heat transfer in fire-tube boilers - advanced boiler technology-warga ...

HEAT TRANSFER IN FIRE-TUBE BOILERS
Doctoral dissertation
Defended in December 2000 at
University of Ljubljana, Slovenia, Department for Mechanical Engineering
by
Zeljko Warga, Slovenia
Summary
Mentor:
Prof. Peter Novak, Slovenia
Co-mentor:
Prof. Wladimir Linzer, Institute for Thermal Engineering,
Technical University Vienna, Austria
A research of heat transfer in fire-tube boilers was conducted, a mathematical
model for the heat transfer in the fire-tube boilers was developed, and a
corresponding computer software model was written. The model enables a more
accurate analytical assessment of the impact of coiled-wire turbulence promoters
on the heat transfer and pressure drop in boiler tubes, it allows for a more
accurate determination of the mean radiant temperature in boiler segments and
takes in account the heat transfer from impinging jet since the latter was found to
have a significant impact on heat transfer in fire-tube boilers. The model was
verified on several test boilers of sizes ranging from smaller residential hot water
units to a larger industrial steam unit boiler. The comparison between the
modeled values with those measured was very good.
1

ACKNOWLEDGMENT
The author thanks the following who contributed in making of this work:
• Prof. Peter Novak, Department of Mechanical Engineering, University of
Ljubljana, Slovenia;
• Prof. Wladimir Linzer, Institute for Thermal Engineering, Technical University
of Vienna, Austria;
• Dr. Vincenc Butala, Prof. Joze Duhovnik, late Dr. Joze Zupancic and late Prof.
Hinko Muren, Department of Mechanical Engineering, University of Ljubljana,
Slovenia
• Weishaupt GmbH, a corporation of Germany;
• R.W. Beckett Corporation, Ohio, U.S.A.;
• Dr. Thomas Butcher, Brookhaven National Laboratory, Energy Sciences and
Technology Department, New York, U.S.A.;
• Tucson Electric Power Company, a subsidiary of UniSource Energy
Corporation, Tucson, Arizona, U.S.A.;
• Tucson Medical Center, Tucson, Arizona, U.S.A.;
• inoMETAL, a company of Slovenia;
• C o m e t , a c o m p a n y o f S l o v e n i a ;
• Burton E. Waite Jr., Phoenix, Arizona, U.S.A.;
• Prof. Alex Dely, Tucson, Arizona, U.S.A.
2

1. Foreword
A general mathematical model for heat transfer calculations in boilers shall
enable the following:
• Lessen the errors associated with determining the amount of heat transfer in
boilers descending from the use of actual mean flame temperatures;
• Simplify the heat exchanger factor in cases of convection and simplification of
the view factor determination in case of radiation;
• Determine an analytical method for the assessment of the impact of
turbulators on the heat transfer and pressure drop;
• Assess portions of the heat exchanged by convection and radiation in each
boiler section;
• Determine an accurate wall temperature.
Simplifications are allowed for the following cases:
• Heat exchanger factor for convection;
• View factor for radiation;
• Wall temperature for convection on the water side.
2. Definition of the problems
2.1 Heat transfer by thermal radiation in boilers
To determine the radiant and convective part of heat transfer in the fire-tube
boilers the effective temperatures in each boiler section must be known. In case
of convection that temperature is transferred to mean logarithmic temperature
difference. For the radiation from the hot flame and flue-gases to a cooled
enclosure of simple geometry, which is the case with cylindrical shape of the
furnace in fire-tube boilers, the literature quotes an approximate approach to
calculate the total radiant heat flow from hot flame and gas to the boiler walls by
defining an effective flame temperature. That approach is not based on the laws
of physics, it was introduced solely due to its simplicity compared to other
methods for large water-cooled furnaces. However, it tends to be less accurate
since it cannot accurately asses a combination of flame and gas radiation. A
more accurate yet simple equation for mean radiant temperature (MRT) has not
yet been published.
2.2 Turbulators in boiler tubes
Turbulence promoters (hereafter refereed to as turbulators) are inserts which
increase the rate of convection in the tubes compared to those which are empty.
As their name implies, their function is to increase the turbulence of the hot gases
flow by breaking up the laminar boundary layer and thereby increasing
convection. These devices appear in different shapes. Meanwhile, a more
accurate heat transfer calculation in boilers calls for a more accurate analytical
assessment of turbulators’ effect on the total heat transfer. This is of particular
importance in order to appropriately operate the boiler system, including the
burner. An inappropriate assessment of turbulators’ impact on pressure drop can
3

cause the choking of the burner because its fan would be unable to overcome the
increased pressure drop in the boiler due to an inaccurate assessment of the
turbulators’ combined effect on heat exchange and pressure drop.
Coiled-wire turbulators are, today, the most widely used but this area has not yet
been researched enough, and the available theory needs to be extended. The
latest published results of experiments and theory is limited to coiled-wire
turbulators with smaller diameters which is not applicable to boilers where the
tubes are of larger diameter.
2.3 Heat transfer by convection in boilers
Convection in boilers takes place simultaneously with radiation. In tubes of firetube
boilers more than 90% of heat exchange takes place by the convection. In
the furnaces the radiant part is greater than in tubes. Calculation of convection is
conducted by standard equations for flows in straight tubes and channels. This is
also true for the boiler furnaces, whether they are circular or rectangular in crosssection.
This picture totally changes when gas flow directly hits the surface
involved in convection, such as in the case of the rear of the furnace (Figure 1).
The rate of convection is much higher in these cases and cannot be assessed by
classical equations for straight flow in tubes and channels. Tests showed much
lower gas exit temperatures from the furnace than had been calculated which
was found to be attributable to lacking of taking into account the heat transfer
from impinging jet of the flue-gases. This type of the heat transfer (Figure 2) had
been investigated, but its role in boilers has not yet been investigated enough.
Another special case of heat transfer in boilers is the use of a cylinder of high
temperature and corrosion resistant material in reversing type furnaces. This
cylinder improves the long-term boiler performances (less scaling of surfaces by
unburned fuel sulfur, less soot), while also affecting the radiation and convection
(Figure 3). Further, tests show an overall improvement in heat transfer in boiler
with such a cylinder. However, the theoretical background of cylinder’s impact on
local heat exchange has not yet been researched enough.
4

cooled door inside wallarea
of heat transfer
from impinging jet
circular
furnace
flue-gases path
in boiler
tubes
water
furnace rear wall -
area of heat transfer
from impinging jet
Figure 1: Areas of heat transfer from impinging jet in fire-tube boilers
wall
wall-jet region
jet
Z
D
d
nozzle
transition region
region of fully developed jet flow
stagnation point
stagnation or
impingement region
Figure 2: Impinging round jet to flat plate
5

door area
inside cylinder:
- area of no
convection
even when door
is cooled
- area of
increased MRT
door area
outside cylinder:
- area of increased
convection by
increased flue-gas
velocity from ring-slot
(when door is cooled)
- area of reduced MRT
furnace wall
inside cylinder:
- area of no convection
- area of reduced radiation
- area of increased MRT
cylinder of high temp.
resistant material
in furnace
part of furnace between
cylinder's end and rear wall
- area of increased MRT
and increased convection
ring-slot outside
cylinder and
furnace walls:
- area of increased
convection to furnace
- area of radiation to furnace
and cylinder at reduced MRT
furnace rear wall:
- area of increased
convection due to
increased flue-gas
velocity from cylinder
- area of increased MRT
Figure 3: Heat exchange in furnace of boiler with cylinder of high temperature resistant material
6

3. The experiments
The tests were performed on actual residential-sized hot water boilers of 450 kW
and steam boiler 4MW 10 bar by the direct method. These boilers were built in a
two-pass with reversing furnace design (Figure 1, Figure 5).
3.1 The experimental program
The research program consisted of the following:
1. Research of the heat transfer on conventional residential-sized hot water
boilers with these configurations:
• empty furnace and empty tubes;
• cylinder of high temperature resistant material in the furnace;
• coiled-wire turbulators of varying geometry in tubes of varying diameters.
2. Research of the pressure drop on conventional residential-sized hot water
boilers with these configurations:
• empty furnace and empty tubes;
• cylinder of high temperature resistant material in the furnace;
• coiled-wire turbulators of varying geometry in tubes of varying diameters.
3. Research of the heat transfer and pressure drop on conventional residentialsized
hot water boilers of varying internal geometry with these configurations:
• hot water boiler in conventional design;
• above boiler with added cylinder of high temperature resistant material in
the furnace in conventional boiler design;
• above boiler with added cooled door;
• above boiler with added short cooled-cylinder on inside side of cooled
door.
4. Research of heat transfer on industrial size steam boiler with cylinder of high
temperature resistant material in the furnace and cooled door.
7

4. Summary of test results
Mean radiant temperature
The introduced approximation for MRT delivers an improved accuracy in heat
transfer calculations in fire-tube boilers. As is evident from the table below, while
holding all other parameters constant, the introduced equation for MRT delivers a
higher radiation which results in a better agreement between the predicted and
measured values of heat transfer in test fire-tube boilers.
MRT by
existent
equation
°C/K
MRT by
introduced
equation
°C/K
difference in
MRT
%
difference
in calculated
radiation
%
calculated/
measured
furnace exit
temp. with
MRT by
existent
equation
°C
calculated/
measured
furnace exit
temp. with MRT
by introduced
equation
°C
1251/1524 1477/1750 +18/+14.8 +43.7 727/609 630/609
Coiled-wire turbulators in boiler tubes
The analytical assessment of the impact of the coiled-wire turbulators on heat
transfer and pressure drop in straight tubes, enables a higher degree of accuracy
as is evident from tables below.
measured boiler heat
output
kW
calculated boiler heat
output
kW
absolute difference between
measured and calculated heat
transfer in boiler
%
446.7 442.2 -1
measured gas side
pressure drop in boiler
Pa
calculated gas side
pressure drop in
boiler
Pa
absolute difference between
measured and calculated
pressure drop
%
97 88 -9.2
The optimum dimensions of the coiled-wire turbulators to assure the maximum
accuracy in calculations was found as shown in Figure 4. Additionally it was
found that pitch, wire, and turbulator diameter do not significantly affect the
accuracy of heat transfer calculations with the exception of an extremely low
pitch. The turbulator length was found to be of importance and corresponding
coefficients taking the length of turbulator into account were found.
8

D
min. 0.8 D
min. 1/3 L
L
max. 0.5 mm
Figure 4: Optimum dimensions of coiled-wire turbulator to assure +10-15% accuracy in pressure
drop estimation
Role of heat transfer from impinging jet in boilers
Research on test boilers showed the role of heat transfer from impinging jet is
highly significant in the analytical assessment of heat exchange in the furnace.
As is evident from the table below, the heat transfer from impinging jet can
contribute to more than 80% of all convection in the furnace. Temperatures of
surfaces exposed to heat transfer from impinging jet are more than 100% higher
than those of surfaces not involved in heat transfer from impinging jet..
convective
heat transfer
coefficient
W/m 2 K
convective heat
transfer
coefficient
from jet
impingement
W/m 2 K
convection
from jet
impingement
kW
total
convection
kW
convection
from jet
impingement vs.
total convection
%
11.7 165.2 150.4 183 82.1
9

