WISE software

Use of water reservoirs for energy storage

Work package 4 2016

2 WISE software

Title:

WISE software,

Use of water reservoirs for energy storage

Authors :

Kristian Duch Søndergaard, Sweco Denmark A/S

Elham Ramin, Sweco Denmark A/S

Publisher:

Miljøstyrelsen

Strandgade 29

1401 København K

www.mst.dk

Photo:

-

Illustration:

-

Year:

2016

Map:

-

ISBN nr.

WISE software 3

Ansvarsfraskrivelse:

Miljøministeriet offentliggør rapporter og indlæg vedrørende forsknings- og udviklingsprojekter inden for miljøsektoren,

som er finansieret af Miljøministeriet. Det skal bemærkes, at en sådan offentliggørelse ikke nødvendigvis betyder, at det

pågældende indlæg giver udtryk for Miljøministeriets synspunkter. Offentliggørelsen betyder imidlertid, at

Miljøministeriet finder, at indholdet udgør et væsentligt indlæg i debatten omkring den danske miljøpolitik.

Må citeres med kildeangivelse.

4 WISE software

Contents .................................................................................................................... 4

1. Introduction ....................................................................................................... 5 1.1 Driving force - Why this project? ....................................................................................... 5

1.2 The concept: What can WISE do ........................................................................................ 5

1.3 Workflow for developing the WISE program .................................................................... 6

1.4 Vulnerability of abstraction wells and water quality ......................................................... 6

2. Description of WISE ........................................................................................... 8 2.1 WISE, software architecture ............................................................................................... 8

2.2 Optimization method .......................................................................................................... 9

2.2.1 Optimization algorithm ...................................................................................................... 9

2.2.2 Constraints and configurations ........................................................................................ 10

2.2.3 Convergence and robustness ............................................................................................. 11

2.3 Optimization process ......................................................................................................... 11

2.4 WISE interfaces ................................................................................................................. 12

2.4.1 Control systems .................................................................................................................. 12

2.4.2 NordPoolSpot ..................................................................................................................... 12

2.4.3 Energinet.dk ....................................................................................................................... 12

3. Proof of concept ............................................................................................... 13 3.1 Proof of concept, level 1 - Conceptual model .................................................................... 13

3.2 Proof of concept, level 2: HOFOR case ............................................................................. 15

3.2.1 Problem formulation .........................................................................................................16

3.2.2 Optimization results .......................................................................................................... 17

3.2.3 Savings ............................................................................................................................... 18

3.3 Proof of concept, level 3 - VCS case...................................................................................19

3.3.1 Problem formulation ......................................................................................................... 21

3.3.2 Optimization Results ........................................................................................................ 22

3.3.3 Savings ............................................................................................................................... 22

3.3.4 Usage of potential storage volume ................................................................................... 24

4. Conclusions ...................................................................................................... 25

5. Perspective ....................................................................................................... 26 5.1 Business plan .................................................................................................................... 26

5.2 Business opportunities ..................................................................................................... 26

5.3 Communication and presentation activities .................................................................... 26

Contents

WISE software 5

The current report describes the development of WISE software, a Water Intelligent Saving Engine

developed by SWECO in collaboration with VandCenter Syd, HOFOR and Aarhus Vand, which are

among the largest water utilities in Denmark. Development of WISE is part of the “Future Water”

project funded by VTUF and the Ministry of Environment and initiated by VandCenter Syd (VCS)

with the purpose of developing tools to support efficient and sustainable operation in drinking

water utilities (http://futurewatercity.com/).

1.1 Driving force - Why this project?

In the framework of Climate Action by European commission, an objective is set to limit global

warming by reducing the emission of greenhouse gases to at least 50% of 1990 levels by 2050.

Drinking water utilities can contribute to this vision by reducing the CO2 emissions of their

pumping systems. To do so it is necessary to develop optimization tools that help water utilities to

use storage volumes at waterworks and in the distribution network in a way that enables them to

postpone energy consumption to “green” hours with lower CO2 emission factors. In fact, water

supplies have the possibility of using more green energy and at the same time achieving economical

savings. Furthermore, many utilities in Denmark have the vison of becoming CO2-neutral in the

near future. For the above reasons the initiative was taken to develop a software that is able to help

water supply utilities to:

1) Take advantage of smart grid (water reservoirs = batteries)

2) Exploit more fossil free energy

3) Move from constant energy prices to market based prices (= economical savings)

4) Reduce total energy costs as much as possible

5) Document and reduce CO2 emissions in relation to energy consumption

6) Automatic and on-line control water production and pumping in relation to chosen priorities.

