Agrivoltaic system for energy-food production: A symbiotic approach on strategy, modelling, and optimization

Abstract

Agrivoltaic system is a symbiotic approach to minimize land use conflicts for energy-food production. This production is dynamically affected by environmental parameters. Hence, an efficient system design is needed by adopting a suitable optimization model. The objective of this study is to discuss the application of artificial neural network and genetic algorithm to an experimental 0.675 kWp agrivoltaic system with turmeric crops for optimum production. Turmeric is a medicinal crop used primarily as spices and traditional medicine. The input parameters for the model are solar radiation, temperature, humidity, and rainfall, while the output parameters are energy and food. After training, validation, and testing of the model, the error has been found as ­0.03711 and ­0.00494 for energy and food systems respectively. For system performance, the established model generates a minimum mean squared error and maximum regression correlation coefficient indicating a favourable relationship between measured and predicted values. The predicted energy and food production has been obtained at optimal conditions as 104.9097 kWh and 9.0955 kg, respectively. The performance indicators such as land equivalent ratio and payback period have been found as 1.73 and 9.49 respectively. Thus, the system is techno-economically viable and feasible worldwide satisfying the targets of sustainable development goals.