Copyright © AWWA 2010

Distribution System Energy Distribution System Energy Management and Operations Management and Operations

Optimization System Optimization System Implementation at Gwinnett CountyImplementation at Gwinnett County

Distribution System Energy Distribution System Energy Management and Operations Management and Operations

Optimization System Optimization System Implementation at Gwinnett CountyImplementation at Gwinnett County

Brian M. Skeens, P.E.Brian M. Skeens, P.E.Brian M. Skeens, P.E.Brian M. Skeens, P.E.

AcknowledgementsAcknowledgements

≈ Neal Spivey, Gwinnett County

≈ Angela Dotson, Gwinnett County

≈ Derceto

Need for Real Time Energy OptimizationNeed for Real Time Energy Optimization

Rising Energy Costs – (Second largest cost behind labor)

Manage energy cost in real-time rate environment– Many systems have option of real-time hourly pricing– Systems in deregulated markets can negotiate rates and structure

Reduce energy use (and GHG emissions)– 85% energy use is pumping - manage pumping for best efficiency

Operate more consistently to best utilize & protect assets– Do not breach system constraints (pressures, etc)

Typical Power Use in a Water System Typical Power Use in a Water System

Range of SolutionsRange of Solutions

≈ Capital equipment upgrades

≈ Standard operating procedures (SOPs) development

≈ Partial automation (PLC coordination)

≈ Full, real time automation (Energy Management Operations System)

≈ Understand potential savings and calculate the payback period for capital costs

Cost Reduction TechniquesCost Reduction Techniques

≈ Moving Energy (kWh) in Time (Energy Load Shifting)

≈ Reducing (peak) Demand Charges (kW)

≈ Generating Efficiency Gains Selecting most efficient pumps or

combination of pumps≈ Selection of lowest production

cost sources of water≈ Selection of lowest cost transport

path for water Pump lifecycle costs

Why a Real Time Optimization System?Why a Real Time Optimization System?≈ Interfaces directly to existing SCADA with minimal equipment,

instrumentation or hardware changes≈ Operational tool to schedule pumps/valves to achieve lowest overall

cost (without breaching constraints)≈ Solves mass-balance first (i.e. must deliver water)≈ Aims to minimize costs of energy (best use of off-peak rates)≈ Aims to maximize energy efficiency of pumps (BEP)≈ Can improve water quality by managing turnover≈ Runs in real time – Like an autopilot≈ Recalculates schedule (next 24-48 hours) every 1/2-hour, adapting to

changing conditions of the day≈ Energy Management and Operations (EM&O) Optimization System

takes max advantage of off peak rates

Energy Load ShiftingEnergy Load Shifting

Max Peak-Time kW (May 2006)

0

2000

4000

6000

8000

10000

12000

14000

16000

18000

20000

1-M

ay

2-M

ay

3-M

ay

4-M

ay

5-M

ay

6-M

ay

7-M

ay

8-M

ay

9-M

ay

10-M

ay

11-M

ay

12-M

ay

13-M

ay

14-M

ay

15-M

ay

16-M

ay

17-M

ay

18-M

ay

19-M

ay

20-M

ay

21-M

ay

22-M

ay

23-M

ay

24-M

ay

25-M

ay

26-M

ay

27-M

ay

28-M

ay

29-M

ay

30-M

ay

31-M

ay

Day

Su

m o

f kW

at

WT

P a

nd

Inta

kes

5 MW Reduction

Peak kW ReductionPeak kW Reduction

Target the Highest Efficiency PumpsTarget the Highest Efficiency Pumps

Chooses Efficient Pump CombinationsChooses Efficient Pump Combinations

Pump Efficiency ImprovementsPump Efficiency Improvements

Typical Real Time Optimization Project OverviewTypical Real Time Optimization Project Overview

≈ Phase 1 – Feasibility Study: 3 months

≈ Phase 2 – Detailed Design: 4 months≈ Phase 3 – Implementation 8 months

SW Configuration /FAT (6 months) Field Installation/SAT (2 months)

≈ Phase 4 – Ongoing Support and Maintenance

Derceto AQUADAPT Utility Case StudiesDerceto AQUADAPT Utility Case Studies

Energy Management InstallationsTotal Utility Population

Served

Annual Savings(US$)

Energy Cost Savings(%)

Annual CO2 Reduction

(Ton)

