Statistical estimation the thermal conductivity of MWCNTs-SiO2/Water-EG nanofluid using the ridge regression method

• Published in: Physica A Vol. 537; p. 122782
• Main Authors: Xiaohong, Dai, Huajiang, Chen, Bagherzadeh, Seyed Amin, Shayan, Masoud, Akbari, Mohammad
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.01.2020
• Subjects: Nanofluid; Ridge regression; Solid volume fraction; Temperature; Thermal conductivity; Ridge regression; Solid volume fraction; Thermal conductivity; Temperature; Nanofluid
• ISSN: 0378-4371; 1873-2119
• DOI: 10.1016/j.physa.2019.122782

Ridge regression is a regularization method which evaluated according to the experimental results of the hybrid nanofluid containing SiO2 and MWCNTs suspended in water and ethylene glycol as the base fluid concerned to the thermal conductivity versus different amounts of nanoparticles concentration and temperature. The novelty of this study is that this method is very useful for decreasing the variance of the fit and improves its future predictions Also, it is applicable especially in handling the small training data sets. Meanwhile, if the training data set is small, ridge regression can find a solution based on the cross validation and the ridge regression penalty. The findings showed that the fit is almost perfect because the fit line is almost identical with the Y=T line indicating the ideal fit. Also, the slope and y-intercept values of the fit line are 0.98 and 0.0076, respectively. •Ridge regression is a method which evaluated according to the experimental data.•This method is applicable especially in handling the small training data sets.•The error between the experimental and numerical data is less than 2%.•The training data set has acceptable predictions for nanofluid thermal conductivity.