Recent state of nanofluid in automobile cooling systems

Abstract

Nanofluid that made up of fluid and solid nanoparticles has gained attention from diverse fields due to its superior thermophysical properties to enhance the performance of different systems which require flowing medium with excellent heat transfer behavior. Many past researchers have proven that conventional heat transfer fluid can be replaced by the rising nanotechnology–nanofluid which showed astonishing performance under different circumstances. In this paper, we attempt to present a recent review on the consequences of implantation of nanofluid, especially in vehicle engine cooling system and other heat transfer applications such as solar collector, electronics cooling system, flow boiling and thermal energy storage system. Thermophysical properties and heat transfer performance of nanofluids obtained in simulation, test rigs and even real vehicle engine experiments are discussed thoroughly. Models and correlations used by past researchers to compute thermophysical properties are also included. In the last part, various advantages from using nanofluid are summarized, and suggestions for research gap between past studies are discussed to further improve the investigation work in the future.

Introduction

With expeditious development of science and technology, nanotechnology is utilized in diverse applications, especially heat exchanger, electronics cooling, solar energy, biomedical, refrigeration and thermal energy storage system. The miniaturization of devices and materials is prevailing trend nowadays. One of the most common nanoscience applications today is called nanofluid. Nanofluid was first proposed by Choi [1], and it is basically a two-phase system, which consists of base fluid and suspended solid nanoparticles. Nanoparticles typically sized less than 100 nm and usually made up of materials listed in Fig. 1.

Fig. 1

Common material for base fluid and nanoparticles

As nanoparticles are dispersed into base fluid, rheological behavior and thermophysical properties of base fluid will be greatly influenced. Among all thermophysical properties, thermal conductivity plays a vital role in most engineering applications as it represents the capability of a material to transfer heat. Figure 2 shows the difference of thermal conductivity among commonly used components in nanofluids [2]. It can be seen that thermal conductivity of fluid is significantly lower when compared to that of solid particles. Thus, many past researchers studied on the effects of mixing solid nanoparticles into fluids and observed that nanofluids show superb heat transfer characteristics [3,4,5] and heat transfer performance [6,7,8,9,10] compared to base fluid.

Fig. 2

Thermal conductivities of different polymers, liquids and solids [2] (License Number: 4344040672984)

Up to now, there is only a comprehensive review focused on automobile engine cooling system with utilization of nanofluid as nanocoolant [11]. To the best of authors’ knowledge, there are some past researches about applications of nanofluid in real vehicle engine and automobile radiators. It is found that there is still no review reported on behavior of nanofluids in different types of automobile radiators. Hence, authors are inspired to include this discrepancy and provide an extensive review on nanofluids in automobile radiator cooling system and other major applications which involve impregnation of nanofluid to improve thermophysical properties, efficiency and heat transfer performance.

Thermophysical properties of nanofluid

Due to exceptional thermal properties of solid material compared to fluid, Stephen Choi [1] expected nanofluid could be possible substitute for conventional heat transfer fluid. Since then, nanofluid started to draw attention of researchers from different fields and they looked up the outcome from using nanofluid. One of the earliest efforts in measuring thermal conductivity was carried out by Lee et al. [12]. Transient hot-wire method was used for measuring purpose, and it was found that low concentration of oxide nanofluid showed surprisingly high thermal conductivity than base fluid. Since then, many researchers conducted studies on thermophysical properties of nanofluid in different environments.

Experimental studies

In 2010, Kole and Dey [13] measured viscosity of water/propylene glycol mixed with alumina nanoparticles using Brookfield programmable viscometer. Their results showed that relationship between viscosity and temperature of the nanocoolant agreed well with empirical correlation proposed by Namburu et al. [14] in year 2007, with maximum deviation of less than 2%. In Fig. 3, dots indicate their obtained results, while dotted lines are Namburu’s correlation.

Fig. 3

Relationship between log viscosity and temperature [13]

A major study by Wang et al. [15] included the dispersion of two different nanoparticles (Al₂O₃ and CuO) into water, engine oil, ethylene glycol and vacuum pump fluid. Steady-state parallel-state method was used to measure thermal conductivity, and they reported that thermal conductivities of all nanofluids were higher than respective base fluids. Another novel study on investigating the relationship between temperature and thermal conductivity was carried out by Das and his team [16]. From their result, it was observed that the enhancement of thermal conductivity for 4 vol% Al₂O₃–water nanofluid was increased from 9.4 to 24.3% when temperature increased from 21 to 51 °C.

Chen and Jia [17] reported the enhancement of thermal conductivity of 3% when mass fraction of TiO₂ in water/ethylene glycol was varied from 0.5 to 5%. Measurement of thermophysical properties of 13 nm Al₂O₃–water/ethylene glycol as car radiator coolant was taken by Elias and his squad [18] in 2014. Their result revealed that maximum enhancement for thermal conductivity, viscosity and density was 8.30, 150 and 2.91%, respectively, at 1 vol% Al₂O₃ in the range of 10–15 °C.

Kh and his team [19] investigated thermophysical and rheological properties of water/ethylene glycol nanofluid with functionalized graphene nanoplatelets. Their results showed that the thermal conductivity of 0.2 mass% nanoplatelets was about 58% higher than that of base fluid at 65 °C; meanwhile, dynamic viscosity showed 4.86% of increment at the same conditions. Selvam and his group [20] reported that thermal conductivity of 0.45 vol% graphene in water/ethylene glycol nanofluid increased thermal conductivity by 18% but decreased specific heat capacity by 8%.

Multi-walled carbon nanotube (MWCNT) nanofluid in a square enclosure was studied by Garbadeen et al. [21]. They measured thermal conductivity and viscosity using KD2 Pro and SV10 Sine-wave Vibro Viscometer, respectively. For 0–1 vol% of MWCNT, maximum enhancement of thermal conductivity and viscosity is found to be 6 and 58%, respectively, when compared to water. Thakur et al. [22] also measured thermophysical properties of same nanofluid at temperature of 30–70 °C and concentration of 0–0.8 vol%. They reported that 23% of enhancement in thermal conductivity was found at 0.8 vol% and 70 °C. Moreover, specific heat was found to decrease when nanoparticle concentration was increased which in conformity with experimental work by Ilyas et al. [23] who worked on MWCNT–thermal oil.

Modification of surface properties of SiO₂ nanoparticles by depositing copper was carried out by Amiri et al. [24]. For 50–80 nm modified SiO₂ nanoparticles produced, they used transient hot-wire method to obtain thermal conductivity of the nanoparticles dispersed in water. From their experiment, thermal conductivity could be enhanced up to 11% by using less than 1 vol% modified nanoparticles. In addition to that, they claimed that this new nanocomposite has better resistance against oxidation in air compared to pure metal.

Few researchers investigated the relationship between base fluid ratio and thermophysical properties. Abdolbaqi and his team [25] prepared water-/bioglycol-based SiO₂ nanofluid in 20:80 and 30:70% base fluid ratio. Temperature and nanoparticle concentration were varied between 30–80 °C and 0.5–2.0 vol% in the experiment. For 20:80% base fluid nanofluid, 7.2% of thermal conductivity enhancement was obtained at 2.0 vol% SiO₂ and 70 °C. Besides that, viscosity was increased up to 29.8 and 53.4% at 30 °C and 60 °C, respectively, for 30:70% base fluid nanofluid with 2.0 vol% SiO₂. With the same intention, Chiam et al. [26] prepared Al₂O₃ nanofluids with different base fluid ratios (40:60, 50:50 and 60:40) and tested them from 30 to 70 °C. They highlighted that higher portion of ethylene glycol in mixture led to increment of thermal conductivity and decrement of viscosity. Their results revealed 12.8% of thermal conductivity enhancement for 40:60 base fluid and 50% increment of dynamic viscosity for 60:40 base fluid, with 1.0 vol% Al₂O₃.

