Вы находитесь на странице: 1из 35

Elsevier Editorial System(tm) for Nuclear

Engineering and Design


Manuscript Draft

Manuscript Number:

Title: Steady State Flow Analysis of Two-Phase Natural Circulation in


Multiple Parallel Channel Loop

Article Type: Full Length Article

Corresponding Author: Prof. J B Joshi,

Corresponding Author's Institution: Institute of Chemical Technology

First Author: Vishal H Bhusare, M Tech

Order of Authors: Vishal H Bhusare, M Tech; Ritesh K Bagul; J B Joshi;


Arun K Nayak; Umasankari Kannan; D S Pilkhwal; P K Vijayan

Abstract: In this work, steady state flow analysis has been carried out
experimentally in order to estimate the liquid circulation velocities and
two-phase pressure drop in air-water multichannel circulating loop.
Experiments were performed in 15 channel circulating loop. Single phase
and two-phase pressure drops in the channels have been measured
experimentally and have been compared with theoretical model of Joshi et
al. (1990). Experimental measurements show good agreement with model.
Cover Letter

Dear Dr. Jason Chao,

I am enclosing manuscript entitled “Steady state flow analysis of two-phase

natural circulation in multiple parallel channel loop”. In this paper multiple parallel channel loop,

consisting of riser-air-water separation drum-downcomer, similar to that of main heat transport

system of the advanced nuclear reactors is studied. The experimental measurements were

performed in order to measure liquid circulation velocities in the channels. A systematic

investigation for a case of 15 channels (risers) is done. The liquid circulation was investigated

over a wide range of superficial gas velocities and liquid level in the air-water drum. For all these

cases, the force balance or energy balance across the loop helped in the development of the

model for prediction of circulation velocities. Experimental measurements show good agreement

with model predictions.

I should be grateful to you if you could kindly consider it for publication in the Nuclear

Engineering and Design.

With kind regards

J. B. Joshi
Emeritus Professor and J. C. Bose Fellow
Homi Bhabha National Institute,
Anushakti Nagar, Mumbai 400 094
INDIA
Tel: +91-22-33612106/+91-22-25597625
Highlights (for review)

Highlights

1. Liquid circulation velocity increases with increasing superficial gas velocity.

2. Total two-phase pressure drop decreases with increasing superficial gas velocity.

3. Channels with larger driving force have maximum circulation velocities.

4. Good agreement between experimental and model predictions.


*Title Page

Steady State Flow Analysis of Two-Phase Natural Circulation in

Multiple Parallel Channel Loop

V. H. Bhusare,1,2 R. K. Bagul,2 J. B. Joshi,1,3,* A. K. Nayak,1,2 Umasankari Kannan,1,4 D. S.

Pilkhwal,2 P. K. Vijayan1,2

1. Homi Bhabha National Institute, Anushaktinagar, Mumbai 400094, India

2. Reactor Engineering Division, Bhabha Atomic Research Centre, Trombay, Mumbai


400085, India

3. Department of Chemical Engineering, Institute of Chemical Technology, Matunga,


Mumbai 400019, India

4. Reactor Physics Design Division, Bhabha Atomic Research Centre, Trombay, Mumbai
400085, India

______________________________________________________________________________
*Author to whom correspondence may be addressed
Email address: jbjoshi@gmail.com
*Manuscript

Abstract

In this work, steady state flow analysis has been carried out experimentally in order to estimate

the liquid circulation velocities and two-phase pressure drop in air-water multichannel circulating loop.

Experiments were performed in 15 channel circulating loop. Single phase and two-phase pressure

drops in the channels have been measured experimentally and have been compared with theoretical

model of Joshi, J. B., Ranade, V. V., Gharat, S. D., Lele, S. S., 1990. Sparged loop reactors. The

Canadian Journal of Chemical Engineering, 68, 705-741. Experimental measurements show good

agreement with model.

