Managing Risk in Hydro-Based Portfolios: the Brazilian Experience - - PowerPoint PPT Presentation

Markets, I nvestments and Risks in Hydro vs Thermal-Dominated Systems

The Energy Centre – U of Auckland Business School

Mario Pereira mario@psr-inc.com

Managing Risk in Hydro-Based Portfolios: the Brazilian Experience

2

Topics

• Brazilian system overview
• Hydrothermal scheduling
• Risks and challenges
• Tools for risk management

3

Topics

• Brazilian system overview
• Hydrothermal scheduling
• Risks and challenges
• Tools for risk management

4

The big numbers...

Brazil

Argentina Peru Colombia Venezuela Bolivia Uruguay Paraguay

Surface area: 8 million km2 (= continental US + 1/2 Alaska) Population: 185 million GDP: US$ 800 billion Installed capacity (2006): 100 GW Production (2006): 50 GW average Peak load: 65 GW

5

Energy production sources

Hydro (85%): large plants in cascade, in several river basins,

with multiple weather patterns

Thermal (15%): natural gas (combined and simple cycle); coal;

heavy fuel; diesel; nuclear; sugarcane biomass cogen

CEMIG FURNAS AES-TIETÊ CESP CDSA Consórcios COPEL TRACTEBEL

ITAIPU Binacional

Rio Grande Rio Paranaíba Rio Tietê Rio Paranapanema Rio Iguaçu CEMIG FURNAS AES-TIETÊ CESP CDSA Consórcios COPEL TRACTEBEL

ITAIPU Binacional

Rio Grande Rio Paranaíba Rio Tietê Rio Paranapanema Rio Iguaçu

6

Transmission network

Country is interconnected by 80 000 km of HV lines (>230 kV) 2200 MW interconnection with Argentina Long transmission lines (> 1 000 km) 15 000 km of new lines added in the past five years Auctions for the construction of grid reinforcements

Source: ONS, www.ons.org.br

3500 km 2800 km

7

G, T and D Sectors

• Generation

– 11 major utilities + several smaller companies – 15% private (energy produced) – Total revenues (2005) : US$ 13 billion

• Transmission

– 35 companies (27 private) – Total revenues (2005) : US$ 3 billion

• Distribution

– 64 utilities – 80% private (energy consumed) – Total revenues (2005): US$ 27 billion

8

Investment needs

• For a GDP growth of 4%, it is necessary to

install 3200 MW average of new firm energy per year ⇒ US$ 6 billion/year in investments

Main objective: to ensure an efficient capacity increase

9

Resources for generation expansion

 Northeast (NE):

• offshore natural gas and
• il; LNG and coal

imports; biomass (sugarcane); wind

 Southeast (SE):

• Hydro; Bolivian gas + local
• ffshore gas fields

(Campos and Santos); biomass (sugarcane)

 South (S):

• Electricity and gas

imports from Argentina; local coal; binational hydro plants; LNG

 North (N)

• Substantial hydro (170 GW);

limited natural gas

10

Topics

• Brazilian system overview
• Hydrothermal scheduling
• Risks and challenges
• Tools for risk management

11

System dispatch

• The National System

Operator (ONS) controls the production

• f all hydro and thermal

plants

• Hydro plants are

dispatched as a portfolio, to take advantage of hydrological diversity (export from “wet” to “dry” basins)

12

The Hydrothermal (HT) scheduling problem

• Formulated as a stochastic DP recursion

– Objective: minimize the present value of expected operation cost (fuel cost for thermal plants + penalties for rationing) taking into account inflow uncertainty – State variables: reservoir storage levels and observed lateral inflows at each reservoir

• For a system with 50 hydro plants and an

autoregressive lag-3 model, this results into 200 state variables ⇒ Discrete stochastic DP cannot be used (curse of dimensionality)

13

The SDDP scheme

• A stochastic dual dynamic programming algorithm

(SDDP) is used to solve the dispatch problem

– the future cost function (FCF) is represented by piecewise linear hyperplanes (Benders cut)

