Available online at www.sciencedirect.

com

Procedia Engineering 31 (2012) 834 – 843

International Conference on Advances in Computational Modeling and Simulation

Full-scale Mathematical Model and Simulation of Marine

Natural Recirculation Drum-boiler
Tao Honga, He Nia,b*, Gang Chenga
a
Research certain of naval power plant simulation, Naval University of Engineering, Wuhan, china
b
Department of Power Engineering, Naval University of Engineering, Wuhan, China

Abstract
Through some reasonable hypothesis and simplification of conflagation, heat and mass transfer, a full-scale dynamic
mathematic model of marine natural recirculation drum-boiler is established by modular modeling method via the
basic working elements of objects. The dynamic feature of a certain type natural recirculation drum-boiler was
studied by this model as an instance, by contrasting the simulative results and real values, the precision and real-time
capability of the developed model was testified. This study was helpful for the performance research, working
optimize and control strategy building of various natural recirculation drum-boilers.
Keywords: natural recirculation drum-boiler, simulation, mathematic model, dynamic feature of thermal-power system

1. Introduction

Drum-boiler is the important part of marine steam power plant[1]. It provide steam which possesses
the needed pressure and temperature for every equipments in power plant, its working circs directly
influence the safety, stability and economization. As a effective method that study the dynamic feature of
object, the full-scale mathematical model of marine drum-boilers have basilic significance and applied
value.
Scholars have conducted extensive research in the aspects of modeling and simulation of marine
boiler. Out of mass conservation law, Astrom[2] Established a model for water level of drum boiler on
the assumption that the mass share of steam of evaporation tubes is linear distribution. The operation
result of this model was good agreement with experimental data in the trends, but the static simulation
accuracy is not satisfied. Then Adam and Kim established the distrbuted parameter models for the
evaporation tubes respectively used the homogeneous models of steam-water mixture[3] and vapor and
liquid two-phase flow theory[4], the models were both described the flow of steam-water mixture in
drum-boiler more detailed, and reflect the characteristics of boiler water recirculation more reasonable.
Tian L established a centrain nonlinear general model[5,6] for drum-boiler through the analysis of emergy

* Corresponding author:Tel.: +86-13871124621; fax: +86-27-83444759.

E-mail address: [email protected].

1877-7058 © 2011 Published by Elsevier Ltd. Open access under CC BY-NC-ND license.
doi:10.1016/j.proeng.2012.01.1109
 Tao Hong et al. / Procedia Engineering 31 (2012) 834 – 843 835

balabces that exist in the heat storage of boiler and steam pipes. Those moedls have simple form, and
were successfully applied in the analysis of boiler control system and the study of boiler load-pressure
characteristics, but they are not the full-scale model of objects, can not reflect the dynamic characteristics
changes in the processes of boiler startup, boiler stop or load change widely
For a more comprehensive study of the boiler characteristics, scholars usually requires the dervations
of the thermal process of boilers by mechanism modeling [7~9]. This paper adopted modular modeling
method, and divided the object into several independent modulars by temporarily cut off the interrelations
between every parts of marine natural recirculation drum-boiler. Then the models of every modulars were
established according to the basic working principle of object. Finally, through the module bonding, the
whole mathematic model of marine natural recirculation drum-boiler was structured.

2. Modeling process

2.1. module division

Out of the comprehensive consideration of volum, reliability and other aspects, most marine steam
power plant adopt natural water recirculation boiler[10]. This paper adopt modular modeling method to
study the modeling and simulation of marine natural water recirculation boiler, divided the whole boiler
into six parts: drum, combustor, water wall, convective evaporation pipe bundles, superheater and decline
pipes. Among them, combustor is the starting point of the mathematical moding of object, the heat flux of
thermal evaporation and superheat system can be derived by calculate the gas temperature distributions
outflow combustor, then the temperature and heat flux distribution of water wall, convective evaporation
pipe bundles and superheater can be calculated, that provide the boundary conditions for the calculation
of flow parameters and distribution. Drum modular was mainly used to calculate the dynamic changes of
steam pressure, water level and other parameters. The relationships between each modules showen in Fig.1.

