200 Design Background
200 Design Background
200 Design Background
Abstract
This section contains the basic concepts which need to be part of your thinking as
you begin to design or prepare a specification for a new fired heater or retrofit. The
various design parameters are discussed in Section 210, Process and Economics
Considerations, where recommended values and rules of thumb are given.
Section 220, Combustion and Fuels, covers heating values, how to calculate the
combustion products a heater has to handle, and the effects of fuel type and quality.
The various types of Heat Transfer are described in Section 230; typical values and
methods of estimation are given. Thermal Efficiency, Section 240, includes the
effects of stack gas temperature and excess air and gives quick methods for esti-
mating efficiency. Waste Heat Recovery (Section 250) is restricted to those tech-
niques applicable to a fired heater; impacts on design and operation are included.
Emissions, Section 260, covers the major pollutants produced by a heater, how they
can be controlled, performance standards, and where to find information on regula-
tory approval. The mechanism of Draft is explained in Section 270, together with its
importance, measurement, and adjustment. Section 280, Computer Program
Abstracts, gives descriptions of two useful programs, FRNC5 and REFORM3, for
design and analysis of fired heaters.
Contents Page
When the bank of tubes is placed in the center of the firebox and is fired uniformly
from both sides, the ratio of peak to average flux will be 1.2. The advantage of
double firing is clear: For a given peak flux, the average flux will be 50% greater
than in the case of single firing. (This subject is covered further in a later section.)
In fire tube heaters, flux varies along the length of the tube, but not around the
circumference.
Typical peak flux values are in the range 15,000 to 22,000 Btu/hr(ft2). Average flux
values are 8,000 to 15,000. Recommendations are listed in Figure 200-1.
Inside Film Temperature. Flux, either peak or average, is a useful variable to fix
when specifying a heater or beginning a design. Excessively high flux will cause
tube overheating and damage to refractory. But what we are really concerned about,
in most cases, is maximum inside fluid film temperature. This is the maximum
temperature that the process fluid sees.
The maximum temperature difference across the inside film is computed as the
peak flux (adjusted to the basis of the inside tube area) divided by the inside
operating costs are justified by the increased resistance to tube coking. In hot oil
system heaters, where the oil is heated repeatedly over a long period of time, cold
oil velocity should be 7 to 14 ft/sec.
The cost for pumping a liquid is usually a secondary concern. The recommended
value of cold oil velocity may raise pumping cost above what might be considered
optimum. If this reduces coking, the extra cost will be more than repaid. Neverthe-
less, the pump should be specified on the basis that a layer of coke ¼-inch thick
will form.
Two-Phase Velocity. Suppose that the tube diameter of a crude unit vacuum heater
coil has been set by the liquid velocity criterion above. Through most of the coil,
the degree of vaporization will be below 5 weight percent and bulk velocity will be
in the range of 20 to 50 ft/sec. Pressure drop will be moderate, 0.1 to 0.2 psi per
foot of tube length. The flow regime discussed below should be acceptable.
In the last few tubes of the coil, after appreciable vaporization has occurred, mixed-
phase velocity will reach the range of 100 to 250 ft/sec. Pressure drop will become
excessive, 0.3 to 0.4 psi/ft, and sonic velocity may be approached. (See Section
473, “Transfer Line Sizing.”) A larger diameter should be considered for these
tubes. The benefit is that pressure drop is reduced, vaporization increases, and resi-
dence time at the highest temperatures decreases. However, you should check care-
fully to be sure that expected benefits are obtained. Where coking is the concern,
inside film temperature rather than bulk fluid temperature is the important factor. If
the larger tube cross- section results in decreased velocity and a lower inside heat
transfer coefficient, the inside film temperature will increase.
Gas Phase Velocity. When a gas is being heated (or a gas-liquid mixture with a
high volume fraction of gas), allowable pressure drop may be specified rather than
velocity. Power for compression is an important operating cost, particularly in
plants with gas recycle loops. Flow is typically turbulent and (if liquid is present) in
a dispersed (good) flow regime. Allowable pressure drop in the heater is deter-
mined by economics and the pressure drops found elsewhere in the process. We
typically use 80 psi for a hydrocraker heater operating at 2000 to 2500 psi. For
hydrotreaters at lower pressures, 50 to 60 psi is more common. In a catalytic
reformer at 150 psi, 15 psi would be specified as a total for three heaters in the
reactor loop.
Flow Regime. For two-phase mixtures, proper flow regime is important to prevent
vibration, pass stalling, and (for temperature-sensitive materials) coking. The
dispersed flow regimes, mist and froth, are considered best. Stratified flow (vapor
and liquid moving smoothly but tending to separate) is less desirable. The various
forms of slug or intermittent flow are worst and should be avoided. Check your
proposed designs or operations against the “Talmor Flow Maps,” Figures 200-2 and
200-3 Nomenclature is attached to the figures. Note: these figures are for use only
within Chevron and its subsidiaries.
