Artificial neural network modeling of nanofluid flow in a microchannel heat sink using experimental data

The present paper deals with the artificial neural network modeling (ANN) of heat transfer coefficient and Nusselt number in TiO₂/water nanofluid flow in a microchannel heat sink. The microchannel comprises of 40 channels; each channel has a length of 4 cm, a width of 500 μm, and a height of 800 μm. In the ANN modeling of heat transfer coefficient and Nusselt number 23 and 72 datasets have been used, respectively. The experimental Nusselt number has been calculated based on three different thermal conductivity models, four volume fractions of 0, 0.5, 1, and 2%, two values of Reynolds number i.e. 400 and 1200 and three different heating rates including 50.6, 60.7, and 69.1 W. Therefore, the inputs that are introduced to the neural network are volume fraction of nanoparticles, Reynolds number, heating rate, and model number while the output of network is the Nusselt number. It is elucidated that an appropriately trained network can act as a good alternative for costly and time-consuming experiments on the nanofluid flow in microchannels. The average relative errors in the prediction of Nusselt number and heat transfer coefficients were 0.3% and 0.2%, respectively.     

Thermal conductivity of Cu/TiO2–water/EG hybrid nanofluid: Experimental data and modeling using artificial neural network and correlation

In the present paper, the thermal conductivity of hybrid nanofluids is experimentally investigated. The studied nanofluid was produced using a two-step method by dispersing Cu and TiO₂ nanoparticles with average diameter of 70 and 40 nm in a binary mixture of water/EG (60:40). The properties of this nanofluid were measured in various solid concentrations (0.1, 0.2, 0.4, 0.8, 1, 1.5, and 2%) and temperatures ranging from 30 to 60 °C. Next, two new correlations for predicting the thermal conductivity of studied hybrid nanofluids, in terms of solid concentration and temperature, are proposed that use an artificial neural network (ANN) and are based on experimental data. The results indicate that these two new models have great ability to predict thermal conductivity and show excellent agreement with the experimental results.     

A hybrid artificial neural network-genetic algorithm modeling approach for viscosity estimation of graphene nanoplatelets nanofluid using experimental data

Predicting the viscosity of graphene nanoplatelets nanofluid with the help of multi-layered perceptron artificial neural network and genetic algorithm was the main aim of this study. In order to achieve the experimental results nanofluid which contains graphene nanoplatelets and deionized water at 20 to 60 °C and 0.025, 0.05, 0.075, and 0.1 wt% is used. Furthermore, genetic algorithm in artificial neural network is used to improve the learning process. In other words, different weights have been chosen for neurons' relations. Also, the bias preoccupation is based on improvements by genetic algorithm. On the other hand, for analyzing the accuracy of the presented model which gives us the nanofluid viscosity predictions MAPE, RMSE, R², and MBE indexes were used. The values of the presented indexes are 0.777, 0.086, 0.985, − 0.0009 respectively. In case of comparison the results show that the presented model which is the combination of genetic algorithm and artificial neural network is compatible with experimental work.     

Rheological characteristics of MgO/oil nanolubricants: Experimental study and neural network modeling

The present paper deals with the rheological behavior of MgO nanoparticles suspended in synthetic motor oil. First, the viscosity of prepared nanofluids is measured at different concentrations and temperatures. The experiments are performed in the temperatures ranging from 5 °C to 65 °C, shear rates approximately up to 13,000 s^− 1, and concentrations of 0.25%, 0.5%, 0.75%, 1%, 1.5%, and 2.0%. The viscosity measurements revealed that all nanofluid samples exhibit shear-thinning behavior. The consistency and the power law index were obtained by curve-fitting. The curve-fitting results show that all power law indices were in the range of 0.8 to 0.91. Finally, artificial neural network (ANN) is used to model the experimental results.     