Key words: Natural circulation, multichannel, circulation velocity, two-phase flow, pressure drop.
2

1. Introduction

Many of the chemical, biochemical, pharmaceutical, agricultural, and nuclear industries find

applications of the same principles as that of the natural circulation loop (NCL). In natural circulation,

flow is mainly developed due to the buoyancy force, which is due to the difference in densities of the

two-phase fluid in riser and single phase fluid in downcomer. Two-phase flow in NCL can be obtained

either by injection of gas into the riser (most commonly airlift loop reactors) or it can be generated

within by boiling the liquid (nuclear reactors). Of all these applications natural circulation shows great

promise to the nuclear energy sectors now a days and can be a design criterion. In nuclear reactors,

where the steam is generated in a core and two-phases, i. e., water and steam are carried to the steam

drum through the riser channels or tail pipes. In steam drum, steam is separated from water and sent for

electricity generation. The separated water from steam drum is sent back to the core again through the

downcomer channels. Thus, there is a density difference between the riser and downcomer which leads

to the gravity driven circulating flows in the loop between core and steam drum. Thus, the gravity

driven flows are generated in the loop. Circulation flow velocities in the reactor are very low to

moderate, which is very much essential for the safe and efficient operation of reactor.

In natural circulating loops, introduction of gas phase to the riser gives rise to the driving force,

which establishes circulation in the loop. For steady state circulation in the loop, driving force is

balanced by the resistive force. Resistive force mainly includes frictional force, losses in straight part

of the riser and downcomer and the other pipe fittings such as bends, expansions and contractions etc.

An approach for liquid circulation velocity calculation based on the energy balance over the loop is

developed by many researchers (Chisti et al., 1988; Joshi et al., 1990; Mohamed et al., 1998). The

reactor height plays important role in prediction of circulation velocity. The circulation velocity is

increased by increasing gas-liquid dispersion height to the power of 0.5. Mohamed et al. (1998) have
3

measured liquid circulation velocities in pilot scale air water loop. The experimental results were

validated by using the model of Chisti et al (1988). They also characterized the different regimes, i.e.,

homogeneous regime, heterogeneous regime and transition regime from the behavior of local and mean

gas hold-up against superficial gas velocity in the riser.

Joshi et al. (1990) have developed a comprehensive model over a wide range of superficial gas

velocity in the riser (VGR) as well as downcomer (VGD), height to diameter (riser) ratio, area ratio of

riser to downcomer and the role of pipe fittings. They have shown excellent agreement between the

model predictions and all the data reported in the published literature.

Lele et al. (1993) have studied the hydrodynamics and mass transfer in airlift loop reactors. It

has been observed that the performance of the loop is mainly dependent of liquid circulation velocity,

which is a function of driving head for liquid circulation and hence critical sparging location. The

effects of various design parameters, such as area ratio of riser to downcomer, height to diameter ratio

of the riser, riser and downcomer gas flow rates on the critical sparging location and total mass transfer

rate for slow and fast reactions are studied.

Choi (1996) performed experimental measurements of liquid circulation velocity in external

loop airlift reactor by using tracer injection techniques. He studied the effect of horizontal connection

length, ratio of downcomer to riser cross-sectional area, superficial gas velocity on the circulation

velocity. It has been observed that, as the horizontal length between the two legs of loop is increased, it

gives better gas-liquid separation. Hence, increasing the driving force, which increases the circulation

velocity. Increasing area ratio for downcomer-to-riser decreases the liquid circulation velocity due to

increase in frictional resistance to flow. Choi obtained a correlation for circulation liquid velocity

relating all the above parameters.

Couvert et al. (1999) have experimentally characterized the air water loops by studying axial

and radial local gas holdup, global gas holdup, liquid velocity and bubble diameter. They have also
4

proposed a four cell mathematical model based on hydraulic laws for gas-liquid system. Airlift reactor

is considered as four cells which are riser, downcomer, gas-liquid separator and bottom of airlift

reactor. It solves mass and momentum balance equations for riser and downcomer. It also includes a

closure for slip velocity and pressure drops at the top and bottom of the reactor. Felice (2005) presented

basic approach to calculate the circulation rates in air water loops for two and three-phase systems

based on pressure balance, where the difference in densities between two vertical columns of airlift

reactor provides the driving force for liquid circulation and under steady it is balanced by the losses

due to the flow itself. He proposed analytical model to relate circulation rates with gas flowrate. The

liquid circulation rate was found to increase exponentially with an increase in the gas flowrate because

it is a function of gas holdup, which increases in an exponential manner. In case of three-phase system,

introduction of solid phase reduces liquid circulation rate in the loop because of frictional energy

dissipation of the solid-liquid interface. Jones et al. (2010) have performed experimental studies in

order to understand the hydrodynamics of external loop airlift reactors with modified downcomer.