5. Conclusion
In conventional 3- and 4-pass fire-tube boilers, only a smaller portion of total heat
is transferred in the furnace as it has as much as 90% and more heat transfer
surfaces in the tubes. The radiation in the tubes is almost nonexistent compared
to convection, while in the furnace, the radiation can be even smaller than
convection, as the test boilers demonstrated. This is in total contradiction to
water-tube boilers where convection represents less than 20% of total amount of
heat exchange in furnace. As the test boilers demonstrated, the total heat
exchange in the furnace can be as high as 80%. Hot water test boilers showed
additionally that the percentage of surfaces in tubes could be close to that found
in the furnace. The industrial-sized steam test boiler has as low as 2.3 times
more area in the tubes then in the rest of the boiler. Furthermore it was proved
that the convection in the furnace of fire-tube boilers can be made even higher
than the radiation.
The number of boiler tubes is limited by burner fan capability to overcome
internal pressure loss. By that fact, the general direction in designing fire-tube
boilers is given; namely to install only as many tubes as are necessary. This
requires the exact analytical assessment of heat transfer in particular boiler
sections to which this dissertation was devoted. Thus, by proper design of the
boiler (for which the in-deep knowledge of heat transfer is of primary importance),
as demonstrated in this dissertation, a sizable intensification of heat transfer and
noticeable savings in boiler manufacturing cost can be achieved.
Three units of the test boilers, designed under the consideration of the new
findings described in this dissertation (steam boiler of 4MW 10bar) were built and
delivered to actual customer in USA in 2001 (www.wargaboiler.com). Compared
with conventional design the noticeable improvements are evident as f.i. the
number of boiler tubes which is reduced for more than 70%.
Figure 5: Steam test boilers in actual operation
10

1. APPARATUS AND EXPERIMENTAL UNCERTAINTY USED FOR TESTING
OF TEST BOILERS
The assessment of accuracy of the testing line for boiler thermal output and
efficiency was performed according to Policy on Reporting Uncertainties in
Experimental Measurements and Results.
1.1 The apparatus used for testing of hot water test boilers
The hot water boiler testing line was built according to German Standard DIN
4702, Teil 2. The tests on the boilers were run under the same conditions (steady
state, 30-minute minimum test duration etc.). The average difference between
boiler thermal output calculated from repeated tests under the same conditions
was less than ±0.5%.
Data acquisition:
Hewlett Packard® 9000/226 digital data acquisition system:
• digital scanner HP 3497 A
• digital voltmeter HP 3456 A
• line printer HP 2631 A
• computer 9000/300
Temperatures:
• thermocouples NiCr - Ni B 1475, PT 100 by Degussa®, Italy
Flow rates:
• natural gas flow: gas flow meter G65 DN80 by Rombach®m, Germany
• water rates: analog scale of 0-1000 kg range, KG II by Libela-Celje®, Slovenia
• oil rates: analog scale of 0-50 kg range, KG II by Libela-Celje®, Slovenia
Time intervals:
• stop-watch of 0-60s and 0-30 min. range by Zaquet®, Germany
Flue-gases analysis:
• electrochemical analyzer Combilyzer 2000 by Afriso®, Germany
Boiler internal flue-gases side pressure drop:
• U-tube differential pressure gauge with liquid
Flue draft:
• tube micro manometer with liquid
Fuel heating value:
• oil: calorimeter C-4000 by Ika-Werk®, Germany
• natural gas: calculated according to chemical analysis (97.3% Methane CH4)
1

1.2 Assessment of measurement uncertainty for hot water test boiler
• temperatures in 22.1-81.0°C range (calibration results):
at t1=22.1°C: -0.32°C (-1.44%)
at t2=81.0°C:-0.47°C (-0.58%)
• water rate (calibration results): -0.5%
• liquid fuel rate (calibration results): -0.025%
• gaseous fuel rate (calibration results):
min. rate: -0.29%
max. rate: -0.39%
average: -0.34%
• liquid fuel heating value: -1.5% (calibration results)
• natural gas (>97% CH4) heating value: calculated as 100% methane
• CO - emission: ±10% of displayed value
• rest oxygen: ±8% of displayed value
• NOx - emission: ±10% of displayed value
• differential pressure and boiler draft: ±0.5 mmH2O (±5 Pa)
Maximum expected uncertainty in calculating the boiler thermal input:
Qfuel = B ⋅ Hi
[W]
• PQ/Q=0.005 (from average tests results)
• BB/B=0.0034 (from calibration)
• BH/H=0.015 (from calibration)
U
Q
Q
fuel
=
⎛
⎜
⎝ Q
2
2
B ⎞ Q ⎛ BB
⎞ ⎛ BH
fuel
⎛ P
⎜
⎝ Q
Q
fuel
⎟
⎠
⎞
⎟
⎠
2
=
=
⎜
⎝
B
⎛ B
+ ⎜
⎝ Q
⎟
⎠
0.
0034
Q
fuel
+ ⎜
⎝ H
⎞
⎟
⎠
2
2
2
⎞
⎟
⎠
+ 0.
015
=
2
0.
005
2
=
0.
0153
+ 0.
0153
=
2
1.
53%
=
0.
016
= 1.
6%
Maximum expected uncertainty in calculating the boiler thermal output:
Q
.
= mH
O ⋅ c ⋅ ( t − t ) [W]
boiler
2
p
H 2O,
2 H 2O,
1
• ∆t=65±5°C (average tests results)
• PQ/Q=0.005 (from average tests results)
• B’t1=

⎛ B
⎜
⎝ Q
Q
boiler
2 2 2
⎞ Bm
Bc
Bt
Bt
Bt
Bt
⎟ =
⎠ m c t t t t
⎛ ⎞
⎜ ⎟ +
⎝ ⎠
⎛ ⎞
⎜ ⎟ +
⎝ ⎠
⎛ ⎞
⎜ ⎟ +
⎝ ⎠
⎛ ⎞
⎜ ⎟ −
⎝ ⎠
⎛ ⎞
⎜ ⎟ ⋅
⎝ ⎠
⎛
' '
⎞
1
2
1 2
2 ⎜ ⎟
∆ ∆ ∆ ⎝ ∆ ⎠
Q
U
( ) ( ) ( ) ( )
= 0. 005 + 0. 005 + 0. 32 / 65 + 0. 47 / 65 − 2 ⋅ 0. 05 / 65 ⋅ 0. 05 / 65
= 0. 0112 = 112% .
Q
boiler
=
⎛ P
⎜
⎝ Q
2
2 2 2 2
Q
boiler
⎞
⎟
⎠
2
⎛ B
+ ⎜
⎝ Q
Q
boiler
⎞
⎟
⎠
2
=
2
0.
005
2
+ 0.
0112
2
=
0.
0122
Maximum expected uncertainty in calculating the boiler efficiency:
Qboiler
ηboiler = ⋅100 [%]
Q
Uη
=
η
⎛ U
⎜
⎝ Q
Q
boiler
⎞
⎟
⎠
2
⎛ U
+ ⎜
⎝ Q
Q
fuel
⎞
⎟
⎠
2
=
fuel
0.
0122
2
+ 0.
016
1.3 The apparatus used for testing of steam test boiler
2
=
0.
0201
=
= 1.
22
2.
01%
The tests on the steam boiler were run in an actual boiler room for steam heating
(see Figure 5) in steady state, with a minimum duration of 30 minutes. The
average difference between the boiler thermal output calculated from the tests
was less than ±0.5%.
Omega® DAQ-12 data acquisition system:
• DAQP-12A terminal strip
• PCMCIA card UIO-37 with 12 bit 8/16 channel analog input
• CP-DAQP cable
• OMX-R250, 250 ohm precision resistors
• Portable computer Toshiba® Satellite 310 series
Steam & gas flow rates and temperatures of media (steam, gas, feed water):
• Rosemount & Fisher® 3095 M multivariable mass flow orifice type transmitter
with remote feed water temperature pt 100 sensor
Flue-gases analysis:
• EGA Sampling system by Autoflame® (came as part of burner equipment)
Boiler internal flue-gases side pressure drop:
• U-tube differential pressure gauge with liquid
Flue draft:
• tube micro manometer with liquid
Fuel heating value:
• natural gas: calculated according to chemical analysis (97.3% Methane CH4)
3

1.4 Assessment of measurement uncertainty for steam test boiler
• temperatures (calibration results): at t=81.0°C: -0.47°C (-0.58%)
• steam and gas rate (calibration results): +/-0.8% (average)
• natural gas (>97% CH4) heating value: calculated as 100% methane
• differential pressure and boiler draft: ±0.5 mmH2O (±5 Pa)
Maximum expected uncertainty in calculating the boiler thermal input:
Qfuel = B ⋅ Hi
[W]
• PQ/Q=0.005 (from average tests results)
• BB/B=0.01 (from calibration)
• BH/H=0.01 (from calibration)
U
Q
Q
fuel
=
⎛ P
⎜
⎝ Q
⎛
⎜
⎝ Q
Q
fuel
2
2
B ⎞ Q ⎛ BB
⎞ ⎛ BH
fuel
2
⎟ ⎞
⎠
⎟
⎠
=
=
⎜
⎝
⎛ B
+ ⎜
⎝ Q
Q
fuel
B
⎟
⎠
+ ⎜
⎝ H
2 2
0.
01 + 0.
01 =
⎞
⎟
⎠
2
=
2
⎞
⎟
⎠
0.
005
2
0.
014
+ 0.
014
= 1.
4%
2
=
0.
0148
= 1.
48%
The maximum expected uncertainty in calculating the boiler thermal output
according to the manufacturer of the steam flow measurement equipment:
Q = V steam ⋅ ( h − h )
.
[W]
Q
U
Q
boiler
=
⎛ P
⎜
⎝ Q
Q
boiler
⎞
⎟
⎠
2
boiler
⎛ B
+ ⎜
⎝ Q
⎛ B
⎜
⎝ Q
Q
boiler
Q
boiler
⎞
⎟
⎠
2
=
⎞
⎟
⎠
steam
2
= 1%
0.
005
2
fw
+ 0.
01
2
=
0.
0111
=
1.
11%
Maximum expected uncertainty in calculating the steam boiler efficiency:
Qboiler
ηboiler = ⋅100 [%]
Q
Uη
=
η
⎛ U
⎜
⎝ Q
Q
boiler
⎞
⎟
⎠
2
⎛ U
+ ⎜
⎝ Q
Q
fuel
2
⎟ ⎞
⎠
=
fuel
2
2
0.
0111 + 0.
0148 =
0.
0185
= 1.
85%
4

2. THE EXPERIMENTS
Introduction
The tests were performed on actual residential-sized hot water boilers based on
standardized boiler testing line and arrangements according to German standard
DIN 4702 Part 2 by the direct method. The final tests were performed on
industrial sized steam boiler in boiler room, with the results being evaluated per
above standard.
2.1 Description of hot water boiler testing line
Figure 1 shows a testing line for measuring the hot water boiler output and
efficiency by direct method. Figure 2 depicts the testing line for measuring the
boiler’s internal pressure drop on the flue-gases side. Figure 3 shows the
arrangement for fuel flow measuring. Section 1 lists the apparatus used and
provides an estimation of measurement uncertainty.
outflow
boiler water
barrel
tv
constant
level
barrel deaeration
scale
boiler
RV2
constant level
supply water barrel
te
P
RV1
tr
constant
water
pressure
valve
RV
fresh
water
supply
Figure 1: Testing line for heat output and efficiency of hot water boilers
All pipes containing the boiler water are thermally insulated and their heat losses
are known. The pump (P) pushes the water through the boiler. The fresh water
first reaches the water barrel and maintains a constant level, which is located
usually at the attic. The desired fresh water flow into barrel is adjusted by valve
(RV) and then maintained by valve (RV1). Before the fresh water with
temperature, tE, enters the boiler, it is premixed with boiler water of temperature,
tV, by valve (RV2), which adjusts the premixing rate and temperature difference
5