East Bay Municipal Utility District, Oakland CA (2004) 1.3 M $370k 13% 800

Eastern Municipal Water District, Perris CA, Stage 1 (2006) 0.6 M $125k 10% 300

Eastern Municipal Water District, Perris CA, Stage 2 (2007) 0.6 M $150k 15% TBA

Washington Suburban Sanitary Commission, Laurel MD (2006) 1.7 M $775k 11% 4,500

WaterOne, Kansas City KS (2006) 0.4 M $800k 20% 4,800Region of Peel, ON (2009)* 1.1 M ~1M+* 16% TBA

Gwinnett County, GA (2009)* 0.4 M ~$460k* 10% TBA

* Estimated savings on recent Installations

Gwinnett County, GeorgiaGwinnett County, Georgia

800,000 Population Served 2 Water Filtration Plants, 145 MGD Peak Demand 8 Pressure Zones,19 Storage Tanks, 9 RCVs 17 Pumping Plants with 44 Pumps $ 4.6 M power cost in 2006

Gwinnett Optimization DriversGwinnett Optimization DriversNeal Spivey, Director of Water Production quoted their key operational drivers for pursuing an optimization project as being:

≈Operational Consistency Overall objectives of distribution system operation were met, but Each operator had his own preference for an operational scheme

≈Asset Management Required better information on pump efficiency, run hours, lead-lag, so better pump schedule

selection based on “best fit” could be made coincident with Gwinnett’s Asset Management concept.

≈Energy Cost and Usage Power cost and the economy were both considerations. Cost was $4.6 M in 2006 and anticipated $ 6.1 million in 2010 Energy management and cost reduction were prudent strategies

Real-Time Market PricesReal-Time Market Prices

2007 Analysis of Options2007 Analysis of Options

≈ Optimization feasibility study completed by CH2M HILL and Derceto concluded: Short Term: Savings estimated at $235k of “Incremental

Energy” Long Term: Savings on “Standard Bill” Risk Mitigation: Operating on real-time market Other Savings: Pump operating efficiency gains Other Benefits: Water turnover, predictable operation &

planning

Will Take Advantage of Real-Time Energy PricingWill Take Advantage of Real-Time Energy Pricing

0

50

100

150

200

250

300

350

15/0

6/20

06 0

:00

15/0

6/20

06 1

2:00

16/0

6/20

06 0

:00

16/0

6/20

06 1

2:00

17/0

6/20

06 0

:00

17/0

6/20

06 1

2:00

18/0

6/20

06 0

:00

18/0

6/20

06 1

2:00

19/0

6/20

06 0

:00

19/0

6/20

06 1

2:00

20/0

6/20

06 0

:00

20/0

6/20

06 1

2:00

21/0

6/20

06 0

:00

21/0

6/20

06 1

2:00

Date

MG

D a

nd

$/M

Wh

-10

-5

0

5

10

15

20

25

30

35

40

MG

Derceto Raw Pumps from SC intake Historical Raw Pumps from SC intake Day Ahead Price

Derceto Raw Water Storage Min/Max Raw Lanier Storage

Raw Pumping – Summer Alt 1Raw Pumping – Summer Alt 1

0

50

100

150

200

250

300

350

15/0

6/20

06 0

:00

15/0

6/20

06 1

2:00

16/0

6/20

06 0

:00

16/0

6/20

06 1

2:00

17/0

6/20

06 0

:00

17/0

6/20

06 1

2:00

18/0

6/20

06 0

:00

18/0

6/20

06 1

2:00

19/0

6/20

06 0

:00

19/0

6/20

06 1

2:00

20/0

6/20

06 0

:00

20/0

6/20

06 1

2:00

21/0

6/20

06 0

:00

21/0

6/20

06 1

2:00

Date

MG

D a

nd

$/M

Wh

-20

-15

-10

-5

0

5

10

15

20

25

30

35

40

MG

Historical Raw Pumps from SC intake Derceto Raw Pumps from SC intake Day Ahead Price