Steam or Water Injection. We can increase velocity, improve flow regime, and
reduce coking tendency by injecting steam or water into process coils. Injection is
usually done only when heating residual stocks.
Example. The following example illustrates how flow velocity, flux, and multiple
pass considerations influence radiant section design.
Service Crude Unit Vacuum
Column Feed
Feed rate 100,000 BPOD
Radiant Section Duty 100 MMBH
Outlet Conditions 800°F, 20 psia
Specified Average Flux 10,000 Btu/hr(ft2)
Desired Cold Oil Velocity 10 ft/sec
Consider standard tube sizes, beginning with 4-inch tubes. Assume that the wall
thickness is that of Schedule 40 pipe. All tubes will be the same diameter.
The feed rate, at 60°F, is 6.50 ft3/sec. A 4-inch tube has an internal cross-section of
0.0884 sq ft. Use of eight parallel passes would give a flow velocity of 9.2 ft/sec,
close enough to the desired velocity.
Radiant section duty, 100 MMBH, divided by average flux, 10,000 Btu/hr (ft2),
gives 10,000 ft2 of tube surface required. A 4-inch tube has 1.178 ft2 of surface per
foot of length; therefore, the length of tube per pass is 1,061 feet. If we use 60-foot
tubes, there will be 17.7 tubes per pass.
Therefore, one solution is 4-inch tubes, 8 parallel passes, 18 radiant tubes per pass.
This is feasible; we have similar heaters. Is it optimum?
Figure 200-4 shows the results of using other tube sizes. For each size considered,
cold oil velocity dictates the number of parallel passes necessary. Average flux then
determines the number of tubes in each pass.
Fig. 200-4 Choosing Optimum Tube Size: Comparison of Various Tube Sizes for Example Given in Section 214
Velocity Number of Number of Total Number Volume of
Tube Size, in. ft/sec Passes Tubes per Pass of Tubes Tube Metal, ft3
3 9.7 13 14 182 169
3.5 9.5 10 16 160 179
4 9.2 8 18 144 190
5 9.4 5 23 115 206
6 8.1 4 24 96 223
6 10.8 3 32 96 223
8 9.4 2 37 74 259
12 8.4 1 50 50 328
Pumping cost (pressure drop) does not differ greatly from one case to another.
Installed cost of a tube coil is proportional to quantity of tube metal. A comparison
based only on volume of tube metal probably overstates the cost advantage of
smaller tubes; there are tube hanger and construction costs dependent on number of
tubes. Nevertheless, the coils of larger diameter tubes probably cost more, and
should be evaluated in a cost estimate if necessary.
It is possible, by using 12-inch tubes, to have a single pass heater; however, this is
far from optimum. The coil would be much more expensive than those made of
smaller tubes. Further, thermal cracking would increase as “residence time”
increased. Because the residence time of the process fluid (the time the process
fluid is in the tube) is roughly proportional to the number of tubes per pass, resi-
dence time in the 12-inch tubes would be about three times as long as in 3- to
4-inch diameter tubes.
For tube sizes below 4 inches, the capital cost and residence time benefits become
marginal. Eight is the largest number of parallel passes we have accepted for this
service.
Example Continued. Using 4-inch tubes, 8 parallel passes, and 18 tubes per pass
as our choice, we now look at the effect of increasing the diameter of the last two
tubes of each pass (Figure 200-5). The purpose is to decrease pressure drop,
promote vaporization, and decrease residence time at high temperature in these last
two tubes. (See the paragraph, Two-Phase Velocity, in Section 212.)
Referring to Figure 200-5, with 4-inch tubes throughout, sonic velocity prevents
obtaining the desired outlet pressure of 20 psia. Using either 5- or 6-inch tubes
removes this constraint. Pressure drop in the last two tubes decreases greatly as
diameter increases, but total pressure drop decreases only a little; the stock vapor-
izes more in the earlier tubes of the coil. Vaporization does not increase further as
tube size changes from 5 to 6 inches. This is because the outlet pressure is fixed
(assumed to be limited by pressure drop in outlet manifolds and the transfer line.)
Equivalent residence time decreases moderately as we move to 5-inch tubes but
only a little as we go to 6-inch tubes. Equivalent residence time is a relative
measure of thermal cracking tendency: the time at 800°F that would produce the
same amount of cracking. It is based on bulk fluid temperature in the tubes.
Finally, we note the tube diameter effects on inside heat transfer coefficient and
inside film temperature; both are undesirable. A 21°F increase in film temperature
(the difference between the 4-inch and 6-inch diameter cases) will significantly
increase the rate of tube coking.
In summary, increasing diameter of the last two tubes will decrease pressure drop
and thermal cracking in the bulk fluid; it will decrease cracked gas production in a
crude unit. On the other hand, tube wall temperature and coking will increase. We
recommend you increase diameter to the extent needed to eliminate a sonic flow
constraint, but not further.