They have considered two different downcomer configurations and studied effects on gas hold up and

superficial velocities in both riser and downcomer. Basu et al. (2013) have performed experimental

studies on the hydrodynamics of natural circulation advanced heavy water reactor (AHWR). Scaled-

down AHWR prototype was fabricated as an experimental facility with one steam drum, two risers,

one downcomer and one gas sparging chamber. Experiments were performed by using air and water as

test fluids. Two phase flow was generated by injection of gas. The superficial air velocity was varied

from 1.27 m/s to 6.8 m/s. At very low gas flowrate, the dispersion was in the bubbly flow regime. As

the gas flowrate increases the gas hold-up also increases and the dispersion passes through sludge to

churn to annular flow regimes. Further, the extent of entrainment (carryover) also increases with an

increase in the superficial gas velocity. In all the flow regimes, it was observed that the liquid

circulation velocity is totally dependent on the gas flow rate in the column. As the gas flow rate in the

column is increases, the liquid circulation velocity increases upto certain gas flow rate and then falls
5

with a further increase in the gas flow rate. This is the case because the slip velocity between two

phases is much higher at higher gas flow rates. Gas slips so fast that the amount of liquid driven with it

is very low and hence, liquid circulation velocity reduces.

From the foregoing discussion, it is clear that the liquid circulation pattern has been understood

over a wide range of superficial gas velocity (VG), height to diameter ratio (H/DR) and area ratio

(AR/AD) riser to downcomer. For all these cases, the force as well as energy balance have also been

well understood. However, very scanty information is available for a case where the riser comprises of

multiple channels (at least 10). Therefore, it was thought desirable to undertake a systematic

investigation for a case of 15 channels as a riser. The liquid circulation was investigated over a wide

range of VG and liquid level in the air-water drum.

2. Experimental setup and procedure

2.1 Experimental setup

In this work, Figure 1 shows the isotropic view of experimental setup for air-water

multichannel circulating loop. It consists of a water storage tank (1), 15 tail pipes (62.7 mm ID) (2),

air-water drum (3), one downcomer (97.15 mm ID) (4), and 15 air injection lines (5). Water storage

tank has dimensions of 1.2 m × 0.925 m × 2.5 m. Front wall of the tank has transparent window (6) of

acrylic glass in order to observe the water level in the tank. Water level in the tank is measured by

using level transmitters based on hydrostatic pressure developed by liquid column. Water from the

storage tank is carried to air-water drum through tail pipes. 15 channels or tail pipes are categorized

into 3 types based on their specifications, i.e., the outermost channels (channels no. 1, 2, 3, 4, 5) are

categorized as type I, the innermost channels (channel no. 11, 12, 13, 14, 15) are categorized as type III

and the intermediate channels (channel no. 6, 7, 8, 9, 10) are categorized as type II. The channel

arrangements in the loop are as shown in Figure 1 and their specifications are provided in Table 1.

Each tail pipe has three different sections, i.e., vertical down flow section, single phase horizontal
6

section and vertical rising section. Each tail pipe has a transparent section (7) to visualize the two-

phase flow. Air is injected near the bottom of vertical rising section of the tail pipes. Two-phase, air-

water, mixture rises upwards and enters the air-water drum. Tail pipes are grouped into three sets based

on their dimensions and each consisting of 5 tail pipes. The details of each channel specification are

given in Table 1. Air-water drum is provided with 15 tail pipe connections. Further, the drum has three

plane surfaces (8) and one curved circular surface arc (9) as shown in Figure (1). The two plane

adjoining sides are made of transparent material and the third plane side and curved surface are made

of SS. In air-water drum, two-phase mixture is separated, air being vented to atmosphere (10) and

water is returned (11) to storage tank through downcomer. The total loop occupies a space of 4 m × 7

m × 8 m (floor area × height). Loop is operated at 2 bar pressure and atmospheric temperature. Single

phase and two-phase pressure drops in the loop are measured by using low range differential pressure

transmitters. Different parts of air-water loop and the flows in it are represented schematically in Figure

2.

2.2 Experimental procedure

Initially, complete loop is filled with water. Constant water level 500 mm is maintained in an

air-water drum and storage tank. Now, air is sparged at a constant flow rate of 0.833 L/s through

sparger pipe at the bottom of each riser arm of 15 tail pipes (locations shown in Figure 2 and Table 1).

