Analysis Nature Food (2024) Cite this article Metrics Abstract Plant factories with artificial lighting (PFALs) can boost food production per unit area but require resources such as carbon dioxide and energy to maintain optimal plant growth conditions. Here we use computational modelling and artificial intelligence (AI) to examine plant–environment […]
Click here to view original web page at www.nature.com
- Analysis
Nature Food (2024)Cite this article
Abstract
Plant factories with artificial lighting (PFALs) can boost food production per unit area but require resources such as carbon dioxide and energy to maintain optimal plant growth conditions. Here we use computational modelling and artificial intelligence (AI) to examine plant–environment interactions across ten diverse global locations with distinct climates. AI reduces energy use by optimizing lighting and climate regulation systems, with energy use in PFALs ranging from 6.42 kWh kg−1 in cooler climates to 7.26 kWh kg−1 in warmer climates, compared to 9.5–10.5 kWh kg−1 in PFALs using existing, non-AI-based technology. Outdoor temperatures between 0 °C and 25 °C favour ventilation-related energy use reduction, with outdoor humidity showing no clear pattern or effect on energy use. Ventilation-related energy savings negatively impact other resource utilization such as carbon dioxide use. AI can substantially enhance energy savings in PFALs and support sustainable food production.
This is a preview of subscription content, access via your institution
Access options
Access through your institution
Access Nature and 54 other Nature Portfolio journals
Get Nature+, our best-value online-access subscription
$29.99 / 30 days
cancel any time
Subscribe to this journal
Receive 12 digital issues and online access to articles
$119.00 per year
only $9.92 per issue
Buy this article
- Purchase on SpringerLink
- Instant access to full article PDF
Prices may be subject to local taxes which are calculated during checkout
Additional access options:
Data availability
The mean monthly outdoor climate data used in this study are publicly available on timeanddate (https://www.timeanddate.com/weather). The electricity price data for the locations in the United States were obtained from the US Energy Information Administration (https://www.eia.gov/electricity/data/state/); the electricity price data for Reykjavik and Dubai were obtained from Climatescope by BloombergNEF (https://www.global-climatescope.org/markets/is/). All other data are provided in the paper and Supplementary Information. Source data are provided with this paper.
Code availability
All the Python scripts and data necessary to run the analyses presented in this study are publicly available at https://github.com/PEESEgroup/PFAL-DRL.
References
Ruel, M. T., Garrett, J., Yosef, S. & Olivier, M. in Nutrition and Health in a Developing World (eds de Pee, S., Taren, D. & Bloem, M.) 705–735 (Humana Press, 2017).
Béné, C. Resilience of local food systems and links to food security—a review of some important concepts in the context of COVID-19 and other shocks. Food Secur. 12, 805–822 (2020).
Godfray, H. C. J. et al. Food security: the challenge of feeding 9 billion people. Science 327, 812–818 (2010).
Chung, M. G. & Liu, J. International food trade benefits biodiversity and food security in low-income countries. Nat. Food 3, 349–355 (2022).
Beacham, A. M., Vickers, L. H. & Monaghan, J. M. Vertical farming: a summary of approaches to growing skywards. J. Hortic. Sci. Biotechnol. 94, 277–283 (2019).
Orsini, F., Pennisi, G., Zulfiqar, F. & Gianquinto, G. Sustainable use of resources in plant factories with artificial lighting (PFALs). Eur. J. Hortic. Sci. 85, 297–309 (2020).
Kalantari, F., Tahir, O. M., Joni, R. A. & Fatemi, E. Opportunities and challenges in sustainability of vertical farming: a review. J. Landsc. Ecol. 11, 35–60 (2018).
Gómez, C. et al. Controlled environment food production for urban agriculture. HortScience 54, 1448–1458 (2019).
Lages Barbosa, G. et al. Comparison of land, water, and energy requirements of lettuce grown using hydroponic vs. conventional agricultural methods. Int. J. Environ. Res. Public Health 12, 6879–6891 (2015).
SharathKumar, M., Heuvelink, E. & Marcelis, L. F. Vertical farming: moving from genetic to environmental modification. Trends Plant Sci. 25, 724–727 (2020).