Fig.1. relationship of model modules of natural recirculation drum-boiler

In Fig.1: a refers to the outlet gas temperature of combustor, b refers to the flue gas themperature distribution at the enter of convective
evaporation pipe bundles, c refers to the parameters of the steam-water mixture that outflow water wall, d refers to the parameters of the
steam-water mixture that outflow convective evaporation pipe bundles, e refers to the flue gas themperature distribution at the enter of
superheater, f refers to the parameters of superheated steam, g 1~ 3 respectively refer to the outer temperatures of water wall, convective
evaporation pipe and superheater, h1~ 3 respectively refer to the heat fluxs of water wall, convective evaporation pipe and superheater,
i refers to the mass flow rate of superheated steam, j refers to the drum pressure, boiler water enthalpy and density, k refers to the mass
flux and enthalpy of supply water, l refers to the outlet steam flow rate of drum, m refers to the boiler water level, n refers to the
parameters of the water that outflow decline pipes.
2.2. modeling principles

Boiler plant is a nonlinear distributed parameter object which possesses typical characters such as pure
836 Tao Hong et al. / Procedia Engineering 31 (2012) 834 – 843

lag, large inertia and so on, its working process is very complex. In this paper, the modeling assumptions
are as follows:
(1) steam and water separate completely in drum
(2) refrigerant mixed uniformly, can be treat as lumped parameter
(3) steam and water uniformly mixed in water wall and convective evaporation pipe bundles, and ignore
fluid inertia of the mixture
(4) do not consider the radiation heat transfer of convective evaporation pipe bundles and superheater,
the radiation of combustor is completely absorbed by water wall
(5) ignores the axial heat conduction of tubes
(6) the circumferential heat of convective evaporation pipe bundles and superheater are uniform
(7) when the model involves the structure sizes of object, using the average
(8) treat combustor as a gray body that composed by gray wall and gray gas
(9) treat the combustion process as heating process, treat gas as ideal gas

3. Modular models

3.1. drum

In order to correctly reflect the dynamics of steam pressure, water level and temperature of drum, paper
divided the internal volum of boiler into liquid and vapor two regions. Because the water storage capacity of
marine drum-bouler is usually larger, and the water supply method usually adopts direct injection mode, so
when boiler stable running, the boiler water in drum can be approximatly considered to be saturated[2,7,11].
(1) liquid region
The flows enter and exit the liquid region of drum are supply water flow Wgsh decline pipes inlet
flow Wd the moisture content of drum inlet steam-water mixture flow Wsl (1  F sl 2 )  Wdl (1  F dl 2 ) , dynamic
evaporation and condensation flow Wdv , and boiler sewage flow Wex 
According to assumption (2), established the mass conservation equation as follows:
dVbw dt X bw [Wgsh  Wsl (1  F sl 2 )  Wdl (1  F dl 2 )  Wd  Wex  Wdv ] (1)
Where: Vbw is boiler water volume in drum F sl 2 and F dl 2 respectively refer to the dryness of outflow
steam-water mixture of water wall and convective evaporation pipe bundles, X bw is specific volume of
saurated water under drum pressure p  b

Only adopt formula (1) to calculate boiler water level may cause a larger errors, there are two main
reasons: (a)when steam-water mixture flow through steam-water separator, there would be some steam
was entrained into drum liquid region by water droplets, (b) the dynamic evaporation and condensation of
boiler water is not only occurring in the vapor and liquid two-phase surface, but appeared in the entire
liquid region, that make the liquid region expansion thereby bring the ³falsehood water level´. These two
factors should be taken into account when calculating the actual volume of drum liquid region:
*
Vbw Vbw  (Wsl  Wdl ) F bwXbs 'tbs I  Wdv (Xbs  X bw ) (2)
*
Where: Vbw is the actual volume of drum liquid region, X bs is specific volume of saurated steam under drum
pressure p  F bw is the mass ratio of steam entrained by water droplets, I is boiler Recycling Ratio, 'tbs is
b

the average residence time that steam remain in liquid region, can be determined by experience or test.
When the accuracy is not critical, we can ignore the internal structure and external protrusions of drum,
and approximate it as a lying cylinder, then the boiler water level can be calculated by the following formula:
­°S ( 4  S  ( 4  S ) 2  32Vbw* Lb rb2 ) 16 ˈVbw* d S rb2 Lb 2
lb rb (1  cos D )ˈ D ® (3)
2 * 2 * 2
°̄S (12  S  S  24S  16  32Vbw Lb rb ) 16 ˈVbw ! S rb Lb 2
 Tao Hong et al. / Procedia Engineering 31 (2012) 834 – 843 837

