Grimm

Transmission and Generation Investment in Electricity Markets: the Impact of Market Splitting and Network Fee Regimes

Veronika Grimm Barcelona February 3, 2015 joint work with A. Martin, M. Schmidt, C. Sölch, M. Weibelzahl, and G.Zöttl

A Grand Challenge ● Abandoning nuclear energy requires complete reorientation of power supply schemes. ● Old plants get dismanteld or need repowering. ● A lot of fluctuating renewable sources have been installed.

● We need market rules that generate adequate investment incentives: => adequate capacities

=> at the right locations Prof. Dr. Veronika Grimm, FAU & EnCN Economy

Quelle: Bundesamt für Strahlenschutz 2

Transmission constraints become an issue  Transmission constraints become relevant – both within

Congestion

and between countries.

No congestion

 Possible solutions include: gas power plants, network capacity,

demand side management, storage facilities and smart technologies  The locations and capacities of generation facilities have crucial relevance for the network expansion.

Source: EWI, Trendstudie 2022. Case: high wind in-feed.2022. Prof. Dr. Veronika Grimm, FAU & EnCN Economy

3

Related Literature “Old World”: Integrated planer problem  Optimal expansion for generation and transmission capacities Gallego et al. (1998), Binato et al. (2001), and others 



Investment models for generation facilities (e.g. peak load pricing literature, “Capacity-market”-discussion).  typically disregards network and network expansion (“copper plate”) Gaszewicz & Poddar (1997), Murphy & Smeers (2005), Grimm & Zöttl (2013) 

Models on optimal transmission planning in anticipation of private investment  Typically asume optimal management of the network (nodla pricing) Sauma and Oren (2006, 2009), Jin and Ryan (2011), Jenabi et al. (2013)

Prof. Dr. Veronika Grimm, FAU & EnCN Economy

05.02.2015

#4

Related Literature 

Models analyzing impact of different network management regimes (nodal pricing, zonal pricing, redispatch)  typically focus on the short run perspective (given network & generation facilities) Hogan (1999), Ehrenmann and Smeers (2005), Neuhoff et al. (2005), … 

Analyses of the ability to exercise market power under different network management regimes Oren (1997), Jing-Yuan & Smeers (1999), Oggoni & Smeers (2013)



This paper: models transmission planning by a regulator in anticipation of private investment in an energy only market with redispatch

Prof. Dr. Veronika Grimm, FAU & EnCN Economy

05.02.2015

#5

Questions we have in mind (examples) 

What is the quantifiable impact of adopting a different transmission management regime (e.g. price zones,…, nodal pricing) taking into account long run investment?



what is the impact of changed way of charging network fees on generation investment and associated network expansion? 

This paper: lump sum, capacity based, energy based



Current work: G- and L-Component, regional differentiation



Incentives under Cost Based vs. Market Based redispatch (different paper, short run)



What are the incentives to invest in responsive consumption units and what is the impact on optimal transmission investment?



We present a computable equilibrium framework which allows to analyze those issues

Prof. Dr. Veronika Grimm, FAU & EnCN Economy

6

Roadmap of this talk

(1)

Introduction

(2)

Computational Equilibrium Framework

(3)

Testexample (6-node-network)

(4)

Very first results on Germany&Neighbours

(5)

Conclusion

Prof. Dr. Veronika Grimm, FAU & EnCN Economy

05.02.2015

#7

What we have in mind Model Components

Illustration

 Network expansion by social planner  Competitve Firms invest in different production technologies throughout the network  Demand at the nodes (net of renewable feed-in) can be fluctuating and uncertain.  We want to explicitly take into account impact of different network management regimes (redispatch, market splitting) Main purpose: to identify the impact of market rules on investment decisions (overall system optimization is just a benchmark!) Prof. Dr. Veronika Grimm, FAU & EnCN Economy

8

Model: Timing  The transmission system operator chooses to realize line investments from set of options (integer decisions).  Competitive firms choose how much to invest in available production technologies at each node t=1,2,… ,each technology (kt,ct) has marginal cost of production ct, marginal cost of investment kt at the supply node.  Spot market competition  Management of network congestion by cost based redispatch.

Transmission Investment (Planner)

Generation Investment (Firms)

Prof. Dr. Veronika Grimm, FAU & EnCN Economy

Spot Markets (with coupling/splitting) and redispatch after each market

05.02. #9 2015

Model Components: modelling the physical network ● We consider the usual linear lossless DC-Approximation:

resistance: 2 therm. capacity: 40 MW flow: 25 MW resistance : 1 therm. capacity: 40 MW flow: 25 MW

100 MW resistance : 1 therm. capacity: 80 MW flow: 75 MW

Prof. Dr. Veronika Grimm, FAU & EnCN Economy

100 MW

10

Model Components: Network Management Regimes Cost based Redispatch:  All bids at the spot markets are made entirely independently of network constraints, we obtain a uniform price accross the entire market.  Quantities traded may be physically unfeasible. Then the TSO has to find the cheapest possible re-dispatch to make final quantities physically feasible. Market Splitting:  The market region is divided into price zones, potential congestion among zones (but not within zones!) is already taken into account at the spot markets.  Remaining physical infeasibilities are still resolved through redispatch.

