Experimental study on the flow interaction of two parallel rectangular jets through exits with sudden contraction

Accepted Manuscript Experimental study on the flow interaction of two parallel rectangular jets through exits with sudden contraction Pengyuan Liu, Hai Zhang, Yuxin Wu, Man Zhang, Junfu Lu PII: DOI: Reference:

S0894-1777(17)30214-5 http://dx.doi.org/10.1016/j.expthermflusci.2017.07.018 ETF 9160

To appear in:

Experimental Thermal and Fluid Science

Received Date: Revised Date: Accepted Date:

8 May 2017 21 July 2017 23 July 2017

Please cite this article as: P. Liu, H. Zhang, Y. Wu, M. Zhang, J. Lu, Experimental study on the flow interaction of two parallel rectangular jets through exits with sudden contraction, Experimental Thermal and Fluid Science (2017), doi: http://dx.doi.org/10.1016/j.expthermflusci.2017.07.018

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Experimental Study on the Flow Interaction of Two Parallel Rectangular Jets through Exits with Sudden Contraction Pengyuan Liu, Hai Zhang*θYuxin Wu, Man Zhang, Junfu Lu Key Laboratory for Thermal Science and Power Engineering of the Ministry of Education, Department of Thermal Engineering, Tsinghua University, Beijing 100084, China

Abstract: The flow characteristics of two parallel jets ejecting from the nozzles with sudden contraction exits were studied using particle imaging velocimetry (PIV) technique. The measured mean velocity components, velocity vector contours, turbulence characteristics and Reynolds shear stresses showed that the flow fields adjacent to the exits are significantly different from those of the dual flows ejecting from the nozzles with constant cross section area. The sudden contraction at the exits strongly promotes the merging of the two parallel jets and enhances the turbulence intensity in the merging region. The location of the combined point locates much closer to the exits when the contraction ratio in area varies from 1 to 1.5 and thereafter changes slowly when the contraction ratio further increases. The measured results also showed that the parallel jets ejecting from the nozzles with sudden contraction exits creates much stronger turbulence in the inner and outer shear layers of the two parallel jets than the jets ejecting from the nozzles with constant cross-section area.

Key words: parallel jets, flow measurements, sudden contraction, PIV, turbulent flow

1

Nomenclature u

streamwise velocity

u0

jet exit velocity

umax

maximum value of u

urms

Root mean square of u

v vrms a L1 , L2

lateral velocity Root mean square of v jet width inlet length and outlet length of nozzle

D

jet spacing, the distance between the centers of the two jets

Re

Reynolds number

x

x axis perpendicular to the main flow direction passing through the two nozzle centerlines

u

y axis perpendicular to the symmetric line between the nozzles

z

z axis parallel to the main flow direction

Zcp

the distance between the combined point and the nozzle exit

u 'v '

Reynolds shear stress

CR

The ratio of the inlet length to the outlet length

2

1. Introduction Parallel rectangular jets have been used extensively in many industrial applications, such as chemical processing, power generation, heating and air-conditioning [1-7]. Understanding the characteristics of the flow adjacent to the nozzle exits of these parallel jets is of importance to the industrial designs, and the associated studies have being conducted for decades. In general, as shown in Fig. 1, the flow field adjacent to the nozzle exits of two parallel jets can be divided into three regions along the flow path [1, 2]. The first one is the converging region, from the nozzle exit to the point at which the inside shear layers merge while the axial velocity equals to zero (denoted as the merging point, MP). The second region is called as merging region, from MP to the combined point (CP), at which the axial velocity reaches its maximum. The last region is called the combined region, locating in the downstream of CP where the two jets merge as a single jet. D

