Articles
Uncertainty analysis and validation of an IoT-based condensation model for apples in the storage bin
Article number
1456_14
Pages
111 – 118
Language
English
Abstract
This study presents the uncertainty analysis of an IoT-based condensation model using Monte-Carlo simulations.
Seven key input parameters, namely, air temperature, apple surface temperature, relative humidity, mass transfer coefficient for condensation, air speed, apple diameter, and apple surface area, were varied within their absolute uncertainties to evaluate their impact on model output through sensitivity.
The model showed moderate accuracy in predicting model output (condensation amount and retention time). Differences between predictions and experimental data were mainly due to condensation on bin surfaces and simplified model assumptions.
Model inputs with greater uncertainty tended to have a stronger influence on the output.
Through sensitivity analysis, the relative humidity was found to be the most influential parameter, significantly affecting both condensation mass and retention time.
Apple surface area was equally important for condensation mass, while air and surface temperature had moderate but opposite effects, with their difference playing a key role in the condensation process.
Together, relative humidity and surface area accounted for 48% of the uncertainty in condensation.
Retention time increased by 16 min with a 1% rise in relative humidity, by 7 min with a 0.06°C rise in air temperature, and decreased by 7 min with a 0.06°C rise in surface temperature, each with other variables held constant.
The strong influence of live sensor inputs, particularly those with greater uncertainty, such as relative humidity, underscores the importance of using accurate, well-calibrated sensors.
Seven key input parameters, namely, air temperature, apple surface temperature, relative humidity, mass transfer coefficient for condensation, air speed, apple diameter, and apple surface area, were varied within their absolute uncertainties to evaluate their impact on model output through sensitivity.
The model showed moderate accuracy in predicting model output (condensation amount and retention time). Differences between predictions and experimental data were mainly due to condensation on bin surfaces and simplified model assumptions.
Model inputs with greater uncertainty tended to have a stronger influence on the output.
Through sensitivity analysis, the relative humidity was found to be the most influential parameter, significantly affecting both condensation mass and retention time.
Apple surface area was equally important for condensation mass, while air and surface temperature had moderate but opposite effects, with their difference playing a key role in the condensation process.
Together, relative humidity and surface area accounted for 48% of the uncertainty in condensation.
Retention time increased by 16 min with a 1% rise in relative humidity, by 7 min with a 0.06°C rise in air temperature, and decreased by 7 min with a 0.06°C rise in surface temperature, each with other variables held constant.
The strong influence of live sensor inputs, particularly those with greater uncertainty, such as relative humidity, underscores the importance of using accurate, well-calibrated sensors.
Authors
A.D. Sonawane, T.G. Hoffmann, R. Jedermann, M. Linke, P.V. Mahajan
Keywords
apple storage, cold chain, sensors, Monte Carlo, sensitivity, retention time
Online Articles (48)