ky 17 description

SYSTEM ID: KY 17

NARRATIVE DESCRIPTION

The KY 17 system is based on a real-world water distribution system in Kentucky. Itserves about 13,000 customers and sells water at a rate of $5.19 to $7.70 per 1,000gallons. The system has an average demand of 4.3 MGD with a peak of 6.09 MGD, butthe water treatment plant has a design capacity of 9 MGD. The network was used byHoagland et al. (2015) as part of a classification study. A general schematic of thesystem is shown below. The system had one reservoir, five pumps, and three elevatedstorage tanks. Water loss within the system is estimated to be 27%.

NETWORK SCHEMATIC:

Water Distribution System Database

HISTORY OF THE NETWORK FILE

The KY 17 system was originally created by Steven Hoagland in 2015 as part of anarticle “Classification of Water Distribution Systems for Research Applications” whichwas presented in 2015 in the World Environmental and Water Resources Congress.

ORIGINAL REFERENCE:

Hoagland, Steven & Schal, Stacey & Ormsbee, Lindell & Bryson, Lindsey. (2015).Classification of Water Distribution Systems for Research Applications. 696-702.10.1061/9780784479162.064.

ABSTRACT: Water distribution system models can aid utilities in achieving morereliable and optimal operations of their system. They are also useful in research effortsaimed at improving the planning, design, and operation of systems. This paper outlinesthe development, classification process, and analysis of 15 water distribution systemsfor the purpose of creating a database of system models which can be used among theresearch community to test newly developed algorithms. Differences in basic systemcharacteristics based on configuration are also examined to determine if certaincharacteristics (e.g. number of tanks, average pipe diameter, etc.) vary systematicallyby configuration. The study aims to help quantify differences in the three main systemconfigurations beyond the general layout differences. Such a classification may beuseful in generalizing the economic performance, reliability, resiliency, or requiredcharacteristics (e.g. number of pumps, tanks, etc. per total system demand) of suchsystems. Such statistics may also be useful in helping to forecast system expansionneeds (pipe, tanks, etc.), and security needs (i.e. number of water quality sensors, etc.)as the system continues to grow and expand.

ADDITIONAL CITATIONS:

The original publication of Hoagland et. al. (2015) and by inference the KY 17 systemhave been cited by 7 additional authors. These may be accessed by moving yourcursor over the following link while simultaneously depressing the CTRL key on yourkeyboard: 7 Citations


https://scholar.google.com/scholar?cites=9996444632074407592&as_sdt=4000005&sciodt=0,18&hl=en

AVAILABLE INFORMATION

Physical attributes YesSchematic diagram YesNetwork geometry data YesGIS data file YesBackground map YesElevation data YesPipe data YesPipe material YesPipe age YesPipe pressure class YesNominal or actual diameters NominalPump data YesUseful horsepower YesPump operating curves No

Tank data YesElevation data YesStage storage curves NoWater quality information NoValve data NoPRV/FCV dataIsolation valve dataHydrant dataDemand data YesTotal system demand YesNodal demand data YesTemporal data demands YesSystem leakage YesHydraulic data YesHydraulically calibrated model YesField hydraulic calibration data YesWater quality data NoDisinfection methodChlorine residual dataBooster station dataFluoride/Chloride field dataWater quality calibrated modelOperational data NoSCADA datasetsOperational rules


SYSTEM CLASSIFICATION:

PIPE/LOOP HISTOGRAM:

Hoagland et al. (2015) designed a network classification algorithm for use in classifyingwater distribution systems as either “branched,” “looped,” or “gridded” based on theobserved frequency of network loops with different numbers of distinct pipe segments.The frequency distribution for the KY 17 system is provided below. Using thisinformation, Hoagland et al., classified this system as being a GRIDDED system.

# Total Pipes: 6567# Branch Pipes: 2713

Ratio (Branch Pipes / Total Pipes): 0.41


Figure 3.4. Classification Algorithm (Hoagland et al., 2015)

Hoagland, Steven & Schal, Stacey & Ormsbee, Lindell & Bryson, Lindsey. (2015).Classification of Water Distribution Systems for Research Applications. 696-702.10.1061/9780784479162.064.

NETWORK STRUCTURE METRICS:

Building on the work of Hoagland et al., (2015), Hwang & Lansey (2017) created anexpanded classification system that allows for further classification of a system as beingeither a transmission or distribution branched, looped, gridded, or hybrid system. Theiralgorithm streamlines the classification system by removing unnecessary nodes that donot contribute to the structure of the system while still retaining their use as intermediatepoints for demand data entry. A full description of the algorithm can be found in thecited reference.