Glycerin (G13) was mixed by Sundari and his group [27] with 0.05–0.15 vol% of Al₂O₃ to measure the thermophysical properties from 30 °C to 50 °C. They varied the operating temperature from 30 °C to 50 °C, and the viscosity was decreased by 33.84%. At 40 °C, 0.15 vol% Al₂O₃ increased thermal conductivity for 46.15%. Furthermore, it was found that surface tension and pH value were inversely proportional to temperature.

Nabil et al. [28] prepared and measured a hybrid nanofluid that made up of TiO₂–SiO₂ (50:50) and water/ethylene glycol (60:40). Maximum error of 1.6% was found when the measured data were compared to ASHRAE. For 3.0 vol% nanoparticles, maximum thermal conductivity was enhanced by 22.8% and average relative viscosity obtained 62.5% increment. They suggested that this hybrid nanofluid could benefit heat transfer applications with the addition of at least 1.5% vol% nanoparticles concentration.

Functionalized single-walled carbon nanotube (F-SWCNT) was dispersed in water/ethylene glycol by Adhami et al. [29] for determining the thermophysical properties. From 0.025 to 0.65% volume fraction, temperature showed huge impact on thermal conductivity when volume fraction is more than 0.53%. Additionally, they compared alumina–water/ethylene glycol nanofluid with existing nanofluid and found that only 5% of increment in thermal conductivity from alumina nanofluid, whereas F-SWCNT nanofluid showed 52.7% under identical conditions.

Iqbal and his team [30] presented thermal conductivities of different deionized water-based nanofluids at same volume concentration. At 1 vol% nanoparticles, respective thermal conductivity increment for Al₂O₃, SiO₂ and ZrO₂ was 10.13, 6.5 and 8.5%. As other researchers, this team also mentioned that viscosity is directly proportional to nanoparticles concentration.

For hybrid Fe₂O₃/MWCNT water-based nanofluid, Chen et al. [32] varied the concentration of Fe₂O₃ nanoparticles and measured the thermal conductivity. With 0.02 mass% Fe₂O₃ and 0.05 mass% MWCNT, they obtained 27.7% of enhancement in thermal conductivity which was higher than 0.02 mass% MWCNT and 0.02 mass% Fe₂O₃ nanofluid alone. However, when they added more Fe₂O₃ (> 0.02 mass%) in the hybrid suspension, thermal conductivity decrement was observed. They proposed that high concentration of nanoparticles would lead to agglomeration easily, in which affecting heat transfer performance significantly.

Summary of experimental studies on thermophysical properties of nanofluids reviewed earlier is tabulated in Table 1, with some other undiscussed researches.

Table 1 Summary of experimental studies of thermophysical properties for nanofluids

Empirical correlations and equations

Based on the literature review, different models were used by former researchers to compute thermophysical properties. Maxwell model [42] is one of the earliest models to compute thermal conductivity of solid–liquid mixture and commonly modified by former researchers to develop new thermal conductivity models. Hamilton and Crosser [43] modified Maxwell model and proposed shape factor (n) which can be used for other nanoparticles shapes rather than spherical. Yu and Choi [44] expanded Maxwell model by assuming a solid-like nanolayer with thickness (h) surrounded a spherical nanoparticle with radius (r) and thus form a bigger radius of r + h nanoparticle. Koo and Kleinstreuer [45] added the effects of mixture temperature, nanoparticle size and concentration, Brownian motion of nanoparticles into Maxwell model. Pak and Cho [46] suggested that thermal conductivity is mainly affected by dispersion of nanoparticles. On the other side, Maiga et al. [47] suggested that some past researchers underestimated viscosity and thermal conductivity of nanofluids, and thus, they proposed few correlations using least-square curve fitting of past experimental data. Similarly, Corcione [48] extracted experimental data from past studies and proposed effective thermal conductivity and viscosity correlations with about 1.85% of standard deviation error.

For dynamic viscosity models, Einstein model [49] is the earliest model developed. It was then modified by Brinkman to add in viscosity and volume fraction of both nanoparticles and base fluid. Nanoparticles concentration was taken into account by Wang et al. [15] when calculating viscosity of nanofluid. Batchelor [50] considered the effect of Brownian motion in a suspension containing rigid spherical nanoparticles. On the other hand, Gherasim et al. [51] considered only spherical nanoparticles in their proposed model. Brownian motion of nanoparticles was considered as one of the factors affecting effective viscosity, as proposed by Tiwari and Das [52].

Khanafer and Vafai [53] tested the reliability of the first density equation in Table 1. Also, they developed a correlation for density of aluminum oxide based on experimental data from Ho et al. [54]. For specific heat capacity models, the authors [53] compared the first and second models in Table 1 with experimental data from past researchers. It was found that second model could give more accurate results due to an assumption in the equation which considers thermal equilibrium between base fluid and nanoparticles.

Although there are many available equations and correlations in the literature, different assumptions were made, and this led to distinct results among past researchers even though similar approaches were used on same nanoparticles, as reported in the past study [53]. Hence, general correlations for thermal conductivity and viscosity of Al₂O₃–water were developed by combining past experimental data:

$$\frac{{k_{\text{eff}} }}{{k_{\text{f}} }} = 0.9843 + 0.398\varphi_{\text{p}}^{0.7383} \left( {\frac{1}{{d_{\text{p}} }}} \right)^{0.2246} \left( {\frac{{\mu_{\text{eff}} \left( T \right)}}{{\mu_{\text{f}} \left( T \right)}}} \right)^{0.0235} - 3.9517\frac{{\varphi_{\text{p}} }}{T} + 34.034\frac{{\varphi_{\text{p}}^{2} }}{{T^{3} }} + 32.509\frac{{\varphi_{\text{p}} }}{{T^{2} }}$$

(1)

where 0 ≤ $\varphi_{p}$ ≤ 10%, 11 nm ≤ d_p ≤ 150 nm, 20 °C ≤ T ≤ 70 °C. This correlation was tested with experimental data [16, 55, 56], and results were in good agreement.

$$\mu_{\text{eff}} = - 0.4491 + \frac{28.837}{T} + 0.574\varphi_{\text{p}} - 0.1634\varphi_{\text{p}}^{2} + 23.053\frac{{\varphi_{\text{p}}^{2} }}{{T^{2} }} + 0.0132\varphi_{\text{p}}^{3} - 2354.735\frac{{\varphi_{\text{p}} }}{{T^{3} }} + 23.498\frac{{\varphi_{\text{p}}^{2} }}{{d_{\text{p}}^{2} }} - 3.0185\frac{{\varphi_{\text{p}}^{3} }}{{d_{\text{p}}^{2} }}$$

(2)

where 1% ≤ $\varphi_{\text{p}}$ ≤ 10%, 13 nm ≤ d_p ≤ 131 nm, 20 °C ≤ T ≤ 70 °C. This correlation was proven in line with results from past studies [46, 57,58,59].

