Deep Reinforcement Learning based Smart Energy Management Strategy for an Integrated Energy System with Wind Energy

Abstract:
Integrated energy system (IES) has great potential to enhance the flexibility of power system to accommodate more RE such as wind and solar.
The key motivation of this project is to apply deep reinforcement learning (DRL) -based intelligent energy management (EM) strategy for an IES with wind power, which is coupled with electric power network, natural gas network and distributed heat network.
The significance of this project includes development of the proper combination of DRL methods and IESs, development of the DRL methods-based smart EM strategy for IESs, and helps system operators with decision support in the emerging IESs. In this project, the combination of the state-of-the-art DRL algorithms and IESs with different optimization targets will be explored. Firstly, an intelligent decarbonization economic EM Strategy for an integrated electricity-gas network based on an improved deep reinforcement learning method is developed. Then, a data-driven approach towards autonomous low-carbon economic energy management strategy in electricity-gas coupled energy systems based on transformer and deep reinforcement learning is investigated. Finally, a physical-model-free optimal energy management strategy for a grid-connected hybrid wind-microturbine-PV energy system via deep reinforcement learning approach is developed.
The outcome of this project is a new optimization control method. By applying this control strategy, complex uncertainties including renewable energy (RE), load demands and energy price signals are achieved. Besides, multiple objectives, such as RE curtailment, operating cost and carbon emission, can be effectively optimized by coordinating the output of each distributed generators and energy coupling units.

Funding: CSC Scholarship