November
www.epa.gov/research
United States
Environmental Protection
Agency
Application of TEVA-SPOT for
Prioritizing Security Enhancements
at Utility Facilities and for
Protection of Critical Facilities
J j
Office of Research and Development
National Homeland Security Research Center
-------
November 2014
EPA/600/R-14/433
APPLICATION OF TEVA-SPOT FOR PRIORITIZING SECURITY
ENHANCEMENTS AT UTILITY FACILITIES AND FOR PROTECTION OF
CRITICAL FACILITIES
Page 1 of 93
-------
Disclaimer:
The U.S. Environmental Protection Agency through its Office of Research and Development (funded and
managed) or (partially funded and collaborated in) the research described here under (contract number)
or (assistance agreement number) to (contracting company name). It has been subjected to the
Agency's review and has been approved for publication. Note that approval does not signify that the
contents necessarily reflect the views of the Agency. Mention of trade names, products, or services does
not convey official EPA approval, endorsement, or recommendation.
Page 2 of 93
-------
Definitions:
Collection - Describes a group of applications or ensembles within TEVA-SPOT-GUI (graphical user
interface version of TEVA-SPOT).
Contamination Scenario - Describes a contaminant's injection or release at a location in the network
(node ID), the contaminant's transport and fate in the network, and the exposure of individuals by the
ingestion of tap water obtained from the network.
C-Values - Consequence Values. Consequences can be estimated, for example, as public health
consequences or extent of contamination, such as number of pipe feet contaminated.
Dose Level - A measure of public health consequences defined as the size of the population receiving an
ingestion dose above some level.
Ensemble - Defines a group of unique contamination scenarios that are to be simulated within TEVA-
SPOT-GUI. Also, describes a user defined application of TEVA-SPOT-GUI.
EPANET-MSX - EPANET's water quality component is limited to tracking the transport and fate of a
single chemical species, such as chloride used in a tracer study or free chlorine used in a disinfectant
decay study. EPANET-MSX provides an extension to EPANET that allows the user to model any system of
multiple, interacting chemical species. This capability has been incorporated into both a stand-alone
executable program as well as a toolkit library of functions that programmers can use to build custom
applications. This set of software tools is referred to as EPANET-MSX, where MSX stands for Multi-
Species Extension.
GRASP heuristic - GRASP heuristic algorithm performs sensor placement optimization without explicitly
creating a MIP formulation. GRASP uses much less memory than PICO and usually runs much faster in
comparison to PICO. Although the GRASP does not guarantee that a globally optimal solution is found, it
has proven effective at finding optimal solutions to a variety of large-scale applications.
Injection Definition - Describes the contaminant release or injection that will be simulated in TEVA-
SPOT-GUI, including contaminant name, contaminant's mass injection rate (mg/min), and the start and
stop times for the contaminant release or injection.
Lagrangian - Heuristic algorithm that uses the structure of the p-median MIP formulation to find near-
optimal solutions while computing a lower bound on the best possible solution. Uses less RAM memory
than GRASP.
Node Set Definition - Defines a set of network locations (node IDs) where the contaminant will be
release or injected.
Node Injections - Represents the addition of a contaminant's "Injection Definition" with a "Node Set
Definition" to create a set of "Node Injections".
Population Dosed - Number of people who receive a dose greater than a given dose level.
Sensor Placement Optimization Statistic - The statistical measure used to optimize the placement of
sensor monitoring stations given the distribution of consequences that are defined by the ensemble of
contamination scenarios. The mean statistic is the expected consequence given a set of contamination
scenarios. Worst case is the maximum consequence obtained with a design for some specified
conditions.
Scenario Set - Represents one or more "Node Injections" to create a folder of contamination
simulations. One or more scenario sets specify a unique set of contamination scenarios.
TSO-2-lmpact - Precursor file for sensor placement optimization.
Page 3 of 93
-------
Acronyms:
ATUS - American Time Use Survey
CWS - Contamination Warning System
DL- Dose Level
EC - Extent of contamination (feet of pipe)
Erd - EPANET results database
GPD - Gallons per person per day.
GRASP - Greedy randomized adaptive search procedure
GUI - Graphical User Interface
HIA- Health Impacts Analysis
ID - Infectious dose
MA - Infrastructure Impacts Analysis
Kg- Kilograms
LD- Lethal dose
MIP - Mixed integer program
NHSRC- National Homeland Security Research Center
NZD - Non-zero demand nodes
PICO - Parallel Integer and Combinatorial Optimizer
pd - Population dosed
pk- Population killed
RAM - Random Access Memory
SP - Sensor Placement
SRS - Surveillance and Response System
TEVA-SPOT - Threat Ensemble Vulnerability Assessment, Sensor Placement Optimization Tool
Tso - Threat simulator output. In an earlier version of TEVA-SPOT, the acronym "tso" was used to
describe the output from EPANET, i.e., contaminant concentration values at each receptor node and
time step given a contaminant injection at an injection node. The tso formalism has since been replaced
by "ERD" (EPANET Results Database).
WQ-Water quality
Page 4 of 93
-------
1.0 INTRODUCTION.
The purpose of this manual is to describe two applications of the Threat Ensemble Vulnerability
Assessment, Sensor Placement Optimization Tool (TEVA-SPOT) software program to the design of a
Surveillance and Response System (SRS). The first to prioritize security enhancements to utility facilities
located in the distribution system (e.g., tanks and pump stations). The second is to design a water
quality monitoring sensor network to monitor the water serving critical facilities (e.g., hospitals or
government buildings). This manual supports implementation of the Enhanced Security Monitoring and
Online Water Quality Monitoring components of an SRS under the Water Security Initiative, a program
developed by U.S. Environmental Protection Agency's Water Security Division.
TEVA-SPOT is a software tool developed by the U.S. EPA, National Homeland Security Research Center
(NHSRC) to evaluate contaminant threats to drinking water distribution systems. Contaminant threats
are posed by both intentional and unintentional releases. The TEVA-SPOT software tool quantifies
consequences from an array of contaminant-release scenarios. The emergency planner's ability to
prioritize enhancements to distribution system monitoring & surveillance and response capabilities
through deployment of a water quality SRS, hinges on effective consequence analysis. The
quantification of consequences is needed to support the design of an effective SRS. One key component
of SRS is the sensor monitoring network, which enables timely detection of water quality anomalies
indicative of a contaminant release. Quantified consequences of contaminant releases - of various
contaminants, in various amounts and from various locations - are required for the design and
evaluation of a robust sensor monitoring network and for prioritization of security enhancements at
utility facilities located in the distribution system. The TEVA-SPOT software tool fulfills pivotal roles in
the development of an effective SRS for the protection of drinking water distribution systems.
Public health consequences can be estimated in terms of infections and fatalities, or more generically as
simply exposures or doses. Infrastructure consequences of a particular contamination injection event
can be estimated based on the length of pipe (pipe feet) that witnesses (or exhibits) contamination
above some level (specified as a contaminant water concentration in units of milligrams per liter).
The prioritization of security enhancements at utility facilities is implemented in TEVA-SPOT using the
consequence modules to calculate Utility Facility-based consequence values (C-Values). Utility facilities
include, for example, tanks, pump stations, and valves. A C-Value for a particular utility-owned facility is
defined by the consequences that could result when contamination is introduced at such a facility. A
group of utility facilities are ranked based on the magnitude of consequences that result when
contamination is introduced at the facility. Once C-Values are determined, users (e.g., utility managers
or engineers) can use the information to prioritize their SRS efforts for increased surveillance and facility
hardening.
Another application of TEVA-SPOT during development of an SRS is to identify optimal locations for
placing water quality monitoring sensors to detect contamination incidents. The design of a water
quality monitoring sensor network to monitor the water serving critical facilities is implemented in
TEVA-SPOT using the sensor placement optimization tool (SPOT) components. In order to avoid
Page 5 of 93
-------
confusion with the C-Values analysis demonstrated on utility facilities, we restrict our attention here to
critical facilities that are NOT utility facilities. For example, a user may determine that their critical
facilities include hospitals, police and fire stations, and government buildings. A user can identify a set
of locations as critical facilities and TEVA-SPOT will determine the set of upstream network locations
that supply water to and, therefore, could potentially contaminate a critical facility. Using this approach
critical infrastructure, other than or in addition to utility infrastructure, can be identified and TEVA-SPOT
can be used to help protect them. By determining the upstream locations that could affect the critical
facilities, TEVA-SPOT allows the user to evaluate the consequences from each of the identified upstream
locations (evaluated as possible contamination injection locations) to prioritize SRS efforts at those
upstream locations that present the greatest risk to identified critical facilities. This feature can also be
used to help determine the best locations to place monitoring stations to protect the critical facilities.
This manual begins with a brief background discussion of TEVA-SPOT and TEVA-SPOT-GUI. In Section 2,
we provide installation instructions for TEVA-SPOT-GUI and some references to supporting journal
articles and reports that can be examined for additional information if needed. We describe the TEVA
analysis approach for determining Utility Facility C-Values in Section 3 through a step-by-step illustration
process using an example network and example biotoxin. In Section 4 we describe how to use TEVA-
SPOT-GUI to help protect user-identified critical facilities and how to use TEVA-SPOT-GUI to optimally
locate sensor monitoring devices. In Section 5, we provide tips on using and applying the TEVA analysis
approach to determining C-Values and protecting critical facilities. Section 6 provides supporting
references.
2.0 INSTALLATION AND BACKGROUND.
This manual is not meant to replace the TEVA-SPOT-GUI Users Manual. The TEVA-SPOT-GUI Users
Manual is available at:
http://cfpub.epa.gov/si/si public record report.cfm?address=nhsrc/&dirEntrylD=202703.
TEVA-SPOT uses EPANET 2.00.12 as its hydraulic and water quality modeling engine [1]. Two versions of
TEVA-SPOT are available for download. TEVA-SPOT Toolkit is a command-line based set of software
tools designed for software developers or researchers. The TEVA-SPOT graphical user interface (GUI)
version provides an easy to use graphical user interface for inputting the necessary parameters and for
outputting and viewing results. This manual is based on using the TEVA-SPOT-GUI version. TEVA-SPOT-
GUI and TEVA-SPOT Tool kit are available for download at: http://www.epa.gov/nhsrc/index.html [2].
Users should contact Robert Janke at ianke.robert@epa.gov to ensure that they have the latest update
to TEVA-SPOT-GUI prior to using the web version of TEVA-SPOT-GUI.
TEVA-SPOT incorporates the multi-species extension to EPANET (EPANET-MSX), which provides the
ability to simulate and model more sophisticated contaminant interactions and the influence of those
interactions on estimating public health and infrastructure damages. The use of the EPANET-MSX
capabilities requires the specification of the necessary kinetic reactions that describe the interactions of
the contaminant in the water distribution system. EPANET-MSX is available from the same link as TEVA-
SPOT: http://www.epa.gov/nhsrc/index.html, see "EPANET Extensions" [3].
Page 6 of 93
-------
The distribution executable (.exe file) contains the TEVA-SPOT-GUI program and Users Manual. The
prerequisite software can be obtained from the FTP site below or from the internet by searching for the
required items below. Prerequisite software includes the Java Development Kit (JDK) (e.g., Version 1.6
update 20) and the Python Scripting Language (version 2.6 or later). The JDK and Python products need
to be installed first prior to the installation of TEVA-SPOT-GUI. The JDK and Python versions installed
need to be consistent with the computer's capabilities. In other words, use the 64 bit versions of JAVA
and Python if hardware and software on the host computer is 64 bit compatible. TEVA-SPOT-GUI has
been tested on Windows XP and Windows 7, not Windows Vista. Below are the JDK and Python
installers needed depending on the computer hardware and operating system that TEVA-SPOT-GUI will
be installed on. The latest versions should be installed, these are for example only.
• jdk-6u41-windows-586.exe (32 bit version)
• jdk-6u41-windows-x64.exe (64 bit version)
• python-2.7.3.msi (32 bit version)
• python-2.7.3amd64.msi (64 bit version)
After the prerequisites of JAVA and Python are installed, the TEVA-SPOT-GUI software is installed and
configured to the c:\Program Files\TEVA-SPOT folder. Output files generated by TEVA-SPOT-GUI are
stored in c:\TEVA-SPOT-Database. During installation the drive location can be changed if desired, e.g.,
D:\Program Files\TEVA-SPOT.
Updates to TEVA-SPOT-GUI can be installed by downloading the most recent update from the EPA TEVA-
SPOT Research Program website (http://www.epa.gov/nhsrc/index.html). TEVA-SPOT-GUI can also be
obtained from the following File Transfer Protocol (ftp) website: ftp://scienceftp.epa.gov/ and entering
the appropriate login and password information which can be obtained via email from
ianke.robert@epa.gov.
Each TEVA-SPOT-GUI installation (.exe) file is tagged with a date of release, e.g., TEVA-SPOTInstaller-
2.3.1b-20130723.exe, signifying a release date of July 23, 2013 (last eight digits in the file name). The
initial two numbers in the file name describe the version number, i.e., 2.3, refers to the major release
number (2) and the minor release number (3) and are consistent with the version numbering for the
TEVA-SPOT toolkit. The final number in the file name that describes the version number, i.e., 1,
designates a sub-minor release which is specific to TEVA-SPOT-GUI.
Older versions of TEVA-SPOT-GUI should be removed first using the following procedure before
installing the updated version:
• Choose STOP TEVA-SPOT Services from START Programs/TEVA-SPOT.
• If you have previously setup a contaminants file, move the "contaminants.xml file" from the
"client" folder under Program Files/TEVA-SPOT/client to suitable location. During installation of
a TEVA-SPOT update the contaminants.xml file is replaced with a blank version. After
installation replace the blank contaminants.xml file with the saved contaminants.xml file.
• Choose Uninstall TEVA-SPOT from START Programs/TEVA-SPOT.
• Navigate to Programs/TEVA-SPOT and Remove the TEVA-SPOT folder and its contents.
• Install the TEVA-SPOT update by double-clicking on the TEVA-SPOT executable icon.
Page 7 of 93
-------
Updates are installed without loss of older ensemble data residing in the TEVA-SPOT-Database folder.
2.1 Supporting Documents:
We provide a listing of some relevant journal papers and reports that provide supplemental information
supporting the underlying models within TEVA-SPOT and the application of TEVA-SPOT to water
distribution systems.
• "Model for estimating acute health impacts from consumption of contaminated drinking water"
provides the underlying methodology for assessing public health effects of contaminants in
potable distribution systems [4].
• "Importance of exposure model in estimating impacts when a water distribution system is
contaminated" provides the motivation for investigating and developing a realistic timing model
to describe when people are most likely going to drink tap water each day [5].
• "Development of a probabilistic timing model for the ingestion of tap water" develops the
methodology for the American Time Use Survey (ATUS) based probabilistic timing model for the
likely timing of ingestion of tap water by water system users [6].
• "Patterns in potential impacts associated with contamination events in water distribution
systems" provides a case study analysis of potential public health consequences for 12 real and
diverse distribution systems [7].
• "Assessing Potential Impacts Associated with Contamination Events in Water Distribution
Systems: A Sensitivity Analysis" extends the analysis provided in [7] to provide a better
understanding of the nature of adverse public health consequences that could be associated
with contamination events [8].
• "A Framework for Estimating the Adverse Health Effects of Contamination Events in Water
Distribution Systems and Its Application" develops and describes a flexible analysis framework
for estimating the magnitude of public health consequences given detailed information or very
limited information [9]. This paper builds on the work in [7] and [8].
• "Robustness of Designs for Drinking-Water Contamination Warning Systems under Uncertain
Conditions" examines the performance of contamination warning system based sensor network
designs [10]. Specifically, the manuscript analyzes performance in the context of changed
conditions, i.e., conditions different from those used in the design. The analysis helps users of
the TEVA-SPOT software to develop sensor network designs that are more robust while reducing
computational requirements.
• "Influence of Network Model Detail on Estimated Health Effects of Drinking Water
Contamination Events" examines the influence of distribution system model detail on estimating
public health effects from drinking water contamination events [11]. This paper examines the
loss of model (distribution system) detail on estimating public health consequences and helps to
guide the TEVA-SPOT user how to minimize such adverse effects by restricting attention to the
high percentile contamination scenarios.
Page 8 of 93
-------
"Modeling reaction and transport of multiple species in water distribution systems" provides the
methodological basis for EPANET-MSX along with examples for the user [12].
"Sensor Network Design for Drinking Water Contamination Warning Systems: A Compendium
of Research Results and Case Studies Using TEVA-SPOT Software/' provides supporting
documentation to assist in the design of contamination warning systems [13].
"Securing Water and Wastewater Systems, Chapter 2: Protecting Water Supply Critical
Infrastructure: An Overview/' provides a literature review of threats facing water systems with
a focus on contamination threats. The book chapter discusses general principles and
characteristics of water and waste water system security and summarizes current research as it
relates to improving the security of water systems from intentional contamination threats [14].
3.0 TEVA ANALYSIS APPROACH FOR DETERMINING C-VALUES.
We describe a step-by-step method to determine C-Values for an example biotoxin and a real, but
anonymous, water distribution system model using the TEVA-SPOT-GUI software program. C-Values
can be used to help a water utility prioritize their resources and their efforts at hardening their facilities
or at installing surveillance and monitoring equipment. Using this approach, a group of utility facilities
can be identified and analyzed for consequences that could result if contamination were introduced
independently at each of the utility facilities. For the purpose of this manual, we define utility facilities
separately from critical facilities. With utility facilities the user is evaluating consequences from the
introduction of a contaminant at a particular water utility facility. With critical facilities, the user is
evaluating those upstream contaminant injection nodes that produce consequences at a particular
critical facility.
To determine the scope of the C-Values analysis, the user must decide the scope of the utility-based
consequence assessment that they want to investigate. Specifically, the user decides whether they
want to examine the consequences of contamination being introduced at some group of utility facilities,
e.g., all tanks, pump stations, or valves, or some subset of utility facilities, in order to determine those
utility facility locations that will result in the greatest consequences if contamination were introduced.
First, the user imports their EPANET-based model (*.inp file) [1]. The EPANET model should be capable
of running extended period simulation. The user will also input population information and define how
people are served by the water system, as described by the network model. This is typically the
population served by the water system, but could represent some fraction of the population served.
Next, the user decides how they want to examine consequences. TEVA-SPOT-GUI provides the user with
the capability to examine consequences in terms of public health measures (e.g., infections, fatalities,
exposures or doses) and/or as feet of contaminated pipe.
Next, we describe how to open TEVA-SPOT-GUI, setup of a "Collection", naming an "Ensemble",
importing an EPANET ".inp" file, and setting up the population distribution. Next, we describe and
demonstrate the use of five modules within TEVA-SPOT-GUI to determine C-Values or evaluate critical
facilities: (1) Initialization: Population, (2) UpDownStreamNodes, (3) EPANET Simulations, (4) Health
Page 9 of 93
-------
Impacts Analysis, and (5) Infrastructure Impacts Analysis. Table 3.0.1 provides an overview of the steps
for determining C-Values.
Table 3.0.1. Overview of steps for determining C-Values
Step
Steps 3. 1.1 -Step 3. 1.10
Step 3. 2.1 -Step 3. 2.5
Steps 3.3.1 -Step 3.3. 2
Steps 3.4.1 -Step 3.4.2
Steps 3.5.1 -Step 3.5. 11
Feature
Opening and Input Generation
Scenarios
• EPANET Simulations
Health Impacts Analysis (HIA)
Infrastructure Impacts Analysis
(MA)
Obtaining results
• Map tab
• Table tab
Task
Opening TEVA-SPOT-GUI and setting up
and running "Initialization: Population"
and "Upstream & Downstream Nodes".
Setting up and running EPANET
simulations.
Setting up and running HIA analyses.
Setting up and running IAA analyses.
Retrieving results and determining C-
Values.
Utility facilities are selected by the user as a set of contamination scenarios and simulated in the
"EPANET Simulations" module. Consequences are determined using the "Health Impacts Analysis"
module or the "Infrastructure Impacts Analysis" module or both modules.
To demonstrate the use of TEVA-SPOT-GUI to examine public health and infrastructure consequences
and determine C-Values, we use an example biotoxin. For our illustrative example, we use a real, but
anonymous, distribution system model. First, we demonstrate how to create an "Ensemble" to
determine C-Values for Utility Facilities (UF). For demonstration purposes only, we use the example
network's group of tanks as the UF contamination scenarios. After the creation of the ensemble,
"Network_4_Utility_Facilities", we demonstrate how to determine consequences.
3.1 Collections, Ensembles, Population: Initiation, and Upstream and Downstream Nodes:
A TEVA-SPOT-GUI analysis application begins with setting up a "Collection" and creating an "Ensemble".
