Energy Management for Plant Factory with Deep Learning and Predictive Control

Abstract

This study introduces a novel cyber-physical-biological system for energy management and crop production in plant factories. The goal is to minimize energy consumption while maintaining operational efficiency in sustainable food production. The CPBS integrates various control variables, including temperature, humidity, lighting, and CO₂ levels, and accurately captures plant biological dynamics while predicting crop growth within controlled environments. To achieve this, physics-informed deep learning techniques are utilized to develop computationally efficient digital twins of the plant factory's internal microclimate and crop states. By incorporating Artificial Intelligence (AI), the control objectives are fine-tuned to minimize energy usage and resource expenses, ensuring sustainable crop production rates under different daytime scenarios. The results demonstrate an impressive 8.75 % reduction in energy usage compared to alternative control methods, enhancing operational efficiency and promoting sustainability in plant factories. The proposed approach offers a promising solution for achieving sustainable food production with minimized energy consumption in controlled environments.