Application of the Hwang and Lansey classification algorithm to the system yields thefollowing statics and associated classification:

Parameter ValueEdges 6570Pipes 6565Nodes 6251Average Diameter 6.4Reduced Nodes 3580Reduced Edges 3899Branched Edges 2671Branched Index 0.4Meshed Connectedness 0Reduced Meshed Connectedness 0.04Loop Density 0Average Node Degree 2Hwang & Lansey Classification Distribution Hybrid

Figure 7. Water Distribution System Classification Flowchart (Hwang & Lansey, 2017)

Hwang H. & Lansey, K. (2015) “Water distribution system classification using systemcharacteristics and graph theory metrics.” Journal of water resource planning andmanagement 143(12) https://doi.org/10.1061/(ASCE)WR.1943-5452.0000850


https://doi.org/10.1061/(ASCE)WR.1943-5452.0000850

DETAILED DATA SUMMARIES

PHYSICAL ASSETS:

Asset Type: # of AssetsMaster Meters 7

Tanks 3Pumps 5

Water Sources 1

NETWORK CHARACTERISTICS:

# Total Pipes: 6567# Junctions 6242# Reservoirs 1

# Tanks 3# Regulating Valves Unknown# Isolation Values Unknown

# Hydrants UnknownElevation Data YES

PIPE DATA:

Diameter (in) Length (ft)1 12272 150670

2.5 499243 2251116 6468428 211563

10 11949012 2433316 1953020 3384924 4547

PUMP DATA:

Pump Horsepower YESPump Curves: NO


DEMAND STATISTICS:

Demographic Type Population HouseholdsDirectly Serviceable: 31,336 12,856

Indirectly Serviceable: 11,824 4,647Total Serviceable: 43,160 17,503

Production StatisticsTotal Annual Volume Produced (MG): 1,444.777Total Annual Volume Purchased (MG): 56.895Total Annual Volume Provided (MG): 1,501.672

Estimated Annual Water Loss: 27%

Water CostsCustomer Type Cost per 1000 gallons

Customers within the municipality $5.19Customers outside the municipality $7.70

CUSTOMERS AND USAGE:

Customer Type Customer Count Average Demand (MG)Wholesale: 2 255.461Residential: 12,680 659.633Commercial: 865 121.424Institutional:Industrial: 23 64.612

Other:Total Customers: 13,570

Flushing, Maintenance& Fire Protection:

Total Water Usage: 1,101.130


DATA FILE ATTRIBUTES:

ATTRIBUTE UNITSPipe Length & Diameter X Feet & inches

Pipe Age X Year InstalledNode Elevation X FeetNode Demand X GPM

ValvesHydrants

Tank Levels X FeetTank Volume X Cubic Feet

PRVsWTP

WTP Capacity X GPDPump Data X HP


CALIBRATION REPORTWATER QUALITYWater Treatment Plant DischargesThe average daily demand in 2010 for the Water treatment plant was 4,349,000 gallons.The maximum daily demand in 2010 was 5,782,237 gallons. Figure 4 and Figure 5illustrate the average daily demand and maximum daily demands between 1989 and2010. As can be seen from the figures, the total capacity of the system was increasedfrom 6.0 MGD to 9 MGD in 2009.

Figure 1. KY 17 Average Day Water Production


Figure 2. KY 17 Maximum Day Water Production

Water Quality Monitoring

Continuous water quality testing is performed at the KY 17 Water Treatment Plant.Water is tested for turbidity, alkalinity, hardness, iron, manganese, fluoride, pH,corrosiveness, and disinfectant residual. Table 1 summarizes Fluoride and Chlorinelevels in the annual Water Quality Report.

Table 1. Summarized Water Quality Report

System DemandThe design capacity of the KY 17 Water Treatment plant is 9 MGD. The average dailyproduction is approximately 4.3 MGD with a high daily production of 6.09 MGD. Thetotal annual volume produced is about 1320 MG. The estimated water loss is about13.17% of the total annual volume.

DATA COLLECTIONC-Factor TestAll pipes were categorized into different calibration groups based upon material, size,and age. The breakdown of calibration groups, along with the percentage of thedistribution system that each calibration group encompasses, is shown below in Table 2.The pipes are first divided by pipe material classification and then further broken downby pipe diameter within each pipe material group. The age classification of eachcalibration group is also included, along with the total length of pipe in each group andthe percentage each encompasses of the entire system.

Table 2 Calibration Groups

These pipes were then assigned an initial roughness value, found from C-factors tests,to be placed in the uncalibrated hydraulic model. The results of the C-factor tests can beWater Distribution System Database

used to assign C-factors to other pipes in the system with similar characteristics (EPA,2005). The goal of the sampling locations was to try and perform a C-factor test for eachof the calibration groups. This was not possible due to accessibility of hydrants and lackof available, suitable locations for a given pipe material. Ten C-factor tests wereexecuted in KY 17, representing pipes from four different calibration groups.