A "Collection" is a folder that contains one or more "Ensembles". An "Ensemble" represents a user-
defined application of TEVA-SPOT-GUI. A contamination scenario is described by the injection or release
of a contaminant at a location in the network (node ID), the contaminant's transport and fate in the
network, and the exposure of individuals by the ingestion of tap water obtained from the network. An
"Ensemble" defines a unique set of contamination scenarios that are simulated in TEVA-SPOT-GUI. For
example, an "Ensemble" of Utility Facilities" could describe a contaminant injection at all the nodes in
the distribution system model representing valves, tanks, or pumps or the "Ensemble" could describe a
contaminant injection at all possible locations in the model (all nodes). Figure 3.1.1 provides a
schematic illustrating the TEVA-SPOT-GUI process to define a contamination event, identify
contamination injection locations, and define a unique set of contamination scenarios. The figure also
describes relationship between an "Ensemble" and a "Collection".
Page 10 of 93
-------
Figure 3.1.1. TEVA-SPOT-GUI process for defining contamination scenarios
ENSEMBLE: A named instantiation of a TEVA-SPOT-GUI application
Use of Multiple Folders:
Within each folder is a single set of
"node Injections".
* A set of unique contamination
scenarios is defined by the addition of
each folder's "Node Injections" to
creates unique list.
» Each "Node Injection" or
contamination scenario is an
independent contamination event.
Use of Single Folders:
Combining multiple "node Injections"
into a "Scenario Set" folder creates a
set of multiple, simultaneous
» The number of such simultaneous
contamination scenarios is trie
product of the scenarios defined.
COLLECTION:
A group of user-
defined ensembles.
After a "Collection" and an "Ensemble" are created (steps 3.1.1-3.1.4), the next step is to import a
distribution system model into the "Ensemble" (step 3.1.5). The distribution system model must be an
EPANET model (EPANET formatted text file with the .inp extension) that is capable of running extended
period simulations. Please refer to TEVA-SPOT-GUI User's Manual, Appendix A, Water Utility
Requirements for Using EPA's TEVA-SPOT for a discussion of the requirements and steps that must be
followed to use TEVA-SPOT appropriately [2]. Our model is named Network 4. Network 4 is a real, but
anonymous, distribution system model. This model has been used elsewhere; see [7], [8], [9], and [10].
In step 3.1.6 the user clicks the "Edit" menu item and then chooses "Ensemble Options". The "Ensemble
Options" box enables the user to specify various simulation parameters. More discussion of these
parameters is available in Section 5.
After the network model is imported and the simulation parameters are reviewed, the network
population is initialized (step 3.1.7). TEVA-SPOT allows the user to specify the population distribution
using either (1) a demand based method or (2) a census based method. The demand-based method
requires a user to input of the number of gallons of water used per person per day (GPD). A default
value often used is 200 GPD, which corresponds to the quantity of water used per person per day for all
uses. For more information on drinking water consumption rates, the user should seek other sources
for the most up-to-date and relevant information. Alternatively, the user could import a text file that
details the number of people who reside at each non-zero demand node (node ID) in the model (i.e., a
text file containing two columns of data). Non-zero demand nodes are used because exposure and
ingestion of contaminated water can only occur at those locations where water is used.
We assume that the population distribution (i.e., the number of people residing at each non-zero
demand node) is based on demand or water usage and those 200 gallons per person per day is usage
Page 11 of 93
-------
rate. Given this per capita water usage rate, the total population served by the water distribution
system model is approximately 152,000 people.
Step 3.1.8 describes the use of the "UpDownStreamNodes" module. This module determines the
upstream and downstream nodes corresponding to a set of contamination scenarios or injection
locations. The number of upstream or downstream nodes associated with a particular injection location
(node) is a function of the time step of the simulation. The default simulation duration for this module
is the simulation length specified in the "Ensemble Options" box. Most uses of this module would use
the default simulation length.
Steps 3.1.9 and 3.1.10 demonstrate the execution and completion of the "Input Generation" modules.
Opening and Input Generation
STEP 3.1.1.
Open TEVA-SPOT-GUI by choosing START Menu and then clicking on the TEVA-SPOT folder and then TEVA-SPOT.
_cj TEVA-SPOT (No Ensemble Loaded)
Ensemble File Map Mode Edit About
pll Charts | Tables,
Q, X (^ |D Google Map^ Road Map -»-]| Map [Default ^] Filtering | Manage ] |NO Filtering
No ensemble loaded
Initial screen - indicates "No Ensemble Loaded".
For more explanation, see TEVA-SPOT-GUI User's Manual.
Page 12 of 93
-------
Opening and Input Generation
Step 3.1.2.
Under Ensemble Menu, select "Collection Management".
*J TEVA-SPOT (No Ensemble Loaded)
a
Ensemble File Map Mode Edit About
Map | Chans | Tables
fit X 4 |U Google Maps Road Map • | Map Default
» Filtering | Manage 11 No Filtering
l±J Collection Management !• ^ '1
Available Ensemble Collections
TEVA-SPOT-Tests [Main]
.
Ensembles
Regret Analyses
Total Size
| New
N/A
N/A
N/A
] Rename | j Move Delete
| Close |
:l
No ensemble loaded
Collections are defined by the user and represent a set of ensembles.
Collections and their associated Ensembles are located in the folder: TEVA-SPOT-Database.
Page 13 of 93
-------
Opening and Input Generation
Step 3.1.3
Click "New" and Name your "collection". For instance, name your collection "C-Values". Click "OK".
Click "Close" to close the "Collection Management" box.
iiii £Vi]Urtian ktMUtW
Avrt«dbkD.wtrl;U?i:-yl«
j*pnl_2Q3Q_rosts [Man]
ApFll_2Qll (Main]
r
[Mumbles
Regret Ana^»6
Total Siz«
•Hri
BGMfl
WA
N/A
| Nan j | S»nwni | | MM* '
nan
.^ New Ensemble CoHectun
FnwrrWeColiictiMUftnitior Man »
Ennnttia Colacbor Hnmo C-Vdui^
2S '"
'*-
C3
*
This is the "Collection" folder where "Ensembles" of contamination scenarios are located.
Step 3.1.4.
Under "Ensemble" menu, click "New".
Name the "Ensemble" as "Network_4_Utility_Facilities". Click "OK".
Ensemble File Map Mode Edit About
Map Charts Tables
|D Google Maps|Road Map ->|| Map [Default
,^, New Ensemble
Ensemble Collection c-Values CMainl
Ensemble Name NetVs
-------
Opening and Input Generation
Step 3.1.5.
Under "Ensemble Menu", select "Import EPANET .inp File". The dialogue box will appear and select or browse to the
location for your model. Click "Open".
sj TEVA-SPOT (Network_4_Utility_Facilities in C-Values [Main!)
Ensemble File Map
Map | charts Tables
g
Mode Edit About
X tSi | Google Maps) Road Map * Map Default
l&JOpen
Look in:
,
£.
•£•
Recent
Items
Desktop
My
Documents
dBj
*»
Computer
•
Network
2S
^^
1. TEVA-SPOT T, $ °@H
AnalysisServer
I Client
config
l DBServer
ExecutionController
1 RMIRegistry
SensorAnalysis
Server
ServerBroker
; tevaLogs
J Network_4.inp
File name: Network_4.inp Open
Files of type: EPANET input file (".inp) ^^^^^_ " Cancel
If there are errors within the .inp file: TEVA-SPOT-GUI may indicate problems with loading the inp file and the file will not be
loaded.
TEVA-SPOT-GUI will indicate where the problems are. The user should open their model using a text editor (e.g., Microsoft®
Notepad) and try to fix the problem indicated.
Note for the discussion here, the "Network_4.inp" file was first copied to the C:\Program Files\TEVA-SPOT\Client directory.
This is the directory that TEVA-SPOT-GUI first opens to.
Page 15 of 93
-------
Opening and Input Generation
Step 3.1.6.
Under the "EDIT" Menu, select "Ensemble Options". Use the default parameters shown below.
Type the text shown below into the description box. Also, type "Example Biotoxin" into the "Decay Name" box. Click "OK".
Under the "Ensemble" menu, select "Execution Control" to display the "Execution" dialogue box.
* Dftering ' Manage No ntt«rtng
,i| Fnwmble Options
iVMiiplwn
Tm*' Pa
fluirfioi
Hyflrwk l irrwswp 1
IHJW-J
mtmitM
Juun
-,0. s
t ant MMn0 nwdet; Us* swnano
i Override if set
HydMife Pw4in*itiir,
Maximum I ruls 4V
tHfcuu-Ktq |a
MAKOtffK 1 fl
tr-riF.iM'T {i.o
Unbolfwce' Para m««rs
««f "»*"f««pOWMn
Bblk * "
React-on Coefficients
Bufic 0.0
Wall O.D
Unntu*g LoncentratKin O.U
Wail Co*ff. Corretaban o.u
QA P««m-[iT'.
bailable
;ipu; -S<
9 ^gaSBH&iS^E
Q / Upstream & Uav/nsreain Modes Wit
Pwform Mow Boiaraee Ans^sts
IftjecOon Location cxdusiws{0) ^ S«ltct |
ScwunoB
Q F?A»JFT ^at.iHjIniri'.
htadutrr.
O "h impacts AlM*y«S [ Edit"']
O ! 'T50 to kipacts Ajialyss
Q !nl i.-. . '
denser fiacement
For more information, see TEVA-SPOT-GUI User's Manual.
Page 16 of 93
-------
Opening and Input Generation
Step 3.1.7.
Click EDIT box for the "Initialization: Population" module. In the "Edit Population Parameters" dialogue box, for the
"Population Model" parameter, choose "Demand-based" from the drop down menu. Within the "Per Capita Usage (GPD)"
box, type "200" and then click the "Calc" button. Finally, Click "OK".
Execution
Available Executions —
Input Generation
0 OInitialization: Population | Edit
0 Li Upstream & Downstream Nodes Edit
Status
Not Running
0 n EPANET Simulations
O LI Health Impacts Analysis
0 QTSO to Impacts Analysis
0 L] Infrastructure Impacts Analysis
0 I_J Sensor Placement
Edit |
Edit |
Edit |
Edit |
Edit |
Estimated Time Remaining
Estimated Total Time
Estimated Completion Time
Execute Terminate
l^j Edit Population Parameters
Population Model
Estimated Population
Demand-based •»•
Demand-based population assignment is demonstrated. A "Census" based approach could also be used by choosing "Census
based" and then importing a text file that provides the number of people for each non-zero demand node.
Note: TEVA-SPOT-GUI will check the user's uploaded population file to determine if people are placed at nodes with zero
demand. TEVA-SPOT-GUI will notify the user in a message box as shown below:
LfeJ Edit Populati
Population Model
|
sn Para...
Census-based ^J
Import File
Cancel |
Message
WARNING: the following nodes have a population specified, but
have no demand and will therefore not be included in any impacts
No demand:
JUNCTION-0, JUNCTION-1, JUNCTION-ID, JUNCTION-100
OK
Page 17 of 93
-------
Opening and Input Generation
Step 3.1.8.
Click "EDIT" box for "UpDownStreamNodes" module. Notice that the box "Edit Upstream Node Calculation Parameters"
appears. Click cancel, this uses the default simulation length specified in the "EDIT/Ensemble Options" menu.
tlVA-iPOI (Ntlmt*
.y-
«| m
O : mwr
O H««»n»ai/u»lv)>i bji
4) ISO 10 InlpKll Alu»VJ4 I 6d»
| Ul ]|
The default duration period (hrs) is used to capture the full extent of the exposure period, which is one wk or 168 hrs.
Page 18 of 93
-------
Opening and Input Generation
Step 3.1.9.
Click the two boxes associated with: "Initialization: Population" & "UpDownStreamNodes". Check marks will appear, then
click "EXECUTE". Both modules should complete running, i.e., the RED Exclamation points will turn GREEN.
A K A
O Hl
.
,_,.
Page 19 of 93
-------
Opening and Input Generation
Step 3.1.10.
Observe the completion of the two modules. "Initialization: Population" should run quickly, i.e., take a few minutes.
"Upstream & Downstream Nodes" may take a while (from a few minutes to an hour or more) depending on the size of the
network model.
nsembic Frtc Map Mode Edit Aixjui
Now GREEN checkmarks appear indicating the successful completion of the Input Generation Modules.
3.2 Contamination Scenarios:
TEVA-SPOT-GUI simulates contaminant injection or release into a drinking water distribution system and
then performs the required hydraulic and water quality analysis using a modified version of EPANET to
predict contaminant concentrations (mg/L) throughout the distribution system. A contamination
scenario is defined by the contaminant, its injection specifications, and the locations where the
contaminant is injected in the system. More specifically, the contamination scenario is given by: (1)
contaminant mass release rate (mg/min), (2) contaminant release duration, given a start time and a stop
time in units of seconds, minutes or hours, and (3) specification of the threat ensemble (node IDs) to
represent contamination injection locations. For the purpose of determining C-Values, we define the
locations for contamination injection as user-specified Utility Facilities.
The contamination source for the hypothetical biotoxin is described by the arbitrary mixture of 300
grams of the example biotoxin into 7,500 liters of water to result in a biotoxin bulk water concentration
of 40 mg/L. It is assumed, for the purposes of this demonstration, that the example biotoxin is released
Page 20 of 93
-------
at the time of 12:00 AM and injected for a period of 480 minutes or 8 hours for a total mass of 300
grams. TEVA-SPOT simulates contamination injections in a water distribution system using mass per
minute for the input quantity (e.g., mg/min), i.e., assuming no volume added. A total mass of 300 grams
(300,000 mg) and an injection duration of 480 minutes results in a mass injection rate of 625 mg/min.
For additional information on the design of a contamination scenario please refer to Section 5.3.
The creation of an ensemble of contamination scenarios is the coupling of an "Injection Definition"
(steps 3.2.1 and 3.2.2) with a "Node Set Definition" (step 3.2.3) to create a "Node Injection" (See Figure
3.1.1). "Injection Definition" defines the contaminant injection parameters, namely the contaminant
name, mass injection rate (mg/min, or organisms or cells/min), and the start and stop time of the
injection (specified in the units of sees, mins, or hrs). "Node Set Definition" defines the locations where
the contaminant will be injected.
"Node Injections" is then added using "Add" to a folder called a "Scenario Set" (step 3.2.4). "Node
Injections" represent the summation of the injection definition with the set of injection locations. Each
"Scenario Set" represents a set of contamination scenarios. Multiple "Scenario Set[s]", with each set
containing one "Injection: Example_Biotoxin / Nodes: Tanks (34)" "Node Injections", can be created and
the number of simulations will be additive. Combining multiple "Node Injections" within a "Scenario
Set" folder will result in the product of the number of scenarios described by each "Node Injections"
definition to result in the unique set of simultaneous contamination scenario simulations. Combining
multiple "Node Injections" within a "Scenario Set" folder will result in a number of combinations of
simultaneous contamination scenarios. For example, a "Scenario Set" containing two instances of our
"Injection: Example biotoxin/Nodes: Tanks (34)" will result in 34 * 33 combinations of simultaneous
contamination scenarios. This will result in 595 unique scenarios.
Once an ensemble of contamination scenarios are created within the "Edit" box associated with
"EPANET Simulations", the "EPANET Simulations" will be ready (i.e., not grayed out) to "Execute". Step
3.2.5 shows the "Execution" and completion (green check mark) of the EPANET simulations.
Page 21 of 93
-------
Contamination Scenarios
Step 3.2.1.
In the "Scenarios" box, Click "EDIT" button next to "EPANET Simulations" and the dialogue box appear.
TEVA SPOT (Network 4 Utility Faolrties in C Values [M*
I
ton-Arc &cm»nd JunOwns (IHi;
| th-Me j
Input Generation
£ v tottaKzatwn: ropulatlon j Edit
IKi:0»:tlO:[XI
Estimwed Total Time OO-OOrOOrH
-- Upstrwm 8> Downstream Mows I EgK j IHatmMwl fotn[itntein IKIH- Uft/.'V.'m ( i
O WMhUipKKAMlfla j M« j
O j:. iBMpBCtsAnilKOS I UK j
u.t
For more explanation, see TEVA-SPOT-GUI User's Manual.
Page 22 of 93
-------
Contamination Scenarios
Step 3.2.2.
Under the "Injection Definition", click "Add". Enter "Example_Biotoxin" for the NAME and the parameters: Mass Injection
Rate of 625 mg/min and for the START and STOP Times of the injection, select "minutes from the drop down menu, and
enter 0 and 480 minutes, respectively. In the description box, type "8 hour injection" since 480 minutes corresponds to an !
hour injection. (For descriptions of the other parameters shown in this dialogue box, see below.) Click "OK".
•• 'MW HfMl
* Httnnfl MaMflt Ho Htenog
Mass fc)«UM Rate
WQ Tcitrine* (rofti'L)
8 VJU
o
tI.N«k.(_£aJ fcllmlri Com,*™, TmcM/JMIM! !!:!!:!! TO
E~oft
.-.Trim
T5iicimp«BMilI«
Kjm
»
"Concentration Zero Threshold (mg/L)" is the concentration of contaminant in the distribution below which it will not be
tracked. This feature behaves like a detection limit, that is, any concentration value below this value will be recorded as
zero.
"WQ Tolerance (mg/L)" refers to the EPANET water quality tolerance parameter. This parameter specifies the smallest
change in water quality that will result in a new parcel of water being created in a pipe. The default setting in EPANET is 0.01
mg/L for chemicals [1]. The water quality tolerance determines whether the quality of one parcel of water is the same as
that of another parcel. For a chemical contaminant the value could be the detection limit of the analytical method used to
measure the chemical contaminant, adjusted by a suitable safety factor [1]. Generally, lowering the water quality tolerance
will provide increased accuracy of the water quality modeling results at the expense of increased simulation time.
Inserting a WQ Tolerance (mg/L) value in this step (within the Injection Definition box pictured above) will override the WQ
Tolerance specified in the EDIT/ Ensemble Options box; however no value need be specified here.
Note that the actual START time of the contaminant injection is defined in the .inp file, which for the example model being
used here is 12:00 AM local time.
Page 23 of 93
-------
Contamination Scenarios
Step 3.2.3.
Click "Add" under the "Node Set Definitions" and select "Facility Nodes". In the Name field, type "Tanks". Check the "Tanks"
component. Click "OK". Note that in addition to all tanks, connected nodes are also included.
jj TEVA-SPOT (Network_4_Utility_Facilities in C-Values [Main])
Ensemble File Map Mode Edit About
Google Maps| Road Map ^||Map Default
^ Filtering
Isl Edit BaseEnsemble Parameters
Import... | | Export...
Injection Definitions
''Jon-Zero Demand Junctions
injunctions (3323)
\\\ Nodes (3358)
Scenario Sets
0 Simulations
Remove || New |
Execut
Availe
Inpir
Name Tanks
Facility Types
J Selects all tanks and nodes connected to all tanks
:_ Valves
LJ Pumps - Upstream Nodes
!_ Pumps - Downstream Nodes
The example here demonstrates how a "Utility Facility" node set is created. The use of "Tanks" as the only component of
the "Utility Facility" node set is arbitrary. All utility facilities could be selected to represent all the water system components
as locations for injection.
Page 24 of 93
-------
Contamination Scenarios
Step 3.2.4.
Select "Example Biotoxin" & "Tanks". Click "Create" "Node Injections". Click "New" under "Scenario Sets". A folder
appears. In the Node Injections box, select: "Injection: Example Biotoxin/Nodes: Tanks (34)". Below the Scenario Sets box,
click "New". The folder name "Scenario Set" will appear; click on "Scenario Set". Click "Add" to add to "Scenario Sets" folder.
Click "OK" and the "Edit BaseEnsemble Parameters" box disappears.
_ TEVA-SPOT (Network_4JJtility_Faalities in C-Va!ues [Main])
;•.-,: V,™ EC- .--»-',,-
Map charts | Tables]
* Filtering ^Manage | Ho Filtering
Import... | Export...
ijci. lull c lln luha
(Non-Zero Demand Junctions (1621)
All Junctions (3323)
AJI Nodes 1,3358:
~5^n
Edit
Delete
_Creste J
Hu cL.jeaiulla
JS^^^H^^^^^^S^^^^^^fi
|fte?iKws|
| OK ] | Cancel
^S3?5^^^^B
^dnailb j=
• Injection: Example_Biotoxin / Nodes: Tanks £34}
t \-\ \
34 Simulations
[ Remove j| Mew ]
Execution
^ [7| Initialization: Population
& [*] Upstream & Downstream Nodes
„
0 Q EPANET Simulations
0 _iHeal(h Impacts Analysis
O L! ISO to Imparts Analysis
0 1 1 Infrastructure Impacts Analysis
0 l_i Sensor Placement
1 Ed" |
r^n
[ Edit |
Ciii]
r E*'
| Edit |
fain
Estimated Time Remaining 00:00:00:00
Estimated Tola! Time 00:00;00:14
Estimated Completion Time 06/25/2013 12:31:38 PM
| Execute | Te n * i
Example here demonstrates how a "Utility Facility" node set is combined with an "Injection Definition" to create an
ensemble of contamination scenarios.
