Experimental study and dynamic simulation of the continuous two-phase instable boiling in multiple parallel microchannels

Highlights

• Continuous two-phase instable boiling in parallel multiple micro-channels.

• Evaluation of heat transfer and pressure drop correlations with measured data.

• Observed backflow and movement of liquid and gas interface in unstable boiling.

• New proposed dynamic simulation model considering phenomena and mechanisms.

• Some proposed methods for suppressing unstable flow boiling in microchannels.

Abstract

Continuous vapor liquid two-phase flow means that both subcooled single phase and saturated flow boiling always exist in microchannels with the moving interface between them. The instable boiling phenomena and mechanisms are very complicated and needed to be deeply investigated and well understood. In the present study, both experiments and dynamics simulation of the continuous two-phase instable boiling in multiple parallel microchannels were systematically conducted. In the experimental aspect, characteristics of the unstable flow boiling in the microchannels were experimentally investigated using acetone as the working fluid and flow boiling processes were visualized with a high speed video camera simultaneously. The test section includes 16 parallel microchannels having a hydraulic diameter of 104.3 µm. The test mass flux ranged from 437.2 to 868.1 kg/m² s and the test heat flux ranged from 0 to 100 W/cm². The measured flow boiling heat transfer coefficients and two-phase frictional pressure drops are analyzed according to the observed flow phenomena and mechanisms. They were further compared to the existing flow boiling heat transfer and two-phase pressure drop prediction methods. The top three flow boiling heat transfer methods predict more than 65% of the heat transfer coefficients within ±30% while the top two pressure drop methods only predicted 43% of the measured data within ±30%. It shows that the continuous two-phase instable boiling mainly occurs at the mass flux larger than 607.6 kg/m² s and the heat flux larger than 30 W/cm². The backflow causes instability of the liquid and the two-phase interface which change the thermal resistances of the fluid. In the theoretical aspect, a dynamic simulation model was developed by using the thermal network method and was used to simulate the dynamic process of the instable boiling process. The simulated results were compared to the experimental data in this study and from the literature and the relative errors are within ±24.4%. According to the simulation and analysis, the mechanisms are that the axial thermal conduction in the channel walls generates a negative differential zone in the heat power characteristic curves, where the wall temperature decreases with increasing the heat flux in the negative differential zone. In order to suppress the oscillations in the boiling process, it is essential to eliminate the negative differential zone. Enhancing single phase forced convection heat transfer by 100% or enhancing flow boiling heat transfer by 50% can suppress the continuous two-phase instable boiling in the microchannels. Elongating the microchannel length and enhancing the axis thermal conduction can reduce the amplitudes of the oscillations by 60–80% at the conditions of G = 700 kg/m² s, q_eff = 70 W/cm² and T_in = 30 °C.

Introduction

The requirements of high performance electronic chips are increasing sharply in the wake of rapid developments of micro-electronic industry. Improving integration level and packaging density are important ways to ensure the high performance of chips but also generate extremely high heat flux density. As an efficient technology for clean energy utilization, fuel cells and their energy conversion can generate high heat fluxes as well. The effective cooling methods are important for increasing service life of the relevant critical device in fuel cells as pointed out by Wang [1]. Due to the confined spaces, multiple parallel microchannels are generally used in fuel cells, understanding the fluid flow and thermal management in the microchannels are critical in fuel cell scaling-up [1], [8], [9]. Furthermore, microchannel technology becomes a cutting edge technology for heat transfer and fluid flow in many engineering applications but there are also big challenges [1], [3], [5], [8], [9]. Besides the channel size, shape, surface condition, fluid physical properties and operation parameters, understanding the fluid distribution and instability becomes more important in understanding the physical phenomena and technology development but not yet well understood. In the fluid distribution issues, Wang [1] presents a comprehensive review on the relevant theory and practice of flow flied designs in fuel cell. In this paper, the research mainly focuses on the experimental and modeling aspects of the continuous two-phase instable boiling in the multiple parallel microchannels as it is impractical to cover all aspects in one paper. However, it is suggested to consider flow distribution issues in flow boiling and instability study in future.

