Elsevier

Powder Technology

Volume 342, 15 January 2019, Pages 85-98
Powder Technology

A decision-making based method to optimize energy efficiency of ecofriendly nanofluid flow inside a new heat sink enhanced with flow distributor

https://doi.org/10.1016/j.powtec.2018.10.007Get rights and content

Highlights

  • Energy efficiency of a heat sink operated with a biological nanofluid is optimized.

  • A three-objective optimization and a decision-making based method are implemented.

  • Objective functions are surface temperature, pumping power and entropy generation.

  • Effect of concentration on pumping power is lower than its effect on other outputs.

  • Temperature and entropy generation are affected by Re number rather similarly.

Abstract

In this contribution, the hydrothermal and irreversibility attributes of a biological nanofluid in a new heat sink enhanced with flow distributor are investigated. By increasing either concentration or velocity, the surface temperature reduces, and the cooling uniformity improves. The velocity has a more significant effect on the temperature in comparison with the effect of concentration. Increasing the velocity and concentration decreases the thermal entropy generation. Moreover, frictional entropy generation intensifies with the increment of concentration and velocity. Besides, the friction has a minor contribution in overall irreversibility of the heat sink compared with heat transfer. In the heat sink, a lower fraction of total entropy is generated in the solid walls, and chief fraction occurs in the fluid. Finally, a decision-making based approach is used for optimization in addition to the three-objective optimization. The optimization is carried out on the models obtained from the neural network to reach minimum values for the surface temperature, pumping power and irreversibility. The optimal cases are reported considering different priorities of the objective functions. The results of optimization reveal that the nanofluid employed shows the promising views to be applied in electronics cooling based on both first and second laws of thermodynamics.

Introduction

The heat transfer attributes are limited by the fluids that are employed in thermal devices. The concept of employing solid nanoparticles as additive in ordinary fluids has been mentioned in the recent years for the purpose of heat transfer enhancement. This concept results in synthesis of the suspensions with excellent thermal characteristics termed nanofluids [[1], [2], [3], [4], [5], [6], [7], [8]]. Many scholars have evaluated the performance of nanofluids in different thermal devices [[9], [10], [11], [12]]. Moreover, their thermophysical properties including viscosity, heat capacity, latent heat, thermal conductivity, surface tension and critical heat flux have been examined extensively. Murshed et al. [13] measured the thermal conductivity of a TiO2–water nanofluid, and reported a great increment of the thermal conductivity. Amiri et al. [14] prepared a graphene nanoplatelets nanofluid, and indicated greater thermophysical features in comparison with the pure water. Shi et al. [15], in their study on direct vapor generation through localized solar heating via carbon nanotube nanofluid, showed that this nanofluid has a broadband absorption in visible light.

In addition, due to the excellent characteristics of nanofluids, many scholars have studied their heat transfer efficiency in natural convection, forced convection as well as pool boiling systems. Mirzaei et al. [16] numerically evaluated the laminar flow and heat transfer of a water–Al2O3 nanofluid. Jamal-Abad et al. [17] experimentally examined the convection heat transfer of Cu–water and Al–water nanofluids under the laminar flow regime with constant thermal boundary condition. Moreover, some investigators have studied the advantages of using nanofluids in heat sinks for electronics cooling. Sarafraz et al. [18] researched the thermal performance of a heat sink working with gallium, water, and CuO–water nanofluid. The CuO–water nanofluid demonstrated a higher thermal efficiency than the water, while had lower pressure drop and pumping power compared with the gallium. Al-Rashed et al. [19] examined the influence of nanofluids on the performance of a micro heat sink for cooling of an electronic processor. For the range of mass flow rates and heat loads under investigation, the heat transfer enhancement was obtained up to 7.7% in the case of applying nanofluids rather than water. Khaleduzzaman et al. [20] evaluated exergy and entropy generation of TiO2–water nanofluid for cooling of an electronic device. It was found that the surface temperature reduces by increasing the flow rate, and increases with addition of the nanoparticles. Moreover, with the flow rate increment, exergy loss and thermal entropy generation rate decreased, whereas frictional entropy generation rate increased.

