Design of a Flexible DC to AC Converter and its Control Using Reinforcement Learning
CovenanteeÉcole polytechnique fédérale de Lausanne
Document typeMaster thesis
Rights accessOpen Access
The use of DC to AC converters in renewable energy systems to transform direct to alternating current is necessary to couple these elements to the grid. In this project, a design for a DC to AC converter plant has been developed with future research purposes. In addition, direct heuristic dynamic programming, a reinforcement learning algorithm within the family of adaptive critic designs, has been analysed and simulated as a supplementary controller to improve the harmonic compensation of a shunt active filter.