DISTRIBUTION SYSTEM MODELA hydraulic model representative of the current water distribution system in KY 17 wascreated using the KYPIPE Program (Pipe 2010). KYPIPE was developed by civilengineering professors at the University of Kentucky. The program allows users tocreate a model of a system consisting of pipe links, internal nodes, and end nodes. Thepoint where pipes intersect is represented as a junction, and locations where a demandoccurs are shown as nodes. Background maps and drawings can be input in vector andraster formats. The model can be created to precisely match the conditions present inthe system. The program can be used to simulate numerous different scenarios in thesystem, analyzing the network through an iterative process utilizing the mass balanceconcept. The process provides results for pressures, velocities, hydraulic grade lines,etc. in pipes and nodes throughout the system. The program can be utilized to analyzeboth steady state and extended period simulations.

Table 3 Typical Hazen Williams C-Factor Coefficients


The purpose of the C-factor test is to measure all factors in the Hazen Williams equationduring hydraulic testing and then solve for the unknown C-factor. The flow rate ismeasured in the field, along with parameters to find the corresponding head loss, inorder to calculate the unknown C-factor (EPA, 2005).

MODEL CALIBRATIONCALIBRATION METHODSCalibration SetupHydraulic calibration is a long process that involves a lot of changes and, so far, is notwell automated by programming (AWWA, 2005). The calibration started by using theresults of the C-factor tests as input for pipes in the model to be used as a starting pointfor the actual calibration. As mentioned earlier each pipe was assigned an individualpipe calibration group 0-9. The results of the C-Factor test were used to assign aC-factor to each pipe calibration group before additional calibration was performed. Onthis step was complete The Fire Flow data was input into the system by setting up tencases in KYPIPE, one for each fire flow test, so as to apply the appropriate boundarycondition and demand patterns. The boundary conditions were set up as changepatterns, which override the setting at specified nodes or pipes with a new value that isapplied to that case. For example FF-9 was modeled as case 9, and the HGL for all ofthe tanks are “changed” to the HGL’s recorded for that test for KYPIPE to use those inthe hydraulic analysis. If not performing extended period simulations, KYPIPE usesthese demand patterns and change patterns as a series of steady state simulations. Tomodel the fire flows, junction nodes were added at the locations in the model of thehydrants used in testing. In the change pattern, the newly assigned junction nodes weregiven a demand equivalent to flow produced from the flowing hydrant during the FireFlow test In order to analyze both static and residual pressures, two sets of change datawere used that were exactly the same except one changed the demand at thesejunctions to zero. Using this procedure, KYPIPE will report the results of all thesimulation runs (all the fire flow tests) simultaneously. This setup allows the modeler toeasily see the effect of changes on all tests; which is advantageous since changingpump or pipe attributes will affect all other simulation runs, if only slightly.

Pump Curve CalibrationIn order to calibrate the high service pumps at the water treatment plant, themanufacturer’s pump curves were obtained. The pump curve is subject to change dueto wear and stress on the impeller over time. If pumps are not tested periodically toupdate the pump curve, the actual pump curve could vary dramatically from the curveprovided by the manufacturer. This attribute is commonly altered during the calibrationprocess (AWWA, 2005).KY 17 WTP had recently undergone improvement within the past 5 years to two of theirhigh service pumps and therefore the pumps were reasonably close to the curve themanufacturer had provided. Data was collected from the SCADA system that measuredboth the discharge pressure of the pump and the flow of the pump to help fine tune thepump curves. The suction pressure of the pump was calculated from the hydraulicgrade line of the clear wells. KY 17 had 5 high service pumps. One of the pumps was


only used as a backwash pump and was not used for the hydraulic model. The resultsof the finalized pump curves can be shown below in Table 4.

Table 4 Finalized Pump CurvesPump- 1 and 2 Pump- 3 Pump- 4Head(ft)

Flow(GPM)

Efficiency(%)

Head(ft)

Flow(GPM)

Efficiency(%)

Head(ft)

Flow(GPM)

Efficiency(%)

600 200 0 550 800 40 500 1120 50500 1600 78 500 1360 55 470 1480 60490 1700 81 470 1760 65 450 1760 65470 1800 82 450 2100 72 430 2160 75460 1900 82.5 430 2600 80 410 2600 80440 2000 83 410 2840 83 390 2880 83420 2100 83 380 3200 83.5 360 3200 84400 2200 82 340 3500 83.5 320 3520 83380 2300 81 320 3640 82.5 300 3700 82350 2400 78 290 3900 81 270 3900 80300 2600 73 240 4200 73 240 4100 77

