Upload
h5349h
View
251
Download
1
Tags:
Embed Size (px)
DESCRIPTION
Supreme
Citation preview
Vienna, Austria, September 10-13, 2007
BARCBARCBARCBARC
Thermo-fluid dynamics and pressure drops in various geometrical configurations
M.R. Gartia, P.K. Vijayan D.S. Pilkhwal and D. Saha
Reactor Engineering DivisionBhabha Atomic Research Centre
Mumbai, India
4th RCM on the IAEA CRP on Natural Circulation Phenomena, Modelling and Reliability of Passive Safety Systems that Utilize Natural Circulation
Vienna, Austria, September 10-13, 2007 2
BARCBARCBARCBARC
Introduction
A large number of single-phase and two-phase flow pressure drop correlations can be found in literature. Important pressure drop relationships can be found in the IAEA technical document for “Thermohydraulic relationships for advanced water cooled reactors” (IAEA-TECDOC-1203).
Most of the pressure drop correlations are developed from data generated in forced circulation systems.
The mechanism of flow in natural circulation loop may be complex due to buoyancy effect and formation of secondary flows.
Therefore, there is a need to give a closer look to pressure drop phenomena under natural circulation, which is both complex and
important.
Vienna, Austria, September 10-13, 2007
BARCBARCBARCBARC
Definition
Pressure drop is the difference in pressure between two points of interest in a fluid system. In general, pressure drop can be caused by resistance to flow, changes in elevation, density, flow area and flow direction .
It is customary to express the total pressure drop in a flowing system as the sum of its individual components such as distributed pressure loss due to friction, local pressure losses due to sudden variations of shape, flow area, direction, etc. and pressure losses due to acceleration and elevation.
An important factor affecting the pressure loss is the geometry.
Other factors are concerned with the fluid status, the flow nature, the flow pattern, the flow direction, flow type, flow paths and the operating conditions
Vienna, Austria, September 10-13, 2007
BARCBARCBARCBARC
An important focus of this phenomenon is the geometric conditions that hinder the establishment of fully developed flow especially when the fluid in question is a mixture of steam, air and water. This complex thermo-fluid dynamic phenomenon warrants special attention.
Though in many systems like the primary system of a nuclear power plant, flow is mostly not fully developed, pressure drop relationships used in these systems are invariably those obtained for developed flow. This practice is also experimentally proved to be more than adequate in most of the cases. However, in some specific cases like containment internal geometry, it is necessary to consider thermo fluid dynamics in the developing region.
Normally the pressure loss inside a device depends on the nature of flow through the device and not on the nature of driving head causing the flow. However, under some circumstances, because of local effects, the pressure loss may get influenced by the nature of driving force.
Definition
Vienna, Austria, September 10-13, 2007
BARCBARCBARCBARC
Scenario
In a flowing system there are two components of total pressure drop
1. Irreversible pressure drop 2. Reversible pressure drop
The irreversible pressure drop is called pressure loss. This is due to irreversible conversion of mechanical energy (the work of resistance force) into heat. This includes Friction loss and Local loss.
There are also reversible component of pressure drop such as elevation pressure drop and acceleration pressure drop.
Vienna, Austria, September 10-13, 2007
BARCBARCBARCBARC
Scenario
Friction loss
Due to the viscosity (both molecular and turbulent) of real liquid and gases in motion, and results from momentum transfer between the molecules (in laminar flow) and between individual particles (in turbulent flow) of adjacent fluid layers moving at different velocities.
For two-phase flow, an additional frictional pressure drop may be due to the inter-phase friction between gas-liquid or steam-liquid phases.
Local losses
Caused by local disturbances of the flow; separation of flow from the walls; and formation of vortices and strong turbulence agitation of the flow
Vienna, Austria, September 10-13, 2007
BARCBARCBARCBARC
Scenario
Acceleration pressure drop
Due to the energy spent in accelerating the molecules of the fluid. This reversible component of pressure drop is caused by a change in flow area or density.