(tV-tR). The premixed water then enters the boiler with the temperature, tR. The
same quantity of water that enters the boiler also leaves it. Boiler water exiting the
boiler and fuel flow are weighted separately by scales.
The water temperatures are measured by thermocouples, as is the ambient
temperature. The flue-gases temperatures in the test boiler were measured by
protected thermocouples at following points:
• furnace exit in leftmost, rightmost, uppermost and lowermost tube;
• boiler exit.
As shown on Figure 2, the testing line arrangement for measuring the internal
flue-gases pressure drop in the boiler is simple and consists of only one line. One
end penetrates the flue-gases exit from boiler, while the other end is in the
furnace. The color-liquid filled gauge on the top (where all line-ends meet) shows
the differential pressure, ∆p, in mm of water column, thereby representing the
internal pressure drop in boiler.
from furnace
boiler
flue
Figure 2: Testing line for pressure drop in boiler
liquid fuel
barrel
scale
h
boiler
Figure 3: Liquid fuel flow measuring arrangement
6

2.2 Description of steam boiler testing line
The tests on the steam boilers were performed during the actual operation as
depicted in Figure 5 by the direct method. Fresh water comes to chemical water
treatment (4) from where it goes to the boiler feed water tank (3) with deaerator
where it gets pre-warmed and deoxidized afterwards it enters boiler (1) via the
feed water pump. The additional feed water tank (5) serves the purpose of storing
the extra feed water to cover the demand at peak loads.
The fuel, natural gas, is delivered to burner (2) through gas train (8) according to
ASME Boiler and pressure vessel code CSC-1. The valves are controlled by
Honeywell®, model 7800, electronics located in the burner command box. Boiler
(1) is equipped with two sets of water gages, McDonnell&Miller® and
Honeywell®. The boiler feed pump is controlled by a McDonnell&Miller® water
level gage. In the boiler flue, the Exhaust Gas Analysis (EGA) electrode for O2,
CO2, CO, NOx and temperature probes are located transmitting signals to an
Autoflame® unit located in the burner control box, displaying and calculating the
combustion efficiency. On the basis of a comparison between the measured and
pre-set values, this unit controls the servomotors for air and fuel delivery to the
burner.
The generated steam in the boiler comes to the steam distributing and steam
pressure reducing valve (7) directing a necessary part of it to steam heating
system, the rest to the atmosphere. The condensate from the heating system
then goes to condensate tank (6). The amount of steam generated and the
incoming gas are measured by flow multivariable orifice type transmitters placed
in steam line (10) and gas train, respectively. The feed water temperature is
measured by a thermocouple located in the feed water-delivering pipe. The test
data are acquired via portable computer and data acquisition system (see Figure
4), converting 4-20mA current coming from flow transmitters to 1-5V of electric
force via resistors (see section 1) for list of apparatus and estimation of
measurement uncertainty). The engineering values are then calculated from
voltage, written to the file and used to calculate boiler input, output, and
efficiency.
Figure 4: Data acquisition system
7

5
water level gage
feed water
6
condensate
M M
4
water
level
gage
pressure gage
pressure gage
feed water
pressure
gage
Y
thermometer
Y
Y
Y
deaerator
PC
LEGEND
PC
pressure control
3
pressure gage
30 X D
9
to computer
p, p, t
p
temp. control
TC
7 X D
p t
D
p
Z
Z
AUTOFLAME
8
feed water
2
switch for
p pressure
gage
natural gas
line
EGA - exhaust gas analysis
1 steam boiler 4MW 10bar, 6t/h steam
2 dual fue burner Weishaupt RGL 60 2-A with Autoflame combustion control for O2, CO2, CO
3 feed water tank with deaerator 6t/h
4 chemical feed water treatment 3 t/h
5 feed water tank 20 m3
6 condensate tank 2 m3 with pump set 6 t/h
7 steam distributing and steam pressure reducing valve from p=8bar to p=2bar
8 natural gas train DN 50 with pilot gas train DN 25
9 gas flow measuring device ROSEMOUNT&FISHER 3095
10 steam flow measuring device RESEMOUNT&FISHER 3095
safety
valve
line
steam line
water
level pressure
gage
controls set
Figure 5: Steam boiler testing line
M
30 X D
venting
valve
line
water level in boiler
1
7 X D
EGA
O2, CO2
CO, t,
comb.eff.
to computer
p, p, t
p
p t
10
D
deposits from blow down
blow down deposits barrel
pressure gage
steam line
pressure control
7
Z
Z
8

2.3 Equations for evaluation of test results
The boiler net thermal output and efficiency were calculated according to DIN
4702 Part 2 by following equations:
• Hot water boiler net thermal output:
( t − t )
.
boiler = mH
O ⋅ cp
⋅
2 , H O H 2O,
2 H 2O,
1
Q + Q − Q
2
Equation 1
testline
Qboiler – boiler power [W]
mH2O – water flow through boiler [kg/s]
cp,H2O – specific heat of boiler water [kJ/kgK]
tH2O,1 – temp. of water entering boiler [°C]
tH2O,2 – temp. of water leaving boiler [°C]
Qtestline – heat losses of boiler testing line [W]
QE – auxiliary power consumed during tests (water pump, burner...) [W]
• Steam boiler net thermal output:
boiler
.
steam
( hsteam
− hfw
) Qtestline
- QE
Q = V ⋅
+
Equation 2
Vsteam – steam flow from boiler [kg/s]
hsteam – enthalpy of steam leaving boiler [kJ/kg/K]
hfw – enthalpy of feed water [kJ/kg/K]
• Heat released from fuel:
Q = B ⋅ H [W]
fuel
Equation 3
i
E
[W]
[W]
Qfuel – released heat from combustion of fuel [W]
B – fuel rate [kg/s, m 3 /s]
Hi – lower fuel heating value [kJ/kg/ kJ/m 3 ] (or higher as used US)
• Boiler efficiency:
η
Q
=
boiler
boiler
Q fuel
Equation 4
⋅100
[ % ]

THE EXPERIMENTAL PROGRAM
The research program consisted of the following:
a) Research of the heat transfer on conventional residential-sized hot water
boilers with these configurations:
• Empty furnace and empty tubes;
• Cylinder/hot tube of high temperature resistant material in the furnace;
• Coiled-wire turbulators of varying geometries in tubes of varying
diameters (30 tubes of 36.5mm and 26 tubes of 51.2mm).
b) Research of the pressure drop on conventional residential-sized hot water
boilers with these configurations:
• Empty furnace and empty tubes;
• Hot tube of high temperature resistant material in the furnace;
• Coiled-wire turbulators of varying geometries in tubes of varying
diameters.
c) Research of the heat transfer and pressure drop on conventional
residential-sized hot water boilers of varying internal geometries with these
configurations:
• Hot water boiler in conventional design (see Figure 4);
• Above boiler with added hot tube of high temperature resistant material
in the furnace in conventional boiler design (see Figure 5);
• Above boiler with added cooled door (see Figure 6);
• Above boiler with added short cooled cylinder on inside side of cooled
door (see Figure 7).
d) Research of heat transfer on industrial size steam boiler with hot tube of
high temperature resistant material in the furnace and cooled door (see
Figure 9).
1. Choosing the test boiler design
The research was focused on fire-tube boilers as these represent the most
commonly-used type of boiler today, which are up to 25MW power and 25bar
pressure for hot water and steam applications. There are several families of
them, including:
• two-pass design (see Figure 2);
• three-pass design (see Figure 1);
• four-pass design (same as the three-pass design, but with one additional
tube assembly).
Basically, all these families are pretty much the same. The primary difference
is in how many times the flue-gases change direction inside the boiler and
how many tube assemblies there are.
1

In the case of the two-pass design, there is a single tube assembly, so the
flue-gases must usually make a 180° turn at the rear of furnace back towards
the door before being able to enter the tubes 1 . Boilers of a three-pass design
(see Figure 1) feature two assemblies of tubes, so there is no turn at the
furnace rear back to the door as the flue-gases go directly into first tube
assembly and from there, turn 180° as they enter the second tube assembly.
The four-pass design is the same as the three-pass only with one more tube
assembly resulting in more tubes and direction changes.
Interestingly, the number of passes does not affect the extraction of heat from
the flue-gases on their way through the interior of the boiler. Rather, the
amount of extracted heat depends solely on how much cooled heat transfer
surface area there is, how it is distributed (f.i. how much area is located in
furnace and in tubes) and on the velocity of flue-gases. More passes do not
mean more efficiency; it only results in more reversals of the flue-gases flow.
Thus, from the heat transfer point of view, there is no advantage of the three-
and four-pass designs over the two-pass one. Further, more passes means
higher cost and weight of the boilers because of the introduction of more tube
assemblies, which are up to three times more expensive than boiler steel
plates.
tubes
furnacefirst
pass
direct tubessecond
pass
front
FRONT VIEW
LONGITUDINAL CROSS-SECTION
return tubes-
third pass
REAR VIEW
Figure 1: Fire-tube steam boiler in conventional three-pass design
rear
steam space
water line
direct tubes-
second pass
1 This is so called reversing furnace type. There are also two-pass boilers where flue-gases take the
same path as in three-pass boilers. These boilers require the flue at the front of boiler, above the burner.
2

Three- and four-pass fire-tube boiler designs have only one advantage over
the two-pass one, they eliminate the need for a 180° flue-gases reversal at
the rear of furnace. As a result, the dwell-time of the combustion products
and, hence, the reactants involved in NOx formation, is shorter resulting in
approximately up to 20% lower NOx - emissions compared to the two-pass
reversing design. As some reductions in NOx - emission can be achieved by
relatively simple and inexpensive means (see Figure 7), along with the new
generation of burners, lower NOx - emissions regardless the type of boiler are
possible. Hence, there is nothing advantageous in the three- and four-pass
boiler designs over two-pass 2 one.
Thus, a two-pass boiler design with reversing furnace was chosen for test
boilers (see Figure 2) and several actual-sized two-pass fire-tube, 450 kW, hot
water boilers were built and a series of tests was run on them. For the
purpose of more extensive validation of the algorithm, a 4MW 10bar industrialsized
steam boiler was also built and put to the tests. Since contemporary
commercial fan burners were applied, the 100% combustion efficiency was
achieved eliminating an additional error source due to unburned fuel as in the
case of solid fuels. The test boilers were completely thermally insulated,
including their doors, in order to minimize the jacket losses, which were
measured separately and taken into account in the calculations.
burner
opening
front side of
furnace
(area: 0.227m 2 )
646mm
880mm
tube assembly/
second pass
(area: 6.124m 2
)
1429mm
26 tubes of 51.4mm diameter
and 2.6 mm wall thickness
1400mm
exit from furnace -
flue-gases temp. measuring points no.1-4
(protected thermocouples)
flue-gases flow direction
furnace/first pass
(cylindrical area: 3.936m 2
)
exit from boiler -
flue-gases temp.
measuring point
no.5 (protected
thermocouple)
furnace
rear wall
(area: 0.327m 2 )
smoke chamber
at exit
Figure 2: Dimensions of fire-tube hot water test boiler in a two-pass design with a reversing
furnace
2
The two-pass boiler with reversing furnace turns into semi three-pass one by adding the hot tube in the
furnace according to Figure 5.
3