190 4400 65 210 4320 71

Macro Level CalibrationThe calibration of the KY 17 model involved incremental changes in pipe roughness.The pipes were changed as groups, and these groups are classified by diameter, ageand material as previously shown. This aids calibration, as pipes of similar size andmaterial should have similar roughness. However, pipes with similar attributes will weardifferently over the years depending on their location in the system, causing theirroughness values to diverge. The initial model used the results from the C-factor testsfor each calibration group so as to develop a baseline from which to make adjustments.Once the C-factors from the C-factor testing were input and the pump curve calibrated,a macro calibration was performed on the system. The macro calibration consists ofentering all 10 fire flow tests and measuring such parameters as the static pressuresand the corresponding residual pressures for the given fire flow. “If any of the measuredstate variables are different from the modeled valves by 30% or more, then it is likelythat the cause for the difference may extend beyond errors in the estimates for piperoughness or nodal demand. Possible causes for such differences are many but mayinclude: 1) closed or partially closed valves, 2) inaccurate pump curves or tanktelemetry data, 3) incorrect pipe diameter, 4) incorrect pipe length, 5) incorrect geometry” (Lindell Ormsbee, 1997).It was discovered through the macro calibration that severalof the modeled pressures were excessive in comparison to the measured pressures.The data in the system was reviewed and a wide range of fixes were applied to themodel. Some of the fixes include but are not limited to:

• Connectivity errors- pipes appear to be connected but were not• Incorrect shape data for tank- input error• Demand allocation- demand had been placed on the wrong pipe


• Demand pattern- global demand factor was incorrect• Incorrect pump operations- wrong pump was turned on• Incorrect flow entered – input error• Broken valve was found• Incorrect hydrant elevation• Incorrect placement of hydrant

After the macro calibration was performed, the modeled pressures were much closer tothe measured pressures.Micro-Level CalibrationThe micro calibration was performed into two different steps. The first step of the microcalibration was a steady state analysis and the second was a 24 hour extended periodsimulation. The steady state analysis consisted of fine tuning the C-factors of individualpipes through a trial and error approach. Once the C-factors were assigned to individualpipes a 24 hour extended period simulation (EPS) was performed to determine how wellthe model parameters predicted the flows and tank levels. The EPS allowed themodeler to determine if tanks were filling at the appropriate level. In some cases thetanks would all be over filling which indicates an incorrect system demand. Anotherscenario is when one tank is filling faster than another tank which would indicateincorrect C-factors or incorrect demand pattern in a particular region of the system. TheMicro level calibration consisted of an iterative process until the model reasonablypredicted pressures in the steady state analysis as well as tank levels in the extendedperiod simulation.

CALIBRATION RESULTSFinal Calibrated C-FactorsThe calibration process resulted in alterations to the C-factors for pipes in the systemmodel. These new C-factors, developed for each calibration group, were assigned topipes in the model and varied slightly from the original C-Factors calculated from theC-Factor field tests. These results are shown below in Table 5.

Table 5 C-Factor Calibration ResultsGroup

PipeMaterial

Diameter(in)

AverageAge (yr)

AgeLowEnd (yr)

AgeHighEnd (yr)

AssignedC-Factor

1 AsbestosCement

4, 6 40 N/A 40 123.5

2 PVC 2,3 15 2 30 1113 PVC 4,6 15 2 30 1314 PVC 8,10,12 5 1 10 1395 Ductile Iron 10,12,16 30 20 45 986 Ductile Iron 6,

16,20,2415 15 20 112.6

7 AsbestosCement

8,10 40 N/A 40 115.81


8 Cast Iron 4,6 40 40 60 99.49 Cast Iron 8,10,12 40 40 60 117.4

Comparison of Pressures between Model and Field TestsResults from each fire flow test performed were also found using the model andcompared to real field results. The values for pressure (for residual and flow hydrants)were observed for both model and field results, and the difference between themeasured and predicted values were calculated. The results are shown below in Table6.

Table 6 Fire Flow Calibration Results

The percent difference between model and field results for pressure measurements wasalso calculated, and the results are shown below in Table 7. It was desired to haverelative convergence between model and field results, and Table 7 shows the percentdifference were reasonable.Each fire flow sampling location has been given a Test Site ID. Each test site contains 2hydrants. One hydrant is the designated flow hydrant and the other hydrant is theresidual hydrant. Each individual hydrant has been given an ID for this project (assigned


in KYPIPE Model). Figure 3 shows the location for each fire flow site, and Table 8shows descriptions of all sites used for fire flow tests.