215 Turndown
Turndown is the ability to operate at lower than design rate. Turndown is not often a
problem on the firebox side. Fuel oil burners may be turned down to typically 33%
of design capacity, gas burners to 10%.
A heater with a single-phase process stream will have less turndown difficulty than
one with a two-phase stream. This is because low velocity in the two-phase case
may result in a poor flow regime, pulsating flow, and increased cracking and
coking. If the furnace was designed with tube velocity at the upper end of the
recommended range, turndown capability will be greater. Steam injection is often
added to the radiant coil of a heavy oil heater to raise the bulk flow velocity and
increase turndown ability.
If multiple passes are flow controlled, turndown can move the controllers out of
their optimum range, leading to pass stalling. Horizontal tubes with flow downward
through the coil are most resistant to stalling. Vertical tubes are least resistant.
Redundant fan drivers are not installed. However, if more than one induced draft
fan is used, electric motors can be connected to different power supplies to improve
reliability. Alternatively, one or more drivers can be steam turbines. In many cases,
a normally forced draft heater can be operated at reduced duty under natural draft.
The higher or gross heating value (HHV) is that obtained with the product water
present as a liquid.
The lower or net heating value (LHV) assumes the water remains in the vapor state.
The difference is the water's “latent heat of vaporization.” In actual practice, the
lower heating value is used in designing and rating process heaters because the
product water normally is not condensed on heat transfer surfaces and leaves the
stack as a vapor.
For hydrocarbon gases and mixtures of hydrocarbon gases, the lower heating value
on a volume basis (Btu/scf) is linearly related to the specific gravity of the fuel
(air=1) as shown in Figure 200-6. But on a mass basis, the net heating value (LHV)
of all hydrocarbon gases is close to 20,000 Btu/lb.
Fig. 200-6 Heating Value Versus Specific Gravity for Hydrocarbon Mixtures
Fuel oil always contains some noncombustible materials like water, ash or metals.
So the heating value of fuel oil is generally a little lower, typically around
18,000 Btu/lb. (Energy is often expressed in Btu per “equivalent fuel oil” barrels
(EFO). Note that an EFO is approximately equivalent to 6MM BTU's.)
is very fast and that of carbon slow. The slower burning free carbon is heated to a
bright yellow incandescence.
Oil flames are typically characterized by a very bright yellow incandescence
because there is so much free carbon around. Somewhat yellow flames are typical
of the heavier gaseous fuels, too, but yellow flames in gas burning are more often a
tipoff of poor or inadequate air mixing.
Process gas (fuel gas) has always been preferred because it is easier to handle,
burns cleaner and is less expensive. Many locations use some natural gas. However,
most natural gas contracts provide that industrial gas can be cut off on short notice
whenever the demand by domestic users exceeds the supply. Consequently, those
locations have some provisions for backup oil burning in selectedgas is curtailed.
Steam boilers are often chosen for supplemental oil firing when the local facilities
are subject to possible natural gas curtailment because they are normally attended,
generally use larger quantities of fuel, and operate more nearly at constant load. If
more units have to be switched to oil, the process heaters usually selected are large
users in less severe services with steady loads, like atmospheric crude heaters.
Design of oil fired heaters is more complicated than that for gas fired heaters:
• The oil requires separate oil guns and steam for atomization, along with sepa-
rate controls.
• The flame for a given heat release is longer and of a less predictable shape than
the flame in a gas fired heater, which affects firebox geometry. For example,
when heating a temperature-sensitive process fluid with gas, we often specify a
clearance of 5 feet between burner centerline and nearest tube surface. With oil
the figure is 6 feet.
• Because oil burning produces ash and soot, design of the heater's convection
section should include provision for tube cleaning by onstream water washing
or soot blowing.
• Choice of refractory material may be affected due to chemical attack from
components in the oil. See a specialist for more information.
• Because of the soot produced with oil, the maximum number of fins per inch
of convection tubes is only three (maximum is five with gas). This affects
thermal efficiency and draft.
• Based on operating experience, the design excess air with gas fuel is 10%; with
oil, 15%. This also affects the draft requirement.
• Fuel oil has a higher sulfur content than gas. This may affect the minimum
temperature allowed for steel surfaces in direct contact with flue gas. The
concern is sulfuric acid condensation on metal surfaces, which may present
corrosion and air pollution problems. The table below tells you the minimum
required steel surface temperature (not flue gas temperature) which is neces-
sary to prevent acid condensation. The temperature varies depending on the
sulfur content of fuel (measured in percent of weight).
Sulfur Temperature
1% 325°F
<1% 300°F
<0.1% 275°F
• The vanadium often present in fuel oil may cause severe corrosion of hot metal
parts like tube supports.
• Oil guns foul quickly, requiring frequent attention.
• Oil firing requires much more maintenance time and operator attention.
226 Naphtha
Burning naphtha is not new to the oil industry. In Europe, naphtha has been burned
in process heaters for a long time. It is unusual in American refineries only because
fuel gas has always been in ample supply and naphtha has been regarded as more
valuable as a gasoline blend stock.