Page 25 of 93
-------
Contamination Scenarios
Step 3.2.5.
There is no need to re-run the "Initialization: Population" and the "Upstream & Downstream Nodes", so unclick those boxes
if they are still checked. Click the EPANET Simulation EDIT box and then click EXECUTE. Execution time could take a while
(from a few minutes to many hours or even days) depending on the network model, the number of contamination scenarios
being executed, and the computer that is being used. The simulation of contamination scenarios is running properly when
the scenarios are being counted down, e.g., "running 2 of 34...". EPANET Simulations are executed and completed. Green
checkmark appears.
rwmble hip MAP Mode Edit About
I Cough MqriftNHlM.t, - Map &
!itl f.ifrr-.ij
O Hwtth Imports Analysis [ E*t .]
O I SO flo ktiDMIs Antfss | Mt |
Q rt^uct^pAmAn.**; [~iir|
s?
Note that multiple applications of TEVA-SPOT-GUI (each application is termed an "Ensemble") can be opened at once and, if
desired, executed together. Executing the desired modules within multiple ensembles will result in TEVA-SPOT-GUI running
all the modules in a sequential fashion, i.e., one after another in the order they were started, until all are complete. After
the Utility Facility ensemble is created, it will be executed. If modules are not executed together, each application should be
unchecked, if needed, after it is run to avoid the time-consuming re-running of the application as other applications are
selected for execution.
Page 26 of 93
-------
3.3 Health Impacts Analysis:
Within the Health Impacts Analysis module, TEVA-SPOT-GUI provides for the selection and input of (1)
dose calculation parameters and (2) dose response method and parameters, including disease
progression parameters. TEVA-SPOT-GUI version 2.3.1 estimates public health consequences from the
ingestion of contaminated tap water.
The health impacts analysis module provides the capability to assess public health consequences in
terms of dose received or a health effects end point using contaminant-specific dose response
information entered by the user. A check mark in the box "Calculate Dose-Response" will enable or, if
absent, disables the dose response capabilities.
To "Calculate Dose-Response", the user first selects the type of contaminant, i.e., "Chemical/Toxin" or
"Biological". Next, the user selects the dose response model to be used. TEVA-SPOT-GUI provides two
such models: (1) probit slope and (2) sigmoid, 4 parameter model. We demonstrate the use of the
probit slope method here. For "Chemical/Toxin" contaminants, health consequences are based on a
lethal dose (LD). For "Biological" contaminants, health consequences are based on an infectious dose
(ID). For "chemical/toxin", the user enters their contaminant-specific LD that would result in death to
50% of the population (LD50) given such a dose. For "Biological", the user enters their contaminant-
specific ID dose that would result in 50% of the population becoming diseased or infected (ID50) after
receiving such a dose. For "Chemical/Toxin" or "Biological" the user is required to enter a contaminant-
specific probit slope factor. For "Chemical/Toxin" the user is also required to enter a body weight since
LD50 values are specified in terms of mg of chemical or toxin per kilogram of body weight.
Since health consequences for biological contaminants are estimated as infections, the user is required
to enter a "Fatality Rate" (fraction from 0 to 1) to predict fatalities. For "Chemical/Toxin", the dose
response model is predicting fatalities so a fatality rate is not needed. For both "Chemical/Toxin" and
"Biological" contaminants, the user is also required to enter a "Latency Time" and "Fatality Time" in
units of hours (integer only). The user can simplify the process for entering contaminant specific dose
response information by using the "Contaminant Defaults" box. The user prepares an XML formatted
file (i.e., contaminants.xml) containing dose response data for a group of contaminants. The
"contaminants.xml" file is then placed in the following folder: (C:\Program Files\TEVA-SPOT\Client).
If "Calculate Dose-Response" is disabled, the only public health measure available will be "Dose
Thresholds". If the dose threshold-based approach is desired, the user enters dose thresholds into the
dose thresholds panel by clicking the associated EDIT button, entering a dose threshold value and then
clicking ENTER after each value. Using dose thresholds, public health consequences are reported as the
number of people who receive a dose greater than the indicated threshold. The "Dose Thresholds"
method can be used with the "Calculate Dose-Response" method. The methodology for using dose
thresholds is described in [6], [7], [8], [9], [10], and [11].
Our example biotoxin has a lethal dose to 50 percent of the population (LD50) of 6.00 x 10"6 mg/kg. We
use a probit dose response model with a beta slope factor of 0.6711 to characterize health effects (i.e.,
fatalities) for a given dose of the biotoxin. The fatality rate is determined by the probit dose response
Page 27 of 93
-------
curve. For demonstration purposes, we consider the typical adult person's body mass to be 71.8 kg and
their mean per capita water ingestion rate to be 1.41 liters per day. We assume a person's dose of the
biotoxin is solely from the ingestion of contaminated tap. We use the "five fixed times" ingestion timing
model in TEVA-SPOT [6]. The contaminant description and distribution system model used are for
demonstration purposes.
The setup and execution of the health impacts analysis module requires two steps: (1) step 3.3.1,
entering the needed parameters and (2) step 3.3.2.
Page 28 of 93
-------
Health Impacts Analysis
Step 3.3.1.
In the "Execution" dialogue box, in the "Modules" box, click "EDIT" button next to the "Health Impacts Analysis" or HIA
module. Input the parameters shown in the screen capture to the right to complete the data inputs for the HIA
module. Click "OK".
Us) Edit Health Impacts Analysis Parameters
Contaminant Name Example_Biotoxin
Dose Calculation Parameters
Dose Calculation Method Inqestion of Tap Water -r
Ingestion Timing Model [ATUS
Ingestion Volume Model Fixed Volume
Mean Volume (liters/person/day) 1.41
Dose Thresholds
Dose Thresholds
Response Thresholds
Edit
Miscellaneous
Number of worst-case fatality scenarios to save 0
Number of worst-case dosage scenarios to save 0
Use one server per node
|V] Calculate Dose-Response
Dose Response Options
Contaminant Defaults
Contaminant Type
Contaminant Type
Select...
Chemical/Toxin
Average Body Mass (kg) 71.8
Dose Response Method
Dose-Response calculation method Probit T
LD50 (mg/kg)
6.00E-6
Beta
0.6711
Disease Progression Parameters
Latency Time (hrs)
Fatality Time (hrs)
48
Miscellaneous: Placing a number (integer only) in the boxes in the Miscellaneous section, i.e., next to either "Number
of worst-case fatality scenarios to save" or "Number of worst-case dosage scenarios to save" will save contaminant
concentration data per time step of the simulation (fatality scenarios) and/or dosage data per time step of the
simulation (dosage scenarios) for the number of highest ranked scenarios indicated.
The purpose of "fatality scenarios" and "dosage scenarios" in TEVA-SPOT-GUI and, as described here, is only for the
purpose of saving and retrieving contaminant concentration data or dosage data for high percentile consequence
contamination scenarios.
The resulting "fatality scenarios" and "dosage scenarios" based text files, one for each scenario saved, will be saved to
the C:\TEVA-SPOT-Database\C-Values\Network_4_Utility_Facilities\Health Impacts Analysis folder. For large networks
(>10K nodes), each file saved can be large; therefore, the storage requirement to save many such files maybe very
large. It is important that a relatively small number be placed in these boxes box, i.e., typically on the order of 10 or
less.
Use one server per node: Check this box only if your computer's resources (processing speed, number of computer
cores and memory are very limited.
Dose Thresholds: An alternate method for determining public health consequences.
Page 29 of 93
-------
Health Impacts Analysis
Step 3.3.2.
If checked, unclick the EPANET Simulations as there is no need to execute this scenario again. Click the "Health Impacts
Analysis" module and then click "EXECUTE". Execution time could take a while (from a few minutes to a number of
hours) depending on the size of the network model, the number of contamination scenarios being run, and the
computer that is being used. Green checkmark will appear when the Health Impacts Analysis module completes
successfully.
£ns«mble File Map Mode Edit About
M°Pi Chattel Tables
^| V A Google MapsWlMap -r|| Map Default
Input Generation
® Initialization: Population
& Upstream & Downstream Nodes I Edit
Scenarios
£ EPANET Simulations
Modules
I -^ I Estimated Time Remaining 00:00:00:00
' Estimated Total Time 00:00:00:09
Estimated Completion Time 06/25/2013 03:34:02 PM
cute
® V Hearth impacts Anah/sts
Q 'so to Impacts Analysis
O infrastructure Impacts Analysis
Sensor Placement
O n Sensor Placement
«
The Health Impacts Analysis (HIA) module completes.
Note: It is faster (more computationally efficient) to run the EPANET Simulations along with the "Impacts
Analysis Modules", i.e., HIA and IIA.
Page 30 of 93
-------
3.4 Input of Infrastructure Impacts Analysis Data:
TEVA-SPOT-GUI provides for the analysis of infrastructure consequences as determined by the length of
pipe (ft) that witnesses contaminant concentrations above a user specified concentration threshold.
Steps 3.4.1 and 3.4.2 demonstrate this module.
Infrastructure Impacts Analysis
Step 3.4.1.
In the "Execution" dialogue box, under the "Modules" box, select EDIT next to the Infrastructure Impacts Analysis Module.
Place cursor in box labeled "Concentration Threshold" and type "0.01"; press ENTER. Next, type "0.1" and press ENTER. Click
"OK" to close the "Edit Concentration Thresholds" dialogue box and again to close the "Edit Infrastructure Impacts..."
dialogue box.
Edit Infrastructure Impacts Analysis
Contaminant Name BioToxin
Concentration Thresholds
Concentration Thresholds
OK
Cancel
Edit Concentration Thresholds
Concentration Thresholds
0.01
0.1
Concentration Threshold
OK I I Cancel
Multiple contaminant concentration thresholds can be specified. Enter the value and press ENTER. The values specified
above are only for demonstration purposes.
Page 31 of 93
-------
Infrastructure Impacts Analysis
Step 3.4.2.
Unclickall boxes. Then click the "Infrastructure Impacts Analysis" module to check it and then click "EXECUTE".
TEVA-SPOT (NetwxM.WI'tyJaciKies in C-Values [Ma.
**moie File Mac* Mode fctM AtxxA
•-. Uiarti i'.it-*f
AvnilnM^
input Gi»r*r,n-jFi
0
M
O T» B
® ^
GREEN CHECK MARKS denote that the Infrastructure Impacts Analysis (HIA) module has completed running.
3.5 Determining Utility Facility C-Values:
Public health and/or infrastructure consequences can result from the injection of a contaminant at a
particular utility facility. Through the analysis of the example biotoxin threat (Sections 3.1 to 3.4) TEVA-
SPOT-GUI enables the user to estimate public health and infrastructure consequences that could result
from the intentional injection of the contaminant at each of the utility facilities selected. The
consequences that are estimated to result from the injection at each utility facility represent each
particular facility's C-Value.
Once the utility facility C-Values are determined they can be used to rank the utility facilities that
represent the greatest concern for intentional contamination given the injection of the example
biotoxin. Ranking of utility facilities can be based on public health C-Value and infrastructure damage
(e.g., extent of pipe feet witnessing contamination) C-Value.
Page 32 of 93
-------
The C-Value associated with each facility can be obtained in a variety of ways, e.g., tables or maps,
available from within the TEVA-SPOT graphical user interface or, depending on the data, within the
TEVA-SPOT-Database for the particular ensemble.
We demonstrate the retrieval of utility facility C-Values from TEVA-SPOT-GUI and the ranking of utility
facilities by their public health C-Value that could result if contamination were introduced at the
particular facility. We also demonstrate how to retrieve utility facility C-Values based on infrastructure
contamination and show how to use the information to rank the utility facilities based on infrastructure
consequences. For any given ensemble of contamination events, the nth percentile injection node is the
injection node that yields the nth percentile consequence or C-Value. TEVA-SPOT-GUI ranks
contamination injection locations by public health consequences, e.g., fatalities, and by infrastructure
consequences, e.g., feet of pipe that witness contamination above a certain level. TEVA-SPOT-GUI ranks
a given ensemble's contamination injection locations according to six percentile scenarios, namely 10th,
25th, 50th, 75th, 90th and 100th. More detailed information is provided for this ranked contamination
injection locations. Table 3.5.1 provides a tabular description of some of the TEVA-SPOT-GUI outputs
that can be used for retrieving and evaluating utility facility C-Values. Table 3.5.1 also provides a
roadmap to the TEVA-SPOT-GUI illustrative screen shots presented in this section.
Table 3.5.1: Description of Selected TEVA-SPOT-GUI Outputs to Determine, Rank, and Evaluate C-Values
Step
Map or Table
Use
3.5.1
Map: Estimated Fatalities by Injection
Location
Obtain a map of utility facility C-Values.
Map provides a visual ranking of utility
facilities, color-coded by largest public
health C-Value.
3.5.2
Maps: Estimated Fatalities by Receptor
Location -various percentile (25, 50, 75, 90,
and 100th) maps.
Obtain detailed information about where
public health consequences occur in the
network given contaminant injection at a
particular utility facility.
3.5.3/3.5.4
Table: Injection Impacts Table: Summary
Table
Obtain tabular utility facility C-Values
results. Use these results to rank utility
facilities by largest C-Value.
3.5.5
Table: Estimated Fatalities Receptor Nodes
Table
Examine ranking of utility facilities by C-
Value (e.g., examine number of fatalities
by receptor locations for six, ranked
percentile scenarios ranging from 10th to
100th percentile). Use data to evaluate
highest ranked utility facilities.
3.5.6/3.5.7
Map: Infrastructure Impacts Analysis: Feet
of Pipe with Concentrations Over 0.01 by
Injection Location. Although two IIA analysis
concentration thresholds were specified, we
illustrate only the case of 0.01 mg/L here.
Obtain a map of utility facility C-Values
based on infrastructure impacts. Map
provides a visual ranking of utility
facilities by the contaminant injection
location that results in the largest feet of
pipe that witnesses contamination above
0.01 mg/L.
Page 33 of 93
-------
Step
3.5.8
3.5.9 to 3.5.11
Map or Table
Map: Infrastructure Impacts Analysis: Pipes
with Contamination Over 0.01 - 90th
Percentile
Table: Upstream and Downstream Nodes: 3
tables. (1) Downstream Node Counts, (2)
Downstream Population, and (3) Upstream
Node Counts
Use
Obtain detailed information about which
pipes witness contamination over a user-
specified concentration for a particular
contamination injection (utility facility)
scenario.
Obtain additional information about the
water distribution system network. Rank
utility facilities by their upstream or
downstream connectivity.
Retrieving and Using C-Values
Step 3.5.1.
If the "Execution" panel at bottom remains open, you should close the "Execution" dialogue box to begin this section to
provide more space to viewing the maps and tables. Click on the MAP tab. Using the drop down menu next to the "Map"
field bar, select the map titled, "Estimated Fatalities by Injection Location". Note that injection locations (nodes) are color
coded by the number of fatalities. Tank-3327 represents the tank that if contaminated could result in the largest public
health consequences.
[Mjmjtfd f iLilitwi by ln>Ktlon lociUon
These results can be exported from TEVA-SPOT-GUI to a dbase IV file and then incorporated into a geographical information
system (GIS). See "ENSEMBLE MANAGEMENT", "EXPORT MAP DATA".
Page 34 of 93
-------
Retrieving and Using C-Values
Step 3.5.2.
Scrolling down through the drop down menu next to the "Map" field bar, select the map titled, "Estimated Fatalities by
Receptor Location - 90th percentile". The RED SQUARE corresponds to the injection location, Tank-3329. Note that this map
details estimated receptor information for a specific scenario, i.e., the estimated 90* percentile scenario based fatalities. The
legend corresponds to the number of fatalities by receptor location. The map below has been altered for better viewing
here, i.e., the legend has been placed on the left.
On the slider bar associated with the legend, using your mouse right click and then choose "Set min/max". Enter the number
"1" in the "min" box. This will restrict the view of fatalities by receptor location to just those locations that experienced one
or more estimated fatalities. This will not be necessary with later versions of TEVA-SPOT-GUI (i.e., 8-07-13 and later).
to TEVA-SPOI (Nfl*«MJ>liMy-F*;*li« 10 C-VHaK IMxiD I
!nsOTt>» file Map Mode Eo-1 About
% X A <•<• •
Estimated Fatalities by Receptor Location
5.0 £0 hcmrs ft hours}
These results can be exported from TEVA-SPOT-GUI to a dbase IV file and then incorporated into a geographical information
system (GIS). See "ENSEMBLE MANAGEMENT", "EXPORT MA DATA".
Page 35 of 93
-------
Retrieving and Using C-Values
Step 3.5.3.
Select the TABLES tab. Then, using the drop down menu next to the table field bar, select the "Injection Impacts Table". This
table provides each contamination scenario's (i.e., each tank) consequences in terms of "Estimated Fatalities" (see RED
CIRCLE). Other information is available, but we restrict our attention to the public health based C-Values.
This table can be used to determine the utility-based locations (node IDs) that would result in the greatest consequences if
contaminated.
*J ItVA SPUI (NL'lwuifc 4 UliMy kiulrLic:* in t Vdluei MdinlJ
Ensemble File Table Mode Edit About
Map] Charts; mMee
Tobte THHfth impacts Arwtysls]: HjetWn impacts TnWe
b.K-tt
TAMK
fAMK:
TAHK
TAWK
TAJJPt
TAWK
TAJJK
TANK
r'AHK
TAtJK
T.-:K
T/i!JS-
T^'l\
I.-:K
r ••. • ix -
TfJHK
\fi3HK-
TAHK-
TAMK-
TAHK.
TAtJH
FAHK-
^^ *«^ Summary Table
mi I. mil- ln|<-{ linn HiHli' Type
35 ^|625.0 (0 hours
3 [WSJJ (U hours -
-.( 625.« TO l-.UU'i
3J bZ'j.O (U hours
1 fi^-j.ij ftf hi.ur-.
3 625.0 te tvours
3 jwif.g (u hours
1 [635.0 in hour!;
J CiJj.'J [0 hours
3 625.0 [0 hours
3+4 '&2i,Q (D hours
1 rtP-vf, !fl hinir-.
3 £25,0 (6 hour&
^ [635.0 (Olwur,
348 1625.0 (ft teurs
J4'j |wi,« (U hours •
3.34 6?5.rt {ftjjnur'. -
174 |fi3S.O{fHw3ur'. -
350 625.0 ffl hours
JS'j JwiTu (U houre -
^^lS fi?1.fl ;ft hinir-. -
D^U 62&.D fO hours
T52 J625.0 (0 hours
3i4 |b2i.g (D hours -
HS3 IH5.Q (n luurs -
347 ,625.0 (0 hours
341 i&2^.o (o hours -
356 J625.Q (0 hours
2Xi Itzj.o (0 hours -
Iwurs)^^
hojrs:
fTOW9>
hows)
hn«S}
hours;
how-s)
hpwro
ho^rs)
hm .si-.;
ho-.rS;
hgursj
IlLI'.il-.'
hg.fS'
hours)
hoursj
hovrs)
haitr-j}
hours)
hoiB-ij
hows}
hours}
hour-,}
T
hours) p"
hciur-.) it
hours) ]T
hours) [t
hcm-;> |T
hours) rr
hojrsj
hc^j rs1;
hOiirsJ
t
r
T
MaxCtMi... Mm linli.. NimdC-... Number n! Ffdimalwl Falalj^. tkHli-.ltx 40% Fstimol^J FfltflMii-r; Nodi", with Fnlrttitu-.
0-147) 6.623^2.399.352 42.399.352! 55«' 8««
U.UAi 2.<4tJiJ JJ.14/.^ys 33,H/.bSHi 4i(j
0-076 Z.408' 33.147.1761 33.147.176j 486
D.Q2S JS.Ilb 21
0-0*5; 13.631 7
,t33.9/V 29.UJ.9/5 &2
,t?7.310i 77,177 .?H| 3Sfi
0.013 7.327 Z4.047.053i 24.Q47.053[ 547
o.uilT B,MS.
O.QSfl TP.fi^T1 1
U,Uaj 1K.3W 1
n n-!T n.^si 1
U.U7? 13.574 1
U.U/4 1J.09
n m I?.I«B i
0.143 39.461
n iTifl -S7.O57
0.036 9-5.429
g.gj/ jif.yia
n.oifl ?,'1.7so
0.21 20.603
u filt 70.1 Rfi
0 1S6 54.JJ74
U.UV U'j.i^O
n.n=i3 m.j>")o
O.U-JO 277.003
0.1 */ 14 3-91 W
0.139 35.242
U.l'jli IW.ub/
0-11?] 54.6?li
F&
1212
3,B94.bU 2J,S9-f.&a[ 2bJ i/1
1,774.753 lft,374.752| Mflj ?n
.b.-j li/ l/,B3i,12/, 209: 3/1
JI35.177 17,835.1371 JO* 171
.aae.eei i4,«sB.e8i lee! 245
•1 . BOO . J 'j 14,8W*,J1 lbb[ 24^
,1J3.7M IZ.lU.TOa «| 148
.33;. 337 8,351.237 99
,-iin .ISA 7,4
-------
Retrieving and Using C-Values
Step 3.5.4.
