WEB-BASED REAL-TIME MONITORING AT PERRIS DAM USING IN-PLACE INCLINOMETERS AND PIEZOMETERS WITH AN AUTOMATIC NOTIFICATION
SYSTEM
John Lemke, PE, GE1 Mike Driller, PE, GE2
Dan Wilson, PhD3
ABSTRACT A real-time monitoring system was used to monitor subsurface movements and groundwater levels at Perris Dam during the construction of a liquefaction remediation test section. The test section considered herein included dewatering, driving sheet piles, excavation and replacement of soil, and deep soil cement mixing. The California Department of Water Resources (DWR) implemented a geotechnical instrumentation program to automatically and continuously monitor changes in subsurface movements and groundwater levels using in-place inclinometers and electronic piezometers, respectively. Each remote station at the project site automatically transmitted readings to a server at selected time intervals using wireless Internet modems. Plots of in-place inclinometer displacement profiles and water elevation were monitored by DWR project team members via a secure web page during and after construction. Water levels and displacement profiles were calculated and compared automatically, in real-time, to pre-established threshold levels. The server computer was programmed to initiate phone calls to a list of designated project members whenever threshold levels were exceeded. Each automated message included the project name, remote station number, and relevant reading from the in-place inclinometer or piezometer instrumentation. The real-time monitoring system used at Perris Dam provided continuous feedback during remediation construction work and provided a cost effective way to achieve continuous field observation with an automatic notification system. This paper will discuss the specification, set up, and performance of the geotechnical monitoring system during the test section monitoring.
INTRODUCTION Perris Dam, located in western Riverside County, California, is a 125-foot high earth dam extending 2.2 miles in length (Figure 1). A residential neighborhood is located immediately downstream of the dam. A 2005 California Department of Water Resources (DWR) study of the Perris Dam foundation indicated that thin sandy layers in the dam foundation are potentially susceptible to liquefaction and severe loss of strength during a large earthquake event. A test program was initiated to evaluate planned repair work scheduled for 2013. This paper describes a real-time monitoring system used to measure ground water levels and lateral subsurface displacements in an effort to ensure public safety during the test program construction.
BACKGROUND While the foundation conditions at Perris Dam were considered adequate by the standards of practice during its design phase in the late 1960s and early 1970s, significant advances in soil liquefaction engineering, including soil sampling, testing, and computer modeling methods, have
resulted iperformeindicatedto liquefafoundatiorelease fr
In 2007, to remedand replacement dconstructsections idesign. Tdewaterinequipmenshows thdownstreplace incexcavatioduring, avaries buexcavatio
in new interped a detailed d that the preaction and seon soils coulrom Perris L
Figure 1. Pe
DWR and thiate the lique
acement of thdeep soil mixtion of the plimmediatelyThe test sectng of the fount, procedure layout of t
eam of the declinometer anon area and fand after the ut on averageon was perfo
pretations ofstudy of the
esence of thinevere loss ofld potentially
Lake.
erris Dam Pr
he Division oefaction riskhe upper fouxing (CDSMlanned reme
y downstreamtion work incundation andes, and mix the dewaterinewatering tend piezometeformed an inconstruction
e is about 3.5ormed.
f foundation e Perris Damn sandy layef strength duy lead to slop
roject Locati
of Safety of k at Perris Daundation soil
M), and constediation meam of the damcluded one dd one full-scadesign woulng test excav
est area is an er instrumenntegral part on of the dewa5 feet below
conditions am foundationers in the damuring a large pe failure of
ion Image w
f Dams (DSOam. The plals, strengthentruction of a asures, it wasm toe, as showdewatering teale CDSM teld produce thvation area. indication o
ntation was iof the real-timatering test a
w ground surf
and predicten in 2005 (DWm foundationearthquake
f the dam and
with Proposed
OD) jointly wan included aning of deepdownstream
s decided to wn in Figureest excavatioest section tohe required sThe vegetati
of the seepaginstalled at thme monitoriand excavatiface in the ar
ed risk in perWR 2005). n are potentievent. Liqud uncontroll
d Remediatio
worked out aa combinatioper foundatiom berm. Prio
perform twoe 2, to evaluon to evaluao verify thatsoil improveion visible imge conditionhe dewaterining system oion. The grorea where th
rformance. DThe 2005 stially suscept
uefaction of tled reservoir
on Zone
a conceptualon of removaon soils usinor to full o types of tesuate aspects oate the t CDSM ement. Figummediately s at the site.ng test operation priound water the dewaterin
DWR tudy tible the r
l plan al
ng
st of the
ure 3
In-
ior, table ng test
F
The instrduring coadvance meet thes
Figure 2.
