Manuel Martínez Gómez ● Profesor Asistente

[vc_section el_class="container mx-auto align-items-center circle--pattern" css=".vc_custom_1648956589196{padding-top: 3rem !important;}"][vc_row el_class="pb-5"][vc_column][vc_wp_custommenu nav_menu="6"][uoh_breadcrumb_component automatic_breadcrumb="true"][uoh_title_component title_dropdown="big" title_decorator="true"]{{title}}[/uoh_title_component][vc_column_text css=""]This project will address the implementation of distributed controllers in intelligent agents within AC/DC electrical microgrids. Specifically, this project will address open issues in the distributed control literature for microgrids; these include optimal parameter tuning and resilience to communications disturbances such as transport delays, packet loss, and communication failures due to cyberattacks. All of these are important components that prevent the proliferation of microgrid projects throughout the country and the world. Microgrids have the potential to improve the energy management of renewable resources and the resilience of current and future electrical systems. Furthermore, they aid in decarbonization and benefit the energization of isolated communities and national industries. Based on the above, the main objective of this research is to formulate, implement, and validate distributed intelligent controllers, using reinforcement learning, in agents that comprise interconnected AC/DC microgrids, in order to achieve optimal operation concerning available energy resources despite disturbances and failures in communication channels. To achieve this objective, the following specific objectives are specified: (i) investigate the state of the art in the use of reinforcement learning algorithms in cooperative multi-agent system control and their application to microgrids; (ii) design a deep reinforcement learning algorithm to auxiliary control an ILC of a hybrid AC/DC microgrid with communication loss and variable time delays; (iii) design a distributed controller with parameter and structure adjustment capability through deep reinforcement learning algorithms for the agents of an AC/DC multi-microgrid with communication losses and variable time delay; (iv) design a robust distributed controller through deep reinforcement learning algorithms that allows agents in an AC/DC multi-microgrid to be resilient to heterogeneous and variable transport delays, loss of data packets, and DoS cyber-attacks; (v) build a prototyping platform for multi-agent-based intelligent agent control schemes with digital twin co-simulation; (vi) implement and validate the proposed reinforcement learning controllers in an AC/DC multi-microgrid experimental setup.[/vc_column_text][/vc_column][/vc_row][/vc_section][vc_section css=".vc_custom_1649209804184{background-color: #f6faff !important;}" el_class="p-md-0 pt-md-5"][vc_row el_class="container mx-auto align-items-center p-md-0 pt-5"][vc_column el_class="p-0"][/vc_column][/vc_row][/vc_section][vc_section css=".vc_custom_1649210787516{background-color: #f6faff !important;}" el_class="p-md-0 pt-md-5 pb-md-5"][vc_row el_class="container mx-auto align-items-center"][vc_column][/vc_column][/vc_row][/vc_section]