A different number and size of tubes were used on the test boilers. The test
boilers featured varying internal. Emissions were measured and evaluated
using a 3% rest oxygen in the flue-gases as a reference point in order to
assure comparability of results.
measuring point no.3
for exit temp. from
furnace
boiler shell
measuring point no.1
for exit temp. from
furnace
tubes
measuring point no.4
for exit temp. from
furnace
furnace
measuring point no.2
for exit temp. from
furnace
Figure 3: Flue-gases exit temperature-measuring points in furnace of hot water test boiler
4

2. Test boiler No. 1: 450 kW hot water boiler in conventional 2-pass
design
The boiler shown in Figure 4 has a cylindrical reversing furnace, and one set
of straight tubes. Here, the flue-gases travel all the way to rear of the furnace
where they make a 180° turn. From there, the flue-gases return to the door in
front where they make another 180° turn into the tubes. From these tubes, the
flue-gases assemble in the exit smoke chamber from where they enter the
stack. It should be noted that in all cases, the test boilers were entirely
thermally insulated including the door to minimize jacket-losses, which were
also measured and taken into account in the calculations.
uncooled door
coiled-wire turbulators
Figure 4: Test boiler No. 1
reversing furnace
(first pass)
thermal insulation
rear smoke chamber
stack connection
flue-gases path in boiler
tube assembly
(second pass)
5

3. Test boiler No. 2: Hot tube in furnace
As shown in figure the first test boiler’s furnace was added a hot tube of 3 mm
thickness made of fire-resistant steel.
ring-slot around hot tube
Figure 5: Test boiler No. 2
hot tube
6

4. Test boiler No. 3: 450 kW hot water boiler with cooled door
Added to second test boiler configuration was a water-cooled door (see Figure
6). The cooled door 3 presents additional cooled surfaces in the furnace, which
retains its same dimensions. In this case, the cooled door area represents
more than 10% of the furnace area. The following effects were expected:
• An intensified heat exchange in the furnace due to more cooled surfaces;
• An improved efficiency attributable to the intensified heat exchange;
• A lowered surface temperature resulting in reduced jacket losses.
Figure 6: Test boiler No. 3
The cooled-door design contains two tubes (9, 10) made of rubber 4 capable to
resist the temperature of boiler water (since it is tube-shaped it is also is
capable of resisting he pressures higher than what the boiler itself is designed
to withstand). The door walls are plain and strengthened by anchors (4). The
boiler water (1) comes in the door (5) passing an elbow (17), placed into the
3
A practical advantage of the cooled door is an absence of the refractory which lowers the maintenance
cost of a boiler.
4
If regulations do not permit the rubber a bent steel tube can be used instead.
7

eturn tube (16) causing an injection effect, tube (15), another elbow (14),
flexible tube (9) and third elbow (7). There is also a built-in vertical plate
separating the door halves herewith enabling directed water flow within the
door. Water (1) leaves the door on the top through elbow (11) and flexible
tube (10) into advance flow tube (12) having inside of it an elbow (13) of same
direction as that in the return tube causing a suction effect of water from the
door.
The flexible tube connections to the door and water jacket were realized by
flanges (8). The waterside resistance is not noticeably increased despite of
the additional elements. Thus, there are no additional requirements on the
pump. Tests showed the water temperatures in the door and rest of the boiler
to be almost equal.
5. Test boiler No. 4: Increase of cooled door area
Figure 7 shows a boiler door design, which further intensifies the heat
exchange in the furnace by further increasing the cooled door area. There is a
short cylinder (6) on the inner side (3) of the boiler door (2), which is cooled
(4) by the water circulating in the door (2). The cylinder (6) is horizontally
taperless and partly embraces the flame (8). The thickness of water-flow
cross-section (4) in the cylinder (6) was 30 mm (to prevent water circulation
problems and local evaporation). On the test boiler, the outer diameter of the
cylinder (6) was 490 mm, the inside diameter was 406 mm, and its length was
250 mm.
The following additional effects were expected:
Due to cooling (4) of cylinder (6) and, therefore, the vicinity of additional
cooled surfaces to the flame (8), a more intense cooling of the initial part of
the flame (8) is enabled causing a decrease in NOx - emissions. The flame
temperature is also partly reduced due to a cooling effect caused by the fluegases
expansion while exiting the cylinder (6) because its inside (5) diameter
is smaller than that of the furnace (9). There is further a local recirculation of
the (partly cooled) flue-gases (7) back to the flame. The following process
accomplishes this:
1. The inside diameter (5) of cylinder (6) is smaller than that of the furnace;
2. The same quantity of flue-gases pass in the same time period as first short
cross-section (5) of the cylinder (6) and then the larger section of the
furnace;
3. Due to the diameter differential, the flue-gases velocity at the exit of the
cylinder (6) accordingly slows down;
4. Slowing down the flue-gases flow generates a local pressure increase at
the cylinder (6) exit;
5. Because there is higher local pressure at the exit of the cylinder (6) than at
its interior (5), part of the already partly cooled flue-gases is sucked back
into the flame (8), further lowering its temperature and, thus, NOx
production.
8

The cylinder’s 5 (6) inner and outer diameters were adapted to match the
sprinkling angle of the burner nozzle to avoid a contact of flame with
surrounding walls of cylinder (6). This would cause an under-cooling of the
flame and, thereby, production of soot and increased CO emissions.
Figure 7: Test boiler No. 4
5 With a new generation of low NOx burners this cylinder is no longer necessary as the effect, it causes,
is generated by burner’s head design. This boiler served only as test boiler.
9

6. Test boiler No. 5: Industrial-sized 4MW 10 bar steam boiler
Since, the hot water test boilers were relatively small, the test results obtained
on them cannot be treated as generally applicable for all sizes of boiler. Thus,
for the purpose of extended checking of the thesis explained before, an
industrial-sized steam boiler of 4MW, 10 bar was designed. The boiler was
built according to the ASME Boiler and Pressure Vessel Code, Section I
(1998).
The steam test boiler depicted in Figure 8 and Figure 9 was built the same
way as its hot water counterparts (2-pass design with reversing furnace) and
was also fully functional 6 . Added features used on test boilers No. 2 and No. 3
are the cooled door, the hot tube of high temperature resistant material (see
Figure 11) in the furnace, which is corrugated, and full-tube-length coiled-wire
turbulators. There are 75 tubes of 62.6 mm in with 4.213 m in length as shown
in Figure 16. The furnace is 3.946 m long and 1.52 m in diameter (1.35 m
when corrugated).
During the tests, a fire-resistant steel was used as hot tube material in test
boiler No. 5 (see Figure 11) as well in other test boilers. In actual application
the steel was found not applicable in boilers beyond 500 kW because the hot
tube starts sagging over time under its own weight when hot. This is avoided
by use of ceramic and the hot tube is then made in segments (see Figure 9
and Figure 12). With a reference to steel, the ceramic used has a lower
thermal conductivity but a higher surface emissivity, which does not change
with the temperature as this is the case with steel (it is additionally several
times lighter and cheaper then fire-resistant steel). It was observed that with a
ceramic hot tube the heat exchange in the furnace was nearly the same,
namely not affected, compared to the hot tube made of steel.
Thus, the test boiler No.5 in actual application (Figure 9 and Figure 10) has
the hot tube in the furnace made of ceramic (see Figure 12). Such a hot tube
in this boiler is made-up of 7 segments, each 530 mm long and 35 mm thick.
Each segments consists of 4 parts, 2 upper and 2 lower parts with legs
attached by a side and at the bottom of the lower 2 segments. In boiler the
segments are lined-up with each other so they measure 3.71 m in length in
total.
Boiler (1) is placed on legs (4). Boiler water below water line (23) is in the
boiler (1) and the door (13). The steam space in the boiler and door is located
above water line (23). On their outsides, door (13) and boiler (1) are clad with
a thermally insulated jacket (25). The flame is observed through aperture (15)
located in the door (13). The burner is attached to the door (13) on mounting
plate (14). Water level gauges (16) are located on boiler (1) and door (13).
The flue-gases exit through outlet (2) at the rear of the boiler (1). On the inside
of the door (13), the flue-gases are diverted into a tube assembly where a
6 Three of them were built, sold and installed at Tucson Medical Center Healthcare in Tucson, Arizona,
U.S.A. in spring 1999. Their performances are monitored on an ongoing basis. See
www.wargaboiler.com.
10

Figure 8: Test boiler No. 5
partition (24) made of fire-resistant steel prevented the hot flue-gases to enter
the boiler (1) and door (13) above the water line (23).
A portion of the boiler water travels from boiler (1) bottom via pipes (6), (7),
(9), (11), and (12) into the door (13) via the connected vessels principle. The
water in door (13) turns into steam above water level line (23), exits the door
(13) through pipes (18) and (20), enters the boiler (1) steam area above water
line (23) and leaves the boiler (1) via the steam outlet (22). The part (21)
11

elow pipe (20), which delivers the steam from the door (13) into the boiler (1)
prevents the steam coming from the door (13) from getting in touch with the
boiler water, thereby preventing it from getting wet.
Pipes (7), (11), (18), and (20) are flanged (5), (10), and (19) for the purpose of
making it possible to open the door (13), and for the removal of deposits from
the boiler (1) bottom. Connections (3) and (8) are provided on the underside
of the boiler (1) for draining the boiler water, and for the removal of deposits
from boiler (1) bottom. The solids are in an ongoing process removed from
boiler and door (13) from water surface (23) via pipe (17).
In the case of hot water boilers of such design, a pump (26) to pipe (20) is
added thereby assuring sufficient water flow in and out of the door (13). Also,
a valve (27) makes it possible to open the boiler door (13) without emptying
entire boiler (1).
pressure controls
connection
blockade to prevent
hot gases contact
uncooled area
water column
connection on door
(another one side-
mounted on boiler)
water space
in door
diagonal
stay-bars
steam space
in door
front flame observing
aperture
steam connection
of door with boiler
door steam flow
dispeller
safety valve
connection
burner
main steam valve connection
removal of water
droplets in steam
man hole
(another two at each
boiler side)
water connection of
boiler with door
primary low water cut-off connection
secondary low water cut-off connection
in combination with pump controller
steam space in boiler
boiler leg
thermal insulation
water space in boiler
Figure 9: Test boiler No. 5 as in actual operation
lifting loop
removable cylinder made of
ceramic segments
stainless steel jacket
feedwater inlet connections
completely submerged
corrugated furnace
bottom blow-down connection/
boiler filling/draining
venting valve connection
flue-gases exit box
with stack connection
condensate draining
during cold start
flue-gases path in boiler
rear flame observing
aperture
rear manhole
12
water line
coiled-wire turbulence
promoters made of
fire-resistant steel
smoke tubes

Figure 10: Test boiler No. 5 as in actual operation with door open
Figure 11: Hot tube of fire-resistant steel in furnace of test boiler No.5 as during tests
13