The results in Step 3.5.3. can be exported to EXCEL by SELECTING MENU TABLE and then choosing "EXPORT TO EXCEL...". A
dialogue box appears to indicate where the .xls file should be saved.
ItVA-SPOT (MeHtort:J.Utii.tv.f*C'lilie'fl C-Values iM*nj>
Ft* |T«Mej Moo* Edit About
MI? \ Own. ** tuport to bexL.
Tittle {HttKtt tnpacfe Anilysi]: kijcdwi Irtpectj fitt*
Summary Table
Mm Con... fete 14,, (fester**, IM**. i lift
M5.0 t*
TAIIS-J353 «J.O (
;•:• ,~ i
•Of. .T
nCg
:•: " T
'.: ". 1
••; ". . T
;.',r I
fffc-.g
:•:•- i
(»fc— JT
:;• -• i
.;,- . ,
-
rt»_T
{OK_T
••:• f. i
.•:• >. T
: - :
i'-.f, ':
:•: - T
•0 f- . '
•i > i
[eiv-.:T
SEE
•
wlOf
SEI
tjOpcn
lootm:
P*Cfr'^
Xeois
^»
Docum*nt5
»•
•Compuler
»
^1
Nrtwcrk
bn
j fin* »• ? 'Qf3
t doct
tj«
1 lite
l modules
I prop!
* Con1»mirian1sjirn3
«ii*mbl*jid
nwdUeuid
- runclienLftst
sfxxsd
(na^MMuad
MM nanH. Open
«*•««« «H« ^j | SmoH j
^
85.2*1 I.MiS.«J
IM.401
';
Note: Typically under Windows 7 operating system TEVA-SPOT-GUI IS NOT ALLOWED to export an output table to the
directory: C:\Program Files\TEVA-SPOT\Client. Therefore, the user should browse to an accessible location.
Page 37 of 93
-------
Retrieving and Using C-Values
Step 3.5.5.
Select the TABLE drop down menu. Next, select the "Estimated Fatalities Receptor Nodes Table".
This table provides the number of fatalities at each receptor location for six, ranked percentile scenarios ranging from 10th to
100th (ONLY 75th PERCENTILE SHOWN).
' .'•Vli'l •
tincmlsif File Table Mod* Edit Atamt
Mep, Charts T«dlBB
•• •'!-•-• IH^ »i JnwH3i;^iah^:E«UmflCrtMll«»»l«C«^MadMT^
Estimated Fatalities
UWUWN-
UIJCTJON
UHTTK1N-
UIK1KJN-
UtJCIJON
Note that the table above has been truncated for viewing here.
Columns (except first column) represent the percentile (ranked utility facilities). Rows represent the receptor locations for
each utility facility contamination injection location. Each cell in the table represents the number of people expected to die
(fatalities).
These results can be exported to EXCEL (.xls) file using the approach described in Step 3.5.4.
Page 38 of 93
-------
Retrieving and Using C-Values
Step 3.5.6.
Select the MAP tab. Next, select the "Infrastructure Impacts Analysis" map titled "Feet of Pipe with Contamination Over 0.01
by Injection Location". Infrastructure damage (pipe contamination) can be examined by Injection Location.
1£VA SPOT mrma*,4JM:ty_r*aeurs n C VAjn SW«m|)
Emenou Me Map Moor Eat ADaut
~p Chrti ru*.
f«t of Pip* With Conc*ntra«ont Over 0.01 by Injection Location
\
Note: For viewing purposes, this screen shot has been modified to fit this format. All MAP data can be exported for
incorporation into GIS using ENSEMBLE MANAGEMENT and EXPORT MAP DATA.
Note: Recall that the contaminant concentration value of 0.01 mg/L was specified in Step 3.4.1. The values used are for
demonstration purposes only.
Page 39 of 93
-------
Retrieving and Using C-Values
Step 3.5.7
Click on the water faucet icon within the MAP TAB and the "EPANET Components" panel will appear. Next, with the water
faucet icon SELECTED, click on the darkest red tank and notice the panel "Select Water Network Element" appears. The
arrow icon, in between the magnifier glass and the faucet, can be used to move the map. Within the "Select Water Network
Element" box, use scroll bar to maneuver to bottom to see that this color-code Tank could be Tank-3343 or Tank-3344. Use
the magnifier glass to ZOOM in to see the two tanks (not illustrated here).
Ensenb* File Map Mode Edit About
Number of Pipes With Concentrations Over 0.01 by Injection Location
HM«ril|
««*f
Use the "Water Faucet" icon to examine model elements in more detail. Notice that the map above is for "Number of Pipes
with Concentrations Over 0.01 Injection Location"; whereas, the map in the previous step was "Feet of Pipe...". Both versions
of maps are available.
Note: Recall that the contaminant concentration value of 0.01 mg/L was specified in Step 3.4.1. The values used are for
demonstration purposes only.
Page 40 of 93
-------
Retrieving and Using C-Values
Step 3.5.8.
Select the MAP tab. Next, select the "Infrastructure Impacts Analysis" map titled "Pipes with Contamination over 0.01 - 90th
Percentile". Infrastructure damage (pipe contamination) can be examined for specific, percentile ranked facilities, such as
the 90* percentile scenario highlighted here.
Ensemble File Map Mode Edit About
.,i:onicf-M.TsKr.,viyip «• Mop [Wrostrurturt kr^rts AA«ly«); Pip« wlh O^^inoton ww 0.01 - «tii Peranoie
Pipes with Contamination Over 0.01
90th percentile w*e (110143,43 Total)
Injection Node: TANK 3339
Injection Parameters: 625,0 (0 hours - 8 hours)
All MAP data can be exported for incorporation into GIS using ENSEMBLE MANAGEMENT and EXPORT MAP DATA.
Page 41 of 93
-------
Retrieving and Using C-Values
s
s
b
IS
T
(
e
t
I-
T
a
tep 3.5.9.
elect the TABLES tab. Next, select the table "Upstream & Downstream Nodes: Downstream Population Table". The table
elow is available as a result of executing the first two modules: "Initialization: Population" and "Upstream & Downstream
odes".
=j TEVA-SPOT (Network_4_Utility_Facilities in C-Values [Main])
Ensemble File Table Mode Edit About
Map| Charts] Tables
Table [Upstream & [
ownstream Nodes (UpDownStreamNodes)]: Downstream Population Tabte
Downstream Population
Simulation Length: 168 hours
Total Population: 151938
Injection Node ~
TANK-3357
TANK-3356
TANK-3355
TANK-3354
TANK-3353
TANK-3352
TANK-3351
TANK-3350
TANK-3349
TANK-3348
TAHK-3347
TANK-3346
TANK-3345
TANK-3344
FANK-3343
TANK-3342
TANK-3341
TANK-3340
TANK-3339
TANK-3338
TANK-3337
FANK-3336
TANK-3335
TANK-3334
TANK-3333
TANK-3332
TANK-3331
FANK-3330
TANK-3329
TANK-332S
TAHK-3327
FANK-3326
TANK-3325
TANK-3324
RESERVOIR-3323
# of Downstream Nodes
3357
6
7
48
30
71
464
464
464
464
18
12
464
464
464
43
15
50
701
701
701
157
83
3357
3357
3357
3357
3357
3357
3357
3357
3357
3357
86
3357
Dead End?
N
N
N
N
N
N
N
H
N
N
N
N
N
M
H
N
ti
H
N
N
M
N
N
M
U
N
N
N
N
N
M
a
N
N
N
Downstream Population
151938
163
136
1459
1063
2034
1B031
1B031
18031
18031
766
122
16031
18031
1B031
2085
266
2723
30339
30339
30339
3966
3543
151938
151938
151938
151938
151938
151938
151938
151938
151938
151938
3543
151938
Downstream Population %
100.0%
0.1%
0.1%
1.0%
0.7%
1.3%
11.9%
11.9%
11.9%
11.9%
0.5%
0.1%
11.9%
11.9%
11.9%
1.4%
0.2%
1.8%
20.0%
20.0%
20.0%
2.6%
2.3%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
2.3%
100.0%
his table provides for each node: (1) number of downstream nodes, (2) whether the injection node is a dead-end node - yes
V) or no (N), (3) downstream population, and (4) downstream population as a percentage of the total population. A dead-
nd node is any node that does not have a downstream node. In TEVA-SPOT-GUI it is determined hydraulically and,
lerefore, is dependent on the simulation time.
ere injection node is every node in the model NOT just the injection nodes described by the ENSEMBLE definition.
hese results can be exported to EXCEL by SELECTING MENU TABLE and then choosing "EXPORT TO EXCEL..". A dialogue box
ppears to indicate where the Excel.xls file should be saved.
Page 42 of 93
-------
Retrieving and Using C-Values
Step 3.5.10.
Select the MAP tab. Select "Upstream Node Counts".
^, TEVA-SPOT (NetwotM.UlilityJacilities in C-Values [Main])
Ensemble File Map Mode Edit About
Gooate Mnp^ Roan Hap - Map [upaream 8. Downstream Mates (upOovmslreaiTiMocles)]; upareum Mode COunls
Upstream Node Count
Simulation Length: 168 hours
Note: For viewing purposes, this screen shot has been modified to fit this format. This map illustrates the areas of the
network (color coded) that are fed by the largest number of upstream nodes. For example, the dark red nodes are fed by the
largest number of upstream nodes. All MAP data can be exported for incorporation into GIS using ENSEMBLE MANAGEMENT
and EXPORT MAP DATA.
Page 43 of 93
-------
Retrieving and Using C-Values
Step 3.5.11.
Select the MAP tab. Select "Downstream Node Counts".
41 TEVA-SPOT (Network_4_Utility_Fadlities in C-Values [Main])
Ensemble File Map Mode Edit About
Charts Tables
Google Maps Road Map " Map [Upstream & Downstream Nodes (UpDownStreamNodes)]: Downstream Node Counts
Thematic Legend
Dovmstream Nods
Downstream Node Count
Simulation Length: 168 hours
Note: For viewing purposes, this screen shot has been modified to fit this format. This map illustrates the areas of the
network (color coded) that feed the largest number of downstream nodes. For example, the dark red nodes feed the largest
number of downstream nodes. All MAP data can be exported for incorporation into GIS using ENSEMBLE MANAGEMENT and
EXPORT MAP DATA.
Page 44 of 93
-------
4.0 PROTECTING CRITICAL FACILITIES.
In this section we describe the process for using TEVA-SPOT-GUI to evaluate a set of user-defined critical
facilities for their upstream vulnerability to contamination. Specifically, a user can identify a set of
locations as critical facilities and TEVA-SPOT-GUI will determine the set of upstream network locations
that supply water and, therefore, could potentially contaminate a critical facility. For example,
hospitals, schools, or other critical infrastructure can be identified and TEVA-SPOT-GUI can be used to
help protect them. Since utility facilities are separately identified by TEVA-SPOT-GUI, they are not
considered critical facilities. By determining the upstream locations that could most affect the critical
facilities, TEVA-SPOT-GUI helps the user to prioritize SRS efforts and resources to protect critical
facilities. This feature can also be used to help determine the best locations to place monitoring stations
to protect the critical facilities. Here, again, we define critical facilities to not include utility facilities in
order to avoid confusion. Utility facilities were examined in Section 3.0.
Critical Facilities can be evaluated after executing the module "UpDownStreamNodes". The
"UpDownStreamNodes" module must be run prior to the "EPANET Simulations" module in order for the
critical facilities capability to be available. In Step 3.1.8 we demonstrated how to run this module. Here,
we demonstrate how to select or upload a list of critical facilities into TEVA-SPOT-GUI and then use the
consequence assessment modules to evaluate the ensemble of contamination scenarios. The user
selects or inputs a list of critical facility nodes, which TEVA-SPOT-GUI then determines the ensemble of
associated contamination scenarios that could affect the critical facilities identified. The contamination
scenarios are then simulated in the "EPANET Simulations" module. Consequences are determined using
the public health and infrastructure modules described earlier. Similarly, consequence results can be
retrieved from TEVA-SPOT-GUI as described in Section 3.5 and Table 3.5.1.
The user identifies critical facilities by node ID. The node identifier for a particular critical facility could
be obtained from the geographical information system (GIS) database for the water distribution system
model. Alternatively, the user could identify the model node for each critical facility by using the
available mapping features with TEVA-SPOT-GUI. The mapping features are described in Section 5.0.
We also demonstrate how to use the sensor placement module to identify the preferred locations for
monitoring a group of critical facilities. The goal in designing such a monitoring strategy is to minimize,
subject to a constraint on the number of sensors, the consequences associated with contamination
event(s) directed at the critical facilities.
The process described will base the selection of the monitoring locations on minimizing public health
consequences (fatalities). The process described assumes that all the upstream locations identified are
equally likely locations for intentional contamination and, therefore, the monitoring locations selected
will seek to minimize the expected or mean consequences across the network. This means that sensor
monitoring locations are selected to minimize the mean public health consequences given the ensemble
of contamination events simulated.
Additionally, we demonstrate how to//x a sensor at a critical facility to ensure that a monitoring sensor
is placed at each critical facility to prevent consequences from occurring at the critical facility. Without
Page 45 of 93
-------
fixing a sensor at each critical facility TEVA-SPOT optimizes the placement of sensors upstream of the
critical facility to minimize the mean consequences determined from the group of upstream
contamination scenarios. In other words, each critical facility would be protected only from the extent
of minimizing the mean consequences associated with the upstream contamination scenario locations.
Therefore, without fixing a sensor at each critical facility, consequences could still occur at the critical
facility. For a given set of sensor monitoring stations, TEVA-SPOT will optimize the placement of sensors
for the number of sensors that remain after fixing at identified critical facilities. For example, if a user
sets a sensor set size in TEVA-SPOT to be 15 sensors and the water system being evaluated has 10
critical facilities with a sensor monitoring station fixed at each critical facility, TEVA-SPOT will optimize
the placement of the remaining 5 sensors.
Sensor placement optimization involves a two-step process: (1) generation of threat simulator output
tso-to-impact (tso-2-impact) files and (2) optimization of sensor locations. The generation of tso-2-
impact files (one for each combination of consequence measure, response time, and detection limit) is
the precursor to the sensor placement optimization step. The impact file is what the sensor placement
algorithm uses to optimally locate sensor monitoring stations to minimize a chosen objective. An impact
file stores the results of the consequence assessment. Specifically, for each contamination scenario, the
impact file contains a list of the locations in the network where a sensor could detect contamination.
Locations (nodes) that do not detect the contaminant are not included in the impact file for that
particular contamination scenario. For each location (node) that detects the contaminant, the impact
file contains the detection time and the total consequence given a sensor at that location is the first to
detect contamination from that scenario. For more information on the generation of impact files and
sensor placement optimization, we refer the reader to the User Manual for TEVA-SPOT-Toolkit [2]. For
more information on sensor network design using TEVA-SPOT [13].
Page 46 of 93
-------
Table 4.0.1: Overview of TEVA-SPOT-GUI procedure to identify critical facilities, determine and evaluate
possible upstream contamination events, design a sensor network to protect critical facilities, and
retrieve the sensor placement results.
Step
Feature
Task
Steps 4.0.1-Step 4.0.2
Ensemble Menu
• Save As
• Load
Creating a "Critical_Facilities"
ensemble using the "Save As"
feature to save an existing ensemble
under a new name, and then loading
the newly created ensemble.
Step 4.0.3
Input Generation
• Initialization: Population
• Upstream &
Downstream Nodes
Running the Input Generation
modules on the "Critical_Facilities"
ensemble.
Steps 4.0.4-4.0.8
Scenarios
• EPANET Simulations
Setting up the EPANET Simulation
module for the "Critical_Facilities"
ensemble.
Steps 4.0.9-4.0.10
Modules
• Health Impacts Analysis
• Infrastructure Impacts
Analysis
Setting up the Health Impacts
Analysis and Infrastructure Impacts
Analysis modules for the
"Critical Facilities" ensemble.
Steps 4.0.11-4.0.16
Sensor Placement
Setting up the Sensor Placement
module for the "Critical_Facilities"
ensemble.
Step 4.0.17
Execution
• EPANET Simulations
• HIA
• MA
• Sensor Placement
Executing the EPANET simulations,
HIA, MA, and Sensor Placement
modules.
Steps 4.0.18-4.0.19
Obtaining Results
• Map tab
Retrieving a sensor network design
map: "Sensor Locations". Retrieving
a sensor network design map:
"Sensor Event Detection: Links and
Nodes".
Steps 4.0.20-4.0.22
Obtaining Results
• Table tab
Retrieving sensor placement
performance results: "Trade-Off
Analysis Data" & "Sensor Placement
Summary", "Impacts and Detection
Time", "Sensor Ranking" tables.
Step 4.0.23
Obtaining Results
• Table tab
Retrieving sensor placement results:
"All Sensors" table & "Sensor
Counts" tables.
Page 47 of 93
-------
Protecting Critical Facilities
Step 4.0.1.
Under the ENSEMBLE MENU, select "Save As" while in the "Network_4_Utility_Facilities" Ensemble. In the "Ensemble
Name" box, type "Network_4 _Critical_ Facilities". Click "OK".
IbVA STOJ [Network 4 Uldily f-iiulifci.-* in C Vdluci |
• c Mjp Much- Ulll About
,-•! Save Ensemble As
Cns«nbl« collection c-vriuu TKHtt
Rlrjinhtr Ham- Hl-hMlrk -1 frail-a^ FrtcllitH".
,-,:»,
AvntoUe t
Inp-jt G«neral»n
Status
Estimated T*ne Remfl.filpq
F-Jiriidli'dTeilal Tinir
Ml
5 tKAIIfcT
Madrie
Q ~SQ to ImpaCTE fifutfos EdB ]
^ . IrA-ostruCture ImpuctE Ananas bAE j
Censor F-lBcement
O S.^H™™* .' » j
Use the "Save As" feature to replicate an ensemble. Using the "Save As" feature does NOT replace the current
ensemble but instead only saves the current or loaded ensemble with a new name. It does not save or replicate
any results - only the complete description of the ensemble.
Page 48 of 93
-------
Protecting Critical Facilities
Step 4.0.2.
Under the ENSEMBLE MENU, select "LOAD" and load the just saved ensemble to examine critical facilities. Next, choose
"Execution Control" under the ENSEMBLE MENU. Alternatively type the keys: Ctrl+E.
Ensemble File Map Mode Edit Abou
Map charts j Tables
3 Road Map * ) MsP Default
f Filtering | Manage | [NO Rite
Execution
Available Executior
Input Generation
O [_! Initialization: Population
0 EJ Upstream &. Downstream Nodes |
-Scenarios
0 OEPANET simulations
Modules
O IZ Health Impacts Analysis
0 |_ . TSO to Impacts Analysis
O Infrastructure Impacts Analysis
Sensor Placement
0 [J Sensor Placement
nated Total Time
Estimated Completion Time
Execute I Terminate
Page 49 of 93
-------
Protecting Critical Facilities
Step 4.0.3.
Click the two boxes associated with: "Initialization: Population" and "Upstream & Downstream Nodes" and execute the
modules.
Ensemble File Map Mode Edit About
Map Charts | Tablesj
Ji Google Maps Road Mep "• Map Default
••• Filtering , Manage j | No Filtering
Available Executions
Input Generation
O 0 Initialization: Population
MOT Running
Estimated Time Remaining
^=^J Estimated Total Time
I7J Upstream & Downstream Nodes | Edit Estimated Completion Time
O D EPANET Simulations
Modules
0 |_J Health Impacts Analysis [ Edit
O D TSO to Impacts Analysis | Edit
Q |_J Infrastructure Impacts Analysis Edit
Sensor Placement
LJ Sensor Placement
This step was demonstrated earlier for the "Utility_Facilities" ensemble.
Page 50 of 93
-------
Protecting Critical Facilities
Step 4.0.4.
Click "EDIT" button next to "EPANET Simulations" and the dialogue box appears.