Where: Lb and rb respectively refer to the equivalent length and radius of drum, D is water lever angle.
(2) vapor region
The flows enter and exit the liquid region of drum are saturated steam flow Wbs 2 , superheater inlet
steam flow Wbgr , drum inlet steam flow Wsl F sl 2  Wdl F dl 2 dynamic evaporation and condensation flow Wdv ,
and steam flow of discharge valve Wexh  According to mass conservation, we have the following formula:
*
(Vb 0  Vbw ) Kbps wpb wt Wdv  Wsl Fsl 2  Wdl Fdl 2  Wbs 2  Wbgr  Wexh (4)
Where: K bps wU bs wpb is the compress coefficient of saturated steam, can be determined by corresponding
water and water vapor function, Vb 0 is the gross volume of drum.
Boiler¶s dynamic evaporation and condensation have a great relationship with the enthalpy of boiler
water, can be calculated by the follow formula:
Wdv K dv ( hw  hbw ) (5)
In fromula (5), the specific enthalpy of saturated water can be expressed as the single-valued
function of drum pressure
' ' '
hbw hphw ( pb ) | hbw  ( pb  pb ) dhbw dpb (6)
Where: h w is the specific enthalpy of boiler water in drum, superscript ' refer to the previous time.
When boiler load changes, drum pressure changes is certainly faster than the enthalpy of boiler water.
Therefore, we can thus believe that the specific enthalpy of boiler water h w is the saturated water specific
'
enthalpy under the pressure of the previous time, there are: hw hbw . Then fromula (5) can be rewritten as:
' '
Wdv  K dv ( pb  pb ) dhbw dpb (7)
Substituting fromula (7) into fromula (4), and processed by Euler method:
pb 'W (Wsl Fsl 2  Wdl Fdl 2  Wbs 2  Wbgr  Wexh ) [(Vb 0  Vbw ) Kbps  Kdv 'W dhbw dpb ]  pb (8)
* ' '

Where: 'W is calculation time step.

3.2. combustor

The combustors of marine boiler are extremely compact, the distribution characteristics of internal
temperature field are not obvious when working, so the combustion flue gas can be treated as a lumped
parameter object. According to assumption (9), established the energy conservation equation as follows:
0.5mg Cpg dTg dt Wo Ho (1 kb )  Qbra  (Wo Wa )(Tg  T0 ) (9)
Where: Tg is the average temperature of gas in combustor, Wo and Wa are the fuel and air feed flows of
boiler, H o is the chemical exergy of fuel, kb is the thermal coefficient of boiler combustor, m g and C pg are
gas mass and specific heat capacity at constant pressure.
According to assumption (4), radiant heat transfer rate can be calculated by the following formula:
4 4
Qbra V Asl (Tg  H slTsl1 ) (10)
Where: V is boltzmann coefficient, Asl and Tsl 1 respectively are the effective radiation area and average
wall temperature of water wall, H sl is the grayscale of water wall, usually taken as constant.
Limited by testing technologies, the temperature distribution of outflow gas is difficult to obtain by
experiment, if adopt theoretical calculations, the problem will be very complicated for the complexity of
internal structure and combustion process. Therefore, paper adopted parabolic empirical formula[12] that
commonly used in engineering when calculated the outflow gas temperature of combustor. Assuming that
the gas temperature presented parabolic distribution along the width of flue, the highest and lowest
temperatures were respectively appeared in the middle and both sides of the flue:
2
Tgx Tg  xyd (1  H x ) 6  H x xyd 1 12 (11)
838 Tao Hong et al. / Procedia Engineering 31 (2012) 834 – 843

Where: Tgx is the average gas temperature in the width of the flue, H x is horizontal flue gas temperature
deviation coefficient, usually take the empirical value, x yd  [ 0, 0.5] is the horizontal relative position
between calculated point and the midpoint of the flue.
The changes of gas temperature on the height of the flue has some relationship with the arrangement
of combustor shape and evaporative pipe bundles, but in general there is a decreasing trend with the
increase of relative height, and it is essentially a linear relationship[11,13]. So in this paper, a linear model
was adopted to approximately simulate the changes of gas temperature along the height of flue:
Tgxy Tgx (1  H y 2  H y y yd ) (12)
Where: Tgxy is the flue gas temperature, H x is vertical flue gas temperature deviation coefficient, usually
take the empirical value, y yd  [ 0, 1] is the relative height of calculated point.

3.3. water wall

According to assumption (3)and(4), took the outlet parameters as lumped parameters, established the
mass, momentum and energy conservation equation as follows:
dX W2 dh h h
Vsl sl X sl2 (Wsl1  Wsl 2 )ˈ [ sl sl 2 ( pd  pb )X sl  'H sl gˈ Vsl sl X sl [(Wsl1  Wsl 2 ) sl wd  Qsl ] (13)
dt 2 dt 2
Where: Vsl , X sl , hsl , Q sl respectively are the volume, specific volume, specific enthalpy and heat adsoration
of steam-water mixture in water wall, Wsl 1 and Wsl 2 are the fluid flow that inflow and outflow water wall,
p d and hwd are the outlet pressure and enthalpy of decline pipes, 'H sl and [ sl are the relative height and
pressure loss coefficient of water wall.
Pressure loss coefficient can be taken as constant when accuracy is not critical. Yet the heat adsoration
of steam-water mixture can be calculated by the following empirical formula[14]:
4/3 3
Qsl Ksl Asl psl (Tslm 2  Tsl ) (14)
Where: Tsl is the saurated temperature of steam-water mixture under drum pressure p , Tslm 2 and K sl areb