The traditional convectional cooling methods have already reached the limits of heat transfer, thus the micro-heat sinks become key technologies for safe and stable operation of chips. Microchannel proposed by Tuckerman and Pease [2] is one of the most effective and potential forms of micro-heat sinks. According to the micro-scale definition proposed by Kandlikar [3], the hydraulic diameters of microchannels are generally smaller than 200 μm, thus the surface area of per unit flow volume is very large, which is in favor of high efficiency of heat transfer. However, the oscillations of temperature and pressure caused by instable boiling brings extra burden to fluid loop, weakens heat transfer efficiency and possibly damages chips [4]. Mechanisms of instable boiling in microchannels and macro-channels are not exactly the same because of the microchannel confinement effects, which have already attracted many scholars’ attention [5]. Besides decreasing the heat transfer performance and service life in the practical applications, instable flow boiling in microchannels may lead to the big deviations in the heat transfer and pressure drop measurements which may be one the main reasons why the proposed prediction methods give very big errors [5], [45]. A lot of experimental and numerical investigations have been carried out to study and suppress instable boiling over the past decades. However, due to the complicated instable flow boiling phenomena in microchannels, mechanisms, complete theory and models have not yet been well understood so far and need to be continuously investigated in experimental, theoretical and modeling aspects.

The main mechanisms of dynamic instable boiling in macro-channels include the pressure drop oscillation, the density wave oscillation and the thermal oscillation, which also exist in microchannels [6]. These instabilities can be affected by many factors, such as two-phase interaction, flow distribution uniformity, pressure drop, thermal load, wall surface properties and fluid properties etc. Recently, Karayiannis and Mahmoud [7] have presented a comprehensive review on the research of flow boiling in microchannels including instable flow boiling phenomena and theory exploration in microchannels. Quite a few studies of the boiling instabilities in straight and smooth microchannels have indicated that the micro-scale effects make boiling instabilities more sensitive by enhancing disturbances of gas phase [7]. The non-uniform distribution of mass flux has very complex dynamic characteristics, and the inappropriate or oversimple microchannels usually intensify this phenomenon [8]. The non-uniform distribution of mass flux is highly possibly triggered by drastic disturbance of bubbles and then leads to the density wave oscillation and the thermal oscillation in microchannels, which have negative effects on service lives of heating elements [9]. The microchannels with small hydraulic diameters and straight structures can barely induce turbulences for interrupting thermal boundary layers, which possibly limit further improvement of heat transfer performance of single phase convection [10]. The smooth wall surfaces are also lack of active nucleation cavities, and the thin liquid film is hard to maintain because of surface tension and poor wettability [11]. As a result, the superheated degree of wall is huge in high thermal load condition. Moreover, the restriction of the channel wall is significantly enhanced in the microchannels and changes the way in which bubbles grow and move. The newly formed bubbles can grow up to be similar size of cross-section of microchannel very easily, after that, bubbles have to expand to upstream and downstream directions [12], [13], [14], [15], [16]. More recently, Li and Hrhjak [14], [15] and Tuo and Hrnjak [16] have conducted careful investigation on the flow reversal phenomena and modeling of bubble dynamics in a single microchannel. Their important research has further provided deep insights to understand the very complicated instable flow boiling in microchannels. If the thermal load is high enough and newly formed bubble is closed to inlet, some vapor is able to rush out of channel via inlet, and then the phenomenon backflow is formed. The uniform and stable liquid film usually has good heat transfer abilities [17], [18], but the backflow may disturb the heat transfer ability and increase the flow resistance significantly, which could cause more complicated dynamic instabilities.

In general, the instable boiling phenomenon can be characterized as oscillations of mass flow rate, pressure drop, and wall temperature and so on. Xu et al. [19] noticed that three types of oscillations appeared in the microchannels include the large amplitude long period oscillation (LALPO), the small amplitude short period oscillation (SASPO) and the thermal oscillation (TO). SASPO can be superimposed upon LALPO, or exist alone. The flow pattern of SASPO was continuous two-phase with periodic movement of liquid/two-phase interface. Wu and Cheng [20] observed similar flow pattern under moderate heat and mass fluxes. Besides, at low heat and mass fluxes, the liquid/two-phase alternating flow was observed; at high heat and low mass fluxes, the liquid/two-phase/vapor alternating flow was observed. These two types of oscillations usually generate large amplitudes and long period. Wang and Cheng [21] designed several microchannels with different lengths and structures of inlet and observed continuous two-phase instable boiling appeared in their experiments. They also proposed stability map based on their experimental results. The stable and instable regions were divided by the outlet vapor quality in their map. Qu and Mudawar [22] experimentally investigated instable boiling in parallel microchannel with square cross-sections. The working fluid was water. SASPO and LALPO were observed in their studies. Experimental results showed that LALPO were mainly caused by compressible volume in the upstream of fluid loop and drastic evaporation of working fluid. Lee and Mudawar [23] studied the effects of flow resistance of fluid loop on instable boiling. The fluid in the microchannels will exchange momentum with fluid in the compressive volume spontaneously. They thought that throttling elements installed between compressible volume and inlet of microchannel can eliminate oscillations.