Naphtha is highly flammable. It flashes quickly in contact with hot lines or equip-
ment. So burners are equipped with safety interlocks. The safety interlocks prevent
burner guns from being pulled before they are purged free of liquid fuel. Naphtha
can be fired either as a liquid or as a vapor in a conventional gas burner. But the two
phases cannot be interchanged or mixed: they must be 100% one or the other
(liquid or vapor). Today's trend is strongly toward liquid burning when naphtha
is used.
Liquid naphtha is burned in conventional steam-atomized oil guns. The fuel pres-
sure must be kept high enough to keep the naphtha liquid until it reaches the burner.
Small amounts of vaporized naphtha vapor lock the atomizer. The result is unstable
burner operation and blowout.
The following may be considered in evaluating liquid versus vaporized naphtha: A
vaporizer and steam-traced/insulated piping are not needed in a liquid system, so it
usually costs less. However, the cost of new liquid naphtha burners may offset the
cost of the vaporizer, and a vaporized naphtha system may therefore be economic.
In order to use the old gas burners without modification, the vaporized naphtha
heating value has to be reduced to that of the gas. The needed large amounts of
diluent steam imply a significant operating cost penalty.
Where sour process gas is sweetened in amine units, then used as fuel gas, there is
the possibility of contamination of the fuel supply. Amine plant upsets can allow
some H2S breakthrough. Alternatively, the fuel gas may become extremely dirty if
sour gas is bypassed around an amine unit.
We have had little luck with additives. One operating center reported success in
dealing with a scale plugging problem. But at another center, use of additives
proved disappointing.
Potential solutions to fuel quality problems are:
• Enlarge the burner ports (possibly you get poor flames)
• Clean up the gas
• Install strainers (they must be kept clean)
• Try chemical cleaning, a good solution, if the deposits come from debris accu-
mulated over years. But if particles or globs are just flowing with the gas at a
rate sufficient to plug burners in weeks, chemical cleaning is not a solution.
• Try a hardware solution: go to a waste oil burner (huge ports like in a flame
thrower) for burning gas. You would get no turndown, excess air would go up,
efficiency would go down, and you might have to rebuild the fireboxes so the
flames would fit into them. But the burners probably wouldn't plug.
• Change from our typical raw gas burners (small ports) to premix burners
(larger ports). This may relieve the symptoms, but operating centers have plug-
ging problems with premix burners, too. Also, premix burners are intolerant
of the large quantities of hydrogen that we often encounter in refinery fuel
systems. They backfire and blow out when the hydrogen content is too high.
Unfortunately, no foolproof method of preventing burner plugging has yet been
developed. In actual practice operators periodically must clean burners manually to
assure even combustion and optimum heater efficiency.
231 Radiation
Radiation is the primary mode of heat transfer in the firebox and a secondary effect
in the hotter portions of the convection section. The variables used in correlating
radiant heat transfer are described here, with typical values.
The Basic Equation. Radiant heat transfer from hot surface “a” to cold surface “b”
is given by the following equation:
qR = σ AF (Ta4 - Tb4)
(Eq. 200-2)
where:
qR = radiant heat transfer rate, Btu/hr
σ = the Stefan-Boltzmann constant, 1.73 × 10-9
A = area of either surface, ft2
F = radiant exchange factor (discussed in the following example)
Ta = absolute temperature of surface a, °R
Tb = absolute temperature of surface b, °R
Note that degrees Rankine (°R) equals degrees Fahrenheit (°F) plus 460°.
(°R = °F + 460°).
The radiant exchange factor, F, depends on the relative area and arrangement of the
two surfaces and on the emissivity and absorptivity of each.
Equation for a Firebox. The basic equation is adapted to the radiant section of a
fired heater. Surface “a” becomes a radiating volume of hot flue gas. Surface “b”
becomes a tube bank, called a “cold plane.”
spacing by the effective length of a tube. The product of α and Acp is called the
“equivalent cold plane surface.”
Radiant Exchange Factor. F is roughly proportional to flue gas emissivity. It also
depends on the ratio of equivalent cold plane surface to total refractory surface. As
that ratio decreases, F increases. Explanation: Being able to “see” hot refractory
surface increases heat absorption by the cold tubes. Values of F are usually in the
range 0.3 to 0.7. (See Appendix B for more information.)
Gas Emissivity. This factor is very roughly proportional to the partial pressure of
radiating components, P, and to the “mean beam length”, L, both of which are
explained just below. Emissivity decreases as gas temperature increases. We have
little control over it. Appendix B contains a graphical correlation. Values in the
range 0.5 to 0.6 are common.
Partial Pressure of Radiating Components, P. Only water vapor and carbon
dioxide radiate significantly. For a typical fuel gas and 10% excess air, P equals
0.26 atm. More air decreases P which decreases heat transfer, another reason to
control excess air.