a

a

o Recirculation zone

x Inner shear layer Outer shear layer

Converging region

Merging point u=0

Merging region Combined point u = umax Combined region umax z

Fig.1 Flow structure of parallel jets 3

A number of experimental and numerical studies have been conducted on the interaction of the dual parallel jet flows, including the influencing factors on CP and flow structure and momentum transfer in the parallel jet flows [3-7]. The mixing of the two parallel jets in the converging region was experimentally investigated by Wang et al. [2] using particle image velocimetry (PIV). The authors found the strong momentum transfer happened in the merging region. A long, narrow low-fluctuation region was found in the space between the two jets with the turbulence intensity higher at the edges of the jets. Abdel-Salam and Tiwari [3] numerically found that the ratio of D/a and the nozzle width are related linearly to the locations of MP and CP, with the velocity having no effect on their location. Ozalp et al. [4] studied the flow of a Newtonian fluid in an axisymmetric 4:1 sudden contraction geometry using PIV. The results showed that as the fluid approaches the contraction, there is a complete change in the shape of the axial velocity profiles. The structure and the turbulence of the flow field were also studied by Miller and Comings [5], Nasr and Lai [6], Anderson and Spall [7], Fujisawa et al. [8] and others. However, most of existing studies are on the parallel jets ejecting from the nozzles without contraction or only on the single jet flow through an exit with sudden contraction. Based on the previous studies for the single jet flow through an exit with sudden contraction, the flow characteristics, including the velocity contours and profiles could be obviously different from those through exits with constant area duct. Therefore, in the present study, the flow characteristics and interaction of two parallel rectangular jets through exits with sudden contraction are studied. Under different 4

contraction ratios (CRs), i.e., the area of incoming flow over that of exit nozzle, the velocity contours, average velocities, merging characteristics and turbulence properties of the two parallel jets are measured by PIV. The results are then compared to those of two parallel rectangular jets through exit of constant cross section area.

2. Experimental 2.1 Experimental system The schematic experimental system is shown in Fig. 1. It mainly consists of an air supply unit, a test chamber unit, and a PIV measurement unit. The air was supplied by a Roots blower and its flow rate was controlled and measured by a flow rate meter. The test chamber was a cubic of 800 mm high, 300 mm wide and 300 mm deep. Two rectangular slot nozzles were mounted on the top cover of the chamber. Experiments were conducted at room temperature and ambient pressure. The structure of the nozzle with sudden contraction exit is shown in Fig. 2. Five sets of rectangular nozzles with different inlet length L1 and same outlet length L2 were used in the experiments. In dimension, L1 = 200, 100, 75, 60 and 50 mm, respectively and L2 = 50 mm. The width of the nozzles was 10 mm. Correspondingly, the contraction ratio in cross-section area (CR) was 4, 2, 1.5, 1.2 and 1, respectively. The separation distance between the nozzle centerlines, D, was set to 5 times of the nozzle width, i.e., D = 50 mm. According to our numerical simulation, the size of the test chamber was large enough to avoid the interference between the jet flows and the wall.

5

During the experiments, two sections were selected for flow field visualization and PIV measurements. Section 1 was along the midpoint of the nozzle length, and it was used to investigate the merging of the two parallel jets. Section 2 was along the midpoint of the nozzle width, and it was used to investigate the expansion of the jet along the streamwise jet. The exit velocity, u0, was kept at 15 m/s in all the experiments. Correspondingly, Reynolds number ( Re = u0 d0 / v , where d0 is the nozzle equivalent diameter and v is

the kinetic viscosity of air at 298 K) at the nozzle exit was ~17500. The two jets had the same discharge velocity in all experiments. Air nozzle

Tracing particles

Laser modulator Section 2 Section 1 \ Collector

[

] CCD camera

Fig. 2 The schematic of experimental system

L1

L2 Contraction ratio=L1/L2 Fig. 3 Schematic of the nozzle structure 6

2.2 Flow field visualization and PIV measurements The flow field adjacent to the exits of the two parallel jets was measured using PIV technique. The particle motion and the interactions were illuminated by laser sheet at 532 nm wavelength using an ND:YAG laser (New Wave Research Solo 120XT) with a typical energy of 120 mJ/pulse and a 10 ns pulse width. In the experiments, the laser beam was extended using a group of cylindrical lenses to form a thin laser sheet that passed through the jet test section. SiC particles were used as the tracing particles to show the distribution of gas flow field. The relative refractive index of SiC was 2.65, which was good for scattering signal [9]. The size of tracing particles was ~1 μm. Calculated by Eq. (1), Stokes number was less than 0.01, indicating the particles followed the fluid very well [10].