,_, TEVA-SPOT (Network_4_Critical_Facilities in C-Values [
I ' ^
I Ensemble File Map Mode Edit About
Map charts Tables
Google Map Road... "• Map Default
Execution
Available Executions
Input Generation
J Edit BaseEnsemble Parameters
1 Add |
Edit ]
Delete
Node set Definitions
Non-Zero Demand Junctic
All Junctions (3323)
All Nodes (3358)
^ [ iii | t
Add ]
:_
Delete |
[ Create » |
Node Injections
Scenario Sets
[ Add »
I Remove Mew
[v] Initialization: Population
[V] Upstream &. Downstream Nodes Edit
0 LJ EPANET Simulations
Modules
0 [_J Health Impacts Analysis [ Edit
0 DTSO to Impacts Analysis | Edit
0 | | Infrastructure Impacts Analysis EdK
Sensor Placement
Q Sensor Placement
Estimated Total Time 00:00:00:14
Estimated Completion Time 05/01/2013 03:58:17 AM
Page 51 of 93
-------
Protecting Critical Facilities
Step 4.0.5.
Under the "Injection Definition", click "Add". Enter "Example_Biotoxin" for the NAME and the parameters and information
specified. For "Start Time" and "Stop Time", select "minutes" from the drop down menu. Click "OK".
-,_. TEVA-SPOT (Network_4_Critical_Facilities in C-Values
Ensemble File Map Mode Edit About
[SOX Jil Google Maps!Road... "IMap Default
fif [^1 Initialization: Population
& |yj Upstream a. Downstream Nodes
„
Q n EPANET Simulations
Q . Health Impacts Analysis
Q TSO to Impacts Analysis
0 Infrastructure Impacts Analysis |
^ | ! Sensor Placement
Edit
Edit
Edit
Edrt
Edit
Edit
Edit
Import... ] Export-
All Junctions (3323)
All Nodes (3359)
4 \ ".f J *
Add !
El
-i.il-
Add ]
Create » j
| Remove |
[ OK ] | Cancel |
[ Add » ]
0 Simulations
[ Remove [ New ]
Injection Definition
Name Example_Biotoxin
Mass Injection Rate (mg/min) 535
Concentration Zero Threshold (mg/L)
WQ Tolerance (mg/L)
Start Time
Stop Time
Description
8 hr iajecCion
Page 52 of 93
-------
Protecting Critical Facilities
Step 4.0.6.
Under "Node Set Definitions", click "Add". Next, click "Critical Facilities". To demonstrate we arbitrarily select the first
node, JUNCTION-0 and click "Add". Type in the "Name" of the Node Set as "Critical_Facility". Click "OK".
^_, TEVA-SPOT (Network_4_Critical_Facilities in C-Values [Main])
Map charts Tables
•»• Filtering | Manage | No Filter...
;xample_Biotoxin
Add
Node Set Definitions
Non-Zero Demand Junctions (162:
All Junctions (3323)
All Nodes (3358)
Node Injections
.^...
D EPANET Simulations
Modules
G Health Impacts Analysis | Edit
D TSO to Impacts Analysis | Edit
Qj Infrastructure Impacts Analysis | Edft
Sensor Placement
Scenario Sets
Name Critical_Facility
-Nodes—
Available
Selected
Junction
Junction
Junction
Junction
Junction
Junction
Junction
Junction
JUNCTION-1
JUNCTION-10
JUNCTION-100
JUNCTION-1000
JUNCnON-1001
JUNCTION-1002
JUNCTION-1003
JUHCTION-1004
Junction JUNCTION-0
Add*
[ Remove
Add From File Save Upstream Nodes
Sensor Placement
Page 53 of 93
-------
Protecting Critical Facilities
Step 4.0.7.
(All loaded "Scenario Sets" will run. So, when beginning a new exercise, it may be less confusing to first delete the previous
entries, e.g., the "Injection Definition". Therefore, if you have previously completed the Utility Facilities portion of this
tutorial, please "Delete" the "Injection Definitions" entry, the "Node Injections" entry, and the "Scenario Set" entry from the
boxes. Then, proceed with the Critical Facilities exercise.)
Select "Example Biotoxin" & "Critical Facilities". Click "Create" "Node Injections". Under "Scenario Sets", click "New". A
new Folder appears. (If you have previously run the Scenario Set for Utilities, those folders will also show in the "Node
Injections" and "Scenario Sets" list of folders.) Select Node Injections: "Example Biotoxin/Nodes: Critical_Facilities (1593)"
and the new "Scenario Set" folder. Click "Add" to add to "Scenario Sets" folder. Click "OK".
TEVA-SPOT {Network_4_CriticaLFacilities in C-Values [Main])
Insemble File Map Mode Edit About
Google MapsRoad Map *• Map Default
•* Filtering Manage I [NpF
I Edit BaseEnsemble Parameters
Add ]
Edit
Delete 1
Node Set Definitions
Non-Zero Demand Junctions (1621
All Junctions (3323)
Add J
Edit
[Create » |
Node Injections
[niection: Example_Biotoxin / Modes
Remove
Scenario Sets
* Injection: Example_Biotoxin / Nodes: Critica
1593 Simulations
Remove New
Note: The above box is truncated. Note too that all such boxes in TEVA-SPOT-GUI can be enlarged if needed by
clicking the left mouse button at a corner and dragging the box to enlarge.
Page 54 of 93
-------
Protecting Critical Facilities
Step 4.0.8.
Select "Ensemble" from the Menu and Click "Save".
ij TEVA-SPOT (Network_4_Critical_Facilities in C-Values
Ensemble File Map Mode Edit About
New... Ctrt+N
Load.., Ctrl+O
Save
Collection Management...
Ensemble Management...
Import EPANET ,inp File... Ctrl+I
Import EPANET ,msx File...
HIA Sensors
•/ Execution Control Ctrl+E
Available Executions
Input Generation
Estimated Time Remaining 00:00:00:00
Estimated Total Time 00:00:00:14
Estimated Completion Time 05/01/2013 08:58:17 AM
Execute I Terminate
@ Initialization: Population
/j Upstream St Downstream Nodes Edit
Q D EPANET Simulations
_ Health Impacts Analysis
n TSO to Impacts Analysis
Infrastructure Impacts Analysis
Page 55 of 93
-------
Protecting Critical Facilities
Step 4.0.9.
See Section 3.3 and Step 3.3.1 for setting up the Health Impacts Analysis Module. In the "Execution" dialogue box, under
"Modules", select "Edit" for "Health Impacts Analysis". Verify that entries match and close the "Edit" box.
j TEVA-SPOT {Network4_Critical_Facilities in C-Values [Mai
» Filtering | Manage | No Filtering
uu=c LaiLUiauuii rmDiuci.ei;t
Dose Calculation Method Irfgestion of Tap Water -
Ingestion Timing Model ATUS »
Ingestion Volume Model Fixed Volume »
Mean Volume [liters/ person/day) ' 1.41
D&seTmesriulas
Dose Thresholds | Edft |
Response Thresholds Edit
Number of worst-case fatality scenarios to save 0
Number of worst-case dosage scenarios to save 0
l_! Use one server per node
•/ Calculate Dose-Response
Contaminant Defaults Select...
Contaminant Type
Contaminant Type Chemical/Toxin v
Dose Response Method
Dose-Response calculation method Probtt •»
LD50 (mg/kg) 6.0E-6
Beta ;0.6711
Disease Progression Parameters
Latency Time (hrs) 1
Fatality Time (nrs) 46
U EPANET Simulations
O D Health Impacts Analysis | Edit |
0 ~'TSO to Impacts Analysis [ Edit J
O Q Infrastructure Impacts Analysis [ Edit |
Sensor Placement
i .Sensor Placement
Page 56 of 93
-------
Protecting Critical Facilities
Step 4.0.10.
See Section 3.4 and Step 3.4.1 for setting up the Infrastructure Impacts Analysis module. In the "Execution" dialogue box,
under "Modules", select "Edit" for "Infrastructure Impacts Analysis". Verify the entries and close the "Edit" box.
TEVA-SPOT (Network_4_Critical_Facilities in C-Va!ues [Main]) •
rnsemble File Map Mode hdit About
Map charts T
'"-: ;> '•;--"-•-•-•.--.--•--• - Map Default
•» Fiftering | Manage | | Ho Filtering
,z j Edit Infrastructure Impacts Analysis Parameters
O D Health Impacts Analysis | Edit |
0 Lli TSO to Impacts Analysis [ Edit j
O LI Infrastructure Impacts Analysis [ Edit
Sensor Placement
0 [31 Sensor Placement
Page 57 of 93
-------
Protecting Critical Facilities
Step 4.0.11.
In the "Execution" dialogue box, next to "Sensor Placement", click the EDIT button. The panel "Edit Sensor Design
Parameters" will appear. Starting with "Solvers" Click the "EDIT" button and then Click "Add Solver". Two options: "GRASP
(heuristic)" and "Lagrangian" will be able to be selected. Select "GRASP (heuristic)" and click "OK".
Edit Sensor Design Parameters
[&\ Choose Solvers
Solvers
Grasp (heuristic)
Sensor Design Parameters
Solvers Grasp (heuristic)
Sensor Set Size
Objectives
Dose Thresholds
Response Times
Constraints
Location Cateog
Costs
Detection Limits
Selection Metro
- m -
Add Solver
Cancel
Lagrangian: This solver can be used for very large models that cannot be easily solved using the GRASP
algorithm. Although the lagrangian's performance is less accurate (i.e., the sensor locations chosen will not
perform as well at reducing the chosen consequence measure) than that of the GRASP it requires less computer
memory.
PICO: Parallel Integer and Combinatorial Optimizer (PICO). PICO is a mixed integer linear program (MILP) based
optimization solver and is the only TEVA-SPOT solver that can provide guaranteed optimization. PICO is
currently not available in TEVA-SPOT-GUI but is available in the TEVA-SPOT toolkit. Please refer to the TEVA-
SPOT Toolkit Users Manual for more information [2] and "Sensor Network Design for Drinking Water
Contamination Warning Systems: A Compendium of Research Results and Case Studies Using TEVA-SPOT
Software" [13].
Page 58 of 93
-------
Protecting Critical Facilities
Step 4.0.12.
Click the EDIT button next to "Objectives" and the "Edit Objectives" box appears. From the list appearing under "Goal"
select each of the indicated "Objectives" one at a time and clicking "Add" with each one to populate the panel as shown.
Ensure the "Statistic" selected is "Mean".
J Edit Sensor Design Parameters
Sensor Design Parameters
Solvers Grasp (heuristic)
Sensor Set Sizes
Objectives
Dose Thresholds
Response Times (minutes
Constraints
Location Cateogories
Costs
Detection Limits
Selection Methods
3 Sens
Mean Estimated Population Killed (pk_mean)
Add Objective
Goal
Pppu|atipn Kjlied
After adding an objective, the objective can be removed by double clicking on the objective.
Constraints: This capability allows the user to define a side constraint on the objective. For example,
population killed (pk) could be constrained to minimize the extent of contamination (ec) objective subject to an
upper bound on extent of contamination. The user would insert an upper bound for ec in units of pipe feet.
The user is referred to the TEVA-SPOT Toolkit Users Manual for more information [2].
Page 59 of 93
-------
Protecting Critical Facilities
Step 4.0.13.
Click the EDIT button next to "Sensor Set Sizes" and the "Edit Sensor Set Sizes" box appears. In the "Sensor Set Size" box
type the number "5" and press ENTER on the keyboard. Next, type the number "15" and press ENTER on the keyboard. The
"0" sensor set size appears automatically. Click "OK".
-,
Edit Sensor Design Parameters
Sensor Design Parameters
Solvers
Sensor Set Sizes
Objectives
Dose Thresholds
Response Times (minutes
Constraints
Location Cateogories
Costs
Detection Limits
Selection Methods
Grasp (heuristic)
& Edit Sensor Set
Sensor Set Sizes
0
5
15
Sensor Set Size
OK
Cancel
uired
Edit
Edit
Edit
Edit
Edit
Edit
Edit
Edit
Edit
Edit
"Sensor Set Size" defines the number of sensors being considered in the monitoring strategy to protect the
critical facilities. Multiple "Sensor Set Sizes" (as shown above) can be considered together.
Page 60 of 93
-------
Protecting Critical Facilities
Step 4.0.14.
To "Edit" the "Location Categories" parameter requires the creation of an "Existing Sensor(s) Location File". Open Windows
"Notepad" or a suitable text editor and create the text file shown below.
Example_Network_4_Critlcal_Facility_Junction-0.txt - Notepad
File Edit Format View Help
name: Network_4_Fixed_Critical_Faci hty
easib le ALI
JUNCTION-0
Please name the file: "Example_Network_4_Critical_Facility_Junction-0.txt" and save the file to: C:\Program
Files\TEVA-SPOT\Client (or to a location from which you will be able to upload the file).
Notice the format of the contents of the file and the key words: #name:", "feasible ALL", and "fixed [Node ID]".
Word separations are SPACE or TAB.
TEVA-SPOT-GUI will use either the "#name:" value or the filename (if the "#name:" is not specified in the file) as
the location category name that is appended to the output results. The keyword "#name:" is ignored by the
TEVA-SPOT Toolkit [2].
For the key words "feasible ALL" TEVA-SPOT-GUI identifies all nodes in the model as being feasible for the
placement of sensors.
TEVA-SPOT-GUI executes the key word "fixed" and the associated location(s) that are to be fixed. Multiple node
IDs can be placed on the same line and/or multiple rows can be used. For each row, the keyword "fixed" must
precede the node ID(s).
Notice that capitalization is important: TEVA-SPOT-GUI will not recognize the fixed location if the node
"Junction-0" is NOT capitalized exactly as it is in the model.
A keyword NOT used here is "infeasible". The keyword "infeasible" specifies those node ID(s) that can NOT be
considered in sensor placement. The "AN" in "feasible AN" or "infeasible AN" can be replaced with node IDs. For
more information, refer to the TEVA-SPOT Toolkit Users Manual [2].
Page 61 of 93
-------
Protecting Critical Facilities
Step 4.0.15.
Click the "EDIT" button next to the "Location Categories". Then click "New" and "SP Format". The "Open" dialogue box will
appear allowing the upload of an "existing sensor(s)" text file.
Solvers
Sensor Set Sizes
Objectives
Dose Thresholds
Response Times (minuts
Constraints
Location Cateogones
Costs
Detection Limits
Selection Methods
5 Executions
neration
Initialization: Population [ Edit ]
. Upstream & Downstream Nodes I Edit I
EPANET Simulations
Health Impacts Analysis [ Edit ]
] ISO to Impacts Analysis [ Edit ]
Infrastructure Impacts Analysis | Edit |
Estimated Time Rem<
Estimated Total Time
Estimated Completion
I Execub
Look in:
5
311
Recent
Items
•
Desktop
r
Documents
•v
Computer
^
Network
I docs sp-xsd
, jar teva-base.xsd
i libs
, modules
i props
* contaminants.xml
* ensjemplate.xml
ensemble.xsd
• Example Network 4 Critical Facility Junction-O.txtH
modules.xsd
- runclientbat
4 \ II! | >
-ile name: Example_Network_4_Critical_Facility_Junction-O.M Open
^ilesoftype: /jl files -wj | Cancel |
This step uploads the critical facility location file. Fixing a sensor at each critical facility will ensure that each
sensor network design includes a sensor monitoring location at the critical facility. For example, if you
identified five critical facilities, you would start with five fixed sensors.
Location Categories: Refers to the various categories of feasible, infeasible, and fixed locations that TEVA-SPOT-
GUI will consider in the determination of the optimal placement of sensors. Choosing "All Locations" will result
in the best overall performance (i.e., selection of sensor locations which best reduce the chosen consequence
measure(s)) across the contamination scenarios ensemble. However, choosing "All Locations" may substantially
increase computation time and computer memory requirements to complete.
Page 62 of 93
-------
Protecting Critical Facilities
Step 4.0.16.
The "Sensor Design Parameters" consisting of "Response Times" (minutes) and "Detection Limits" (e.g., mg/L) can be
modified as described by Step 4.0.13. For information on these remaining "Sensor Design Parameters", users should consult
the TEVA-SPOT Toolkit Users Manual [2].
_^j Edit Sensor Design Parameters ^pfl
Sensor Design Parameters
Solvers
Sensor Set Sizes
Objectives
Dose Thresholds
Response Times (minutes)
Constraints
Location Cateogories
Costs
Detection Limits
Selection Methods
i_
i^ ua.r
Grasp (heuristic)
0, 5, 15
pk_mean
None defined - no pd goal: none required
None defined - using
0
None defined - Optional
Exa m p 1 e_N etwo rk_4_Criti ca l_Fa ci 1 ity_J u n cti o n- 0 .txt
H/A
None defined - using
0
N/A
Edit
| Edit |
[ Edit ]
Edit
[ Edit |
[ Edit |
i Edit |
Edit
[ Edit~|
[ Edit ]
O Create All Impacts
3 Sensor Designs will be generated
[ OK |
Cancel
"Response Time (minutes)": Refers to the delay in time required to confirm the detection of contamination and
stop further exposure and consequences, either through public notification and/or cessation of drinking water
service.
"Detection Limits": Refers to a minimum water concentration (e.g., mg/L or organisms/L) that will be required
for sensor detection at a particular sensor monitoring location.
"Create All Impacts" should not be checked if fast runtimes are needed. Checking the box results in the
generation of additional tso-2-impact files (located in the TEVA-SPOT-Database) for any objective not already
selected, which can be then used for Regret Analyses.
Regret Analysis (Menu item "Mode", choose "Switch to Regret Analysis") provides a means to select the best
sensor design among many. For example, multiple ensembles can be created, e.g., different contaminants,
release times, mass release rates, response delays, detection limits, and with each of these ensembles sensor
designs are created. The impact files and sensor designs can be loaded into TEVA-SPOT-GUI Regret Analysis and
evaluated to compute a matrix of consequences. Every sensor design chosen is evaluated against every impact.
The performance measures (Preferred Overall Design and Preferred Design to Minimize Maximum Deviation) are
used to compute the difference between the best design (which is the design created for the particular
consequence measure) and all the other designs. There are no parameters to enter like there is in the
"Ensemble Analysis Mode". Please refer to the TEVA-SPOT-GUI Users Manual for more information [2].
Page 63 of 93
-------
Protecting Critical Facilities
Step 4.0.17.
Verify entries into the "Edit Sensor Design Parameters" dialogue box and click "OK". Modules are now ready for execution.
Check the "EPANET Simulations" scenario, the "Health Impacts Analysis" module, and the "Infrastructure Impacts Analysis".
Click "Execute". Execution could take some time (hours or even days) to complete. The execution of all the modules
concurrently on the Example Network_4 demonstrated here takes approximately 40 minutes using a 64-bit, Windows 7,
quad-core, 2.6 gigahertz laptop computer with 32 gigabytes of random access memory.
Once complete, the exclamation points in red circles should become check marks in green circles. (Troubleshooting tip: If
the marks fail to change upon completion of the execution, go to the "Ensemble" menu and select "Load" to reload the
data.)
Execution of "Sensor Placement". First, make sure that the completed executions are unchecked so that they do not run
again. Now click the "Sensor Placement" box with a check mark and click EXECUTE. Two processes will be executed. First,
the precursor to sensor placement will run called "tso-2-impacts". Next sensor placement will run. Once both processes are
complete the screen below will be displayed. In order to view the results, minimize the "Execution" dialogue box.
_ IIVA-STOI lNri«i»K.4.CriLn»U null Lies ill (.-Value
"«!'
GtiHuttan
HvWuMe txecuoons status
Input GenerolKjri |
®£stimM«l Tims RerMinmg
.MHlBatWl. ttip.rf.iHn | *<* , f^tni rad Tilrll.
S Upstream & Uownstreatn Modes | tid* | CsthnBteJ ComptettDn Tune
ScHtanOs
^rcPANtrsrm*^ w
& HrallhlrlTads Analysis
Q I •-.(] to rtnpsrt'. Annly'jl
-Q InfrJlsrrurtur* InpnrtE MatyCK
Sensw Pla«me[rt
Precursor to sensor placement - "tso-2-impacts" will be run for each combination of objective, response time,
and detection limit.
For sensor placements, the total number includes the null or zero sensors case.
Notice in the screen shot above that the "ISO to Impacts Analysis" module under the "Modules" section is
grayed out and RED. This module for running "ISO to Impacts Analysis" is only for generating stand-alone
impact files that could then be used with the TEVA-SPOT Toolkit or for other purposes. "TSO-to-lmpact"
analyses are also performed within the "Sensor Placement" module.
Page 64 of 93
-------
Protecting Critical Facilities
Step 4.0.18.
Retrieving sensor placement results: 5 sensor location design. Click MAP tab, next choose the "Sograsp-NSOOS-
Obpk_mean...Sensor Locations".
^, TEVA-SPOT (Network_4_Critlcal_Fadlities in C-Values [Main])
Ensemble File Map Mode Edit About
Map Charts | Tabtesj
._:; • ::_ ,- : _•_: :-_-•: jj
;j» Filtering | Manage ; [Sensor Pla
Sensor Locations: SOgrasp-NS005-OBpk_mean-RTOOO-LCExample_Network_4_Critical_Facility_Junction-0.txt-...
