Jump to Main Content
Action dependent heuristic dynamic programming based residential energy scheduling with home energy inter-exchange
- Xu, Yancai, Liu, Derong, Wei, Qinglai
- Energy conversion and management 2015 v.103 pp. 553-561
- algorithms, batteries, dynamic programming, electricity, energy conservation, energy efficiency, environmental protection, issues and policy, neural networks, prices, renewable energy sources
- Residential energy scheduling is a hot topic nowadays in the background of energy saving and environmental protection worldwide. To achieve this objective, a new residential energy scheduling algorithm is developed for energy management, based on action dependent heuristic dynamic programming. The algorithm works under the circumstance of residential real-time pricing and two adjacent housing units with energy inter-exchange, which can reduce the overall cost and enhance renewable energy efficiency after long-term operation. It is designed to obtain the optimal control policy to manage the directions and amounts of electricity energy flux. The algorithm’s architecture is mainly constructed based on neural networks, denoting the learned characteristics in the linkage of layers. To get close to real situations, many constraints such as maximum charging/discharging power of batteries are taken into account. The absent energy penalty cost is developed for the first time as a part of the performance index function. When the environment changes, the residential energy scheduling algorithm gains new features and keeps adapting in real-time operations. Simulation results show that the developed algorithm is beneficial to energy conversation.