Thermal conductivity of ethylene glycol and water mixture based Fe3O4 nanofluid

Thermal conductivity of ethylene glycol and water mixture based Fe₃O₄ nanofluid has been investigated experimentally. Magnetic Fe₃O₄ nanoparticles were synthesized by chemical co-precipitation method and the nanofluids were prepared by dispersing nanoparticles into different base fluids like 20:80%, 40:60% and 60:40% by weight of the ethylene glycol and water mixture. Experiments were conducted in the temperature range from 20 °C to 60 °C and in the volume concentration range from 0.2% to 2.0%. Results indicate that the thermal conductivity increases with the increase of particle concentration and temperature. The thermal conductivity is enhanced by 46% at 2.0 vol.% of nanoparticles dispersed in 20:80% ethylene glycol and water mixture compared to other base fluids. The theoretical Hamilton–Crosser model failed to predict the thermal conductivity of the nanofluid with the effect of temperature. A new correlation is developed for the estimation of thermal conductivity of nanofluids based on the experimental data.     

The effect of alumina/water nanofluid particle size on thermal conductivity

This study examines the effect of particle size, temperature, and weight fraction on the thermal conductivity ratio of alumina(Al₂O₃)/water nanofluids. A Al₂O₃/water nanofluid produced by the direct synthesis method served as the experimental sample, and nanoparticles, each of a different nominal diameter (20, 50, and 100 nm), were dispersed into four different concentrations (0.5, 1.0, 1.5, and 2.0 wt%). This experiment measured the thermal conductivity of nanofluids with different particle sizes, weight fractions, and working temperatures (10, 30, 50 °C). The results showed a correlation between high thermal conductivity ratios and enhanced sensitivity, and small nanoparticle size and higher temperature. This research utilized experimental data to construct a new empirical equation, taking the nanoparticle size, temperature, and lower weight fraction of the nanofluid into consideration. Comparing the regression results with the experimental values, the margin of error was within ?3.5% to +2.7%. The proposed empirical equation showed reasonably good agreement with our experimental results.     

Experimental thermal conductivity of ethylene glycol and water mixture based low volume concentration of Al2O3 and CuO nanofluids

Thermal conductivity of ethylene glycol and water mixture based Al₂O₃ and CuO nanofluids has been estimated experimentally at different volume concentrations and temperatures. The base fluid is a mixture of 50:50% (by weight) of ethylene glycol and water (EG/W). The particle concentration up to 0.8% and temperature range from 15 °C–50 °C were considered. Both the nanofluids are exhibiting higher thermal conductivity compared to base fluid. Under same volume concentration and temperature, CuO nanofluid thermal conductivity is more compared to Al₂O₃ nanofluid. A new correlation was developed based on the experimental data for the estimation of thermal conductivity of both the nanofluids.     

Thermal performance of multi-walled carbon nanotubes (MWCNTs) in aqueous suspensions with surfactants SDBS and SDS

The surfactants of sodium dodecylbenzene sulfonate (SDBS) and sodium dodecyl sulfate (SDS) are used in multi-walled carbon nanotubes (MWCNT) aqueous solution respectively due to the hydrophobic nature of MWCNT. Thermal conductivities of nanofluid solutions are measured via the LAMBDA measuring system by transient hot wire method and compared as function of dispersing two different surfactants. MWCNT (hereinafter sometime referred to as CNTs) nanofluid gets a good dispersion and long time stability with both surfactants within 3/1 relative ratio mixture. However, the thermal conductivity of nanofluid decreases with increasing the concentration of both surfactants, and CNT nanofluid with SDBS exhibits better thermal conductivity than that with SDS dispersant. Finally the proper mixture ratio of CNT nanofluid with SDBS and pH value is examined and results show that 0.5 wt.% CNT nanofluids with 0.25 wt.% SDBS, at pH ≈ 9.0 condition display the best thermal performance which increases by 2.8% totally on thermal conductivity compared with that of base fluid distilled water (DW).     