Colhoun, J. Effects of environmental factors on plant disease. Annu. Rev. Phytopathol. 11, 343–364 (1973).
Graamans, L., Baeza, E., Van Den Dobbelsteen, A., Tsafaras, I. & Stanghellini, C. Plant factories versus greenhouses: comparison of resource use efficiency. Agric. Syst. 160, 31–43 (2018).
Weidner, T., Yang, A. & Hamm, M. W. Energy optimisation of plant factories and greenhouses for different climatic conditions. Energy Convers. Manage. 243, 114336 (2021).
Avgoustaki, D. D. & Xydis, G. Indoor vertical farming in the urban nexus context: business growth and resource savings. Sustainability 12, 1965 (2020).
Graamans, L., Tenpierik, M., van den Dobbelsteen, A. & Stanghellini, C. Plant factories: reducing energy demand at high internal heat loads through facade design. Appl. Energy 262, 114544 (2020).
Zhang, Y., Kacira, M. & An, L. A CFD study on improving air flow uniformity in indoor plant factory system. Biosyst. Eng. 147, 193–205 (2016).
Summers, H. M., Sproul, E. & Quinn, J. C. The greenhouse gas emissions of indoor cannabis production in the United States. Nat. Sustain. 4, 644–650 (2021).
Kitaya, Y., Tsuruyama, J., Shibuya, T., Yoshida, M. & Kiyota, M. Effects of air current speed on gas exchange in plant leaves and plant canopies. Adv. Space Res. 31, 177–182 (2003).
Ashrae, A. Ventilation and Acceptable Indoor Air Quality in Low-rise Residential Buildings (American Society of Heating, Refrigeration and Air-Conditioning Engineers, 2016).
Chinn, R. Y. & Sehulster, L. Guidelines for environmental infection control in health-care facilities: recommendations of CDC and Healthcare Infection Control Practices Advisory Committee (HICPAC). MMWR Recomm. Rep. 52, 1–42 (2003).
Eaton, M., Shelford, T., Cole, M. & Mattson, N. Modeling resource consumption and carbon emissions associated with lettuce production in plant factories. J. Clean. Prod. 384, 135569 (2023).
Zhang, Y. & Kacira, M. Comparison of energy use efficiency of greenhouse and indoor plant factory system. Eur. J. Hortic. Sci. 85, 310–320 (2020).
Harbick, K. & Albright, L. in VIII International Symposium on Light in Horticulture (eds Currey, C.J. et al.) 1134, 285–292 (ISHS, 2016).
Kozai, T. in Smart Plant Factory: The Next Generation Indoor Vertical Farms (ed. Kozai, T.) 3–13 (Springer, 2018).
Berenguel, M., Yebra, L. J. & Rodríguez, F. in 2003 European control conference (ECC), 2747–2752 (IEEE, 2003).
Van Straten, G. Acceptance of optimal operation and control methods for greenhouse cultivation. Annu. Rev. Control 23, 83–90 (1999).
Kamp, P. G. H. & Timmerman, G. J. Computerised Environmental Control in Greenhouses: A Step by Step Approach. 2nd edn (PTC+, 2003).
Xu, D., Ahmed, H. A., Tong, Y., Yang, Q. & van Willigenburg, L. G. Optimal control as a tool to investigate the profitability of a Chinese plant factory-lettuce production system. Biosyst. Eng. 208, 319–332 (2021).
Morimoto, T., Torii, T. & Hashimoto, Y. Optimal control of physiological processes of plants in a green plant factory. Control Eng. Pract. 3, 505–511 (1995).
Wang, X. J., Kang, M. Z., Lewlomphaisarl, U., Hua, J. & Wang, H. Y. in 2022 Australian & New Zealand Control Conference (ANZCC) 166–170 (IEEE, 2022).
Van Delden, S. et al. Current status and future challenges in implementing and upscaling vertical farming systems. Nat. Food 2, 944–956 (2021).
Sutton, R. S. & Barto, A. G. Reinforcement Learning: An Introduction (MIT Press, 2018).