Figure 3. Pe
INS
rumentation onstruction, the state of pse goals, the
Perris Dam
rris Dam De
STRUMEN
and monitorevaluate perpractice for f monitoring
Dewatering
ewatering Te
TATION A
ring programrformance dufuture DWRprogram wo
Test Excava
est Excavatio
AND MONIT
m for Perris During test sec
R projects whould require
ation and CD
on, looking d
TORING P
Dam had thrction construhere monitorinstrumenta
DSM Test S
down from t
PROGRAM
ree main goauction activiring is requiration measur
ection
the crest.
als: ensure saities, and red. In orderements at tim
afety
r to me
intervals and excathis projereal-timevalues. Kof sensorsensor resensor rereduced dthe reducauthorizeservice a Instrum Ground wpiezometinstalled Instrumeto monitolocation ameasuredtransducediaphragmatmosphesubmergemodule (wireless
frequent enovation constect to meet the data evaluaKey componrs deployed aeadings fromeadings at thedata and comced data exceed users via and a commo
entation
water levels ters and in-pat the projecntation Nortor changes inand depth ofd at the timeer is an electm to an analeric pressureence. A cen(GST) conneInternet mod
Figu
ough to effetruction activhe continuou
ation and notnents of the aat the site, fr
m the project e time of tranmparison to teeded threshan interactiv
on web brow
and slope stplace inclinomct site. Ten mthwest were n water levef each piezomof installati
trical resistanlog voltage oe, so pressurentral controllected to eachdem.
ure 4. Piezo
ctively evaluvity. An autus sampling tification whautomated mrequent sampsite to a webnsmission tothreshold cri
hold levels, ave software p
wser.
tability perfometer instrummodel numbplaced at thel within the meter is showon are also pnce type senoutput. Eache acting on ter module (G
h piezometer
ometer Eleva
uate the perftomated monand reportin
hen reading lmonitoring sypling of the b server como the server citeria, automand delivery program at a
ormance of Pmentation. A
ber PS9801 ee bottom of casing over wn in Figurepresented in nsor that convh piezometerthe piezometGCM) acquir and transmi
ations for P-
formance of nitoring systng requiremelevels exceedystem includsensors, wir
mputer, real-tcomputer, co
matic notificaof meaning
any time on a
Perris Dam wA total of 14electronic piopen standpthe period oes 4 and 5. GFigures 4 an
nverts water pr utilizes a vter sensor reires readingsits the data t
1, P-3, P-5,
the dam durtem was seleents and provded establish
ded a relativereless transmtime data evomputer evaation for anyful reduced any compute
was monitore4 monitoringiezometers mpipe PVC pieof constructioGround watend 5. The mpressure acti
vent tube thaepresents thes from a tranto a server co
P-7 and P-9
ring dewaterected for usevide a meanhed thresholely large num
mission of thevaluation of taluation of thy situation wdata plots toer with Intern
ed using g stations wemanufacturedezometer cason activity. er levels
model PS980ing on a
at references e depth of wansducer interomputer via
ring on
ns of ld mber e the he
where o net
ere d by sings The
1
ater rface
a
A total owithin Pewere manconnecte1, IPI-2, with sensaxial tilt Thereforsensors. developmreadings.
Fig
f four in-plaerris Dam annufactured bd end-to-endIPI-3 and IPsors spaced sensor consie, subsurfacA unique fe
ment resultin.
Figure 6.
gure 5. Piez
ace inclinomend at the toe by Geodaq, Id forming a
PI-4 was 42.012-inches apisting of Mice monitoring
eature of thisng from a den
In-Place Inc
zometer Elev
eters, designof the dam a
Inc. and conslong continu0, 59.0, 135.part. Each secro-Electro-Mg included as monitoringnse spatial d
clinometer E
vations for P
nated IPI-1, Ias shown in sisted of numuous network5 and 169.5 ensor level iMechanical total of 384
g project incldistribution o
Elevations fo
P-2, P-4, P-6
IPI-2, IPI-3 Figure 6. T
merous modek of tilt sens feet below gncludes a temSystems (M
4 temperatureluded enhancof sensors co
or IPI-1, IPI-
, P-8, P-10
and IPI-4,wThe in-place el INC500 m
sors. The totground surfamperature se
MEMS) accele sensors anced detectionombined with
-2, IPI-3 and
ere installedinclinometer
modules tal depth of Iace, respectiensor and a blerometers. d 768 tilt n of shear plh frequent
d IPI-4
d rs
IPI-ively, bi-
lane
Each chacasing usprovide amodule wincludes Figure 8 battery. enclosuremaintain
ain of INC50sing a centraaccurate senswith centralizone controllillustrates thAll piezomees, so no exted a low pro
Fig
00 modules walizer device sor alignmenzer is shownler module (Ghree field enceter and incliternal compoofile.