St =

Dp2 r p u0

(1)

18m L

where D p is the particle diameter, r p is the particle density, m is the fluid viscosity, and L is the characteristic length of the fluid. The flow structure of the parallel jets was recorded by a CCD camera with a resolution of 2048 ´ 2048 pixels at a frame rate of 7 frames per second. The physical resolution of each image was 78.2 μm/pixel. The noise signals were eliminated by a band-pass optical filter (center wavelength = 532 nm, half bandwidth = 3 nm) attached to the camera. The time period between each pair of images was 50 μs . The Nd:YAG Laser and the CCD camera were synchronized using a pulse delay generator to capture the transient particle motion.

7

The digital images were processed and analyzed using the DynamicStudio software. In each run, 100 images were selected. The velocity field was calculated using an adaptive-correlation technique with 32 ´ 32 interrogation windows. Spurious

velocity vectors were removed using the local median-filter technique [11,12]. For each run, time-averaged velocity and turbulence quantities, such as the urms and Reynolds stress, were obtained from the information captured in the images (N = 100) and calculated as follows: Streamwise average velocity u

Lateral average velocity v

Normalized urms

u ( x, y ) =

v( x, y) =

1 N

å

u ( x, y)

(2)

1 N å n v ( x, y ) N n =1

(3)

1 N

urms = u0

N n

n =1

N

å (u

n

- u )2

n =1

(4)

u0 N

Normalized Reynolds shear stress

u 'v ' u02

=

1 N

å (u

n

- u )(vn - v )

n =1

u02

(5)

3. Results and discussion The uncertainty in the PIV measurements was estimated based on two different measuring scale factors, which were used to convert the distance unit from pixel to mm. The velocity profiles at the plane of z = 140 mm when CR = 1 are shown in Fig. 3. Two data sets agreed very well with most measured zones. The errors of PIV are less than 3% except for those points of velocities close to zero.

8

0.0

u/u0

0.2 0.4 0.6 0.8 1.0 -8

Scale factor=10.49 Scale factor=7.48

Z=140mm

-6

-4

-2

0 x/a

2

4

6

8

Fig. 4 u profile measured by PIV based on two different scale factors 3.1 The effect of sudden contraction exit on single jet flow Figure 5 shows the typical flow field of a single jet flow through the nozzles when CR = 1 and 2. It can be seen that the flow is symmetric with respect to the centerline of the nozzle for both CRs. The maximum velocity locates at the axis of the jet. The outer boundary grows as the surrounding static fluid is continuously entrained into the main flow. However, when CR = 2, the lateral expansion of the jet becomes more evident. The observed behavior is governed by the pressure field induced by the shear layer vortices. The sudden contraction drives the flow from sides to the central area in the nozzle exit plane. This effect leads to high pressure on the outer shear layer

0

0

20

20

40

40

60

60

z,mm

z,mm

and higher spreading rate in the lateral direction.

80

80

100

100

120

120

140

140

0

20

40

60

80

100

120

140

0

x,mm

20

40

60

80

x,mm

9

100

120

140

(a) CR = 1

(b) CR = 2

Fig. 5 Time-averaged velocity vector contour of single jet flow in Section 1 (m/s) 50

z, mm

40

30

20

10

0 0

10

20

30

40 x,mm

50

60

70

Fig. 6 Schematic of the flow pattern with sudden contraction exit (CR=2) Figure 6 shows the schematic of the flow pattern with sudden contraction exit. The direction of the velocities in the middle of the stream does not change. Away from the centerline, the flow intends to go to the center of the nozzle exit. A vortex forms at the corner of the duct wall and the contraction surface. In addition, near the contraction, the angle between the velocity vectors and contraction surface is less than 45o, which means that the velocity in the lateral direction is larger than that in the streamwise direction.