This air injection gives rise to two phase flow in the riser section and hence a driving force for liquid

circulation. The two phase flow in the 15 riser enters the air-water drum (3). The air and water get

separated. The water returns back to the storage tank (1) through downcomer and then enters the tail

pipes including the riser and finally into drum, thus completing the circulation loop. Once, an air-water

drum and storage tank levels come to steady state, measurements are performed for the two-phase

pressure drops in riser section of tail pipes, single phase liquid velocities and single phase pressure

drop in the horizontal section of the tail pipes, dispersion (swell) level in the air-water drum and level

in the water storage tank. Loop is allowed to operate at this condition for few minutes and data are
7

recorded by using data acquisition system. Then keeping the initial water levels of air-water drum and

storage tank constant, air flow rate is varied in the range of 0.833 L/s to 5 L/s and the complete

procedure is repeated at a constant liquid height. After completing the experiments for all the gas flow

rates, the liquid level in the air-water drum is changed to three additional levels of 600, 700 and 800

mm. Since the flow in the horizontal pipe may not be fully developed, the flow rate versus pressure

drop was calibrated. The flow rate was measured by using ultrasound flow meter and the pressure drop

by differential pressure cell across horizontal section lengths of the tail pipes.

3. Modeling

In this study a model developed by Joshi et al. (1990) was employed to predict the overall

liquid circulation velocities in the loop. This model is based on the force balance across the loop. The

driving force in the circulation is mainly due to the density difference between the riser and the

downcomer. The resistive forces in the loop are due to the friction in the pipelines and pipe fittings

such as contraction, bends, expansions, etc.

Since the riser heights are different for the three groups of channels, the method of calculation

has been written for Type I channels (refer Figure 2). The same methodology is applicable for

calculation of circulation velocity for other channels. H1 is the height of dispersion inside the drum. H2

is the riser height above the point of injection and upto the entrance of the drum. H4 is the height of the

water level in the tank above point A and H5 is the height of channel below point A as shown in Figure

2. The static pressure at point A (due to the single phase liquid) is given by,

PA  H 4  L g (1)

The static pressure at point B (due to the two-phase dispersion) is given by:

PB  H1 L ,DR  L  G ,DR  G g  H 2 LR  L  GR G g (2)

Therefore, the pressure driving force for the liquid circulation for Type I channels is given by:
8

PA  PB  H 4  L g  H1 L ,DR  L  G ,DR G g  H 2 LR  L  GR G g 

PA  PB  H 4  H1 L ,DR  H 2 LR  L g  H1 G ,DR  H 2 GR  G g (3)

It is observed that there is a carry-under of gas in the section D-M of downcomer. Due to the presence

of gas in section D-M, there is reduction in driving force equivalent to GD  L  G gH DM  across

section DMNE. Hence, while calculating actual driving force for circulation, driving force of section

DMNE has to be subtracted from Eq. (3). Thus, the actual driving force is given as:

PA  PB  H 4  H1 L,DR H 2 LR  L g  H1 G ,DR  H 2 GR  G g  GD  L   G gH DM (4)

In Eq. (4), gas holdup in the section D-M of downcomer GD is calculated by taking phase mass

balance given by following equation.

VG V L
VS    (5)
G L

Where, VS is the slip velocity between gas and liquid and for gas-liquid systems it is around 0.23. In

Eq. (5), VS is considered positive in the upward direction since the natural bubble velocity is in the

upward direction. VG is considered positive when it flows in the direction of VS, whereas VL is

considered positive when it is in the reverse direction of VS. Since there is carry-under of gas with

liquid in section D-M, the gas velocity is considered zero in this section.

The resistive force in the loop comprises of the frictional losses along the loop TQRSDMNET

(pressure drop in the tank has been assumed to be negligible), which includes both single phase

frictional losses (along path T-Q, Q-R, R-B, D-M, M-N, and N-E) and two phase frictional losses along

path B-S (again, pressure drop in the drum has been assumed negligible). Resistive force also consists

of losses due to the pipe fittings like bends (at Q, R, M, and N), losses due to area expansion at the

entrance of air-water drum (at S) and water storage tank (at E), losses due to area contraction at the exit
9

of the air-water drum (at D) and water storage tank (at T). Summation of all these forces will give the

total resistive force along the loop.