Derceto Raw Water Storage Min/Max Raw Lanier Storage

Raw Pumping – Summer Alt 2Raw Pumping – Summer Alt 2

0

50

100

150

200

250

300

350

15/0

6/20

06 0

:00

15/0

6/20

06 1

2:00

16/0

6/20

06 0

:00

16/0

6/20

06 1

2:00

17/0

6/20

06 0

:00

17/0

6/20

06 1

2:00

18/0

6/20

06 0

:00

18/0

6/20

06 1

2:00

19/0

6/20

06 0

:00

19/0

6/20

06 1

2:00

20/0

6/20

06 0

:00

20/0

6/20

06 1

2:00

21/0

6/20

06 0

:00

21/0

6/20

06 1

2:00

Date

MG

D a

nd

$/M

Wh

-10

-5

0

5

10

15

20

25

30

35

40

MG

Derceto Lanier FP Flow Rate Historical Lanier FP Flow Rate Day Ahead Price

Derceto Lanier CW Storage Historical Lanier CW Storage Min/Max CW Storage

Lanier FP Flow & CW Storage – SummerLanier FP Flow & CW Storage – Summer

WFP Flow Rate Change ExampleWFP Flow Rate Change Example

2008 Detailed Design2008 Detailed Design≈ Gwinnett decided to proceed with an optimization

system design and implementation project Based on industry proven software solution (Derceto

Aquadapt) Timing driven by budget availability (drought / economy)

≈ Key elements of detailed design phase included: Operational workshops and constraints specification Hydraulic modeling/analysis and pump curve calibration Review of optimization benefits (savings estimate

updated to $400k+) and revised operating strategies Energy management software configuration

specification SCADA / HMI interface & IT design Project WBS, schedule and implementation timelines

Paul West/ATL worked in NZ with Derceto during Detailed Design

HAZOP Meeting & Constraints HAZOP Meeting & Constraints≈ Hazards and Operability (HAZOP) Workshops≈ HAZOP Summary≈ Constraints for each Asset at All Facilities

Reservoir operating levels Pressure constraints Pump station constraints Valve constraints WFP constraints Other constraints Other modes of operation

≈ Fill Valves Recommendations

Example Calibrated Pump CurveExample Calibrated Pump Curve

VFD Pump Curve RecalibrationVFD Pump Curve Recalibration

Efficiency Improvements Analysis Efficiency Improvements Analysis

EM&O DETAILED DESIGN - Savings ReviewEM&O DETAILED DESIGN - Savings Review

Original projected ROI of 32 months Projected energy cost increase 32% since study

Annual Energy Bill Savings Type $ Savings % Savings

$4.6 M Load Shifting $ 235K 5.1%

Efficiency Gains $ 160K 4.0%

TOTAL $ 395K 9.1%

Annual Energy Bill Savings Type $ Savings % Savings

$6.072 M Load Shifting $ 300K 3.5%

Efficiency Gains $ 160K 4.0%

TOTAL $ 460K 7.5%

2009 Implementation2009 Implementation≈ Key elements of implementation phase;

Aquadapt software configuration Optimizer server hardware procurement Integration with calibrated hydraulic model Transdyn SCADA / HMI interface testing Transdyn RTU/PLC updated for optimization operating mode Set-up robust “real-world” test environment for factory acceptance testing (FAT) On-site installation and operator training and site acceptance testing (SAT)

Key Aquadapt ModulesKey Aquadapt ModulesWater Utility

SCADA System

PC on LAN

Application Manager218

PC onLAN

Dashboard210

OPC

Current day / real-time

Data Cleaner206

SCADA Interface203

PC onLAN

Operator Panel201

Operations Simulator209

Hydraulic Model208

Primary Database (Live Server)

Backup Database

GCDWR Operator Panel – Lanier Central High Service PumpsGCDWR Operator Panel – Lanier Central High Service Pumps

2009 Implementation2009 Implementation

≈ Deliverables Fully configured, tested and implemented energy

management and operations optimization system Fully trained operations team (certification contact hours) Fully documented implementation project and user guides

Overall Outcomes Overall Outcomes

≈ Aquadapt software has basically run the Gwinnett County system since December 2009, with few surprises or difficulties.

≈ Settings have been modified quickly to correct competing actions (too many pumps for conditions, etc).

≈ Support has been excellent. ≈ Project is a good story to tell (public relations).≈ Initial evaluation of savings from the first 4 months of 2010

(compared to 2009): Total system pumping ~ 11 MGD more Energy bills are ~$100,000 lower, so far Expecting much more savings as water demand increases

Preliminary Savings ResultsPreliminary Savings Results

Preliminary Savings ResultsPreliminary Savings Results

Preliminary Savings ResultsPreliminary Savings Results

Preliminary Savings ResultsPreliminary Savings Results

Preliminary Savings ResultsPreliminary Savings Results