5 Sensors
Response Time: 0 minutes
Objectives: Mean Estimated Population Killed
Constraints: None
Thematic Legend
Existing
C Ignored
Selected
Notice the naming convention for sensor network designs:
• SOgrasp: Sensor Objective-grasp
• NS005: 5 locations sensor network design
• OBpk: Objective-population killed (pk)
• Mean: Mean objective statistic.
• RTOOO: Response time of zero minutes
• LC: Location Category(s). The name that appears after LC is the file name of the uploaded "SP
formatted" text file
• GREEN STARS designate optimally selected locations. BLUE STARS designate "fixed" location(s). RED
STARS would designate "infeasible" locations (if used, not shown here).
Notice that scrolling within the MAP, TABLE, and CHART menus of TEVA-SPOT-GUI can be tedious due to the
many output products generated. Use the "MANAGE" feature within each tab to create filtered sets of results.
Results can be filtered for either Health Impacts or Sensor Placement. Please refer to the TEVA-SPOT-GUI Users
Manual for more information [2].
Page 65 of 93
-------
Protecting Critical Facilities
Step 4.0.19.
Retrieving sensor placement results. Next, choose the map below (NS = 5 and Detected Events, Links and Nodes) from under
the MAP TAB.
Ensemble File Map Mode Edit About
Map charts' Tables!
Soogle Maps Road ... » [Map [SC^rasp-f4SOD5-oapk_mean-RTOQQ-LC£xample_Netvjork_4_Critical_Fa<:ilityJunction-O.M-SMstd]: Detected Events (links and nod..
Sensor Event Detection
SOgrasp-NS005-OBpk_niean-RTOOO-LCExample_Network_4_Critical_FacilitY_Junction-0.txt-SMstd (Grasp (heuristic))
5 Sensors
Response Time: 0 minutes
Objectives: Mean Estimated Population Killed
Constraints: None
Thematic Legend
Detected Events
— • JUNCTION-0
JUNCTION-1019
— •jlJNCTION-1609
JUNCTION-2Q30
JUNCTION-2284
— NO_DFTECTION
— UNKNOWN
^C UNCTIOH-Q
UNCTION-1019
^ UNCTION-1609
UNCT!ON-2030
JUNCTION-2284
This MAP provides sensor event detection information associated with each sensor location.
RED STAR shows the location of the Critical Facility. Multiple critical facilities will be displayed using different
colors.
Although it is difficult to see each COLORED STAR, these sensor locations are depicted in Step 4.0.18. Associated
with each STAR (Sensor Location) are similarly colored pipes and nodes. The correspondingly colored pipes and
nodes represent the event detection information associated with the particular sensor location.
Right clicking on the legend (see RED circles), "Detected Events" or "Sensors" allows the removal from view of
selected information. The data is only removed from view.
Note: Event Detection Maps are also available for "Nodes Only" and "Pipes Only".
Page 66 of 93
-------
Protecting Critical Facilities
Step 4.0.20.
Retrieving sensor placement performance results: TRADE-OFF ANALYSIS DATA & SENSOR PLACEMENT SUMMARY tables.
HHI^^d
' hto ir^rt Page Layout
Pa;
, *Cut Arial
^ Copy
J Format Painter B z
H -\
sp_trade-off.xis [Compatibility Mode] - Microsoft Excel
Formulas Data Review View
;9
E -| &
Clipboard ra ]| Font
Al ^^ ^
1
2
3
4
A B
Name
SOgrasp-NS01 5-
OBpkjnean-RTOOO-
LCExample_Network_4_Criti
cal_Facility_Junction-0.txt-
SMstd
SOgrasp-NSOQB-
OBpk_mean-RTOOO-
.CExample Network 4 Criti
cal Facility Junction-0.txt-
SMstd
SOgrasp-NSOGO-
OBpk_mean-RTOQO-
.CExample Network 4 Criti
cal Facility Junction-0.txt-
SMstd
A" A'
-A-
fa
* • m
*H
^ &
Afignm
Acrobat
5? Wrap Text
^ Merge & Center -
errt a I
General
$ - % • ||TdS ilS
Number &
S si
Conditional Format as
Formatting - Table •
Normal
Neutral
Si!
Bad
Calculation
tes
Name
C D E F
Solver
Grasp
(heuristic)
Grasp
(heurist c)
Grasp
(heuristic)
Number of
Sensors
15
s
0
Objectwe
Mean
Estimated
Population
Killed
Mean
Estimated
Population
Killed
Mean
Estimated
Population
Killed
Objective ID
pk_meari
pk mean
pk mean
Response
Time
0
0
0
G H 1 J K
Constraints
None
None
None
Location Categories
Example_Network_4_Cri
tical_Facility_Junction-
O.txt
Example Network 4 Cri
Sical Facility Junction-
0-txt
Example Network 4 Cri
tical_Facility_Junction-
O.txt
Costs
N/A
N/A
N/A
Detection
Limits
0
0
0
L
Selection
Method
Standard
Standard
Standard
Impact
45.040298
234938
32985 699
Table Above: "Trade-Off Analysis", EXPORTED TO Micosoft EXCEL and then formatted.
;nr i
®" HO™ insert
1. - A cut
Paste
: pbos. i
AB
A
• sp_summaryj
-------
Protecting Critical Facilities
Step 4.0.21.
Retrieving sensor placement results: "Impacts and Detection Time" tables. Click on the TABLE TAB. Scroll to the table titled:
"SO...NS005-...OBpk... Impacts and Detection Time".
TEVA-SPOT (Network 4 Critical Facilities in C-Va!ues [Main]) »
Map . Charts Tables
Table [SOgrasp-NS005-OBpk_mean-RTOOO-LCExBmple_Netv^ork_4_CriticaLFacilit¥Junction-0.txt-SMstd]: Impacts and Detection Time v\ Filtering [ Manage No Filtering V
Impacts and Detection Time: SOgrasp-NS005-OBpk_mean-RTOOO-LCExample_Network_4_Critical_Facility_Junction-0.txt-S,..
5 Sensors
Response Time: 0 minutes
Objectives: Mean Estimated Population Killed
Constraints: None
Injection
UUNCT10N-Q
JUNCTION-
IJUNCTION- 0
JUNCTION- 30
JUNCTION- 001
JUNCTION- 002
•JUNCTION- 003
JUNCTION- 004
JUNCTION- 005
JUNCTION- 006
)UNCTK>N- M7
JUNCTION- 008
•JUNCTION- 009
JUNCTION- 010
JUNCTION- Oil
Detected by
JUNCTION-0
JUNCTION- 1019
JUNCnON-1019
JUNCTION- 1609
JUNCTION- 101 9
JUNCTION- 1019
JUNCnON-1019
JUNCTION-2030
JUNCTION-1019
JUNCTION-2030
JUNCTION-1019
JUNCTION-2030
JUNCTION-2030
JUNCTION-2030
JUN"ION-2030
JUNCTION- 013 JUNCTION-2030
JUNCTION- 014
JUNCTION- 015
JUNCTION- 016
JUNCTION-2030
JUNCTION-2030
JUNCTION-2030
Detection Time
0
60
240
60
60
60
60
240
60
180
60
80
80
80
80
80
80
240
300
Impacts
0
20.199
49.412
22.018
6.253
2.815
4.169
21.844
4.669
3.999
3.615
6.058
11.382
14.398
22.242
30.791
5.45
58.575
85.368
Note that this table provides information on the 5-sensor network design that was optimized for minimizing the
mean population killed objective. The four columns are:
• Column 1 "Injection" provides the injection node
• Column 2 "Detected by" details which sensor (node ID) detected in the event in Column #1
• Column 3 "Detection Time" details the time (minutes) before the scenario (Column #1) was detected by
the sensor at the node ID specified in Column #2
• Column 4 "Impacts" provides the number of people killed (pk) up to the detection and/ therefore/
cessation of public health consequences for the scenario specified in Column #1
Note that this table has been sorted in TEVA-SPOT-GUI by clicking on the first column/ "Injection" to order
alphabetically. Therefore/ note that the Critical Facility/ JUNCTION-0/ has zero impacts. In TEVA-SPOT/ fixing a
sensor at each Critical Facility ensures there will be no impacts at the facility.
Page 68 of 93
-------
Protecting Critical Facilities
Step 4.0.22.
Retrieving sensor placement results: "Sensor Ranking" tables.
TEVA-SPOT (Network 4 Critical Facilities in C-Values [Main]) ~«
Ensemble File Table Mode Edit About
Map Charts Tables
Table [SOgrasp-NS005-OBpk_mean-RTOOO-LC£xample_Network_4_Critical_Facility_Junction-0.txt-SMstd]: Sensor Ranking ^ Filtering Manage No Filterin
Sensor Ranking: SOgrasp-NS005-OBpk_mean-RTOOO-LCExample_Network_4_Critical_Facility_Junction-0
5 Sensors
Response Time: 0 minutes
Objectives: Mean Estimated Population Killed
Constraints: None
Sensor
!No Sensors
JUNCTTON-2030
JUNCTIOM-2284
JLINCTIOH-1609
JUNCTION-1019
JUNCTION-0
Cumulate Impact Incremental Reduction
32,985.711jO.O%
1,339.08995.9%
663.362i50.5%
422.33236.3%
234.95J44.4%
234.938|0.0%
Cumulative Reduction
0.0%
95.9%
98.0%
98.7%
99.3%
99.3%
txt-
Sensor Ranking tables provide sensor-by-sensor performance information. If no sensors were placed, the
cumulative impact (Column #2) would be approximately 33,000 people. The majority of benefit provided occurs
due to the sensor at location JUNCTION-2030.
Note that while 4 of the 5 sensors provided measurable benefit, the fixed sensor at JUNCTION-0 did not seem to
provide any meaningful benefit. This, of course, is just an illustrative example of a critical facility.
Step 4.0.23.
Retrieving sensor placement results: OTHER TABLES. See "All Sensors" table and "Sensor Counts" table.
£ i TEVA-SPOT (Network 4 Critical Facilities in C-Values [Main]) J
Ensemble File Table Mode Edit About
' Map | Charts ! Tables
Table [SensorPlacem
A
SOgra
SQqra
SOgra
sntSummary]: All Sensors _^^^^^^^^^^^^^^^^^^^^^^^^^^^^-
[SOgrasp-NSno5-OEpk_rnean-RTuOO-LCExample_Netvvork_'!r_CriticaI_Facility_Junction-0.txt-SMstd]: Impacts and Detection Time >
[SOgrasp-NS005-OBpk_mean-RTOOO-LCExarnple_Neti,vork_4_Critical_Facility_Jund:ion-0.txt-SMstd]: Sensor Ranking
[SOgrasp-NS015-OBpk_mean-RTOOO-LCExample_NetVsfork_4_Crittcal_Facility_Jundion-0.txt-SMstd]: Imparts and Detection Time
[SQgrasp-NSQ15-GBpk mean-RTOQQ-LCExample Network 4 Critical Facility Junction-Q.txt-SMstd]: Sensor Ranking
[SensorPlacementSummary]: All Sensors
[SensorPlacementSummary]: Sensor Counts
[SensorPlacementSumrnary]: Sensor Placement Summary
[SensorPlacementSummary]: Trade-off Analysis Data
— .
— "
The table "All Sensors" provides node IDs for all sensor designs created.
The table "Sensor Counts" provides a two-column table: Column #1 (Node ID) and Column #2 (Times Selected). This table
provides a count how many times each node was selected in a sensor design.
Page 69 of 93
-------
5.0 SUPPLEMENTAL GUIDANCE AND INFORMATION FOR USING TEVA-SPOT-GUI.
We provide some additional discussion to help the user setup, run, analyze, and interpret results from
TEVA-SPOT-GUI. This section is separated into five subsections as follows:
5.1 EPANET Simulation Parameters
5.2 TEVA-SPOT-GUI Simulation Parameters
5.3 Contamination Scenario
5.4 Quality Assurance Features within TEVA-SPOT-GUI
5.5 Using Google Maps within TEVA-SPOT-GUI
5.6 Trouble-shooting TEVA-SPOT-GUI Applications
5.1 EPANET Simulation Parameters:
The EPANET simulation parameters are generally specified in the Menu item "EDIT" and then by
selecting "Ensemble Options". The simulation length (hrs or min) should be sufficiently long for
consequences to stabilize. This is generally a week (168 hrs) or longer.
The water quality time step should be as short as reasonably achievable given computational resources
and available time. Generally a 1 minute water quality time step is preferred. If injection duration
shorter than 1 minute is needed, a shorter water quality time step may be needed. See Figure 5.1.1
below. An error message is given when the user tries to define a contaminant injection shorter than the
water quality time step. The error message instructs the user as to how the injection mass rate should
be adjusted to accommodate the longer water quality time step.
Shorter water quality time steps will increase computation time and resources (memory requirements).
Figure 5.1.1. Injection Definition Error Message.
|_dbj Edit BaseEnsemble Parameters
Import... Export...
Injection Definitions
Node Set Definitions
Non-Zero Demand Junctions (1621]
All Junctions (3323)
All Nodes (3358)
Critical_Facility (1593)
Add
Edit
Add
h : -
' Delete 1
l^J Injection Definition r" ^ ^1
Name
Mass Injection Rate (mg/min)
Concentration Zero Threshold (mg/L)
WQ Tolerance (mg/L)
Start Time
Stop Time
Status:
ERROR: The injection interval param
the water quality timestep (1 minute)
to approximately 500.000 mg/min anc
should be changed to 0 minutes- 1 m
Description
xyz
0 [seconds •*•
30 [seconds •*•
tersdo not align with
In order to
on rate should be changed
the injection interval
| OK ] | Cancel |
--
22
I±J ,. Scenario Set
1593 Simulations
Remove [ New |
The hydraulic time step must be less than or equal to the reporting interval, otherwise an error message
will be given. Shorter hydraulic time steps will increase computation time.
The hydraulic time step must be less than or equal to the reporting time step or TEVA-SPOT-GUI displays
an error message in the "Ensemble Options" under the "Edit" menu item. Similarly, the hydraulic time
step must also be less than or equal to the pattern time step or TEVA-SPOT-GUI will display an error
Page 70 of 93
-------
message. The error message will appear in the "Ensemble Options" which is under the "Edit" menu
item. Additionally, as part of the error message indicating that the hydraulic time step is not less than or
equal to the pattern time step, TEVA-SPOT-GUI will report what the pattern time step currently is in
parentheses.
Generally a reporting time step of 1 hr is used. A warning will be given if the hydraulic/reporting interval
parameters generate an exceedingly large number of time steps. Proceeding without change can result
in long computation time and/or exhaustion of computer resources, including processing capabilities,
available memory, and storage (hard drive) space.
Within the "Ensemble Options" box is a "Water Quality Parameters" section. Here the box "Average
concentration over reporting interval" should be checked. Un-checking would result in the
contamination concentration occurring at the end of the time step being reported. The "WQ Tolerance
Value" refers to the water quality tolerance, specified in EPANET. This parameter specifies the smallest
change in water quality that will result in a new parcel of water being created in a pipe. The default
setting in EPANET is 0.01 mg/L for chemicals (Rossman 2000). The water quality tolerance determines
whether the quality of one parcel of water is the same as that of another parcel. In EPANET, water flow
is discretized into finite volumes of water. The volume of water in a parcel is assumed to be
homogeneous. The water quality tolerance specifies the acceptable tolerance of each parcel. For a
chemical contaminant the value could be the detection limit of the analytical method used to measure
the chemical contaminant, adjusted by a suitable safety factor (Rossman 2000). Generally, lowering the
water quality tolerance will provide increased accuracy of the water quality modeling results but will
also increase computation time. It is recommended that a value of 0.001 mg/L be used. Finally, within
the "Water Quality Parameters" section the user can globally set the mixing model for the system's
tanks. The default is "Use Scenario", which corresponds to using what is specified in the INP file for the
water distribution system.
The water quality tolerance can also be specified in the INJECTION DEFINITION. Inserting a WQ
Tolerance (mg/L) value here will override the WQ Tolerance specified in the EDIT/ Ensemble Options
box; however no value need be specified with the INJECTION DEFINITION.
Clicking the faucet icon while in the MAP tab will bring up the majority of the INP file in a side panel on
the right of the screen. See Figure 5.1.2 below. The "faucet" functionality can be used to explore the
water system model. While in the default map, different network attributes can be highlighted. Within
any MAP network, users can zoom to features. Zooming out can be accomplished by right clicking the
mouse on the MAP and then choosing the appropriate zoom out level.
Page 71 of 93
-------
Figure 5.1.2. Accessing EPANET Components: Using the Faucet Icon.
TFVA-SPOT (Network 4.UUrty FaoiitmmC-Valu« EMaintt
his Map Mod* (flit About
WMCTION-1W5
.'.'.CTIO'J H-jS
JUUCTON 1007
hmclwil 3UHC1KJM-100B
Kincton )IM IICHJ-IIKM
lUI*-nO!JH 0]
EPANET network
attributes
The START TIME of the model is specified in the *.inpfile and can be accessed via the faucet icon. Under
"Data Type" scroll to "Options", and next "Times" (see Figure 5.1.3).
Page 72 of 93
-------
Figure 5.1.3. Accessing EPANET Component: Start Time.
TEVA-SPOI (Hetwort 4 UHity.Fariliwsi in C-Valuss (Main!)
c Map Mode Edit Abwjl
Map Cham taW^s
(^ Xffl MOB* Mp
Notice Start Time is "0" which
corresponds to 12:00 AM given
Start Time of model.
The START TIME specifies the start of the simulation. TEVA-SPOT-GUI recognizes the START TIME as
local time. For the example Network_4, the START TIME specified is 12:00 AM. If a contaminant is to be
injected at 6:00 AM, then the user would need to specify within the INJECTION DEFINITION the START
TIME for the injection, which in this case would be 6 hours, i.e., NOT zero hours. Similarly, an injection
occurring at 6 PM local time would be simulated by defining the START TIME for the injection to be 18
hours, i.e., 12 hours later than the 6 AM injection. The user should be careful that if the START TIME for
the injection is much later than the start time of the model the total simulation length may need to be
increased to be sure that consequences still stabilize. There is not a minimum recommended time
between the simulation start time and the injection start time.
The EPANET 2.00.12 hydraulic adjustment parameters are specified within the "Ensemble Options" box.
These parameters can be adjusted as advised in EPANET 2.00.12 to decrease simulation time and/or
help eliminate hydraulic errors and warnings. EPANET's reaction and decay parameters are also
specified in the "Ensemble Options" box.
EPANET multispecies (EPANET-MSX) reaction parameters are specified in an .MSXfile, which is uploaded
into TEVA-SPOT-GUI immediately after loading of the network INP file [2], [3], and [12].
Page 73 of 93
-------
It is highly recommended that initial model testing be performed in EPANETand then once the
preferred hydraulic and water quality parameters are identified, tested, and evaluated the revised
model be incorporated into TEVA-SPOT-GUI for analysis.
5.2TEVA-SPOT-GUI Simulation Parameters.
Most of the TEVA-SPOT-GUI simulation parameters are specified in each execution module. Others are
specified within the Ensemble Options box. Most of the TEVA-SPOT-GUI simulation parameters were
discussed with the discussion of each module. Here, we provide an overview of some additional TEVA-
SPOT-GUI simulation parameters.
Within the "Ensemble Options" box under "Simulation Parameters" are two features: (1) check box
indicating whether all servers will be used and (2) drop down box for selecting the number of
simultaneous sensor network designs that will be executed within the Sensor Placement module. For
the check box, this generally should always be checked. For the number of simultaneous "SP
executions", some thought is warranted depending on the size (# of nodes) of the model, number of
injection scenarios evaluated, and number of feasible sensor locations being used. Generally, if the
number of injection scenarios is 12,000 or less and the number of feasible nodes for sensor placement
(i.e., nodes in the model) is also 12,000 or less, then the default of "MAX" should suffice. This means
that TEVA-SPOT-GUI has allocated 2 GB of computer memory for each sensor placement. If, for
example, one of these two numbers double then a smaller number of simultaneous "SP executions"
should possibly be chosen. Specifically, the user should select a number that corresponds to allocating
approximately 10 GB of computer memory per sensor network design.
Also within the "Ensemble Options" box but under the "QA Parameters" are two features: (1) check box
for performing "mass balance analysis" and (2) "SELECT" button for selecting or uploading a list of node
ID(s) representing nodes that will be eliminated from the particular "Node Set Definition" and,
therefore, eliminated from being simulated or evaluated. These parameters will be discussed in the
Quality Assurance section, 5.4.