gure 7. In-Pla
was installedwith four w
nt over the dn in Figure 7GCM) and oclosures witinometer reaonents were
ace Inclinom
d inside a stawheels to tracdepth of the i7. A monitorone Internet h all the nec
adings were tvisible and a
meter and Da
andard 2.75-ck the inner ginclinometerring station amodem (MD
cessary data transmitted tall the monit
ata Collection
inch diametegrooves of thr casing. That the groundDM) as showcollection hthrough the toring instru
n Hardware
er inclinomehe casing an
he INC500 d surface
wn in Figure ardware andplastic field
umentation
eter nd
7. d d
Data Co The data power toplace incnetwork power toInternet mcommanda sleep m Sample F A signififrequent between parameterule of thadequateconditionor less shquality ocapturing
Figure
llection and
collection a the in-place
clinometer orusing a Con the in-placemodem (MDd from the se
mode drawing
Frequency
cant benefit time intervasamples) is
er being meahumb, a samply characteri
n for a slope hould be impf a monitorin
g unexpected
e 8. In-Place
d Transmiss
and transmisse inclinometer GST to obttroller Area
e inclinometeDM) and tranerver and dig low curren
of an automals. Real-timfrequent eno
asured duringple rate of atize the eventmay occur o
plemented. Ing program d events.
Inclinomete
sion Procedu
sion sequencer or GST (ptain a set of rNetwork (Cer or GST mnsmits data tsconnects pont from the b
mated monitome monitorinough to captug the monitot least 10 timt of interest. over a periodIncreasing thor the readin
er and Piezom
ure
ce included tpiezometer sreadings, c)
CAN) commumodule, e) the
o a web servower to the Mbattery until
oring systemng is effectivure the expe
oring programmes the highe
For exampd of one dayhe frequencyngs obtained
meter Monit
the followinstations), b) treadings are
unication bue GCM applver, f) the GCMDM, and gthe next sam
m includes theve when the sected rate of m, includingest anticipat
ple, if you exy, then a samy of sample ed and can pro
toring Statio
ng steps: a) ththe GCM inse acquired ovus, d) the GClies power toCM receivesg) the GCM mple event.
e ability to ssample interchange of ea
g unanticipatted frequencyxpect that a limple interval
events does ovide the ben
ons
he GCM appstructed the ver a digital
CM disconneo the wireless a new places itself
sample at rval (time ach physicalted events. Ay should imiting failuof about 2 hnot reduce thnefit of
plied in-
ects ss
f into
l As a
ure hours he
Figure 9 levels usifield persconstructautomatetemporal
Figure 10captured earthquakcompleteinterval rresults usmonthly magnitudprovidedacquired,water levdoes not eventualltrend of athat may
illustrates hoing sample isonnel to rection during ted monitorinl resolution o
0 provides anan unexpect
ke in Baja Ce set of readiranging fromsing a monthpiezometer r
de of changed in the lower, a complete
vel as a resulsuggest a doly leveling oa lowering wactually exp
ow an automintervals as ocord water lethe dewaterinng system, thof the real-tim
Figu
n illustrationted drop in p
California, Mings from the
m every 2 to hly sample inreadings tha
e in this examr plot from tly different clt of the magownward treout. The uppwater level anplain why th
mated monitooften as everevel readingsng tests. Wi
he cost savinme automate
ure 9. Typic
n of how reapiezometer p
Mexico. The e automated 6 hours. Thenterval. It woat the water lmple is very the automateconclusion e
gnitude 7.2 end. After th
per plot with nd the evente water leve
oring systemry 3 minutess for 10 obseithin a few wgs and the teed monitorin
al Piezomete
al-time monitpressure readlower plot smonitoring
e upper plot ould be reaslevel is gradusmall). How
ed monitorinemerges. Thearthquake. Bhe earthquak
monthly reatual leveling el changed in
m can captures. It would bervation welweeks of impechnical advng system be
er Readings
toring with fdings due to shown in Fig
system at piin Figure 10onable to coually decreawever, with ng system whhe lower plotBefore the e
ke, the plot shadings corre off, but it d
n the first pla
e changes inbe unreasonalls by hand dplementationvantages of inecame appar
frequent samthe magnitu
gure 10 proviezometer P-0 shows the sonclude fromasing over timthe additionhere frequent clearly ind
earthquake, thows a sligh
ectly capturedoes not recoace. This sim
n ground watable to expecduring n of the ncreased
rent.