Mean Beam Length, L. This is a measure of the depth of the blanket of flue gas
which is radiating heat to the tubes. For fireboxes which do not differ greatly from
cubes, use:
L = 3.6 V/Ae
(Eq. 200-4)
where:
L = mean beam length, ft
V = firebox volume, ft3
Ae = total firebox envelope (inside surface) area, ft2
If one or two dimensions greatly exceed the other(s), the mean beam length is a
function of just the smallest dimension. For a vertical cylindrical heater, use the
diameter for L.
Temperatures in the Radiant Section. Our basic firebox model assumes that the
firebox is “well mixed.” Hot flue gas from the burners is immediately mixed with
the total flue gas volume of the firebox; thus, the firebox is of uniform temperature
throughout. Further, the gas is non-luminous; there are no flames. The result is that
Tg in Equation 200-3 is the firebox outlet or “bridgewall” temperature. We know
that this is overly conservative in most cases; much of the tube surface “sees” a gas
that is hotter. In hand calculations and in FRNC5, we apply an increment to the
bridgewall temperature to get Tg. In a box heater, the increment might be 50°F. In a
vertical cylindrical firebox, in which flue gas comes closer to plug flow, the incre-
ment could be 200°F.
Steam reformer heaters are constructed to provide control of heat flux in “zones”
along the length of vertical tubes. A more rigorous design technique is used, which
involves a different flue gas temperature in each zone.
For hand calculation of temperature in well-mixed fireboxes, Tt, the tube surface
temperature, is approximated as the average process fluid temperature (average of
temperatures in and out of the radiant coil) plus 100°F.
Example Firebox Calculation. Consider a box heater similar to that shown in
Figure 200-3. The firebox calculation is in fact an iterative, trial and error proce-
dure. In the current case, we will make the necessary assumptions and do one
iteration.
• Firebox dimensions: 15 feet high × 15 feet wide × 40 feet long (parallel to
centerlines of tubes). Convection section is 5 feet wide.
• Tubes: 13 in each bank, 6-inch inside diameter on 12-inch spacing, 40 feet
long.
• Fuel: Gas with 10% excess air.
• Temperatures: Process in, 600°F; process out, 780°F; bridgewall, 1600°F. (In
an iterative calculation, bridgewall temperature is adjusted to obtain a heat
balance on the flue gas.)
Total refractory surface area, Ae, is the sum of the following areas: side walls, floor,
roof except entry to convection section, and end walls. Ae = (2)(15)(40) + (15)(40) +
(2)(5)(40) + (2)(15)(15) = 2650 ft2.
Firebox volume, V = (15)(15)(40) = 9000 ft3.
Cold plane area, Acp = (2)(13 - 1)(1.0)(40) = 960 ft2.
Mean beam length, L, = (3.6)(9000)/2650 = 12.23 ft = 3.73 m (from
Equation 200-4).
Partial pressure of radiating components, P = 0.255 atm (from Appendix A,
Figure A-3).
PL = (0.255)(3.73) = 0.95 atm-m.
Effective gas temperature, Tg, is bridgewall temperature + 50°F = 1600 + 50 =
1650°F = 2110°R = 899°C.
Gas emissivity = 0.475 (from Appendix A, Figure A-4).
Center-to-center distance/tube OD = 12/6.625 = 1.81.
Absorption efficiency, a = 0.925 (from Appendix A, Figure A-2).
α Acp/Ae = (0.925)(960)/2650 = 0.335.
Radiant exchange factor, F = 0.71 (from Appendix A, Figure A-5).
Effective tubewall temperature, Tt, = (600 + 780)/2 + 100 = 790°F = 1250°R.
Finally, from Equation 200-3, the heat transfer rate is found: qR = (1.73 × 10-9)
(0.925)(960)(0.71)(21104 - 12504) = 18,960,000 Btu/hr.
Since the total tube area is 1804 ft2, average flux is 10,510 Btu/hr(ft2).
232 Convection
Convection is a small but significant part of firebox heat transfer. It is a major
component of transfer to shock tubes and the dominant mechanism elsewhere in the
convection section. Heat transfer inside tubes is by forced convection; losses from
the furnace casing and stack surface occur by natural convection.
Hand calculation methods for convection to bare tubes are in Appendix A;
Appendix B covers extended surface (finned and studded) tubes.
Convection in the Firebox. While burners are the major influence on circulation
patterns within the firebox, convection currents are significant. Flue gas, cooled by
convection to the tubes, tends to flow downward in the vicinity of tube banks.
These currents can have unexpected effects on burner flames, causing them to
“lean” or changing their shape; this may cause flame impingement on tubes.
Firebox convective heat transfer is handled simply, by adding a term to
Equation 200-3:
Heat Transfer to Finned Tubes. Appendix B contains methods suitable for use by
hand. Equation 200-7 is applied to finned tubes. The flue gas mass velocity is
computed on the basis of the area left after subtracting cross sectional areas of both
the tube and the fins (but not the space between fins). The value of do is not
affected by the fins. The effectiveness of the fins is reduced by conductive resis-
tance as discussed in Section 233.