5.3 Contamination Scenario:
Contamination of a water distribution system could occur through the dumping of toxic chemicals,
pesticides, biological organisms, toxins, or radioactivity into a water tank or through the inadvertent or
purposeful injection of contaminants. Contaminants can be injected into a water distribution system
from a service connection by reversing the flow of water for the connection using water pressure
greater than that in the network. The factors that define a contamination scenario are (1) the
contaminant and its behavior in a water distribution system, (2) the quantity of contaminant released,
(3) the start time of the contaminant injection, and (4) the duration of the contaminant injection. For
factor (1) the TEVA-SPOT-GUI user decides whether to consider contaminant loss (e.g., using a first order
decay/loss rate coefficient or analyzing multi-species interactions by using the EPANET-MSX capability
Page 74 of 93
-------
within TEVA-SPOT-GUI. For factors (2) through (4), the TEVA-SPOT-GUI user specifies the contamination
scenario parameters of the mass injection rate and the start time and stop time of the contaminant
injection. Simple contaminant loss/decay can be set in TEVA-SPOT-GUI under the Edit menu bar and
then select Ensemble Options and specify the appropriate rate coefficient under bulk or wall rate
coefficients boxes. For more information on determining and entering the most appropriate rate
coefficients refer to the EPANET Users Manual [1] and [9]. For using the EPANET-MSX capability in
TEVA-SPOT-GUI, the user must first load the .MSXfile after importing the EPANET (inp file) model.
Please refer to the EPANET-MSX User Manual for more information [3]. Recall that the mass injection
rate and the start and stop times for a contamination scenario are specified within the Injection
Definition which is under the EPANET Simulations module.
A contamination scenario together with the possible locations within the water distribution system
where the contaminant could be injected describes a threat. Recall that the contamination injection
locations are specified as a Node Set Definition within the EPANET Simulations Module.
Users of TEVA-SPOT should determine the relevant threat or threats for their application (e.g.,
prioritizing resources and efforts to support facility hardening or designing a sensor monitoring network
to support SRS efforts). Unfortunately, the interpretation of threat information is imperfect, and there
is no single definition of a threat to the water sector. Thus, judgement and consideration of local
circumstances must be considered when deciding on the threat. For example, a utility concerned
primarily with local or insider threats might consider contaminants that could be stolen or purchased in
large quantities from local sources. On the other hand, utilities concerned about international terrorism
threats might consider exotic contaminants that are more difficult to acquire, but would produce more
severe consequences. Resources are available to help in this decision-making process. The Water ISAC
does track and report potential threats to the water sector through their secure portal. Clark and Hakim
[14] provide some supporting information, including an overview of research related to understanding
and characterizing intentional threats to water systems.
In Sections 5.3.1 through 5.3.4 we provide information related to each of the factors that define a
contamination scenario. This information is provided to help users design a contamination scenario for
use in TEVA-SPOT to meet their needs in the most computationally efficient manner. In Section 5.3.5 we
provide a table with three example contamination scenarios for possible use in TEVA-SPOT. We also
provide some supporting information about how TEVA-SPOT simulates contaminant releases into tanks
and pressurized back flow of contaminants at nodes. Finally, we provide some thoughts on designing
robust sensor monitoring networks using TEVA-SPOT.
5.3.1 Contaminant:
The potential for adverse public health consequences increases as contaminant toxicity increases.
Potential public health consequences for highly toxic contaminants can vary substantially between
networks but contaminants with lower toxicities potential public health consequences have been shown
to be similar for most networks [7], [8], [9]. For contaminants with high toxicities, public health
consequences increase with network population, particularly for scenarios with injection locations that
Page 75 of 93
-------
produce consequences near the upper end of the distribution [7], [8]. For less toxic contaminants
potential public health consequences have been shown to not be particularly sensitive to network
population, even across networks that vary by more than a factor of 100 in population size [7], [8]. This
is because low toxicity contaminants tend to produce consequences that are localized, and which are
limited by the mass of the contaminant available rather than by the population served by a utility. Davis
et al. [10] show that sensor monitoring networks should be designed for the most toxic contaminant
likely to be of concern.
Contaminant decay can have a significant influence on the magnitude of estimated consequences
resulting from a contamination event [9]. Davis et al. [9] provide a table (Table III) listing various
potential water contaminants, along with literature values of first-order decay or loss coefficients, K.
The contaminants, decay mechanisms, and environmental conditions in the reference were selected to
illustrate a range of K values for a range of possible water contaminants [9].
Research has shown that sensor monitoring networks that are designed assuming no contaminant decay
or loss will likely perform better (i.e., better able to reduce public health consequences) than if decay or
loss is assumed to occur [10]. Therefore, it is recommended that no decay or loss of contaminant be
considered for most applications. The use of a simple decay/loss approach in TEVA-SPOT-GUI does not
increase computational times; however, the use of the EPANET-MSX functionality can increase
computational time significantly.
Davis et al. define adverse public health consequences (impacts) as the size of the population receiving
an ingestion dose of contaminant above some level [9]. Such an approach avoids any reference to a
specific contaminant, but provides results that can be related to specific contaminants if desired [9].
Use of this approach in TEVA-SPOT-GUI is simple and provides the ability to examine a wide-range of
potential contaminants at the same time and in a single ensemble. Estimated impacts have less
uncertainty than estimated number of fatalities or illnesses [9]. The approach is more flexible than use
of a particular contaminant since a range of contaminants is considered. Overall, the approach provides
a convenient and computationally efficient means for estimating and comparing potential adverse
public health consequences of a wide range of contamination events in water distribution systems.
5.3.2 Contaminant Quantity:
Adverse public health consequences can be widespread given a sufficiently large mass injection [7], [8].
Consequences can be expected to be localized for a small injection mass [7], [8]. For chemicals and
toxins, the injection mass in TEVA-SPOT is specified in units of milligrams. For biological contaminants
the injection quantity must be based on the number of organisms injected instead of the mass of the
contaminant used for chemicals or toxins. In Davis et al. [9], the authors show how to convert a range of
dose levels developed for a chemical or toxin to a comparable range for biological contaminants.
Solubility and organoleptic properties of a particular contaminant (taste, odor, color) should be
considered in determining the injection mass that should be used. An evaluation of the actual
availability of the contaminant being considered for injection should also be performed.
Page 76 of 93
-------
The total mass of a contaminant that can be reasonably obtained depends on the specific contaminant,
and is also related to the threat information discussed under 5.3.1. Generally speaking, toxic chemicals
might be available in quantities of several thousand kilograms (Kg), while more exotic, yet highly toxic,
contaminants will generally only be available in much smaller quantities (e.g., less than 10 Kg for tightly
controlled materials).
Unless localized consequences are of concern, the injection mass for a sensor monitoring network
design should be as large as feasible with no losses assumed to occur in the system [10]. Sensor
monitoring networks generally perform best when the actual injection mass is smaller than that used in
the sensor network design [10].
In Davis et al. [9], the authors outline a process for determining a simple contaminant mass bound on
adverse consequences. The equations in Davis et al. [9] can be used to help determine a suitable, lower
limit on mass injection rate for a particular contaminant. The mass of contaminant ingested can be
related to the mass of contaminant injected as shown in Equation (1):
M, = q*M/Q (1)
Where M, is the mass of contaminant ingested, q is a constant per capita water ingestion rate, M is the
mass of contaminant injected, and Q is a constant per capita water usage rate. U. S. EPA provides an
estimate of the per capita water ingestion rate (1 liter per person per day (L/d)) for the United States
[15]. In 2000, the United States Geological Survey (U.S.G.S) provided an estimate of water use (677.6
L/d) in the United States [16]. The maximum number of people (N) who can receive an ingested dose d
equals the mass ingested divided by the dose, shown in Equation (2):
N = q*M/Qd (2)
N should be set to the population served as represented by the water distribution system model. The
dose d is the dose resulting from the tap water ingestion of the target contaminant. An individual's
target dose for a particular contaminant can be determined from available dose response data for the
contaminant along with the individual's body mass. For instance, the LD dose at which 50% of the
population who receives such a dose dies defines the LD50 for a contaminant. LD50 values for chemicals
and toxins are usually expressed in units of mg of contaminant per Kg of body mass. Using contaminant-
specific LD50 information and an average body mass for the population (N), a target dose for a
contaminant can be determined as shown in Equation (3):
Target Dose for Contaminant = LD50 (mg/Kg) * Body Mass (Kg) (3)
Equation (2) can be solved for M to give:
M = (d*N*Q)/q (4)
As an example, consider a target contaminant's LD50is 0.001 mg/Kg, the population served by the water
system is 1,000,000, average body mass of the population served is 70 Kg, Q is 677.6 L/d, and q is 1 L/d
Page 77 of 93
-------
results in a target contaminant injection mass of 47 Kg. The estimate of 47 Kg would be the minimum
injection quantity needed since it assumes:
• Injection of the contaminant is at a location (node in the model) with at least the given
number of people downstream of the injection location. In other words, assumes the
injection location is a very high consequence injection location.
• Everyone ingests tap water with the same concentration of contaminant.
• There are no losses in the system.
This approach would be expected to work best for less toxic contaminants in which high consequence
injection locations are likely easiest to find and, generally, consequences are small.
5.3.3 Contaminant Injection Start Time:
The sensitivity of the performance of a water quality sensor monitoring network to the time of
contaminant injection depends on the network. The performance can be affected substantially if a
contaminant injection occurs at a time different from that used for the design [10]. Although an
example contamination scenario described here uses an injection time of 12:00 AM, research has shown
that a water quality sensor monitoring network based on contaminant injection at 6:00 AM may be least
sensitive to changes in injection time [10].
5.3.4 Contaminant Injection Duration:
Generally speaking, considering a large ensemble of injection locations, adverse public health
consequences resulting from 24-h injections are larger than those for 1-h injections because there is an
increased opportunity for exposure. The difference between the magnitude of consequences from 24-h
and 1-hr injections tends to increase with the decreasing toxicity of the contaminant, especially as the
importance of the injection location decreases [8]. However, there are considerable differences among
water distribution system networks [8]. The magnitude of adverse consequences which result for a
given contamination scenario is due to the complex relationship between the contaminant's toxicity,
mass injected, and injection rate as well as the hydraulic and water quality dynamics associated with a
particular injection location.
Overall, the performance of a sensor network design based on a 24-h injection appears less sensitive to
decreases in injection duration than the performance of a 1-h design to duration increases [10].
However, relative sensitivity to contaminant injection duration is network dependent [10].
Once the injection duration, D, is determined, the target contaminant mass derived in Section 5.3.2 can
be used to determine a corresponding mass release rate, MR (mg/min), which is used by TEVA-SPOT-
GUI. The mass release rate is shown in Equation (5):
MR = (M * 1 x 10s mg/Kg)/(60 min/hr * D (min)) (5)
5.3.5 Supporting Information:
Page 78 of 93
-------
In TEVA-SPOT, the direct addition of a contaminant (mass) to a tank is simulated utilizing a modified
version of EPANET. The current version of EPANET does not allow the direct addition of contaminant
mass to a tank.
EPANET simulates the pressurized back flow of a contaminant at a junction or node in the water
distribution system network model. A node in a network model generally represents multiple service
connections. The pressurized back flow of contaminants will contaminate water used at the injection
node and also any water flowing downstream from the node. EPANET, and therefore TEVA-SPOT,
models pressurized backflow or contaminant injection as a mass injected per unit time (mg/min).
EPANET does not have the ability to model volume additions to pressurized pipe flow. As a result, the
assumption is that any volume changes are insignificant.
If a water distribution system model is sufficiently detailed that all non-zero demand (NZD) nodes
represent terminal nodes in the water distribution system or individual service connections, people
located at these nodes will only receive water from the water distribution system. As a result, injection
of a contaminant at such nodes will only affect people at the injection node. Using TEVA-SPOT to
simulate contaminant injections at terminal nodes or nodes that only receive water will likely under-
predict consequences for two reasons [8]. First, not everyone at such nodes who would likely ingest
contaminated water would be exposed due to the removal of all contaminated water by demand at the
node before ingestion occurs. Second, in the water distribution system model, contaminants injected at
terminal nodes have no means for conveyance back into the distribution system. Any ingestion of the
contaminant can only occur at the injection node.
The performance of a water quality sensor monitoring network is sensitive to factors such as the
population distribution and the ingestion model used in the design and, as a result, sensor monitoring
networks should be designed using the most realistic descriptions available for population distribution
and the timing of when people ingest tap water [10]. Sensor network performance (i.e., the ability to
reduce public health consequences through the optimal placement of sensors to detect contaminants)
can decrease substantially when conditions used for design differ from what actually occurs during a
contamination event [10]. However, some sensor network designs perform better under changed
conditions than others [10]. For example, sensor monitoring networks based on the mean statistic
generally outperform worst-case designs when minimizing worst-case impacts when conditions differ
from those assumed in the designs [10].
Table 5.3.5.1 below provides three example contamination scenarios.
Page 79 of 93
-------
Table 5.3.5.1. Example Contamination Scenarios
Contaminant &
Behavior
Example Toxin
Used in this
Manual
(assumed no
decay or loss)
Example Less
Toxic Chemical
(assume no
decay or loss)
Very toxic
chemical or
toxin (assume
no decay or
loss)
Very toxic
biological
(assume no
decay or loss)
Mass or
Quantity
300 g
>47Kg
=> 500 g
=> 1014
organisms
Mass or
Quantity
Release Rate
625 mg/min
32,939
mg/min
=> 8330
mg/min
4.2 x 1012
organisms/min
Time of
Injection
12:00 AM
local time
6:00 AM local
time
6:00 AM local
time
6:00 AM local
time
Duration
of
Injection
8h
24 h
Ih
24 h
Public Health
Consequence
Measure
pk
(LD50 = 6.0E-6
mg/Kg)
pk
(LD50= 0.001
mg/Kg)
LD50 <= 0.001
mg/kg
Public health
consequences
for dose Levels
(DL) ranging
from 0.0001 to
1 mg
ID50<= 1,000
organisms
Public health
consequences
for dose Levels
(DL) ranging
from 1 to 10s
organisms
Sensor
Network
Design
Approach
Objective
=pk
SP statistic =
mean
Objective =
pk
SP statistic =
pk
Objective
=pk or pd
(use with
DL)
Statistic =
mean
Objective
=pk or pd
(use with
DL)
SP statistic =
mean
DL: Dose level (see definitions).
LD50: Refers to mean lethal dose resulting in 50% mortality.
pd: Population dosed, number of people who receive a dose greater than a given dose level.
pk: Population Killed, sensor placement optimization objective in TEVA-SPOT.
SP statistic: Sensor placement optimization statistic (see definitions).
Page 80 of 93
-------
5.4 Quality Assurance Features within TEVA-SPOT-GUI:
Here we describe three quality assurance (QA) features that a user can utilize to decrease errors and to
improve the accuracy of the results: (1) use of description boxes, (2) running "Perform Mass Balance
Analysis", and (3) removing potentially problematic contamination scenarios using "Injection Location
Exclusions".
5.4.1. Use of Description Boxes:
Description boxes are available within various modules of TEVA-SPOT-GUI to allow the user to add
additional information to describe the purpose and scope of the analysis. For example, within the
"Injection Definition" box the user can enter descriptive or derivative information about the
contaminant being modeled. Such information could be pertinent to the mass injection rate being used.
This description is in addition to the contaminant name specified both in the "Injection Definition" and
within the "Health Impacts Analysis" module. A more complete description of the Ensemble as well as
the consequence assessment and sensor network design being developed can be outlined in the
description box associated with the "Ensemble Options" box.
It highly recommended that sufficient documentation be included in the various description boxes to
enable a quick and easy recognition of what was done in case the documentation is needed months or
years after execution.
5.4.2. Running "Perform Mass Balance Analysis":
For EPANET inp files (models) that do not run in EPANET without errors or warnings, the user is advised
to investigate the errors and warnings. The user may also want to click the box "Perform Mass Balance
Analysis" within the "Ensemble Options" dialogue box and run the mass balance analyses. Figure 5.4.2.1
is a screen shot from TEVA-SPOT-GUI displaying the "Perform Mass Balance Analysis" check box (circled
in red). Checking this box will enable the mass balance analyses to be run when the "EPANET
Simulations" module is run.
Re-open the ensemble "Network_4_Utility Facilities" and check the box associated with "Perform Mass
Balance Analysis". Re-run the "EPANET Simulations" module.
The results of the mass balance analyses will appear as text (.txt) files within the "BaseEnsemble"
directory, for instance within:
C:\TEVA-SPOT-Database\C-Values\Network_4_Utility_Facilities\BaseEnsemble.
Page 81 of 93
-------
Additionally, two other mass balance text files will appear in "BaseEnsemble" folder:
(1) massData.[ENSEMBLE_NAME],
(2)massDataRatio.[ENSEMBLE_NAME].
Open the file "massDataRatio[ENSEMBLE_NAME] with Microsoft Excel. This table
"massDataRatio.[ENSEMBLE_NAME]" provides the results of a mass balance analysis for each
contamination scenario executed within the "EPANET Simulations" module. The file contains six
columns of data. Column 1 specifies the contamination scenario or node ID. The second column reports
the ratio of mass removed to mass injected ((Mass Removed + Mass in Pipes + Mass in Tanks)/Mass
Injected)) for each contamination scenario. Columns 3 through 6 details the mass of contaminant
injected, removed, in tanks, and in pipes, respectively, at the end of the simulation.
The mass balance analysis results used to create the table of mass removed to injected ratios is
contained in the file "massData.Network_4_Utility_Facilities". Within this file are the complete mass
balance results for each contamination scenario and for each time step of the simulation. Here is a
complete, bulleted description of the "massData.[ENSEMBLE_NAME]" file.
• The first row is a header row describing the contents of each column. The headings starting at
column 3 (or "column C" when opened in Excel) is the time step of the simulation and extends
to the right of the table n columns, with n denoting the number of time steps of the simulation
• Column 1 specifies the node ID of the contamination scenario.
• Column 2 specifies the type or category of data that appears in columns 3 through n. There are
four categories of data associated with each contamination scenario: (1) mass removed, (2)
mass injected, (3) mass in pipes, and (4) mass in tanks.
• Columns 3 through n provide the mass balance data associated with each contamination
scenario and category of data at each time step of the simulation.
• The table has been automatically constructed to show those scenarios (node IDs) that have the
highest ratio of mass removed to mass injected first.
Page 82 of 93
-------
5.4.3. Running "Perform Mass Balance Analysis":
The contamination scenarios (node IDs) that have a ratio greater than 1.1 or less than 0.9 should be
investigated further. For more information on investigating such scenarios the user is referred to [9].
For those contamination scenarios that have a ratio greater than 1.1 or less than 0.9, a text file (.txt)
should be created from the file "massDataRatio.Network_4_Utility_Facilities". The resulting text file
must contain a single column of node IDs corresponding to those nodes that have a ratio of mass
injected to mass removed greater than 1.1 or less than 0.9. The resulting file must be saved to a suitable
location and then uploaded through the "Select" box associated with "Injection Location Exclusions" box
within the "Ensemble Options" dialogue box. A screen shot of the "Injection Locations Exclusions"
upload dialogue box is shown in Figure 5.4.3.1. The text file generated of excluded contamination
injection locations could be used for other ensembles using the same network model.
Once the file of excluded scenarios is uploaded,
the "EPANET Simulations", any consequence
modules and sensor placement module will need
to be re-run.
Figure 5.4.2.1. Screen shot of "Perform Mass Balance Analysis" check box.
LsJ Ensemble Options IW^WB
wBHCnpDOH
. r v 1 f Ul" F 1 ~ 1
i ...
Time Faiameters
Duration 168 ] hours T
Quality Timestep 1 minutes •»•
Hydraulic Timestep 1 hours •*•
Reporting Interval 1 hours *
Status: OK
.
f/J Average concentration over reporting interval
WQ Tolerance Value 0.01
Tank Mixing model: Use Scenario ^]
LJ Override if set
Hydraulic Parameters
Maximum Trials 40
CHECKFREQ 2
MAXCHECK 10 f
DAMPLIM1T 0.0 f
Unbalanced Condition CONTINUE »
OK
\v] Use all servers (Uncheck for extremely large networks)
Number of simultaneous SP executions Max T
Decay Parameters
Decay NameExampie^Biotoxin
Reaction Order
Bulk 0 -r
Wall |0 ^]
Tank[o »]
Bulk 0.0
Wall 0.0
Limiting Concentration 0.0
^(ifrToeff. CorrelatitHTTT^^^^
~^
^^iPerform Mass Balance Anajjj^re
InjectonracSfolffifclusions fO) | Select |
| Cancel |
Page 83 of 93
-------
Figure 5.4.3.1. Screen shot of "Injection Location Exclusions" dialogue box.