Figure 12: Ceramic hot tube segment in furnace of test boiler No. 5 as in actual operation
7. Choosing the turbulators’ geometry
The coiled-wire turbulator geometry was chosen according to following criteria
and findings:
• The turbulators shall tightly fit into each tube. Hence, the basis turbulator
diameter was chosen to be at most 1 mm less than the internal tube
diameter;
• The pitch must be of sufficient length so the basis pitch was chosen to be
approximately the same as the tube net diameter;
• Each turbulator must be of sufficient length so the minimum length was
chosen to be one-third tube length, whereas experiments were also
conducted with turbulators of two-thirds and full tube lengths.
A series of tests was run on a test boilers with coiled-wire turbulators of
various diameters, lengths, wire diameters, and pitches (see Figure 13, Figure
14, Figure 15 and Figure 16). The purpose of these tests was to find out how
the heat transfer and pressure drop are affected by these varying parameters,
14

and also which pitch/wire-diameter/diameter/length combination is in best
agreement with the analytical predictions of the heat transfer and pressure
drop in the tubes.
36.5
36
500
40
1429
Figure 13: The dimensions of tubes and turbulators used in 30-tube hot water test boiler
51
48
500
50
1360
Figure 14: The dimensions of tubes and turbulators used in 26-tube hot water test boiler
d
h
l
δ
L
l=1/3 L, 2/3 L, L D=51 mm
d=45, 48, 50 mm L=1470 mm
δ = 6, 8 mm
h=10, 25, 35, 53, 65 mm
Figure 15: Additional palette of coiled-wire turbulators used in 26-tube hot water test boiler
10
5
D
15

h
d D l=L
l
L
δ
D=62.6 mm
L=4213 mm
h=53 mm
d=62 mm
δ = 5 mm
Figure 16: Dimensions of tubes and turbulators in 75-tube 4MW steam test boiler
8. Heat transfer tests
The data collection began after each test boiler reached the steady state
condition. Average values, boiler outputs, and efficiencies were calculated by
equations given earlier in this section. The tests were run at differing boiler
loads with fuel oil and natural gas. The test results were compared to
calculated.
Experimental results and discussion on Mean Radiant Temperature, MRT
The tables and graphs to follow show the representative results of comparison
between the measured and calculated heat transfer in test boilers, at different
boiler loads, using introduced, new and existing, old equations for MRT. It is
quite evident that, for all parameters being held constant, the heat transfer
results are in better agreement with the measured values when MRT is
calculated by the new equation.
Compared to the case when MRT is calculated by the new equation, the
differences between the measured and calculated furnace exit temperatures
are reduced to a minimum because the introduced equation delivers a higher
MRT (17.4% in average in °C). It is also evident that differences between the
calculated and measured flue-gases exit temperatures from the furnace are
higher at lower boiler loads, which means the lower the boiler load the greater
the error when using the old equation for MRT.
General observations
With the values obtained by the new equation for MRT the conditions in the
furnace, when calculated by old equation are noticeably closer to the
measured values (see Table 1, Graph 1, Graph 2, and Graph 3):
• The flue-gases exit temperature from the furnace, when calculated by the
old equation for MRT is up to 41.5% higher from that obtained by
introduced equation;
• The calculated total heat exchange in the furnace, when calculated by the
old equation for MRT is up to 13.8% less;
• The radiant heat exchange in the furnace, when calculated by the old
equation for MRT is up to 43.7% less.
16

oiler
load
%
Graph 1: MRTs in furnace obtained by different equations
Table 1: Measured and calculated conditions in furnace of test boiler by different equations for
MRT
measured
furnace exit
temp.
°C
MRT
[K]
1
calculated
furnace exit
temp. with
MRT by old
equation for
MRT
°C
2
calculated
furnace exit
temp. with
MRT by new
equation for
MRT
°C
relative
difference
(1 vs. 2)
%
relative
difference
(3 vs. 4)
%
measured
energy output/
energy input of
boiler
kW
adiabatic
combustion
temp.
°C
117 609 727 629 +15.5 +6.7 526.3/576.4 2047
100 567 713 564 +26.4 +10.2 453/493.4 2019
55 447 647 457 +41.5 +13.8 248.5/269.2 1887
Note: Lower fuel heating value (42700 kJ/kg) was used to calculate the energy input (higher
fuel heating value was 45294 kJ/kg)
furnace
exit
temp.
[°C]
1800
1750
1700
1650
1600
1550
1500
1450
1400
117 100 55
boiler load %
MRT by old equation MRT by new equation
850
700
550
400
117 100 55
boiler load %
furnace exit temp. with MRT by old equation
furnace exit temp. with MRT by new equation
measured furnace exit temp.
Graph 2: Exit temperatures from furnace with MRTs obtained by different equations
17

Conclusions
radiative
heat
exchange
in
furnace
[kW]
220
200
180
160
140
120
100
117 100 55
boiler load %
with MRT by old equation with MRT by new equation
Graph 3: Radiation in furnace with MRTs obtained by different equations
The analytical and experimental work described above permits the following
conclusions on presented procedure for calculation of MRT:
• Application of a principle of radiation as a fourth power of absolute
temperature from Stefan-Boltzman’s law in derivation of the new equation
for MRT gives better results than old equation, which is especially evident
when there is higher difference between the two temperatures used to
calculate MRT from, as found in furnaces of fire-tube boilers;
• Using the new equation for the MRT, the heat exchange in the furnace and
other sections of fire-tube boilers can be assessed with an improved
accuracy. When MRT is calculated using the new equation for MRT the
boiler cannot be optimally sized. This is especially true when determining
the number of boiler tubes since there must be a higher number of tubes in
order to lower the internal pressure drop. This cannot be accurately
calculated well enough because the calculated furnace exit temperatures
are higher than actually measured.
• The test results imply that the new equation for MRT can also be applied
for luminous flames since the agreement between values, obtained from
tests on several test boilers of different sizes, and calculated values was
very good, and of virtually the same accuracy for both fuels used in tests
(natural gas and oil).
Hence, the introduced equation for MRT can be generally applied in simplyshaped
enclosures, as this is the case in the furnaces and tubes of the firetube
boilers, since it yields better agreement between measured and
calculated values of heat transfer in boiler.
18

8.1 Test boiler No. 1 with 30 empty tubes
The first series of tests was run on test boiler No. 1 with an empty furnace and
30 empty tubes of 36.5mm net diameter each. The tests were run at differing
loads of which Table 5 shows representative results. The first test was
conducted with fuel oil, while the second and third tests were run with natural
gas.
test
No.
Table 2: Heat transfer in test boiler No. 1 with 30 empty tubes
boiler
load
%
measured
heat transfer
in boiler
kW
calculated
heat transfer
in boiler
kW
measured/calculated
furnace
exit temp.
°C
relative
discrepancy
%
1 40 181.0 187.2 442/398 +3.4
2 50 230.6 239.9 - 7 +4.0
3 95 426.7 418.1 652/684 -2.0
450
400
350
boiler 300
output 250
[kW] 200
150
100
50
0
Graph 4:Output of test boiler No. 1 with 30 empty tubes
furnace
exit
temp.
[°C]
Graph 5: Furnace exit temperature from test boiler No.1 with 30 empty tubes
7 The display of results failed.
800
700
600
500
400
300
200
100
0
40 50 95
measured calculated
measured calculated
boiler load %
50 95
boiler load %
19

Discussion and conclusions
The results in the preceding and following tables and graphs show relatively
good agreement between the calculated and measured heat transfer rates.
The largest discrepancy was +4.0% (test No. 1), while the smallest was -0.7%
(test No. 3). It is evident that in the furnace, a part of the cooled surfaces is
exposed to increased thermal loads, namely where the heat transfer from
impinging jet takes place. Thus, its surface temperature is up to 100% higher
than those of the surrounding walls (test No. 3 in below tables), this value
growing with the boiler load.
The convection in the furnace also increases with the boiler load and can
even exceed the radiant part. Meanwhile, the lower the boiler load, the smaller
the amount of energy transferred in the tubes. By increasing the boiler load,
the part of the heat transfer in the tubes grows, while at the lower loads, the
amount of the heat exchange in the furnace can be as high as 90% of boiler
heat output.
According to Table 5, the velocity in the tubes at the nominal load is 4.4 times
higher than at reduced loads. This results in more than doubling the
percentage of heat transfer versus the transfer at the reduced loads. The
radiation is, due to the very small volume of the tubes, reduced to a negligible
amount and represents less than 5% in the nominal load case and 12% in the
reduced load case. At reduced loads, the computed total heat transfer in the
boiler was higher than actually measured. It was found that the relative
discrepancy between the calculated and measured amounts of heat
transferred in the boiler was not the same at all loads. It was generally
observed that the discrepancy between the actual and calculated values of
the furnace exit temperature becomes higher at reduced boiler loads. A
plausible explanation follows.
At the lower boiler loads, the flame is much shorter so there is a so-called
short cut effect caused by the buoyancy of the flue-gases, which causes the
flue-gases to reach the rear wall with lessened intensity. Hence, the heat
transfer from the impinging jet at the rear wall is greatly reduced and
conditions approach the flow along the plane vertical plate. So, the convection
in furnace is noticeably reduced resulting in an efficiency drop at the reduced
loads. Further, the velocity in the tubes is not equal in all tubes; the upper
tubes and those nearer the vertical center line of the furnace get more fluegases
than those at each side because of flue-gases buoyancy, which comes
more into effect at reduced loads.
Table 3 shows the heat exchange in the furnace assuming the heat transfer
from the impinging jet to the rear wall is not reduced. It is evident that due to
heat transfer from impinging jet, the convection is increased from 9 up to 25
times as it is flue-gases velocity dependent. Despite the rear wall area
represents only 13% of total surface in the furnace, the amount of the
convection makes 65%-83% of the total convection there and grows with fluegases
velocity.
20

Table 3: Calculated heat exchange in furnace of test boiler No. 1 with 30 empty tubes
Test No. in
Table 2
flue-gases
velocity 8
m/s
convective heat
transfer
coefficient
W/m 2 K
convective heat transfer
coefficient from jet impingement
on rear wall
W/m 2 K
radiation heat
transfer
coefficient
W/m 2 K
1 0.8 4.1 36.8 35.1
2 0.7 3.7 34.0 25.1
3 2.0 6.1 154.0 30.9
Test No. in
Table 2
convection
from jet
impingement
kW
total
convection
kW
total
radiation
kW
% of total heat
transferred in
boiler
1 34.1 44.0 172.7 90.3
2 22.8 33.3 139.1 91.1
3 125.2 150.2 187.9 79.7
Test No. in
Table 2
temp. of rear
wall
°C
temp. of furnace
cylindrical walls
°C
1 129.2 107.9
2 117.9 101.3
3 187.5 99.8
Table 4 shows the calculations of heat exchange in the furnace at reduced
loads (tests No. 1 and 2) performed under the assumption the heat transfer
from impinging jet to the rear wall was lessened by 50%. It is clearly evident
that the flue-gases temperature at the furnace exit increased, as
demonstrated by tests discussed in the section to follow. This reduces the
overall heat transfer in the boiler and lessens the efficiency as result.
As there is a part of the area in the tubes with a lessened intensity of the
convective heat transfer due to flue-gases buoyancy, the efficiency of the
boiler is further reduced, which is more apparent in cases of numerous tubes.
This gives the general direction in designing boilers. Namely, there must be
only as many tubes as are necessary. This imposes the need for an accurate
heat transfer analysis in any particular boiler.
Table 4: Calculated heat exchange in furnace of test boiler No. 1 with 30 empty tubes at 50%
reduction of jet impingement
Test
No.
in
Table
2
fluegases
velocity
m/s
convective
heat transfer
coefficient
from jet
impingement
W/m 2 K
convective
heat transfer
coefficient
W/m 2 K
convection
from jet
impingement
kW
total
convection
kW
calculated
furnace exit
temp. at 50%
reduced jet
impingement
°C
calculated
furnace exit
temp. at 100%
jet impingement
°C
1 0.9 18.9 4.2 19.0 30.0 482.6 410.3
2 0.7 17.3 3.8 11.9 24.6 403.1 357.6
8
The flue-gases velocities were calculated at average flue-gases temperature in the particular boiler
section.
21

