Professional Documents
Culture Documents
Boiling Heat Transfer, Two-Phase Flow and Circulation: The Babcock & Wilcox Company
Boiling Heat Transfer, Two-Phase Flow and Circulation: The Babcock & Wilcox Company
Chapter 5
Boiling Heat Transfer,
Two-Phase Flow and Circulation
A case of heat transfer and flow of particular inter- ture (Tsat) is a unique function of pressure. The Ameri-
est in steam generation is the process of boiling and can Society of Mechanical Engineers (ASME) and the
steam-water flow. The boiling or evaporation of wa- International Association for the Properties of Steam
ter is a familiar phenomenon. In general terms, boil- (IAPS) have compiled extensive correlations of thermo-
ing is the heat transfer process where heat addition physical characteristics of water. These characteristics
to a liquid no longer raises its temperature under con- include the enthalpy (or heat content) of water, the
stant pressure conditions; the heat is absorbed as the enthalpy of evaporation (also referred to as the latent
liquid becomes a gas. The heat transfer rates are high, heat of vaporization), and the enthalpy of steam. As
making this an ideal cooling method for surfaces ex- the pressure is increased to the critical pressure [3200
posed to the high heat input rates found in fossil fuel psi (22.1 MPa)], the latent heat of vaporization declines
boilers, concentrated solar energy collectors and the to zero and the bubble formation associated with boil-
nuclear reactor fuel bundles. However, the boiling ing no longer occurs. Instead, a smooth transition from
phenomenon poses special challenges such as: 1) the liquid to gaseous behavior occurs with a continuous in-
sudden breakdown of the boiling behavior at very high crease in temperature as energy is applied.
heat input rates, 2) the potential flow rate fluctuations Two other definitions are also helpful in discussing
which may occur in steam-water flows, and 3) the ef- boiling heat transfer:
ficient separation of steam from water. An additional
feature of boiling and two-phase flow is the creation 1. Subcooling For water below the local saturation
of significant density differences between heated and temperature, this is the difference between the
unheated tubes. These density differences result in saturation temperature and the local water tem-
water flowing to the heated tubes in a well designed perature (Tsat – T ).
boiler natural circulation loop. 2. Quality This is the flowing mass fraction of steam
Most fossil fuel steam generators and all commer- (frequently stated as percent steam by weight or
cial nuclear steam supply systems operate in the pres- %SBW after multiplying by 100%):
sure range where boiling is a key element of the heat m steam
transfer process. Therefore, a comprehensive under- x = (1)
water + m
m steam
standing of boiling and its various related phenom-
ena is essential in the design of these units. Even at where
operating conditions above the critical pressure, where
water no longer boils but experiences a continuous steam = steam flow rate, lb/h (kg/s)
m
transition from a liquid-like to a gas-like fluid, boil-
ing type behavior and special heat transfer charac- water = water flow rate, lb/h (kg/s)
m
teristics occur.
Thermodynamically, this can also be defined as:
H − Hf H − Hf
Boiling process and fundamentals x = or (2)
H fg Hg − Hf
Boiling point and thermophysical properties where
The boiling point, or saturation temperature, of a H = local average fluid enthalpy, Btu/lb (J/kg)
liquid can be defined as the temperature at which its Hf = enthalpy of water at saturation, Btu/lb (J/kg)
vapor pressure is equal to the total local pressure. The Hg = enthalpy of steam at saturation, Btu/lb (J/kg)
saturation temperature for water at atmospheric pres- Hfg = latent heat of vaporization, Btu/lb (J/kg)
sure is 212F (100C). This is the point at which net
vapor generation occurs and free steam bubbles are When boiling is occurring at saturated, thermal
formed from a liquid undergoing continuous heating. equilibrium conditions, Equation 2 provides the frac-
As discussed in Chapter 2, this saturation tempera- tional steam flow rate by mass. For subcooled condi-
tions where H < Hf, quality (x) can be negative and is critical heat flux (CHF), departure from nucleate boil-
an indication of liquid subcooling. For conditions where ing (DNB), burnout, dryout, peak heat flux, or boil-
H > Hg, this value can be greater than 100% and repre- ing crisis. The temperature response of the surface un-
sents the amount of average superheat of the steam. der this condition depends upon how the surface is
being heated. In fossil fuel boiler furnaces and nuclear
Boiling curve reactor cores, the heat input is effectively independent
Fig. 1 illustrates a boiling curve which summarizes of surface temperature. Therefore, a reduction in the
the results of many investigators. This curve provides heat transfer rate results in a corresponding increase
the results of a heated wire in a pool, although the in surface temperature from point D to D′ in Fig. 1.
characteristics are similar for most situations. The heat In some cases, the elevated surface temperature is so
transfer rate per unit area, or heat flux, is plotted high that the metal surface may melt. If, on the other
versus the temperature differential between the metal hand, the heat input or heat transfer rate is depen-
surface and the bulk fluid. From points A to B, con- dent upon the surface temperature, typical of a
vection heat transfer cools the wire and boiling on the nuclear steam generator, the average local tempera-
surface is suppressed. Moving beyond point B, which ture of the surface increases as the local heat trans-
is also referred to as the incipient boiling point, the fer rate declines. This region, illustrated in Fig. 1 from
temperature of the fluid immediately adjacent to the points D to E, is typically referred to as unstable film
heated surface slightly exceeds the local saturation boiling or transition boiling. Because a large surface
temperature of the fluid while the bulk fluid remains temperature increase does not occur, the main conse-
subcooled. Bubbles, initially very small, begin to form quences are a decline in heat transfer performance per
adjacent to the wire. The bubbles then periodically unit surface area and less overall energy transfer. The
collapse as they come into contact with the cooler bulk actual local phenomenon in this region is quite com-
fluid. This phenomenon, referred to as subcooled boil- plex and unstable as discrete areas of surface fluctu-
ing, occurs between points B and S on the curve. The ate between a wetted boiling condition and a steam
heat transfer rate is quite high, but no net steam gen- blanketed, or dry patch, condition. From position E
eration occurs. From points S to C, the temperature through D′ to F, the surface is effectively blanketed
of the bulk fluid has reached the local saturation tem- by an insulating layer of steam or vapor. Energy is
perature. Bubbles are no longer confined to the area transferred from the solid surface through this layer
immediately adjacent to the surface, but move into the by radiation, conduction and microconvection to the
bulk fluid. This region is usually referred to as the liquid-vapor interface. From this interface, evapora-
nucleate boiling region, and as with subcooled boil- tion occurs and bubbles depart. This heat transfer
ing, the heat transfer rates are quite high and the region is frequently referred to as stable film boiling.
