Experimental analysis of novel ionic liquid-MXene hybrid nanofluid's energy storage properties: Model-prediction using modern ensemble machine learning methods
Identifikátory výsledku
Kód výsledku v IS VaVaI
<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F68407700%3A21220%2F22%3A00358157" target="_blank" >RIV/68407700:21220/22:00358157 - isvavai.cz</a>
Výsledek na webu
<a href="https://doi.org/10.1016/j.est.2022.104858" target="_blank" >https://doi.org/10.1016/j.est.2022.104858</a>
DOI - Digital Object Identifier
<a href="http://dx.doi.org/10.1016/j.est.2022.104858" target="_blank" >10.1016/j.est.2022.104858</a>
Alternativní jazyky
Jazyk výsledku
angličtina
Název v původním jazyce
Experimental analysis of novel ionic liquid-MXene hybrid nanofluid's energy storage properties: Model-prediction using modern ensemble machine learning methods
Popis výsledku v původním jazyce
The current work employs two modern ensemble machine learning algorithms, Matern 5/2 Gaussian process regression (GPR) and quadratic support vector regression (SVR), to model-predict the thermal conductivity, specific heat, and viscosity of novel Ionic liquid-MXene hybrid nanofluids. The data for model development was obtained in the lab at various temperatures and mass concentrations. The obtained thermal conductivity value for the pure aqueous Ionic liquid (IL) solution at 20 °C was 0.443 W/m⋅K which rose to 0.82 W/m⋅K due to the inclusion of 0.5 wt% MXene nanomaterial. The achieved result for the specific heat capacity of the pure aqueous IL solution is 1.985 J/g⋅K, while this value enhances up to 2.374 J/g⋅K due to the inclusion of the highest loading concentration (0.5 wt%) of MXene nanomaterial. The acquired data was divided into two portions, with 70% of the data utilized for model training and 30% (hold-out) data used for model testing. Several statistical indicators were used to evaluate the GPR and SVR-based prognostic models created for specific heat, viscosity, and thermal conductivity. For SVM models, the correlation coefficient (R) ranged from 0.9741 to 0.9958, while for GPR-based models, it ranged from 0.9942 to 0.9998. The inaccuracy in the prediction model was quantified using root mean squared error, which ranged from 0.0129 to 0.0398 for SVM-based models and 0.004 to 0.105 for GPR-based models. The predictive effectiveness of the models was tested using Willmott's Index of Agreement, which was 0.9833 to 0.999 for SVM-based models and 0.9834 to 0.9999 for GPR-based models, respectively. The close to unity regression coefficient, low model error, and good prediction effectiveness illustrate the robustness of both SVM and GPR prognostic models. However, the GPR outperformed the SVM on all the statistical indices. The experiment results were also used to compute the Mouromtseff number (<4) and Figure of merit (>1). The results showed that the studied nanofluids can successfully substitute water in specified applications for solar energy.
Název v anglickém jazyce
Experimental analysis of novel ionic liquid-MXene hybrid nanofluid's energy storage properties: Model-prediction using modern ensemble machine learning methods
Popis výsledku anglicky
The current work employs two modern ensemble machine learning algorithms, Matern 5/2 Gaussian process regression (GPR) and quadratic support vector regression (SVR), to model-predict the thermal conductivity, specific heat, and viscosity of novel Ionic liquid-MXene hybrid nanofluids. The data for model development was obtained in the lab at various temperatures and mass concentrations. The obtained thermal conductivity value for the pure aqueous Ionic liquid (IL) solution at 20 °C was 0.443 W/m⋅K which rose to 0.82 W/m⋅K due to the inclusion of 0.5 wt% MXene nanomaterial. The achieved result for the specific heat capacity of the pure aqueous IL solution is 1.985 J/g⋅K, while this value enhances up to 2.374 J/g⋅K due to the inclusion of the highest loading concentration (0.5 wt%) of MXene nanomaterial. The acquired data was divided into two portions, with 70% of the data utilized for model training and 30% (hold-out) data used for model testing. Several statistical indicators were used to evaluate the GPR and SVR-based prognostic models created for specific heat, viscosity, and thermal conductivity. For SVM models, the correlation coefficient (R) ranged from 0.9741 to 0.9958, while for GPR-based models, it ranged from 0.9942 to 0.9998. The inaccuracy in the prediction model was quantified using root mean squared error, which ranged from 0.0129 to 0.0398 for SVM-based models and 0.004 to 0.105 for GPR-based models. The predictive effectiveness of the models was tested using Willmott's Index of Agreement, which was 0.9833 to 0.999 for SVM-based models and 0.9834 to 0.9999 for GPR-based models, respectively. The close to unity regression coefficient, low model error, and good prediction effectiveness illustrate the robustness of both SVM and GPR prognostic models. However, the GPR outperformed the SVM on all the statistical indices. The experiment results were also used to compute the Mouromtseff number (<4) and Figure of merit (>1). The results showed that the studied nanofluids can successfully substitute water in specified applications for solar energy.
Klasifikace
Druh
J<sub>imp</sub> - Článek v periodiku v databázi Web of Science
CEP obor
—
OECD FORD obor
20303 - Thermodynamics
Návaznosti výsledku
Projekt
—
Návaznosti
I - Institucionalni podpora na dlouhodoby koncepcni rozvoj vyzkumne organizace
Ostatní
Rok uplatnění
2022
Kód důvěrnosti údajů
S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů
Údaje specifické pro druh výsledku
Název periodika
Journal of Energy Storage
ISSN
2352-152X
e-ISSN
2352-1538
Svazek periodika
2022 (52)
Číslo periodika v rámci svazku
05
Stát vydavatele periodika
GB - Spojené království Velké Británie a Severního Irska
Počet stran výsledku
13
Strana od-do
—
Kód UT WoS článku
000807731100001
EID výsledku v databázi Scopus
2-s2.0-85130572918