Measuring Performance of Long-Term Power Generating Portfolios - Green Energy and Efficiency: An Economic Perspective - page 387

Measuring Performance of Long-Term
Power Generating Portfolios

Jos M. Chamorro, Luis M. Abadie and Richard de Neufville

é

Abstract We propose a model for assessing the performance of generation mixes
in a mean-variance context. In particular, we focus on the expected price of elec-
tricity and the price volatility that result from different generating portfolios that
change over time (because of investments and retirements). Our valuation model
rests on solving an optimization problem. At any time it minimizes the total costs of
electricity generation and delivery. A distinctive feature of our model is that the
optimization process is subject to the behavior of stochastic variables (e.g. load,
wind generation, fuel prices). Thus we deal with a problem of stochastic optimal
control. The model combines optimization techniques, Monte Carlo simulation over
the decades-long planning horizon, and market data from futures contracts on
commodities. It accounts for uncertain dynamics on both the demand side and the
supply side. The aim is to assist decision makers in trying to assess electricity
portfolios or supply strategies regarding generation infrastructures. To demonstrate
the model by example we consider the case of Great Britain s generation mix over

the next 20 years. In particular, we compare three future energy scenarios and the
contracted background, i.e. four time-varying generating portfolios. Major British
power producers are covered by the EU Emissions Trading Scheme (ETS), so they
operate under binding greenhouse gas (GHG) emission constraints. Further, the UK
Government has announced a

'

oor price for carbon in the power sector from 1
April 2013. The generation mix is optimally managed every period by changing
input fuel and electricity output as required.