0.05 -

5.26

100 0.95

This value is the percent of the total oxygen in the target zone replenished every minute at this example flow rate of air.

110

Appendix B

-

Time Series Pressure Plots

111

B.1 Wind Experiment

B.1.1 BAS

Location 1

40

45

40

20

50 100 150 200 250

Time (seconds)

300 350 400 450 S0

Location 2

30

50 100 150 200 250

Time (seconds)

300 30 400 450 500

112

Location 3

28

26

24

29

50

100 150 200 250

Time (seconds)

300 350 400 450 500

Location 4

526

24

23

0

50

100 150 2uu

Time (seconds)

113

"o

Location 5

285

27

26

25

2 1

23.5

0 50 100 150 200 250

Time (seconds)

100 350 400 450 500

Location 6

26 .-..

25

24

21

20 p 50 100

150 200 250

Time (seconds)

100 350

400 450 500

114

Location 7

275-

27

26-A

26

25

25

24 5

0 50 100 150 200 250

Time (seconds)

300 350 400 450 500

27.5

27

26.5

26 54

E2fI/

Location 8 lSfs

20ft/s

0.25.5

25

24 5

24

0 50 100

150 200 250

Time (seconds)

300 350 400 450 500

115

Location 9

25.5

25

27.4

27.2

27

268

26.6

214

S26

272

0 50 100

150 200 250

Time (seconds)

300 350 400 450 ,00

Location 10

25.8

50 100 150 200 250

Time (seconds)

300 3SO 400 450 500

116

Location 11

26.4

26

2

26

2

5.8

25.6

&2542

25

26

268

26.2

0

268

26.6

26A

276

274

27.2-

27

--

50 100 150 200 250 lime (seconds)

300 350 400 450 500

Location 12

L0t/EI/

20ft/s

24.6 -7-

24A

24 2

0 50 100

160 200 250

Time (seconds)

300 350 400 450 500

117

28

Location 13

27,5 a.

26,5

29

90 100 1S0 200 250

Time (seconds)

300 3S0 400 4S0 00

Location 14

0 0 100 150 .00 290

Time (seconds)

300

350 400 490 900

118

Location 15

29

28

27

25

24

23

22

21

20

0 50 100 150 200 250

Time (seconds)

300

-

350 400 450 S00

Location 16

29

27

25

23

21

19

17

0 50 100 10 200 250

Time (seconds)

300 350 400 450 500

119

B.1.2 BASSW

50

45

40

30

Location 1

Time (seconds)

Location 2

30

25

50 10r 15i

200 .50

Time (seconds}

300 350 400 450

500

120

Location 3

30

28

L2i

26

S25

.

24

I

I

25.

2

24.

5

2

7 23.

IL 2 3

22,

21.

2 2

2

0

IL E

A&A

21, f"r

21

20

0

50 100 150 200 250

Time (seconds)

300 350 400 a

I

I

450 500

Location 4

*F

50 100 150 200 250

TIme (seconds)

300 350 400 450 500

121

Location 5

225

215

21

25 5

25

24.5

24

*7 23.5

23

100 150 200 250

Time (seconds)

300 350 400 450

500

Location 6

2 S

25--

24 5

24

235

23

-

0 50 100 150

200 250

Time Iseconds)

300 350 400 450 Soo

122

Location 7

25 5

25

24.S

24

23.5

23

22 5

0

25+5

25

50 100 150 200 250

Time (seconds)

300 350 400 450 500

Location 8

10ff/

20ff/s

24

S23.5

22

0 100 200 300

Time (seconds)

400 500 600

123

Location 9

24.5

24

23,5

23

22.5

0 50 100 150

200 250

Time

(seconds)

300 350 400 450

5300

Location 10

24

I

I

100 150 200 250

Time

(seconds)

300 350 400 450 )00

124

Location 11

25

24.5

265

26

25

24

23.5

0

50 100 150 200 250

Time (seconds)

300 350 400 450 500

Location 12

Slo/s [2If§/

27,5

27

26.5

26

7 25 5 jr 25

245

23

24

23

0

50

100 150 200 250

Tkne (seconds)