Elevation pressure drop
Because of the work needs to be done against the gravity to raise the fluid molecules to a height. This reversible component of pressure drop is caused by the difference in elevation.
Vienna, Austria, September 10-13, 2007
BARCBARCBARCBARC
Scenario
The pressure loss components are inseparable. However, for ease of calculation they are subdivided into components like local losses, frictional losses etc.
It is also assumed that the local losses are concentrated in one sectioned although they can occur virtually over a considerable length
Most of the pressure drop correlations reported in literature had been developed from steady state experimental data and mostly under adiabatic conditions.
Vienna, Austria, September 10-13, 2007
BARCBARCBARCBARC
Hardware: where it occurs?
Channel type Vessel typeDistributed pressure drop - Feeder and tail pipe
- Bare bundle- Core and core bypasses- Surge line- Steam Generator (SG) tubes
Local pressure drop - Fuel bundle assembly- Various header connections- Valves and rupture disc locations
- Pump inlet, outlet and inside- Pressurizer and surge line connections
Safety system pressure drop
- Accumulator outlet line- ECCS header to water tube connection- Advanced fluidic device- Gravity Driven Water Pool (GDWP) to ECCS header connection
-Accumulator connections- ECCS connections
Geometries of interest to Nuclear Power Plants (NPPs) are only considered here.
Vienna, Austria, September 10-13, 2007
BARCBARCBARCBARCSingle Phase Pressure Drop: flow
under transition regime• In many transients, the flow may change from laminar to turbulent, or vice versa, necessitating a switch of correlations.
• Numerical calculations, often encounter convergence problems when such switching takes place due to the discontinuity in the friction factor values predicted by the laminar flow and turbulent flow equations.
• Well established correlations for friction factor do not exist in transition region.
Few ways:
1.Calculate both fTURBULENT and fLAMINAR. If fT > fL then f = fT. This procedure, however, causes the switch over from laminar to turbulent flow equation at Re1100.
2. f = (fT)4000 for 2000 Re 4000 where (fT)4000 is the friction factor calculated by the turbulent flow equation at Re = 4000.
Vienna, Austria, September 10-13, 2007
BARCBARCBARCBARCSingle Phase Pressure Drop: flow
under diabatic condition
fNON-ISOTHERMAL = fISOTHERMAL* F
1. F in terms of temperature correction:
F=1+ C Tf ;
Tf is the temperature drop in the laminar layer (q”/h). Constant C = 0.001-0.0025
2. F in terms of viscosity ratio:
F = ( bulk / wall )- 0.28
fNON-ISOTHERMAL = fISOTHERMAL with properties evaluated at film temperature
Tfilm = 0.4 (Twall - Tbulk) + Tbulk
Vienna, Austria, September 10-13, 2007 12
BARCBARCBARCBARC
Two-phase
Two-phase pressure drop relationships- adiabatic Empirical correlation based on the homogeneous model Empirical correlation based on the two-phase friction multiplier concept Direct empirical models Flow pattern specific models
Void fraction relationships
Slip ratio models K- models Correlations based on drift flux models
Vienna, Austria, September 10-13, 2007
BARCBARCBARCBARC
Models using interfacial friction
Another form of two-phase pressure drop correlations are that uses interfacial friction models. The two-fluid model used in many of the advanced system codes require correlations for interfacial friction in addition to wall friction.
Flow under diabatic condition
The correlations discussed so far are applicable to adiabatic two-phase flow. The effect of heat flux on two phase pressure drop has been studied by Leung and Groeneveld (1991), Tarasova (1966) and Koehler and Kastner (1988).
Two-phase
Vienna, Austria, September 10-13, 2007
BARCBARCBARCBARC
Assessment of two-phase pressure drop correlationsThe table given below gives the assessment of pressure drop correlations by various authors and their recommendation.
