Using a Digital Twin for Greenhouse Tunnel Temperature Management and Prediction

Unpredictable weather patterns caused by climate change is impacting agricultural productivity worldwide. This threatens sustainability and may lead to food insecurity, especially in poorer regions. Affluent countries can afford costly investments toward mitigating the affects of climate change on food production. However, poorer countries tend to lag due to a lack of resources. To improve climate resilience evolving technologies, such as the Internet of Things (IoT), have been proposed and developed for climate-smart farming. In this paper we present a technological solution towards smarter farming using a digital twin to mitigate climate change and eliminate inefficiencies in agricultural production. A greenhouse tunnel in Stellenbosch, South Africa was instrumented to monitor temperatures inside the tunnel and control the cooling fan and wet wall. These measurements were used with a Support Vector Regression algorithm to develop and validate a data-driven thermal model that accurately predicts the temperatures inside the tunnel. The study was successful in producing a data-driven model that can simulate internal temperatures to an RMSE value of 1.76 °C and an
$R^{2}$
value of 0.9 for a 1-hour ahead simulation. With this early success, the model can be developed into a 24-hour ahead horizon simulator for improved decision-making by researchers.