Optimizing energy market participation with batteries
Document typeMaster thesis
Rights accessRestricted access - author's decision
All rights reserved. This work is protected by the corresponding intellectual and industrial property rights. Without prejudice to any existing legal exemptions, reproduction, distribution, public communication or transformation of this work are prohibited without permission of the copyright holder
Due to the fact that the energy sector is in transition, there are goals for lowering the energy cost with the use of renewables and batteries. This presents challenges to the system and the solution is the issuing of energy communities that can be used to make electricity provision more clean and secure. It is also to see how energy flexibility elements or elements on the consumption side can make the system more efficient and cheaper, which is being done in this paper concerning the day-ahead bid and batteries. Traditional day-ahead bidding methods have become costly, mainly when the forecasted energy consumption differs from the actual consumption, which has to be resolved by penalizing with an imbalance cost. This thesis is part of a more significant project (Layered Energy System) that is to be deployed in Spain. Applying such changes to the electricity system first requires becoming familiar with and understanding Spain's context. The first part of this thesis provides research to understand the Spanish regulatory framework, how the market works, and the status of these technologies in Spain. Following that, this thesis's primary work is to explore how day-ahead market bid could be improved through the use of batteries for better planning and error assumptions. It mentions several day-ahead bidding strategies in the context of energy and batteries. And then selects a subset (three) of the studied strategies and implements them, comparing their performance on actual electricity data. Finally, selects the one that best fits various scenarios and requirements. A particular objective function is opted to be minimized with respect to the battery constraints that involve the variables. A linear program will find the values that best fits those variables at every time step $t$ of a single day. The methodology is an improvement over traditional predictive models. After comparing different strategies, Results show that strategy one, namely "Stochastic Chance-constraint optimization", yields the best results. In this strategy, the battery would have the freedom to maximize profit even if it sometimes increases imbalance. The preferred error distribution for this strategy is the Gamma distribution. Using a battery to offset imbalances can help to minimize total energy cost for a whole day (up to 26%). The last part of the thesis is ongoing research about capacity traders and market performance. It surveys the literature on trading strategies in various contexts and markets relevant to capacity traders. The market performance in capacity trading needs to consider how well the buildings can reach their desired capacity through bidding and selling. Performance metrics that are typically used to evaluate those trading strategies were documented. This feature is being worked on with python, but it will not be able to be shown.
DegreeMÀSTER UNIVERSITARI EN INNOVACIÓ I RECERCA EN INFORMÀTICA (Pla 2012)