The single phase frictional pressure drop along the path T-Q is given by,

PW SP,TQ  2 f SP  LVL H 3


2
(6)
DR

Similarly, the single phase frictional losses in other sections (Q-R, R-B, M-N) of the loop are given by:

PW SP,QR  2 f SP  LVL L1


2
(7)
DR

PW SP,RB  2 f SP  LVL H D1


2
(8)
DR

2 f SP  LVL2 L2
PW SP,MN  (9)
DD

In above equations, single phase friction factor, f SP , is calculated from friction velocity as follows:

12
 f 
V  VL  SP 
*

 2 

2
V * 
f SP  2  (10)
 VL 

where, the physical significance of friction velocity V * for the single phase flow is known to be equal to

the fluctuating velocity in the transverse direction.

The two-phase frictional pressure drop is calculated by using two-phase friction multiplier (  L ).
2

The two-phase friction multiplier is defined as the ratio of the frictional pressure drop for two-phase

flow to that of the single phase liquid flow and in section B-S (Figure 2) it is given by:
10

PW TP ,BS


L2  (11)
PW SP,BS

In the two-phase region (section B-S and air-water drum), the presence of bubbles increase the pressure

drop. The first reason for the increase is because of the increase in liquid velocity in the presence of

bubbles. Also, the level of turbulence is much higher as compared to that of the single-phase flow. This

additional turbulence increases the friction factor due to the corresponding increase in the friction

velocity. Thus the overall pressure drop in two-phase region is the combined effect of the increased

friction factor and the true liquid velocity. Thus, the two-phase frictional pressure drop along the path

B-S is given by Gharat et al. (1992):

PW TP,BS  PW 'AT ,BS  PW SP,BS (12)

where, PW SP,BS is the single-phase pressure drop (in the section BS) is given by:

PW SP,BS  2 f SP  LVL H 2


2

(13)
DR

and, PW 'AT ,BS is the additional pressure drop due to the turbulence generated by bubbles and is given
by:

P  '


2 f '  L H 2 VL LR 
2

(14)
W AT , BS
DR

'
where, the friction factor due to additional turbulence, f , is calculated in the same way as that given in

Eq. (10) based on true velocity,

2
V ' 
f  2 
'
(15)
V 
 S
11

Single phase frictional pressure drop for the flow of liquid in section B-S is given by,

2 f SP  LVL2 H 2
PW SP,BS  (16)
DR

From Eq. (11) to (14) and Eq. (16), we get the two-phase friction multiplier given by,

2
1 2 U ' 
  2 
2   (17)
L
L f SP V 
 L 

Thus, from Eqs. (11), (16) and (17), the single phase frictional pressure drop is multiplied by two-phase

friction multiplier in order to get the two-phase frictional pressure drop. For the two phase flow, U ' can

be calculated as given by Joshi et al. (1990):

U '  1.5 G VS (For homogeneous regime) (18)

1
   V  3
U '  0.25 gDR  VG  G L  G VS  (For heterogeneous regime) (19)
  L 

where, VS is the slip velocity and is given by Eq. (5).

Resistive forces also includes the other losses in the circulation loop, such as losses in 90ᵒ elbows in the

pipe, losses due the flow area contraction and flow area expansion. These were estimated according to

the procedure described by McCabe et al. (1993).

VL2
hf  k f (20)
2

In Eq. (20), kf is the frictional coefficient. For different types of losses the different values of kf are

given in Table 2. Thus, the total resistive force along the loop TQRSDMNET is given by:
12

 PW SP,TQ  h fb,Q  PW SP,QR  h fb,R  PW SP,RB  PW TP ,BS 
 
Resistive force =   h fex,S  h fcon,D  PW SP,DM  h fb,M  PW SP,MN  h fb, N  (21)
 
  PW   h fex,E  h fcon,T 
 SP, NE 

Under steady state conditions, equating the driving force [Eq. (4)] and the total resistive force [Eq.

(21)] gives the value of average circulation velocity.

4. Results and discussion

The liquid circulation velocities, single phase pressure drop and two-phase pressure drop were

measured experimentally in the multichannel loops and are compared with the model predictions [Eq.

(1) to (21)].

The effect of superficial gas velocity on the pressure drops (both two-phase and single-phase),

and liquid circulation velocity in the loop are shown in Figures 4 to 6, Figure 8 and Figure 10,

respectively. Figure 3 shows the schematic of the two-phase gas-liquid dispersion section Bꞌ-Sꞌ, where

experimental measurements were performed to measure total two-phase pressure drop. It includes both:

the two-phase hydrostatic head and two-phase frictional pressure drop. The effect of superficial gas

velocity on the two-phase pressure drop is shown in Figures 4 to 6. It is observed that the total pressure

drop in the riser decreases with an increase in the superficial gas velocity. Since the superficial gas

velocity increases the two-phase frictional pressure drop in the channel increases [calculated in similar

way as in Eq. (11), (16), and (17)], however, the hydrostatic head decreases because of increase in gas

hold-up. Thus, the total pressure drop decreases as per the following equation.