3.2 The effect on flow characteristics of two parallel jets

3.2.1 On general flow field

Figure 7 shows the instantaneous scattering images of the silicon carbide jet when CR = 1, 2 and 4. For all the cases, the scattering signal is strong near the parallel jets zone and decays as the flow proceeds downstream. The flow fields exhibit obvious turbulent characteristics. The jet flow expands much faster and wider when the CR increases from 1 to 4. 10

Distance from the nozzle exit, Z, mm

0 40 80 120 160 200 240 280 320

(a) CR = 1

(b) CR = 2

(c) CR = 4

Fig. 7 Instantaneous scattering images for the two parallel jets at various CRs

3.2.2 On velocity field in the section along the nozzle width

Figure 8 presents the velocity vector contours adjacent to the exit in Section 1 when CR = 4, 2, 1.5, 1.2 and 1, respectively. The fixed scaling was set to make 1 mm of arrow equals 1 m/s. The velocity vector is calculated using measured axial and lateral components u and v. Close to the exit, the two jets are symmetric with respect to the own centerline at all tested CRs. There is a negative pressure region between the jets as fluid is entrained into the shear layers. Therefore, the two jets attract each other. Ejecting from the exits with sudden contraction, the jets expand rapidly and the negative pressure region is relatively small. There are two small recirculation areas between the two nozzles, as shown in Fig. 8(a). Not far from the exit, the two jets begin to merge to and interact with each other. The cross area of the jets rapidly 11

increases. Moreover, under different CRs, the flow characteristics are very different, as shown in Fig. 8(b~e). For a small CR, the expansion of jets is weaker. When z = 30 mm (z is the distance from the exit) and CR ≥ 1.5 the two jets clearly move towards each other, but when CR = 1.2, the mutual influence is weak and the two jets nearly develop along the original direction. When the nozzle is free of sudden contraction (CR = 1), a long and narrow negative pressure region is found in the space between the two jets. The jet expansion becomes slower. The above differences are due to the different flow fields inside the nozzles. In case of a contraction exit, corner vortices are induced by the nozzle wall and the contraction surface [13-16], as shown in Fig. 6. The size of the vortices is affected by both CR and Reynolds number (Re). For a given Reynolds number, it increases with increasing values of CR. Also, for a given CR, the vortex size decreases with increasing Re [4]. In the present study, Re was fixed at 17500. When CR varied from 1 to 4, the vortex increased in size. Clearly, the angle between the streamline near the wall and the contraction surface, α, decreases and the streamlines near the wall are more curved. The exit with sudden contraction restricts the flow along the jet direction. The fluid is squeezed at the outlet, enhancing the radial direction extension of the jets. Hence, the merging process of the parallel jets is enhanced for a higher CR.

12

0 20 40

z,mm

60 80 100 120 140 0

20

40

60

80

100

120

140

x,mm

0

20

20

40

40

60

60

z,mm

z,mm

(a) CR = 4 0

80

80

100

100

120

120

140

140

0

20

40

60

80

100

120

0

140

20

40

60

0

0

20

20

40

40

60

60

80

100

120

120

140

140

40

60

80

120

140

120

140

80

100

20

100

(c) CR = 1.5

z,mm

z,mm

(b) CR = 2

0

80

x,mm

x,mm

100

120

140

0

20

40

60

80

100

x,mm

x,mm

(d) CR = 1.2

(e) CR = 1

Fig. 8 Time-averaged velocity vector contours in Section 1 at various CRs (m/s)

13

0

0

20

20

40

40

60

60

z,mm

z,mm

3.2.3 On velocity field in the section along the nozzle length

80

80

100

100

120

120

140

140

0

20

40

60

80

100

120

140

0

20

40

60

y,mm

0

0

20

20

40

40

60

60

80

100

120

120

140

140

40

60

80

120

140

120

140

80

100

20

100

(b) CR = 2

z,mm

z,mm

(a) CR = 4

0

80

y,mm

100

120

140

0

20

40

60

y,mm

80

100

y,mm

(c) CR = 1.5

(d) CR = 1.2

0 20 40

z,mm

60 80 100 120 140 0

20

40

60

80

100

120

140

y,mm

(e) CR = 1 Fig. 9 Time-averaged velocity vector contours in Section 2 at various CRs (m/s) 14