1.2 The concept: What can WISE do

Optimization of pumping operation in water distribution systems is not a trivial task. It requires

managing storage facilities at water works and in the network at elevated locations. Storage facilities

are mainly used to pressurize the system and provide storage for reducing max. pumping flow and

emergency situations (Figure 1). However, reservoirs can also be used to optimize the operation of

pumps. Most utilities have enough pumping capacity to pump more water when the price is low (and

green) to fill their reservoirs. Then during peak hours with high electricity prices, they can supply the

consumers mostly with the stored water in the elevated reservoirs and thereby minimize pumping

from water works. These reservoirs play the role of "batteries", which are able to store the hydraulic

energy with an efficiency close to 100%. This concept also applies to the reservoirs at each water work

to optimize the operation of pumps at the abstraction wells.

1. Introduction

6 WISE software

FIGURE 1 ELEVATED RESERVOIRS AS BATTERIES FOR STORAGE OF HYDRAULIC ENERGY

Using the most out of storage volumes requires the use of advanced tools that can take into account

the time-varying consumption rates, electricity tariffs, CO2 emission factors, as well as various

constraints on water abstraction, pumping capacity and storage volumes.

The WISE software is capable of connecting to external sources such as SCADA/SRO systems as well

as online servers to retrieve dynamic data on water consumption, electricity price and CO2 emission

factors etc. Based on these data, WISE performs an optimization analysis every hour to find the most

economical operational strategies for the next 24 hours.

1.3 Workflow for developing the WISE program

The following steps have been performed for development of the WISE program:

Requirement analysis for the optimization of water supply systems

Architectural design of the software

Implementation and coding

Testing (Proof of concept)

The WISE program has been evaluated through three levels of proof of concept to confirm that the

developed software can deliver the required capabilities:

Level 1: Using a conceptual model to test the performance of the software in terms of

programmability, expandability, simplicity, stability/robustness and speed.

Level 2: Optimizing a simple existing water supply system using online data to test the trends

and savings.

Level 3: Optimizing a complex existing water supply system using online data to test the

automation of operation control through SCADA systems.

1.4 Vulnerability of abstraction wells and water quality

An even abstraction, to ensure a high and a stable quality of drinking water, has been the rule of

thumb in Denmark for years. This practice is being challenged when using WISE, because it wants

to increase abstraction when electricity is cheap and green.

WISE software 7

Therefore, the influence of fluctuating abstraction on the long term water quality has been

investigated as a sub project by Aktor Innovation. The results have been published in the Danish

technical magazine “DanskVand” in October 2014 with the title “Jævn indvinding er godt for

vandkvaliteten – en sandhed med modificationer”.

The overall conclusion is that most of the aquifers exploited for drinking water purposes in

Denmark will not experience deteriorated water quality due to fluctuating abstraction. WISE does

not so far carry out any automated water quality calculations. Therefore, before applying WISE the

vulnerability of all well fields involved should be investigated through an external study using the

categorization of aquifers and the water quality model developed by Aktor Innovation in the sub-

project.

8 WISE software

2.1 WISE, software architecture

The architecture of WISE is component based, meaning that every component of the software functions independently of the others. An overview of the whole component design of the software is shown in FIGURE 2, which also illustrates the integration of WISE to the external systems in order to obtain the information for the optimization analysis. The external systems are mainly the SCADA/OPC system of the water distribution network, as well as online sources publishing the electricity price and CO2 emission factors, namely Nordpoolspot and Energinet.dk. Data from the SCADA/OPC system are retrieved instantly, whereas the 24-hour data for the electricity price and CO2 emission factor are published the day before.

FIGURE 2 WISE SYSTEM COMPONENT DESIGN AND CONNECTIONS TO EXTERNAL SERVERS

Optimerization algorithm (DLL)

Input database

Output database

Input-functions

Output-funcions

User interface

1) Key values

2) Control strategies

3) Alarms / Error messages

Virtuel OPC

OPC

WISE

Fire

wal

l

Input-user interface

Data validation

Controlers

Nordpoolspot (kr/kwh)Energinet (CO2/kWh)

WISE – boundry interface

2. Description of WISE

Optimization algorithm (DLL)

WISE software 9

2.2 Optimization method

Optimization of water distribution systems in WISE is formulated as a multi-objective

optimization problem, in which the best combination of the values of the decision variables needs

to be identified in terms of certain objectives, such that a number of constraints are satisfied.

The two main objectives are to minimize cost and CO2 emissions of the water works operation.

The decision variables are the water levels in reservoirs and the distribution fraction between the

waterworks. The constraints are pumping and storage capacities as well as abstraction limitations

from the wells.