Table 7 Percent Difference between Model and Field Data

Table 8 Fire Flow Testing Locations

Test Site ID Residual Hydrant Pipe Diameter (in)Flow Hydrant

FF-1 R1 8FH101 8

FF-2 R2 8FH102 8

FF-3 R3 8FH103 8

FF-4 R4 6FH104 6

FF-5 R5 6FH105 6

FF-6 R6 8FH106 8


FF-7 R7 8FH107 8

FF-8 R8 10FH108 6

FF-9 R9 8FH109 8

FF-10 R10 8FH110 8


Figure 3. Fire Flow Testing Locations


MODEL VALIDATION24 hour-EPS SimulationIn order to validate the model calibration process, an Extended Period Simulation (EPS)was performed on the calibrated model. A 24 hour period was examined in 1 hourintervals. Specifically, the water levels at all three storage tanks, and pump flow rateswere examined. SCADA data was collected for a few days, and the data for tank levelsand flow rate was compared to the EPS performed on the system model. Thiscomparison can be seen graphically; the water levels in each tank measured by theSCADA system and the levels predicted in the model EPS for October 10, 2011 areshown in Figure 4. The same comparison for October 11 along with October 13 isshown in Figure 5 and Figure 6.The goal of the EPS is to show that all values predicted by the model are reasonablyclose to measurements taken during field testing. The exact values for water levels ofthe tanks under the EPS simulation and recorded by the SCADA system, along with thecalculated differences in water level, are shown for each of the three days.

Figure 4 EPS vs. SCADA Data for 10/10/2011


7Figure 5 EPS vs. SCADA Data (Tank Levels) 10/11/2011

Figure 6 EPS vs. SCADA Data 10/13/2011


SURVEYING PROCEDURES AND DATAC-FACTOR SURVEYINGC-Factor Survey ProcedureDuring the C-Factor Testing prescribed for this project, it was necessary to determinethe difference in elevation between hydrants. The following provides an outline of theappropriate procedure for performing a C-Factor Hydrant Elevation Survey. NOTE: Priorto any surveying activities, the proper care and operation of the total station and itsaccompanying equipment should be studied and reviewed.1) Identify the hydrants that are designated as the flow and the residual and position aLeica TC400NL Total Station and its tripod so that the machine can have a clear line ofsight to both hydrants.2) Level the total station and measure the instrument’s height using a tape measure,yardstick or similar device. Duplicate this height on the prism rods. In situations where itis impractical or undesirable for the instruments and rods to have the same height,record each individual height for use in future calculation.3) At this point, the total station is turned on and the rods are placed at their respectivehydrants. Place the rods on top of the nut located in the center of the desired flownozzle. This will approximate the elevation at the center of the hydrant’s flow.4) Once the rods are placed and steady, the total station operator can take ameasurement by sighting the center of the prism and pressing the “DISP” button on theinstrument. After a few moments, the total station will display the slope distance,horizontal distance, vertical angle and the vertical distance between the prism and theinstrument’s sight. The vertical distance should be recorded for this hydrant on theC-Factor Surveying Data Log.5) Step 4 should be repeated, leaving the total station in place and simply turning ittowards the second hydrant.6) Once the vertical distances for both hydrants have been measured and recorded,elevations will be assigned to each residual hydrant to differentiate which hydrant islocated at a higher elevation (with the lower elevation being assigned a 0 ft elevation).The difference between elevations of the residual hydrants will be calculated andrecorded on the same C-Factor Surveying Data Log.

C-Factor Surveying ResultsTable 9 shows the elevation difference and distances between residual hydrants neededfor C-factor calculations.


Table 9 C-Factor Surveying Results

Site ID FlowHyd ID

ResidualHyd #1ID

ResidualHyd #2ID

ElevationResidual#1 (ft)

ElevationResidual#2 (ft)

ElevationDifference (Elev2 - Elev1) in ft

DistancebetweenResidualHydrants (ft)

C-1 F-1 P-21 P-1 958.511 962.036 3.525 424C-2 F-2 P-22 P-2 927.315 936.087 8.772 769C-3 F-3 P-23 P-3 938.591 919.705 -18.886 809C-4 F-4 P-24 P-4 922.197 916.64 -5.557 682C-5 F-5 P-25 P-5 893.655 882.335 -11.32 2236C-6 F-6 P-26 P-6 948.241 949.591 1.35 372C-7 F-7 P-27 P-7 1009.39 989.81 -19.58 317C-8 F-8 P-28 P-8 914.596 921.091 6.495 443C-9 F-9 P-29 P-9 916.782 907.06 -9.722 492C-10 F-10 P-30 P-10 935.213 955.51 20.297 763

C-FACTOR CALCULATIONSIn order to calculate the C-factor, the procedure shown below was followed. The headloss was first calculated using the pressures recorded at the two residual hydrants afterthe flow hydrant is opened, along with the elevations of each residual hydrant. Theequation used to calculate the head loss between the two residual hydrants was foundusing the Bernoulli Equation (1). The hydrant labeled Residual Hydrant #2 on theC-Factor Data Collection Log will be the upstream hydrant, while the hydrant labeledResidual Hydrant #1 is located downstream

