Development of a Predictive Model of Water Hyacinth Drying in Solar Parabolic Dome with Artificial Neural Networks

Abstract

Water hyacinth benefits humans and the environment. Cellulose makes water Hyacinth a biofuel. It encourages eco-friendly furniture, décor, and toys. This study dries hyacinths of various sizes in a solar parabolic dome to investigate how they dehumidify and then builds an artificial neural network model to forecast relative humidity. This study includes statistical and artificial neural network analysis at all experimental stages. Drying conditions in solar parabolic greenhouses reveal that all three hyacinth sizes dehumidify over time, decreasing the moisture ratio. These differences may be due to a tighter internal fiber structure or crust formation on the material's surface during baking, which slows water evaporation from the inside to the outside. This illustrates that material size affects dehumidification performance. Operating time and Hyacinth size affect the drying rate. We assessed the model's performance using R² and MSE values. These studies show that a few nodes may regulate automated and industrial drying processes. ANNs simplify machine learning and optimization by revealing complex data patterns and linkages, which are crucial for time series prediction, classification, and regression. Big data can help them improve over time.