Baechler, M. C., Gilbride, T. L., Cole, P. C., Hefty, M. G. & Ruiz, K. Guide to Determining Climate Regions by County, Building America Best Practices Series vol. 7.3 (Pacific Northwest National Laboratory, 2015).
Endo, A., Tsurita, I., Burnett, K. & Orencio, P. M. A review of the current state of research on the water, energy, and food nexus. J. Hydrol. Reg. Stud. 11, 20–30 (2017).
Kozai, T. Towards sustainable plant factories with artificial lighting (PFALs) for achieving SDGs. Int. J. Agric. Biol. Eng. 12, 28–37 (2019).
Kozai, T. in Smart Plant Factory: The Next Generation Indoor Vertical Farms (ed. Kozai, T.), 15–29 (Springer, 2018).
Rachmilevitch, S., Cousins, A. B. & Bloom, A. J. Nitrate assimilation in plant shoots depends on photorespiration. Proc. Natl Acad. Sci. USA 101, 11506–11510 (2004).
Jackson, W. & Volk, R. Photorespiration. Annu. Rev. Plant Physiol. 21, 385–432 (1970).
Ogren, W. L. Photorespiration: pathways, regulation, and modification. Annu. Rev. Plant Physiol. 35, 415–442 (1984).
Weidner, T., Yang, A., Forster, F. & Hamm, M. W. Regional conditions shape the food–energy–land nexus of low-carbon indoor farming. Nat. Food 3, 206–216 (2022).
Gao, L. & Bryan, B. A. Finding pathways to national-scale land-sector sustainability. Nature 544, 217–222 (2017).
Xie, Z. et al. Conservation opportunities on uncontested lands. Nat. Sustain. 3, 9–15 (2020).
Graham-Rowe, D. Agriculture: beyond food versus fuel. Nature 474, S6–S8 (2011).
Calvert, K. & Mabee, W. More solar farms or more bioenergy crops? Mapping and assessing potential land-use conflicts among renewable energy technologies in eastern Ontario, Canada. Appl. Geogr. 56, 209–221 (2015).
Nie, Y. et al. A food-energy-water nexus approach for land use optimization. Sci. Total Environ. 659, 7–19 (2019).
Leung Pah Hang, M. Y., Martinez-Hernandez, E., Leach, M. & Yang, A. Insight-based approach for the design of integrated local food–energy–water systems. Environ. Sci. Technol. 51, 8643–8653 (2017).
Namany, S., Al-Ansari, T. & Govindan, R. Optimisation of the energy, water, and food nexus for food security scenarios. Comput. Chem. Eng. 129, 106513 (2019).
Martin, M., Elnour, M. & Siñol, A. C. Environmental life cycle assessment of a large-scale commercial vertical farm. Sustain. Prod. Consum. 40, 182–193 (2023).
Kozai, T., Uraisami, K., Kai, K. & Hayashi, E. in Plant Factory Basics, Applications and Advances (eds Kozai, T., Niu, G. & Masabni, J.) 197–216 (Academic Press, 2022).
Horomia, K. & Gordon-Smith, H. Global CEA Census Report (WayBeyond and Agritecture, 2021).
Van Henten, E. J. Sensitivity analysis of an optimal control problem in greenhouse climate management. Biosyst. Eng. 85, 355–364 (2003).
Chen, W.-H. & You, F. Semiclosed greenhouse climate control under uncertainty via machine learning and data-driven robust model predictive control. IEEE Trans. Control Syst. Technol. 30, 1186–1197 (2021).
Ioslovich, I. Optimal control strategy for greenhouse lettuce: incorporating supplemental lighting. Biosyst Eng 103, 57–67 (2009).
Zelkind, M., Livingston, T. & Verlage, V. in Plant Factory Basics, Applications and Advances (eds Kozai, T., Niu, G. & Masabni, J.) 295–305 (Academic Press, 2022).
Kozai, T. Resource use efficiency of closed plant production system with artificial light: concept, estimation and application to plant factory. Proc. Jpn Acad. B 89, 447–461 (2013).