2.2.1 Optimization algorithm

Among many multi-objective optimization algorithms, genetic algorithm is shown to be an

effective method in optimization of water distribution systems. The “state-of-the-art” multi-

objective genetic algorithm, non-dominated Sorting Genetic Algorithm (NSGA-II, Deb et al.,

20021), is used in WISE. The optimization process in the NSGA-II method is summarized in

Figure 3. The additional features of NSGA-II besides selection, crossover, and mutation as

compared to the traditional genetic algorithms are assuring elitism by combining parent and child

and generating a global population. Additionally, crowding distance comparison on the solutions

with the same ranking is performed. Finally, an efficient constraint handling method referred to

as constrained tournament method by Deb et al. (2002)1 is used to give feasible solutions a

priority over infeasible solutions.

FIGURE 3 OPTIMIZATION PROCESS USING MULTI-OBJECTIVE GENETIC ALGORITHM NSGA-II (WU ET AL., 2010)

1 Deb, Kalyanmoy, et al. "A fast and elitist multiobjective genetic algorithm: NSGA-II." Evolutionary

Computation, IEEE Transactions on 6.2 (2002): 182-197.

10 WISE software

2.2.2 Constraints and configurations In WISE it is possible to apply several constraints that the optimization algorithm has to take into account. The following constraints and configurations are applied:

Water towers / elevated reservoirs:

Level/volume graph

Max. and min. levels (limits)

In service: Yes/No Well fields:

EQ- graph, well pumps

Max. possible production (capacity)

Max. and min. abstraction (limits)

Max. and min. abstractions (capacity) Yearly abstraction permit

Yearly min. production (limit) Water works:

EQ- graphs, pumps

Max. and min. pumping (capacity)

Max. and min. pumping (limits)

Control pressure in pipe network

Min. pressure in pipe network

Reservoirs: o Level/volume graph o Max. and min. levels (limits) o In service: Yes/No

Booster pumps:

Yearly intake permit

EQ- graph, pumps Max. and min. pumping (capacity)

Max. and min. pumping (limits)

Control pressure in pressure zone

Min. pressure in pressure zone

Most of the constraints and configurations are applied via the WEB user interface made for

WISE. The WEB interface is showed in Figure 4.

FIGURE 4 WEB USER INTERFACE FOR APPLYING CONSTRAINTS AND CONFIGURATIONS IN THE

WISE OPTIMIZATION.

WISE software 11

2.2.3 Convergence and robustness

The efficiency of genetic algorithm is very much dependent on its parameter values. The main

parameters of NSGA-II are:

Number of population

Number of generation

Cross-over probability

Mutation probability

To identify the optimum combination of parameters for WISE, several sensitivity tests have been

performed. It is found that the number of population and the mutation probability are the most

crucial parameters for convergence and robustness of the algorithm. High numbers of

populations and low mutations result in high explorations of the possible values for the decision

variables and increase the chance of finding an optimum solution.

2.3 Optimization process

Figure 5 illustrates the components and data flow in the optimization process that is

programmed in WISE.

FIGURE 5 OPTIMIZATION PROCESS AND DATA FLOW IN WISE PROGRAM

The optimization of water supply network is performed for the following hour, which has both

the electricity price and CO2 emission factor available on the online servers. The maximum

optimization period is set to be no more than 24 hours due to computation time constraints.

Constraintshandling

Total pumping flow

Tower volume change

Discharge flow

Waterworksvolume change

Energy

Abstraction flow

CO2 COST

Distribution Fraction

Import/export

Actual water Levels

Water towerwater level

Waterworks water level

Decision variablesSCADA/OPC

EL price

CO2 factor

Online server

EQ curve

HV curve

Static data

12 WISE software

We must note that the water consumption is obtained instantly from the SCADA system for the

current hour. In order to forecast the hourly values of import/export of water in the system for

the following hours, the values from the last three weeks of historical data stored in the database

are used. This way we also take into account the correlation of daily consumption profile to the

type of the day (weekday vs. weekend).

2.4 WISE interfaces

The input data to WISE software are imported through an interface from the following servers:

Control systems (SCADA, OPC)

NordPoolSpot

Energinet.dk

2.4.1 Control systems

Most of the water utilities supervise and control the operation of their machineries using a series of

control systems and tools such as SCADA (Supervisory Control and Data Acquisition), which is the

equivalent of SRO (Styring Regulering Overvågning) in Denmark. The communication of WISE

with SCADA is through an OPC (OPL for Process Control) interface.

