Experimental Research of Dynamic Instabilities in the Presence of Coiled Wire Inserts on Two-Phase Flow

The aim of this study is to experimentally investigate the effect of the coiled wire insertions on dynamic instabilities and to compare the results with the smooth tube for forced convection boiling. The experiments were conducted in a circular tube, and water was used as the working fluid. Two different pitch ratios (H/D = 2.77 and 5.55) of coiled wire with circular cross-sections were utilised. The constant heat flux boundary condition was applied to the outer side of the test tube, and the constant exit restriction was used at the tube outlet. The mass flow rate changed from 110 to 20 g/s in order to obtain a detailed idea about the density wave and pressure drop oscillations, and the range of the inlet temperature was 15–35°C. The changes in pressure drop, inlet temperature, amplitude, and the period with mass flow rate are presented. For each configuration, it is seen that density wave and pressure drop oscillations occur at all inlet temperatures. Analyses show that the decrease in the mass flow rate and inlet temperature causes the amplitude and the period of the density wave and the pressure drop oscillations to decrease separately.


Introduction
In many industrial systems in which boiling heat transfer exists, the �ow instabilities occurring based on the boiling heat transfer cause certain parts where heat is transferred to breakdown. e fact that the system pressure, �ow rate, and similar parameters create oscillations shortens the life of the process systems. It is possible to order the events decreasing the existence of the thermal systems operated by two-phase �ows as thermal fatigues, mechanical vibrations, the di�culty of control caused by high transient temperatures, and the burn-out occurring on the test pipe surface [1,2]. e main property categorising the type of two-phase �ow is the shapes that interfaces occurring between twophases take. e effect of the �ow direction on these occurring shapes is highly distinctive. Two-phase �ows are clas-si�ed as horizontal, vertical, and inclined. e �ow regimes and the characteristics occurring in horizontal, vertical, and inclined tubes are different from each other. e direction that the gravitational force affects according to the �ow direction brings together the main classi�cation [1,3,4]. Menteş et al. [5] and Kakaç et al. [6,7] studied two-phase �ows in different �ow directions.
Considerable efforts have been made to investigate twophase �ow instabilities by researchers for many years because the instabilities and thus the oscillations shorten the life of the systems. In different types of test tubes and tubes with different cross-sections, Coleman and Garimella [8] and Leung et al. [9] investigated the instabilities of two-phase �ow systems. e forced convection boiling in a horizontal tube was studied by Çomaklı et al. [10]. All types of dynamic instabilities were observed at all conducted temperatures, and the appearance boundaries of the oscillations were determined. It was observed that the instability of the system increased with the increase of the inlet temperature. Besides this trend, the periods, and the amplitudes of the pressure drop and density wave type oscillations decreased with the decreasing mass �ow rate and increased with the decreasing inlet temperature. Furthermore, it was noticed that the channel length had an important effect on two-phase dynamic �ow e Scienti�c �orld �ournal  instabilities. In narrow channels, Tadrist [11] conducted an experimental study to examine two-phase �ow instabilities. Parallel channel and single channel oscillations were clearly determined. Bao et al. [12] investigated experimentally the gas-li�uid two-phase �ow in a narrow channel. �eat transfer characteristics and pressure drop were studied for nonboiling single phase �ows.
In single phase and two-phase �ow systems, water is the most commonly used wor�ing �uid. �xcept for water, there are some other �uids and �uid mixtures. e R-11 wor�ing �uid in a two-phase horizontal boiling system was used by Çomaklı et al. [13] and Kakaç and Cao [14]. e amplitudes and the periods of the pressure drop type and the density wave type oscillations of R-11 were higher than those of water [13]. ermal oscillations and pressure drop type oscillations occurred for all thermal power levels. e periods and the amplitudes of the oscillations increased with the increase of thermal power and inlet subcooling [14]. Oil-water mixtures were used by Poesio [15] and Sotgia et al. [16] in two-phase �ow systems to analyse the combined effect of both �uids on dynamic instabilities. In order to investigate the two-phase �ow dynamic instabilities, different approaches have been employed besides experimental methods. �omogenous and dri� �ow models and the Wo�tan-�rsenbacher-ome �ow model were used by Moreno Quibén and ome [17] and Kakac and Bon [18] in a horizontal test tube, and the model results were con-�rmed with experimental models. Yu et al. [19] investigated two-phase pressure drop, forced convection boiling heat transfer and the critical heat �ux of water in a horizontal test tube. e results were successfully related to the correlations improved for different types of working �uids and boiling water in both narrow and wide channels.
