(484d) Optimization-Based Estimation and Control of Renewable Energy-Powered Greenhouse Systems

With the globally rising demand for sustainable food sources, the need for optimizing crop growth with minimum water and energy use has significantly increased. The remarkable potential of greenhouse systems to conserve resources (due to their enclosed environment that can be sufficiently controlled [1]) makes them a reasonable solution for year-round crop development [2, 3]. Automatic control systems have improved crop yield in greenhouses by 10-20% while reducing energy consumption by 25-35% [4]. The current paradigm for achieving greenhouse overall economy and sustainability objectives requires combining information management and decision-making in one optimization-based setting rather than a hierarchical separation of information and commands.

The optimal control and decision-making strategies are often implemented in practice by model predictive control (MPC) due to its relative conceptual simplicity, flexibility, performance, robustness, and ability to efficiently handle complex multivariable systems with terminal and path constraints [5, 6]. However, the perfect state information and parameters values cannot always be assumed when employing MPC. Therefore, an estimation method must be utilized to deal with the uncertainties that arise from imperfect measuring and unknown system parameters. Among the several state and parameter estimation methods, moving horizon estimation (MHE) has proven to be an efficient approach with a similar dynamic optimization architecture as MPC [7-9]. Applications of MPC to greenhouse climate control have been growing in recent years, allowing for improved quality and efficiency [10]. A mixture of control strategies has also been applied to greenhouse systems, mainly MPC, fuzzy logic control, and adaptive control [11, 12].

This work is focused on developing an optimization-based nonlinear output feedback control framework for semi-closed renewable energy-powered greenhouse systems. The proposed control architecture integrates MHE and MPC to regulate greenhouse environment temperature and humidity at desired values and optimize crop growth subject to minimum energy use and operating costs. MHE is employed to estimate the unmeasurable states and unknown parameters required by MPC using the feedback from the available measurement sensors. A predictive multiscale model is developed to describe the integrated dynamics of the renewable energy generation system, the greenhouse climate, and crop growth. The model consists of two components; (1) renewable energy generation (wind and solar energy) component describing the power generation dynamics and (2) greenhouse environment component predicting the temperature of the greenhouse cover, plant canopy, soil, and inside air, the concentration of CO2, the air humidity inside the greenhouse, the canopy cover, soil moisture, soil nitrogen concentration, and biomass growth. The resulting model is then employed as the basis for the nonlinear integrated MHE and MPC design. The closed-loop performance of the proposed control architecture is compared to a traditional MPC with a Kalman filter in the presence of varying environmental conditions and disturbances in the manipulated inputs and measured outputs.

References:

[1] T. Kozai, “Resource use efficiency of closed plant production system with artificial light: Concept, estimation and application to plant factory,” Proceedings of the Japan Academy Series B: Physical and Biological Sciences, 89(10), 447–461, 2013.

[2] D.D. Avgoustaki and G. Xydis, “Plant factories in the water-food-energy Nexus era: a systematic bibliographical review,” Food Security, 12(2) 253–268, 2020.

[3] Y. Kikuchi, Y. Kanematsu, N. Yoshikawa, T. Okubo, and M. Takagaki, “Environmental and resource use analysis of plant factories with energy technology options: A case study in Japan,” J. Clean. Prod., 186, 703-717, 2018.

[4] A. De Gelder, J.A. Dieleman, G.P.A. Bot, and L.F.M. Marcelis, “An overview of climate and crop yield in closed greenhouses,” J. Hortic. Sci. Biotechnol., 87(3), 193-202, 2012.

[5] J.B. Rawlings, D.Q. Mayne, and M.M. Diehl, Model Predictive Control: Theory, Computation, and Design. 2nd Edition, Nob Hill Publishing, 2019.

[6] M. Morari and J.H. Lee, “Model predictive control: Past, present and future,” Comput. Chem. Eng., 23(4-5), 667–682, 1999.

[7] J. Rawlings and B. Bakshi, “Particle filtering and moving horizon estimation,” Comput. Chem. Eng., 30(10-12), 1529–1541, 2006

[8] D. Copp and J. Hespanha, “Simultaneous nonlinear model predictive control and state estimation,” Automatica, 77:143–154, 2017.

[9] C.V. Rao, J.B. Rawlings, D.Q. Mayne, “Constrained state estimation for nonlinear discrete-time systems: Stability and moving horizon approximations, “ IEEE Trans. Autom. Control, 48(2), 246-258, 2003.

[10] T. Weidner, A. Yang, and M.W. Hamm, “Energy optimisation of plant factories and greenhouses for different climatic conditions,” Energy Convers. Manag., 243, 114336, 2021.

[11] E. Iddio, L. Wang, Y. Thomas, G. McMorrow, and A. Denzer, “Energy efficient operation and modeling for greenhouses: A literature review,” Renewable and Sustainable Energy Reviews, 117, 109480, 2020

[12] W.H. Chen and F. You, “Data-driven robust optimization for greenhouse temperature control using model predictive control,” Chem. Eng. Trans., 81, 721-726, 2020.