2.4.2 NordPoolSpot

Nord Pool is one of the oldest and leading electricity exchange markets, which has a market share of

about 70% in the Nordic region. The pricing is based on the intersection between the aggregated

supply and demand curves, when every day at noon, the participants submit their bids for every

hour of the succeeding day.

2.4.3 Energinet.dk

Energinet is a non-profit company in Denmark owned by the Ministry of Climate, Energy and

Building that is responsible for maintaining the security of electricity and gas supply in Denmark.

Energinet calculates the emission of greenhouse gases of the energy system as a whole in Denmark.

Therefore, Energinet supports the use of fossil free energy sources such as wind turbines, biomass

and solar panels. Energinet has an online FTP server to get the values on the calculated hourly

emission factors.

WISE software 13

This section presents the results of three level testing of the WISE software to confirm its

performance:

Level 1: Using a conceptual model to test the performance of the software in terms of accuracy,

stability/robustness, speed, simplicity, programmability and expandability.

Level 2: Optimizing a simple existing water supply system using online data to test the trends

and savings. Herlev supply area, which is a suburb to Copenhagen and a part of HOFOR, was

used as test facility.

Level 3: Optimizing a complex existing water supply system using online data to test the

automation of operation control through SCADA systems. VandCenterSyd supply area, which is

supplying most of Odense, the third largest city in Denmark, was used as test facility.

3.1 Proof of concept, level 1 - Conceptual model

The optimization algorithm was verified based on the following criteria:

Accuracy: Delivering reliable results, trends and correlations

Stability/robustness: Produces constantly acceptable “accuracy”

Speed: Fast performance to apply to real-time control

Simplicity: Easy to understand and maintain

Programmability: Easy to manipulate the code and adapt to new functionalities

Expandability: Able to handle Multi-zone and other complex cases. Further development is

possible.

For this purpose, a simple conceptual model was set up consisting of a water tower and a pump. The

conceptual model was then tested using arbitrary dynamic data for consumption rate and electricity

prices (see Figure 6) in three different scenarios. The results are listed in Figure 7. Most

importantly, the results show the right trends, meaning that when the electricity prices are low,

reservoirs are filling and when the price is high, reservoirs are emptying.

From this assessment (Table 1) it was concluded that NSGA II had the right performance and that

the WISE software therefore should be based on this optimization algorithm.

TABLE 1 ASSESMENT ON THE PREFORMANCE OF THE OPTIMIZATION ALGORITHM, NSGA II

3. Proof of concept

Optimization algorithm, NSGA-II

Accuracy High

Stability/robustness High

Speed Conceptual model: 10 sec (calculated)

HOFOR case: ~ 5 min (estimated)

VCS case: ~ 30 min (estimated)

Multi-zone cases: ~ 40 min (estimated)

Simplicity Low / Medium

Programmability Medium

Expandability Very high

14 WISE software

FIGURE 6 THE CONCEPTUAL MODEL WITH DYNAMIC WATER CONSUMPTION AND ELECTRICITY PRICE.

THE CO2 EMISSION FACTOR IS SET TO A CONSTANT VALUE OF 366 [G/kWh].

Accuracy Convergence and stability

Scenario 1: Constant flow rate and variable El price

Scenario 2: Variable flow rate and constant El price

Scenario 3: Variable flow rate and El price

FIGURE 7 OPTIMIZED VOLUME CHANGE IN WATER TOWER TO MINIMIZE THE COST DURING

ELECTRICITY PRICE AND CONSUMPTION RATE VARIATION IN 24 HOURS.

WISE software 15

3.2 Proof of concept, level 2: HOFOR case

FIGURE 8 LOCATION OF HERLEV MUNICIPALITY AND THE DIVISION IN TWO HYDRAULIC

DISCONNECTED PRESSURE ZONES (PZ) NAMED “PZ1” (NORTH) AND “PZ2” (SOUTH). HERLEV WATER

TOWER IS LOCATED IN PZ1.

Herlev municipality imports all of its drinking water from HOFOR’s water supply network through

two pressure zones, which in normal operation mode are hydraulically disconnected. Key figures

covering Herlev pressure zone 1 (North) are listed in Table 2. Hjortespring booster pump station

delivers all drinking water to pressure zone 1 (North) in normal operation. However, in emergencies

water can be supplied from “PST Klausdalsbrovej” (pump station), “MB Borgerdiget” (District

metering well) and “MB Violinvej” (District metering well). Herlev water tower (also called

Hjortespring Water tower) connects to the booster station to control the water supply in pressure

zone 1. Figure 9 shows the conceptual model of Herlev water supply system in pressure zone 1 that

was applied in the WISE-model and used in the proof of concept level 2.