Test No.
in
Table 2
Table 5: Calculated heat exchange in 30 empty tubes of test boiler No. 1
boiler
load
%
fluegases
velocity
m/s
convective
heat transfer
coefficient
W/m 2 K
convection
kW
radiation
heat transfer
coefficient
W/m 2 K
radiation
kW
% of
total heat
transferred
in boiler
2 40 2.9 11.7 13.9 1.4 1.7 8.2
3 95 12.9 37.7 72.2 1.7 3.8 19.6
The amount of boiler surfaces with reduced convection (the above theory) can
only be guessed, thereby making the accurate assessment of efficiency
deterioration at reduced boiler loads nearly impossible. In the case of fire-tube
boilers, especially those with a reversing furnace, the efficiency drop at
reduced loads seems inevitable. Therefore, the solution to reduce or eliminate
efficiency drop due to less intense convection at reduced loads was
introduced as discussed in following section. Tests on three- and four-pass
fire-tube boilers with numerous tubes were not conducted as it was assumed
that similar results would occur since such boilers have up to 90% of their
internal surface located in the tubes.
8.2 Test boiler No. 2 with 30 empty tubes and a hot tube in furnace
Internal geometry of the test boiler No. 1 was altered by adding the hot tube
(see Figure 5) of high temperature resistant material to the furnace 9 . The
tubes were empty to assure comparability of results with test boiler No. 1.
Tests were run at reduced and nominal loads. The heat transfer in the boiler
was analyzed and compared to measured values as in the tables below.
It was observed that due to the presence of the hot tube, radiation was
lessened by 2.5% at reduced loads and by 6.5% at the nominal load. The
convection, on the other hand, increased by 20%. Since the heat gain in the
furnace due to increased convection was greater than loss on radiation, the
total heat transfer in the furnace increased by 2% at reduced loads up to 5.8%
at nominal loads. This leads to an efficiency increase at the same energy
input in terms of fuel rate as in the case when the hot tube is not present.
Table 6: Heat transfer in test boiler No. 2 with 30 empty tubes and hot tube in furnace
test
No.
boiler
load
%
measured
heat transfer
in boiler
kW
calculated
heat transfer in
boiler
kW
relative
discrepancy
%
measured/
calculated
furnace exit
temperature
°C
1 40 189.6 193.0 +1.7 388/366
2 50 220.6 224.4 +1.7 410/387
3 50 234.7 232.5 -0.9 - 10
4 70 301.6 299.6 -0.6 -
5 95 427.6 431.0 +0.7 585/620
9 Fire-resistant steel was used on test boilers.
10 The display of results failed.
22

Graph 6: Output of test boiler No. 2 with 30 empty tubes and hot tube in furnace
Graph 7: Furnace exit temp. from test boiler No. 2 with 30 empty tubes and hot tube in
furnace
Table 7: Calculated heat exchange in furnace of test boiler No. 2 with 30 empty tubes and hot
tube in furnace
Test No.
in
Table 6
boiler
power
[kW]
fluegases
velocity
inside hot
tube
m/s
450
400
350
300
250
200
150
650
600
flue gas 550
exit temp. 500
from 450
furnace
400
350
[°C] 300
250
200
150
fluegases
velocity
in ringslot
around
hot tube
m/s
40 50 50 70 95
measured calculated
convective
heat
transfer
coefficient
in ring-slot
W/m 2 K
boiler load %
40 50 95
measured calculated
boiler load %
convective heat
transfer
coefficient
from jet
impingement
on rear wall
W/m 2 K
radiation
heat
transfer
coefficient
inside hot
tube
W/m 2 K
radiation
heat
transfer
coefficient
outside
hot tube
W/m 2 K
1 1.6 3.7 7.1 37.4 41.8 26.6
5 4.3 10.2 12.0 176.3 49.2 31.6
23

Test No.
in
Table 6
convection
from jet
impingement
kW
convection
in ring-slot
around
hot tube
kW
total
convection
kW
radiation to
hot tube
kW
total
radiation
kW
% of
total
heat
transfer
in boiler
1 24.0 12.1 40.3 52.4 135.7 91.1
5 143.9 26.7 181.2 68.8 176.6 83.0
Test No.
in Table 6
temp. of
rear wall
°C
Influence of hot tube on emissions
surface temp. of
hot tube
°C
temp. of furnace
cylindrical walls
°C
1 123.6 810 101.6
5 200.8 906 98.0
As shown in Table 8, the application of a hot tube in the furnace improves the
quality of the combustion. This is attributed to reductions of the CO and soot
due to increased temperature of the hot tube, while NOx - emission is
practically not affected (in the best case it is only marginally reduced). There
was even an additional side benefit observed, namely the scaling of furnace
surfaces by unburned sulfur from oil was almost eliminated. It turns out that
the sulfur burns-out on the inside hot tube walls, which heated up red/yellow
(see Table 7).
Table 8: Emissions from natural gas fired test boiler No. 2 with 30 tubes with and without hot
tube in furnace
excess-air CO NOx
soot
-
ppm ppm
Ba
without hot tube 1.19 40 47 0-trace
with hot tube 1.15 10 43 0
Discussion and conclusions
The calculated values shown in Table 6 are in very good agreement with
those actually measured. The largest discrepancy, which was found at
reduced loads, was +1.7%. The differences between the actual and calculated
furnace exit temperatures were as low as 22°C at reduced loads and 35°C at
nominal load. This proves the importance of properly considering the heat
transfer from impinging jet in boilers.
The hot tube causes the radiation in a furnace to drop compared to that in the
furnace without the hot tube. The hot tube’s surface temperature is, compared
to that of surrounding walls, much higher. The hot tube intensifies convection
by increasing the flue-gases velocity, this taking place in a ring-slot around the
hot tube. Since the hot tube reduces the cross-section of the furnace, the fluegases
velocity inside the hot tube also increases (up to 230% in the test
boilers) contributing to intensified convective heat transfer from impinging jet
to rear wall and door (when cooled).
24

The increase in the convective part of the heat transfer in the furnace is higher
than the decrease in the radiant part, resulting in the higher total heat transfer
in the furnace at the same fuel rate and quality of combustion than in the case
without hot tube in the furnace.
Another series of tests with oil under different loads was conducted with the
hot tube installed. The results are displayed in Table 9. Graph 8 shows a
uniform growing efficiency trend even at reduced loads (boiler efficiency does
not drop when the boiler load is reduced). This is due to the fact that fluegases
are forced to go all the way to rear of the furnace before returning to
enter the tubes, thereby maintaining the intensity of the heat transfer from
impinging jet on the rear wall.
So, in the fire-tube boilers, at increased loads, the boiler efficiency always
slightly drops due to an increase in flue-gases exit temperature, while in firetube
boilers with reversing furnace, at reduced loads, the presence of hot tube
in the furnace maintains the efficiency at all loads due to the intensity of heat
transfer from impinging being not only preserved but even intensified.
Table 9: Efficiency of test boiler No. 2 with 30 empty tubes and hot tube in furnace at different
loads
test
No.
boiler
load
%
measured boiler
output 11
kW
calculated heat
transfer in boiler
kW
measured
boiler
efficiency
%
calculated
boiler
efficiency
%
1 55 249.4 (±3.0) 248.6 92.6 (±1.86) 92.3
2 100 449.4 (±5.4) 453.6 91.0 (±1.82) 91.9
3 115 524.3 (±6.3) 525.4 90.9 (±1.82) 91.1
boiler
efficiency
%
95
94
93
92
91
90
89
88
87
86
85
55 100 115
boiler load %
measured efficiency calculated efficiency
Graph 8: Efficiency of test boiler No. 2 with empty 30 tubes and hot tube in furnace at different
loads
11 The values in brackets show range with the maximum measurement uncertainty of 1.22% for a boiler
output and 2.01% for a boiler efficiency, respectively. Example: 1.22% of 524.3kW is 6.3kW and 2.01%
of 90.9% is 1.8.
25

With regard to the influence of hot tube on the heat transfer in the boiler, the
following general conclusions are drawn:
• An application of a hot tube in a furnace of a fire-tube boiler with a
reversing furnace impairs the radiation and intensifies the convection;
• The increase in convection is larger than loss in radiation so total heat
exchange in a furnace is intensified;
• The so called short-cut effect caused by the flue-gases buoyancy at
reduced loads is eliminated, while the flue-gases are forced to go all the
way to rear wall of a furnace maintaining the heat transfer from impinging
jet at all loads;
• A decrease of the furnace exit temperature results in an increased boiler
efficiency compared to cases when a hot tube is not present;
• The presence of the hot tube does not affect emissions of NOx while CO
and soot are reduced due to the high temperature of hot tube’s inner
surface.
• A side benefit caused by the hot tube is the almost complete elimination of
scaling of furnace’s internal surfaces by unburned sulfur from oil.
8.3 Test boiler No. 2 with 30 tubes and coiled-wire turbulators of 1/3 rd tube
length
The next series of tests was run on the test boiler No. 2 (see Figure 5) with
coiled-wire turbulators of one-third tube length and a wire diameter of 5mm
(see Figure 13). Representative results are shown in Table 10 and Table 11.
During all the tests the furnace was installed the hot tube per Figure 5. Table
12 shows the results of the heat transfer analysis in the tubes at various
loads.
Table 10: Results of test boiler No. 2 with 30 tubes with coiled-wire turbulators of 1/3 tube
length at reduced loads
test
No.
boiler
load
%
measured heat
output
kW
calculated heat
output
kW
relative
discrepancy
%
1 50 244.6 242.3 -0.9
2 50 235.9 242.1 +2.5
3 70 297.5 295.8 -0.5
Table 11: Results of test boiler No. 2 with 30 tubes and coiled-wire turbulators of 1/3 tube
length at nominal loads
test
No.
boiler
load
%
measured heat
output
kW
calculated heat
output
kW
relative
discrepancy
%
1 100 446.7 442.2 -1.0
2 105 486.0 487.3 +0.2
3 110 498.1 483.7 -2.8
26

