Experimental study of pressure drop oscillations in parallel horizontal channels

Highlights

• PDO limit cycles in parallel channels differ notably from the single channel ones.

• Higher outlet wall temperatures during oscillations can be achieved.

• The system gets more stable than for single channel systems.

Abstract

Two-phase flow instabilities are an undesirable phenomenon present in many fields and scales, ranging from large heat exchangers and boilers in industrial applications to micro-scale heat exchangers for high density power electronics. In the present study pressure drop oscillations in a two parallel horizontal channels system have been experimentally investigated, focusing in the individual behavior of each channel. The balanced (same characteristic pressure drop curve) and unbalanced cases have been analyzed finding different limit cycles than the typical single channel case. No pressure drop oscillations with both channels following the typical limit cycle were found. The oscillation mode detected consisted in one channel performing the usual limit cycle, while the other was always oscillating in the superheated vapor region.

Introduction

Two-phase flow boiling instabilities are an undesirable wide spread phenomenon present in many fields and scales, ranging from large heat exchangers and boilers in chemical and nuclear applications to micro-scale heat exchangers for high density power electronics in space applications. These types of instabilities can cause operational problems, affecting the efficiency and control of the processes, and in extreme situations, can lead to burnout and breakage of the equipment. Therefore the importance of having a clear understanding of its behavior. Two phase flow instabilities have been reviewed by several researchers (Boure et al., 1973, Kakaç and Bon, 2008, Tadrist, 2007). These instabilities are usually divided into two main groups: static and dynamic instabilities. Among the static instabilities, the most widely known is the Ledinegg type instability. For the case of dynamic instabilities, the most common ones and widely studied are density wave oscillations (DWOs) and pressure drop oscillations (PDOs). Pressure drop oscillations are actually compound instabilities, since they are dynamic instabilities triggered by a static instability. These kind of instabilities need in order to occur a compressible volume upstream the heated section (Cao et al., 2000, Doĝan et al., 1983, Gürgenci et al., 1983, Kakaç et al., 1990, Kakaç et al., 1977, Mawasha and Gross, 2001, Mentes et al., 1983, Padki et al., 1991). The focus of this work is placed on PDO in parallel channels.

Very few studies have dealt with PDO in parallel channels (Manavela Chiapero et al., 2012). Most of the experimental, theoretical and numerical studies performed on two-phase flow instabilities in parallel channels have dealt with DWO (Aritomi et al., 1979, Clausse et al., 1989, Fukuda and Kobori, 1979, Guido et al., 1991, Guo et al., 2010, Hirayama et al., 2006, Lee and Pan, 1999, Podowski et al., 1990, Xiao et al., 1993, Yun et al., 2008, Zhang et al., 2009). Among these few studies dealing with PDO oscillations in parallel channels, Kakaç et al. (1977) performed an experimental study reporting oscillations in the inlet pressure of a setup of four parallel channels with cross connections. The stability limits were investigated but no information about the flow behavior in each channel was reported. It can be seen from their results that the behavior of the four channels together is too complex to understand the interaction between channels in the system. Experiments on two parallel channels could have brought a clearer idea of the behavior of the parallel boiling channels system. Ozawa et al. (1989) analyzed in great detail flow instabilities in an analogous air–water mixture system. The set up used was a twin parallel adiabatic system, with the possibility of adding compressible volumes in the gas and liquid feed lines in each channel. Even though this was an adiabatic system, the similarity of the flow oscillations with PDO in a boiling channel when adding a compressible volume to the air feed lines was remarkable (Ozawa et al., 1979). When the compressible volume was added only to the air feed lines the typical single channel pressure drop oscillations were found with both channels oscillating in phase with the same amplitude. However, when the compressible volume was added to the liquid feed lines two new modes of oscillation were found, named as “U-tube mode” (quasi-static out of phase oscillation) and “multi-channel mode”. In the “U-tube mode” the oscillation was a quasi-static oscillation, with the gas flow and the liquid flow oscillating $180°$ out of phase in the two channels and the pressure drop approximately constant. The “multi-channel mode” was also a quasi-static oscillation, but with the gas flow in each channel oscillating independently, i.e. with different periods. Even though only two channels were used in this study, which made clear the behavior of the parallel system, the absence of heating and phase change makes the results hard to extrapolate to boiling parallel channels.

During the last ten years boiling in microchannels has gained the focus of many researchers due to the need of cooling of high-power density electronic devices and flow instabilities like PDO have been reported in several of such systems (Tadrist, 2007, Wu and Cheng, 2003, Zhang et al., 2010). Qu and Mudawar (2003) reported severe PDO in a two-phase microchannel heat sink, showing all channels oscillating in phase. Much effort has been made trying to stabilize these systems, for example by placing inlet restrictions (micro-orifices) (Szczukiewicz et al., 2013b, Szczukiewicz et al., 2013a). Again for the case of microchannels, the number of channels studied is too large to see any interaction between channels. Besides, this micro-systems usually have large common flow restrictions at the inlet and outlet of the parallel arrangement, making the channels move all together as a whole. Furthermore, the instrumentation is set in most of the cases in order to measure global variables and not the individual channels behavior.

Manavela Chiapero et al. (2011) analyzed numerically PDO oscillations in two parallel channels arrangements under balanced and unbalanced heat loads. The oscillation modes found in both cases were in phase. The unstable region was divided into two regions, “Region 1” where the oscillations take place, and “Region 2”, dominated by stable maldistributed solutions. The oscillation mode with unbalanced heat loads found was with the most heated channel oscillating in the superheated vapor outlet region and the other channel following the typical PDO limit cycle. In a later study, Manavela Chiapero et al. (2013) considered the thermal capacity of the heated pipes in the model, finding the maldistributed oscillation mode also for the case with balanced parallel channels. These modes of oscillations are strongly connected to the global characteristic curve of the system. This global characteristic curve is the graphical representation of all the possible equilibrium points for the total pressure drop along the parallel channels arrangement as a function of the total mass flow rate. Akagawa et al. (1971) performed a broad theoretical and experimental analysis on the possible solutions and their stability in a three parallel boiling channels system. In this analysis the global behavior and flow distribution for equally heated channels and unbalanced heated channels was clearly shown and explained based on the characteristic curves of each channel. In the last decade many studies have also dealt with flow distribution and pressure drop vs. mass flux behavior for a parallel arrange of boiling channels (Minzer et al., 2006, Minzer et al., 2004, Natan et al., 2003, Pustylnik et al., 2006, Taitel et al., 2008).

From the available literature it can be seen that there is very little information regarding the different modes of oscillation for PDO in parallel boiling channels and the physics behind the different behaviors. Furthermore, the experimental results reported are few and show contradictory effects. On the numerical side, few analysis have been made showing new results and limit cycles but there is still not detailed experimental data available to confirm these results.

In the present study, pressure drop oscillations are experimentally analyzed for two parallel horizontal heated channels. The aim of the study is to analyze two parallel boiling channels with the focus placed on the flow behavior at each channel in order to provide new basic knowledge about the different possible limit cycles in such configurations. In Section 2 the experimental facility is described together with the experimental procedure and the accuracy of the measurements. Section 3 shows the results obtained for the different cases analyzed. In Section 4 the main results and the limitations of the study are discussed. Finally, in Section 5 the main conclusions from the work are drawn.