On the other side, the effect on Nusselt number using different thermal conductivity models was numerically compared by Ogut and Kahveci [60] in 2016. Nusselt number of Al₂O₃–water/ethylene glycol nanofluid was analyzed in a square enclosure with lid-driven. Four models studied were Pak and Cho model [46], Yu and Choi model [44], Ghanbarpour et al. model [61], Maxwell model [42], and Timofeeva et al. [62] model. Among all models, Pak and Cho model gave the highest average number of Nusselt number. Noted that M3 in two graphs shown below is Pak and Cho model (Fig. 4).

Fig. 4

Average Nusselt number for a 0:100% and b 40:60% ethylene glycol/water mixture (License Number: 4344041420177)

In short, each model has different considerations on different aspects, and thus, many modified models and new correlations were proposed by former researchers specifically for certain nanoparticles and working parameters. Some of the earliest models, modified models and correlations adopted are summarized in Table 2, with additional information obtained from the past study [63].

Table 2 Models used to compute thermophysical properties

Artificial neural network (ANN) modeling in predicting thermophysical properties

Not only traditional method can be used to measure thermophysical properties, but artificial neural network (ANN) is able to present experimental data in shorter time and even more accurate than existing mathematical model. It simulates human brain neural network as artificial neurons in data processing. Artificial neurons are made up of different artificial neurons in which each unit represents specified input, computation/function and output [79] (Fig. 5).

Fig. 5

Concept diagram of ANN

Hemmat et al. [80] determined thermal conductivity of hybrid nanofluid using transient hot-wire method with KD2 Pro conductivity meter. Firstly, MWCNTs with inner diameter of 3–5 nm and outer diameter of 5–15 nm were mixed with 10–30 nm ZnO nanoparticles and water/ethylene glycol mixture. Thermal conductivity was then measured from 0.02 to 1 vol%. Among the collected results, maximum thermal conductivity (TCR) was obtained at concentration of 1 vol% and temperature of 50 °C. Their proposed correlation and ANN were able to provide maximum error of 97.4 and 98%, and R-squared value of 0.9864 and 0.9968, respectively, in predicting TCR.

Thermal conductivity of CuO–single-walled carbon nanotubes (SWCNTs) dispersed in water/ethylene glycol was investigated by Rostamian et al. [79]. Temperature was altered from 20 to 50 °C, and nanoparticles concentration ranged from 0.02 to 0.75 vol%. They found that concentration variation was more dominant than operating temperature in thermal conductivity increment. Moreover, ANN gave more precise and accurate answer than their proposed correlation to estimate thermal conductivity of the hybrid nanofluid as maximum deviation of ANN and correlation was 0.544 and 4%, respectively.

Zhao and Li [81] predicted the thermal conductivity and viscosity of alumina–water nanofluid using ANN-RBF (radial basis function) model. Four different concentrations (1.31, 2.72, 4.25 and 5.92%) of nanofluids were prepared. Within 296–313 K, the ANN-RBF model has been having mean absolute percent error of 0.5177 and 0.5168% to estimate thermal conductivity and viscosity. Their finding showed that viscosity was mainly dependant on Al₂O₃ concentration, whereas thermal conductivity relies on both nanoparticles concentration and temperature. Esfe et al. [82] used ANN model to predict thermal conductivity and viscosity of ferromagnetic–ethylene glycol nanofluid. Total of 72 experimental data were compared with ANN model, and the outcome showed 2 and 2.5% maximum error in the prediction of thermal conductivity and viscosity, respectively.

For hybrid Fe₂O₃/MWCNT nanofluid, Chen et al. [32] varied the concentration of Fe₂O₃ nanoparticles and measured the thermal conductivity. With 0.02 mass% Fe₂O₃ and 0.05 mass% MWCNT, they obtained 27.7% of enhancement in thermal conductivity. However, when they added more Fe₂O₃ (> 0.02 mass%) in the hybrid suspension, thermal conductivity decrement was observed. They proposed that high concentration of nanoparticles would lead to agglomeration easily, in which affecting heat transfer performance significantly.

It can be seen that most of the former researchers reported enhanced thermal conductivity and viscosity when nanofluid was used. When concentration of nanoparticles is increased, both thermophysical properties increase as well. The behavior of nanoparticles in improving base fluid is favorable to elevate current systems from various applications. Although higher concentration of nanoparticles can show more improvement and more efficient than conventional fluid, excessive amount of nanoparticles can still corrupt thermal conductivity which directly affects heat transfer performance [83].

Engine cooling and vehicle radiator system

For few decades till today, vehicle engine system is becoming more advanced due to men’s incessant pursue of higher-performance engines. However, heat generated from engine block system is a huge drawback on overall performance. In a car engine cooling system, coolant is initially pumped into engine block system from radiator via flowing tube. Then, the coolant absorbs heat generated from the engine block which mainly results from friction due to the movement of pistons to turn the crankshaft for rotating vehicle wheels. Engine blocks are usually made up of cast iron or aluminum alloy. After that, coolant flows to radiator when it reaches certain temperature by triggering thermostat. Thermostat is located between engine block system and radiator. It acts as a valve to regulate the coolant flowing to radiator so temperature of the engine block system can be controlled. The mechanism behind it is that the wax at the thermostat melts when the temperature reaches its preset value and a rod connecting to the valve will be pushed away due to thermal expansion of wax and then opening the valve. Before hot coolant enters radiator, the high temperature may cause high pressure in the flowing tube and expansion tank with cap or pressure regulating valve is used to release the excessive pressure. At the last stage, hot coolant flowing inside the radiator will be cooled by surrounding air with the aid of fan behind the radiator (Fig. 6).

Fig. 6

Automobile engine cooling system

Failure of efficient heat transfer could lead to overheating of engine and next damaging engine block body. Thus, maintaining temperature of engine block system is crucial to strengthen its life span and performance. Radiator acts as heat exchanger in vehicle cooling system to transfer heat away from the engine block system to surrounding. In order to attract more users, many improvements have been done by engine companies on radiator system since back then, such as adding fins to increase surface area, changing radiator material and using different configuration of tubes (cross flow, counter flow, parallel flow and shell-and-tube). Nonetheless, there are limitations on these renovations in which few consequences have to be taken into consideration: size of radiator, burden on car, cost of material, durability of material and more.

In the history of development, water was first used as coolant in vehicle cooling system. In some countries with extremely cold weather, water tends to freeze and causes damage to flowing tubes and engine block due to its volume expansion. Then, antifreeze agent was introduced as additive to make up deficiencies of water which has unsatisfying freezing point and low boiling point. Advantage of increasing boiling point is that coolant is allowed to absorb more heat to reach higher operating temperature; thus, more heat can be rejected in a cycle, which mean higher power engine can be implemented. Nowadays, water/ethylene glycol mixture is used as conventional automobile coolant because water and ethylene glycol alone are poor heat transfer fluid. Table 3 and Fig. 7 show the properties of water, ethylene glycol and water/ethylene glycol mixture.

Table 3 Properties of water and ethylene glycol

Fig. 7

Boiling and freezing points result from ratio of ethylene glycol to water

According to comprehensive review from Saidur et al. [84], thermal conductivity is one of the main factors which contribute to the enhancement of heat transfer performance in various applications. Up to authors’ literature review, implantation of nanofluid into vehicle cooling system was initiated by Choi et al. [85] in 2001. They measured thermal conductivity of metal and oxide nanofluids produced themselves. It was found that the measured values were much higher than expected values, and they proposed that nanofluid could enhance vehicle thermal performance. This has led former researchers to start exploring the superior performance of nanofluid as coolant in automobile cooling system.