Voogd, K. L., Allamaa, J. P., Alonso-Mora, J. & Son, T. D. Reinforcement learning from simulation to real world autonomous driving using digital twin. IFAC Pap OnLine 56, 1510–1515 (2023).
Balaji, B. et al. in 2020 IEEE International Conference on Robotics and Automation (ICRA) 2746–2754 (IEEE, 2020).
Van Henten, E. Validation of a dynamic lettuce growth model for greenhouse climate control. Agric. Syst. 45, 55–72 (1994).
Ravishankar, E. et al. Organic solar powered greenhouse performance optimization and global economic opportunity. Energy Environ. Sci. 15, 1659–1671 (2022).
Zhang, Y. & Kacira, M. in Plant Factory Basics, Applications and Advances (eds Kozai, T. et al) 391–400 (Elsevier, 2022).
Ahmed, H. A., Yu-Xin, T. & Qi-Chang, Y. Optimal control of environmental conditions affecting lettuce plant growth in a controlled environment with artificial lighting: a review. S. Afr. J. Bot. 130, 75–89 (2020).
Zhang, X., He, D., Niu, G., Yan, Z. & Song, J. Effects of environment lighting on the growth, photosynthesis, and quality of hydroponic lettuce in a plant factory. Int. J. Agric. Biol. Eng. 11, 33–40 (2018).
Waycott, W. Photoperiodic response of genetically diverse lettuce accessions. J. Am. Soc. Hortic. Sci. 120, 460–467 (1995).
Van Henten, E. Greenhouse Climate Management: An Optimal Control Approach (Wageningen University and Research, 1994).
Haarnoja, T. et al. Soft actor-critic algorithms and applications. Preprint at arXiv https://doi.org/10.48550/arXiv.1812.05905 (2018).
Meggers, F., Ritter, V., Goffin, P., Baetschmann, M. & Leibundgut, H. Low exergy building systems implementation. Energy 41, 48–55 (2012).
Weng, J. et al. Tianshou: a highly modularized deep reinforcement learning library. J. Mach. Learn. Res. 23, 1–6 (2022).
Acknowledgements
B.D.-N. acknowledges partial support from Schmidt Sciences via an Eric and Wendy Schmidt AI in Science Postdoctoral Fellowship to Cornell University, and the Natural Sciences and Engineering Research Council of Canada (NSERC). The authors gratefully acknowledge the Cornell Institute for Digital Agriculture (CIDA) for partial funding support. This work was partially supported by the Specialty Crop Research Initiative (award no. 2022-51181-38324) from the US Department of Agriculture National Institute of Food and Agriculture.
Author information
Authors and Affiliations
Systems Engineering, Cornell University, Ithaca, NY, USA
Benjamin Decardi-Nelson & Fengqi You
Robert Frederick Smith School of Chemical and Biomolecular Engineering, Cornell University, Ithaca, NY, USA
Fengqi You
Cornell Institute for Digital Agriculture, Cornell University, Ithaca, NY, USA
Fengqi You
Contributions
B.D.-N. and F.Y. conceived the study. B.D.-N. developed the models. B.D.-N. and F.Y. conducted the study, discussed the results, and wrote and reviewed the paper and Supplementary Information files. F.Y. secured the funding and supervised the research.
Corresponding author
Correspondence to Fengqi You.
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
Nature Food thanks Michael Martin, Giulia Martini and Aidong Yang for their contribution to the peer review of this work.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Extended data
Extended Data Fig. 1 Outdoor environmental conditions for the various locations considered in this study.
The outdoor CO2 concentration was kept constant at 400 ppm. The first seven plots are arranged according to seven climate zones in the USA. Generally, as we move across the zones, the mean monthly temperature decreases and the number of cold months increases, starting from Miami (Zone 1 A) and ending with Fargo (Zone 7 A). The letters A, B and C denote moist, dry, and marine conditions respectively. a, Miami, Florida (Zone 1 A). b, Phoenix, Arizona (Zone 2B). c, Los Angeles, California (Zone 3B). d, Seattle, Washington (Zone 4 C). e, Chicago, Illinois (Zone 5 A). f, Milwaukee, Wisconsin (Zone 6 A). g, Fargo, North Dakota (Zone 7 A). h, Ithaca, New York (Zone 6 A). i, Reykjavik, Iceland. j, Dubai, United Arab Emirates (UAE).