mple events ude 7.2 ides the -8 with a samsame piezom
m the plot of me (note thenal informatint readings wicates a dropthe lower ploht lowering trs the overall
ord informatimple examp
ter ct
mple meter
on
were p in ot rend, l ion le
provides system ca
Fig Real-Tim Three thrmonitorinLevel 1 wwith possnotificatidesign terequiringand desiginclinoma 24-houas 0.1, 0.thresholdincreasedadditionaelevationpiezometof if it ex
an illustratioan increase o
gure 10. Com
me Evaluati
reshold criteng, each orgwas considersible increasion and promeam depending immediate gn team inclu
meter instrumr period. Th2, and 0.3 in
d criteria wasd by 2 feet bal criteria wan exceeded ater P-4 exceexceeded 3-fe
on of how frone’s ability
mparison of
ion and Thr
ria were estaganized into tred a variatioses to future mpt review ong on the resnotification,uding the Di
mentation werhe maximumnches for levs used for pietween the las added dura specified leeded elevatioet over the l
requent sampy to accuratel
f Piezometer
reshold Crit
ablished for three categoon from normsample frequ
of the data wsults of the d, data reviewivision of Sare based on a
m cumulativevel 1, 2, and iezometer mlast two readring the perioevel. For exaon 1466 feetast three rea
pling with a ly interpret o
P-8 Reading
teria for Saf
both the in-pories based omal requirinuency. Leve
with possible data review. w, and mandaafety of Dama maximum
e displaceme3 thresholdsonitoring. A
dings or 3 feeod of excavaample, durint, or if it excadings, then a
well designobserved beh
gs at Differe
fety Monito
place inclinon the relativ
ng notificatioel 2 was conconsultation Level 3 waatory consul
ms. The critecumulative
ent for inclins, respectivelA trigger waset between thation, and a tng excavatioceeded 2-feean automate
ned automatehavior.
ent Sample F
oring
ometer and pve significancon and reviewnsidered signn with the coas consideredltation with teria for the indisplacemen
nometer monly. A slightls activated ihe last threetrigger was an, if the watet over the la
ed notificatio
ed monitorin
Frequencies
piezometer ce of the reaw of the datanificant requontractor andd very importhe contracton-place nt occurring
nitoring was ly different f the water l
e readings. Aactivated if tter level at ast two read
on was initia
ng
ading. a
uiring d rtant or
over set
level An the
dings, ated.
In the event any data reduced from the instrumentation readings exceeded the established threshold levels, a computer generated phone message would automatically be sent with a description of what station exceeded the threshold criteria along with the measured value. A detailed plan was added to the project specifications that described what action would be required for a given notification. The required action varied with each threshold level. For example, a level 1 notification would be sent only to the Instrumentation Engineer and DWR Project Manager for review. In contrast, exceeding the level 3 threshold required the construction foreman to be on site within 30 minutes and key staff from the contractor and DWR to be on site within 60 minutes of notification. The automatic notification system was deactivated for some instrumentation at various times during construction based on the nature of the construction activity and the anticipated measurement changes. For example, the notification system was disabled for two of the in-place inclinometers when sheet piles were installed adjacent to the inclinometers. Similarly, the notification system was disabled for the piezometers during several periods during construction when it was anticipated that dewatering activity would initiate false notifications. The majority of construction work requiring automated monitoring occurred during the period of time beginning in January and extending through the middle of March 2010. Figure 11 shows the maximum cumulative displacement results for in-place inclinometer IPI-4. The displacement threshold levels are also included on Figure 11 for comparison with the measured results. The results indicate that IPI-4 never exceeded the lowest displacement threshold criteria of 0.1 inches over a period of 24 hours. Cumulative displacements shown in Figure 11 and identified as “24-Hr Average” were calculated by averaging 12 hours of readings and comparing the resulting averaged profile with a set of readings (averaged over a 12 hour period) occurring 24 hours prior to the first set. In addition to the 12 hour averaging method, maximum cumulative displacements were also calculated using a single inclinometer reading and one reading 24-hours prior to the first. Figure 11 illustrates that using single reading events results in increased displacement calculations as expected without averaging, but both calculation methods produced displacements below the established level 1 criteria. None of the inclinometer instrumentation readings exceeded the level 1 threshold for this project.