300 350 400 450 500

125

Location 13

27

265

26

25

23.5

7 7

27 S

0

50 100 ISO 200 250

Time (seconds)

300 350 400 450 o5o

Location 14

265

25 --

225

23

0 0 100 150 200 2$SO

Time

(seconds)

300

350 400 450 500

126

Location 15

27

26

25

23

22

21

20

0

27

50

100

150 200 250

Time (seconds)

300 350 400 450 500

Location 16

-

25

23

21

19

17

15

0 50 100 ISO 200 250

Time (seconds)

300 350 400 450 500

127

B.1.3 SOG

Location 1

60

50 c 45

S40

30

20 i 50 100 150 200 250

Time

(seconds)

300 350 400 450

Location 2

-730

241

20

50

100 150

200 "50

Time (seconds)

300

350 400 450 500

128

Location 3

32

30

3 28

26

24

31

0 50 100 150 200 250

Time

(seconds)

300 350 400 450 500

Location 4

28

727

$26

24

23$

22

0

50 100 150 200 250

Time

(seconds)

300 350 400 450 500

129

Location 5

26

2S

24

28

26

50 100 150 200 250

Time Iseconds)

300

350 400 450 500

Location 6

50 100 15r 200 250

Time (seconds)

300 350 400 45U 500

130

10 ff/s

Location 7

5/s2f/

28

271

27

26

255

245

24

0 s0 100 150 200 250

Time (seconds)

300 350 400 450 500

Location 8

21

24.1

26

245

24

0

28

27.5

27

26.1

50 100 ISO

200 250

Time

{seconds)

300 350 400 450 500

131

Location 9

265

24,5

24

25.5

25

27

26 5

100

200 300

Time (seconds)

Location 10

400 500 600

25.5

24

C) 100 200 300 ime (seconds)

400 500 600

132

Location 11

27 5

IL

26

25.5

2S

24.5

100 200 300

Time Isecond5)

400

Location 12

500

600 700

26

25.5

25

24,$

27.5

27

26.5

24

23.5

0 100 200 300

Time (seconds)

400 500 600 700

133

Location 13

26.5

26

25

5

25

24 5

24

23

22 5

0 100 200 300

Time (seconds)

Location 14

400

500

600

21

I I

50 100 150 200 250

Time (seconds)

300 350 400 450 500

134

B.1.4 SOGSW

45

10l/

Location 1

5Wts 20 ft/s

40

35

30

25

20

0 50 100 150 200 250

Time (seconds)

300 350 400 450 500

Location 2

34

32

30

26

24

50

100 150 200 "50

Time (seconds)

300 350 400 450 500

135

Location 3

28 a

27

29

28

24

31

50

100

150 200 250

Time

(seconds)

300 350 400 450 -00

Location 4

25

24

50

100

150 200

250 ime (seconds)

300 350

400 450 OuU

136

28-

Location 5

10fts 15ft/s2oftls

28.5

28

27 5

27

27

26.5

26

25.5

25

24

5

24

0 50 100 150 200 250 inme

(seconds)

300 350 400 450 500

1ft/s

Location 6

[ 5It/I

2Lft/

25.5

245

245

0

50 100 150 200 250

Time

(seconds)

300 350 400 450 500

137

Location 7

27

26

25.5

245

50 100 150

200 250

Time (weods)

300 350 400 450 Soo

Location 8

24

5

24

100 200

Time

(seconds)

138

WO

Location 9

28

27,5

26

25.5

26.5

26

275

27

25

0 50 100 150 200 250

Time (seconds)

300 350 400 450 500

Location 10

25

24

0

100 200 300

Time (seconds)

400 500

600

139

Location 11

2?