Experimental studies on nanofluid in automobile radiator cooling system

In order to determine the thermal and flow performance of nanofluids in automobile radiator, many test rigs were set up and studied based on the actual condition of engine radiator system. Non-oxide ceramic material, SiC, was first dispersed into automobile engine coolant by Li et al. [38] to investigate its thermal conductivity. At 50 °C, 0.5 vol% of SiC nanoparticles could enhance thermal conductivity up to 53.81%. For 0.2 vol% concentration, it was found that the overall effectiveness achieved about 1.6, which means that it could act as better engine coolant than conventional water/ethylene glycol.

Selvam et al. [6] investigated the amount of enhancement by dispersing graphene nanoplatelets into water/ethylene glycol mixture in a louvered fin flat tube. Their results revealed that the combination of these parameters, namely nanoparticle concentration (0.5 vol%), nanofluid flow rate (62.5 g s⁻¹) and ambient air velocity (5 m s⁻¹), contributed to about 104 and 81% of enhancement at inlet temperatures of 35 and 45 °C, respectively. In the same year, Selvam and his team [86] seek to deepen the understanding on the performance of the same nanofluid and tube. From the obtained results, they strengthen the statement from their previous work in which mass flow rate of nanocoolant was more dominant than nanoparticle concentration in the increment of pressure drop. For convective heat transfer coefficient, 51 and 20% of improvement were found at 45 and 35 °C, respectively, with nanoparticle concentration of 0.5% and mass flow rate of 100 g s⁻¹.

Islam et al. [7] investigated the effects of nanocoolant used in a 2.4-kW Proton Exchange Membrane Fuel Cell. ZnO was chosen as working nanoparticle due to its better stability and low electrical conductivity compared to Al₂O₃ and TiO₂. From their observation, using 0.5 vol% ZnO nanocoolant enhanced the cooling performance by 29%, reduction in radiator size by almost 27% and increment of less than 10% in pumping power.

Azmi et al. [8] investigated heat transfer performance of water/ethylene glycol nanofluid containing TiO₂ under turbulent flow in a circular tube. Compared to base fluid, the nanofluid showed 28.9% of enhancement at 70 °C when concentration of TiO₂ was increased from 0.5 to 1.5 vol%. The team developed correlations for Nusselt number, and friction factor and the average error were 4.9 and 3.3%, respectively. Then, Azmi and his co-workers [87] compared the convective heat transfer coefficient of Al₂O₃ and TiO₂ dispersed in water/ethylene glycol (60:40) mixture. Three different operating temperatures were considered, and it was found that at 30 °C, heat transfer coefficient of Al₂O₃ has higher value than TiO₂ nanofluids. Meanwhile at 70 °C, Al₂O₃ and TiO₂ nanofluids showed 23.8 and 24.2% of heat transfer enhancement at 1 vol% of concentration compared to base fluid.

Using a flat tube radiator, Alosious et al. [88] conducted experimental and numerical study on the hydrodynamic and heat transfer performance of two water-based nanofluids mixed with Al₂O₃ and CuO nanoparticles. Both prepared nanoparticles having diameter of less than 50 nm, 0.05 vol% concentration, forced to flow within 136 < Re < 186 and fixed inlet temperature of 90 °C. On the other side, Reynolds number range remained the same, but volume concentration was varied from 0.05 to 1% in numerical study. From their experimental outcome, Al₂O₃ and CuO nanofluids contributed 0.5 and 0.38% of enhancement in overall heat transfer coefficient, respectively. Result obtained from the numerical study showed that 1% of volume concentration of CuO and Al₂O₃ nanofluids at Reynolds number equal to 816 led to 13.2 and 16.4% of heat transfer coefficient improvement, respectively. For the same amount of heat released by water, 1 vol% CuO and Al₂O₃ could reduce the area of radiator by 2.1 and 2.9%. Lastly, they suggested that volume concentration of 0.4–0.8% was the optimum value where pumping power could be neglected.

Goudarzi and Jamali [89] tested the effect when both nanofluid and wire coil insert were used in a car cooling system. Different amounts of Al₂O₃ nanoparticles were dispersed in ethylene glycol to produce three different concentrations (0.08, 0.5 and 1%) of nanofluids. The copper wire coil inserts have been having 1.3 cm width and 0.3 mm thickness. Up to 9% of heat transfer augmentation was reported for using wire coil inserts, and when nanofluid was used together with the inserts, the effect boosted for 5% more.

The consequences of implementing CuO–water nanofluid in a four-stroke diesel engine were identified by Senthilraja et al. [90]. For 0.05, 0.1 and 0.2% CuO nanoparticles concentration, specific fuel consumption was reduced by 8.6, 15.1 and 21.1%, followed by emission of NO_x at 881, 853 and 833 ppm, respectively. At the same time, Muruganandam and Kumar [91] tested MWCNT–water nanofluid as coolant used in four-stroke diesel engine as well. As a result, exhaust temperature was decreased by 10% and brake thermal efficiency was increased by 15%.

Effectiveness of MWCNT nanofluid in an air-cooled radiator was determined by Oliveira and his party [92]. The constant parameter in the experiment was air flow rate of 0.175 kg s⁻¹ and four inlet temperatures from 50 to 80 °C. Hot fluid which was to be cooled by air was varied from 30 to 70 g s⁻¹. Viscosity was increased for 54% at 30 °C and 0.16 mass% concentration. However, heat transfer deterioration was observed at 0.16 mass% MWCNT as distilled water presented higher heat transfer rate.

Water-/ethylene glycol-based alumina nanofluid was adopted by Gulhane and his partner [93]. The parameters altered were nanoparticle concentration (0.1–0.4 vol%), flow rate (2–5 L min⁻¹) and inlet temperature (50, 60 and 70 °C). When compared to base fluid, 45.87% of enhancement in heat transfer coefficient was acquired, due to increased nanoparticles concentration. Their explanations were consistent with those made by Sheikhzadeh et al. [94] who worked on the same nanofluid but different parameter values. Laminar flow was considered by setting 9, 11 and 13 L min⁻¹ for volume concentration of 0.003–0.012%. At 13 L min⁻¹, 0.012 vol% alumina nanofluid raised Nusselt number for about 9%. Likewise, empirical correlation for Nusselt number developed was able to do prediction with 3% maximum error.

Some researchers reported heat transfer enhancement of adding TiO₂ in water/ethylene glycol. Chen and Jia [17] varied the nanoparticle concentration from 0.5 to 5 and 10% of improvement was observed. Jagadishwar and Sudhakar [95] prepared the nanofluid in 0.1, 0.2 and 0.35% concentration and evaluated from 6 to 16 L min⁻¹. Their product showed heat transfer inclination of 42.5% with only 0.35% of TiO₂. Using the same base fluid and coolant flow rate, Kumar and Appalanaidu [96] dispersed ZnO and tested with inlet temperature of 50–80 °C. Surprisingly with the similar result, 0.4 vol% of ZnO brought up the enhancement of heat transfer rate to 46% when compared to base fluid.

γ-Al₂O₃/water nanofluid was experimentally studied by Moghaieb et al. [97]. Diameter of nanoparticles ranged from 21 to 37 nm. Four parameters were taken into consideration: nanoparticles concentration (0–2 vol%), coolant flow velocity (1–2 m s⁻¹), heat flux (100–400 kW m⁻²) and bulk temperature (60–80 °C) in experiment. From their inspection, maximum heat transfer coefficient of 78.67% was reported at 1 vol%, 80 °C and 2 m s⁻¹ when compared to water. Convective heat transfer coefficient decreased with increasing coolant temperature and increased with coolant flow velocity.

