Dynamic pricing in district heating systems and the impact of heat storageNina Dupont, Bjarne Bach and Mikael Togeby

0

100

200

300

400

500

Stu

dst

rup

v…

Lisb

jerg

Ren

o S

yd

Skan

der

bo

rg

Aar

hu

svæ

r…

Jen

s Ju

ulv

ej

Gel

leru

p

Har

lev

No

de

13

9

No

de

13

0

No

de

14

5

MW

2035 Seasonal heat storage

Heat storage SSV

Electric boiler

Light oil boiler

Wastewater HP

Drinkwater HP

Seawater HP

Industrial surplus HP

Geothermal HP

Solar thermal

Wood pellet boiler

Wood boiler

Wood pellet ST

Straw ST

Waste ST

0

50

100

150

200

250

300

-100 100 300 500

Mar

gin

al h

eat

pri

ce (

DK

K/G

J)

Electricity price (DKK/MWh)

2035 – Without storage

0

50

100

150

200

250

300

-100 100 300 500

Mar

gin

al h

eat

pri

ce (

DK

K/G

J)

Electricity price (DKK/MWh)

2035 – With storageElectric boiler -StudstrupværketLight oil boiler - Gellerup

Wood boiler - Skanderborg

Wood pellet boiler -SkanderborgWaste ST - Lisbjerg

Light oil boiler - Jens Juulvej

Straw ST - Lisbjerg

Solar thermal - Reno Syd

Geothermal HP - Harlev

Seawater HP - Aarhusværket

Wood pellet boiler -StudstrupværketWood boiler - Reno Syd

Drinkwater HP - Node 130

Industrial surplus HP -StudstrupværketWastewater HP - Node 139

Marginal price - Balmorel

0

200

400

600

800

1000

1200

1400

1600

1800

2015 2035

Tota

l hea

t sa

les

(mill

ion

DK

K)

Marginal heat price

Average heat price

0

50

100

150

200

250

300

0 200

Mar

gin

al h

eat

pri

ce (

DK

K/G

J)

Average heat price (DKK/GJ)

2035

Introduction

Modelling

Discussion

Results: Marginal pricing vs average pricing Results: Impact of heat storage on heat prices

Using dynamic pricing is well known in the power sector. The focus of this

poster is the impact of using hourly marginal pricing in district heating

systems in stead of the heat tariffs currently in use.

Definition of marginal heat price: The extra total cost associated with

a marginal increase of heat demand.

Marginal heat price depends on:

• Heat demand, which changes

the level of production in the

merit-order curve.

• Electricity price, which

influences the revenue the

production units.

Impact of heat storage on the marginal heat price is a major focus point

in this poster.

The Aarhus district heating system ismodelled in the least-cost optimizationmodel Balmorel for the years 2015 and2035.Heat demandThe annual heat demand is assumed tobe 2.86 TWh in both years. The totaldemand is distributed over 47interlinked demand nodes.

Heat productionIn 2015, existing units are modelled,where the units in 2035 represent apossible future for the Aarhus districtheating sector, and technologies suchas solar thermal heat and heat pumpsare introduced in the mix.

As the marginal heat price

uses the marginal cost of the

most expensive supplier, it’s

always higher than the

average short-term heat costs.

The variation of the marginal

price is larger (in the order of

50% to 100% larger) than that

of the average price.

0

50

100

150

200

250

300

0 200

Mar

gin

al h

eat

pri

ce (

DK

K/G

J)

Average heat price (DKK/GJ)

2015

Using marginal pricing reflects the true

hourly marginal cost. The total revenue

for the heat producers when using

marginal prices is larger than the total

variable costs. This surplus could be

used to finance (part of) the fixed costs.

The difference is bigger in 2035 than in

2015. It is still possible to keep the non-

profit obligation.

When comparing theresults with storage tothose without storage,several of the marginalprices now lay apartfrom the marginal costlines. In all the hourswhere the realisedmarginal prices (blackdots) are not on asupplier line, one ofthe storages is settingthe price.This happens when theHeat storage is themarginal unit (scenario3 in the table).