PBS  H 6 LR  L  GR G g  PW SP,BS L2,BS (22)

All the channels in type I, since they have same specifications, behave in the same manner as

shown in Figure 4A-4D at initial drum levels of 0.8 m, 0.7 m, 0.6 m and 0.5 m, respectively. Similar

results were observed for channels type II and type III and are shown in Figures 5 and 6, respectively.
13

Two-phase lengths of the type I channels are longer than those in type II and type III as given in Table

1. The driving head for channels in type I is more than that of type II and driving head for channels of

type II is more than that of type III as given in Table 1. This causes the reduction in total two-phase

pressure drop in the channels of type II compared to type I and further reduction in total pressure drop

in type III at constant superficial gas velocity. It is observed that there is no significant change in

pressure drops with varying the initial air-water drum levels. Since configurations and operating

conditions are same for every channel of the respective type, they must behave in a same manner.

Hence, here onwards the behavior of each channel is represented by the channel type itself.

For the estimation of total pressure drop (frictional as well as static head), the knowledge of gas

hold-up and liquid circulation velocity are needed. In order to obtain these, the solution procedure as

shown in Figure 7 is used. This procedure is explained in stepwise manner as follows:

1. The channel gas hold-up ( G ) and superficial liquid circulation velocity (VL) are assumed.

2. Eq. (4) and Eq. (21) are solved to obtain the driving force and resistive force, respectively. An

iterative loop is operated for VL at constant gas hold-up to get steady state circulation velocity at

driving force equals to resistive force.

3. Obtained steady state liquid circulation velocity and gas hold-up are used to calculate two-

phase

pressure drop across section Bꞌ-Sꞌ. In order to validate the values of VL and G calculated two-

phase pressure drop is compared with experimental data.

4. If calculated data do not show much deviation from experimental data, then the validated VL

and G are steady state operating values. Else another iterative loop is operated for new value of

G , and steps (1) to (3) are repeated.

In Figure 8, driving forces and resistive forces were calculated by using the methodology

provided in Figure 7 at different liquid superficial velocities and are plotted against superficial liquid
14

velocity for channel type I at gas flow rate of 0.833 L/s. As the liquid velocity increases, the gas holdup

in the riser channel decreases as liquid carries more and more gas with it. As a result, the driving force

for liquid circulation decreases. On the other way, as liquid superficial velocity increases, the resistive

forces increases because of increases frictional losses. Driving force and resistive force when plotted

against liquid superficial velocity, the point of intersection gives the actual liquid circulation velocity

under steady state operation. Similar to this the driving and resistive forces are calculated for all the

channels by varying the gas flow rates from 0.833 L/s to 5 L/s to get the liquid circulation velocity

under steady state operation. Figure 9 shows the liquid circulation velocity in the channel type I, type II

and type III. It can be seen that the liquid circulation velocity increases with an increase in the

superficial gas velocity. As superficial gas velocity in the channel increases, the gas holdup in the

channel also increases. This increased gas hold up reduces the mixture density in the channel i.e.,

produces more driving force as per Eq. (4). It is also observed that the channel type I has maximum

liquid circulation velocity as compared to other channels since it has higher two-phase head and hence

the driving force, while liquid circulation velocity for channel type III is minimum as it has lowest two-

phase driving force. In Figure 9, the experimental values, represented by marker symbols, are plotted

against superficial gas velocity. An excellent agreement (Standard Deviation = 2%) can be seen

between the experimental measurements and model predictions.

In Figure 10, experimental two-phase pressure drop in the section Bꞌ-Sꞌ, at air-water drum level

of 0.8 m, is compared with model predictions. The modeled two-phase pressure drop is calculated by

using Eqs. (11), (16) and (17) for the section Bꞌ-Sꞌ. Actual liquid circulation velocity and gas hold-up

obtained by using the methodology explained in Figure 7, are used to calculate the modeled two-phase

pressure drop. Results are compared for channel type I, type II and type III. The model predictions

show good agreement (Standard Deviation = 1%) with the experimental measurements.