Heat transfer enhancement of bioglycol that has higher boiling point and lower freezing point was mixed with water and TiO₂ nanoparticles. Abdolbaqi and his squad [98] tested the nanofluid in a flat tube under constant heat flux. Authors had reported augmentation of Nusselt number for about 28.2% when compared to water. Surprisingly, decline of Nusselt number was about 3% at 2.0 vol% TiO₂ and 30 °C operating temperature. Also, friction factor increased by 6.1 and 14.3% at 1.0 and 2.0 vol% nanoparticles.

Sajedi et al. [99] in 2016 suggested that ignoring hydraulic effect might cause miscalculation in heat transfer performance of nanofluid when compared to base fluid. In their experimental setup, pumping power and Reynolds number were fixed for turbulent flow in a finned air-cooled heat exchanger. Three different concentrations (0.5, 1 and 2.5%) were compared at different temperatures. In their analysis, maximum difference of 15% was obtained for 2.5 vol% SiO₂ in water base fluid and 40 °C operating temperature. Based on their result, they concluded that considering constant pumping power as criteria for computing heat transfer performance is appropriate instead of constant Reynolds number.

Experimental studies above are summarized in Table 4, and some other studies are included as well.

Table 4 Summary of experimental studies of nanofluid in automobile cooling system

Experimental studies on various types of automobile radiator

Heat exchangers are broadly utilized in various engineering applications, such as waste heat recovery, air-conditioning system, refrigerator, automobile cooling system, chemical and food industries [107]. Heat exchanger installed in vehicle is usually called as radiator. Generally, there are many types of radiators used in automobile engine cooling system: parallel flow, counter flow, cross flow and shell-and-tube heat exchanger. However, different configurations of heat exchangers are still limited by durability of material, in which restricted high thermal performance and compactness of radiator. Thus, former researchers studied on performance of nanofluids flowing in different types of heat exchangers, and the outcome is tabulated in Table 5.

Table 5 Summary of performance of nanofluid in heat exchangers

Experimental studies using real vehicle engine

One of the earliest experimental studies using actual vehicle components was carried out by Tzeng et al. [124]. His team determined the heat transfer performance of transmission oil with the addition of CuO, Al₂O₃ and antifoam in a four-wheel-drive (Mazda brand) transmission system. The experiment was carried out at four different rotating speeds, starting from 400 to 1600 with 400 rpm increment for each interval. Their result reported that CuO had the best heat transfer performance at both high and low rotating speeds because its distribution of temperature was the lowest one. Two years later, Zhang and his colleagues [125] tested the performance of heavy-duty-diesel engine by adding 3% concentration of nanographite into coolant. It was found that the cooling capability of the nanocoolant was 15% higher than the original coolant itself.

In 2014, radiator from Toyota Yaris 2007 was used to identify the forced convection heat transfer of Al₂O₃ nanofluid by Ali et al. [126]. The nanofluid was prepared with different volume concentrations: 0.1, 0.5, 1.0, 1.5 and 2%. Their finding showed that the optimum heat transfer coefficient was found at 1% volume concentration and heat transfer deterioration occurred when the concentration was further increased. 14.72 and 9.51% of maximum increment were found for Nusselt number and heat transfer coefficient of the coolant, respectively.

M’hamed et al. [127] investigated heat transfer performance of a Proton Kelisa 1000 cc engine system. MWCNT was mixed with water/ethylene glycol base fluid and used as coolant in their study. From their results, it was found that the nanocoolant with 0.50% of volume concentration yielded about 196% of maximum heat transfer coefficient enhancement in laminar flow condition. In 2006, Devireddy et al. [128] used a car radiator which was commercially available to demonstrate the performance of TiO₂ at different concentrations. They obtained heat transfer improvement of 35% at 0.5 vol% TiO₂ and proposed that Brownian motion might be the main contributor to the enhancement instead of thermophysical properties.

Not only four-wheel car engine system has been studied till now, Mathivanan and his team [129] used Aprilia SXV 450 (motorcycle) engine in their experimental setup. Various nanoparticles were mixed separately with distilled water and compared to each other. The nanofluids prepared included 1–100 nm of MWCNT, Al₂O₃, SiC and TiC nanoparticles. At 1% of nanoparticle concentration, TiO₂ nanofluid dissipated heat the most among all tested nanofluids and showed 31.9% of better heat dissipation capacity than water at 3.5 GPM of flow rate.

Numerical studies on nanofluid in automobile radiator cooling system

Sahoo et al. [130] analyzed the capability of brine-based nanocoolant in wavy finned radiator. They compared the heat transfer performance between two different types of nanoparticles (Ag and Al₂O₃) mixed with propylene glycol and ethylene glycol. It was found that both Ag and Al₂O₃ nanofluids with propylene glycol performed better than Ag and Al₂O₃ nanofluids with ethylene glycol. Apart from that, size and pumping power required by radiator could be reduced up to about 4 and 25.5% when Ag–propylene glycol nanofluid was used instead of base coolant without nanoparticles.

Another numerical study using Fluent software was carried out by Vajjha et al. [131] on flat tubes of radiator. Al₂O₃ and CuO nanofluids with water/ethylene glycol were compared. 10 vol% Al₂O₃ nanofluid increased average heat transfer coefficient for 94% and expected to decrease required pumping power by 82% when compared to base fluid. On the other side, 6 vol% CuO contributed 89% enhancement in heat transfer coefficient and could reduce pumping power up to 77% on the basis of same amount of heat transfer from base fluid. Four years later, Vajjha et al. [132] studied on the thermal performance of same nanofluids in radiator flat tubes with similar dimensions. Both Al₂O₃ and CuO nanofluids with 3 vol% concentration increased the average heat transfer coefficient for 36.6 and 49.7% at Reynolds number equal to 5500. As a result, Al₂O₃ nanofluid allowed more reduction in terms of pumping power compared to CuO nanofluid.

Few years later, researchers from China and Iran [133] examined on four different nanomaterials mixed with water/ethylene glycol. Their result showed good agreement with correlation from past researchers [131]. A 3D vehicle radiator flattened tube was modeled using ANSYS and analyzed using Fluent. Laminar flow (500–2000 Reynolds number) and different shapes (cylindrical, spherical, platelet and brick) of nanoparticles were tested. It was noticed that CuO and TiO₂ could transfer heat better than Fe₃O₄ and Al₂O₃. Besides that, nanoparticles shape with minimum and maximum Nusselt numbers can be arranged as spherical, brick, cylindrical and platelet accordingly.

Five different hybrid nanofluids were produced by Sahoo and Sarkar [134] using Al₂O₃ and five different nanoparticles in an experiment. 1 vol% nanoparticle water/ethylene glycol nanofluids were tested in a louvered fin radiator. The authors found that 0.5% Ag mixed with 0.5% Al₂O₃ gave the highest value in heat transfer rate, effectiveness and pumping power. Meanwhile, highest performance index was obtained from the combination of SiC and Al₂O₃. They computed that reduction in radiator size could reach 3.7% for constant coolant flow rate and heat transfer rate, whereas coolant flow rate can be decreased by 3.1% for fixed heat transfer rate and radiator size, when Ag hybrid nanofluid was compared to base fluid.

Leong and his team [135] examined heat transfer behavior of Cu–ethylene glycol nanofluid as coolant in flat tubes of a diesel engine (TBD 232V-12). From their research, variation of Reynolds number for air has more obvious effect on thermal performance of the radiator than nanocoolant. When the Reynolds number of nanocoolant and air was set to 5000 and 6000, respectively, 2 vol% of copper was sufficient to increase heat transfer by 3.8% compared to base fluid. In addition, frontal area of radiator was estimated to have 18.7% of contraction.

