Modeling of the frost deposition by natural convection on horizontal ultra-low-temperature surfaces

The present work develops a reliable predictive model for precise estimation of the frost layer thickness by free convection on horizontal ultra-low-temperature surfaces. Wall temperature, relative humidity, time, and air temperature are considered as the input variables, and six well-known heuristi...

Full description

Saved in:
Bibliographic Details
Published in:Journal of thermal analysis and calorimetry 2019-09, Vol.137 (6), p.2029-2043
Main Authors: Zendehboudi, Alireza, Hosseini, S. H.
Format: Article
Language:English
Subjects:
Air temperature
Algorithms
Analytical Chemistry
Chemistry
Chemistry and Materials Science
Comparative analysis
Free convection
Genetic algorithms
Genetic vectors
Humidity
Inorganic Chemistry
Low temperature
Measurement Science and Instrumentation
Physical Chemistry
Polymer Sciences
Relative humidity
Sensitivity analysis
Support vector machines
Thickness
Wall temperature
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The present work develops a reliable predictive model for precise estimation of the frost layer thickness by free convection on horizontal ultra-low-temperature surfaces. Wall temperature, relative humidity, time, and air temperature are considered as the input variables, and six well-known heuristic models are developed to estimate the desired output. The comparative results demonstrate that the least square support vector machine incorporating the genetic algorithm (GA-LSSVM) outperforms the other approaches. The coefficient of determination of 0.9998 and 0.9976, average absolute relative deviation of 0.8536% and 9.4002%, root mean squared error of 0.0115 and 0.0486, and relative root mean square error of 1.4479 and 5.8989 are the results of training and testing stages of the suggested model, respectively. A new test condition is studied to verify applicability of the proposed approach in computing the values that have not been evaluated in the experiments. It is observed that a decrease in the wall temperature causes a decrease in the frost layer thickness on horizontal surfaces under ultra-low-temperature conditions. The non-dominated sorted genetic algorithm II is also employed and combined with LSSVM model to study a sensitivity analysis. According to the Pareto optimal solutions, the time, wall temperature, air temperature, and relative humidity are, respectively, the most influential parameters on the frost layer thickness.
ISSN:1388-6150
1588-2926
DOI:10.1007/s10973-019-08087-x