TY - GEN
T1 - Sim-to-Real Deep Reinforcement Learning for Maximum Power Point Tracking of Photovoltaic Systems
AU - Wang, Kangshi
AU - Ma, Jieming
AU - Man, Ka Lok
AU - Huang, Kaizhu
AU - Huang, Xiaowei
N1 - Funding Information:
This research is supported by the Natural Science Foundation of China (Grant No. 61702353,62002296), the Natural Science Foundation of Jiangsu Province (Grant No. BK20200250), the Suzhou Science and Technology Project-Key Industrial Technology Innovation (Grant No. SYG202006), Qing Lan Project of Jiangsu Province, the Key Program Special Fund of Xi’an Jiaotong-Liverpool University (XJTLU), Suzhou, China (Grant No. KSF-A-19, KSF-E-65, KSF-P-02, KSF-E-54), the XJTLU Research Development Fund (Grant No. RDF-17-02-04), and the XJTLU Postgraduate Research Scholarship (Grand No. PGRS1906004).
Publisher Copyright:
© 2021 IEEE
PY - 2021
Y1 - 2021
N2 - The maximum power points (MPPs) of photovoltaic (PV) panels vary with atmospheric conditions (solar irradiance, ambient temperature, shading conditions, etc.). Simulation platforms are widely used by engineers and scientists to create models, analyze data, and develop algorithms for the maximum power point tracking (MPPT) of PV systems. However, the behavior, which the algorithm develops in simulations, is usually specific to the characteristics of the models. Strategies that succeed in simulations may not be victoriously transferred to the real world because the modeling errors and sensor noise. In this paper, we create a dynamic model of the PV system only based on the data-sheet from manufacturers. A sim-to-real transfer strategy is proposed for the maximum power point tracking of PV systems. By dynamically randomizing the environments for the agent during the training, the proposed policy can adapt to very different atmospheric conditions. Simulations show that the proposed strategy can maintain a similar level of performance when deployed on the real PV panels.
AB - The maximum power points (MPPs) of photovoltaic (PV) panels vary with atmospheric conditions (solar irradiance, ambient temperature, shading conditions, etc.). Simulation platforms are widely used by engineers and scientists to create models, analyze data, and develop algorithms for the maximum power point tracking (MPPT) of PV systems. However, the behavior, which the algorithm develops in simulations, is usually specific to the characteristics of the models. Strategies that succeed in simulations may not be victoriously transferred to the real world because the modeling errors and sensor noise. In this paper, we create a dynamic model of the PV system only based on the data-sheet from manufacturers. A sim-to-real transfer strategy is proposed for the maximum power point tracking of PV systems. By dynamically randomizing the environments for the agent during the training, the proposed policy can adapt to very different atmospheric conditions. Simulations show that the proposed strategy can maintain a similar level of performance when deployed on the real PV panels.
KW - deep reinforcement learning
KW - dynamics randomization
KW - maximum power point tracking
KW - photovoltaic systems
UR - http://www.scopus.com/inward/record.url?scp=85126431635&partnerID=8YFLogxK
U2 - 10.1109/EEEIC/ICPSEurope51590.2021.9584821
DO - 10.1109/EEEIC/ICPSEurope51590.2021.9584821
M3 - Conference Proceeding
AN - SCOPUS:85126431635
T3 - 21st IEEE International Conference on Environment and Electrical Engineering and 2021 5th IEEE Industrial and Commercial Power System Europe, EEEIC / I and CPS Europe 2021 - Proceedings
BT - 21st IEEE International Conference on Environment and Electrical Engineering and 2021 5th IEEE Industrial and Commercial Power System Europe, EEEIC / I and CPS Europe 2021 - Proceedings
A2 - Leonowicz, Zbigniew M.
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 21st IEEE International Conference on Environment and Electrical Engineering and 2021 5th IEEE Industrial and Commercial Power System Europe, EEEIC / I and CPS Europe 2021
Y2 - 7 September 2021 through 10 September 2021
ER -