27

5

27

S26 5

26--

2 5.5

25

0 100 200 300

Time

(seconds)

1aft/s

Location 12

5t

400

2fs

500 600

24

23 5

25

24

5-

26

5

26

S25.5

275

7

50 100 150

200 250

Time (seconds)

300 350 400 450 500

140

Location 13

28

27

23

21

25

24

23

29

28

27

26

S25

0 50 100 150 200 250 nme (seconds)

300 350 400 450 500

Location 14

500

Time (seconds)

141

B.2 Depressurization Experiment

B.2.1 BAS_0.08%

20

Location 1

10

10

15

20

S

0 20 40 F0 100 140 160 200 220 260 2 0 100 ;20

---

Stbburfav

L72 a

Time (seconds)

Location 2

15

10

10

15

0 20

40 80 100 120

19.4PO

160 IS

220 2-0 280 300 120 340

-BUb(dm

L7.3i

Time (seconds)

142

Location 3

20

10

-10

-15

20

10 is

20

0

0

20 40

80 100

140 160 200 220

27.2P1

50 280 300 320 340

Time (seconds)

Location 4

20

15

10

5

0 20

40

1

10.9L

80 100 140 160 200 220 i 60 280 300 320

-Buildmig

Subsurface rime (seconds)

143

Location 5

10

10

0

20 40

19A PO

80 100 1 0

160

"Jo.e,

200 220 60 280 300 320

-iuIlding

--

StI SbsIrf atCe

Time (seconds)

Location 6

10

10

0 20 40 6 100 114 140 160 00 220

4 26 280 300 320

Sh us dC

27.2 P

Time (seconds)

144

Location 7

20

10

0

0 20 40 80 100 140 160 200 220

10

15

20

Time (seconds)

Location 8

280 100 320

-Buiding

-Subsurface

-10

^NwrWY

10

0 20 40 80 100 140 160 jo 220 260 280 300 320

-Building

-Subsurface

27.2 Pf

Time (seconds)

145

10

Location 9

20

1s

10

0 20 0 80 100

140 160 200 260 280 300 320

Bulldln ,

-50lbsurftan

Time (seconds)

Location 10 vr7.2A"

IS

15

10.9Pa

0 20 40 FA 80 100 140 160

220 26 280 300 320 40 360 380

-suddin,

-substflace

27.2 Pa

Time (seconds)

146

20

IS

10

Location 11

20

15

10

CL

0 20 40

80 100 120 160 18

10

27.2PO

Time (seconds)

Location 12

260 280 300 320

Butding

-Subsurf act:

W-V*NA

10

0

0 20 40 60 100 120

30

180 20C 240 260 300 320 340 360

Time (seconds)

147

Location 13

WN-VAII-l

10

10

1'1

0 20 40 0 100 :0 160 180

220 2 230 300 320 340

27.2fP

Time (seconds)

Location 14

20

10

0

20

40

Eu 100

140 160 180 200 220 2

280 300 320 340

-suldincg

-SUbsuArfaf

Time (seconds)

148

Location 15

20

10

CL

U

0 20 40 so 100

EV14

140 160 200 220 260 280 300 320

-Budding

10 is

Tme (seconds)

Location 16

10

-5

0 20 40

10

80 100

140 160 200 220

260 280 300 320

Time (seconds)

149

10

1

Location 17

20

V-v6wjvf%.d-

10

L~E1

0

0 20 40

80 100 14

160 1 l o0 220

280 300 320 140

-

BoIidfn

Time (seconds)

06j 4

150

B.2.2 BAS_0.5%

40

35

Location 1

25

20

10

5.8PL

10.1 P

0

0 20 40 60 80 100 120 140 160 180 200 220 240 260 280

Time (seconds)

300 320

Location 2

--

-Buildng

Subsurface

~YYVS~Av~~

EkI

10

5

V_*A

E~Fi

0

0

20 40 60 80 100 120 140 160 180

Time (seconds)

200 220 240 260 280 300 320

-

-Building

Subsurface

151

Location 3

20

15

10

-buldng,

-

S

OSU I m k

1 0.1

Pa

25

0 20 40 60 80 100 120 140 160 180 200

Time (seconds)

220 240 260 280 300 320 340

Location 4

20 p.8Po

I."