In order to enhance the heat transfer in two-phase �ows, there are numerous techniques as in the single phase �ows. At the head of these techniques, inserting inner elements with different geometry and con�guration as turbulence promoter and surface area enhancer is the most preferred among the passive heat transfer enhancement techniques. Yılmaz et al. [20,21] examined the heat transfer enhancement methods used in two-phase �ow in detail and presented basic �ndings related to heat transfer enhancement in two-phase �ows. e effects of enhanced surfaces on two-phase �ow instabilities in a horizontal boiling system were investigated by Widmann et al. [22]. Dynamic instability types occurred in all different tube types. Çomakli et al. [4] and Karagoz et al. [23] studied the dynamic �ow instabilities in a forced convection boiling system using a smooth tube and a tube with different insertions. By the addition of the insertions into the �ow �eld, it was seen that the periods and the amplitudes of the pressure drop type and density wave type oscillations of the test tube with inner elements were higher than those of the smooth tube, and the smooth tube had a very stable structure in terms of �ow instability. It was also noticed that the stability of the system increased with the decrease of the equivalent diameter for the same type enhanced surface element. In this experimental work, the coiled wire insert with two pitch ratios was used as turbulator in addition to the smooth tube in order to investigate the effects on two-phase �ow instabilities. Besides the inner element and thus the pitch ratio, the effect of the inlet subcooling and the mass �ow rate were also examined.

Experimental Set-Up
e experimental set-up is schematically shown in Figure 1, which has been designed to create three main dynamic instabilities (pressure drop type, density wave type and thermal oscillations). In this study, the effects of inlet subcooling, the insertion type, pitch ratio, and the mass �ow rate on two-phase �ow instabilities have been investigated. As seen in Figure 1, the experimental system consists of three main parts: �uid supply section, test section, and �uid recovery section. e working �uid (water) in liquid form supplied from the �uid supply section enters the circular tube and turns into a mixture of liquid and vapour by the effect of the heat input from the tube walls. e �uid that is nearly in the vapour phase at the end of the test section is sent to the �uid recovery section. is �ow process goes on consistently as a close loop.

Fluid Supply
Section. e �uid supply section is composed of a main tank (1), �ow control valve (2), �ow meter (3), and heater (4). e main tank made of stainless steel stores the water used during the experiments. e cylindrical tank has 3 metres of height and 0.7 m 3 volume rates. e �ow control valve is used to set the rate of the �ow to the desired level. e �ow rate is controlled by two �ow meters. e ranges of the �ow meters is 0�400 and 0�1000 l�hour, respectively. e �rst �ow meter is only used in small �ow rates to obtain more sensitive measurements. e inlet subcooling level is remarkably important for two-phase �ow experiments. erefore, a shell-tube heat exchanger is used to send the water at the desired temperature to the test section. e working �uid is heated by electrical heaters while passing through the tube. e temperature of the �uid heated by two heaters (each heater is 4 kW) is controlled with a digital thermometer.