Figure 9 shows the velocity vector contours adjacent to the exits in Section 2. For the jets ejecting from straight nozzles, as shown in Fig. 9(e), they are basically free jets, with constant cross-section area at the exit. However, when the nozzles are with sudden contraction exits, the jet flow converges after issuing into the free space. Then at a certain location (as called the turning point), the flow stops converging and begins to expand. With the increase of CR, the flow converges more rapidly. The results show that the sudden contraction could strongly influence the flow field even when the outlet area is only reduced in a small amount (e.g., CR = 1.2). 4

4

3

3

Z=20 Z=53 Z=148

2

1 v,m/s

v,m/s

1 0

0

-1

-1

-2

-2

-3

-3

-4

-4

0

20

40

60

80 x,mm

100

120

Z=20 Z=58 Z=148

2

140

(a) CR = 1.5

0

20

40

60

80 x,mm

100

120

140

(b) CR = 2

Fig. 10 Lateral velocity profiles at various z locations Figure 10 presents the lateral velocity (v) distribution at different z locations when CR = 1.5 and 2.0. v is positive on the left hand side and negative on the right hand side of Section 1 at z = 20 mm, indicating the jet is converging. Along with the jet, the magnitude of v decreases. When CR = 2, the maximum v is higher than that when CR = 1.5. The flow converging roughly stops at z = 53 and 58 mm for CR = 1.5 and 2, respectively (v ≤ 1 m/s). At z = 148 mm, v ≤ 0 in the zone of 60 ≤ x ≤ 79 mm and positive in the zone of 79 ≤ x ≤ 98 mm. It means that v changes the direction. The 15

results indicate that at a larger CR, v at nozzle exit is higher. The flow converges more obviously and the turning point appears further away from the nozzle.

3.3 The effect on flow interaction of two parallel jets 3.3.1 On velocity field at different z locations

0 40 80

Z, mm

120 160 200 240 280 320 (a) CR = 1

(b) CR = 1.5

(c) CR = 4

Fig. 11 Velocity vector distributions at different z locations in Section 1 (m/s) The velocity vector distributions at different z locations are plotted in Fig. 11 when CR = 1, 1.5 and 4, respectively. The results represent the spatial development of the flow field. When CR = 1 (contraction free), the streamwise velocity is significantly greater than v. Thus, the jet expands slowly and the location of combined point is far from the nozzle. The two peaks in the velocity profile gradually decrease along the streamwise direction. At z ≈ 290 mm, they disappear and a single peak

16

presents. When CR > 1, the velocity vector in the middle jet zone points vertically downstream but in the outer zone, it points outwards. The phenomenon implies that at a given location v is higher when CR = 1.5 and 4 than that when CR = 1. The stronger expansion the jet results in a shorter merging length.

3.3.2 On the mean streamwise velocity profiles along the symmetry axis

Figure 12 presents the profiles of the mean streamwise velocity u along the symmetric axis of the two parallel jets for various CRs. The jet velocities are normalized with respect to the average exit velocity u0. The peak velocities locate at the centerlines of the two jets. The dotted lines represent the location of combined point. u/u0 increases up to the maximum at the combined point (Zcp) and then decreases as the jets proceed at different CRs. When z≥Zcp , u decays. As indicated in Fig. 10, Zcp = 7.5a, 11a, 12a, 20a and 29a for CR = 4, 2, 1.5, 1.2 and 1, respectively. The two jet flows merge faster at a larger CR. The maximum u/u0 is also different at different CR. The maximum u/u0 for CR ≥ 1.5 is greater than 0.5 and is higher than that for CR = 1.2 and 1. The contraction causes strong lateral movements of the flow near the nozzle exit. The squeezing effect along the jet accelerates the expansion of the flow so that the two jets merge into a single jet more quickly with more intensified momentum transfer for a higher CR.