Prof. Dr. Veronika Grimm, FAU & EnCN Economy

11

Model Components: Network Fees The TSO is facing the following cost:  Network expansion investment  Cost of redispatch In our framework TSO is supposed to not make any profits, the above spendings have to be recovered by network fees. We consider the following cases:  lump sum  energy based fees (e.g. Germany, 5 €ct/KWh)

 capacity based fees  Fees payed either by generators or by consumers

Prof. Dr. Veronika Grimm, FAU & EnCN Economy

12

Illustration of our 3-stage approach

Network Expansion (social planner) Investment in Generation Facilities Trading at Spot Markets (competitive companies) Redispatch taking into account renewable production (social planner)

Prof. Dr. Veronika Grimm, FAU & EnCN Economy

13

Our 3-stage approach, more formally Max Welfare(N,K,S,R) s.t. K,S is competitve equilibrium, s.t. Traded quantities S can be produced by capacities K Min REDCost(N,K,S,R) s.t. quantities can be transmitted by network and can be produced by plants

Prof. Dr. Veronika Grimm, FAU & EnCN Economy

Network expansion-stage: Social planner chooses network(expansion) maximizing WF Market-stage: Competitive Firms choose capacities and Spotmarket-bids to maximize profits.

Redispatch-stage: Social planner chooses Redispatch R to minimize Redispatchcost REDCost, s.t. all quantities are feasible.

14

Benchmark: system optimization / first best Max Welfare(N,K,S,R) s.t.

Integrated perspective: Social planner chooses network(expansion), generation investment and production to maximize Welfare

Production schedule is feasible

Transmission is feasible

Prof. Dr. Veronika Grimm, FAU & EnCN Economy

s.t. feasibility constraints.

15

Computational Results, 6 node test example ● To test our equilibrium framwork we consider a common 6-node-example (adapted for long run decisions). ● Lines connecting nodes 1,2,3 and nodes 4,5,6 have sufficient capacities. Only lines 1-6 and 2-5 cause problems. Potential line investment 1-6 and 2-5. ● Three demand nodes (3,5,6). ● Investment in generation facilities only at the supply nodes (1,2,4) ● Notice: Storage facilities are not (yet) included.

Demand

3 1

1 1

Supply 1

2 Supply

2 Demand

2

6

5 Demand

1 1

1

4 Supply

existing line candidate line Prof. Dr. Veronika Grimm, FAU & EnCN Economy

16

Computational Results, 6 Node Test Example

Demand

3 1 Inv.: 700 €/MW

1

Inv.: 600 €/MW

1

Supply 1

2 Supply

Var.: 10 €/MWh 2 Demand

Var.: 15 €/MWh

2

6

5 Demand

1 1

1

● We used German spot market data from 2011 to generate 52 demand scenarios.

4 Supply Inv.: 200 €/MW Var.: 42.5 €/MWh

Prof. Dr. Veronika Grimm, FAU & EnCN Economy

existing line candidate line

17

6 node test example, scenarios analyzed Scenario: Single Zone

Scenario: Two Zones

Zone North

Spot- & Redispatch-Markt

Zone South

Prof. Dr. Veronika Grimm, FAU & EnCN Economy

18

Computational Results, 6 Node Test Example Benchmark (fist best)

Single Zone

Two zones

Welfare (norm.):

 1

 0.93

 0.98

Generation. Invest.:

 All locations

 Only node 1

 Only nodes 1 and 4

Network Invest.:

 Build no line

 Build both lines

 Build 2-5

Prof. Dr. Veronika Grimm, FAU & EnCN Economy

19

 Redispatch leads to underinvestment in nodes 2 and 4.  Energy based fees could potentially aggravate problems of

overinvestment in node 1.  Splitting in separate zones (only) partially overcomes those problems!

Prof. Dr. Veronika Grimm, FAU & EnCN Economy

20

Summary of Results

Prof. Dr. Veronika Grimm, FAU & EnCN Economy

21

Prices

Prof. Dr. Veronika Grimm, FAU & EnCN Economy

22

6 node test example, Summary of Results ● Under Cost Based Redispatch Regime investment in generation facilities in the „South“ is too low and network investments are too high (relative to the first best). ● Energy based fees potentially aggravate problems of overinvestment in the „North“. ● Consideration of different regions already at the spot market (market splitting) would aleviate but not eliminate distortions ● Perspective: Our framework allows to precisely quantify all those differences, also for detailled calibration of specific market regions.

