Digital experimental environmentas a concept for improving greenhouse microclimate control efficiency

Abstract

The article substantiates the concept of a digital experimental environment as a tool for improving greenhouse microclimate control under current energy, economic, and environmental constraints. The study is motivated by the need to test advanced control strategies in a safe virtual space before their transfer to real greenhouse facilities, where errors may lead not only to excessive energy consumption but also to crop losses. The proposed environment integrates simulation models, SCADA/HMI visualization, data acquisition modules, and hierarchical decision-making logic into a unified research and engineering framework. 

The lower control level is intended for stabilization of key microclimate parameters, primarily temperature, humidity, and CO₂ concentration, whereas the upper level performs optimization of operating modes with regard to weather disturbances, crop requirements, technical limitations, and the cost of energy resources. Special attention is paid to the possibility of combining greenhouse climate models with crop-oriented indicators, which allows the transition from simple setpoint tracking to economically justified and biologically meaningful control. In this interpretation, the digital environment serves not only as a simulation tool but also as a prototype of a future digital twin for greenhouse production. 

The paper also develops the practical idea of using secondary technological steam as an alternative heat source in greenhouse heating systems. The integration of waste-heat recovery into the digital environment makes it possible to assess the stability of thermal control loops, compare operating scenarios, and estimate the potential reduction in primary energy demand and environmental burden. The obtained results show that such an environment increases reproducibility of experiments, supports data-driven optimization, and creates a foundation for further implementation of model predictive control, machine learning, and artificial intelligence methods in greenhouse management. In addition, the proposed approach improves the preparation of operator interfaces, facilitates parameter identification from experimental data, and reduces the risks associated with direct full-scale trials in industrial greenhouse conditions.