Optimization of vertical shelves sizing affecting sunlight availability for strawberries cultivation

Abstract

The goal of this research is to develop an optimization model for predicting the sunlight availability on each level of the VF (vertical farming) shelves for strawberry plantations inside a greenhouse. Experiments were carried out in a 100 m2 land area in Chiang Mai, Thailand. A greenhouse was constructed, and six VF shelves with three levels on each shelf were built. First, the experiments were carried out by recording sunlight availability on each level of the shelf. A computer model was developed using RHINO (Rhinoceros) to match the physical conditions of the experiment room. Then, the sunlight availability was simulated for each shelf using the Grasshopper plugin with Chiang Mai weather data. It was then validated using the collected data. After the model was validated, it was used to simulate the sunlight availability on each level of the shelves for different designs of the VF shelves. The simulated values are used to train an ANN (Artificial Neural Network) optimization model, which can give out the required design of VF shelves to receive desired values of sunlight availability on each level of the shelf. The main factors affecting sunlight availability are the weather, orientation of the shelves, and vertical and horizontal spacings. Different VF shelves designs are developed by varying the horizontal and vertical spacings between each level. The sunlight availability on each level on the shelves is recorded for each design. These sets of sunlight availability and the spacings are used to train the ANN model. The developed ANN model is capable of estimating the required dimension of the vertical farming shelves when the desired sunlight availability is used as input.