±_\ Ensemble Options
Description
Network 4. Here we determine C-Values for Utility Facilities
hours
minutes
hours
Time Parameters
Duration 163
Quality Timestep 1
Hydraulic Timestep 1
Reporting Interval 1
Status: OK
Water Quality Parameters
|V] Average concentration over reporting interval
WQ Tolerance Value 0.01
hours
Tank Mixing model; Use Scenario
Override if set
Hydraulic Parameters
Maximum Trials 40
CHECKFREQ
MAXCHECK
DAMPLIMIT
10
0.0
Unbalanced Condition CONTINUE
Simulation Parameters
0 Use all servers (Uncheck for extremely large networks)
Number of simultaneous SP executions Max ^
Decay Parameters
Decay Name Example_Biotoxin
Reaction Order
Bulk JO »-|
Wail |o •*]
Tank cT^l
Reaction Coefficients
Bulk 0.0
Wall 0.0
CLICK Here
Limiting Concentrator
Wall Coeff. Correlatio
QA Parameters
| Perform Mass Balance Analysis
Injection Location Exclusions (0) Select
|£-| Injection Node Exclusions
H
>
:enario
> EH
odules
Nodes
Available
Selected
:unction JUNCTION-0
Junction JUNCTION-1
Junction JUNCTION-10
Junction JUNCTION-100
Junction JUNCTION-1000
Junction JUNCTION-1001
Junction JUNCTION-1002
Junction JUNCTION-1003
Add From File
ansor P
OK
Page 84 of 93
-------
5.5 Using Google Maps™ within TEVA-SPOT-GUI:
This section describes how to view a network distribution system model within Google Maps. The
functionality within TEVA-SPOT-GUI to access Google Maps™ and view the distribution system model in
various Google Maps backdrops is accomplished through the MAP Menu item (Figure 5.5.1) and the
selection of "Projection Data..." and "Google Map Settings...", Figures 5.5.2 and 5.5.3, respectively.
Figure 5.5.1. Accessing Google Maps within TEVA-SPOT-GUI.
i:_, TEVA-SPOT (Networtc_4_Utility_Fadlities in C-Values [Main])
Ensemble File [Map] Mode Edit About
Map Charts' Tc Projection Data...
GflX S» Google Map Settings... efauit
In order to properly display the distribution system network, network model projection information is
needed. In order to properly plot a distribution system model within Google Maps, the data must be
defined according to a particular coordinate system. The coordinate system allows the integration of
the data (e.g., network model) with the geographical data of Google Maps. A coordinate system is a
reference system used to represent the locations of the water utility attributes, i.e., pipes, nodes, and
facilities. Each coordinate system is defined by:
• A measurement system or framework that is often geographic, meaning that spherical
coordinates are measured from the earth's center.
• A unit of measurement that is often in feet or meters.
• Map projection definition for project coordinate systems.
• Other measurement system parameters, e.g., spheroid of reference, datum, and project
parameters such as central meridian.
There are two common types of coordinate systems, geographic coordinate system and projected
coordinate system. A geographic coordinate system may be specified as spherical and parameterized by
latitude-longitude. A projected coordinate system is based on a map projection such as transverse
Mercator. TEVA-SPOT-GUI recognizes three projected coordinate systems: (1) state planar, (2) universal
transverse Mercator, and (3) transverse Mercator. Generally, projection information is available within
the geographical information system (GIS) database for the model. Table 5.5.1 provide projection
information for a particular water system. Notice the bolded information in Table 5.5.1. This
information can be entered into TEVA-SPOT-GUI to project the water system model onto a user-selected
Google Map backdrop.
Page 85 of 93
-------
Table 5.5.1. Example projection Information for a water system model.
PROJCS["NAD_1983_StatePlane_Florida_East_FIPS_0901_Feet",GEOGCS["GCS_North_American_1983"
,DATUM["D_North_American_1983",SPHEROID["GRS_1980",6378137.0,298.257222101]],
PRIMEM["Greenwich",0.0],UNIT["Degree",0.0174532925199433]],PROJECTION["Transverse_Mercator"]
,PARAMETER["False_Easting",656166.6666666665],PARAMETER["False_Northing",0.0],
PARAMETER["Central_Meridian",81.0],PARAMETER["Scale_Factor",0.9999411764705882],PARAMETER
["Latitude_Of_Origin",24.33333333333333],UNIT["Foot_US",0.3048006096012192]]
Note: For North America, the spheroid of choice is GRS 1980, upon which the North American Datum 1983 (NAD83) is based.
To demonstrate here, the user selects "State Plane", "GRS 1980", "Feet", and "zone: 0901: Florida-East"
within the panel of options shown in Figure 5.5.2 of TEVA-SPOT-GUI and clicks "OK". Within TEVA-SPOT-
GUI, the user can select one of four Google Map backdrops to display model attributes and output
results. The four backdrops are: (1) road, (2) satellite, (3) terrain, and (4) hybrid. Figure 5.5.3
demonstrates the display of a small portion of a water distribution system model within Google Maps
(road backdrop) using the projection data described in Table 5.5.1.
To display a network model with a Google Maps backdrop in TEVA-SPOT-GUI, nothing is required from
Google Maps. Provided the projection information is correct and access to the internet is available,
TEVA-SPOT-GUI automatically communicates with Google Maps to display the map details requested by
the user. Notice in the example (Figure 5.5.3) the projection of the pipe detail onto the Google Map
road backdrop indicates that several pipes lie underneath houses or businesses. This is likely not the
case and most likely is a result of errors in the projection information or the coordinates associated with
the pipes/nodes in the model has errors. See Figure 5.5.4 (satellite backdrop), which displays a small
portion of another distribution system model. Notice here that the pipes are reasonably located along
or adjacent to roads and the model's tank icon is aligned with the actual tank.
Google Maps sets a limit on the number of times a user can access tiles from Google per day. This limit
is 1,000 tiles. To overcome this limitation, the user may register for a Google API (application
programmer's interface) key. Within the "Google Map Settings" feature of TEVA-SPOT-GUI is a link,
"Get API Key". A Google API key is not required and since retrieved tiles are cached locally, exceeding
1,000 tiles in a single day will likely not be a problem.
Page 86 of 93
-------
Figure 5.5.2. Projection Panel to Access Google Maps (default data shown).
Aj TEVA-SPOT (Network_4_Utility_Facilities in C-Values [Main])
Ensemble File Map Mode Edit About
Map
Charts Tables
^ & Google Mapa Road Map ^ 1 Map Default
i±j Ua»l
Projection: istate Plane _| •»
Ellipse: GRS 1980
Units: Meters
Zone: 0101: Alabama - East
Status: OK
| OK | | Cancel |
^
Figure 5.5.3. Example distribution system model (small portion of model) projection.
Ensemble File Map Mode Edit About
Map ! charts Tables
0 Google Maps j Road Map vj Map j Default"
v Filtering Manage |
Ellipse: GRS I9SG
Zone: 1090i; Florida - East
Osage Qr
-
-
Projectidfilfata
Monmouth Rd
Monrnc.1 1 i - :
Silver Ridge Ln
~ Katherine Rd
6544
Windmill Way
Model Pipes & Nodes
Spring Meadow Or
Spring Me;
m
mouth Rd
Google Maps™ mapping service is a registered trademarks of Google Inc.
Page 87 of 93
-------
Figure 5.5.4. Example of a properly projected distribution system model.
21
Pipes are along or next to
roads, not under properties,
and the model tank is
accurately represented.
(Satellite view)
Figure 5.5.5. Google Map Settings.
1 X i, y Google Maps Hybrid
nsemble File Map Mode Edit About
Maximum age
Maximum cache size (Mb) 200
Clear Cache |
Google API Key
A Google API key Is not required. Without one, however, you will only be able
to retrieve 1000 tiles from Google per day. Since they are cached locally, this
vjill not likely be an issue. If it is, click the "Get API Key" button below and
follow the instructions to obtain one and enter it here.
Get API Key
The maximum age specifies how long a piece of map data will be kept in the local cache. They shouldn't
be kept forever because Google does update the map data periodically and that would ensure that the
Page 88 of 93
-------
user would have relatively fresh data. The "Clear Cache" can be used to clear the cache. The maximum
cache size controls how big the cache can grow. After it reaches that size, the least recently used tiles
are removed.
5.6 Trouble-Shooting TEVA-SPOT-GUI Problems:
Here we provide some trouble shooting tips in using TEVA-SPOT-GUI. In Section 5.6.1 we describe some
trouble shooting tips for ensuring that TEVA-SPOT-GUI is running properly. In Section 5.6.2 we provide
some information about the numerous log files that are created by TEVA-SPOT-GUI that can be used to
identify potential problems.
5.6.1 Execution:
The uncertainty as to whether TEVA-SPOT-GUI is operating properly after clicking "Execute" can
sometimes be vexing to the user. It is important that the network model (.inp file) first be run and
examined in EPANET to identify any problems. For example, if the .inp file has STOP for the unbalanced
hydraulic condition and the system is unbalanced, TEVA-SPOT-GUI will fail.
The following checks can be performed to determine if TEVA-SPOT-GUI is operating or has
malfunctioned or frozen up and, therefore, should be terminated.
• Examine the "DOS Command" box open associated with TEVA-SPOT-GUI. Visually inspect the
DOS screen looking for the keyword "Severe". If "Severe" appears in the "DOS Command"
window box, there is a problem that will need to be investigated. Choose "Restart TEVA-SPOT
Services" program available from the START Menu under TEVA-SPOT. Close TEVA-SPOT-GUI and
investigate the problem.
• During EPANET simulations, check to ensure that a temporary hydraulic file has been created
within one or more of the "server" folders, e.g., C:\Program Files\TEVA-SPOT\Server\Server-l.
Also, check to make sure no error files (error.txt) appear in the same folder, e.g., the "Server-1"
folder. Figure 5.6.1 is a screen shot showing the temporary hydraulic file. Be sure there is
sufficient hard drive space to accommodate the temporary hydraulic file. For very large
networks, the temporary hydraulic file may be several gigabits or more in size. Since multiple
computer cores are likely processing the simulations, the total space required to accommodate
all the hydraulic files could be quite large.
• During sensor placement (SP), use WINDOWS task manager to check whether the service
"Randomsample.exe" is running. If the network is large, the tso-2-impacts file(s) will be large
and it may take some time before "Randomsample.exe" execution(s) begin. On multi-core
computers with sufficient memory, there may be multiple "Randomsample.exe" services
running.
Page 89 of 93
-------
Figure 5.6.1. Screen shot of temporary hydraulic file.
(Cit( jw i > Computer >
OS (C:) > Program Files > TEVA-SPOT > Server > Server-1 >
Organize * Include in library » Share with * Burn New folder
i Favorites
• Desktop
)l Downloads
~* Recent Places
Libraries
Documents
J' Music
- Pictures
9t. Videos
A Computer
ft, OS (C:)
Name Date modified
jar 7/18/2013
7:24 AM
1 libs 7/18/2013 7:24 AM
modules 7/18/2013
7:24 AM
rrnnt 7/18/2013 7:24 AM
^hyd_data.tmy 7/22/2013
12:09 PM
^"neiuiuilUtlnp 7/22/2013 12:09 PM
network-l.rep 7/22/2013
server.Server-l.ref 7/22/2013
12:09 PM
7:15 AM
I] tevaServer.exe 7/16/2013 2:05 PM
Type
File folder
File folder
File folder
File folder
IMP File
INP File
REP File
REF File
Application
Size
77,744 KB
419KB
21KB
1KB
8KB
A key trouble-shooting tool within TEVA-SPOT-GUI is the "Restart TEVA-SPOT Services" program
available from the START Menu under TEVA-SPOT. If TEVA-SPOT-GUI freezes up, "Restart TEVA-SPOT
Services" stops and restarts the required JAVA and TEVA services. Rebooting the computer will also
"Restart TEVA-SPOT Services". Simply closing the TEVA-SPOT-GUI application does NOT stop TEVA-
SPOT-GUI services from running. Stopping a TEVA-SPOT-GUI execution can only be done by (1) choosing
"Terminate" from Execution Controller panel and then closing TEVA-SPOT-GUI, after the execution of
Termination, (2) choosing Stop or re-start TEVA-SPOT services, or (3) rebooting computer. Choosing
"Terminate" may take a while to execute, depending on the particular module that is running.
Generally, only use "Terminate" to stop executions that are running but waiting to begin due to another
execution that is currently running. Otherwise use "Restart TEVA-SPOT Services".
"Stop TEVA-SPOT Services" and "Restart TEVA-SPOT Services" may not fully work on Microsoft Windows
7 operating systems when running sensor placement optimization or regret analyses. Microsoft
Windows 7 does not allow users to alter processes started by an application (in this case TEVA-SPOT-
GUI) unless executed from the START Menu in Windows 7 using "Run as an administrator". Therefore,
RIGHT-CLICKING on "Restart TEVA-SPOT Services" and choosing "Run as an administrator" will need to
be performed.
5.6.2 Log Files:
TEVA-SPOT-GUI generates numerous log files to help the user identify whether TEVA-SPOT-GUI has
malfunctioned. However, to view the log files a live or real-time editor is needed. Windows notepad is
a live editor but is not compatible with easy viewing of the log files output by TEVA-SPOT-GUI, i.e.,
everything looks jumbled. An open source program called emacs can be downloaded from the link
below.
Page 90 of 93
-------
http://www.xemacs.org/Download/win32/index.htmltflnnoSetup-Stable-Download
TEVA-SPOT-GUI log files are contained in various folders located within the "tevalogs" folder:
C:\Program Files\TEVA-SPOT\tevaLogs. Specifically, there are five folders representing the various
TEVA-SPOT-GUI services, with various log files associated with each. Generally, however, the three most
important log files to examine to identify problems are: (1) "ExecutionController", (2) "Server", and (3)
"DBServer".
The number of "Server" folders will correspond to the number of computer cores that were setup by
TEVA-SPOT-GUI at installation to process the simulations, with each numbered sequentially. For
example, for the quad core computer with 32 gigabytes of random access memory used here, four
"Server" folders were established, i.e., named "Server-1" through "Server-4".
Files are kept for 30 days, so when examining the various log files be sure to first sort the folder by the
latest date or the date corresponding to the time of the problem.
The "ExecutionController" log files are described first, followed by the "Server" and "DBServer" log files.
The log files that are BOLDED are recommended for examination. Generally, searching ".log" files
starting with the "ExecutionController" and then within the various "Server" or "DBServer" log files for
the word "severe" will likely identify whether TEVA-SPOT-GUI has malfunctioned.
• ExecutionController
- ExecutionController_[Date].log - Provides communication and status information
about the "ExecutionController". Look here for the key words "Severe Exceptions".
- ExecutionController_[Date].memory.log - Memory utilization information.
- ExecutionController_[Date].execstats.log - This will provide a summary level
description of the completion (True or False) of each EPANET simulation, HIA, and
IIA scenario as well as each tso-2-impact scenario and sensor placement design.
- ExecutionController_EDatel.stats.log - This file will provide a description of the
completion (True or False) of each EPANET simulation, HIA, and IIA scenario as well
as each tso-2-impact scenario and sensor placement design. If a TEVA-SPOT-GUI
application is failing because of insufficient computer RAM memory, simulations will
be classified as "FALSE". For example, examine this file for "False" simulations when
running large networks for sensor placement.
- ExecutionController_[Date].stats.detail.log - This file will provide a detailed,
complete description of the completion of each EPANET simulation, HIA, and IIA
scenario as well as each tso-2-impact scenario and sensor placement design.
- ExecutionController_EDatel.assign.log - Details the location (e.g., "Server-1) for the
stated simulation, analysis, or sensor network design. This is useful to determine
where ("Server") sensor placement designs are being developed.
• Server
- Server.[x]_[Date].compression.log - Provides information on file compression.
- Server.[x]_[Date].log - Provides communication and status information about the
"Server". Look for "SevereExceptions".
- Server.[x]_[Date].memory.log - Provides memory utilization information on the
"Server" processes taking place.
• DBServer
Page 91 of 93
-------
DBServer_[Date].log- Provides communication and status information about the
DBServer. The DBServer is the communicator between simulations running on the
various computer cores and TEVA-SPOT-Database established to store the data and
results. Look for "SevereExceptions" or "unable to connect to DBServer".
DBServer_[Date].memory.log- Provides memory utilization information.
DBServer_[Date].tokens.log - Provides information on the processing of "tokens" or
information between the simulations running on the computer cores and the
"DBServer".
Once it is determined that TEVA-SPOT-GUI has malfunctioned, the user should terminate TEVA-SPOT-
GUI (close the application and ensure that "Restart TEVA-SPOT Services" occurs). If the source of the
problem cannot be identified and corrected (e.g., correcting a problem with the .inp file), the user
should contact Robert Janke at ianke.robert@epa.gov. To help facilitate a quick resolution to the
problem, the user should choose "ZIP Log Files" from the Microsoft Windows START MENU. The ZIPPED
log files can be found at C:\Users\...\Documents\TEVA-SPOT\logs. Please email these log files to
ianke.robert@epa.gov. In order for the files to be successfully sent via email, please rename the .ZIP
extension to ".txt". Coordination through email can be used for any other problems, including installing
or using the TEVA-SPOT-GUI software.
Page 92 of 93
-------
6.0 REFERENCES.
1. Rossman LA. EPANET 2 Users Manual. Cincinnati, OH: U.S. EPA, Office of Research and
Development, National Risk Management Research Laboratory Report, EPA/600/R-00/057, 2000.
EPANET is available for download at: http://www.epa.gov/nrmrl/wswrd/dw/epanet.html.
2. U. S. EPA. Homeland Security Research, Key Topics, TEVA-SPOT, TEVA-SPOT-Graphical User Interface
software and users manual. User Manual for TEVA-SPOT-GUI is available by search for EPA 600/R-
08/147. User Manual for TEVA-SPOT Tool kit is available by search for EPA 600/R-08/041B. Software
and Users Manual are available at: http://www.epa.gov/nhsrc/index.html. Accessed on July 22,
2013.
3. U. S. EPA. Homeland Security Research, Key Topics, EPANET Extensions, EPANET Multi-Species
Extension Software. Available at: http://www.epa.gov/nhsrc/index.html. Accessed on July 22, 2013.
4. Murray, R, Uber, J, and Janke, R. Model for estimating acute health impacts from consumption of
contaminated drinking water. J Water Resour Plann Manag, 2006; 132(4):293-299.
5. Davis MJ, and Janke R. Importance of exposure model in estimating impacts when a water
distribution system is contaminated. J Water Resour Plann Manag, 2008; 134(5):449-456.
6. Davis, MJ, and Janke, R. Development of a probabilistic timing model for the ingestion of tap water. J
Water Resour Plann Manag, 2009; 135(5):397-405.
7. Davis MJ, and Janke R. Patterns in potential impacts associated with contamination events in water
distribution systems. J Water Resour Plann Manag, 2011; 137(l):l-9.
8. Davis MJ, Janke R, and Taxon TN. Assessing Potential Impacts Associated with Contamination Events
in Water Distribution Systems: A Sensitivity Analysis. Cincinnati, OH: U.S. EPA, EPA/600/R-10/061,
2010.
9. Davis MJ, Janke R, and Magnuson, M. A Framework for Estimating the Adverse Health Effects of
Contamination Events in Water Distribution Systems and Its Application, accepted for publication in
Risk Analysis.
10. Davis MJ, Janke R, and Phillips, CA. Robustness of Designs for Drinking-Water Contamination
Warning Systems under Uncertain Conditions. Submitted for publication.
11. Davis MJ and Janke R. Influence of Network Model Detail on Estimated Health Effects of Drinking
Water Contamination Events. Submitted for publication.
12. Shang F, Uber JG, and Rossman LA. Modeling reaction and transport of multiple species in water
distribution systems. Environ Sci Technol, 2008; 42(3):808-814.
13. U. S. EPA. Sensor Network Design for Drinking Water Contamination Warning Systems: A
Compendium of Research Results and Case Studies Using TEVA-SPOT Software. Cincinnati, OH: U.S.
EPA, EPA/600/R-09/141, 2010.
14. R. M. Clark and S. Hakim (eds.), Securing Water and Wastewater Systems, Chapter 2: Protecting
Water Supply Critical Infrastructure: An Overview, DOI: 10.1007/978-3-319-01092-2_2, Springer
International Publishing Switzerland 2014.
15. U. S. EPA. Estimated per capita water ingestion and body weight in the United States an update.
Washington (DC): Office of Water; 2004 Oct. Report No.: EPA-822-R-00-001. Appendix E, Part III,
Table Al. Community water: gender by broad age categories consumers only; p. E-94.
16. U. G. Survey. Estimated use of water in the United States in 2000. Technical report, U.S. Geological
Survey, 2004. Available at http://pubs.usgs.gov/circ/2004/circl268/pdf/circularl268.pdf.
Page 93 of 93
-------
United States
Environmental Protection
Agency
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA
PERMIT NO. G-35
Office of Research and Development (8101R)
Washington, DC 20460
Official Business
Penalty for Private Use
$300
------- |