Web-Ba A useful unreasonsecure dauser. Onprovidingneed to pgraphs anInternet cfor examhome comapplicatio2 capturelocated a This typeApplicatitables anscheme uthe RIA b
Figure 11
sed Data De
automated mnable conditiata storage lone of the innog the featurepurchase andnd tables of rconnection.
mple, project mputer usingon used on ted lateral moadjacent to th
e of web-basion (RIA). Td simple col
used for this based on a re
1. Maximum
elivery
monitoring sons. Once rocation, the novative appres of a desktod install softwreduced dataIn the eventpersonnel cog a web browthis project. ovements durhe sheet pile
sed software The RIA allolor indicator project incluequest from
m Displacem
system shoulreadings havnext step is t
roaches usedop software ware. This fa in real-timet a Level 1 trould access cwser. FigureDisplacemering sheet piwall in the t
approach isows data delicons or gif
uded the follthe user, b)
ments for 24-H
ld deliver resve been colleto deliver th
d for this projapplication i
feature allowe from any crigger notificcurrent resul 12 shows a
ent profiles aile installatiotest excavati
s commonly livery to movf images of plowing stepsthe user can
Hour Period
sults in real-ected at a prohe results in soject was a din a standard
wed authorizecomputer wication occurlts, includingscreen shot
and time-histon and remoion area.
referred to ave well beyoplots createds: a) a body on select vario
d, Inclinomet
-time to any oject site andsome meaninata delivery d web browsed project peith a web brored at night og data-reducof the web-b
tory plots illoval. Inclino
as a Rich Intond static de
d on the serveof raw data ious options r
ter I-4
location witd transmittedngful way tosystem
ser without tersonnel to vowser and anon a weekenced plots, frobased lustrate how ometer IPI-2
ternet elivery of dater side. Theis delivered regarding ho
thout d to a o the
the view n
nd, om a
IPI-was
ta e to
ow
the raw dresponse and the R“data clipor report thousandadditionathe optiohistory p
Figure
Continuindensity omonitorinelements
data should bto user requ
RIA recalculpboard” areadocumentat
ds of readingal navigationn to select vlots.
12. Web-Ba
ng advancesof sensor depng provides were involv
be reduced, cuests, d) the uates and re-p
a within the Rtion. The ap
gs into a simpn tabs containvarious analy
ased Rich In
s in technoloployment, wia reliable w
ved in the au
c) the RIA puser can chaplots the resuRIA and pas
pplication runple tab navigning detailed
ysis methods
nternet AppliInstallati
CON
ogy provide aireless data tay to simplif
utomated mo
erforms calcange variousults, and e) tste directly inns within a wgation systemd plots or da, plot charac
ication (RIAion and Rem
NCLUSION
an ever-incretransfer, andfy data colle
onitoring sys
culations loc aspects abothe user can nto a spreadweb browserm. Each tab ata tables of rcteristics, tim
A) Displayingmoval
NS
easing leveld graphical dection and prstem used at
cally and creout how the d
copy reducedsheet for furr and organiz panel typicareduced data
me windows
g Results Du
of measuremdata presentaresentation. Perris Dam:
eates plots indata is plotteed data fromrther evaluatzes the resulally has sevea. The user , and time-
uring Sheet P
ment ability,ation. Web-b
Several key: 1) a signifi
n ed
m a tion lts of eral has
Pile
, based y cant
number of measurement points captured the physical behavior being monitored and provided an enhanced measurement criteria for identification of potential shear plane development based on closely spaced inclinometer sensors, 2) frequent sampling adequately characterized the performance of the dam during construction and provided continuous observation to increase safety, 3) real-time evaluation of instrumentation measurements provided a robust automatic notification system, and 4) browser based software with interactive charting and analysis tools moved real-time data delivery beyond static data tables and plots delivered as gif images in standard HTML pages, and 5) significant cost savings associated with automation of data collection, data reduction, plotting, and evaluation of data to determine if immediate corrective action is needed. In this project there was no automated alarm triggers and thus no automated notifications or response. The continuous monitoring program was considered a success, however, in that even developing the multi-tiered response strategy contributed to a more thorough communication of risks and concerns to all stakeholders. The test construction phase concluded in March 2010. The data collected during the construction activities is being analyzed by the design teams in evaluating the proposed techniques toward determining the final remediation strategies to be used at Perris Dam.
REFERENCES California Department of Water Resources (2005). “Perris Dam Foundation Study.”