metal surface is only slightly above the saturation In designing steam generating systems, care must
temperature. be exercised to control which of these phenomena oc-
As point C is approached, increasingly large sur- cur. In high heat input locations, such as the furnace
face evaporation rates occur. Eventually, the vapor area of fossil fuel boilers or nuclear reactor cores, it is
generation rate becomes so large that it restricts the important to maintain nucleate or subcooled boiling
liquid return flow to the surface. The surface eventu- to adequately cool the surface and prevent material
ally becomes covered (blanketed) with an insulating failures. However, in low heat flux areas or in areas
layer of steam and the ability of the surface to trans- where the heat transfer rate is controlled by the boil-
fer heat drops. This transition is referred to as the ing side heat transfer coefficient, stable or unstable
film boiling may be acceptable. In these areas, the
resultant heat transfer rate must be evaluated, any
temperature limitations maintained and only allow-
able temperature fluctuations accepted.
Flow boiling
Flow or forced convective boiling, which is found in
virtually all steam generating systems, is a more com-
plex phenomenon involving the intimate interaction
of two-phase fluid flow, gravity, material phenomena
and boiling heat transfer mechanisms. Fig. 2 is a clas-
sic picture of boiling water in a long, uniformly heated,
circular tube. The water enters the tube as a subcooled
liquid and convection heat transfer cools the tube. The
point of incipient boiling is reached (point 1 in Fig. 2).
This results in the beginning of subcooled boiling and
bubbly flow. The fluid temperature continues to rise
until the entire bulk fluid reaches the saturation tem-
perature and nucleate boiling occurs, point 2. At this
location, flow boiling departs somewhat from the
simple pool boiling model previously discussed. The
Fig. 1 Boiling curve – heat flux versus applied temperature difference. steam-water mixture progresses through a series of
tigue failures. Data must also be available to predict boiling components as exemplified by the correlation
the rate of heat transfer downstream of the dryout point. developed by J.C. Chen and his colleagues.4 While
CHF phenomena are less important than the heat such correlations are frequently recommended for use
transfer rates for performance evaluation, but are more in saturated boiling systems, their additional precision
important in defining acceptable operating conditions. is not usually required in many boiler or reactor ap-
As discussed in Chapter 4, the heat transfer rate per plications. For general evaluation purposes, the
unit area or heat flux is equal to the product of tem- subcooled boiling relationship provided in Equation
perature difference and a heat transfer coefficient. 4 is usually sufficient.
Post-CHF heat transfer As shown in Fig. 3, substan-
Heat transfer coefficients tial increases in tube wall metal temperatures are
Heat transfer correlations are application (surface possible if boiling is interrupted by the CHF phenom-
and geometry) specific and The Babcock & Wilcox enon. The maximum temperature rise is of particular
Company (B&W) has developed extensive data for its importance in establishing whether tube wall over-
applications through experimental testing and field heating may occur. In addition, the reliable estima-
experience. These detailed correlations remain propri- tion of the heat transfer rate may be important for an
etary to B&W. However, the following generally avail- accurate assessment of thermal performance. Once the
able correlations are provided here as representative metal surface is no longer wetted and water droplets
of the heat transfer relationships. are carried along in the steam flow, the heat transfer
Single-phase convection Several correlations for process becomes more complex and includes: 1) con-
forced convection heat transfer are presented in Chap- vective heat transfer to the steam which becomes su-
ter 4. Forced convection is assumed to occur as long as perheated, 2) heat transfer to droplets impinging on
the calculated forced convection heat flux is greater than the surface from the core of the flow, 3) radiation di-
the calculated boiling heat flux (point 1 in Fig. 2): rectly from the surface to the droplets in the core flow,
and 4) heat transfer from the steam to the droplets.
′′ Convection > qBoiling
qForced ′′ (3) This process results in a nonequilibrium flow featur-
ing superheated steam mixed with water droplets.
While not critical in most steam generator applica- Current correlations do not provide a good estimate of
tions, correlations are available which explicitly de- the heat transfer in this region, but computer models
fine this onset of subcooled boiling and more accurately show promise. Accurate prediction requires the use of
define the transition region.1 experimental data for similar flow conditions.
Subcooled boiling In areas where subcooled boil- Reflooding A key concept in evaluating emergency
ing occurs, several correlations are available to char- core coolant systems for nuclear power applications is
acterize the heat transfer process. Typical of these is reflooding. In a loss of coolant event, the reactor core
the Jens and Lottes2 correlation for water. For inputs can pass through critical heat flux conditions and can
with English units: become completely dry. Reflooding is the term for the
complex thermal-hydraulic phenomena involved in
∆Tsat = 60 ( q′′ / 106 )
1/4
e − P / 900 (4a) rewetting the fuel bundle surfaces as flow is returned
to the reactor core. The fuel elements may be at very
and for inputs with SI units: elevated temperatures so that the post-CHF, or steam
blanketed, condition may continue even in the presence
of returned water flow. Eventually, the surface tem-
∆Tsat = 25 ( q′′)
1/4
e − P / 6.2 (4b) perature drops enough to permit a rewetting front to
wash over the fuel element surface. Analysis includes
where transient conduction of the fuel elements and the in-
∆T sat = Tw – Tsat, F (C) teraction with the steam-water heat transfer processes.