Test
Section. e test section where the dynamic �ow instabilities are generated included a surge tank (5), inlet �uid control valve (6), test plenum (7), test tube (8), DC power supply (9), ori�ce (10), digital manometer (11), �ow meter (12), and pressure transducer (13). In order to create the compressible volume for two-phase �ow of water, a surge tank with 0.05 m 3 was used. A level viewing glass, which is durable against high pressure up to 30 bar, was added to the surge tank in order to see the variations in both water and the compressible volume. e surge tank also contained a level range to measure the level and a manometer to measure the pressure inside the tank. A turbine type �ow meter to measure the oscillations in water �ow rate, a Bourdon type manometer to measure the pressure of the water at the tube inlet, and a pressure transducer to measure the �uid pressure oscillations at the tube inlet were installed between the surge tank and  the test tube. e inlet temperature of the working �uid at the tune inlet was measured with a T-type thermocouple. An ori�ce plate was installed on the exit side of the tube in order to de�ne the e�ects of exit restriction on �ow oscillations. A Bourdon type manometer was used to measure the pressure di�erence caused by the ori�ce plate.  A stainless-steel pipe of 0.017 m outer diameter, 3.4 mm wall thickness, and 3 m length was heated with uniform electrical heat input from a DC generator (24 kW power supply). To measure the oscillations of the wall temperature, 28 copper-constantan thermocouples were �xed on the outer surface of the test tube ( Figure 2). Half of these thermocouples, that is, 14 thermocouples, were �xed along the top of the test tube� the other half were �xed along the bottom of the test tube. e �uid outlet bulk temperature was measured by placing another thermocouple midstream inside the tube immediately aer the test section. e exit restriction (10) created the necessary pressure drop. In this system, an ori�ce plate was used as exit restriction and the diameter ratio of the exit restriction equal to 0.448 was used. e diameter ratio is de�ned as the ratio of the inner diameter of the ori�ce plate to the inner diameter of the tube ( . Downstream of the exit restriction, the exit pressure was measured with a pressure gauge. e test tube was electrically heated uniformly and insulated with glass wool that can withstand a temperature of 1000 ∘ C. e heat input was determined by measuring the current and the voltage drop across the heated section. In order to avoid �oating voltage e�ects, the thermocouple bead was insulated from the electrically heated tube wall surface with a dab of e �cienti�c World �ournal  electrically nonconductive paste. e test tube was followed by a sight glass for visual inspection of the �ow.

Fluid Recovery Section.
Aer passing through the test tube, the working �uid comes into the �uid recovery section in the vapour phase. e �uid recovery section mainly consists of four components: condenser (14), nitrogen tank (15), regulator (1�), and �uid storage tank (17). e working �uid, nearly all in the vapour phase at the end of the test section, is condensed with a water cooled condenser. e condensed water is sent to the storage tank. e storage water is again sent to the main tank by pressuring with the help of nitrogen gas.

Experimental
Procedure. e experiments were conducted in two different categories: steady and unsteady experiments. In steady experiments, the steady state characteristics were de�ned� likewise, the two-phase �ow dynamic instabilities were investigated in unsteady experiments. In order to investigate the effects of inlet subcooling on steady and unsteady state characteristics, the experiments were conducted for three different inlet temperatures of 15 ∘ C, 25 ∘ C, and 35 ∘ C and three different test tubes under constant heat input (24 kW), system pressure (7.5 bar), and exit restriction (diameter ratio of 0.448). e �rst mass �ow rate for each group is 110 g/s, because the �ow is single phase �ow at high �ow rates, and the mass �ow rate was decreased by 10�12 g/s slight steps to de�ne the characteristic curve. e lowest �ow rate was taken as 20 g/s in the experiments. �ecause of the burn-out possibility, �ow rates lower than 20 g/s cannot be achieved. In Figure 3(a), the heat transfer enhancement surfaces and characteristics are presented. Figure 3(b) shows the heat transfer enhancement elements: rings. e insertions are generally characterised with the effective diameter. Table  1 presents the effective diameter and pitch ratios for all types of tube. e experiments were �rst conducted with a smooth tube (Tube-1) and then repeated for Tube-2 and Tube-3. At the beginning of the experiments, the main tank was pressured with high pressure nitrogen gas, and the system pressure was set by the pressure regulator on the nitrogen tube. In steady state experiments, the nitrogen gas was not used, so the gas inside the surge tank was drained. As for the unsteady experiments, the surge tank was pressured with the nitrogen gas in order to create a constant compressible volume. e water level inside the tank was controlled and observed by the aid of a transparent tube level gauge. For both steady and unsteady experiments, the �ow rate was set to the highest value of 110 g/s with a control valve, and to achieve the desired temperature at the tube inlet the temperature of water leaving the main tank was controlled with a digital thermostat. Later, the cooling water was sent to the condenser, and then the constant heat was applied to the tube walls by a DC power supply. Finally, the system was operated and brought to a steady state. It was decided that the steady state was achieved when no higher change than 0.5 ∘ C was observed on the test tube surface temperatures. Aer the steady state was reached, all measurements were taken and the same procedures are repeated for each other mass �ow rates up to 25 g/s. e fact that quick variations in the pressure value of the surge tank and water level were observed means that the oscillations began. At the unsteady experiments, the mass �ow rate was decreased until the oscillation boundary was reached. In order to de�ne the boundary where the pressure drop type oscillations break up, the periods of the oscillations were observed. e slight periods in the oscillations show the boundary where the pressure drop type oscillations �nished and the independent density wave type oscillations started. �y the decreasing mass �ow rate, the wall temperatures were followed carefully and the creation of the thermal oscillations was provided. e experiments were suddenly halted when the burn-out began.