(1)ℎ𝐿= Head loss𝑃1= Residual pressure at Hydrant #1 (downstream hydrant)𝑃2= Residual Pressure at Hydrant #2 (upstream hydrant)𝑍1= Gage elevation at Hydrant #1𝑍2= Gage elevation at Hydrant #2𝑉1= Velocity in pipe at Residual Hydrant #1𝑉2= Velocity in pipe at Residual Hydrant #2𝛾= Specific weight of water g = Acceleration due to gravity

The difference in velocity heads between the two residual hydrants is considerednegligible. Therefore, only the difference in the pressure head and elevation headbetween the hydrants was considered when calculating the head loss.


(2)

ℎ𝐿= Head loss (ft)𝑃1= Residual pressure at Hydrant #1 (downstream hydrant) in psi𝑃2= Residual Pressure at Hydrant #2 (upstream hydrant) in psi𝑍1= Gage elevation at Hydrant #1 (downstream hydrant) in feet𝑍2= Gage elevation at Hydrant #2 (upstream hydrant) in feet𝛾=62.4lb/ft3

The static pressures recorded at each hydrant prior to flowing the hydrant were alsoused as a check for the validity of the data. The static pressure head between the tworesidual hydrants was calculated and compared to the elevation head between thehydrants. The static pressure head should be equal, or relatively close, to the elevationhead. The flow rate in the particular pipe was also calculated using the dischargepressure and geometry of the hydrant.

(3)

Where,𝑄= Flow Rate (gpm)𝐶𝑑= Coefficient of discharge of hydrant (see Figure 24)𝐷𝑜= Diameter of hydrant/reducer opening (in)𝑃𝑑= Discharge or pitot pressure (psi)

The Hazen Williams Equation was used to calculate the C-factor for each C-factor testperformed.

(3)

Where,ℎ𝐿= Head loss (ft)𝐿= Length of pipe (ft)𝑄= Flowrate (gpm)𝐶= C-Factor𝐷= Diameter of pipe (in)

C-Factor Sensitivity AnalysisWater Distribution System Database

In order to perform an accurate sensitivity analysis, every variable used in the C-Factorcalculation is taken into account. These variables include length of the pipe, coefficientof discharge of the hydrant, diameter of the hydrant opening, discharge pressure,diameter of the pipe, pressure at both residual hydrants, and elevation at both residualhydrants. The equation used to calculate the C-factor including every variable is shownbelow.

(4)

Where,𝐿= Length of pipe (ft)𝐶𝑑= Coefficient of discharge of hydrant𝐷𝑜𝑜= Diameter of hydrant opening (in)𝑃𝐷𝐷= Discharge pressure (psi)𝐷= Diameter of pipe (in)Δ𝑃= Change in pressure (psi)Δ𝑍= Charge in Elevation (ft)

The partial derivatives with respect to each variable in the C-Factor equation werecalculated and normalized. These equations were combined with the uncertainty due tothe precision in measurement of each variable. The following equation illustrates aquantitative uncertainty in the C-Factor (labeled as ΔC) due to the uncertainty in eachvariable.

An example uncertainty analysis is shown below for the data measured at Site C-3.

The uncertainty for each C-factor test was calculated. The data needed for calculationsin the sensitivity analysis is shown below in Table 8, and the results of the C-factorsensitivity analysis are shown in Table 10.


Table 10 C-Factor Uncertainty Data

Site Length(ft)

Diameter (in)

DischargePressure (psi)

ΔP (psi) ΔZ (ft) HeadLoss(ft)

C-Factor

C-1 424 8 64.5 1 3.5 5.81 149C-1 424 8 45 4 3.5 15.05 224C-2 769 6 43.2 21 8.8 55.00 143C-3 809 10 76 6 -28.8 14.94 105C-4 682 6 47 12 -5.6 26.74 233C-4 682 6 47.2 14 -5.6 29.05 211C-5 2236 12 82 10.5 -11.39 12.86 127C-5 2236 12 62 17 -11.39 27.88 144C-6 372 8 50.8 3.4 1.4 9.25 132C-7 317 8 36 10 19.6 3.50 130C-8 443 8 54 -7 6.5 5.05 145C-9 492 6 45 14.25 -8.338 20.23 180C-10 763 6 26.5 43 20.3 119.63 75C-10 763 6 28 42 20.3 117.32 78

Several of the C-factor sites were tested multiple times at different flows because thehead loss was not high enough to collect a sound measurement. Site C-9 was later tobe found to have a broken valve directly downstream of the hydrant which would haveaffected the C-Factor rating. Some C-Factors appear to be much higher than what ispossible for a particular material pipe. Thus a sensitivity analysis was performed tocheck to see what the possible errors were in the calculation. The uncertainty for eachC-factor test was calculated, and the results of the C-factor uncertainty analysis areshown in Table 11.