17

0.7

CR=1, Zcp= 29a CR=1.2,Zcp= 20a

0.6

CR=1.5,Zcp= 12a

umax/u0

0.5

CR=2, Zcp= 11a CR=4, Zcp= 7.5a

0.4 0.3 0.2 0.1 0.0 0

5

10

15 z/a

20

25

30

Fig. 12 Mean streamwise velocity along the symmetry axis of the two parallel jets 35

Experiment Fitting curve

30

Zcp/a

25 20 15 10 5 0 0

0.2

0.4

0.6 0.8 1/CR

1

1.2

Fig. 13 Variation of the combined point with the contraction ratio Shown in Fig. 13, a regression analysis of the experimental data is performed to correlate the streamwise location of the combined point, Zcp/a, with CR. The empirical formula of Eq. (6) is obtained. Obviously Zcp/a decreases with the increase of CR, but not linearly. Zcp decreases sharply when CR varies from 1 to 1.5 and thereafter changes slowly. Z cp a

= 4.5exp(1.8 ´

1 ) CR

(6)

3.3.3 On the decay of the mean streamwise velocity along the nozzle centerline

The decay of the mean streamwise velocity u along the nozzle centerline is 18

plotted in Fig. 14. After the flow exits the nozzle, the initial flow area has a core flow region where viscosity and diffusion has no effect on the flow [17], so u remains fairly constant and equal to that at the nozzle exit. u then decays downstream from this core region. u/u0 is mainly influenced by the nozzle width, the flow curvature and the flow acceleration due to the negative pressure region adjacent to the nozzle plane. 1.2 CR=1 CR=1.5 CR=2 CR=4

1.0

u/u0

0.8 0.6 0.4 0.2

0

5

10

15

20

25

30

z/a

Fig. 14 Mean streamwise velocity along the nozzle centerline For all CRs, u/u0 decays sharply in the initial region and then it gradually decreases with the downstream distance. The sharp decay is due to the presence of the recirculation zone. The size and the strength of the recirculation zone differ for the flow with different CRs. Higher flow acceleration along the symmetry axis of the two parallel jets indicates a lower sub-atmospheric pressure zone. More fluid was then entrained into the area along the symmetry axis of the two parallel jets. Hence, the jet flow along the nozzle centerline downstream from the core area decayed much faster for the flow with the contraction.

19

3.3.4 On the turbulence characteristics

Figure 15 shows the profiles of the normalized streamwise urms/u0, calculated by Eq. (4) for CR = 1 and 2. Two potential cores can be seen near the nozzles with urms ≈ 0 for both CRs, indicating that U in these areas remains essentially constant. However, this core region is wider and longer for CR=2 than that for CR=1. This trend is consistent with the decay of U along the nozzle centerline analyzed in Section 3.3.2. It can be also seen urms is peak along both sides of the jet shear layers. The turbulence, which is produced in the shear layers, is transported by diffusion and convection to the jet centerline. Thus, the peak urms indicates the onset of the shear layer structures at the jet centerline. Both the expansion and converging of the peak urms for CR = 2 are faster than those for CR = 1, indicating higher flow curvature. In addition, when Z ≥ Zcp for CR = 2, the peak urms decreases, indicating that the main momentum transfer between the two parallel jets occurs in the merging region rather than in the combined region. 0

0.3

0.3

0

20

20

0.25

0.25

40

40

0.2

0.2 60

0.15

80 100

z,mm

z,mm

60

0.15

80 100

0.1

120

0.1

120

0.05

0.05

140 0

140

20

40

60

80

100

120

140

160

0

0

x,mm

20

40

60

80

100

x,mm

(a) CR = 2

(b) CR = 1

Fig. 15 urms/u0 profiles for Section 1 at various CRs 20

120

140

160

0

Fig. 16 shows urms/u0 in Section 1 at three different z’s for CR = 1 and 2. When z = 50 mm, the profile has four peaks in the inner and outer shear layers of each jet for both CRs. The two jets interact with each other, but have not combined yet. However, the maximum urms/u0 for CR = 2 is higher than that for CR = 1, reflecting the stronger fluctuations existing in this plane. As the flow proceeds downstream to z = 150 mm, for CR = 2, the two jets combines together to a single jet. No more four peaks exist and urms/u0 ≤ 0.2. At the same time, for CR= 1 the profile still shows the characteristics of two jets. At z = 250 mm, urms distribution for CR = 2 is more uniform and urms/u0 ≤ 0.15. For CR = 1, the two jets combine into one single jet and the four high peaks are reduced to two peaks with similar values. Thus, the urms distribution clearly reflects the mixing intensity of the two parallel jets. 0.3

