Artículos de publicaciones periódicas
Permanent URI for this collection
Browse
Browsing Artículos de publicaciones periódicas by Subject "ALGORITMOS"
Now showing 1 - 2 of 2
Results Per Page
Sort Options
artículo de publicación periódica.listelement.badge A non-centralized predictive control strategy for wind farm active power control: a wake-based partitioning approach(2020-05) Siniscalchi-Minna, Sara; Bianchi, Fernando D.; Ocampo-Martínez, Carlos; Domínguez-García, José Luis; De Schutter, Bart"Owing to wake effects, the power production of each turbine in a wind farm is highly coupled to the operating conditions of the other turbines. Wind farm control strategies must take into account these couplings and produce individual power commands for each turbine. In this case, centralized control approaches might be prone to failures due to the high computational burden and communication dependency. To overcome this problem, this paper proposes a non-centralized scheme based on splitting the wind farm into almost uncoupled sets of turbines by solving a mixed-integer partitioning problem. In each set of turbines, a model predictive control strategy seeks to optimize the distribution of the power set-points among turbines such that the impact of the power losses due to the wake effect is reduced. Then, a supervisory controller coordinates the generation of each group to satisfy the power demanded by the grid operator. The effectiveness of the proposed control scheme in terms of reduction of computational costs and power regulation is confirmed by simulations for a wind farm of 42 turbines."artículo de publicación periódica.listelement.badge Training a gaming agent on brainwaves(2020-12-07) Bartolomé, Francisco; Moreno, Juan; Navas, Natalia; Vitali, José; Ramele, Rodrigo; Santos, Juan Miguel"Error-related potential (ErrP) are a particular type of Event-Related Potential (ERP) elicited by a person attending a recognizable error. These Electroencephalographic (EEG) signals can be used to train a gaming agent by a Reinforcement Learning (RL) algorithm to learn an optimal policy. The experimental process consists of an observational human critic (OHC) observing a simple game scenario while their brain signals are captured. The game consists of a grid, where a blue spot has to reach a desired target in the fewest amount of steps. Results show that there is an effective transfer of information and that the agent successfully learns to solve the game efficiently, from the initial 97 steps on average required to reach the target to the optimal number of 8 steps. Our results are expressed in threefold: (i) the mechanics of a simple grid-based game that can elicit the ErrP signal component, (ii) the verification that the reward function only penalizes wrong steps, which means that type II error (not properly identifying a wrong movement) does not affect significantly the agent learning process; (iii) collaborative rewards from multiple observational human critics can be used to train the algorithm effectively and can compensate low classification accuracies and a limited scope of transfer learning schemes."