Table 12: Heat exchange in tubes of test boiler No. 2 with 30 tubes with coiled-wire
turbulators of 1/3 tube length
boiler
load
%
fluegases
velocity
m/s
convective
heat transfer
coefficient
W/m 2 K
convection
kW
radiation
heat transfer
coefficient
W/m 2 K
radiation
kW
% of
total heat
transferred in
boiler
50 10.3 34.9 39.9 1.0 1.5 17.0
100 20.5 57.3 87.8 1.4 2.7 20.4
As seen in Table 12, the flue-gases tube velocities at reduced and nominal
loads increased by 3.5 and 2.5 times, respectively, compared to the emptytube
configuration. Accordingly, the convective heat transfer coefficient
increased by 2.9 and 1.5 times, respectively. Also the radiant component was
further reduced to less than 3%. The percentage of heat exchange in the
tubes increased at reduced loads and remained approximately unchanged at
increased loads.
boiler
power
[kW]
500
400
300
Graph 9: Output of test boiler No. 2 with 30 tubes of 1/3 tube length and coiled-wire
turbulators
Discussion and conclusions
200
50 70 100 105 110
boiler load %
measured calculated
The calculations showed almost equally good agreement with the measured
data as in the case with no turbulators. The largest discrepancy was -2.8%
(third nominal loads test). The discrepancies were approximately the same at
reduced and nominal loads. According to [14], where a minimum error of
±10% in heat exchange in tubes with turbulators was reported, this represents
a sizable improvement in accuracy being achieved, from a minimum ±10% to
less than ±3%.
27

8.4 Test boiler No. 2 with 26 tubes and coiled-wire turbulators of 1/3 rd tube
length
The test boiler per Figure 5 was made with 26 tubes of 40% greater, 51.2 mm
net tube diameter each. The turbulators of one-third tube length and of 100%
greater, 10 mm wire diameter were used (see Figure 14). The boiler was run
at different loads with hot tube in furnace (see Figure 5). Representative
results are shown in Table 13.
Table 13: Results of test boiler No. 2 with 26 tubes and coiled-wire turbulators of 1/3 tube
length at different loads
test
No.
boiler
load
%
measured heat
output
kW
calculated heat
output
kW
relative
discrepancy
%
1 95 435.8 427.4 -1.9
2 95 421.4 417.5 -0.9
3 80 362.0 353.9 -2.2
4 70 305.4 301.8 -1.1
5 60 281.6 278.1 -1.2
boiler
power
[kW]
Graph 10: Output of test boiler No. 2 with 26 tubes and coiled-wire turbulators of 1/3 tube
length
Discussion and conclusions
450
430
410
390
370
350
330
310
290
270
250
60 70 80 95 95
boiler load %
measured calculated
The largest discrepancy between calculated and measured output of the
boiler was -2.2% (third test) while the minimum was -0.9% (second test). The
discrepancies between the measured and calculated values remained at the
same level as in the previous cases. The accuracy of the calculations did not
change despite the significant change in tube, turbulator, and wire diameters.
This indicates that the rate of heat exchange in tubes with coiled-wire
turbulator may not change with change of turbulator geometry and was further
researched in the section to follow.
28

8.5 Test boiler No. 2 with 26 tubes with coiled-wire turbulators of different
geometries
Additional tests on boiler No. 2 (see Figure 5) with same turbulator length but
of different geometries (see Figure 15) were conducted. The test results led to
the conclusions given below.
Impact of turbulator pitch, diameter, and wire diameter
As is evident from Table 14, the tests clearly show the approach introduced
delivers the approximately same accuracy of calculation regardless of the
turbulator pitch, diameter, and wire diameter so long as the length of
turbulator is kept the same 12 . The exception is test No. 1 with turbulators of an
almost minimum possible pitch, almost equal to the wire diameter.
Table 14: Results of test boiler No. 2 with 26 tubes with coiled-wire turbulators of different
geometries of 1/3 tube length
test
No.
boiler
load
%
turbulator
diameter
mm
wire
diameter
pitch
mm
measured
heat
transfer in
boiler
calculated
heat transfer
in boiler
relative
discrepancy
%
1 100 50 8 10 424.4 445.2 +4.9
2 100 50 8 25 446.9 442.1 -1.0
3 115 50 8 65 524.3 523.2 -0.2
4 50 50 8 65 242.5 235.6 -2.8
5 100 48 6 35 451.7 447.2 -0.9
6 50 48 6 35 228.4 227.2 -0.5
7 50 48 6 53 239.8 238.5 -0.5
8 100 48 6 53 442.7 439.8 -0.6
9 100 45 8 32 453.8 448.9 -1.0
10 50 45 8 32 232.5 233.0 +0.2
boiler
power
[kW]
500
450
400
350
300
250
200
10 25 32 53 65
turbulator pitch [mm]
measured calculated
Graph 11: Output of test boiler No. 2 with 26 tubes with coiled-wire turbulators of different
geometries of 1/3 tube length
12 In this case the turbulator length was one third tube length.
29

Impact of turbulator length
Unlike turbulator configuration, turbulator length affects the amount of
convection in boiler tubes. Heat exchange in boiler tubes with turbulators
longer than one third tube length can be calculated with approximately the
same accuracy as with turbulators of one third tube length when the following
modifications are applied:
1. Turbulators of two third tube length: Heat exchange in tubes is calculated
based on turbulators of on third tube length multiplied by a factor of 1.1;
2. Turbulators of full tube length: Heat exchange in tubes is calculated based
on turbulators of on third tube length multiplied by a factor of 1.2.
Table 15 and Table 16 show the representative test results when turbulators
of two third and full tube length, respectively, were used.
Table 15: Results of test boiler No. 2 with 26 tubes with coiled-wire turbulators of 2/3 tube
length
test No. boiler load
%
measured heat
output
kW
calculated heat
transfer with
coefficient 1.1
kW
relative
discrepancy
%
1 100 453.8 448.9 -1.0
2 100 448.2 450.9 +0.6
3 50 232.5 233.0 +0.2
4 50 237.5 232.7 -2.0
Table 16: Results of test boiler No. 2 with 26 tubes with coiled-wire turbulators of full tube
length
test No. boiler load
%
measured heat
output
kW
calculated heat
transfer with
coefficient 1.2
kW
relative
discrepancy
%
1 100 455.5 458.2 +0.5
2 100 446.2 452.6 -1.4
3 50 237.6 232.7 -2.0
4 50 242.5 235.6 -2.8
Discussion and conclusions
The empirical coefficients suggested on the basis of the test results brought
the calculation accuracy of the impact of turbulators longer than one-third tube
length on heat exchange to a ±2% level. Thus, the same accuracy in heat
transfer calculation in straight tubes with coiled-wire turbulators is maintained
over a wide range of turbulator geometries in terms of its diameter, wire
diameter, pitch, and length. The results lead to conclusion that the turbulator
length affects the heat exchange rate in a linear fashion.
30

9. Tests on pressure drop
9.1 Test boiler No. 2 with 26 empty tubes
The flue-gases side pressure losses in a boiler significantly influence the
operation of the boiler. In cases where there are too high flue-gases side
pressure losses in the boiler, the burner fan cannot overcome the pressure.
This eventually leads to flame extinction. So, the accuracy in the analytical
assessment of pressure drop in the boiler must be maintained. This is
particularly needed when there are turbulators as they increase the heat
exchange but increase the flue-gases side pressure losses in the boiler as
well.
The next series of tests on boiler shown in Figure 5 with 26 tubes was run with
the boiler’s internal flue-gases side pressure losses being measured.
Calculations according were compared to these measured values. Table 17
shows representative results of the pressure-drop tests on a test boiler of
conventional design with 26 empty tubes at different loads with and without a
hot tube in furnace. While calculating the pressure drop in boiler, its internal
geometry was taken into account so the coefficients for the pressure drop
equation were calculated separately for each boiler section. Since the boiler
tubes were empty, the friction coefficient of plane tube was used in the
equation for pressure drop.
Table 17: Pressure drop in test boiler No. 2 with 26 empty tubes with and without hot tube
boiler load
%
measured gas
side pressure
drop in boiler
Pa
calculated gas side
pressure drop in
boiler
Pa
relative
discrepancy
%
without hot tube 50 11 10.6 -3.6
with hot tube 50 14 14.4 +2.8
without hot tube 100 60 53 -11.6
with hot tube 100 97 88 -9.2
pressure
drop
[Pa]
120
100
80
60
40
20
0
50 100 50 100
with hot tube without hot tube
load [%]
measured calculated
Graph 12: Pressure drop in test boiler No. 2 with 26 empty tubes with and without hot tube
31

Discussion and conclusions
The presence of the hot tube in furnace causes an approximately 30%
increase in the flue-gases side pressure drop at reduced loads and an
approximately 60% at nominal loads. Since the boiler internal pressure drop in
the 100 Pa range is considered to be a relatively small, the presence of hot
tube does not make additional requirements on the burner. The extent of the
pressure drop by the hot tube is ring-slot thickness dependent. To maintain a
pressure drop in the furnace to within acceptable limits, the area of ring-slot
flow cross-section was found to need to be a minimum 25-30% of the furnace
cross-section. It was observed that, at higher loads, the percentage of the
pressure drop increase, with the hot tube in furnace, is larger than found at
lesser loads (as was expected due to higher velocities at the increased loads).
The largest discrepancy between the predicted and measured pressure drop
of 11.7%, can be treated as being sufficiently accurate. Such results were
expected as the valid equations were generally applicable to empty tubes and
channels, not ones accommodating turbulators.
9.2 Test boiler No. 2 with 26 tubes with coiled-wire turbulators of different
geometries
Further tests were run on the boiler (see Figure 5) with the coiled-wire
turbulators of varying geometries. Table 18 shows representative results at
these loads.
Discussion and conclusions
As is evident from Table 18, the pressure drop and discrepancy value
increased when the turbulator pitch and diameter, compared to the tube
diameter, decreased. This trend is especially evident at increased loads while
less at reduced loads. It was observed that increasing the turbulator length did
not increase the pressure drop predication accuracy. The general conclusion
is that as the turbulator pitch and diameter get smaller to the tube diameter,
the higher the pressure drop and negative discrepancy between the
calculated and measured values.
Further tests resulted demonstrating that the approach used yields mostly
positive discrepancies in pressure drop calculations in straight tubes with
coiled-wire turbulators. These discrepancies are within maximum +10-15%
range when the turbulator pitch, diameter, length, and wire diameter match
the criteria shown in Figure 17.
As the calculated pressure drop was found to be typically higher than that
which was actually measured, a known safety margin factor can be assumed.
This thereby prevents an overload of the burner fan, which would otherwise
lead to the boiler’s inability to reach a nominal load as the burner would not be
able to overcome the flue-gases side pressure in the boiler.
32

test
No.
D
min. 0.8 D
min. 1/3 L
L
max. 0.5 mm
Figure 17: Geometry of coiled-wire turbulators to assure maximum +10-15% inaccuracy in
analytical assessment of pressure drop in tubes
Table 18: Pressure drop in test boiler No. 2 with 26 tubes with turbulators of different
geometries
boiler
load
[%]
length
[% of
tube]
tube
diameter
[mm]
turbulator
diameter/
wire
diameter
[mm]
pitch
[mm]
measured
pressure
drop
[Pa]
calculated
pressure
drop
[Pa]
relative
discrepancy
[%]
1 115 100 51 50/8 65 355 370 +4.2
2 100 100 51 50/8 65 226 265 +17.2
3 100 66 51 45/8 32 405 235 -41.9
4 50 66 51 45/8 32 110 92 -16.3
5 100 50 51 50/8 25 305 272 -10.8
6 100 33 51 50/8 10 255 284 +11.3
7 100 100 51 48/6 53 200 205 +2.5
8 100 100 51 48/6 35 328 216 -34.1
9 50 100 51 48/6 35 74 47 -36.4
10 50 66 51 50/6 53 52 47 -9.6
11 100 66 51 48/6 53 180 201 +11.6
12 50 66 51 48/6 35 59 41 -30.5
13 100 66 51 48/6 35 278 208 -25.1
14 100 50 51 48/6 35 229 216 -5.6
15 100 33 51 48/6 35 191 228 +25.9
16 50 33 51 48/6 35 40 41 +2.5
10. Tests on boilers of different furnace geometries
The test boiler was gradually redesigned as shown in Figure 6 and Figure 7.
The impact of each design feature on heat transfer and emissions was
investigated and analyzed. During the tests, the hot tube (see Figure 5) was
inserted in the furnace and there were one-third tube length coiled-wire
turbulators in the tubes per Figure 14. All tests were run at nominal boiler
loads. While calculating the emissions of CO and NOx, a 3% of rest oxygen in
the flue-gases was taken a base comparison reference.
33