Experimental Measurements and Uncertainties.
In the experimental facility, the temperatures were measured with T-type copper-constantan thermocouples with a 0.25 mm diameter. e reading measurement of the temperature taken by the thermocouples was within ±%0.5 ∘ C. e thermocouples were placed in a copper pipe for inlet and outlet temperature readings and �xed on the tube outer wall for de�ning two-phase �ow regimes and oscillations. �n �dvantech data reading card was used for conversion of the signals taken from thermocouples, pressure transducer, and �ow meters, separately. e total uncertainty in readings based on chosen acquisition level was in the range of 0.1-0.5 ∘ C. e measured pressures in the experimental study are the pressure of the main tank, surge tank, and nitrogen tank. e pressures at the ori�ce inlet and outlet were also measured. e uncertainty read on the analogue manometers was ±%0.1 bar. e outlet pressure of the ori�ce was measured with a digital manometer. e pressure transducer was only used to measure the oscillations occurring at the test tube inlet and outlet; the pressure uncertainty of the transducer is ±%0.1 bar.
On the determination of the oscillation and the stability boundaries, the mass �ow rate measurements should be taken carefully. �wo �ow meters ha�ing �ow rate setting on itself are used to measure and set the mass �ow rate. e ranges of the �ow meters are 0�400 and 0�1000 l/hour, respecti�ely. �y these �ow meters, the uncertainty rate is ±%0.4 l/hour. e total uncertainty of the readings taken from the turbine type �ow meter used to measure the �ow rate oscillations is ±%0.05 l/hour. e power rate of the DC power supply was set with the setting button on the supplier, and the heat input values were read from the digital volt and current gauges. e DC power supply was controlled within ±%0.2. of pressure drop versus mass �ow rate for each tube con�guration. On measuring the pressure drop, the pressure drop between the surge tank and the test tube outlet is used. Also, the steady state curves depending on the inlet subcooling are plotted on de�ning the steady state characteristic curves. ese curves are used for understanding the �ow characteristics.

Results and Discussion
e positive sloped parts of the curves that correspond to high mass �ow rate values represent the single phase �ow region. e �rst bubbles are observed at the point on which the value of the pressure drop is at a minimum with a decreasing mass �ow rate and the slope of the curve is negative. On the same point, two-phase �ow starts by the observation of the �rst bubbles. �uch more vapour �ows with liquid phase as long as the number of the bubbles increases in two-phase �ow. is increase in the number of the bubbles causes the density to decrease in terms of the liquid phase and thus the pressure drop to increase. As a result of the fact that the mass �ow rate is decreased, the parts of the curves with a negative slope up to the saturated vapour boundary have a positive slope and the pressure drop starts to decrease. On the steady state characteristic curves, the region between the dashed lines represents the unstable region where the pressure drop type (p. d. o.) and the density wave type (d. w. o.) oscillations occur at the same time; this region is called the superimposed region. e le side of the superimposed region (the le side of the dashed line on the le hand side) indicates only the density wave type (d. w. o.) �ow oscillations.
Using inner elements such as rings and twisted tapes in the �ow �eld causes more friction than the smooth ones and also causes the pressure drop to increase. On the other hand, the creation of the turbulence and the removal or destruction of the boundary layer enhance the heat transfer rate. In Figures 4, 5, and 6, implies that no heat input is applied and this case presents the single phase steady state characteristics. e highest pressure drop is observed in tube-2, whereas the lowest pressure drop value is in tube-1. On the comparison of the pitch ratios, the pressure drop increases with the increase in the pitch ratio. Tubes are ordered from the highest pressure drop to the lowest value as tube-2, tube-3, and tube-1.