Comparison between CuO, Al₂O₃ and TiO₂ water-based nanofluids was made by few researchers. Togun et al. [136] looked up turbulent heat transfer of these three nanofluids in an annular concentric pipe. Firstly, they prepared Al₂O₃ nanofluid experimentally and validated numerical results from experimental results. k − ε turbulence model was considered in ANSYS Fluent software. At expansion ratio of 2, heat transfer augmentation and pressure drop for TiO₂, CuO and Al₂O₃ were 45.2, 47.3, 49 and 62.6, 65.4, 57.6%, respectively. Khan and his partner [137] tested all nanofluids; CuO nanofluid exhibited highest heat transfer rate. It was noticed that heat transfer rate was increased with increasing concentration, which is in line with the finding of Ahmad et al. [138] who compared Cu, Al₂O₃ and SiO₂ water-based nanofluids. In this study, the boundary conditions set were consistent heat flux (18,000 W cm⁻²) and laminar flow (Re = 100–1000).

Hussein et al. [139] considered inlet temperature of 60–90 °C and Reynolds number of 10,000–100,000 in a 500-mm radiator flat tube. The maximal value for Nusselt number and friction factor increment was 18 and 12%, respectively, for 4 vol% TiO₂ nanoparticle in water-based nanofluid. Performance of nanocoolant in class 8 truck engine was determined by Saripella et al. [140]. The coolant was made up of CuO and water/ethylene glycol. They observed that low engine and coolant temperature gave high heat transfer coefficient.

Using ANSYS Fluent software, Fsadni et al. [141] investigated thermal and flow performance of Al₂O₃–water nanofluid in a helically coiled tube heat exchanger with curvature of 0.032–0.052. Single-phase homogeneous model was used to compute the turbulent flow condition with constant heat flux and Reynolds number of 10,000–60,000. Heat transfer performance and pressure drop were increased by 7–34 and 11–63%, respectively, at concentration of 1–4 vol%. Furthermore, increasing curvature ratio could boost extra 2.5 and 4.7% in heat transfer performance and pressure drop.

Other applications

Not only cooling system in vehicle is concerned nowadays, various applications such as solar collector, processors for electronic devices and thermal energy storage unit are dependent on excellent heat transfer fluid to achieve high efficiency with better thermal and flow properties. Until today, the demand of nanofluid in different fields is kept on increasing. A number of papers related to nanofluid in Scopus are shown in Fig. 8. It is clearly shown that the demand on using nanofluid in engineering field is leading among other fields, and thus, it is important to further explore the subtle mechanism of nanofluid enhancing performance in the approach.

Fig. 8

Number of published papers related to ‘nanofluid’ in Scopus till January 2018

Solar thermal collector

Solar thermal collector is used to capture thermal energy emitted from solar radiation, and it is usually called solar energy. Solar energy is one of the most commonly used renewable resources to promote greener environment, and thus, many modifications were done on geometrical part of absorbers in solar collectors to increase thermal efficiency. Anyhow, there is limitation on the optimization method due to significant increment of pressure drop [142]. Thus, excellent heat transfer fluid is vital to ensure further increment of thermal efficiency of solar collectors, whereby nanofluid is used instead of conventional working fluid–water. For instance, increasing concentration of CuO–water nanofluid was reported to give positive impact on thermodynamic efficiency and energy efficiency in a solar-driven combined cooling, heating and power (CCHP) system [143].

Specific heat capacity of salt-based nanofluids was examined by Hu and his co-workers [144]. Firstly, Al₂O₃ nanoparticles with diameter of 20 nm were mixed with water; then, NaNO₃ and KNO₃ which are solar salts are added into the suspension. With 2.0% of nanoparticle concentration, the specific heat capacity was enhanced up to 8.3%. They also found that Coulombic energy was the main contributor of the increment of specific heat capacity.

Jin and Jing [145] proposed a novel liquid optical filter for hybrid photovoltaic/thermal (PVT) application. They prepared magnetic electrolyte nanofluids (ENFs) which contained magnetic Fe₃O₄ and water/ethylene glycol. Then, methylene blue and copper sulfate were added separately to produce two different nanofluids, which are denoted as ENF-1 and ENF-2, respectively, in Fig. 9. Their results revealed that both nanofluids have better thermal conductivity than base fluid in tested temperature of 20–60 °C and performed better than conventional core/shell nanoparticle nanofluid filters.

Fig. 9

Thermal conductivities of ENFs and base fluid

To determine thermal performance, thermal conductivity and viscosity of Ag–water nanofluid, Koca et al. [146] carried out experiment on a single-phase natural convection mini loop. The nanofluid contained 5 mass%, 15 nm and spherical-shaped Ag nanoparticles and 1.25 mass% polyvinylpyrrolidone (PVP) surfactant. In their analysis, they found that the effectiveness of mini loop was enhanced to 11% with 1 mass% of Ag nanoparticle. Besides that, their results obtained were consistent with their previous work which stated that PVP was the barrier on heat transfer performance at ambient temperature. At 23 °C, thermal conductivity of nanofluid decreased for 11.5% when the Ag concentration was 1 mass%, whereas viscosity was increased by 4.81% at 20 °C with the same concentration.

In another experimental study conducted by Tahat and Benim [147], they investigated thermophysical and rheological properties of hybrid nanofluid which contained Al₂O₃ and CuO nanoparticles in flat plate solar collector. Water and ethylene glycol were served as the base fluid in ratio of 25:75 by mass. The volume concentration of nanoparticles was varied from 0.5 to 2, and 45% of enhancement was observed on the efficiency of solar collector.

An experimental study by Manikandan and Rajan [148] involved determination of viscosity and thermal conductivity of sand–propylene glycol–water nanofluid in solar energy collection. 16.3% of thermal conductivity increment and 47% of viscosity decrement were obtained through testing 2 vol% of the nanofluid at 28 °C. They pointed out that the rising of thermal conductivity was caused by Brownian motion, which is in line with the finding from Devireddy et al. [128]. Accordingly, efficiency of solar energy collection was improved by 16.5 and 9% when 2 and 0.5 vol% of the nanofluid were used.

Saidur et al. [149] found out that nanofluid is able to provide superior optical properties and better heat transfer as volumetric absorber in direct solar collector. From their simulation results, alumina–water nanofluid improved the absorption capability at shorter and visible wavelength area when compared to water. When the amount of Al₂O₃ was raised to 1.0 vol%, the nanofluid was nearly non-transparent to light wave. They also proposed that concentration of nanoparticles has a linear proportional relationship with extinction coefficient.

Bellos and Tzivanidis [150] numerically investigated six different oil-based nanofluids in parabolic trough collectors (PTC). In their analysis, concentration of nanofluids (up to 6%), flow rate (50–300 L min⁻¹), inlet temperature (300–650 K) and solar irradiation level were studied. Their results revealed that nanofluid with Cu nanoparticles exhibits the most thermal efficiency enhancement, while SiO₂ the lowest. Maximum thermal efficiency enhancement of 2.2% was found at 6% Cu concentration, 600 K inlet temperature and flow rate of 50 L min⁻¹. In addition, thermal efficiency enhancement did not show a significant increment when nanoparticles concentration exceed 4%. A new evaluation index that includes heat transfer coefficient, density-specific heat capacity and flow rate was found able to determine thermal efficiency enhancement of PTC.