Authors Categories No. of correlations
tested
No. of data
points
Recommended correlation
Weisman-Choe (1976)
Homogeneous model
--- --- McAdams (1942) and Dukler et al. (1964)
Idsinga et al. (1977)
Homogeneous model
18 3500 Owens (1961) and Cicchitti (1960)
Beattie-Whalley (1982)
Homogeneous model
12 13500 Beattie and Whalley (1982)
Dukler et al. (1964) Multiplier concept
5 9000 Lockhart and Martinelli (1949)
Idsinga et al. (1977)
Multiplier concept
14 3500 Baroczy (1966) and Thom (1964)
Friedel (1980) Multiplier concept
14 12868 Chisholm (1973) and Lombardi-Pedrocchi (1972)
Two-phase
Vienna, Austria, September 10-13, 2007
BARCBARCBARCBARC
Snoek-Leung (1989)
---- 9 1217 Friedel (1979)
Vijayan et al. (2000) --- 14 424 Lockhart and Martinelli (1949) with Chexal et al. (1996) for
void fraction.
Weisman-Choe (1976)
Flow pattern specific
11 Separated flow: Agrawal et al. (1973) and Hoogendoorn (1959)
10 Homogeneous flow : McAdams (1942), Dukler et al. (1964) and Chisholm (1968)
7 Intermittent flow: Dukler (1964), Lockhart-Martinelli (1949) and Hughmark (1965)
6 Annular flow: Dukler (1964) and Lockhart- Martinelli (1949)
Two-phase
Vienna, Austria, September 10-13, 2007
BARCBARCBARCBARC
Mandhane et al. (1977)
Flow pattern specific
14 10500 Bubbly: Chenoweth and Martin (1956)
Stratified: Agrawal et al. (1973)
Stratified wavy: Dukler et al (1964)
Slug: Mandhane et al. (1974)
Annular, annular mist: Chenoweth and Martin
(1956)
Dispersed bubble : Mandhane et al. (1974)
Two-phase
Vienna, Austria, September 10-13, 2007 17
BARCBARCBARCBARC
Natural and Forced Circulation Pressure Drop
For forced circulation loops, the driving force is the head developed by the pump which is generally far greater than the buoyancy driving head.
The buoyancy being the driving head, natural circulation flows are characterized by low driving head and consequent low mass flux.
Due to buoyancy effect and presence of secondary flows, the velocity profile in a heated pipe may get modified which also depends on the orientation of the pipe (horizontal, vertical upward or downward).
The secondary flow may, in turn, affect the friction factor for the pipe, as the friction factor is mainly dependent upon the velocity gradient.
Vienna, Austria, September 10-13, 2007 18
BARCBARCBARCBARC
Natural and Forced Circulation Pressure Drop
Forced Circulation Natural Circulation
Driving head Large Small
Secondary flow Negligible effect Could be significant
Transition from laminar to turbulent flow
Occurs at higher Reynolds number (Re)
Occurs at lower Re due to secondary flow
Pressure drop correlations at low mass flux
Accuracy need not be high
High accuracy required
Transient Relatively fast Sluggish
Flow Relatively high Low
Stratification Not a concern Commonly encountered
Instabilities Less potent High potential
CHF Relatively higher Relatively lower
Vienna, Austria, September 10-13, 2007 19
BARCBARCBARCBARC Pressure Drop under Low Mass Flux, Low Pressure Conditions
• For a natural circulation loop during start-up, the flow builds up virtually from zero flow condition. Hence the friction factor and loss coefficient correlations should be accurate at very low mass flux.
• Natural circulation loops are particularly susceptible to instabilities at low pressure conditions. These flow instabilities may be characterized by repetitive flow reversals.
• There is a need to assess the existing correlation in terms of its applicability for natural circulation loop.
