Future of Global Energy Infrastructures:

1
Future of Global Energy Infrastructures
International Energy Workshop 2004
Model Results of China's Natural Gas Market
Development

• Yaroslav Minullin, Oleg Nikonov, Ivan Matrosov

June 22, 2004 IEA, Paris, France
2
Contents

• Modeling investments decisions
• Model application
• Future development

3
DYN/ECS-IIASA Energy Group (2000-2004)

• Participants
• IIASA (Laxenbrug, Austria)
• Moscow State University (Moscow, Russia)
• Energy Systems Institute (Irkutsk, Russia)
• Energy Research Institute (Moscow, Russia) new
• Urals State Technical University (Ekaterinburg,
  Russia)
• The main results
• 2000 2002 IGOR and G-TIME model, results for
  Turkeys gas market, generalized Nash equilibrium
  solutions
• 2002 G-TIME China with new local-equilibrium
  price formation mechanism
• 2003 Modified theoretical version of the game
  of timing, a new version of software
• 2004 Preliminary results on expansion of the
  game to 8
• players, introduction of
  2-markets game and LNG as a player

4
Russias External Trade
5
General formulation of the problem

• The goal of investors to maximize the future
  profit
• The main variables on which the future profit
  depends are
• the price of gas and costs (fixed, for
  extraction and transportation)
• While the expenses could be estimated, the price
  (under market conditions) depends on the gas
  demand and volume of gas delivered to the market
• The last, in turn, depends on the number of
  participants who operate in the market

6
General formulation of the problem

• Thus, there is a game situation where the key
  parameters are times, when the participants enter
  the market.
• This concept has been formalized in the models
  oriented to Turkeys market and, in principal, to
  other countries-consumers with the market
  economy.

7
Modeling Framework

• Assessments of the market and projects
• Regulation of future supply and optimization of
  current investments (instantaneous supply game)
• Selection of investment and operation scenarios
  (game of timing)

8
Modeling Framework

• N-person game
• N participants (players), i 1, , N xi -
  control variables
• Each participant has his own payoff function
• fi(x1, x2,,xN), which is to be maximized
• by choosing xi
• The N-tuple (x1 x2,,xN) is called Nash
  equilibrium, if for every i the following
  equality is true
• fi(x1,,xi,,xN) max fi(x1,,
  xi,,xN),
• where maximum is taken over all admissible xi

9
Export Routes of Russian Gas
Source Energy Research Institute of RAS, 2002
10
Chinas Gas Market
Source Energy Systems Institute of RAS, 2002
11
Chinas Energy SectorSpecific Features

• Some features of a Planned Economy
• Lack of econometric data
• Low correlation between GDP and sectors incomes

12
Price Formation
13
Gas Market Model
Extraction
Extraction
Deposit 1
Project 1
Deposit 2
Project 2
Transportation
Transportation
Supply
Forecasted demand
Natural-gas market
Forecasted price
Forecasted price elasticity
pp(d0,p0,ep,y)
14
Supply Game One supplier
Maximum payoff
?i
?i max
t - fixed

• ?i(y,yi) p(d0,p0,ep,y)-c(yi)yi

D
ymin
Mi
d0
yi

• At each instant of time the player maximizes
  payoff
• ?i(y,yi) ? max
• yi?D
• and gets an optimal value for supply

15
Supply Game Two suppliers
y2
Nash equilibrium point (y1 (y2) y2 (y1))
Best response y2 (y1)
d0
M2
ymin
Best response y1 (y2)
D
M1
y1
16
Supply Game evolving over time
yi
d0(t)
ymin(t)
yi1(t)
Mi
yi2(t)
t
Each player gets an optimal supply plan in
time yi1(t) and yi2(t)
17
Supply Game Benefits
Upper benefit rate, bi1(t)
bi1(t) bi2(t)
Lower benefit rate, bi2(t)
Bi(t,t-i)
t, t-i
t-i
Substituting optimal supply the player gets
upper benefit rate, bi1(t) ?i(t,y(t),yi1(t)) and
lower benefit rate, bi2(t) ?i(t,y(t),yi2(t))
18
Game of Timing
Pi(t)
Start of making investments
Start of operation
Return of investments
ti
ti0
Ci
t, ti
tiROI
t1ROI (t10 , t20) t0 ? min
t10 t2ROI (t20 , t10) t0 ? min
ti0
Nash equilibrium point of starting construction
times (t10 t20 ti0)
19
Projects
Player Project Length, km Expected starting year Maximal capacity, bcm/year Invest-ments, bln USD Variable costs, (extr. tr.) USD/1000m3
1 Irkutsk 2795 2008 20 6.5 111
2 Sakhalin 1800 2015 12.5 3.5 100.9
3 Sakha 4200 2018 17.5 8.8 166.83
4 Altai 1200 2007 10 2.1 87.4
5 West Siberia 6100 2020 30 15 176.4
6 Turkmenistan 6200 2016 25 13 167.6
7 Kazakhstan 5100 2016 25 12 160.05
8 South-North 1500 2007 25 - 20
8 LNG - - 6-42 - world price
20
New Assumptions

• China as a consumer is considered to be a system
  of 2 gas markets NE and SE. Each market is
  characterized by price, demand and elasticity
  forecasts
• There is a South-North pipeline, connecting both
  markets. It is assumed that gas can flow in both
  directions. It is characterized by maximal
  capacity and cost for transportation

21
New Assumptions

• In the NE region there is an LNG terminal, which
  is characterized by maximal capacity and world
  LNG prices
• Some of the players are present on both markets,
  but have one optimality criterion (coalition)
• Times of penetrating the market are fixed and
  determined by the earliest technical possibility
  (no game of timing)

22
Results of Simulations NE regionSupply curves
23
Results of Simulations SE regionSupply curves
24
Results of Simulations Supply on both markets
25
Results of Simulations North-East pipeline
26
Results of Simulations NE regionReturn on
Investments
27
Results of Simulations SE regionReturn on
Investments
28
Results of Simulations Prices
29
Results of research direction

• An information support for decision-makers and
  energy market analysts, which provides
• A versatile economic evaluation of the selected
  project and the market
• Advices on the optimal timing, investment flow,
  and supply regulation
• An assessment of impacts on the energy balances,
  the economy and the environment in importing as
  well as exporting countries

30
Next Steps

• Modify the model to be able to reflect risk,
  uncertainty, and data scarcity
• To expand the game to 3 markets in China
• To introduce the governments of importing and
  exporting countries as players
• Globalization Europe lt - Russia - gt Asia
• Integration with higher-level models