Convection Inside Tubes. Inside heat transfer coefficients may be calculated by
standard methods found in handbooks and in Appendix C of API-530. It is often
unnecessary when making rough hand calculations. Inside coefficients are usually
in the range 100 to 500 Btu/hr(ft2)(°F); outside coefficients run 2 to 5. Outside resis-
tance is by far the controlling one with bare tubes. Even when extended surfaces are
used, the outside resistance dominates.
An equation for convective transfer to a two-phase mixture is listed here, however,
because it is often furnished to vendors as part of a quotation request:
233 Conduction
The important applications of conduction in heater design and analysis are in
extended surface, in coke layer effects, and (rarely) in the resistance of the tube
wall to heat transfer.
The Basic Equation. For flow through a solid, from one surface to another:
q = k A t/x
(Eq. 200-9)
where:
q = conductive heat transfer rate, Btu/hr
k = thermal conductivity, Btu/hr(ft)(°F)
A = cross sectional area through which heat flows, ft2
t = temperature difference between surfaces, °F
x = thickness of the solid, ft
Some order-of-magnitude values of thermal conductivity are shown in Figure 200-8.
Fig. 200-9 Total Surface Area, Efficiency and Effective Coefficients for Commonly Used Fins
Fin Thickness, Total Surface Ht. Trans.
in. Fin Height, in. Density, Fins/in. Area, ft2/ft. Efficiency, % Coefficient
Bare — — 1.18 — 5.0
0.10 0.75 2.5 7.01 93 28.0
0.10 0.75 3 8.18 93 32.6
0.05 0.75 4 9.38 86 34.9
0.05 0.75 5 11.43 86 42.4
0.05 1.00 5 14.84 79 50.7
The figure was derived from nonproprietary information of the ESCOA Fin-tube
Corporation.
Remember that fouling strongly influences the effectiveness of fins, particularly
closely spaced ones.
Fin tip temperature is important only when it exceeds the temperature limit of the
fin material. Approximately,
Fins are continuously welded to the tubes. There should be minimal bond resistance.
Tube Wall Resistance. Ordinarily, this is a small part of the total heat transfer resis-
tance. Temperature difference through the tube wall is usually 10 to 20°F. Catalyst
tubes of a steam reformer heater are an exception. Thermal conductivity is at the
lower end of the range for ferrous metals, and walls are thick because of low allow-
able stresses. The temperature difference can be as large as 100°F.
Coking Resistance. Hot tube walls are usually the result of coking in tubes. At a
flux of 20,000 Btu/hr(ft2), 0.25 inch of coke will raise tube wall temperature 140°F.
241 General
The thermal efficiency of a heater is the heat absorbed divided by the heat input.
Heat absorbed is heat transferred to process streams, to steam generation and super-
heating coils, to boiler feedwater, etc. Heat input consists mainly of the net or low
heating value of the fuel burned. (See definitions, Section 221.)
We look for efficiency of around 90% in most cases.
Contributors to poor efficiency, in order of importance, are the following:
High Stack Temperature. Efficiency of 90% means cooling stack gas to the 300 to
400°F range. This requires a heat receiver or “sink” whose inlet temperature is
50 to 100°F cooler than the stack outlet temperatures. Regardless of its radiant
section design, a heater must have a convection section to achieve this kind of
efficiency.
In most cases, the most economically attractive waste heat recovery is (1) additional
process heating, followed by (2) steam generation and boiler feedwater heating,
followed by (3) air preheat. Waste heat recovery is discussed in Part 2 of this
manual.
See Figure 200-10, which shows the effect of stack temperature on efficiency.
High Excess Air. Stoichiometric air is the amount needed to burn the fuel
according to the chemical reactions of combustion. Any additional air is “excess.”
Some excess air is needed to ensure safe operation and complete combustion. With
natural draft, a heater typically needs 10% excess air if burning gas, 15% if burning
oil. With forced draft, the figures are 5% for gas and 10% for oil. (“Oil” means any
liquid fuel.)
Excess air is sometimes quoted in terms of oxygen content of the flue gas. Roughly,
10% excess air means 2% oxygen in the flue gas. The exact amount can be found
by combustion calculations; see Section 220. The stack damper and air registers
should be adjusted to obtain only the required excess air. (See Section 274 for
adjustments to natural draft.)
Any air beyond the required excess represents an efficiency loss. This air enters the
heater (or the air preheater) at ambient temperature and leaves at the stack tempera-
ture. The higher the stack temperature, the greater the additional loss caused by
high excess air. If stack gas is cooled to the 300 to 400°F range, the effect of excess
air becomes moderate.
Air Leaks. If the heater is operated properly, with negative pressure throughout,
any leaks will consist of air coming into the heater rather than flue gas leaking out.
Air leakage into the firebox adds to excess air. Burner operation cannot be safely
adjusted to offset it. (Heater casing plates should be seal welded during construc-
tion to minimize air leakage.) Air leakage at the stack does not affect thermal effi-
ciency, but it reduces draft.
