Article ID Journal Published Year Pages File Type
479592 European Journal of Operational Research 2015 16 Pages PDF
Abstract

•The long-run impact of expansion to the evolution of electricity price is sought.•Endogenous and stochastic exogenous models are compared.•Optimised markets lead to moderate electricity prices and carbon emissions.•The assigned load levels are raised in profitable producers.•Endogenous models exhibit lower levels of uncertainty.

A decision support tool is proposed for optimising the expansion planning of a semi-liberalised electricity market, whilst the underlying interaction of the generating mix with electricity prices is researched, in the long-run. A nonlinear stochastic programming algorithm is used for handling multiple uncertainties, optimising the power sector characteristics, both in terms of financial and environmental performance. Two endogenous models and an exogenous one are analysed and compared. The endogenous model results indicate that consumers might benefit by the moderate electricity prices in case the optimal loads and capacity orders are rendered. The exogenous model is insensitive to generating mix variations. The long-term actions suggested for system operators are comprised of the permits issued for new entries. They are affected by the evolution of electricity prices, since the permits granted for conventional technologies are maintained when their profits are rising. The permits granted for renewable technologies are also maintained, thus allowing cleaner electricity production to be induced to the grid. The optimal bid strategies of generators interact with their dispatching schedule and the diversification of their load curves. The relevant bids are primarily driven by the merit order, the plants are dispatched in. The assigned load levels may be raised in profitable producers so that their profit is maximised. They might be restrained instead, in case there are no significant prospects for individual profits. The lognormal distribution of electricity price results is characterised by increasing variance over time, indicating that the model is more robust in the most imminent solutions.

Related Topics
Physical Sciences and Engineering Computer Science Computer Science (General)
Authors
,