Tw = wall temperature, F (C)
T sat = saturated water temperature, F (C) Critical heat flux phenomena
q′′ = heat flux, Btu/h ft2 (MWt/m2) Critical heat flux is one of the most important pa-
rameters in steam generator design. CHF denotes the
P = pressure, psi (MPa)
set of operating conditions (mass flux, pressure, heat
Another relationship frequently used is that developed flux and steam quality) covering the transition from
by Thom.3 the relatively high heat transfer rates associated with
Nucleate and convective boiling Heat transfer in the nucleate or forced convective boiling to the lower rates
saturated boiling region occurs by a complex combi- resulting from transition or film boiling (Figs. 1 and
nation of bubble nucleation at the tube surface (nucle- 2). These operating conditions have been found to be
ate boiling) and direct evaporation at the steam-wa- geometry specific. CHF encompasses the phenomena
ter interface in annular flow (convective boiling). At of departure from nucleate boiling (DNB), burnout,
low steam qualities, nucleate boiling dominates while dryout and boiling crisis. One objective in recirculat-
at higher qualities convective boiling dominates. While ing boiler and nuclear reactor designs is to avoid CHF
separate correlations are available for each range, the conditions. In once-through steam generators, the
most useful relationships cover the entire saturated objective is to design to accommodate the temperature
boiling regime. They typically involve the summation increase at the CHF locations. In this process, the heat
of appropriately weighted nucleate and convective flux profile, flow passage geometry, operating pressure
and inlet enthalpy are usually fixed, leaving mass falling rapidly near the critical pressure; i.e., at con-
flux, local quality, diameter and some surface effects stant heat flux, CHF occurs at lower steam qualities
as the more easily adjusted variables. as pressure rises.
Factors affecting CHF Critical heat flux phenomena Many CHF correlations have been proposed and are
under flowing conditions found in fossil fuel and satisfactory within certain limits of pressure, mass
nuclear steam generators are affected by a variety of velocity and heat flux. Fig. 5 is an example of a corre-
parameters.5 The primary parameters are the operat- lation which is useful in the design of fossil fuel natu-
ing conditions and the design geometries. The oper- ral circulation boilers. This correlation defines safe and
ating conditions affecting CHF are pressure, mass flux unsafe regimes for two heat flux levels at a given pres-
and steam quality. Numerous design geometry factors sure in terms of steam quality and mass velocity. Ad-
include flow passage dimensions and shape, flow path ditional factors must be introduced when tubes are
obstructions, heat flux profile, inclination and wall used in membrane or tangent wall construction, are
surface configuration. Several of these effects are il- inclined from the vertical, or have different inside di-
lustrated in Figs. 3 through 7. ameter or surface configuration. The inclination of the
Fig. 3 illustrates the effect of increasing the heat flow passage can have a particularly dramatic effect
input on the location of the temperature excursion in a on the CHF conditions as illustrated in Fig. 6.6
uniformly heated vertical tube cooled by upward flow- Ribbed tubes Since the 1930s, B&W has investi-
ing water. At low heat fluxes, the water flow can be al- gated a large number of devices, including internal
most completely evaporated to steam before any tem- twisters, springs and grooved, ribbed and corrugated
perature rise is observed. At moderate and high heat tubes to delay the onset of CHF. The most satisfactory
fluxes, the CHF location moves progressively towards overall performance was obtained with tubes having
the tube inlet and the maximum temperature excur- helical ribs on the inside surface.
sion increases. At very high heat fluxes, CHF occurs at Two general types of rib configurations have been
a low steam quality and the metal temperature excur- developed:
sion can be high enough to melt the tube. At extremely
1. single-lead ribbed (SLR) tubes (Fig. 8a) for small
high heat input rates, CHF can occur in subcooled
internal diameters used in once-through subcriti-
water. Avoiding this type of CHF is an important de-
cal pressure boilers, and
sign criterion for pressurized water nuclear reactors.
2. multi-lead ribbed (MLR) tubes (Fig. 8b) for larger in-
Many large fossil fuel boilers are designed to oper-
ternal diameters used in natural circulation boilers.
ate between 2000 and 3000 psi (13.8 and 20.7 MPa).
In this range, pressure has a very important effect, Both of these ribbed tubes have shown a remark-
shown in Fig. 4, with the steam quality limit for CHF able ability to delay the breakdown of boiling. Fig. 7
Fig. 4 Steam quality limit for CHF as a function of pressure. Fig. 5 Steam quality limit for CHF as a function of mass flux.