FIGURE 9 CONCEPTUAL MODEL OF HERLEV WATER SUPPLY NETWORK IN PRESSURE ZONE 1, NORTH.

PZ 1

PZ 2

N

Herlev

Pressure zone 1 (Nord)

Herlev water tower

Hjortespring booster station

only import

M

MB 2

MB 1 (Kildegården)

MMB Borgerdiget

M

PST Klausdalsbrovej

M

M

MB Violinvej

Import and export

Søndersø-Tinghøjtransport pipe

Import and export

16 WISE software

Key Figures 2015, Herlev pressure zone 1 (north) Units Value

Water works No 0

Pressure zones No 1

Water towers and reservoirs in the water supply network No 1

Emergency connections No 3

Water consumption (Pumped into the network) m3/yr 1,073,136

Electricity price (HOFOR) excl. taxes and VAT1) Kr./kWh 0.2113

Energy consumption, PST Hjortespring (pump station) kWh/yr 99,799

Pressure level (= level of water table in water tower)

- Max. Masl 66.10

- Min., in service Masl 62.30

- Bottom plate Masl 57.50

- Normal Masl 64.20

Pressure, suction side of PST Hjortespring Masl 47.59

Avg. lift, PST Hjortespring mWC 16.61

Level of pressure transducer (pressure side of PST Hjortespring) Masl 37.26

1) FlexPricing, Midtjysk Elhandel A/S taken from invoice covering November 2015

TABLE 2 KEY FIGURES FOR HERLEV PRESSURE ZONE 1 (NORTH).

TABLE 3 CONSTRAINTS ON HERLEV WATER SUPPLY SYSTEM.

3.2.1 Problem formulation

The simple formulation of Herlev water supply in pressure zone 1 is illustrated in Figure 10

consisting of a booster pump and a water tower. The capacity constraints on the pump and water

tower are listed in Table 3. The hourly variations in the electricity price, Co2 emission factor and the

consumption rate are shown in Figure 11.

FIGURE 10 FORMULATION OF HERLEV WATER SUPPLY SYSTEM ZONE 1.

Booster pump

Water tower

Consumption

Minimum Maximum

Herlev water tower Volume 800 m3 2900 m3

Hjortespring pumping station

Pumping capacity 140 m3/h or off 1000 m3/h

Target pumping

flow 140 m3/h or off 375 m3/h

WISE software 17

FIGURE 11 DYNAMIC INPUT USED FOR THE OPTIMIZATION OF HERLEV WATER SUPPLY SYSTEM

3.2.2 Optimization results

The Herlev water supply system was optimized by WISE software during February 2016. The results

are presented here in terms of trends and consistency, as well as the achieved savings as compared

to constant operation strategy.

The optimized operation of the booster pump in pressure zone 1 for the last two weeks of February

2016 is illustrated in Figure 12, which shows the suggested pumping flow that follows the opposite

of variations in the electricity price. In terms of trends and consistency, the optimization algorithm

in WISE performs nicely.

19/02 20/02 21/02 22/02 23/02 24/02 25/02 26/02 27/02 28/02 29/02 01/03 02/030

100

200Consumption rate

Flo

w m

3/h

19/02 20/02 21/02 22/02 23/02 24/02 25/02 26/02 27/02 28/02 29/02 01/03 02/03-0.5

0

0.5Electricity price

DK

K/k

Wh

19/02 20/02 21/02 22/02 23/02 24/02 25/02 26/02 27/02 28/02 29/02 01/03 02/030

200

400

600CO2 emission factor

kg C

O2/k

Wh

18 WISE software

FIGURE 12 THE OPTIMIZED OPERATION OF THE BOOSTER PUMPSTATION IN HERLEV BY WISE

SOFTWARE

3.2.3 Savings

Optimization of booster pump operation at Herlev resulted in 3 to 11 % savings in terms of COST

and CO2 respectively. Figure 13 shows the variations in the hourly savings on variation in the

electricity price in the period where the Herlev boosting pump was optimized.

FIGURE 13 SAVINGS IN TERMS OF CO2 AND COST BY OPTIMIZING THE BOOSTER PUMPING

STATION AND WATER TOWER IN HERLEV USING WISE SOFTWARE

The achieved savings is in a very acceptable range considering the limitation in pumping the high flow rates with much higher energy consumption than the low flow rates as compared to the moderate changed in the electricity price.

