Future Energy Systems

Contact: Ashkan Yousefi, [email protected]

Description: High volumes of data are becoming available with the growth of the advanced metering infrastructure and massive deployment of IoT devices. These are expected to benefit the planning and operation of the future energy systems and to help the customers transition from a passive to an active role. In this project, we explore a novel approach using deep reinforcement learning, a hybrid approach that combines reinforcement learning with deep learning, to perform on-line optimization of schedules for building energy management systems. The learning procedure gets benefited from two methods, Deep Q-learning and Deep Policy Gradient. The hybrid approach is capable of handling multiple actions simultaneously. The large-scale Pecan Street Inc. database will be used to validate the proposed approach. The Pecan database is highly-dimensional database includes information about photovoltaic power generation, electric vehicles as well as buildings appliances. In addition, these on-line energy scheduling strategies will be used to provide real-time feedback for electricity customers and prosumers to achieve more efficient use of electricity.