Carbon Dioxide Estimation Using Artificial Neural Networks in Agricultural Greenhouses Case of GHARDAIA

Abstract

To optimize photosynthesis and crop growth in greenhouses, it is essential to predict the concentration and consumption of carbon dioxide. In this study, we aimed to anticipate the CO2 concentration in two distinct greenhouses: one equipped with a cooling system and the other without. Over a month, we meticulously measured temperature, relative humidity, and CO2 concentration in these greenhouses at the Experimental Research Unit for Renewable Energies in Ghardaïa, Algeria. To achieve this, we used an Arduino board coupled with various sensors. The collected data were then used to train an artificial neural network, employing the Long Short-Term Memory (LSTM) algorithm for prediction. The analysis of the obtained results demonstrates the model’s reliability, with R2

and MSE parameter values ranging between 0.95% and.1%. Special attention will be given to the potential use of these models for improving agricultural production, economic evaluation, and environmental impact.