0

200

400F

low

m3/h

19/02 20/02 21/02 22/02 23/02 24/02 25/02 26/02 27/02 28/02 29/02 01/03 02/03-0.5

0

0.5

El p

rice

(D

KK

/kW

h)

El price

Pumping flow

Average consumption rate

-200

0

200

400

Volu

me

cha

ng

e i

n t

ow

er,

m3

19/02 20/02 21/02 22/02 23/02 24/02 25/02 26/02 27/02 28/02 29/02 01/03 02/03-0.2

0

0.2

0.4

Volu

me

cha

ng

e i

n t

ow

er,

m3

El price

Filling the tower

Emptying the tower

19/02 20/02 21/02 22/02 23/02 24/02 25/02 26/02 27/02 28/02 29/02 01/03 02/03-150

-100

-50

0

50

100

150

CO

2 s

avin

g %

Total CO2 saving:3 %

19/02 20/02 21/02 22/02 23/02 24/02 25/02 26/02 27/02 28/02 29/02 01/03 02/03-150

-100

-50

0

50

100

150

Cost

savin

g %

Total Cost saving:11 %

WISE software 19

3.3 Proof of concept, level 3 - VCS case

FIGURE 14 LOCATION OF ODENSE MUNICIPALITY AND THE DIVISION IN WATER SUPPLY AREAS.

FIGURE 15 LOCATION OF WATER WORKS AND ELEVATED RESSERVOIRS OWEND AND RUN BY VCS.

VCS water utility company operates one of the largest and oldest water supply system in Denmark

(Figure 14 and Figure 15). VCS supplies industries and 165,000 residents with about 9.2 million m3

water a year through 1,000 km pipe network and 5 water works. Two reservoirs on elevated ground,

namely Bolbro I + II and Sanderum, are used to handle the daily consumption variations. The water

supply area under VCS operation is divided into 11 different pressure zones due to the elevation

differences; however, it is only pressure zone 1 that has elevated reservoirs (Table 4 and Table 5).

Figure 16 shows the conceptual model of the VCS water supply system that was applied in the

WISE-model and used in the proof of concept level 3.

Odense

Water supply areas in Odense municipality VCS Other water utilities

N

Bolbro 1+2

Sanderum

Storage tanks

20 WISE software

FIGURE 16 CONCEPTUAL MODEL OF THE VCS WATER SUPPLY NETWORK APPLIED IN WISE.

Key figures for 2014 and 2015, VCS Units

Water supply area (all pressure zones)

Pressure zones No. 11

Pressure zones with water towers / elevated reservoirs No. 1

Water consumption (pumped into the network)

-2014 m3/yr. 9,276,002

-2015 m3/yr. 9,125,824

Residents No. 165,000

Electricity price excl. taxes and VAT (2015), set price Kr./kWh 0.3356

Pressure zone 1

Pressure level Masl 41.5

Energy consumption, water works (2014) kWh/yr. 1,226,618

Energy consumption, well fields (2014) kWh/yr. 515,015

Water consumption, 2014 (pumped into the network) m3/yr. 7,761,959 TABLE 4 KEY FIGURES FOR THE WATER SUPPLY RUN BY VANDCENTERSYD (VCS).

Pressure zone 1 Units Holmehave

Værket Hoved Værket

Lunde Værket

Lindved Værket, Z1

Avg. geometric lift Booster pumps at water work mWC 10.7 29.1 45.6 18.8

Well pumps mWC 9.3 0 (-6.5) 7.8 11.0 11.0

Sum mWC 20.0 29.1 36.9 56.6 29.8 TABLE 5: AVERAGE GEMETRIC LIFT FOR THE WATER WORKS LOCATED IN PRESSURE ZONE 1.

WISE software 21

3.3.1 Problem formulation

The optimization was applied to the VCS water supply system including three water towers and the

four water works in pressure zone 1. Only one water tower with the volume equal to the three water

towers at VCS is considered in the calculations. The schematic formulation of the problem is shown

in Figure 1. The VCS waterworks have certain constraints on the pumps, treatment capacity and

abstraction permit, which are summarized in Table 6. The hourly variations in the electricity price,

Co2 emission factor and the consumption rate are shown in Figure 17.

FIGURE 1 FORMULATION OF VCS WATER SUPPLY SYSTEM WITH 4 WATERWORKS AND A WATER TOWERS

TABLE 6 VCS WATER WORKS CONSTRAINTS

Unit Holmehave Hoved-

værket

Lindved-

værket

Lund-

værket

Reservoir capacity

Treatment capacity m3/day 1,000 1,000 1,000 1,000

Volume m3 2,000 2,500 1,450 1,000

Min. Level m 0.90 1.00 1.22 1,00

Max. Level m 2.45 3.50 2,99 3.00

Abstraction capacity

Min. pump capacity m3 150 42.1 47 21.7

Max. pump capacity m 1,000 670 482 165

Target min. pump m 200 283 103 45

Target max. pump m 700 482 348 120

Abstraction permit mio.m3/yr. 5.5 1.8 2.1 1.0

Minimum production mio.m3/yr. 0 0 0 0

Discharge capacity

Min. pump capacity m3 136 195 49.14 37.7

Max. pump capacity m 760 575 450 148

Target min. pump m 136 342 104 63

Target max. pump m 613 360 300 110

22 WISE software

FIGURE 17 DYNAMIC INPUT USED FOR THE OPTIMIZATION OF VCS WATER SUPPLY SYSTEM.