0

10

20

30

40

12

66

53

17

96

1,0

61

1,3

26

1,5

91

1,8

56

2,1

21

2,3

86

2,6

51

2,9

16

3,1

81

3,4

46

3,7

11

3,9

76

4,2

41

4,5

06

4,7

71

5,0

36

5,3

01

5,5

66

5,8

31

6,0

96

6,3

61

6,6

26

6,8

91

7,1

56

7,4

21

7,6

86

7,9

51

8,2

16

8,4

81Sto

rage

vo

lum

e (G

Wh

)

2035

Aarhusværket Studstrupværket

0

20

40

60

80

100

T00

1

T00

7

T01

3

T01

9

T02

5

T03

1

T03

7

T04

3

T04

9

T05

5

T06

1

T06

7

T07

3

T07

9

T08

5

T09

1

T09

7

T10

3

T10

9

T11

5

T12

1

T12

7

T13

3

T13

9

T14

5

T15

1

T15

7

T16

3

Stu

dst

rup

svæ

rket

sto

rage

vo

lum

e (M

Wh

)

2035 - week 3

Where the storage in Studstrupværket is used as a daily storage tobalance daily variations in the demand and heat prices, the storage inAarhusværket is used as a seasonal storage. The storage buys duringthe summer when heat demand is low (and solar heating productionis high) and sells heat in the winter hours where the heat prices arehighest. Due to the storage losses, the storage is not used much forbalancing the (smaller) short-term price variations.

Fixed prices for heat delivered to district heating systems do not reflect the actualvalue of the heat. Using hourly marginal prices would give the correct incentive tonew heat suppliers coming to the district heating market, such as industrial surplusheat, or independent heat generators with e.g. solar heat plants or heat pumps.Dynamic pricing would be technology neutral and economically efficient.Heat storages have a large impact on marginal heat prices. Both seasonal and dailystorages can be used in a district heating system such as in Aarhus.Next steps would be to look at practical implementation of dynamic heat prices in adistrict heating system such as Aarhus.

Unit1Unit2

Unit3

Un

it4

Un

it5

Unit6

Consumption

Market price

Mar

gin

al p

rice

Units heat capacity

• Congestion, which ensures different marginal prices on each side of the

congested line.

• Grid losses and fuel prices.

Heat storages1. The heat storage in Studstrupværket (SSV) exists in 2015 and is a

short-term storage which optimises circularly per week. It has nolosses. Storage volume: 2 GWh – Storage (un)load capacity: 333 MW

2. The storage that’s is added in 2035 in Aarhusværket is a largerseasonal storage and can optimise over an entire year. It has a 12.5%loss. Storage volume: 42 GWh – Storage (un)load capacity: 250 MW

Results: Use of heat storages

All marginal prices on the lines Storage sets price in some hours

87.5%Constant heat price

Storage is Price is set by Illustration

1 Selling Heat supplier with highest marginal costs

2 Buying Heat supplier with highest marginal costs

3 Buying untilmarginal heat plant produces at full capacity

Heat storage (dot’s off the cost lines)

Heat

Price

Storage

Heat

Price Demand + storage

Heat

PriceDemand + storage Other hour

In another hour the storage would sell (1-losses) unit of heat less.

The heat price is then (1-losses) times the heat price in that other hour.

If there is an extra unit of heat demand, the storage would buy one unit less.

When the storage has bought all heat from the marginal heat plant, for one extra unit of heat demand, the storage would buy one unit less heat at that hour. This would mean that in another hour, the storage would sell (1-losses) units less. The heat price is then (1-losses) x heat price in the other hour. Two kinds of new prices appear:

• Constant heat prices where the storage transfers prices from hours where the marginal plant is a CHP/el-boiler/heat pump, independent of the power price.

• Heat prices that are introduced by storages losses