Prof. Dr. Veronika Grimm, FAU & EnCN Economy

Demand

3 1

1 1

Supply 1

2 Supply

2 Demand

2

6

5 Demand

1 1

1

4 Supply

existing line candidate line 23

Regional Model "Electricity Transport 2013"  8784 hours (= year 2012)  20 regions for Germany:  2 regions for off-shore wind energy plants (North and Baltic Sea),  18 regions on the German mainland  9 regions for neighboring countries:   

     

Austria, Belgium, Switzerland, Czech Republic, Denmark (West), France, the Netherlands, Poland, Northern Europe (Denmark East, Norway, Sweden)

Prof. Dr. Veronika Grimm, FAU & EnCN Economy

24

First Results II

Prof. Dr. Veronika Grimm, FAU & EnCN Economy

25

First Results II First best Solution

Market Solution (Cost based redispatch)

26 1

4

8

10

11

1 27

7

6 9

2

3 5

21

26

15

25

17

23

10

27 14

13

12

15

16 18

8

7

11

12

22

4

6 9

14

13

3 5

21

2

25

17 16

19

20 24

Prof. Dr. Veronika Grimm, FAU & EnCN Economy

22

18 23

19

20 24 26

Summary ● We have established a framework where a planner chooses transmission line investment and competitive firms invest in generation facilities. ● The framework allows to explicitly analyze the impact of different network management regimes (network fees, price zones,…) on generation and network investment. ● First qualitative results based on test example: 1)

Redispatch leads to underinvestment in the „South“.

2)

Energy based fees aggravate problems of overinvestment in the „North“.

3)

Splitting in separate zones only partially overcomes those problems!

● Future work: analyze regional differentiation of transmission fees, those might at least partially heal the problems!

Prof. Dr. Veronika Grimm, FAU & EnCN Economy

05.02.2015

# 27

Thank you for your Attention!

Used data and parameters Data for 2012 from:  eex.com (German prices)  entsoe.eu (Consumption)  Transparency homepages of TSOs (solar, wind, cross border physical flow)  Electricity market homepages of neighboring countries (prices)  … Parameters:  Price elasticity: -0.25 => slope of demand function: -4  Generation technologies: Type Nuclear

Investment cost (€ / (MW * a))

Variable cost (€)

no new investment

Lignite

235730

10,00 27,32

Hard coal

202330

40,69

80100

73,68

Gas Prof. Dr. Veronika Grimm, FAU & EnCN Economy

29

First Results I  First best model vs. Redispatch model (single zone, lump sum)  Without net investment vs. (forced) investment in 1 line: high-voltage DC-link  start: Lauchstädt (Saxony-Anhalt)  end: Meitingen (Bavaria)  capacity: 2 GW  length: 450 km  cost: 1.40 m €/km  annuity: 0.11 m €/(km*a)

Prof. Dr. Veronika Grimm, FAU & EnCN Economy

30

Prof. Dr. Veronika Grimm, FAU & EnCN Economy

31

First Results I Benchmark (first best) Welfare (p.c.): No line invest. Forced line invest.

Generation. Invest.: No line invest. Forced line invest.

 100.00 %  99.99 %

Single Zone

 96.37 %  96.38 %

 Build Gas (596 MW) in Baden-Wuerttemberg  No investment  Build Gas (414 MW) in Baden-Wuerttemberg  No investment

Prof. Dr. Veronika Grimm, FAU & EnCN Economy

32

Complex Solution  What is the best mix of the available technologies 

in the short run and



in the long run?

 How does a market environment look like that makes us achieve those goals?  We analyze investment incentives in different market environments

 Solution to a central planer problem

Prof. Dr. Veronika Grimm, FAU & EnCN Economy

33

Neighboring Countries  Derivation of the Export Function

Prof. Dr. Veronika Grimm, FAU & EnCN Economy

34

Neighboring Countries  Welfare Maximization

Pr ice

Pr ice

Expor t pr ice

impor t pr ice

Quant it y

Quant it y Expor t

a) Th e expor t case Prof. Dr. Veronika Grimm, FAU & EnCN Economy

impor t

b) t h e i mpor t case 35

Prices (first best model without line investment) in hour 8784                          

price#8784#1 price#8784#2 price#8784#3 price#8784#4 price#8784#5 price#8784#6 price#8784#7 price#8784#8 price#8784#9 price#8784#10 price#8784#11 price#8784#12 price#8784#13 price#8784#14 price#8784#15 price#8784#16 price#8784#17 price#8784#18 price#8784#19 price#8784#20 priceforeign#8784#21 priceforeign#8784#22 priceforeign#8784#23 priceforeign#8784#24 priceforeign#8784#25 priceforeign#8784#26

Prof. Dr. Veronika Grimm, FAU & EnCN Economy

27.223569 10.841658 27.455101 27.301276 27.758322 27.701336 27.836173 27.589864 28.097541 27.731295 27.851197 27.702865 27.700879 27.699153 28.094892 27.719492 27.845780 28.048212 27.515344 27.757710 27.728219 27.210806 27.443086 27.548400 27.255895 27.307127 36