Thermal Efficiency, %
Stack Temperature, °F
Note:
1. 2% Radiation Loss is Assumed
2. See the text for Conversion to % Excess O2
Note there is a smoke test procedure in the Appendix to detect leaks during turn-
arounds. Infrared scanning is being tested as a method to determine leaks, but there
are no definite conclusions at this time.
Heat Loss from Casing. Your specification for a new heater will give a maximum
temperature for the outside surface under defined ambient conditions. This tempera-
ture is in the range 140 to 180°F, and insulation is designed accordingly. The main
concerns are personnel protection and damage to structural members. Heat loss will
be 1 or 2% of the heat released. If insulation is in poor condition, the loss can be
appreciable.
Generally, 180°F will be sufficiently low to avoid excessive heat loss. Reducing the
casing temperature to 140°F may not be economic due to the extra insulation
required. Note, however, that temperatures over 140°F require protective screens
for safety reasons.
Turndown. If you turn down a heater from its design duty and maintain control of
excess air, thermal efficiency will increase slightly. (It's difficult to control excess
air as a heater is turned down.) Assume that process flowrate and fuel rate are
decreased in proportion. Convective heat transfer coefficients will decrease
according to the 0.7 to 0.8 power of velocity, but the required duty has decreased in
proportion to turndown. The result is a small efficiency gain. Usually the efficiency
of the radiant section increases relative to the convection section.
242 Definitions
Thermal Efficiency: This is defined as the total heat absorbed divided by the total
heat input.
Total Heat Input: This is the net heat of combustion of the fuel plus the sensible
heat of the air, fuel, and fuel atomizing medium. Net heat of combustion is based on
complete combustion of fuel. Carbon burns to carbon dioxide; carbon monoxide
breakthrough is negligible. Sensible heat is measured above the datum temperature,
which is 60°F. These sensible heat effects are usually small and are ignored. An
exception is air which has been preheated by another unit, such as gas turbine
exhaust.
Total Heat Absorbed: This is the heat transferred to the process coils and to utility
coils such as boiler feedwater heating, steam generation, and steam super-heating.
Air preheat, unless it's for another heater, is not a part of heat absorbed; a fired
heater's own air preheater, if there is one, is inside the heat balance envelope.
Total Heat Loss: This is the sum of radiation heat loss and stack gas heat loss.
“Radiation loss” (actually loss by convection from the heater casing) is usually
assumed as 2% of the net heating value of the fuel. The sum of total heat absorbed
and total heat loss equals total heat input.
260 Emissions
In years gone by, burner optimization generally meant complete combustion with
the smallest practical flame volume and minimum practical amount of excess air.
Today we also want to reduce pollutant emissions.
The major pollutants are particulates, carbon monoxide, hydrocarbons, oxides of
sulfur, oxides of nitrogen, aldehydes, ammonia and noise. Normally these pollut-
ants are present in only trace quantities and the effect of their production on
combustion efficiency is insignificant.
The emphasis in this discussion is on the control of nitrogen oxides. Properly oper-
ated and maintained refinery process heaters and burners do not normally emit
amounts of particulates, carbon monoxide, hydrocarbons, aldehydes, ammonia,
smoke or haze beyond legal limits. As far as the oxides of sulfur are concerned, the
sulfur is generally removed from the fuel when necessary. Today the significant
emissions problem that often cannot be adequately controlled by proper operation
and maintenance or by fuel treatment is the oxides of nitrogen.
Noise is covered in Section 500 “Burners and Plenums.”
in many process heaters. Mixing is also a problem. The SNR process reduces NOx
by only 40 to 60%.
When the ammonia is injected into the flue gas and the mixture is then passed at the
proper temperature over a catalyst bed, the process is called Selective Catalytic
Reduction or “SCR.” Sometimes when carbon monoxide limits are very tight, a
CO converter is also included in the Selective Catalytic Reduction Flue Gas Treat-
ment Plant package. Reliable, trouble-free 80 to 90 percent NOx reduction is
typical. The only drawback is that SCR is by far the most costly of the NOx reduc-
tion technologies.
There are a variety of catalysts, designers and suppliers of SCR systems. The
design contractors specify the type of catalyst along with other design details. The
catalyst bed typcially runs in the 600 to 750°F range.
The Company has few SCR systems in operation and experience is scarce. We
know SCR is a reliable but expensive technology, typically costing $5 per pound of
NOx reduced in new construction and $10-$20 per pound for retrofits. Combustion
modification (low NOx burners, flue gas recirculation, etc.) is by far the cheaper
alternative when they can be applied.
270 Draft
271 Definition
Flue gas does not pass through a natural draft heater and rise up the stack of its own
accord. It is pushed by the atmosphere outside, which enters through the burner air
registers. The motive pressure is draft; at any point in the heater, draft is the differ-
ence between internal absolute pressure and pressure of the atmosphere outside, at
the same elevation. Positive draft, or “negative pressure,” is the desired condition
where absolute pressure inside the heater is lower than that outside. “Positive pres-
sure,” the reverse, is to be avoided in order to prevent personnel hazard and damage
to the structure.