Fig. 6 Effect of inclination on CHF at 700,000 lb/h ft2 (950 kg/m2 s).6 Because the ribbed bore tube is more expensive than
a smooth bore tube, its use involves an economic bal-
ance of several design factors. In most instances, there
compares the effectiveness of a ribbed tube to that of is less incentive to use ribbed tubes below 2200 psi
a smooth tube in a membrane wall configuration. This (15.2 MPa).
plot is different from Fig. 5 in that heat flux is given Evaluation CHF is a complex combination of ther-
as an average over the flat projected surface. This is mal-hydraulic phenomena for which a comprehensive
more meaningful in discussing membrane wall heat theoretical basis is not yet available. As a result, ex-
absorption. perimental data are likely to continue to be the basis
The ribbed bore tubes provide a balance of improved for CHF evaluations. Many data and correlations de-
CHF performance at an acceptable increase in pres- fine CHF well over limited ranges of conditions and
sure drop without other detrimental effects. The ribs geometries. However, progress is being made in de-
generate a swirl flow resulting in a centrifugal action veloping more general evaluation procedures for at
which forces the water to the tube wall and retards least the most studied case – a uniformly heated
entrainment of the liquid. The steam blanketing and smooth bore tube with upward flowing water.
film dryout are therefore prevented until substantially To address this complex but critical phenomenon in
higher steam qualities or heat fluxes are reached. the design of reliable steam generating equipment,
B&W has developed an extensive proprietary data-
base and associated correlations. A graphical example
is shown in Fig. 5 for a fossil fuel boiler tube. A B&W
correlation 7 for nuclear reactor fuel rod bundle
subchannel analysis is shown in Table 1.
CHF criteria A number of criteria are used to assess
the CHF margins in a particular tube or tube bundle
geometry.8 These include the CHF ratio, flow ratio and
quality margin, defined as follows:
Fig. 9 Fossil boiler CHF ratio = minimum value of critical heat flux Fig. 10 Nuclear reactor CHF ratio = minimum value of critical heat
divided by upset heat flux. flux divided by design heat flux.
Two-phase flow
Flow patterns
As illustrated in Fig. 2, two-phase steam-water flow
may occur in many regimes or structures. The transi-
tion from one structure to another is continuous rather
than abrupt, especially under heated conditions, and
is strongly influenced by gravity, i.e., flow orientation.
Because of the qualitative nature of flow pattern iden-
tification, there are probably as many flow pattern
descriptions as there are observers. However, for ver-
tical, heated, upward, co-current steam-water flow in a
tube, four general flow patterns are generally recognized
(see Fig. 12):
1. Bubbly flow Relatively discrete steam bubbles are
dispersed in a continuous liquid water phase.
Bubble size, shape and distribution are dependent
upon the flow rate, local enthalpy, heat input rate
Fig. 11 Thermophysical properties of water (English units). and pressure.
2. Intermediate flow This is a range of patterns be-
tween bubbly and annular flows; the patterns are
where also referred to as slug or churn flow. They range
h = heat transfer coefficient, Btu/h ft2 F (W/m2 K) from: a) large bubbles, approaching the tube size
k = thermal conductivity, Btu/h ft F (W/m K) in diameter, separated from the tube wall by thin
Di = inside tube diameter, ft (m) annular films and separated from each other by
G = mass flux, lb/h ft2 (kg/m2 s) slugs of liquid which may also contain smaller
µ = viscosity, lb/ft h (kg/m s) bubbles, to b) chaotic mixtures of large nonsym-
H = enthalpy, Btu/lb (J/kg) metric bubbles and small bubbles.
T = temperature, F (C) 3. Annular flow A liquid layer is formed on the tube
υ = specific volume, ft3/lb (m3/kg) wall with a continuous steam core; most of the liq-
f G υf
2
δP
− = (single-phase friction) (13)
δ l LO Di 2 gc
G 2 δ x υ g (1.0 − x ) υ f
2 2
δP
− = + (accelerration) (14)
δ l a gc δ l α
(1.0 − α )
δP g α (1.0 − α )
−
δ = g sin θ + (static head) (15)
l g c υg υf
G2 υ f
∆Pl = Φ K (local losses) (16)
2 gc
where
Φ and φ LO
2
= appropriate two-phase multipliers
G = mass flux, lb/s ft2 (kg/m2 s)
f = fanning friction factor (see Chapter 3)
Di = tube inside diameter, ft (m)
Fig. 14 Void fraction – quality relationship (homogeneous model, g = acceleration of gravity, ft/s2 (m/s2)
upper bound; separated flow model, lower bound). g c = 32.17 lbm ft/lbf s2 (1 kg m/N s2)
and
γx
α = (18)
1 + x (γ − 1 )
where
γ = (υg /υf)n
n = (0.8294 – 1.1672/P)
P = pressure, psi
υg = saturated steam specific volume, ft3/lb
υf = saturated liquid specific volume, ft3/lb
x = steam quality
Instabilities
Instability in two-phase flow refers to the set of
operating conditions under which sudden changes in
flow direction, reduction in flow rate and oscillating
flow rates can occur in a single flow passage. Often in
manifolded multi-channel systems, the overall mass
flow rate can remain constant while oscillating flows
in individual channels still may occur. Such unstable
conditions in steam generating systems can result in: Fig. 16 Thom void fraction correlation (>3% SBW).12
δ∆P δ∆P
δG ≤ (19)
internal δ G applied
Parameter
Increased Effect on ∆P Comment
Density wave/pressure drop instability Density wave 1. prevent water droplet carryover into the super-
instabilities involve kinematic wave propagation phe- heater where thermal damage may occur,
nomena. Regenerative feedback between flow rate, 2. minimize steam carryunder in the water leaving
vapor generation rate and pressure drop produce self the drum where residual steam can reduce the
sustaining alternating waves of higher and lower den- effective hydraulic pumping head, and
sity mixture that travel through the tube. This dy- 3. prevent the carryover of solids dissolved in the steam-
namic instability can occur in single tubes that con- entrained water droplets into the superheater and
tain two-phase flows. In addition, when multiple tubes turbine where damaging deposits may form.
are connected by inlet and outlet headers, a more com-
The last item is of particular importance. Boiler wa-
plex coupled channel instability, which is driven by
ter may contain contaminants, principally in solution.