In the investigations performed in the area of nanofluids, various nanomaterial types have been utilized as nanoparticles. Owing to the excellent attributes of silver nanoparticles, they are extensively employed in many applications such as thermal systems, medical devices, and clothing. Typically, the method for the synthetizing Ag nanoparticles includes the reduction of silver ions in a solution or in high-temperature gaseous environments [21]. However, it is noteworthy that the reduction of reagents, such as sodium borohydride, can be dangerous and intensify the ecological toxicity [22]. Besides, the production methods in high temperatures cause significant costs. As a result, the development of green techniques for synthesis of Ag nanoparticles by utilizing ecofriendly solvents and nontoxic reagents is very essential. Sun et al. [23] presented a simple, ecofriendly, and cost-effective approach to produce silver nanoparticles by employing tea leaf extract. In this work, the Ag nanoparticles were synthesized using tea extract and silver nitrate, and the reaction was carried out at room temperature for 2 h. Sarafraz and Hormozi [24] used this approach for preparing the nanofluid containing silver nanoparticles. The authors focused on thermal characteristics of the nanofluid flowing in a heat exchanger. A significant enhancement of heat transfer coefficient was obtained up to 67% at concentration of 1%.

Improving the flow uniformity plays a significant role in several practical applications. Uniformity of flow distribution is also one of the most important problems in designing the configuration of heat sinks for electronics cooling. In parallel multi-channel heat sinks, the fluid should be flown evenly to improve the cooling process. This results in a uniform temperature distribution on surface of electronic processors, which is vital for appropriate performance of such devices [[25], [26], [27]].

Manifolds which employ 3D flow distribution systems cause better uniformity, and are essential in accurate applications such as electronics cooling, fuel cells, nuclear reactors, and so forth. Several efforts have been made for designing different distributors to reach the purpose of flow uniformity. Bejan et al. [28] developed the cascade structure to obtain uniform distributions. They indicated that the flow sub-streams from tree-shaped bifurcations of channels are uniform, and carried out analysis and optimization to the dimensions of flow channels forming cascade flow bifurcations. Fan et al. [29] simulated a plate-type distributor based on the theory introduced by Bejan [28], and reached a proper flow distribution. The authors also found that as the distributor has a bifurcated structure, non-uniformity exists owing to the inlet effect and the asymmetry of the streamlines. Li et al. [30] developed the cascade-structure to three innovative designs named 90° tee-type, rounded-type, and slanted-type. Liu et al. [31] assessed these models numerically and experimentally, and reported optimal models. Application of this approach, however, wouldn't always solve the problem of maldistribution particularly at great velocities, thanks to the influence of inertial forces. A minor asymmetry in velocity profile at one bifurcation, because of manufacturing tolerances, would cause noticeable maldistribution in tree-shaped distributors [32,33]. Additionally, these manifolds may need a rather great volume to prevent the adverse influence of flow singularities such as direction variations in manifolds. Leela Vinodhan and Rajan [34] examined flow and heat transfer in four microchannel heat sink outlines. The microchannel heat sinks included four compartments with separate coolant inlet and outlet ducts for each part. The existence of several regions of developing flow in some of these designs led to greater Nusselt number and heat transfer rates. Ramos-Alvarado et al. [35] evaluated the thermal efficiency of liquid-cooled heat sinks with ordinary and new flow field configurations. The flow distribution uniformity in multiple flow channels, temperature uniformity on heating surfaces, and pumping power of heat sinks for the cases of new flow field outlines and ordinary flow field configurations were compared. It was concluded that the new proposed configurations have significant advantages for use in heat sinks.

