Environmental Engineering Reference
In-Depth Information
The traditional framework applies to a generating portfolio that is typically kept
constant over the evaluation horizon (say, 20 years). It can be the current portfolio
in a given country, or a target portfolio assumed to be in place sometime in the
future.
Here we consider a generating portfolio in a dynamic context. We recognize the
fact that the
eet of power plants changes over time as new stations connect to the
electric grid and older ones cease operation. Further, we evaluate the performance
of several generating portfolios in face of a common stochastic path of future
demand. There is more to these real facilities than to
nancial assets, so other
metrics beyond expected price and price volatility can be of interest too. Indeed,
investors, utilities and policy makers aim at different goals, so the most relevant
variables can differ among them.
We develop a valuation model that rests on cost minimization. Our measure of
cost naturally includes that of power generation and of unserved load. Regarding
the former, power producers under the EU ETS face both stochastic fuel prices and
carbon allowance prices. As for the latter, in our model lost load has a non-
negligible cost.
Uncertainty in our model extends beyond economic variables. It affects the state
of physical infrastructures and/or their output. In sum, we deal with a problem of
stochastic optimal control.
At any time, the optimization algorithm provides the level of power generation
by technology, served load, aggregate generation costs, carbon emissions, and
allowance costs, among other variables. The optimization model is nested in Monte
Carlo simulation. A single run determines a number of state variables over 60
20
200 consecutive time steps. Under each setting, the optimization problem is
solved. Therefore, one simulation run involves 1,200 optimizations. We repeat the
sampling procedure 750 times. We thus come up with 750 time pro
¼
1
;
les of each
variable of interest. In particular, our model can assess the performance of a pre-
speci
eet in terms of the resulting expected price and the standard
deviation around that expectation. When several generating portfolios are consid-
ered, comparing their relative performance sheds light on their respective advan-
tages and weaknesses.
We illustrate the model by example. Speci
ed generation
cally, we look at the British power
generation mix over the time horizon 2012
2032. The 2012 Electricity 10 Year
Statement envisages three future energy scenarios alongside the contracted back-
ground. Under Gone Green, the renewable target for 2020 and the emissions targets
for 2020 and 2030 are all met. Under Slow Progression, instead, the 2020 target is
not met until between 2020 and 2025; and the 2030 target is not achieved. Under
Accelerated Growth renewable and carbon reduction targets are all met ahead of
schedule. The Contracted Background portfolio refers to all projects that have a
signed connection agreement with National Grid; reductions and closures with an
explicit noti
-
cation or date are also taken into account. Note that, as of 1 April
2013, the UK Government introduced a carbon price support mechanism. It aims at
a carbon price
oor around 16
₤
/tCO2 in 2013, 30
₤
/tCO2 in 2020, and 70
₤
/tCO2
in 2030 (2009 prices).
Search WWH ::
Custom Search