Investigating apple surface condensation and mass loss with IoT and predictive modelling

Abstract

This study analysed condensation and mass loss in apple storage using an IoT system and predictive modelling. A condensation model, based on the mass transfer coefficient determined through the Sherwood number for the relation between convective and diffusive mass transfer, and a mass loss model incorporating transpiration and respiration, were developed to predict cumulative condensation, retention time, cumulative mass loss, and at the end, the total mass of apples in storage. The model was validated with individual and bulk apples kept under varying air temperature conditions. Data on air and surface temperature, humidity, air speed, and surface wetness were collected via sensors, processed through Raspberry Pi (embedded computer) and Kafka data streaming platform for predictions, and stored in the InfluxDB time-series database for visualisation. The real-time model predictions were effectively aligned with experimental trends. In individual apple trials, 45- and 60-minute on-off refrigeration cycles showed the predicted peak cumulative condensations of 0.05 and 0.12 g kg −1 , approximately 71 % and 77 % of experimental peaks, with mass losses of 0.27 and 0.24 g kg −1 over 12 hours, respectively. In bulk apple trials, predicted peaks were 0.03 and 1.03 g kg −1 , around 30 % and 58 % of experimental values, with mass losses of 0.96 and 0.82 g kg −1 over 60 hours for low and high-temperature fluctuations, respectively. Temperature fluctuations significantly influenced condensation and mass loss, with high fluctuations causing much greater cumulative condensation than low fluctuations. However, experimental peak cumulative condensation values were consistently higher than predicted ones, likely due to the accuracy of sensors, the complexity of the experimental setup, and theoretical model assumptions. Transpiration accounted for a larger portion of the total cumulative mass loss in apples compared to respiration. Additionally, longer condensation retention times resulted in reduced apple mass loss. • IoT model designed for predicting condensation and mass loss prediction. • The model showed robust real-time predictions across temperature fluctuations. • Surface condensation reduced apple mass loss by limiting transpiration. • IoT model could optimize refrigeration, enhancing energy efficiency and shelf life.

Topics & Concepts