In the current study, optimization of energy efficiency for a new heat sink in cooling of an electronic processor using a biologically produced nanofluid is carried out. The purpose is to reach optimal conditions in which the surface temperature, pumping power and entropy generation are minimum. Firstly, the flow and heat transfer attributes are evaluated and then, a multi-objective optimization method and a decision-making based approach (i.e. GA combined with compromise programming) are implemented to find optimal points. To the best knowledge of the authors, this research is the first study that optimizes performance of a heat sink operated with a biological nanofluid in order to reach best cooling process together with minimum energy consumption as well as minimum irreversibility.

Section snippets

Definition of the heat sink and nanofluid

The heat sink under investigation is illustrated in Fig. 1. It is a novel heat sink having T-shaped distributors. The parallel flow channels and flow distributors are evidently perceived in Fig. 1. The material of the liquid block is aluminum. In this outline, the flow is distributed between the channels in four levels (i.e. four steps) before reaching main parallel channels. This leads to a uniform flow distribution between the channels, which can decrease pressure drop and improve temperature

Governing equations

The mass, momentum and energy equations are numerically solved by employing variable effective properties for studying the hydrothermal features and irreversibility attributes of the biological nanofluid flow in the heat sink. Note that the nanofluid is considered incompressible and Newtonian.

Conservation of mass:.ρv=0

Conservation of momentum:ρvv=P+μv

Conservation of energy:ρvcpT=kTwhere ρ is density, k indicates thermal conductivity, μ represents viscosity, and cp is specific heat.

Numerical method and validation

For the numerical simulations, the finite volume method is used. The SIMPLE approach is utilized to couple the pressure and velocity, while second-order upwind scheme is employed for solution of the mass, momentum, and energy equations. The convergence condition is considered 10−5 for all variables by considering the normalized residuals. In order to ensure the mesh independency, the various grids with different cell numbers were assessed. The pressure drop (ΔP) and temperature increment (ΔT)

Entropy generation

Two factors, namely heat transfer and friction, lead to entropy generation and therefore, total entropy generation is calculated by the following equation [36,37]:Ṡg,t=Ṡg,h+Ṡg,fwhere Ṡg,t, Ṡg,h and Ṡg,f indicate respectively total entropy generation rate, thermal entropy generation rate and frictional entropy generation rate.Ṡg,h=kT2Tx2+Ty2+Tz2Ṡg,f=μT{2vxx2+vyy2+vzz2+vxy+vyx2+vxz+vzx2+vyz+vzy2}

The global entropy generation rates (i.e. entropy generated in

Neural network modeling and optimization approach

Multilayer perceptron neural network is employed for developing mathematical models of the surface temperature, pumping power and entropy generation in the heat sink in terms of volume concentration and Reynolds number according to Fig. 4. The backpropagation algorithm is used for training the neural network with the aid of the patterns achieved from the numerical simulations. In order to assess the neural network performance, maximum absolute error, Mean Squared Error (MSE), and coefficient of

Results and discussion

In this research, the hydrothermal and entropy generation characteristics of the biological nanofluid flow in the heat sink are firstly investigated and then, optimization is carried out on the obtained models of these parameters in order to reach minimum values for the surface temperature, pumping power and irreversibility. The simulations are performed at velocities of 0.2, 0.3, 0.35, 0.4 and 0.5 m/s, and concentrations of 0, 0.2, 0.4, 0.6, 0.8 and 1%.

Fig. 5 illustrates the pathlines for flow

Conclusions

In this study, the flow, heat transfer and entropy generation characteristics of the ecofriendly nanofluid inside the new heat sink equipped with flow distributor are firstly evaluated and then, the optimal cases are reported to reach minimum surface temperature, minimum entropy generation and minimum pumping power.

The flow is divided in four stages before reaching main channels and therefore, a uniform flow distribution is obtained in the heat sink, which enhances the temperature uniformity.

References (38)

Cited by (34)

View all citing articles on Scopus
View full text