Extended Data Fig. 2 Relative annual specific energy use of baseline and AI systems for various control activities.
The values were computed by subtracting the annual specific energy use of the PFAL with the AI system from that of the PFAL with the baseline for each control activity. A positive value indicates energy saving using the AI while negative values indicate that the AI strategy uses more energy than the baseline strategy. a, Relative annual specific energy use for the lighting system. b, Relative annual specific energy use for the cooling/heating system. c, Relative annual specific energy use for the dehumidification system. d, Relative annual specific energy use for the ventilation system. e, Relative annual specific energy use for the CO2 dosing system.
Extended Data Fig. 3 Trajectories of the PFAL with the AI system for Ithaca, New York in April.
The AI system effectively regulates the environmental factors within the PFAL without negatively affecting crop growth. a, Trajectory of the lettuce dry weight b, Trajectory of the CO2 concentration with elevated CO2 concentration during the light period. c, Trajectory of the air temperature in the PFAL within the operating bounds (short dashes). d, Trajectory of the relative humidity of the air in the PFAL. e, Trajectory of the light power supplied to the artificial lighting system showing a gradual increase in light intensity as the crop matures. f, Trajectory of the supplemental CO2 supply rate. g, Trajectory of the dehumidification rate. h, Trajectory of the cooling/heating rate. i, Trajectory of the ventilation rate characterized by low ventilation during the light period and high ventilation during dark period.
Extended Data Fig. 4 Trajectories for one day of the PFAL with the AI system for Ithaca, New York in April.
The AI system effectively regulates the environmental factors within the PFAL without negatively affecting crop growth. a, Trajectory of the lettuce dry weight b, Trajectory of the CO2 concentration with elevated CO2 concentration during the light period. c, Trajectory of the air temperature in the PFAL within the operating bounds (short dashes). d, Trajectory of the relative humidity of the air in the PFAL. e, Trajectory of the light power supplied to the artificial lighting system showing a gradual increase in light intensity as the crop matures. f, Trajectory of the supplemental CO2 supply rate. g, Trajectory of the dehumidification rate. h, Trajectory of the cooling/heating rate. i, Trajectory of the ventilation rate characterized by low ventilation during the light period and high ventilation during dark period.
Extended Data Fig. 5 Relationship between mean monthly outdoor climate and monthly energy or CO2 use of the PFAL.
The plots were generated by combining the monthly data across all the locations under the AI strategy. a, Mean monthly outdoor temperature-outdoor relative humidity-energy use relationship. b, Mean monthly outdoor temperature-outdoor relative humidity-CO2 use relationship. The size of each point represents the magnitude of each data point, and the darker colour means a higher point cloud density.
Extended Data Fig. 6
DRL policy training procedure.
Supplementary information
Supplementary Information
Supplementary Figs. 1–8, Tables 1–3, Methods 1 and 2, and Notes 1–5.
Reporting Summary
Source data
Source Data Fig. 2
Numerical source data for Fig. 2b.
Source Data Fig. 3
Numerical source data for Fig. 3.
Source Data Fig. 4
Numerical source data for Fig. 4.
Source Data Fig. 5 and Extended Data Fig. 5
Numerical source data for Fig. 5 and Extended Data Fig. 5.
Source Data Fig. 6
Numerical source data for Fig. 6.
Source Data Extended Data Fig. 1
Numerical source data for Extended Data Fig. 1.
Source Data Extended Data Fig. 2
Numerical source data for Extended Data Fig. 2.
Source Data Extended Data Figs. 3 and 4
Numerical source data for Extended Data Fig. 3 and Extended Data Fig. 4.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Decardi-Nelson, B., You, F. Artificial intelligence can regulate light and climate systems to reduce energy use in plant factories and support sustainable food production. Nat Food (2024). https://doi.org/10.1038/s43016-024-01045-3
Received15 June 2023
Accepted15 August 2024
Published09 September 2024
DOIhttps://doi.org/10.1038/s43016-024-01045-3