density wave oscillations, may occur. Vertical heat flux
These arise from impurities in the makeup water,
distribution is a particularly sensitive parameter in
treatment chemicals and condensate system leaks, as
dynamic instability evaluation.
well as from the reaction of the water and contami-
Density wave oscillations can be predicted by the
nants with the boiler and preboiler equipment mate-
application of feedback control theory. A number of
rials. Even low levels of these solids in the steam (less
computer codes have been developed to provide these
than 0.6 ppm) can damage the superheater and tur-
predictions. In addition, instability criteria, which use
bine. Because the solubility of these solids is typically
a series of dimensionless parameters to reduce the
several orders of magnitude less in steam than in wa-
complexity of the evaluation, have been developed.
ter (see Chapter 42), small amounts of water droplet car-
Effects of operating and design parameters on the
ryover (greater than 0.25% by weight) may result in dra-
density wave instability include:
matically increased solids carryover and unacceptable
Parameter deposition in the superheater and turbine. The deposits
Increased Change in stability have caused turbine damage as well as superheater tube
temperature increases, distortion and burnout.
mass flux improved A cross-section of a horizontal steam drum found
heat flux reduced on a modern high capacity fossil fuel boiler is shown
pressure improved in Fig. 19. This illustrates the general arrangement
inlet ∆P improved of the baffle plates, primary cyclone separators, sec-
inlet subcooling improved (large subcooling) ondary separator elements (scrubbers), water dis-
reduced (small subcooling) charger (downcomer) and feedwater inlets. The blow-
down (water removal) connections are not shown. The
steam-water separation typically takes place in two
Steam-water separation stages. The primary separation removes nearly all the
Subcritical pressure recirculating boilers and steam steam from the water so that very little steam is recir-
generators are equipped with large cylindrical vessels culated from the bottom of the drum through the out-
called steam drums. Their primary objective is to per- let connection (downcomer) towards the heated tubes.
mit separation of the saturated steam from the steam- The steam leaving the primary separators in high
water mixture leaving the boiling heat transfer sur- pressure boilers still typically contains too much liq-
faces. The steam-free water is recirculated with the uid in the form of contaminant-containing droplets for
feedwater to the heat absorbing surfaces for further satisfactory superheater and turbine performance.
steam generation. The saturated steam is discharged Therefore, the steam is passed through a secondary
through a number of outlet nozzles for direct use or set of separators, or scrubber elements (usually closely
further heating. The steam drum also serves to: spaced, corrugated parallel plates) for final water
droplet removal. The steam is then exhausted through
1. mix the feedwater with the saturated water re- several connections. As this figure indicates, success-
maining after steam separation, ful steam-water separation involves the integrated
2. mix the corrosion control and water treatment operation of primary separators, secondary scrubbers
chemicals (if used), and general drum arrangement.
3. purify the steam to remove contaminants and re-
sidual moisture, Factors affecting steam separation
4. remove part of the water (blowdown) to control the Effective steam separation from the steam-water
boiler water chemistry (solids content), and mixture relies on certain design and operating factors.
5. provide limited water storage to accommodate The design factors include:
rapid changes in boiler load.
1. pressure,
However, the primary function of the steam drum 2. drum length and diameter,
is to permit the effective separation of steam and wa- 3. rate of steam generation,
ter. This may be accomplished by providing a large 4. average inlet steam quality,
steam-water surface for natural gravity-driven sepa- 5. type and arrangement of mechanical separators,
ration or by having sufficient space for mechanical 6. feedwater supply and steam discharge equipment
separation equipment. arrangement, and
High efficiency separation is critical in most boiler 7. arrangement of downcomer and riser connections
applications in order to: to the steam drum.
The operating factors include: carrying entrained water droplets into the steam out-
let (Fig. 20a). However, for the same arrangement at
1. pressure,
a higher rate of steam generation (Fig. 20b), there is
2. boiler load (steam flow),
insufficient time to attain either of these desirable
3. type of steam load,
results. Moreover, the dense upward traffic of steam
4. chemical analysis of boiler water, and
bubbles in the mixture may also cause a false water
5. water level.
level indication, as shown.
Primary separation equipment generally takes one of The effect of the riser or inlet connection locations
three forms: in relation to the water level is illustrated in diagrams
a and b of Fig. 21. Neither arrangement is likely to
1. natural gravity-driven separation,
yield desirable results in a drum where gravity alone
2. baffle-assisted separation, and
is used for separation.
3. high capacity mechanical separation.
From an economic standpoint, the diameter of a
Natural gravity-driven separation single drum may become prohibitive. To overcome this
limitation, several smaller steam drums may be used,
While simple in concept, natural steam-water sepa-
as shown in Fig. 22a, although this is no longer com-
ration is quite complex. It is strongly dependent upon
mon. However, in most boiler applications, natural
inlet velocities and inlet locations, average inlet steam
gravity-driven separation alone is generally uneco-
quality, water and steam outlet locations, and disen-
nomical, leading to the need for separation assistance.
gagement of liquid and steam above the nominal wa-
ter surface. Some of these effects are illustrated in Figs.
Baffle-assisted primary separation
20 and 21.
