Modelling investment decision making in the power sector under imperfect foresight
1. Simulating investment decision making
in the power sector under imperfect
foresight
Kris Poncelet
69th Semi-annual ETSAP Meeting
University College Cork
Cork, Ireland
2. 22016-06-01
Reduce computational
complexity
Allows to increase level
of detail
Trade-off LT time
horizon versus level of
ST detail
Either optimization or
simulation paradigm
Capture imperfect
foresight and short-
term focus of decision
makers
More realistic decision
making?
Simulation paradigm
Myopic optimization: limited window of foresight
4. 42016-06-01
Liberalized electricity markets
Investment decision makers are private utilities
Invest if projected revenues allow a reasonable
internal rate of return IRR (incl. risk premium)
Generation assets have a long lifetime
Future revenue streams are uncertain
5. 52016-06-01
Methodological analysis
2 scenarios:
Perfect foresight (PF)
Myopic foresight (MF10)
Focus on period 2020-2055
5-year periods
Power system inspired by
the Belgian system
Increasing carbon price
LUSYM Investment planning
model:
Partial equilibrium
8 representative days
Clustered unit commitment
7. 72016-06-01
Liberalized electricity markets
Investment decision makers are private utilities
Invest if projected revenues allow a reasonable
internal rate of return IRR (incl. risk premium)
SR profits = revenues β operational costs
{πΈ[ππ ππππππ‘π π¦] Γ
1
1+πΌπ π π¦βπ§π¦ } β₯ πππ₯ππ πππ π‘π
8. 82016-06-01
Perfect foresight scenario
Perfect foresight of:
Electricity prices (and capacity remuneration) in each
time step and all years (implicit)
Generation during each time step and all years
Generation costs (fossil fuel prices, maintenance, etc.)
=> Exactly know SR profits
=> {ππ ππππππ‘π π¦ Γ
1
1+πΌπ π π¦βπ§π¦ } = πππ₯ππ πππ π‘π
12. 122016-06-01
Perfect foresight scenario
Factors which can impact the SR profits:
Fossil fuel prices
Carbon price
Timing of decommissioning existing plants
Type, timing and amount of newly built capacity
Technological progress
Evolution of the electricity demand
Policy interactions
Market design changes
Interconnections
Private utilities do not have perfect information
The assumption of perfect foresight is not realistc
14. 142016-06-01
Perfect foresight scenario - conclusions
{ππ ππππππ‘π π¦ Γ
1
1+πΌπ π π¦βπ§π¦ } = πππ₯ππ πππ π‘π
Perfect foresight for private utilities is unrealistic
Can lead to unrealistic simulation of investment
decisions
15. 152016-06-01
Myopic foresight model
ππ ππππππ‘π π¦ Γ
1
1+πΌπ π π¦βπ§
π+ππΉβ1
π¦=π =
πππ₯ππ πππ π‘π β π πππ£πππ π£πππ’π
Calculation of salvage value typically
based on two assumptions:
1. Total discounted value of an asset
equals the total discounted cost
2. value is distributed homogenously
over the assetβs life time
ππ ππππππ‘π π¦ Γ
1
1+πΌπ π π¦βπ§
π+ππΉβ1
π¦=π =
ππππ’ππππ§ππ πππ₯ππ πππ π‘ Γ
1
1+πΌπ π π¦βπ§
π+ππΉβ1
π¦=π
TLIFE = 5
d = 10%
Time horizon = window of foresight
18. 182016-06-01
Myopic foresight scenario - conclusions
ππ ππππππ‘π π¦ Γ
1
1+πΌπ π π¦βπ§
π+ππΉβ1
π¦=π = πππ₯ππ πππ π‘π β
π πππ£πππ π£πππ’π
Salvage value is determined exogenously
Homogeneous distribution of the value of an asset seems
optimistic (implications for dynamic recursive models)
=> No extrapolation of observed trends
Can lead to unrealistic simulation of investment decisions
Additional issues:
What is the window of foresight?
Window of foresight identical for all uncertain parameters
19. 192016-06-01
{ππ ππππππ‘π π¦ Γ
1
1+πΌπ π π¦βπ§π¦ } =
πππ₯ππ πππ π‘π
Private utilities do not have perfect
foresight on short-run profits
Can lead to unrealistic investment
decisions
ππ ππππππ‘π π¦ Γ
1
1+πΌπ π π¦βπ§
π+ππΉβ1
π¦=π
= πππ₯ππ πππ π‘π β π πππ£πππ π£πππ’π
No extrapolation of observed trends
in short-run profits
Can lead to unrealistic investment
decisions
Investment criterion: {πΈ[ππ ππππππ‘π π¦] Γ
1
1+πΌπ π π¦βπ§π¦ } β₯ πππ₯ππ πππ π‘π
Summary and conclusions
Perfect foresight Myopic foresight
Kris Poncelet, kris.poncelet@kuleuven.be
KU Leuven, energy and environment group
http://www.mech.kuleuven.be/en/tme/resea
rch/energy_environment/
Poncelet, K. et al., Myopic Optimization Models
for Simulation of Investment Decisions in the
Electric Power Sector. 13th International
conference on the European Energy Market, 6-
9 June 2016, Porto.