Table 11 C-Factor Uncertainty Results

Site C-Factor withUncertainty

C-1 149 ± 48.6C-1 224 ± 73.1C-2 143 ± 51.9C-3 105 ± 31.0C-4 233 ± 84.5


C-4 211 ± 76.5C-5 127 ± 35.2C-5 144 ± 40.2C-6 132 ± 43.7C-7 130 ± 42.9C-8 145 ± 47.2C-9 180 ± 65.5C-10

75 ± 65.5

C-10

78 ± 28.7

MODEL VALIDATIONThe following tables show the water levels in all three storage tanks measured by theSCADA data compared to the levels found by the model simulation.

Table 12 EPS vs. SCADA Data (Tank Levels) 10/10/2011

Model DATA SCADA DATA Difference(Model-SCADA)

Time(hr)

Elevation of

T1Tank

Elevation of

T2Tank

Elevation of

T3Tank

Elevation of

T1Tank

Elevation of

T2Tank

Elevation of

T3Tank

T1Tan

k

T2Tan

k

T3Tan

k

0 32.16 28.1 26.18 32 28 26 0.16 0.1 0.1

8

1 33.94 29.24 27.89 34 29.7 28 -0.06

-0.46

-0.11

2 36.17 30.93 30.1 36 31.5 30 0.17

-0.57 0.1

3 38.58 32.59 32.43 38 33 33 0.58

-0.41

-0.57

4 39.33 32.47 33.35 39.5 33.5 34 -0.17

-1.03

-0.65

5 37.76 30.43 31.84 39.5 32 33 -1.74

-1.57

-1.16

6 36.91 29.23 31.02 39 31 32 -2.09

-1.77

-0.98

7 37.02 29.44 30.84 38.2 30 31.4 -1.18

-0.56

-0.56

8 37.65 30.19 31.44 38 31.2 31 -0.35

-1.01

0.44

9 35.66 27.99 29.46 36.7 29.5 29.5 -1.04

-1.51

-0.04

10 33.91 25.92 27.62 35 28 27.5 -1.09

-2.08

0.12


11 32.47 24.37 26.21 33 26 26 -0.53

-1.63

0.21

12 30.73 22.44 24.09 31 24.6 24 -0.27

-2.16

0.09

13 28.83 20.37 22.03 29 23.3 22 -0.17

-2.93

0.03

14 27.09 18.61 20.09 27.6 22 20 -0.51

-3.39

0.09

15 27.08 19.36 19.86 27.6 23 20.2 -0.52

-3.64

-0.34

16 27.18 19.86 20.07 27.6 23.5 21 -0.42

-3.64

-0.93

17 27.47 20.41 20.61 28 24 21.5 -0.53

-3.59

-0.89

18 27.88 20.97 21.16 28.1 24.3 22 -0.22

-3.33

-0.84

19 28.2 21.37 21.4 28.3 24.8 22.1 -0.1 -3.43 -0.7

20 28.32 21.5 21.28 29 25 22.3 -0.68 -3.5 -1.0

2

21 28.84 22.26 21.82 29.5 25.6 23 -0.66

-3.34

-1.18

22 29.99 23.8 23.26 30.8 26.8 24 -0.81 -3 -0.7

4

23 31.23 25.29 24.73 32 28 26 -0.77

-2.71

-1.27

24 32.97 27.23 26.85 34 29.5 28 -1.03

-2.27

-1.15


Model DATA SCADA DATADifference(Model -SCADA)

Time

(hr)