For a low rate of steam generation, up to about 3 Simple screens and baffle arrangements may be
ft/s (0.9 m/s) velocity of steam leaving the water sur- used to greatly improve the steam-water separation
face, there is sufficient time for the steam bubbles to process. Three relatively common baffle arrangements
separate from the mixture by gravity without being are illustrated in Fig. 22. In each case, the baffles
drawn into the discharge connections and without provide: 1) changes in direction, 2) more even distri-
Fig. 20 Effect of rate of steam generation on steam separation in a Fig. 21 Effect of location of discharge from risers on steam
boiler drum without separation devices. separation in a boiler drum without separation devices.
bution of the steam-water mixture, 3) added flow re- downcomers virtually free of steam bubbles, the maxi-
sistance, and 4) the maximum steam flow travel length mum net pumping head is available for producing flow
to enhance the gravity-driven separation process. in the circuits. The steam moving upward from the
Various combinations of perforated plates have also cylinder passes through a small primary corrugated
been used. The performance of these devices must be scrubber at the top of the cyclone (see Fig. 24) for ad-
determined by experimental evaluations and they are ditional separation. Under many operating conditions,
typically limited to smaller, low capacity boilers. no further separation is required.
When wide load fluctuations and water analysis
Mechanical primary separators variations are expected, large corrugated secondary
Centrifugal force or radial acceleration is used al- scrubbers may be installed at the top of the drum (see
most universally for modern steam-water separators. Fig. 19) to provide very high steam separation. These
Three types of separators are shown in Fig. 23: the scrubbers are also termed secondary separators. They
conical cyclone, the curved arm and the horizontal provide a large surface which intercepts water drop-
cyclone. The B&W vertical cyclone steam separator is lets as the steam flows sinuously between closely fit-
shown in more detail in Fig. 24. Vertical cyclones are ted plates. Steam velocity through the corrugated plate
arranged internally in rows along the length of the assembly is very low, so that water re-entrainment is
drum and the steam-water mixture is admitted tan- avoided. The collected water is drained from the bot-
gentially as shown in Fig. 19. The water forms a layer tom of the assembly to the water below.
against the cylinder walls and the steam moves to the One to four rows of cyclone separators are installed
core of the cylinder then upward. The water flows in boiler drums, with ample room for access. For
downward in the cylinder and is discharged through smaller boilers at lower pressures [100 psig (0.7 MPa
an annulus at the bottom, below the drum water level. gauge)], the separation rate of clean steam by single
With the water returning from drum storage to the and double rows of cyclone separators is approximately
Fig. 22 Simple types of primary steam separators in boiler drums: a) deflector baffle, b) alternate deflector baffle, and c) compartment baffle.
Superheater
(SH) SH
Economizer
(Econ) Furnace
Walls Circ Furn
(Furn) Pump
Orifices
(a) Natural Recirculation (b) Forced Recirculation
SH
SH
Sep
Circ Furn
Furn Pump
Econ Econ
Natural circulation
g c = 32.17 lbm ft/lbf s2 (l kg m/N s2)
In natural circulation, gravity acting on the den- ∆P = circuitry pressure loss due to friction, fluid
sity difference between the subcooled water in the acceleration and local losses, lb/ft2 (Pa)
downcomer and the steam-water mixture in the tube
circuits produces the driving force or pumping head As the heat input increases, circulation rate in-
to drive the flow. As shown in Fig. 27, a simplified creases until a maximum flow rate is reached (Fig. 28).
boiler circuit consists of an unheated leg or downcomer If higher heat inputs occur, they will result in larger
and heated boiler tubes. The water in the downcomer pressure losses in the heated tubes without correspond-
is subcooled through the mixing of the low tempera- ing increases in pressure differential. As a result, the
ture feedwater from the economizer with the satura- flow rate declines.
tion-temperature water discharged from the steam- Natural circulation boilers are designed to operate
water separators. Steam-water, two-phase flow is cre- in the region where increased heat input results in an
ated in the boiler tubes as a result of the heat input. increase in flow for all specified operating conditions.
Because the steam-water mixture has a lower average In this mode, a natural circulation system tends to be
density than the single-phase downcomer flow, a pres- self compensating for numerous variations in heat
sure differential or pumping pressure is created by the absorption. These can include sudden changes in load,
action of gravity and the water flows around the cir-
cuit. The flow increases or decreases until the pressure
losses in all boiler circuits are balanced by the avail-
able pumping pressure. For steady-state, incompress-
ible flow conditions, this balance takes the form:
g
(Z ρ d )
− ∫ 0Z ρ ( z ) dz =
gc (21)
( ∆Pfriction + ∆Pacceleration + ∆Plocal )
where
Z = total vertical elevation, ft (m)
z = incremental vertical elevations, ft (m)
ρ(z) = heated tube local fluid density, lb/ft3 (kg/m3)
ρ d = average downcomer fluid density, lb/ft3 (kg/
m3 ) Fig. 28 Typical relationship between circulation at a given pressure
g = acceleration of gravity, ft/s2 (m/s2) and steam production (arbitrary scale).
changes in heating surface cleanliness and changes Circulation design and evaluation
in burner operation. The furnace wall enclosure circuits are very impor-
Natural circulation is most effective where there is tant areas in a boiler. High constant heat flux condi-
a considerable difference in density between steam tions make uninterrupted cooling of furnace tubes
and water phases. As shown in Fig. 29, the potential essential. Inadequate cooling can result in rapid over-
for natural circulation flow remains very high even heating, cycling thermal stress failure, or material
at pressures of 3100 psi (21.4 MPa). failures from differential tube expansion. Sufficient
Forced circulation conservatism must be engineered into the system to
provide adequate cooling even during transient up-
In recirculating or once-through forced circulation set conditions. Simultaneously, the rated steam flow
systems, mechanical pumps provide the driving head conditions must be maintained at the drum outlet. Any
to overcome the pressure losses in the flow circuitry. of the circulation methods discussed may be used to
Unlike natural circulation, forced circulation does not cool the furnace waterwall tubes. In evaluating the
enjoy an inherent flow-compensating effect when circulation method selected for a particular situation,
heat input changes, i.e., flow does not increase signifi- the following general procedure can be used:
cantly with increasing heat input. This is because a
large portion of the total flow resistance in the boiler 1. The furnace geometry is set by the fuel and combus-
tubes arises from the flow distribution devices (usually tion system selected. (See Chapters 11, 14, 19 and 21.)
orifices) used to balance flow at the circuit inlets. The 2. Standardized components (furnace walls, headers,
large resistance of the flow distributors prevents signifi- drums, etc.) are selected to enclose the furnace ar-
cant increases in flow when heat absorption is increased. rangement as needed. (See Chapters 19 and 21.)