In Figure 7, the comparison of the steady state characteristic curves for all investigated tubes are presented at 25 ∘ �. As shown in the graph for �ve di�erent test tubes, it is observed that the pressure drop between the maximum and the minimum points of the curves increases and the points where the boiling starts slip towards lower mass �ow rates. Furthermore, the increase in the inlet subcooling increases the slope of the two-phase �ow region and decreases the pressure drop values with the decreasing mass �ow rates. It also can be said that the system becomes more unstable as long as the negative slope angle of the steady state characteristic curves increases. As the negative slope angle of the smooth tube is lower than the tubes with inner elements, it can be concluded that the smooth tube is more stable than the other tubes.
In two-phase �ow systems, as the oscillations occurring under unsteady state conditions harm the systems, the fact that they are kept under control is highly important. Figure  10 presents the comparison of the oscillation boundaries for all tubes. On these curves, the region on the right side of the dashed lines implies the single phase �ow region on high mass �ow rates, and the space between the lines implies the unsteady region where the pressure drop and density wave type oscillations occur as imposed. e le side of the middle region represents the density wave oscillations only. In the single phase �ow region corresponding to the high �ow rates, no oscillation is observed. By the creation of the �rst bubbles, the oscillations begin occurring. ese oscillations are the pressure drop type (p. d. o.) and the density wave type (d. w. o.) oscillations. e beginning and the end points of the p. d. o. slip towards lower �ow rates as long as the inlet subcooling rate increases. By the decrease of the inlet temperature, the single phase �ow region increases and thus the system becomes more stable.
e reason why the system is more stable is the fact that the oscillations start at lower mass �ow rates. e oscillations in the tubes with insertions begin at lower �ow rates than the smooth tube and �nish at higher �ow rates. e insertions used in the horizontal tubes in which the phase separation occurs between the vapour and the liquid by the effect of the gravity disperse the vapour and cause the liquid bubbles to occur on the upper wall besides the vapour. Regarding the oscillation boundaries, the start point of the p. d. o. draws back towards the lower �ow rates as long as the inlet temperature decreases. is case shows that increasing the inlet subcooling makes the system more stable for all tubes investigated. e stability boundary curves are in same manner for all tubes. e unsteady region of the �ow e�pand, that is to say the imposed oscillation region of p. d. o. and d. w. o., grows by the increase of the gap between the boundaries. In Figure 8, it is clearly seen that tube-1 has the widest boundaries. is means that the most unstable �ow occurs in tube-1. e tubes with coiled wire insertions (tube-2 and tube-�) provide a more stable �ow than the smooth tube. �s for the pitch ratio, the fact that the pitch increases makes the system more unstable. In Figures 9, 10,   too long, the inner compressibility is enough for the pressure drop oscillations to occur; if not, the compressible volume is provided by a surge tank located at front side of the test section. In industry, the pressure drop oscillations for the systems including two-phase �ows have a large oscillation period and high oscillation amplitude. is case causes the life of the system to decrease and makes it difficult to control the �ow. erefore, the fact that the pressure drop oscillations are determined and taken under control is vitally important.
In Figures 12 and 13, the e�ects of the mass �ow rate on inlet pressure periods and amplitudes for all investigated tubes at 15 ∘ C are presented. For all tubes, the periods and the amplitudes belonging to the inlet pressure increase by the increase of the mass �ow rate. If an order is made in terms of the periods and the amplitudes, it easily seen that the periods and the amplitudes in tube-1 (smooth tube) are higher than those of the tubes with insertions. e tubes are ordered from the highest period and amplitude to the lowest ones as tube-1, tube-3, and tube-2. According to this order, the smooth tube with the biggest effective diameter has the biggest period and amplitude, while tube-2 with the smallest effective diameter has the smallest period and amplitude. In the other words, it can be said that the period and the amplitude increase as long as the effective diameter increases. Furthermore, the periods and the amplitudes also increase with the increase in the pitch ratio.