CONSUMPTION RATE IS CALCULATED AS: QCONSUMP = QTOTAL CONSUMP + QMUNKDRUP – QDALUM WW, PRODUCTION.

3.3.2 Optimization Results

The optimized operation of the four water works in pressure zone 1 for February 2016 is illustrated

in Figure 18, which shows that filling and emptying of reservoirs follow the opposite of variations in

the electricity price. For the complex water supply system at VCS, the optimization algorithm in

WISE performs just as nicely as in the simple case (Herlev pressure zone 1).

FIGURE 18 VCS UTILITY: OPTIMIZED OPERATION DONE BY THE WISE SOFTWARE.

3.3.3 Savings

Optimization of the operation at VCS resulted in 9 to 8 % savings in terms of COST and CO2

respectively. Transformed into unit cost and emissions and compared to the yearly production in

2014, this adds up to a yearly saving of 18,629 DKK/yr. and 31,824 kg CO2/yr. (Table 7, Table 8

Table 9). The achieved savings are less than the expected potential of 20 % savings. The reason for this is mainly the very low variation in electricity price during the test period combined with the applied constraints on especially the max. and min. pumping limits. Never the less, the project team still

0

1000

2000m

3/h

Consumption rate

0

0.1

0.2

0.3

0.4

Dkk/k

Wh

Electricity price

01/02 03/02 05/02 10/02 13/02 15/02 17/02 19/02 21/02 23/02 25/02 27/02 29/020

600

kg C

O2/k

Wh

CO2 emission factor

Bolbro 1+2 + Sanderum

reservoirs

reservoirs

WISE software 23

believes that 15-20 % savings on the electricity cost (market price) or CO2 emission are achievable. WISE will continue to operate at VCS to verify the full potential of WISE.

VCS - 2014 Yearly water production (2014)

Waterworks m3/year %

Holmehaveværket 2,948,772 38.0%

Lindvedværket 1,399,312 18.0%

Hovedværket 2,732,954 35.2%

Lundeværket 680,921 8.8%

Sum 7,761,959 100%

TABLE 7: YEARLY WATER PRODUCTION ETC. AT VCS IN 2014.

VCS – Average Average hourly

production

(WISE, Feb 2016)

Fraction

(2014)

Cost1) CO21)

Waterworks m3/h % Dkk/h Kg CO2/h

Holmehaveværket 408 38.0% 3,8 7,6

Lindvedværket 67 18.0% 2,5 5,0

Hovedværket 273 35.2% 11,3 6,3

Lundeværket 207 8.8% 3,2 22,3

Sum 955 100% 20,8 41,1 TABLE 8 OPTIMIZED OPERATION AT VCS USING WISE. 1) CALCULATED USING EQ-GRAPHS, VALUES OF ELECTRICITY

PRICE AND CO2 EMISSIONS COVERING FEBURARY 2016 AS LOGGED IN WISE.

Unit COST Unit CO2 COST1) CO21)

Dkk/m3 Kg CO2/m3 Dkk/yr Kg CO2/yr

VCS – Average 0.0258 0.0511 200,259 396,636

VCS – WISE, Feb 2016 0.0234 0.0470 181,630 364,812

Saving 9% 8% 18,629 31.824

TABLE 9: SAVINGS. 1) IN THE CALCULATION, THE YEARLY WATER PRODUCTION IN 2014 IS USED (7.761.959 M3/YR).

24 WISE software

3.3.4 Usage of potential storage volume

FIGURE 19 AN EXAMPLE OF THE VOLUME USAGE IN THE ELEVATED RESERVOIRS AFTER WISE

OPTIMIZATION. LOWER AND UPPER BOUNDS ARE DEFINED BASED ON THE TOTAL PUMPING CAPACITY.