Forced circulation is, however, used where the boil- 3. The local heat absorption is evaluated based upon
ers are designed to operate near or above the critical the furnace geometry, fuel and firing method. Lo-
pressure [3200 psi (22.1 MPa)]. There are instances cal upset factors are evaluated based upon past
in the process and waste heat fields and in some spe- field experience. (See Chapter 4.)
cialized boiler designs where the use of circulating 4. Circulation calculations are performed using the
pumps and forced circulation can be economically at- pressure drop relationships.
tractive. At pressures above 3100 psi (21.4 MPa) a natu- 5. The calculated circulation results (velocities, steam
ral circulation system becomes increasingly large and qualities, etc.) are compared to the design criteria.
costly and a pump can be more economical. In addition, 6. The flow circuitry is modified and the circulation
the forced circulation principle can work effectively in re-evaluated until all of the design criteria are met.
both the supercritical and subcritical pressure ranges.
In forced recirculation there is a net thermal loss Some of the design criteria include:
because of the separate circulating pump. While prac- 1. Critical heat flux limits For recirculating systems,
tically all the energy required to drive the pumps re- CHF conditions are generally avoided. For once-
appears in the water as added enthalpy, this energy through systems, the temperature excursions at
originally came from the fuel at a conversion to use- CHF are accommodated as part of the design.
ful energy factor of less than 1.0. If an electric motor 2. Stability limits These limits generally indicate
drive is used, the net energy lost is about twice the acceptable pressure drop versus mass flow relation-
energy supplied to the pump motor for typical fossil ships to ensure positive flow in all circuits and to
fuel systems. avoid oscillating flow behavior.
3. Steam separator and steam drum limits These
indicate maximum steam and water flow rates to
individual steam-water separators and maximum
water flow to the drum downcomer locations to
ensure that steam carryunder and water carryover
will not be problems.
4. Minimum velocity limits Minimum circuit satu-
rated velocities assure that solids deposition, po-
tentially detrimental chemistry interactions, and
selected operating problems are minimized.
5. Sensitivity The system flow characteristic is
checked to ensure that flow increases with heat
input for all expected operating conditions.
Circulation is analyzed by dividing the boiler into
individual simple circuits – groups of tubes or circuits
with common end points and similar geometry and
heat absorption characteristics. The balanced flow
condition is the simultaneous solution of the flow char-
acteristics of all boiler circuits.
At the heart of a B&W circulation evaluation is a
Fig. 29 Effect of pressure on pumping head. circulation computer program that incorporates tech-
Critical flow
A two-phase flow parameter of particular importance
in nuclear reactor safety analysis and in the operation
of valves in many two-phase flow systems is the criti-
cal flow rate. This is the maximum possible flow rate
through an opening when the flow becomes choked
and further changes in upstream pressure no longer
affect the rate. For single-phase flows, the critical flow
rate is set by the sonic velocity. The analysis is based
upon the assumption that the flow is one dimensional,
homogeneous, at equilibrium and isentropic. These as-
sumptions result in the following relationships:
dP
Sonic velocity = C = d ρ gc (22)
s
dP
Fig. 30 Moody critical flow model for maximum steam-water flow rate.17 Critical flow = Gmax = ρ gc (23)
dρ
niques for calculating the single- and two-phase heat where
transfer and flow parameters discussed above and in
Chapters 3 and 4. With this program, a circulation C = velocity, ft/s (m/s)
model of the entire boiler is developed. Input into the P = pressure, lb/ft2 (Pa)
program is a geometric description of each boiler cir- ρ = fluid density, lb/ft3 (kg/m3)
cuit including descriptions of downcomers, supplies, gc = 32.17 lbm ft/lbf s2 (1 kg m/N s2)
risers, orifices, bends and swages, as well as individual Gmax = mass flux, lb/s ft2 (kg/m2 s)
tubes. Each of the circuits within the boiler is subjected
to the local variation in heat transfer through inputs However, when saturated water or a two-phase
based upon the furnace heat flux distribution. (See steam-water mixture is present, these simplifying as-
Chapter 4.) Given the geometry description and heat sumptions are no longer valid. The flow is heteroge-
absorption profile, the computer program determines neous and nonisentropic with strong interfacial trans-
the balanced steam-water flow to each circuit by solv- port and highly unstable conditions.
ing the energy, mass and momentum equations for the Moody’s analysis17 of steam-water critical flow is per-
model. The results of the program provide the detailed haps the most frequently used. It is based upon an
information on fluid properties, pressure drop and flow annular flow model with uniform axial velocities of
rates for each circuit so that they can be compared to each phase and equilibrium between the two phases.
the design criteria. Adjustments frequently made to A key element of the analysis involves maximizing the
improve the individual circuit circulation rates can flow rate with respect to the slip ratio and the pres-
include: changing the number of riser and supply sure. The results are presented in Fig. 30. The critical
connections, changing the number or type of steam steam-water flow rate is presented as a function of the
separators in the drum, adding orifices to the inlets to stagnation condition. Compared to experimental ob-
individual tubes, changing the drum internal baffling, servations, this correlation slightly overpredicts the
changing the operating pressure (if possible) and low- maximum discharge at low qualities (x < 0.1) and pre-
ering the feedwater temperature entering the drum. dicts reasonably accurately at moderate qualities (0.2
Once the steam-water circuitry is finalized, the detailed < x < 0.6), but tends to underpredict at higher quali-
mechanical design proceeds. ties (x > 0.6).