Vienna, Austria, September 10-13, 2007 20
BARCBARCBARCBARC
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.350.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.10
0.11
0.12
Presure = 21 bar
Mass flux = 558 kg/m2-s Pipe ID = 26.64 mm
Pre
ssu
re d
rop
- k
g/c
m2
Quality
Calculated (Chisholm Model) Measureed
Comparison of measured and calculated pressure drop in a vertical pipe with diabatic flow
Pressure Drop at Low Mass Flux
Vienna, Austria, September 10-13, 2007 21
BARCBARCBARCBARC
0.0 0.4 0.8 1.2 1.6 2.00.0
0.4
0.8
1.2
1.6
2.0 Presure = 21 - 72 bar
Mass flux = 40 - 2000 kg/m2-s
heat flux = 55 - 65.2 W/m2
Pipe ID = 26.64 mm
+10%
+20%+3
0%
-10%
-30%
-20%
P meas=
P cal
Experimental data
Pca
l - b
ar
Pmeas
- bar
Comparison of measured and predicted pressure drop using CNEN (1973) correlation for vertical upward diabatic flow in a tube
Pressure Drop at Low Mass Flux
Vienna, Austria, September 10-13, 2007 22
BARCBARCBARCBARC
Single Phase Natural Circulation
The generalized flow correlation for single-phase loops (Vijayan, 1992) is given by,
r
effmss L
DGrC
Re rpC 2 and br 31
where p and b are given by the friction factor correlation of the form
bpf Re
5.0
1768.0Re
G
mss N
Gr
364.0
96.1Re
G
mss N
Gr
Laminar flow (p=64, b=1)
Turbulent flow (p=0.316, b=0.25) with Blasius correlation
Depending on the value of p and b, the flow correlation is given as
pr
hTrm
CA
HgQDGr
3
20
3
Modified Grashoff number
i
N
ibb
eff
r
tG
ad
l
D
LN
121Geometrical parameter
Vienna, Austria, September 10-13, 2007 23
BARCBARCBARCBARC
106 107 108 109 1010 1011 1012102
103
104
105
6 mm loop 11 mm loop 23.2 mm loop 26.9 mm loop VHHC orientation
f = 0.184 / Re0.2
Blasius correlation
f = 0.316 / Re0.25
f = 55.92 / Re0.9501
f = 64 / Re
Re ss
Grm / N
G
Effect of friction factor on steady state flow rate in a single-phase natural circulation loop as predicted by generalized flow correlation
Generalized Correlation
Vienna, Austria, September 10-13, 2007 24
BARCBARCBARCBARC
Flow dependency on power
0.0 0.5 1.0 1.5 2.0 2.50.000
0.025
0.050
0.075
0.100
Flo
w (
kg/s
)
Power (kW)
Experimental data
Blasius correlation (f = 0.316 Re-0.25)
f = 0.184 Re-0.2
Effect of friction factor on steady state flow rate in a single-phase natural circulation loop
b
Gbr
br
brhT
ss CpN
ADQHg
pW
3
122
02
Vienna, Austria, September 10-13, 2007 25
BARCBARCBARCBARC
Two Phase Natural Circulation
5.0
1768.0Re
G
mss N
Gr
364.0
96.1Re
G
mss N
Gr
A generalized flow correlation of the same form as that of single-phase has been developed (Gartia et al. (2006)) to estimate the steady state flow rate in two-phase natural circulation loops which is given by,
rGmss NGrCRe
Where Ress= Steady State Reynolds Number ; Grm = Modified Grashof Number NG = Geometric Parameter
Laminar flow (p=64, b=1)
Turbulent flow (p=0.316, b=0.25) with Blasius correlation
For density variation, pm
tp hv
1
Vienna, Austria, September 10-13, 2007 26
BARCBARCBARCBARC
Generalized Correlation
Effect of friction factor on steady state flow rate in two-phase natural circulation loops
105 106 107 108 109 1010 1011 1012 1013 1014 1015
102
103
104
105
Turbulent flow equation( C=1.96, r=0.364)
-25%
+25%
Laminar flow equation( C=0.1768, r=0.5)
Re ss
Grm/ N
G
Generalized correlation
Ress
=C [Grm/N
G]r
Apsara loop (ID:1/2") Apsara loop (ID:3/4") Apsara loop (ID:1")
Vienna, Austria, September 10-13, 2007 27
BARCBARCBARCBARC
0 100 200 300 400 5000
1
2
3
4
5
6Pressure=50 bar
Tfeed
=250oC
Mas
s F
low
Rat
e (k
g/s
)
Power (kW)
