Continuous Plume Monitoring and Model Calibration using a wireless sensor network:

flangeeasyΚινητά – Ασύρματες Τεχνολογίες

21 Νοε 2013 (πριν από 3 χρόνια και 10 μήνες)

71 εμφανίσεις



2006 AGU Fall Meeting

Poster No. H41B
-
0420

Continuous Plume Monitoring and Model Calibration using a wireless sensor network:

Proof of Concept in Intermediate
-
Scale Tank Test


Lisa Porta (lporta@mines.edu) & Tissa H. Illangasekare (tissa@mines.edu)

Center for Experimental Study of Subsurface Environmental Processes (CESEP),

Division of Environmental Sciences and Engineering, Colorado School of Mines, Golden, CO

Philip Loden & Qi Han

,
Department of Mathematical and Computer Sciences, Colorado School of Mines, Golden, CO

Denney Liptak &

Anura Jayasumana
,
Dept. of Electrical and Computer Engineering, Colorado State University, Fort Collins, CO


http://cesep.mines.edu

ABSTRACT

3) CURRENT FOCUS: PROOF OF CONCEPT



1) BACKGROUND


TEST BED

5) PRELIMINARY EXPERIMENTAL RESULTS

SUMMARY


The

current

practice

for

accurate

monitoring

of

subsurface

plumes

using

samples

collected

at

monitoring

wells

is

often

limited

by

cost

and

time

constraints
.

This

approach

becomes

impractical

for

continuous

monitoring

of

plumes

that

grow

with

time
.

With

the

development

of

new

sensor

technologies

and

wireless

sensor

networks

(WSNs),

the

potential

exists

to

create

innovative

and

efficient

subsurface

data

collection

for

plume

monitoring
.



The

goal

is

to

automatically

collect

data

from

sensors

and

wirelessly

transmit

the

data

to

computer

platforms

where

inversion

codes

and

forward

simulation

models

reside
.

This

data

can

then

be

used

to

continuously

monitor

and

update

model

parameters

for

the

prediction

of

expected

plume

behavior
.

Many

technological

and

operational

challenges

related

to

sensor

placement

and

distribution,

automation

of

real
-
time

data

collection
,

wireless

communication,

and

modeling

have

to

be

overcome

before

the

implementation

of

complex

plume

monitoring

systems
.



This

preliminary

proof

of

concept

demonstration

study

assesses

this

technology

using

an

aquifer

test

bed

constructed

in

an

intermediate

scale

tank
.

The

test

system

includes

a

set

of

ten

conductivity

probes

individually

connected

to

wireless

sensor

boards

(
motes
)
.

The

tank

was

packed

using

five

well
-
characterized

silica

sands

to

represent

a

heterogeneous

aquifer
.

Bromide

tracer

was

injected

into

a

steady

flow

field

and

tracer

concentrations

at

different

points

in

the

tank

were

measured

with

conductivity

sensors

via

the

motes
.



Eventually,

inverse

modeling

will

be

used

to

determine

subsurface

parameters

needed

for

predictive

modeling
.

This

preliminary

study

is

the

starting

step

in

the

development

of

a

more

complex

wireless

sensing

communication

system

to

be

used

in

field

applications

involving

remediation

design,

performance

assessment,

risk

analysis

and

exposure

assessment
.

Wireless Sensor Networks (WSNs)


Consist of spatially distributed
autonomous

devices using
sensors

to cooperatively monitor physical or environmental
conditions at different locations. Applications include military,
health, traffic, ecosystem monitoring. Each node in a sensor
network is equipped with a
radio

transceiver
, a small
microcontroller
, and an energy source, usually a
battery

(or solar
power in the field).


Why use WSNs?


Reduce wiring for sensor distribution


over a large area


Real time plume monitoring


Intelligent data gathering




2) SCOPE OF PROJECT


AUTOMATED DATA COLLECTION

Migration of dye tracer through tank

Dimensions:


L = 244 cm

H = 40 cm

W = 8 cm

The test bed was built out of plexiglass walls attached to a metal plate. A steady flow
was maintained in the heterogeneous media formed by five well characterized sands.
Ten sensors were distributed in the tank at different depths and in turn connected to
five different motes attached on a rack above the tank.

Conductivity
sensors
packed
within the
tank

Existing sampling technologies



Expensive, involving laboratory analytical methods



Time consuming with respect to collection and analysis



Fail to capture real time plume migration when the concentrations
change



Often not optimally designed with respect to best sampling locations


4) EXPERIMENTAL METHODS

A 200 mg/L NaBr (sodium bromide) solution was injected into a well upstream
(about 10 cm) from the first sensor for 40 min. The slug of tracer migrated
through the tank with the water flow while the sensors reported the measured
change in electrical conductivity. Data were wirelessly communicated to the
computer via the base mote, which sends the reading commands at specified
time intervals.




The motes were effectively used in combination with electrical conductivity sensors in
a two
-
dimensional physical test bed



Real
-
time, quantitative data were obtained to create tracer breakthrough curves



Sensor calibration issues did not permit model calibration at this time

FUTURE WORK



Develop a different sand packing scheme with a random heterogeneous field and
smaller sand lenses to get more valuable data for inverse modeling



Build a screen around the sensors to mimic a well



Extend wireless communication between the motes



Selective data sampling: the motes should report only the sensor values which are
useful for model calibration. This can be done by setting a threshold value for the
motes to report to the computer



Real
-
time inverse modeling: as new data are collected from the sensors over time,
the model can be recalibrated more accurately


Breakthrough curves for tracer experiment in 2-D tank
0
50
100
150
200
250
0.00
1.00
2.00
3.00
4.00
5.00
6.00
7.00
8.00
9.00
10.00
11.00
12.00
13.00
14.00
15.00
16.00
17.00
18.00
19.00
20.00
21.00
22.00
23.00
24.00
Time (hours)
Electrical conductivity (microS/cm)
sensor 1 (#30 sand)
sensor 2 (#16 sand)
sensor 3 (#110 sand)
sensor 4 (#70 sand)
sensor 5 (#16 sand)


Development of a physical test bed for wireless sensing application in
subsurface plume monitoring and evaluation of wireless sensing
equipment efficiency



Sensor to mote connection



Sensor calibration and sensitivity analysis



Development of test bed (two
-
dimensional) with initial modeling




Tracer experiments for wireless sensor testing



Wireless automated data gathering, storage and analysis



Model calibration



Upscaling of methodology (three
-
dimensional)


ACKNOWLEDGEMENTS

Project funded by Army Research Office: Terrestrial Sciences, Environmental Sciences
Division.



Dr. Russell Harmon, Senior Program Manager

Excite sensor
(apply voltage)

Command

Output

Data

Base
Mote

User
interface

Initial flow model for tank packing, and sensor placement in different
sands

The user types in commands (such as
time interval of sampling and sensors to
be excited) which the base mote will
wirelessly communicate to the tank
motes. These apply a voltage to the
sensors which respond with an
instantaneous reading of electrical
conductivity in the sand. Sampling times
are programmable for minimum user
involvement.

The breakthrough curves illustrate the movement of tracer in the tank as captured by
the sensors.