Optimum operating conditions for a water purification process integrated to a heat transformer with energy recycling using neural network inverse

Artificial neural network inverse (ANNi) is applied to calculate the optimal operating conditions on the coefficient of performance (COP) for a water purification process integrated to an absorption heat transformer with energy recycling. An artificial neural network (ANN) model is developed to predict the COP which was increased with energy recycling. This ANN model takes into account the input and output temperatures for each one of the four components (absorber, generator, evaporator, and condenser), as well as two pressures and LiBr + H₂O concentrations. For the network, a feedforward with one hidden layer, a Levenberg–Marquardt learning algorithm, a hyperbolic tangent sigmoid transfer function and a linear transfer function were used. The best fitting training data set was obtained with three neurons in the hidden layer. On the validation data set, simulations and experimental data test were in good agreement (R > 0.99). This ANN model can be used to predict the COP when the input variables (operating conditions) are well known. However, to control the COP in the system, we developed a strategy to estimate the optimal input variables when a COP is required from ANNi. An optimization method (the Nelder–Mead simplex method) is used to fit the unknown input variable resulted from the ANNi. This methodology can be applied to control on-line the performance of the system.     

Effect of fabrication parameters on the thermophysical properties of sintered wicks for heat pipe applications

Porous wicks for use in a loop heat pipe were sintered from copper and Monel powder. These wicks are characterized in terms of their porosity, liquid permeability, capillary pressure and thermal conductivity. The effect of fabrication parameters (particle size and sintering conditions) on these properties is studied. The experimentally measured values of permeability and capillary pressure were compared with correlations available in the literature. The Kozeny–Carman correlation was found to overpredict the experimental values of permeability; while the modified Young–Laplace equation was found to predict within 5% of the measured capillary pressure. Additionally, a model for predicting the thermal conductivity of sintered wicks is developed. First, the ‘two-sphere model’ is used to relate the sintering conditions to the size of the connection (the ‘neck’) between two particles. Then, a finite element simulation is used to determine the thermal resistance of the bonded particles as a function of the neck between them. This thermal resistance is integrated in a random 3D resistor network as a means to model the multiple connections between spheres in a wick and to calculate the effective thermal conductivity of the wick. Results of the model are compared with experimental measurements of thermal conductivity of sintered copper wicks. Agreement between the model and the experimental measurements is good (within 15%) for sintering temperatures below 550 °C, and within 26% for sintering temperatures up to 950 °C. Finally, a generalized thermal conductivity chart is presented, which can be used to estimate the sintering temperature and time required to achieve the desired thermal conductivity.     

Numerical simulation of natural convection of nanofluid in a square enclosure: Effects due to uncertainties of viscosity and thermal conductivity

The present study aims to identify effects due to uncertainties in effective dynamic viscosity and thermal conductivity of nanofluid on laminar natural convection heat transfer in a square enclosure. Numerical simulations have been undertaken incorporating a homogeneous solid–liquid mixture formulation for the two-dimensional buoyancy-driven convection in the enclosure filled with alumina–water nanofluid. Two different formulas from the literature are each considered for the effective viscosity and thermal conductivity of the nanofluid. Simulations have been carried out for the pertinent parameters in the following ranges: the Rayleigh number, Ra_f = 10³–10⁶ and the volumetric fraction of alumina nanoparticles, ? = 0–4%. Significant difference in the effective dynamic viscosity enhancement of the nanofluid calculated from the two adopted formulas, other than that in the thermal conductivity enhancement, was found to play as a major factor, thereby leading to contradictory results concerning the heat transfer efficacy of using nanofluid in the enclosure.     

Enhanced thermal conductivity of ethylene glycol with single-walled carbon nanotube inclusions

In the present work, we report measurements of the effective thermal conductivity of dispersions of single-walled carbon nanotube (SWNT) suspensions in ethylene glycol. The SWNTs were synthesized using the alcohol catalytic chemical vapour deposition method. Resonant Raman spectroscopy was employed to estimate the diameter distribution of the SWNTs based on the frequencies of the radial breathing mode peaks. The nanofluid was prepared by dispersing the nanotubes using a bile salt as the surfactant. Nanotube loading of up to 0.2 vol% was used. Thermal conductivity measurements were performed by the transient hot-wire technique. Good agreement, within an uncertainty of 2%, was found for published thermal conductivities of the pure fluids. The enhancement of thermal conductivity was found to increase with respect to nanotube loading. The maximum enhancement in thermal conductivity was found to be 14.8% at 0.2 vol% loading. The experimental results were compared with literature results in similar dispersion medium. Experimental results were compared with the Hamilton–Crosser model, the Lu–Lin model, Nan’s effective medium theory and the Hashin–Shtrikman model. Effective medium theory seems to predict the thermal conductivity enhancement reasonably well compared to rest of the models. Networking of nanotubes to form a tri-dimensional structure was considered to be the reason for the thermal conductivity enhancement.     