• no discretization necessary
• The hyperplane coefficients are the dual values of the dispatch

problem (hence the name)

• The SDDP scheme has been applied to more than 40

countries in Latin America, Europe, Eurasia and Oceania

14

Spot price

• In addition to energy production schedule, the HT

scheduling model calculates the system short-run marginal cost (SRMC)

– Related to the opportunity cost of water (water value)

• The SRMC is used as a proxy of spot prices in all

wholesale energy market (WEM) transactions

15

Zonal prices

Although bus- level LMPs are calculated, a zonal system with 4 regions is used for WEM transactions

The main transmission network has 3500 buses and 5000 circuits

16

Topics

• Brazilian system overview
• Hydrothermal scheduling
• Risks and challenges
• Tools for risk management

17

Average energy inflow – Southeast region

18

Average storage level – Southeast region

19

Average spot price – Southeast region

20

Energy inflow scenarios – Southeast region

21

Storage scenarios – Southeast region

22

Spot price scenarios – Southeast region

23

The spot price distribution is skewed

50 100 150 200 250 300 350

1 8 15 22 29 36 43 50 57 64 71 78 85 92 99 106 113 120 127 134 141 148 155 162 169 176 183 190 197

US$/MWh média

24

Low prices for a long time, punctuated by “spikes”

10 20 30 40 50 60 70 80 90 100 jan-93 jul-93 jan-94 jul-94 jan-95 jul-95 jan-96 jul-96 jan-97 jul-97

US$/MWh

25

Challenges for new capacity

• Because of price volatility, it is very risky for

any generator (hydro or thermal) to enter the system as a merchant plant

• The uncertainty is compounded by the

variability of load growth

40 45 50 55 60 65 70 2006 2007 2008 2009 2010 2011

GWmédio

[High - Base]: 2700 MW average [Base – Low]: 700 MW average

High GDP 5% Base GDP 4% Low GDP 3.5%

26

Topics

• Brazilian system overview
• Hydrothermal scheduling
• Volatility of spot prices
• Tools for risk management
• Contract auctions
• Forward contracts for hydro
• Call option contracts for thermal plants

27

Supply contracts

• All consumers (free and regulated) should be

100% contracted

– Verified ex-post, for the cumulative energy consumption in the previous year

• Although contracts are financial instruments

(forward or call options), they must be “backed” by a firm energy “certificate”

28

100% contract + firm energy ⇒ expansion

Should be 100% contracted; looks for a genco or a trader

Load increase Genco

New generation

29

Contract auctions

• Discos contract energy through auctions

– Discos are responsible for load forecast; avoids government planners’ “optimism” – Contracts reduce risks for investors; lower prices

• Free consumers can contract as they wish, as

long as they remain 100% covered

– Free consumers are 25% of the market – They serve as “checks and balances” for the regulated sector

30

Auction results 2004-2006

• 5 auctions; US$ 50 billion in contracts

31

Next auctions are scheduled for May 2007

Wind, 14 Small hydro, 31 Hydro, 11 sugar cane BM, 55 Coal, 8 Cogen, 6 Natural Gas, 11 Oil, 69

205 candidate projects; 25 thousand MWs

32

Topics

• Brazilian system overview
• Hydrothermal scheduling
• Volatility of spot prices
• Tools for risk management
• Contract auctions
• Forward contracts for hydro
• Call option contracts for thermal plants

33

Risks of forward contracts for hydro plants

• For thermal plants, forward contracts are Ok

– Hedge against low spot prices – If the spot price is high, the plant will dispatch; the worst expense is the fuel cost

• However, significant risks remain for hydro

34

Risk 1: variable hydro production

5 10 15 20 25 30 35 40 45 1 2 3 4 5 6 7 8 9 10 11 12 H1 5 10 15 20 25 30 35 40 1 2 3 4 5 6 7 8 9 10 11 12 H2 10 20 30 40 50 60 1 2 3 4 5 6 7 8 9 10 11 12 H3

The production of individual hydro plants is quite variable; long periods in which the plant may be “short” on the contract