Section 460, Stacks and Ducts, has additional information on draft calculations.
Proceeding through the heater from a point just outside the burner air registers to a
point just outside the top of the stack, we see that the total pressure change is the
same as that described in the paragraph above, 0.064 psi. Flue gas flows “uphill” as
a result of decreasing pressure. There are sizable pressure drops through the
burners, convection section, and stack damper. Pressure drops through the unob-
structed volumes, the firebox and stack, are much lower than in the atmosphere
outside. Although flue gas has about the same molecular weight as air, the much
higher temperature of flue gas makes it less dense.
The difference between the two curves just described is the draft at any elevation in
the heater. Draft is also plotted in Figure 200-11. Because psi is an inconvenient
unit for small pressure differences, we have switched over to inches of water
column, abbreviated in.wc, the common unit for draft calculations. A cubic foot of
water weighs 62.4 lb, so an inch of water column is 62.4/1728 or 0.0361 psi. We
notice that the point of minimum draft (not minimum pressure) is almost always
just below the convection section. The one exception is a heater operating well
below design; minimum draft may occur just below the damper. The location of
this minimum draft is sometimes called the “pinch point.”
When calculating pressure drop per convection tube row, base G on the minimum
cross-section between tubes, taking into account the additional space taken up by
extended surface. Base stack losses on the stack mass velocity, which is usually in
the range 0.75 to 1.0 lb/sec/ft2. The stack friction loss, in velocity heads, is 0.02 ×
(height/diameter).
Example. Refer again to Figure 200-11. The various components of the draft
profile will be reviewed. We begin at the arch, where negative pressure of 0.05
inches of water column is being maintained. According to Equation 200-12, thirty
feet of height in the firebox at an average temperature of 1500°F provides 0.33 inch
of additional draft or a total of 0.38 inch at the floor.
This is the operating pressure drop of the burners. If the burners were unable to
operate at this draft, then a higher draft (more negative pressure) would have to be
maintained at the arch.
In the convection section, assumed to be at an average temperature of 700°F,
flue gas rate and tube spacing are such that maximum mass velocity equals
0.565 lb/sec/ft2, corresponding to a flue gas density of 0.0342 lb/ft3 and a linear
velocity of 16.5 ft/sec. One velocity head is computed by Equation 200-13:
Twenty tube rows (not the 6 shown schematically in Figure 200-11) at 0.5 velocity
head per row equals 0.28 inches of water column. However, this is partially offset
by the draft effect of 16 vertical feet of flue gas at an average temperature of 700°F.
According to Equation 200-12, this is 0.008 inches of water column per foot or
0.13 inch for the whole section, which leaves a net change of 0.15 inches of water
column. In sum, as flue gas passes through the convection section, draft increases
from 0.05 inch of water column at the arch to 0.20 inch just ahead of the stack
entrance.
We assume that the stack has been sized for a mass velocity of 0.75 lb/sec/ft2,
resulting in a diameter of 4.0 feet. Flue gas temperature is now 400°F, and density
is 0.0462 lb/ft3. According to Equation 200-13, a velocity head is 0.0365 in. wc.
The draft effect of 74 feet of stack at 0.006 inch of water column per foot is 0.44
inch. Friction loss through the stack is (0.02) (74/4) = 0.37 velocity head; entrance
and exit losses total 1.50 heads. Total losses then are (1.50 + 0.37) (0.0365) = 0.07
in. wc. Recalling that draft at the stack entrance is 0.20 inch, pressure drop across
the damper is found by difference: 0.44 - 0.20 - 0.07 = 0.17 in. wc. This is equiva-
lent to 4.6 velocity heads, well above the minimum.
Rules of Thumb. Note the following rules:
• Minimum draft at any point should be 0.05 to 0.10 in. wc.
• Draft loss through a typical natural draft burner (including plenum if present)
is 0.2 to 0.5 in. wc.
Stack height should be such that, at design duty in the heater, draft loss at the
damper is 0.2 to 0.3 in. wc.
Fig. 200-12 Fired Heater Operation Checklist for Control of Draft and Excess O2: Natural Draft Heaters Courtesy of
the John Zink Company
290 References
1. Chem. Eng. vol. 85, no. 18, p. 129.
2. E. Talmor, Predict and Control Combustion Zone Hot Spots in Petroleum
Processing Heaters, Chevron Research Company, Sept. 26, 1978.
3. R. F. Cleverdon, FRNC5 User's Guide, Process Furnace Program, August 25,
1982. Available from ETD Staff Support Services.
4. R. F. Cleverdon, REFORM3 Users' Guide, Steam-Hydrocarbon Reformer
Furnace Program, July 17, 1987. Available from Central Files at CRC.