10.1 Comparison of test boilers No. 2 and No. 3
Influence of cooled door on emissions
As is evident from Table 19, showing representative results, the door cooling
negligibly affects the CO, soot and NOx emissions. The data in the table
below were obtained from tests with oil. Approximately same relative trend
was observed in the case of natural gas. The slight decrease of NOx -
emissions from the boiler with cooled door is based on a decreased average
temperature in the furnace due to the additional cooled surfaces on the boiler
door.
Table 19: Emissions from test boiler with cooled and uncooled door
CO
NOx
soot
ppm
ppm
Ba
boiler without cooled door 40 102 0-trace
boiler with cooled door 45 97 0-trace
Influence on heat transfer
Tables below convey the analyzed heat exchange in the furnace and tubes
with and without the cooled door while holding the other parameters of
influence (e.g. water flow, water temperatures, fuel rate, quality of
combustion) constant. Since in the third boiler segment, the smoke chamber
at rear, there is a negligible heat exchange due to low temperature and low
flue-gases velocity, the heat exchange there was not taken in consideration.
exit
temp.
°C
Table 20: Calculated heat exchange in furnace of test boiler No. 2 without cooled door
flue-gases
velocity
inside/
outside hot
tube
m/s
convective
heat transfer
coefficient
inside/
outside
hot tube
W/m 2 K
convective
heat
transfer
coefficient
at rear
wall 13
W/m 2 K
convection
to rear wall/
total
convection
kW
radiation
heat transfer
coefficient
inside/
outside hot
tube
W/m 2 K
radiation
kW
temp. of rear
wall/
other cooled
surfaces
°C
832.6 3.9/10.4 7.5/12.3 172.0 61.2/164.9 46.5/37.1 129.0 235.2/109.4
Table 21: Calculated heat exchange in tubes of test boiler No. 2 without cooled door
exit
temp.
°C
fluegases
velocity
m/s
convective
heat transfer
coefficient
W/m 2 K
convection
kW
radiation
heat transfer
coefficient
W/m 2 K
radiation
kW
wall
temp.
°C
% of total
heat
transferred
in boiler
319.6 16.5 43.0 109.3 2.3 7.1 101.4 28.0
13 Area of heat transfer from jet impingement.
34

Table 22: Calculated heat exchange in furnace of test boiler No. 3 with cooled door
exit
temp.
°C
fluegases
velocity
inside/
outside
hot tube
m/s
680.0 3.8/
9.6
exit
temp.
°C
convective
heat
transfer
coefficient
inside/
outside
hot tube
W/m 2 K
7.2/
11.7
convective
heat
transfer
coefficient
at rear wall/
door 14
W/m 2 K
165.2/
157.9
convection
to rear wall/
door/
total
convection
kW
53.9/
96.5/
183.0
radiation
heat
transfer
coefficient
inside/
outside hot
tube
W/m 2 K
44.8/
36.8
radiation
kW
temp. of
rear wall/
door/
other
cooled
surfaces
°C
146.9 214.1/
189.3/
106.1
Table 23: Calculated heat exchange in tubes of test boiler No. 3 with cooled door
fluegases
velocity
m/s
convective
heat transfer
coefficient
W/m 2 K
convection
kW
radiation
heat transfer
coefficient
W/m 2 K
radiation
kW
wall
temp.
°C
% of total
heat
transferred
in boiler
280.1 14.8 41.1 84.3 1.8 4.4 99.3 21.0
Discussion and conclusions
The cooled door represents a 15% increase in cooled surfaces in furnace and
intensifies heat transfer there by 10.9%. The flue-gases and exit temperatures
from furnace were reduced by 8.4% and 18.3%, respectively. As the fluegases
exit temperature from the boiler was lowered by 37°C, the slight
increase in boiler efficiency (by 1.7%) results. The wall temperature was
almost unaffected. Radiation increased by 13.8%, while convection increased
by 9%.
While the cooled door was proven as beneficial to the boiler, it affects the NOx
- emissions only slightly while the other emissions are not effected at all. The
door and furnace rear wall are exposed to more intense heat transfer from
impinging jet, their wall temperatures rise well above the other cooled
surfaces. Due to an intensified heat exchange in the furnace, the flue-gases
enter the tubes with a lower velocity resulting in less intense convection and
radiation so the total heat exchange in tubes is reduced. As radiation
represents small part of the total heat transfer in the tubes (less than 6% in
test boiler), its reduction is of no importance and, so it can be considered to be
negligible. Further it is evident that the fire-tube boilers can be designed in
such a way that the convection in the furnace is higher then a radiation.
Despite this design feature was introduced for the purposes of the tests it was
found as fully applicable on hot water boilers for central heating.
10.2 Test boiler No. 4
Next, the test boiler cooled door surface was increased (see Figure 7 and
section 1.1.1.4). A more intense heat transfer and pre-cooling of the flame
caused a reduction in NOx as was expected.
14 Area of heat transfer from jet impingement.
35

Influence on emissions
As is evident from representative results shown in Table 24, with liquid fuels
the cooled cylinder on cooled door contributed to up to 10% less NOx
production, while the CO emission was only slightly increased and the soot
emission remained unchanged. With gaseous fuels there is only a marginal
change in NOx-emission and slight increase in CO-emission with no change
in soot-emission.
cooled door with
cooled cylinder
cooled door without
cooled cylinder
Influence on heat transfer
Table 24: Emissions from test boiler No. 4
oil natural gas
NOx soot CO NOx soot CO
ppm Ba ppm ppm Ba ppm
90 0-trace 38 65 0 33
99 0-trace 28 62 0 51
Tables below depict the heat exchange in the furnace and tubes. As the
cooled cylinder on the cooled door represents an increased amount of cooled
surfaces in the furnace, the heat transfer was more intensified and flue-gases
temperatures further reduced.
exit
temp.
°C
Table 25: Calculated heat exchange in furnace of test boiler No. 4
fluegases
velocity
inside/
outside
hot tube
m/s
645.6 3.8/
9.5
exit
temp.
°C
convective
heat
transfer
coefficient
inside/
outside
hot tube
W/m 2 K
7.1/
11.6
convective
heat
transfer
coefficient
at rear wall/
door
W/m 2 K
163.5/
156.2
convection
to rear wall/
door/
total
convection
kW
52.3/
93.1/
176.8
radiation
heat
transfer
coefficient
inside/
outside hot
tube
W/m 2 K
44.5/
32.8
radiation
kW
Table 26: Calculated heat exchange in tubes of test boiler No. 4
fluegases
velocity
m/s
convective
heat transfer
coefficient
W/m 2 K
convection
kW
radiation
heat transfer
coefficient
W/m 2 K
radiation
kW
temp. of
rear wall/
door/
other
cooled
surfaces
°C
161.1 206.7/
149.5/
105.0
wall
temp.
°C
% of total
heat
transferred
in boiler
270.8 14.4 40.7 78.8 1.7 3.9 98.8 19.5
Discussion and conclusions
A cooled cylinder on the cooled door further intensifies the heat exchange in
the furnace by introducing additional cooled surfaces. The flue-gases average
temperature and their exit temperature from furnace are lowered, thereby
36

educing the heat exchange in the tubes. As the flue-gases exit temperature
from the boiler is lowered, there was slight increase in boiler efficiency of
0.4%. The cooled cylinder also influences the NOx - emissions. The larger the
difference between the flue-gases flow cross-sections of the cooled cylinder
and that of furnace, the greater the effect of local recirculation (see section
1.1.1.4). This contributes to a lowered flame temperature and, herewith,
reduction of NOx.
The inside diameter of furnace/cooled cylinder ratio is, however, limited
because the door opening must not be disturbed, while the inside diameter
and length of the cooled cylinder must be adapted to match the burner oil
nozzle sprinkling angle. The flame must not touch the hot tube’s walls as this
would lead to increase in CO and soot.
Despite this design feature was introduced for the purposes of the tests it
could be applicable in the practice.
10.3 Test boiler No. 5
In order to extend the validation of the presented algorithm to larger boilers,
tests were conducted on an industrial-sized steam boiler in an actual
application (see www.wargaboiler.com). Tables below depict the heat
exchange in boiler furnace and tubes at full loads. The fuel’s thermal energy
was first transferred to the boiler water, which turned into steam without
intermediate heat losses. The boiler jacket losses were measured and taken
into account by deducting them from the flue-gases enthalpy.
boiler
load
%
Table 27: Measured and calculated heat transfer in test boiler No. 5
measured heat
output
kW
calculated
heat output
kW
relative
discrepancy
%
measured
steam flow
kg/h
calculated
steam flow
kg/h
relative
discrepancy
%
97 3831.8 3858.0 +0.68 5643 5703.5 +1.0
exit
temp.
°C
fluegases
velocity
inside/
outside
hot tube
m/s
734.2 8.1/
24.6
Table 28: Calculated heat exchange in furnace of test boiler No. 5
convective
heat
transfer
coefficient
inside/
outside
hot tube
W/m 2 K
11.5/
36.9
convective
heat
transfer
coefficient
at rear wall/
door
W/m 2 K
237.8/
450.4
convection
to rear wall/
door/
total
convection
kW
311.8/
712.5/
1586.0
radiation
heat
transfer
coefficient
inside/
outside/
around hot
tube
W/m 2 K
82.5/
35.3/
6.1
radiation
kW
temp. of
rear wall/
door/
other
cooled
surfaces
°C
1210.1 281.3/
302.8/
215.6
surface
temp.
of hot
tube
°C
1041
37

exit
temp.
°C
Table 29: Calculated heat exchange in tubes of test boiler No. 5
fluegases
velocity
m/s
convective
heat
transfer
coefficient
W/m 2 K
convection
kW
radiation
heat
transfer
coefficient
W/m 2 K
radiation
kW
wall
temp.
°C
% of total
heat
transferred
in boiler
261.7 23.0 53.1 829.4 2.9 64.5 177.9 27.4
Discussion and conclusions
As is evident from the above tables, there was only a negligible difference of
+0.68 - 1% between the measured and calculated values. This proves that the
algorithm presented is fully applicable on fire-tube boilers of arbitrary size and
internal geometries. It is additionally proven that the cooled door on bigger
boilers greatly improves boiler performance as it noticeably intensifies the
heat exchange in the furnace.
Despite the fact the tests were performed on 2-pass fire-tube boiler designs
with a reversing furnace, it can be inferred, on the basis of test results and
basic similarity of all types of boilers, the presented algorithm can be generally
applicable in any fire-tube boiler with the same degree of accuracy.
Such a boiler was designed for the purposes of the tests, has been found as
fully applicable and was therefore introduced to production (see
www.wargaboiler.com).
38