Z=50mm

0.2 0.1

urms/u0

0.0 0.3

CR=1 CR=2

0.2

Z=150mm

0.1 0.0 0.3

Z=250mm

CR=1 CR=2

0.2 0.1 0.0

0

20 40 60 80 100 120 140 160 x,mm

Fig. 16 urms/u0 distributions at different z locations in Section 1 (unit: m/s) The lateral turbulence intensities, vrms/u0, for CR = 2 and CR = 1 are shown in 21

Fig. 17. A relatively low-intensity region appears in the core region. For CR = 2, the peak vrms occurs on the symmetric plane between z = 30 mm and z = 100 mm, locating in the merging region between the two parallel jets. The lateral fluctuations along the symmetric plane gradually weaken when z > Zcp. The result implies that the main momentum transfer between the inner shear layers of the two parallel jets is completed. For CR = 1, the two jets merge over a longer distance and Zcp appears further away from the exit. The lateral fluctuations were not very strong on either side of the jet shear layers. The absolute value of urms differs from that of vrms at the same position in the flow field, indicating the heterogeneity of the turbulence in the two parallel jets and that there was little energy exchange in the lateral direction. 0

0.3

0

0.3

20

20

0.25

0.25

40

40

0.2

0.2

60 0.15

80 100

z,mm

z,mm

60

0.15

80 100

0.1

0.1

120

120

0.05

0.05

140

140 0

20

40

60

80

100

120

140

160

0

0

20

40

60

80

100

x,mm

x,mm

(a) CR = 2

(b) CR = 1

120

140

160

0

Fig. 17 vrms/u0 profiles in Section 1 for both CR.

3.3.5 On the Reynolds stress distribution

Fig. 18 shows the normalized Reynolds stress u 'v' contours for CR = 1 and 2.

Similar to the urms profiles in Fig. 15, peak values occur along the inner and outer shear layers in both jets. When CR = 2, high u 'v' distributes over a larger area than 22

for CR = 1 indicating that the momentum transfer is substantially higher for CR = 2. Furthermore, for both CRs,

u 'v'

> 0 on the outer shear layer of the left jet and the

inner shear layer of the right jet, and u 'v' < 0 negative on the inner shear layer of the left jet and the outer shear layer of the right jet. These are attributed to the velocity gradients in each jet. For CR = 2, the inner shear layers strongly interact as the two parallel jets interact and merge from z = 30 mm to z = 100 mm. u 'v' gradually decreases when z > Zcp as the original outer and inner shear layers gradually disappear. 0

0.03

0.03

0

20

20

0.02

0.02

40

40

0.01

0.01 60

z,mm

z,mm

60 0

80 100

-0.01

0

80 100

-0.01

120

120

-0.02

-0.02 140

140 0

20

40

60

80

100

120

140

-0.03

0

20

40

60

80

100

120

140

-0.03

x,mm

x,mm

(a) CR = 2

(b) CR = 1

Fig. 18 Normalized Reynolds shear stresses in Section 1 at various CRs.

4. Conclusions PIV measurements were conducted on the flow fields adjacent to the exit of two parallel rectangular jets with sudden contraction. Mean velocity vector contour, streamwise velocity variation, root mean square velocities and Reynolds stresses were studied under five contraction ratios. The results revealed that when the nozzles are with sudden contraction exits, the negative pressure zone between the two parallel 23

rectangular jets is significantly smaller. After exiting the nozzle, the jet flow first converges and then expands in the width direction. The contraction induces an earlier combination of the parallel jets. The combined point is closer to the exit at a larger contraction ratio. The location of the combined point decreases sharply when contraction ratio varies from 1 to 1.5 and thereafter changes slowly when contraction ratio increases from 1.5 to 4. The maximum u/u0 along the nozzle centerline increases when contraction ratio increases. The jet flow along the nozzle centerline downstream from the core area decays much faster for the flow with the contracted exits. Furthermore, the turbulence in the inner and outer shear layers is significantly stronger for the flow with the contraction. The measured results indicate that the contraction strongly promotes the merging of the two parallel jets and enhances the turbulence in the merging region. The parallel jet flows mix faster when the exits are with a higher contraction ratio.