References
1. Collier, J.G., and Thome, J.R., Convective Boiling & Con- 4. Chen, J.C., “Correlation for boiling heat transfer to satu-
densation, Third Ed., Oxford University Press, Oxford, rated liquids in convective flow,” Industrial & Engineer-
United Kingdom, 1994. ing Chemistry Process & Design Development, Vol. 5, pp.
2. Jens, W.H., and Lottes, P.A., “Analysis of heat trans- 322-329, 1966.
fer, burnout, pressure drop, and density data for high pres- 5. Kitto, J.B., and Albrecht, M.J., “Elements of two-phase
sure water,” Argonne National Laboratory Report ANL- flow in fossil boilers,” Two-Phase Flow Heat Exchangers,
4627, May, 1951. Kakaç, S., Bergles, A.E. and Fernandes, E.O., Eds., Kluwer
3. Thom, J.R.S., et al., “Boiling in subcooled water dur- Academic Publishers, Dordrecht, The Netherlands, pp.
ing flow up heated tubes or annuli,” Proceedings of Insti- 495-552, 1988.
tute of Mechanical Engineers, Vol. 180, pp. 226-246, 1966.
6. Watson, G.B., Lee, R.A., and Wiener, M., “Critical heat 12. Thom, J.R.S., “Prediction of pressure drop during
flux in inclined and vertical smooth and ribbed tubes,” Pro- forced circulation boiling of water,” International Journal
ceedings of The Fifth International Heat Transfer Con- of Heat and Mass Transfer, Vol. 7, pp. 709-724, 1964.
ference, Vol. 4, Japan Society of Mechanical Engineers, To- 13. Martinelli, R.C., and Nelson, D.B., “Prediction of pres-
kyo, Japan, pp. 275-279, 1974. sure drop during forced-circulation boiling of water,” Trans-
7. Gellerstedt, J.S., et al., “Correlation of critical heat actions of the American Society of Mechanical Engineers
flux in a bundle cooled by pressurized water,” Two-Phase (ASME), pp. 695-702, 1948.
Flow and Heat Transfer in Rod Bundles, Schock, V.E., 14. Zuber, N., and Findlay, J.A., “Average volumetric con-
Ed., American Society of Mechanical Engineers (ASME), centration in two-phase flow systems,” Journal of Heat
New York, New York, pp. 63-71, 1969. Transfer, Vol. 87, pp. 453-468, 1965.
8. Wiener, M., “The latest developments in natural cir- 15. Chexal, B.J., Horowitz, J., and Lellouche, G.S., “An
culation boiler design,” Proceedings of The American assessment of eight void fraction models for vertical flows,”
Power Conference, Vol. 39, pp. 336-348, 1977. Electric Power Research Institute Report NSAC-107, De-
9. Swenson, H.S., Carver, J.R., and Kakarala, C.R., cember, 1986.
“Heat transfer to supercritical water in smooth-bore 16. Ledinegg, M., “Instability of flow during natural and
tubes,” Journal of Heat Transfer, Vol. 87, pp. 477-484, forced circulation,” Die Wärme, Vol. 61, No. 48, pp. 891-
1965. 898, 1938 (AEC-tr-1861, 1954).
10. Ackerman, J.W., “Pseudoboiling heat transfer to su- 17. Moody, F.J., “Maximum flow rate of a single compo-
percritical pressure water in smooth and ribbed tubes,” nent, two-phase mixture,” Journal of Heat Transfer, Vol.
Journal of Heat Transfer, Vol. 92, pp. 490-498, 1970. 87, pp. 134-142, 1965.
11. Hewitt, G.F., and Roberts, D.W., “Studies of two-phase
flow patterns by simultaneous x-ray and flash photogra-
phy,” Atomic Energy Research Establishment Report
M2159, HMSO, London, England, United Kingdom, 1969.
Bibliography
Bergles, A.E., et al., Two-Phase Flow and Heat Transfer Kitto, J.B., “Steam Generators,” Standard Handbook of
in the Power and Process Industries, Hemisphere, Wash- Powerplant Engineering, Second Ed., Elliot, T.C., Chen,
ington, D.C., August, 1981. K., and Swanekamp, R.C., McGraw-Hill, New York, New
Butterworth, D., and Hewitt, G.F., Eds., Two-Phase Flow York, 1998.
and Heat Transfer, Oxford University Press, Oxford, Lahey, R.T., and Moody, F.J., Thermal-Hydraulics of a
England, United Kingdom, 1977. Boiling Water Nuclear Reactor, Second Ed., American
Chen, J.C., Ed., Flow Boiling, Taylor and Francis Group, Nuclear Society (ANS), Hinsdale, Illinois, 1993.
New York, New York, 1996. Lokshin, V.A., Peterson, D.F., and Schwarz, A.L., Stan-
Hsu, Y-Y, and Graham, R.W., Transport Processes in dard Methods of Hydraulic Design for Power Boilers,
Boiling and Two-Phase Systems, Hemisphere, Washing- Hemisphere Publishing, New York, New York, 1988.
ton, D.C., 1976. Tong, L.S., Boiling Heat Transfer and Two-Phase Flow,
Kakaç, S., Boilers, Evaporators and Condensers, John John Wiley & Sons, New York, New York, 1965.
Wiley & Sons, New York, New York, 1991. Wallis, G.B., One-Dimension Two-Phase Flow, McGraw-
Hill, New York, New York, 1969.