Scenario Reduction for Stochastic Optimization Applied to Short-term Trading of PV Power

Abstract

This paper addresses the scenario reduction for stochastic optimization applied to short-term trading of photovoltaic (PV) power. Stochastic optimization becomes a useful technique when leading with problems involving uncertainty. Short-term trading of PV power in electricity markets is an example of a problem involving a high level of uncertainty, namely uncertain parameters as PV power and market prices. As the level of uncertainty grows and the optimization problem becomes more complex, the need to reduce the number of scenarios becomes crucial without losing the representativeness of the original scenarios. Thus, in this paper is proposed an efficient scenario reduction algorithm based on backward method in order to obtain a profitable trading of PV power in electricity markets. The scenario reduction method is applied to a two-period scenario tree, i.e., a scenario fan including uncertainty on day-ahead market prices, on imbalance prices and on PV power. Through a case study is analyzed the performance of the scenario reduction algorithm and the comparison with the original set of scenarios. The results show that the reduced set of scenarios still has a very high level of accuracy.