Automated deep reinforcement learning for real-time scheduling strategy of multi-energy system integrated with post-carbon and direct-air carbon captured system

Alabi，Tobi Michael1,2,3; Lawrence，Nathan P.3,4; Lu，Lin2; Yang，Zaiyue1; Bhushan Gopaluni，R.3

2023-03-01

10.1016/j.apenergy.2022.120633

APPLIED ENERGY   影响因子和分区

0306-2619

1872-9118

The carbon-capturing process with the aid of CO removal technology (CDRT) has been recognised as an alternative and a prominent approach to deep decarbonisation. However, the main hindrance is the enormous energy demand and the economic implication of CDRT if not effectively managed. Hence, a novel deep reinforcement learning agent (DRL), integrated with an automated hyperparameter selection feature, is proposed in this study for the real-time scheduling of a multi-energy system (MES) coupled with CDRT. Post-carbon capture systems (PCCS) and direct-air capture systems (DACS) are considered CDRT. Various possible configurations are evaluated using real-time multi-energy data of a district in Arizona, the United States, and CDRT parameters from manufacturers' catalogues and pilot project documentation. The simulation results validate that an optimised soft-actor critic (SAC) DRL algorithm outperformed the Twin-delayed deep deterministic policy gradient (TD3) algorithm due to its maximum entropy feature. We then trained four (4) SAC DRL agents, equivalent to the number of considered case studies, using optimised hyperparameter values and deployed them in real time for evaluation. The results show that the proposed DRL agent can meet the prosumers' multi-energy demand and schedule the CDRT energy demand economically without specified constraints violation. Also, the proposed DRL agent outperformed rule-based scheduling by 23.65%. However, the configuration with PCCS and solid-sorbent DACS is considered the most suitable configuration with a high CO captured-released ratio (CCRR) of 38.54, low CO released indicator (CRI) value of 2.53, and a 36.5% reduction in CDR cost due to waste heat utilisation and high absorption capacity of the selected sorbent. However, the adoption of CDRT is not economically viable at the current carbon price. Finally, we showed that CDRT would be attractive at a carbon price of 400-450USD/ton with the provision of tax incentives by the policymakers.

Carbon capture Carbon removal Deep reinforcement learning Integrated energy system Zero-emission

SCI ; EI

英语

第一

Natural Science Foundationof China[61873118] ; Shenzhen Committee on Science and Innovations[GJHZ20180411143603361] ; Department of Science and Technology of Guangdong Province[2018A050506003]

Energy & Fuels ; Engineering

Energy & Fuels ; Engineering, Chemical

WOS:000923244900001

ELSEVIER SCI LTD

20230213379280

Carbon capture ; Deep learning ; Energy management ; Learning systems ; Reinforcement learning ; Taxation ; Waste heat

Environmental Engineering:454 ; Ergonomics and Human Factors Engineering:461.4 ; Energy Management and Conversion:525 ; Energy Losses (industrial and residential):525.4 ; Artificial Intelligence:723.4 ; Inorganic Compounds:804.2

ENGINEERING

2-s2.0-85146067228

Scopus

1.Department of Mechanical and Energy Engineering,Southern University of Science and Technology,Shenzhen,China
2.Renewable Energy Research Group (RERG),Department of Building Environment and Energy Engineering,The Hong Kong Polytechnic University,Hong Kong
3.Data Analytics and Intelligent System (DAIS) Laboratory,Department of Chemical and Biological Engineering,University of British Columbia,Vancouver,Canada
4.Department of Mathematics,University of British Columbia,Vancouver,Canada

Alabi，Tobi Michael,Lawrence，Nathan P.,Lu，Lin,et al. Automated deep reinforcement learning for real-time scheduling strategy of multi-energy system integrated with post-carbon and direct-air carbon captured system[J]. APPLIED ENERGY,2023,333.

Alabi，Tobi Michael,Lawrence，Nathan P.,Lu，Lin,Yang，Zaiyue,&Bhushan Gopaluni，R..(2023).Automated deep reinforcement learning for real-time scheduling strategy of multi-energy system integrated with post-carbon and direct-air carbon captured system.APPLIED ENERGY,333.

Alabi，Tobi Michael,et al."Automated deep reinforcement learning for real-time scheduling strategy of multi-energy system integrated with post-carbon and direct-air carbon captured system".APPLIED ENERGY 333(2023).