In the theoretical and modeling aspect, various simulation methods for the dynamic process of instable boiling and mechanisms of heat transfer considering the axial heat conduction in microchannels have also been developed over the past years. Li and Hrnjak [15] proposed a predicting model based on the bubble dynamics for investigating the evolvement of periodical flow reversal. The growth of each bubble in single microchannels is divided into partial confinement stage and full confinement stage. The results of calculation show that the local pressure peak caused by the build-up of downstream flow resistance can cause negative pressure gradient, which induces flow reversal. Zhang et al. [24] developed a dynamic flow oscillation model based on the modern cybernetics. The relation between the mass flux and the pressure drop is a cubic curve. The dynamic model reflects the momentum exchange of fluid between compressive volume and microchannel. An active control strategy based on feedback theory is also proposed. Kuang et al. [25] developed a dynamic model based on the pressure drop curve with negative differential region. Their simulation results show that the negative differential region is mainly caused by the variation of two-phase frictional pressure drop. Zhang et al. [26] numerically investigate the forms of instable boiling at different wall thermal inertia conditions. According to the conservations of energy and momentum, the effects of wall thermal inertia cannot be neglected. The wall temperature oscillation amplitude decreases with increasing wall thermal inertia, which results in smaller thermal stresses. Chiapero et al. [27] theoretically studied on the negative differential zone of the hydraulic characteristic at different working conditions. Their simulation results indicated that the negative differential coefficient is mainly affected by the system pressure and the uniformity of heat flux. For decreasing the instabilities, lots of microchannels with inlet restriction or expanding cross-section are proposed [28], [29], [30], [31]. Furthermore, a number of studies have indicated that the thermal conduction channel walls cannot be ignored in both flow boiling and single phase flow heat transfer in microchannels. Just to name two studies here, Maranzana et al. [32] develop a one-dimensional model to investigate the influence of axial conduction in the channel walls of the microchannels. They have found that the wall heat flux density, for small Reynolds numbers, can become strongly non-uniform. Most of the flux is transferred to the fluid flow at the entrance of the mini-micorchannels. Celata and Cumo et al. [33] conduced careful measurement of single liquid heat transfer in microtubes. Their results show that a decrease of Nusselt number with decreasing diameter, an axial dependence that is linked to thermal entrance effects and a dependence of the Nusselt number also on Reynolds number becomes significant in laminar flow. Therefore, all the affecting factors including single liquid phase flow, two phase flow boiling and their interface moving and other parameters in microchannels should be considered in developing a dynamic model for instable flow boiling in microchannels.

Most of the available studies on the continuous two-phase instable boiling microchannels are based on experiments. The inlet temperature, mass flux, heat flux and local heat transfer coefficients in the up and down streams are proved to be feasible for changing the period and amplitude of oscillation [34], [35], [36], [37], [38]. However, theoretically understanding the instable mechanisms are very limited, which has limited the development of methods for suppressing the continuous two-phase instable boiling. Some proposed physical mechanisms and theory are controversial. This is probably due to the lack of understanding the limiting factors and parameters. Therefore, both experimental investigation and reliable dynamic simulation models are still needed.

The originality of the paper is to provide new accurate experimental flow boiling heat transfer and two phase pressure drop data in the multiple parallel microchannels with very tiny hydraulic diameter of 104.3 μm and the instable characteristics in such microchannels and to further develop reliable dynamic models of the instable two-phase flow boiling in terms of the kinematic and energy analysis together with the observed flow and heat transfer phenomena and measured results. In developing the dynamic model, the actual microchannels with attachments are simplified to a Y-shaped thermal network with several lumped nodes which considers heat capacity and axial thermal conduction of walls. After making some simplifications and hypothesis, a group of differential equations have been established based on this thermal network. The simulation results with the new dynamic model have been compared to both the present experimental data and independent data digitized from the relevant studies in the literature. It shows promising and reasonable results according to the comparisons. Furthermore, the physical mechanisms and suppression methods of the instable two-phase flow are discussed by using various methods and fluids according to the simulation results and the dynamic model. It has thus provided a new approach to investigating the instable two-phase flow boiling and the mechanisms of the instable phenomena.