Based on an hourly analysis of the pumping capacity and the available storage volume in the tower,

the following usage of the storage volume in the water tower could be estimated:

Max. available storage volume (between max. and min. level limitations) 26,350 m3 Max. possible usage due to max. pumping capacity (between upper and lower bound in a day) 4,251 m3 WISE usage of storage capacity (between max. and min. optimized level in a day) 3,763 m3

WISE usage of available capacity 14.3%

WISE usage of possible capacity 88.5%

These values indicate that the WISE optimization of water supply in VCS is mainly limited by the

low pumping capacity at the water works in order to use the most out of the storage volume in the

water towers. In order to increase the usage of storage volume in a longer term, the end water level

in each optimization run is determined based on the average electricity price of that day as

compared to the average price three weeks before. Figure 20 shows the volume change in the tower

during the first two weeks of July, where more than 36% of the tower volume could be used for

optimization, twice more than the 14% usage of storage volume in a day of optimization as shown

above.

FIGURE 20 STORAGE PLANNING AT ELEVATED RESERVOIRS IN JULY 2016 BASED ON AVERAGE DAILY

ELECTRICITY PRICE AS COMPARED TO THREE WEEKS BEFORE.

4800

9800

14800

19800

24800

29800

00:0

0

01:0

0

02:0

0

03:0

0

04:0

0

05:0

0

06:0

0

07:0

0

08:0

0

09:0

0

10:0

0

11:0

0

12:0

0

13:0

0

14:0

0

15:0

0

16:0

0

17:0

0

18:0

0

19:0

0

20:0

0

21:0

0

22:0

0

23:0

0

00:0

0

Wate

r to

wer

volu

me u

sag

e [m

3]

Lower bound Upper bound Optimized

Maximum possible usage of storage capacity

01/07 03/07 05/07 08/07 10/07 12/07 14/072.4

2.6

2.8

3

3.2

3.4

3.6

3.8x 10

4

Wa

tert

ow

er

volu

me [

m3]

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

El p

rice

[D

KK

/kW

h]

Watertower storage planning [m3]

El price

Storage volume [m

3]

Sto

rag

e v

olu

me

[m3]

Storage volume [m3]

WISE software 25

The project has succeeded in developing an optimization tool that enables water utilities to use

reservoirs as energy storage and thereby make it possible to postpone energy consumption to cheap

and “green” hours.

The achievement includes:

1) Finding and developing an appropriate optimization algorithm

2) Developing and programming a flexible software architecture that can be applied to both

simple and complex water supply networks in Denmark and abroad.

3) Developing a WEB interface for changing constraints and configurations as well as

showing up to 20 different graphs for analyzing data.

Futhermore, the 3 levels of proof of concept have showed that WISE are able to:

1) follow the wanted trends. High electricity prices produce reduced pumping and vice

versa.

2) make economical and CO2 savings up to 9-11%. At VCS, this adds up to a yearly saving of

about 19,000 DKK/yr. and 32,000 kg CO2/yr. However, this is less than the expected

potential of 20 % savings. The reason for this is mainly the very low variation in electricity

price during the test period combined with the applied constraints on especially the max.

and min. pumping limits. Never the less, the project team still believes that 15-20 %

savings on the electricity cost (market price) or CO2 emission are achievable. WISE will

continue to operate at VCS to verify the full potential of WISE.

3) automatic and on-line control the water production and pumping in relation to chosen

priorities.

4) connect to external sources such as SCADA/SRO systems as well as online servers to

retrieve dynamic data on water consumption, electricity price and CO2 emission factors

etc.

5) perform an optimization analysis every hour to find the most economical operational

strategies for the next 24 hours.

4. Conclusions

26 WISE software

5.1 Business plan

The WISE software is easily converted into different language versions. This means that the

software is prepared to be exported to drinking water utilities abroad. However, Danish utilities will

be targeted in the beginning to get even more documentation that proves the potential of WISE.

WISE is running in test mode at VCS until December 2016. Results from the final test runs is then

going to be used to initiate a closer dialogue with other Danish utilities.

5.2 Business opportunities

Sweco has a strong network and market position in all of Europe. However, the Scandinavian

countries and especially Sweden will be our main focus in the beginning when taking WISE abroad.

These countries have in many places great elevation differences within water supply areas and they

have varying electricity prices due to implementation of green energy produced from especially

windmills (in Denmark), hydropower etc.

5.3 Communication and presentation activities

Throughout the project period, WISE has been presented on numerous occasions. These include:

1. Presentation at “Dansk Vand Conference” 2014

2. IFAT Messe Munchen. Stand at the Danish pavillon, May 2014

3. Presentation at “Dansk Vand Conference” 2015

4. Presentation at “Young Water Professionals Denmark Conference and Workshop” 2015

In addition, WISE has been presented on several occasions in relation to the mother project “Future

Water” and at the website http://futurewatercity.com/.

5. Perspective

Miljøstyrelsen

Strandgade 29

1401 København K

www. mst.dk