The effect of superficial gas velocity on single phase pressure drop in the horizontal section Qꞌ-

Rꞌ of the tail pipe is shown in Figure 11. Figure 11A shows the single phase pressure drop for initial
15

air-water drum level of 0.8 m. For section Qꞌ-Rꞌ the single phase frictional pressure drop is calculated

by using following equation:

2f SP  L VL2 L Q'R '


PW SP ,Q R
' '  (23)
DR

An excellent agreement can be seen between the experimental and the calculated pressure

drops. Similar comparisons for experimental data and model predictions are made at various initial

drum levels of 0.7 m, 0.6 m and 0.5 m. and are shown in Figure 11B, 11C, and 11D respectively. All

the experimental data shows good agreement (Standard Deviation = 1%) with the model predictions.

5. Conclusion

The experiments were performed in a multichannel air-water loop. It is observed that the single

phase pressure drops, liquid circulation velocities and hence liquid circulation rates, gas holdup in the

channel increases with increase in gas flow rate. Whereas, there is reduction in total pressure of the

riser as two-phase flow increases with increase in the gas flow rate. Initial air-water drum level does

not have significant influence on the liquid circulation rates and pressure drops (both single phase and

two-phase). Experimental data shows good agreement with the estimated results.

Nomenclature

A Cross-sectional area (m2)

D Diameter (m)

f Friction factor

'
f Friction factor based on slip velocity due to additional turbulence

H1 Two-phase dispersion height in air-water drum (m)

H2 Two-phase dispersion height in riser (section B-S) (m)

H3 Height of riser or channel (section T-Q / S-R) (m)


16

H4 Driving height for single phase leg (Tank and section T-A) (m)

hf frictional loss due to the pipe fittings (N/m2)

kf friction coefficient

L1 Length of horizontal section of channel (m)

L2 Length of horizontal section of downcomer (m)

P Pressure (N/m2)

PW Frictional pressure drop (N/m2)

U' Fluctuating velocity (m/s)

V Superficial velocity (m/s)

VS True velocity or slip velocity (m/s)

Greek Letters

ρ Density (kg/m3)

 Phase hold-up

Subscripts and Superscripts

AT Additional turbulence because of bubbles

b Bend

con Area of contraction

D Downcomer

ex Area of expansion

G Gas phase

L Liquid phase

R Riser

SP Single phase

TP Two phase
17

Table 1. Configurations of channels in the air-water loop


Internal Channel Height above Horizontal Length for
Type Channel Nos. Diameter, Height, Air Injection Length, Section
D (m) H3 (m) point, H2 (m) L1 (m) Qꞌ-Rꞌ (m)
I 1,2,3,4,5 0.0627 3.45 2.83 4.1 3.2
II 6,7,8,9,10 0.0627 3.08 2.46 3.3 2.4
III 11,12,13,14,15 0.0627 2.65 2.03 2.9 2

Table 2. Frictional coefficient (kf) for pipes and fittings in air water loop (McCabe et al. [11])

Type Frictional coefficient (kf)


2
 A 
Area expansion at drum 1  R 
 ADrum 
 A 
Area contraction at tank 0.41  R 
 ATank 
2
 A 
Area expansion at tank 1  D 
 ATank 
Area contraction at  A 
0.41  D 
drum  ADrum 
90ᵒ Elbow 0.75
18

Figure 1 The experimental setup for multichannel air-water circulating loop

Figure 2 Schematic of air-water loop (AWL)

Figure 3 Schematics of experimental two-phase pressure drop measurement section

Figure 4 Effect of superficial gas velocity on experimental two-phase pressure drop in

type I channels at initial drum level of 0.8 m (A), 0.7 m (B), 0.6 m (C) and 0.5

m (D)

Figure 5 Effect of superficial gas velocity on experimental two-phase pressure drop in

type II channels at initial drum level of 0.8 m (A), 0.7 m (B), 0.6 m (C) and 0.5

m (D)

Figure 6 Effect of superficial gas velocity on experimental two-phase pressure drop in

type III channels at initial drum level of 0.8 m (A), 0.7 m (B), 0.6 m (C) and 0.5

m (D)

Figure 7 Solution methodology for calculation of liquid circulation velocity

Figure 8 Liquid circulation velocity prediction

Figure 9 Effect of gas flow rate on liquid circulation velocity

Figure 10 Comparison of experimental two-phase pressure drop with model predictions