Experimental study on the thermo-physical properties of car engine coolant (water/ethylene glycol mixture type) based SiC nanofluids

The engine coolant (water/ethylene glycol mixture type) becomes one of the most commonly used commercial fluids in cooling system of automobiles. However, the heat transfer coefficient of this kind of engine coolant is limited. The rapid developments of nanotechnology have led to emerging of a relatively new class of fluids called nanofluids, which could offer the enhanced thermal conductivity (TC) compared with the conventional coolants. The present study reports the new findings on the thermal conductivity and viscosity of car engine coolants based silicon carbide (SiC) nanofluids. The homogeneous and stable nanofluids with volume fraction up to 0.5 vol.% were prepared by the two-step method with the addition of surfactant (oleic acid). It was found that the thermal conductivity of nanofluids increased with the volume fraction and temperature (10–50 °C), and the highest thermal conductivity enhancement was found to be 53.81% for 0.5 vol.% nanofluid at 50 °C. In addition, the overall effectiveness of the current nanofluids (0.2 vol.%) was found to be ~ 1.6, which indicated that the car engine coolant-based SiC nanofluid prepared in this paper was better compared to the car engine coolant used as base liquid in this study.     

Performance characteristics of a microchannel heat sink using TiO2/water nanofluid and different thermophysical models

In this paper, the performance of a microchannel heat sink using TiO₂/water nanofluid is experimentally investigated. The dimensions of the microchannel are 500 μm width, 800 μm height, and 40 mm length, where the number of flowing channels is 40.The effects of uncertainties in thermophysical properties on the Nusselt number and friction factor are investigated by using three different sets of thermophysical models, which are based on experimental and theoretical relations. It is concluded that the use of the model which is based on experimental data is very important to estimate the friction factor, while the use of different models to calculate of thermal conductivity has no considerable effect on the prediction of Nusselt number.     

Experimental investigation and development of new correlation for thermal conductivity and viscosity of BioGlycol/water based SiO2 nanofluids

Nanofluids are a new class of engineered heat transfer fluids which exhibit superior thermophysical properties and have potential applications in numerous important fields. In this study, nanofluids have been prepared by dispersing SiO₂ nanoparticles in different base fluids such as 20:80% and 30:70% by volume of BioGlycol (BG)/water (W) mixtures. Thermal conductivity and viscosity experiments have been conducted in temperatures between 30 °C and 80 °C and in volume concentrations between 0.5% and 2.0%. Results show that thermal conductivity of nanofluids increases with increase of volume concentrations and temperatures. Similarly, viscosity of nanofluid increases with increase of volume concentrations but decreases with increase of temperatures. The maximum thermal conductivity enhancement among all the nanofluids was observed for 20:80% BG/W nanofluid about 7.2% in the volume concentration of 2.0% at a temperature of 70 °C. Correspondingly among all the nanofluids maximum viscosity enhancement was observed for 30:70% BG/W nanofluid about 1.38-times in the volume concentration of 2.0% at a temperature of 70 °C. The classical models and semi-empirical correlations failed to predict the thermal conductivity and viscosity of nanofluids with effect of volume concentration and temperatures. Therefore, nonlinear correlations have been proposed with 3% maximum deviation for the estimation of thermal conductivity and viscosity of nanofluids.     