Elevation of T1Tank

Elevation of T2Tank

Elevation of T3Tank

Elevation of T1Tank

Elevation of T2

Tank

Elevation of T3Tank

T1 T2 T3

0 34 30 28 34 30 28 0 0 0

1 36.04 31.26 30 36 31 30 0.04

0.26 0


2 38.21 32.74 31.95 37.5 33 32 0.71

-0.26

-0.05

3 39.94 34.36 34.34 39.5 34 34 0.44

0.36

0.34

4 38.56 31.53 32.74 39 32.5 33 -0.44

-0.97

-0.26

5 37.7 30.09 31.83 38 31 32 -0.3 -0.91

-0.17

6 36.6 28.86 30.69 37 29.7 30 -0.4 -0.84

0.69

7 35.29 27.42 29.2 35.5 28 28 -0.21

-0.58 1.2

8 33.67 25.65 27.33 33 26.5 26 0.67

-0.85

1.33

9 31.96 23.71 25.34 31 25 24 0.96

-1.29

1.34

10 30.25 21.81 23.45 29 23.4 22 1.25

-1.59

1.45

11 28.39 19.82 21.43 27 21.5 19.5 1.39

-1.68

1.93

12 26.3 17.56 19.1 25 20 17.5 1.3 -2.44 1.6

13 24.35 15.65 17.27 23 18.5 15.5 1.35

-2.85

1.77

14 22.44 13.78 15.41 21 17 13.5 1.44

-3.22

1.91

15 21.94 13.97 14.78 20 16 12 1.94

-2.03

2.78

16 22.81 15.75 15.72 20.5 18 13 2.31

-2.25

2.72

17 23.85 17.4 17.02 21 19 14 2.85 -1.6 3.0

2

18 24.4 18.01 17.47 21.7 19.5 14.6 2.7 -1.49

2.87

19 24.61 18.07 17.68 21.9 19.9 15 2.71

-1.83

2.68

20 24.98 18.38 18.23 21.9 19.9 15 3.08

-1.52

3.23

21 25.15 18.48 18.19 22 19.9 15.2 3.15

-1.42

2.99

22 25.48 18.8 18.7 22.2 20 15.9 3.28 -1.2 2.8

23 25.83 19.15 19.1 23 20.5 16.3 2.83

-1.35 2.8


24 26.22 19.54 19.56 23.4 21 17.2 2.82

-1.46

2.36


Model DATA SCADA DATA Difference (Model - SCADA)

Time

(hr)

Elevation of

T1Tank

Elevation of T2Tank

Elevation of T3Tank

Elevation of

T1 Tank

Elevation of T2Tank

Elevation of T3Tank

T1 T2 T3

0 32.84 28.47 27.19 32.6 28.5 27 0.24

-0.03 0.19

1 34.02 28.57 28.05 33.7 29 28 0.32

-0.43 0.05

2 35.15 29.09 29.1 34.5 30 29 0.65

-0.91 0.1

3 36.29 29.87 30.14 35.5 30.8 30 0.79

-0.93 0.14

4 37.36 30.65 31.08 36.5 31.5 31 0.86

-0.85 0.08

5 38.36 31.43 31.93 37.5 32 32 0.86

-0.57

-0.07

6 39.23 32.14 32.63 38 32.8 32.8 1.23

-0.66

-0.17

7 39.93 32.82 33.13 38.5 33 33 1.43

-0.18 0.13

8 40 33.14 33.14 39.5 33.2 33 0.5 -0.06 0.14

9 40 33.34 33.19 39.5 33.2 33 0.5 0.14 0.1910 40 33.36 33.06 39.5 33.2 33 0.5 0.16 0.06

11 38.03 31.09 31.66 37.3 31.5 31 0.73

-0.41 0.66

12 35.78 28.27 29.51 35.5 29.5 29 0.28

-1.23 0.51


13 34.04 26.11 27.71 33 27.5 26.8 1.04

-1.39 0.91

14 32.14 23.9 25.57 31 25.8 25 1.14 -1.9 0.57

15 30.42 21.98 23.72 29.5 24 23 0.92

-2.02 0.72

16 28.67 20.18 22.01 28 23 21.5 0.67

-2.82 0.51

17 27.98 20.02 21.31 26.8 22 20 1.18

-1.98 1.31

18 28.4 21.05 21.65 26.7 22.6 20 1.7 -1.55 1.65

19 28.64 21.58 21.77 26.8 23.4 20 1.84

-1.82 1.77

20 28.98 22.08 22.19 27 23.8 20.2 1.98

-1.72 1.99

21 29.3 22.49 22.52 27.3 24 20.9 2 -1.51 1.62

22 29.68 22.92 22.96 28 24.1 21.4 1.68

-1.18 1.56

23 30.13 23.43 23.57 28.7 25 22.4 1.43

-1.57 1.17

24 30.94 24.34 24.72 29.5 25.7 23.6 1.44

-1.36 1.12


Pump InformationWater Treatment Plant Pump SetupThe capacity and horsepower of each high service pump is shown below in Table 15.The pump curves for all 5 high service pumps are shown on Figure 8 through Figure12.. The layout of the high service pumps in the WTP (shown in the KYPIPE model) isshown in Figure 7.

Table 15 High Service Pump Information

PUMP Capacity(GPM)

Horsepower

High Service#1 1500 200





Figure 7 WTP Plant Setup


Pump Curve DataThe figures shown below display the pump curves for the 5 high service pumps used inthe KY 17 Water Treatment Plant.

Figure 8 High Service Pump #1


Tank InformationThe following tables show volume and depth ratios based on water levels for eachstorage tank.

Table 16 T1 Tank Data


ky 17 description

Documents