Sutldin

13

0

20 40 60 80 100 120 140 160 180

Tme (seconds)

200 220 240 260 280 300 320 340

152

Location 5

30

15

10 k

LO. i

-Building

-Subsurface

20

I5

10

0

20 40 60 80 100 120 140 160 180

Time (seconds)

200 220 240 260 280 300 320 340

Location 6

I~E1

-Budding

--

Subsurface

0 20 40 60

80 100 120 140 160 180

Time (seconds)

200 220 240 260 280 300 320 340

153

Location 7

20

10

L~I1

N**%V

SubstIdfm

25

0

0 20 40 60 80 100

120 140 160 180

Time

(seconds)

200 220 240 260 280 300 320

Location 8

15

10 vwY

43AM

4

1.|

0 20 40 60 80 100 120 140 160

Time (seconds)

180

200 220 240 260 280 300 320

St l)UfJc

154

Location 9

21)

S15

10

L~E1 Buiding

-Subsurf ce

10.1 Pa

13.8

Pa

10

14

12

10

8

20

0 20 40 60 80 100 120 140 160 180 200 220

Time (seconds)

240 260 280 300 320 340 360

Location 10

P

U.

El.iP

7 1

-

Building

Subsurfact

0

0 20 40 60 80 100 120 140 160

Time (seconds)

180 200 220 240 260 280 300 320

155

Location 11

20

10

0

20 40 60 s0 100 120 140 160 180

Time (seconds)

200 220 240 260 280 300 320 40

Location 12

-1uI)',w

5.8Pa

10

0

20 40 60 S0 100 120 140 160 180

Time (seconds)

200 220 240 260 2 0

100 120 340

156

-widl.

Location 13

VO^-A~

20

15

10

0 20 40 60 80 100 120 140 160 180

Time (seconds)

200 220 240 260 280 300 320

Location 14

20

'"yWVOM.

eL-E1

10

13.89

0 20 40

60 80 100 120 140 160 180 200 220 240 260

Time (seconds)

280 300 320 40

157

-

Budding

Subsurfacc

Location 15

20

MIV N

I.8 P1

L0.AIE

|Lk1

0

20 40 60 80 100 120 140 160 180

Time (seconds)

200 220 240 260 280 300 320

Location 16

-ulling

20

10

F13.8Pt

0 20 40 60

80 100 120 140 160 180

Time (seconds)

200 220 240 260 280 300 320

340

-Buildinp

-Subsurfac,

158

lb

~AAANV

L414

16

Location 17

(A

I01P

0 20 40 60 80 100 120 140 160

Time (seconds)

180 200 220 240 260 280 300 320

159

B.2.3 SOG_0.08%

Location 1

20

10

10

0 20 40 60

20

0 120 160 180 220 240 280 300 320 340

___subsa

Time (seconds)

Location 2

10

CL

10

20 40 80 I0u 140 160 W 220 260

280 i00 320

-br_

20

Time

(seconds)

160

Location 3

30

10

10

0

0 20 40

20

80 100

140 160

LF,6i1

49.4

Fin

220 260 280 300 320 340

-Building

-Subsurface

40

Time (seconds)

Location 4

20

10

CL

0

0 20 40

10

80 100 40 160 18 220 24 280 300 320 340 360 ---

Building

Subsurface

20

30

49.4

Pa

Time (secor

161

Location 5

10 o 20

40

10

80 100 140 160 200 220 260 280

-

300

320 -ShM

Bulding

4l

20

Time (seconds)

Location 6

10

0

.

0

20 40

10 o 100 140 160 1,i

49.4PO

280 300 320 340

-

--u1Idnp

SubsurO,

40

Time

(seconds)

162

20

30

Location 7

30

20

10

10

0

20

40

80 100 140 160

L4.6

P

10 220 260 280 300

Building

320 -Subsrfacc

Time (seconds)

Location 8

49.4 Pa

-20

30

20

10

-10

0 20 40 80 100 10 160

34.6

200 220 260 280 300 320 340 -

Building

Subsurfac

Time (seconds)

49.4P

163

Location 9

30

20

10

CL

10

0 20 40

20

80 100 140 160 200 220 260 280 300 320 340

Time (seconds)

Location 10

30

20

10

10

0 20 40

80 100 140 160

200 220 260 280 300 320 -

K

20

Time (seconds)

164

Location 11

A

..... 01011111

20

10

0

.