Simulation of heat transfer enhancement in nanofluids using dissipative particle dynamics

The current paper applied dissipative particle dynamics (DPD) approach to investigate heat transfer within nanofluids. The DPD approach was applied to study natural convection in a differential heated enclosure by considering the viscosity and the thermal conductivity of the nanofluid to be dual function of temperature and volume fraction of nanoparticles. Experimental data for viscosity and thermal conductivity are incorporated in the current DPD model to mimic energy transport within nanofluids. This incorporation is done through the modification of the dissipative weighting function that appears in the dissipative force vector and the dissipative heat flux. For the entire range of Rayleigh number considered in this study, it was found that the DPD results show a deterioration in heat transfer in the enclosure due to the presence of nanoparticles for φ > 4%. However, some slight enhancement is shown to take place for small volume fraction of nanoparticles, φ ≤ 4%. The DPD results experienced some degree of compressibility at high values of Rayleigh number Ra ≥ 10⁵.     

A new approach using artificial neural networks for determination of the thermodynamic properties of fluid couples

This paper presents a new approach using artificial neural networks (ANN) to determine the thermodynamic properties of two alternative refrigerant/absorbent couples (LiCl–H₂O and LiBr + LiNO₃ + LiI + LiCl–H₂O). These pairs can be used in absorption heat pump systems, and their main advantage is that they do not cause ozone depletion. In order to train the network, limited experimental measurements were used as training and test data. Two feedforward ANNs were trained, one for each pair, using the Levenberg–Marquardt algorithm. The training and validation were performed with good accuracy. The correlation coefficient obtained when unknown data were applied to the networks was 0.9997 and 0.9987 for the two pairs, respectively, which is very satisfactory. The present methodology proved to be much better than linear multiple regression analysis. Using the weights obtained from the trained network, a new formulation is presented for determination of the vapor pressures of the two refrigerant/absorbent couples. The use of this new formulation, which can be employed with any programming language or spreadsheet program for estimation of the vapor pressures of fluid couples, as described in this paper, may make the use of dedicated ANN software unnecessary.     

Estimating effective thermal conductivity of two-phase materials

In this paper, correlations are proposed to estimate the effective thermal conductivity of two-phase materials. For any α, Maxwell equation for 0.0 < c ⩽ 0.10 and phase inverted Maxwell for 0.9 ⩽ c ⩽ 1 are considered. For concentrations between 10% and 90%, and low α (<20), an equation based on the unit-cell approach (constant isotherms) is proposed. For α > 20, three correlations are proposed based on field solution approach which includes three α ranges viz. medium (20 ⩽ α ⩽ 100), high (100 ⩽ α ⩽ 1000) and very high (1000 < α). The predicted effective thermal conductivity of two-phase system is compared with well-established models. Comparison of the predicted values of the correlations with experimental results is also made. The predictions of effective thermal conductivity of two-phase materials match well with the experimental values.     

Transport properties of ultra-low concentration CuO–water nanofluids containing non-spherical nanoparticles

CuO–water nanofluids were prepared from non-spherical CuO nanoparticles by dispersing them in water through the aid of ultrasonication along with the use of Tiron as dispersant. Thermal conductivity enhancements of 13% and 44% have been obtained with 0.016 vol% CuO–water nanofluids at 28 °C and 55 °C respectively, which could be attributed to the high aspect ratio and Brownian motion of nanoparticles. Correlations have been developed to predict the influence of temperature (28–55 °C) and nanoparticles volume concentration (<0.016 vol%) on relative viscosity and thermal conductivity ratio. The results indicate the potential of this nanofluid for thermal management applications.     

An experimental study on thermal conductivity of MgO nanoparticles suspended in a binary mixture of water and ethylene glycol

In the present study, the effects of solid volume fraction and temperature on the thermal conductivity of MgO/water–EG (60:40) nanofluid are discussed. Samples of nanofluid are provided by two step method at different solid concentrations, including 0.1%, 0.2%, 0.5%, 0.75%, 1%, 1.5%, 2% and 3%. The experiments are performed for different temperatures ranging from 20 to 50 °C, using KD2 pro thermal analyzer which employed transient hot wire to measure thermal conductivity. The finding shows that thermal conductivity of nanofluid increases with increasing solid volume fraction or temperature. Based on the experimental data, new correlation for modeling the thermal conductivity of MgO/water–EG (60:40) for different solid volume fractions and temperatures was proposed.