35

5 10 15 20 25 30 35 40 45 1 2 3 4 5 6 7 8 9 10 11 12 H1 5 10 15 20 25 30 35 40 1 2 3 4 5 6 7 8 9 10 11 12 H2 10 20 30 40 50 60 1 2 3 4 5 6 7 8 9 10 11 12 H3

10 20 30 40 50 60 70 1 2 3 4 5 6 7 8 9 10 11 12 H3 H2 H1 10 20 30 40 50 60 70 1 2 3 4 5 6 7 8 9 10 11 12 H3 H2 H1

Idea: total hydro production is more stable Solution: spatial hedging

36

Energy credit x physical hydro production

50 100 150 200 250 300 1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86 91 96 101 106 Month GWh Physical production Energy credit

The spatial hedging scheme (MRE)

• All hydro plants are “shareholders” of a “hedge fund”

called MRE

• The total hydro production is assigned to MRE
• It is then allocated to each plant as an “energy credit”, in

proportion to the shares, not to the physical production

• The energy credits are used for the WEM clearing

37

Risk 2: financial exposure in dry periods

10 20 30 40 50 60 70 80 90 100

jan/00 mai/00 set/00 jan/01 mai/01 set/01 jan/02 mai/02 set/02 jan/03 mai/03 set/03 jan/04 mai/04 set/04 jan/05 mai/05 set/05

storage level (% max) 100 200 300 400 500 600 700 spot price (R$/MWh) Spot price Storage level 10 20 30 40 50 60 70 80 90 100

jan/00 mai/00 set/00 jan/01 mai/01 set/01 jan/02 mai/02 set/02 jan/03 mai/03 set/03 jan/04 mai/04 set/04 jan/05 mai/05 set/05

storage level (% max) 100 200 300 400 500 600 700 spot price (R$/MWh) Spot price Storage level

Hydro plants have a two-sided risk: if they contract too little, they will “starve” in wet periods; if they contract too much, they are “hurt” by high spot prices in dry periods

38

Solution 2a: Contract adjustment in a crisis

• In case of rationing, the contracted amount of all

plants is reduced in the same % as the load curtailement

– Alleviates exposure to very high prices in crisis situations; risks are transferred to consumers

39

Solution 2b: Optimize energy contracted

• For each candidate contract amount, calculate price

that ensures the required return on investment

– e.g. “Value at Risk” on IRR: Pr [IRR > target] > 95% – Stochastic optimization model (OptFolio)

• Select energy amount that maximizes plant

competitiveness in auctions

Preço de Contrato para a contratação máxima (R$/MWh): 121.75 Mínimo Preço de Contrato (R$/MWh): 121.00 Contratação Ótima (%) do Lastro: 96%

118.0 119.0 120.0 121.0 122.0 123.0 124.0 125.0 126.0 Quantidade (% Lastro) Preço Preço 125.6 124.6 123.6 122.6 121.8 121.1 121.0 121.0 121.1 121.3 121.8 90% 91% 92% 93% 94% 95% 96% 97% 98% 99% 100%

40

Topics

• Brazilian system overview
• Hydrothermal scheduling
• Volatility of spot prices
• Tools for risk management
• Contract auctions
• Forward contracts for hydro
• Call option contracts for thermal plants

41

Auctions for “Call option” contracts

• “Call option” contract auctions for thermal plants have

been used since 2005

– Plants bid both the “premium” (fixed annual revenue) and the “strike price” (used as the variable operating cost in the HT dispatch)

• Bids are compared with basis on the estimated

benefit for consumers

– [low premium, high strike] x [high premium, low strike]

• Objective: transfer benefits (and risks) of

hydrothermal optimization to consumers

42

Conclusions

• Load growth uncertainty and spot price volatility

create important risks for generation investors, in particular for hydro plants

• These risks can be handled by a set of technical,

regulatory and financial instruments: – Stochastic optimization for hydrothermal dispatch – “Competition for the market” (long-term contract auctions)

• Discos are responsible for load forecasts

– “Spatial hedging” and forward contract

• ptimization for hydro plants

– Call option contracts for thermal plants