Another numerical study by Bellos et al. [151] was about two different methods to enhance thermal efficiency of linear Fresnel reflector. The use of finned absorber and CuO–thermal oil nanofluid with three different concentrations (2, 4, 6%) was compared under different inlet temperatures and flow rates. It was found that combination of these two approaches showed the highest thermal efficiency enhancement (0.82%), while adding 4% concentration nanofluid and using finned absorber improved thermal efficiency by 0.28 and 0.61%, respectively. Besides that, peripheral receiver temperature variation was reduced using both methods, and this could lead to slower deformation of the receiver. Although pumping work was increased using these enhancement methods, global performance of the collector was still favorable.

Heat sink in electronics cooling system

In electronics field, passive cooling method is commonly used due to the absence of external parts and low cost. There are some passive techniques which have been employed for heat transfer improvement, such as applying corrugated surfaces, rough surfaces, installing flow swirling tools and using porous materials [152]. However, passive method itself is not sufficient for cooling purpose due to rapid growing of complex systems in electronic devices. From Sidik’s review [153] on passive cooling technique for microchannel heat sink, he suggested the implantation of nanofluid may further facilitate cooling performance.

Jang and Choi [154] introduced the combination between nanocoolant and microchannel heat sink in order to provide high cooling performance using active cooling method. They investigated 6-nm Cu nanoparticles and 2-nm diamond nanoparticles, both with 1 vol% concentration dispersed in water under temperature difference of 80 °C between ambient temperature and junction temperature. Cooling performance of the microchannel heat sink was enhanced by 4 and 10% by Cu nanofluid and diamond nanofluid, respectively, due to lower thermal resistance of these nanofluids compared to water as shown in Fig. 10.

Fig. 10

Numerical result obtained at pumping power up to 2.50 kW

Forced convective heat transfer of CuO/water nanofluid in microchannel heat sink was studied by Chabi and his co-workers [155]. At channel entrance region with Reynolds number of 1150, the average heat transfer coefficient of 0.2 vol% nanofluid was 40% higher compared to deionized water. However, they noticed heat transfer deterioration when Reynolds number was further increased. Recent work from Sun and Liu [156] showed the heat transfer coefficient of Al₂O₃ nanofluid and CuO nanofluid in a liquid-cooled central processing unit (CPU) radiator. 0.1–0.5 mass% nanoparticles and Reynolds number of 400–2000 were varied. Convective heat transfer coefficient of Al₂O₃ nanofluid and CuO nanofluid was about 1.1–1.6 times and 1.1–2 times higher than that of deionized water.

A major study on heat transfer performance of ERG aluminum foam heat sink in a computer processor (Intel core i7) was carried out experimentally and numerically by Bayomy and Saghir [157]. They produced γ-Al₂O₃/water nanofluid with volume concentration of 0.1–0.6% and tested in laminar flow (Reynolds number of 210–631) and heat flux of 8.5–13.8 W cm⁻². Their results showed that 0.2 vol% nanofluid yielded the highest local Nusselt number. For average Nusselt number, 37 and 28% of enhancement for 0.2 vol% nanofluid over pure water were seen at Reynolds number of 601.3 and 201, respectively. In addition to that, the numerical results obtained showed low discrepancies up to 3 and 2% for local Nusselt number and local temperature, respectively.

Arjun and Rakesh [158] determined the thermal and flow performance of 23 nm Al₂O₃–water nanofluid in circular microchannel numerically. Reynolds number of 5–11,980 and 0–5 vol% of Al₂O₃ generated heat transfer enhancement of 83%. Convective heat transfer of Al₂O₃ in square microchannel under laminar flow was studied by Irwansyah et al. [159] experimentally. 0.6 and 1.0 vol% water-based alumina nanofluids showed enhancement of 6.9 and 21%, respectively. Effect of three different microchannel shapes was numerically investigated by Toghraie et al. [160]. Their results revealed that sinusoidal microchannel without nanofluid showed higher heat transfer rate than smooth microchannel with nanofluid. Among all shapes, zigzag microchannel is suggested as the best microchannel.

Flow boiling

Boiling is a technique using latent heat transport to increase heat transfer performance in industrial applications such as power plants, electronics cooling system, heat pipes and nuclear reactors [161]. Flow boiling is one of the common mechanisms used in thermal engineering applications. Due to the growing demand from those huge applications, nanofluid has been used to replace or enhance the properties of conventional heat transfer fluid.

Zangeneh et al. [162] reported the effects of nanoparticles synthesis method and subcooled flow boiling on heat transfer performance in a vertical annulus. They observed that ZnO which modified by using amine and UV irradiation gave the highest heat transfer performance (8.14%) compared to water. In addition to that, shape of nanoparticles plays a vital role in heat transfer coefficient as nanotube–nanorod shape performed better than spherical shape.

Refrigerant (R113) mixed with CuO was tested in a smooth copper tube with 150 cm length, 0.07 cm thickness and 0.952 cm outer diameter. 29.7% of maximum intensification in heat transfer coefficient was observed when concentration of CuO was varied in the range of 0–0.5 mass%. The horizontal tube with flow boiling using heating tapes set up by Peng and his team [163] is shown below (Fig. 11).

Fig. 11

Test section where flow boiling takes place [163] (License Number: 4347970093882)

Researchers from Korea [164] investigated flow boiling experiment using 0.01 vol% alumina–water nanofluid in a horizontal rectangular channel. A disk-shaped copper surface was placed below the rectangular channel to perform flow boiling. At 1 and 4 m s⁻¹ nanofluid flow velocity, critical heat flux (CHF) was increased by 24 and 40%, respectively, when compared to water. The increment in CHF was explained in such way that deposition of nanoparticles caused changes in surface wettability. Same conclusion was obtained by Vafaei and Wen [165] who inspected the effect of subcooled flow boiling on critical heat flux in a horizontal 510-μm microchannel. Under mass flow rate of 600–1650 kg m⁻³, 0.1 vol% alumina-deionized water nanofluid remarkably increased CHF for 51%. Another study showed that 0.005 vol% MWCNT nanofluid enhanced heat transfer coefficient of pure water by 4.3% at CHF [166].

Investigation into the flow boiling heat transfer behavior of evaporator vertical surface in a thermosyphon loop was carried out by Yang and Liu [167]. Average 50 nm CuO nanoparticles were suspended into water, and different mass concentrations of nanofluid were prepared (0.1–1.5 mass%). Maximum value of heat transfer coefficient was found at optimal 1.0 mass% CuO, whereas 29% of increment in CHF was observed at 1.5 mass% CuO.

Thermal energy storage system with phase change material

Thermal energy storage system (TESS) functions to store and release energy in the form of latent heat and sensible heat for consequent uses to conserve waste heat from surrounding. Phase change material is commonly used in thermal energy storage system (TESS) due to their excellent capability to store and release energy during density changes. Thus, the criteria for enhanced heat transfer performance in TESS are mainly melting duration, melting temperature and latent heat of fusion of phase change material (PCM).

Behavior of alumina–water nanofluid in TTES was initially studied by Wu et al. [168] experimentally. They found out that 0.2 mass% alumina in water reduced supercooling temperature by 70.9% and total freezing time by 20.5%. In addition, thermal conductivity was enhanced by 10.5%. Years later, Wu and his team [169] made comparison between behavior of Cu, Al and C/Cu nanoparticles when each of them was dispersed into melting paraffin. Best transfer performance was observed when 0.5 mass% Cu nanoparticles were used. Also, latent heat for freezing and melting was reduced by 11.7 and 11.1%, respectively. Figure 12 shows 1 mass% Cu/paraffin PCM with excellent thermal reliability even after 100 cycles of cooling and heating.