e density wave �ow oscillations occur as a result of boiling and the hydrodynamic behaviours of the �ow type. e density wave �ow oscillations take place together with the pressure drop oscillations as superimposed in the region with negative slope. e pure density wave oscillations are observed as long as they move to the upper points of the negative sloped region, towards the unstable region. e period and the amplitude values of the pure density wave oscillations are lower than the pressure drop type. Furthermore, the pure d. w. o. in smooth tubes occurs at higher �ow rates relative to the tubes having insertions. In two-phase �ows, the separation of the phases in the smooth tubes is observed more clearly than in the tubes with insertions.
In Figures 14, 15 and 16, the time dependent variations of both top wall and the bottom wall, the inlet pressure and mass �ow rates are presented. For each ring type, the d. w. o. oscillations are observed and these oscillations cause �uctuations with big amplitudes in the inlet pressure, top and bottom wall temperatures, and the mass �ow rate. Also, oscillations in the mass �ow rate are observed in addition to the inlet pressure. e amplitudes of the bottom wall are higher than at the top wall. When the thermal oscillations occur, the inlet pressure and the mass �ow rate oscillate at lower frequency with large amplitude. Figures 17 and 18 show the variations of the periods and amplitudes with mass �ow rates. For all investigated tubes, the periods and the amplitudes increase as long as the mass �ow rate increases. e amplitude and period of the smooth tube are higher than the other con�gurations. In terms of both the amplitude and the period, the order from the biggest to the smallest is tube-1, tube-3, and tube-2. According to the order made for pure d. w. o., the smooth tube with the biggest effective diameter has the biggest period and amplitude, and the tube-2 with the smallest effective diameter has the smallest period and amplitude. Similar to the results of pressure drop oscillations, the period and the amplitude increase as long as the effective diameter increases. Also, the periods and the amplitudes also increase with the increase in the pitch ratio.

Conclusions
e aim of this study was to experimentally investigate the effect of the coiled wire insertions on dynamic instabilities and to compare the results with the smooth tube for forced convection boiling. e experiments were conducted in a circular tube, and water was used as the working �uid. �wo different pitch ratios ( and 5.55) of coiled wire with circular a cross-section were utilised. e constant heat �ux boundary condition was applied to the outer side of the test tube, and the constant exit restriction was used at the tube outlet. e remarkable conclusions are ordered as presented below.
(i) e rings had a higher pressure drop than the smooth tube and the pressure drop increased with the increase in the pitch ratio. (ii) e system became more unstable as long as the negative slope angle of the steady state characteristic curves increased. As the negative slope angle of the smooth tube was lower than the tubes with inner elements, the smooth tube was more stable than the other tubes. (iii) In the single phase �ow region corresponding to the high �ow rates, no oscillation was observed. e beginning and the end points of the p. d. o. slipped towards lower �ow rates as long as the inlet subcooling rate increased. By the decrease of the inlet temperature, the single phase �ow region increased and thus the system became more stable.
e Scienti�c World Journal (iv) e reason why the system was more stable was the fact that the oscillations started at lower mass �ow rates. e oscillations in the tubes with insertions began at lower �ow rates than the smooth tube and �nished at higher �ow rates. (v) e start point of the p. d. o. drew back towards the lower �ow rates as long as the inlet temperature decreased. is case showed that increasing the inlet subcooling made the system more stable for all the tubes investigated. (vi) e imposed oscillation region of p. d. o. and d.
w. o. grew by the increase of the gap between the boundaries. (vii) e fact that the bubbles occurred because of the too small tube diameter covered both the top and the bottom walls prevented higher temperature differences to comprise between the walls. (viii) For each tube, the p. d. o. oscillations were observed and these oscillations caused the �uctuations with big amplitudes in the inlet pressure, top and bottom wall temperatures, and the mass �ow rate. (ix) e oscillations in the mass �ow rate were observed in addition to the inlet pressure. e amplitudes of the top wall were higher than those of the bottom wall. (x) When the thermal oscillations occurred, the inlet pressure and the mass �ow rate oscillated at a lower frequency with a large amplitude.