0 20 40

10

20

80 100 140 160 200 220 260 280 300 320

-Building

-subsurtatt

49.4

PA

Time (seconds)

Location 12

20

10

0

10

20 40

20

30

18.$Pa

100

W

40 160 200 220 i0 280 300 320

-Busding

-subsuracw

Time (seconds)

165

Location 13

10

20

0 0

20 40

10

20

80 100 140 160

34.-P1

200 220 260 280 300 320

Sukb1)ski fla

40

Time (seconds)

Location 14

20

10

10

0 20 40 80 100

20

S220 240

Time (seconds)

49.4 Pa

280

300 320 340

-Buildng

_

166

Location 15

30

20

10

10

0

0 20 40 a.s5Pa

10 100 140 1>0

34.6iP

200

22 jt,

260 280 00 300 320 340

Building

Time (seconds)

Location 16

20

10

CL

10

0 20 40

20

80 100 140 160

E4.-PE1

0 220 260 280

300 320

--

Building

-subsurfacte

49.4

PO

Time (seconds)

167

Location 17

10 a.

10

0 20 40

10

20

18. 5 PO

80 100 140 160 200 220 260 280 100 320

.

Time (seconds)

168

B.2.4 SOG_0.5%

25

20 is

10

.S Pa

Location 1

~AM~

-Bualding

30

0

20 40 60 80 100 120 140 160 180

Time Iseconds)

200 220 240 260 230 300 320 340

Location 2

20

10

L~i

E~i

0

20 40 60 s0 100 120 140 160 180

Time (seconds)

200 220 240 260 280 300 320

-

-Blding

Subsurface

169

Location 3 rVW~

E2KI

L~~4~Wi

14.8 Pa

10

0 20 40 60 80 100 120 140 160

Time (seconds)

180 200 220 240 260 280 300

320

Location 4

-Suburfcig

10

0 20 40 60 s0 100 120 140 160

Time

(seconds)

180 200 220 240 260 280 300 120

170

-

-Buddinfir

SUturfAOcV

20

1IL

10

Location 5

20

XWAr~

15

10 i.5 i1

14. O I

L~i

30

0

20 40 60 80 100 120 140 160 180

Time (seconds)

200 220 240 260 280 300 320

Location 6

-

-Buiding

Subsurface

Al"E

--

---

Buildng

Subsurfice

0

20 40 60 80 100 120 140 160

Time (seconds)

180 200

220 240 260 280 300 320

171

Location 7

10

10

--

8kidinlfg

-skibsudiace

0 20 40 60 80 100 120 140 160 180

Time (seconds)

200 220 240 260 280 300 320

Location 8

20

15

10

0 20 40 60 80 100 120 140 160

180 200

Time (seconds)

220 240 260 280

300 120

---

B0,m

172

30

Location 9

1

Nr*

20

-

-Building

SubSurdfcC

10

L~II

20

15

10

30

0 20 40 60 80 100 120 140 160

Time (seconds)

180 200 220 240 260 280 300 320

Location 10

25

L~E1

-

Budding

Subsurface

0

20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320

Time (seconds)

173

Location 11

30

ANOA&dnAP-

4110

LF~I1

10

10

0

20 40 60 80 100 120

140 160 180

Time Iseconds)

200 220 240 260 280 300 320

Location 12

~MA'

L~I1

^AW

0 20 40 60 80 100 120 140 161

Time (seconds)

180 200 220 240 260 280 300

320

174

Location 13

20

L.AAPN

24.EPa

10

30

V".P^

0

20 40 60 80 100 120 140 160 180

Time (seconds)

200 220 240 260 280 300 320

Location 14 subuding

-subsurface

20

12~I1

10

20.6 Pa

0

0 20 40 60 80 100 120 140 160

Time (seconds)

180 200 220 240 260 280 300 320

Busdng

-Subsurface

175

Location 15

10

10

EI~i~ kdA.VIPN

C 20 40 60

80 100 120 140 160 180

200 220 240 260 280 300 320 340 ime (seconds)