Acknowledgements The supports provided by the Ministry of Science and Technology (#2015AA04B00) are highly appreciated.

References [1] E. Tanaka. The interference of two-dimensional parallel jets-Experiments on the combined flow of dual jet. Bulletin of JSME, 1974, 17(109):920-927. [2] H. H. Wang, S. Lee, Y. A. Hassan. Particle image velocimetry measurements of the flow in the converging region of two parallel jets. Nuclear Engineering and Design, 2015. 24

[3] T. M. Abdel-Salam, S. N. Tiwari. Mixing characteristics of two-dimensional parallel jets. AIAA 2004-1270. [4] C. Ozalp, A. Pinarbasi, M.S. Fakilar, B. Sahin. PIV measurements of flow through a sudden contraction. Flow Measurement and Instrumentation, 2007, 18(3):121-128. [5] D. R. Miller, E. W. Comings. Force-momentum fields in a dual-jet flow, Journal of Fluid Mechanics, 1960, 7(2): 237-256. [6] A. Nasr, J. C. S. Lai. Comparison of flow characteristics in the near field of two parallel plane jets and an offset plane jet. Physics of Fluids, 2001, 19(9):2919-2931. [7] E. A. Anderson, R. E. Spall. Experimental and numerical investigation of two-dimensional

parallel

jets,

Journal

of

Fluids

Engineering,

2001,

123(2):401-406. [8] N. Fujisawa, K. Nakamura, K. Srinivas. Interaction of two parallel plane jets of different velocities. Journal of Visualization, 2004, 7(2):135-142. [9] MELLING A. Tracer Particles and Seeding for Particle Image Velocimetry[J]. Measurement Science and Technology, 1997, 8(12): 1406. [10] R. C. Flagan, J. H. Seinfeld. Fundamentals of air Pollution Engineering[M]. Prentice Hall, Engle-wood, NJ, pp. 296–304. [11] R. J. Adrian. Particle Imaging Techniques for experimental fluid mechanics. Annual Review of Fluid Mechanics 1991, 23:261–304. [12] J. Westerweel. Efficient detection of spurious vectors in particle image velocimetry data sets. Experiments in Fluids, 1994,16:236–47. [13] T. M. Liou, C. F. Kao, S. M. Wu. The flow in a rectangular channel with sudden contraction and expansion. Chinese Institute of Engineers, 1987, 10(2):139-146. [14] I. Y. Chen, M. C. Chu, J. S. Liaw, C. C. Wang. Two-phase flow characteristics across sudden contraction in small rectangular channels. Experimental Thermal & Fluid Science, 2007, 32(2):696-706. [15] E. B. Christiansen, S. J. Kelsey, T. R. Cater. Laminar tube flow through an abrupt contraction. AIChE Journal, 2010,18(2):372–380 25

[16] J. M. Maia, D. Binding, Influence of elongational properties on the contraction flow of polyisobutylene in a mixed solvent, Rheologica Acta. 1999, 38(2): 160–171. [17] S R. Turns. An Introduction to Combustion. Concepts and Applications [M], WCB/McGraw-Hill Press, 2000.

0

CR=4

20

20

40

40

60

60

z,mm

z,mm

0

80

80

100

100

120

120

140

140

0

20

40

60

80

100

120

140

CR=1

0

20

40

60

80

100

120

140

x,mm

x,mm

Highlights: Ø Experimentally studied the flow characteristics of two parallel rectangular jets through exits with sudden contraction. Ø The sudden contraction remarkably enhances the mixing between the parallel jets. Ø The sudden contraction makes the parallel jets combined earlier. Ø The sudden contraction influences the turbulence characteristics of the jet flows.

26