RELAP5/Mod 3.2 (Two-Fluid Model) Experimental data (ITL) Generalized correlation with Blassius model Generalized correlation with Colebrook model
Variation of friction factor on two phase flow prediction
Effect of friction factor on steady state flow rate in a two-phase natural circulation loop as predicted by the generalized flow correlation
b
Gbr
lb
rbrtpr
ss N
ADQHg
pW
3
12
2
Vienna, Austria, September 10-13, 2007 28
BARCBARCBARCBARC
0 10 20 30 40 50 60 70 80
0.08
0.12
0.16
0.20Colebrook Model for single-phase friction factor
Pressure = 20 barT
sub = 4K
Mas
s fl
ow
rat
e (k
g/s
)
Power(kW)
Homogeneous Lockhart-Martinelli Martinelli-Nelson Chisholm-Laird Sekoguchi
Effect of two-phase friction factor multiplier on steady state flow rate in a two-phase natural circulation loop using the generalized correlation
Effect of Friction Factor Multiplier
Vienna, Austria, September 10-13, 2007 29
BARCBARCBARCBARC
Effect of pressure on steady state flow rate in a two-phase natural circulation loop
Effect of Pressure
10 20 30 40 50 60 700.0
0.4
0.8
1.2
1.6
2.0
2.4Single-phase friction factor used
Experimental : f = 0.569 / Re0.30457
Theoretical : Colebrook Model
Power = 30 kWT
sub = 2 K
Mas
s fl
ow
rat
e (k
g/s
)
Pressure (bar)
Homogenous model Lockhart-Martinelli model Martinelli-Nelson model Chisholm-Laird model Sekoguchi model Experimental data (HPNCL)
Vienna, Austria, September 10-13, 2007 30
BARCBARCBARCBARC
Effect of friction factor on stability
Effect of friction factor on stability in a single-phase natural circulation loop
0 1 2 3 4 5 6 7 8 9 1010-7
100
107
1014
1021
1028
1035
Lt=7.23m, L
t/D=267.29, HHHC Orientation
Stable
Stable
Un
sta
ble
Gr m
Stm
Turbulent flow (p=0.184, b=0.2) Turbulent flow (p=0.316, b=0.25)
Vienna, Austria, September 10-13, 2007 31
BARCBARCBARCBARC
Effect of friction factor on stability
Effect of two-phase friction factor multiplier on the stability of a two-phase natural circulation loop
0 5 10 15 20 25 30 350
20
40
60
80
100
120T
sub= 4K Homogenous model
Lockhart-Martinelli model Martinelli-Nelson model Chisholm-Laird model Sekoguchi model Threshold of Instability
[Furutera (1986)]
Stable
Unstable
Po
wer
(kW
)
Pressure (bar)
Vienna, Austria, September 10-13, 2007 32
BARCBARCBARCBARC
Effect of large flow areas on pressure drops
Although large diameter pipes, large manifolds are used in natural circulation system, still there is no valid correlation for such geometry.
Simpson et al. (1977) compared six pressure drop correlations with data from large diameter (127 and 216 mm) horizontal pipes.
None of the pressure gradient correlations studied predicted the measure pressure drops adequately. In particular, measured pressure gradients for stratified flow differed by an order of magnitude from those predicted by the various correlations.
In view of this, the validity of the existing empirical correlations needs to be checked. However, this is not unique to only natural circulation system.
Vienna, Austria, September 10-13, 2007 33
BARCBARCBARCBARC
Concluding Remarks
Within the range of parameter studied so far, relationships for forced circulation as given in TECDOC-1203 were found to be adequate for natural circulation and stability of natural circulation.
More accurate prediction capability is required at low mass flux and for large area flow paths. However, this issue is not unique to only natural circulation systems.
Applicability of existing correlations to natural circulation needs to be assessed covering wider range of parameters.