Location 16

--

Bwldmng

20

L~E1

10

W4ArPr

U 20 40

60

80 100 120 140 160

Time (seconds)

180 200 220 240 2601 2s0 100 120

176

Location 17

20

I5

10

U

20 40 60

80 100 120 140 160 180 200 220 240 260 280 300 320

Time

{seconds)

-

-Bulding

Subsurface

177

Appendix C

-

SEEP/W Isopotential Contour Diagrams

178

C.1 Wind Experiment

C.1.1 BAS

C.1.1.1 10 ft/s

C.1.1.2 15 ft/s

179

C.1.1.3 20 ft/s

180

C.1.2 BASSW

C.1.2.1 10 ft/s

C.1.2.2 15 ft/s

181

C.1.2.3 20 ft/s

182

C.1.3 SOG

C.1.3.1 10 ft/s

I I

C.1.3.2 15 ft/s

183

C.1.3.3 20 ft/s

184

C.1.4 SOGSW

C.1.4.1 10 ft/s

I

I

C.1.4.2 15 ft/s

I

I

185

C.1.4.3 20 ft/s

L....

186

C.2 Depressurization Experiment

C.2.1 BAS_0.5%

C.2.1.1 17.4 cm 3 /s

C.2.1.2 27.9 cm 3 /s

187

C.2.1.3 37.3 cm 3 /s

188

C.2.2 BAS_0.08%

C.2.2.1 17.4 cm 3 /s

C.2.2.2 27.9 cm 3 /s

189

C.2.2.3 37.3 cm 3 /s

190

C.2.3 SOG_0.5%

C.2.3.1 17.4 cm

3

/s

I

C.2.3.2 27.9 cm 3 /s

191

C.2.3.3 37.3 cm 3 /s

192

C.2.4 SOG_0.08%

C.2.4.1 17.4 cm 3 /s

I I

C.2.4.2 27.9 cm 3 /s

193

C.2.4.3 37.3 cm 3 /s

194

Appendix D

-

Photo Log

195

Photo #1 of

16

Description:

RO-TAP mechanical sieve machine used in grain size analysis

Photo #2 of 16

Description: Porosity laboratory test. 100 mL of 100% saturated soil on the left, the remaining

71 mL of the initial 100 mL of water on the right

196

Photo #3 of 16

Description: Constant head hydraulic conductivity laboratory setup

Photo #4 of 16

Description: Data acquisition system. The blue box is the data acquisition module. The module is placed on top of the power source to the module. The box to the right of the computer monitor is a separate device to display the current input voltage

197

Photo #5 of 16

Description:

Screenshot of Acqlipse

Data Acquisition software interface for data acquisition system

Photo #6 of 16

Description:

Differential Pressure

Transducer calibration apparatus

Description:

Depressurization

Experiment Peristaltic pump setup

198

Photo #7 of 16

Description:

Depressurization

Experiment Peristaltic pump setup

Photo #8 of 16

Description: Wind

Experiment Setup including Dayton propeller fan and data acquisition system

199

Photo #9 of 16

Description: Wind

Experiment Basement configuration

Photo #10 of 16

Description: Wind

Experiment Slab-On-

Grade configuration

200

Photo #11 of 16

Description: Wind

Experiment Slab-On-

Grade configuration with sidewalk

Photo #12 of 16

Description: Wind

Experiment Basement configuration with sidewalk

201

Photo #13 of 16

Description:

Depressurization

Experiment 0.5% foundation crack configuration

Photo #14 of 16

Description:

Depressurization

Experiment 0.07% foundation crack configuration

202

Photo #15 of 16

Description:

Depressurization

Experiment Basement configuration

Photo #16 of 16

Description:

Depressurization

Experiment Slab-On-

Grade configuration

203

Appendix E

-

SEEP/W Data Reports

Available on file

Appendix F

-

Raw Data Files

Available on file

Appendix G Differential Pressure Transducer Manufacturer

Specifications

Available on file

204