Figure 11 Comparison of experimental single phase pressure drop with model predictions

at initial drum levels of 0.8 m (A), 0.7 m (B), 0.6 m (C) and 0.5 m (D)
19

(10)
1: Water Storage Tank
(3) 2: Tail Pipes / Channel
(9)
3: Air–Water Drum
(8)
4: Downcomer
(8)
(11) 5: Air Injection Lines
(7) (1)
(5) 6: Acrylic Glass Window
(5) (6)
7: Tail Pipe Transparent Section
8: Air-Water Drum Plane Surfaces
9: Air-Water Drum Curved Surface
10: Air Venting to Atmosphere
(4)
11: Water Return Line to Storage
Tank

(2)

Figure 1
20

(10) Separated
Air

(3)
(1)
Water Tank
Air-Water
H1 Drum

S Downcomer
D (4) E T
(11) Water
M N
L2 H4
Air-Water
H2 Water
H3

Tail Pipe / Channel (2)


B A
Air Water
Injection Rꞌ Qꞌ H5
(5) R Q
L1

Figure 2
21

Sꞌ

Two-Phase Gas-
Liquid Dispersion
Liquid
Water
H6 Channel

∆PTP, EXP Bꞌ
X Bꞌ

Figure 3
Experimental Two-Phase Pressure Drop, ΔPTP,EXP (N/m2)

Figure 4
Superficial Gas Velocity, VG (m/s)
22
Experimental Two-Phase Pressure Drop, ΔPTP,EXP (N/m2)

Figure 5
Superficial Gas Velocity, VG (m/s)
23
Experimental Two-Phase Pressure Drop, ΔPTP,EXP (N/m2)

Figure 6
Superficial Gas Velocity, VG (m/s)
24
25

Assume G

Assume VL

Resistive
Force = NO
Driving
Force

YES

Steady state G and VL

NO P TP ,EXP


 P TP ,CAL

YES

Final G and VL

Figure 7
26

Figure 8
27

Figure 9

Figure 10
28

Single Phase Pressure Drop, ΔPSP (N/m2)

Superficial Gas Velocity, VG (m/s)


Figure 11
29
*Bibliography
Click here to view linked References

References

Basu, D. N., Patil, N. D., Bhattacharyya, S., Das, P. K., 2013. Hydrodynamics of a natural

circulation loop in a scaled-down steam drum-riser-downcomer assembly. Nuclear Engineering

and Design, 265, 411-423.

Chisti, M. Y., Halard, B., Moo-Young, M., 1988. Liquid Circulation in airlift reactors. Chemical

Engineering Science, 43, 451-457.

Choi, K. H., 1996. Circulation liquid velocity in external-loop airlift reactors. Korean Journal of

Chemical Engineering, 13, 379-383.

Couvert, A., Rousten, M., Chatellier, P., 1999. Two-phase hydrodynamic study of a rectangular

air-lift loop reactor with an internal baffle. Chemical Engineering Science, 54, 5245-5252.

Felice, R. D., 2005. Liquid Circulation rates in two-and three-phase external airlift reactors.

Chemical Engineering Journal, 109, 49-55.

Gharat, S. D., Joshi, J. B., 1992 (a). Transport phenomena in bubble column reactors –I: Flow

pattern. The Chemical Engineering Journal, 48, 141-151.

Gharat, S. D., Joshi, J. B., 1992 (b). Transport phenomena in bubble column reactors –II:

Pressure drop. The Chemical Engineering Journal, 48, 153-166.

Jones, S. T., Heindel, T. J., 2010. Hydrodynamic considerations in an external loop airlift reactor

with a modified downcomer. Ind. Eng. Chem. Res., 49, 1931-1936.

Joshi, J. B., Ranade, V. V., Gharat, S. D., Lele, S. S., 1990. Sparged loop reactors. The Canadian

Journal of Chemical Engineering, 68, 705-741.

Lele, S. S., Joshi, J. B., 1993. Hydrodynamics and mass transfer in airlift reactors. Encyclopedia

of Fluid Mechanics Supplement-2, Cheremisinoff, N.P. ed., Gulf Pub. Co., USA, 409-510.
McCabe, W. L., Smith, J. C., Harriott, P., 1993. Unit Operations of Chemical Engineering: Fifth

Edition. McGraw Hill, 104.

Mohamed, M. E., Narsingh, U., Rouillard, A. E., Judd, R., 1998. Hydrodynamic flow regimes,

gas holdup and liquid circulation in airlift reactors. Ind. Engg. Chem. Res., 37, 1251.

Вам также может понравиться