Fig. 12

Effect of thermal cycles on a phase change temperature and b latent heat of the PCM

Behavior of nanofluids in an unit of upright shell-and-tube TESS was explored by Duan and his partner [170]. They established a computational fluids dynamics model to carry out their numerical study. As novelty in their study, CuO–water nanofluid acted as heat transfer fluid (HTF) and CuO–paraffin nanoparticle-enhanced phase change material (NePCM) was used on shell side of the energy storage unit. Nanoparticle size considered was 10 nm in diameter, and concentration range for HTF and NePCM was 0–7 and 0–4 vol%, respectively. From their analysis, inclusion of nanoparticles not only improved heat transfer coefficient but accelerated melting process of PCM. However, exceeding amount of nanoparticles in PCM could lead to heat transfer and melting rate degradation as viscosity would be increased gradually. Using response surface methodology (RSM), it was found that inlet temperature of HTF was the most compelling parameter compared to concentration of HFT and NePCM.

Sebti et al. [171] carried out numerical investigation on the effect of CuO–NePCM on heat transfer performance in a horizontal annulus with concentric cylindrical shape. 0–0.05 vol% CuO and 5–20 °C temperature difference were their manipulating parameters. Their results showed that higher concentration of CuO reduced solidification time and heat transfer rate was increased.

Tasnim and his team [172] demonstrated scale analysis and numerical study on phase change process of NePCM in a porous-latent HTES. Initially, the extent of entire phase change process was appraised using scale analysis. The analysis provided relationship between various parameters (Rayleigh number, Stefan number, Nusselt number, Fourier number, nanoparticle concentration and porosity of the porous medium). Then, Darcy model was used to solve the melting phenomenon of NePCM in a rectangular enclosure with porous medium under natural convection condition. Exceptionally, both scale analysis and numerical study revealed that the existence of NePCM corrupted the convection and conduction heat transfer performance in the enclosure. Moreover, the melting time of NePCM was longer than PCM.

Conclusions

This paper presents recent review on the effects of implantation of nanofluid in vehicle engine cooling system and other major applications. Based on studies, it is found that nanoparticles can be used to enhance thermophysical properties of working fluid, especially thermal conductivity. Improved thermal conductivity of nanofluid is due to the higher thermal properties of dispersed solid nanoparticles, and this leads to:

• 1. Increased cooling performance/overall system efficiency. Higher thermal conductivity means better capability of a substance to absorb and release heat efficiently. This results in more heat which can be dissipated away from a system.

• 2. Reduced system size. When a more efficient working fluid is used, increment of surface area of heat exchanger or pipe by adding fins or modifying geometric is not needed. This can also help to reduce the drag on car which reduce fuel consumption.

• 3. Reduced storage of heat transfer fluid. Conventional coolants need larger storage volume so that the cooling system does not get overheated easily when heated coolant keep flowing back into storage tank. Nanofluids with better cooling rate require less volume than conventional fluid.

• 4. Strengthen system lifespan. Overheating would result in mechanical failure. Thus, nanofluids with better cooling performance can help to protect body of heat exchanger and cooling system.

• 5. Reduced pumping power when compared to same amount of heat transfer from conventional fluid. When nanofluids are used, both heat transfer rate and pressure drop will increase. As reported by past researchers, heat transfer rate increment is much more significant when compared to pressure drop. Thus, when amount of heat transfer is the same for both conventional coolant and nanofluid, nanofluid is believed to show less pressure drop.

For all of its aforementioned advantages, nanofluid in fact is something of a double-edged sword. Although increasing concentration of nanoparticles can greatly enhance heat transfer performance of cooling system, excessive amount of nanoparticles will lead to high viscosity of nanofluids which contributes to increment of pressure drop and may be followed by deterioration of overall efficiency of a particular system due to clogging and agglomeration of nanoparticles. However, this can be overcome by increasing working temperature which reduces viscosity of nanofluids and in the same time increases thermal conductivity. It has to be noted that extremely high working temperature can lead to mechanical system breakdown.

In short, nanofluids can be a promising working fluid for various cooling systems due to its superior heat transfer performance. Overall, effective and efficiency of a system can be improved when conventional working fluid is replaced by nanofluids.

Current challenges and recommendation

From authors’ review, most of the results are in positive favor in which nanoparticles enhance cooling capability or heat transfer performance of conventional heat transfer fluid in various heat transfer applications. For cases where heat transfer deterioration occurred, few researchers reported that excessive amount of nanoparticles is the factor where agglomeration happens. Among them, very few mentioned on the optimum amount of nanoparticles before deterioration of performance or thermal properties would occur. Consequently, repeating experimental work in the future on the same nanoparticles and similar working parameters is needed to obtain the optimum concentration. Knowing the optimum amount of nanoparticles can obtain the best thermal performance for a system.

Secondly, there are many past researchers that could not reach consensus on the exact factors which affect heat transfer behavior of nanofluid. This may due to different variables used in their respective experiment or numerical study such as working temperature and system pressure. In addition, the challenging part is to produce desirable nanofluids and compare to each other. Some researchers compared the heat transfer enhancement of different nanoparticles by omitting these few aspects:

• 1. Huge difference of mean diameter size. Few past studies showed that smaller size of nanoparticles has better thermal performance due to increased surface area over volume ratio. Anyhow, many researchers are comparing different nanofluids with different nanoparticles sizes. It is inaccurate to judge performance of particular nanoparticles under such circumstances.

• 2. Shape of nanoparticles. From review, thermal performance results from different nanoparticles shapes are quite significant as shown by Hatami et al. [133]. Most researchers carried out the comparison on the performance between different nanofluids with different nanoparticles shapes, which caused lack of agreement in their results.

• 3. Degree of stability. Many past researchers used two-step method to prepare nanofluids. Some of them used sedimentation method to observe the presence of sediments of nanofluids after certain period. Then, nanofluids are considered stable when no sediments are observed. The actual degree of stability is unknown. There is a lack of measurement standard on the stability of nanofluids. Although there are many other inspection methods such as spectral absorbance analysis, zeta potential measurement and pH value adjustment, these methods are not chosen due to either high cost or more steps. Stability of nanofluids should be evaluated using same method before any comparison work is performed.

For future work, main factors affecting thermal conductivity need to be identified. It is important to understand the mechanism which thermal conductivity is greatly affected. The three aspects mentioned above are believed to contribute the most to the magnitude of thermal conductivity of nanofluids. In order to determine impact of each parameter, it is vital to control only one of the variables in experiment. By this way, the main parameter which gives the highest impact to thermal conductivity can be determined more accurately.

In addition, heat transfer performance as a function of parameters such as temperature, pressure and flow rate can be developed as correlation, and impact of each parameter can be determined. To obtain results which are more accurate and reliable, more experimental works have to be carried out systematically. Also, a generalized equation for each type of nanoparticles can be developed. More experimental works using the same preparation method, stability evaluation analysis and physical properties of nanoparticles are needed so that more accurate data are available to develop such equation. This can help future researchers to study behavior of nanoparticles more easily and in the same time provides industry reliable data to commercialize nanofluids into more applications.

Major problems with the inclusion of nanofluids into daily applications are the stability and price of nanofluids. Therefore, more works are needed to improve these drawbacks to implement nanotechnology in this century. In short, current investigation works on nanofluid need to be improved from several aspects and at the same time improve current theoretical model so that feasible use of nanofluids in future researches and industries can be complied.