the aversge temperature and convection heat transfer coefficient of the inner surface of water wall.
The dryness of the outflow steam of water wall can be calcultaed by the enthalpy of steam:
F sl 2 ( hsl 2  hbw ) ( hbs  hbw ) (15)
h h
Where: bs and bw are the specific enthalpy of saurated steam and water under drum pressure b  p
The thermal storage equation can be derived by energy balance:
msl Cmsl dTslm 2 dt Qbra  Qsl (16)
Where: m sl and Cmsl are the mass and specific heat capacity of metal heat-transfer surface of water wall.
According to assumption (5), the temperature of water wall were two-dimensional distribution along
the radial direction of pipe walls. Since the pipe wall of marine boiler is not thick, and the heat conductivity
of metal surface is extremely fast, so we adopted the two-dimensional steady thermal conductivity formula,
take fromula (14) and (16) as the boundary conditions to calculate the temperature distribution of pipe walls:
r 2 w 2Tslm wr 2  r wTslm wr  w 2Tslm wM 2 0 (17)
3.4. convective evaporation pipe bundles

According to the assumption (4) of this paper, there is only convective heat transfer between
convective evaporation pipe bundles and gas, the heat flux of outer surface of pipe bundles is:
qdlxy D dlxy (Tgxy  Tdlxy1 ) (18)
Where: D dlxy is flue gas convective heat transfer coefficient at the ralative position [ x, y] , Tdlxy1 is the outer
 Tao Hong et al. / Procedia Engineering 31 (2012) 834 – 843 839

surface temperature in the appropriate position of pipe wall.

flue gas convective heat transfer coefficient can be calculated by the following formula[10]:
0.65 0.33
D dlxy 0.2Cs Cz Rexy Prxy Ogas ddl (19)
Where: Rexy and Prxy respectively are the Reynolds numbers and Prater numbers of the flue gas at the
ralative position [ x, y] , O gas is thermal conductivity coefficient of the flue gas, d dl is the outer diameter of
convective evaporation pipes, C s and C z respectively are the bundles arrangement correction coefficients
and pipe row correction coefficients.
Because the fluid parameters of flue gas is quite difficult to calculate by theoretical formulas, so
paper adopted avergae convective heat transfer coefficient to calculate the distribution of heat flux.
m
D dlxy | D dl N dl (Wo  Wa ) (20)
Where: N dl is the flue gas heat transfer coefficient of convective evaporation bundles, m is a
coefficient[15], m 0.65 for lateral washed and regularly arranged bundles, m 0.6 for the lateral
washed and staggered bundles, m 0.8 for the lognitudinal washed bundles and large metal flat.
Dividing control bodies along the flow of fluid in convective evaporation pipe bundles, and took the
outlet parameters of control bodies as lumped parameters, the mass, momentum and energy conservation
equations of convective evaporation pipes were established as follows:
2
Vdlxy dX dlxy 2
[ dlxyWdlxy 2 Vdlxy dhdlxy 2 hdlxy 2  hdlxy1
X dlxy (Wdlxy1  Wdlxy 2 )ˈ ( pdlxy1  pdlxy 2 )X dlxy  'H dlxy gˈ X dlxy [(Wdlxy1  Wdlxy 2 )  Qdlxy ]
dt 2 dt 2
(21)
Where: Vdlxy , X dlxy , Wdlxy , hdlxy , pdlxy and Qdlxy respectively are the volume, specific volume, mass flow, specific
enthalpy, pressure and heat adsoration of steam-water mixture at the ralative position [ x, y] , subscript 1
and 2 respectively refer to inflow and outflow, [ dl , 'H dlxy are the pressure loss coefficient and relative
height of convective evaporation pipes at the ralative position [ x, y] .
The calculation of heat adsoration Qdlxy , dryness of outflow steam F dl 2 and inner surface temperature
distribution Tdlmxy 2 are similar to water wall:
43 3
Qdlxy Kdlxy Adlxy pdlxy 2 (Tdlmxy 2  Tdlxy ) ˈF dl 2 (hdl 2  hbw ) (hbs  hbw ) ˈmdlxyCmdl dTdlmxy 2 dt qdlxy Adlxy  Qdlxy (22)
According to assumption (5) and (6), we can take the inner surface temperature distribution and outer
surface heat flux distribution as the boundary conditions to calculate the temperature distribution of pipe
walls by one-dimensional thermal conductivity equation as follows:
2 2
Umdl Cmdl rdTdlmxy dt Omdl [dTdlmxy dt  r d Tdlmxy dt ] (23)
Where: K dlxy , Tdlmxy 2 , Adlxy and m dlxy respectively are the heat transfer coefficient, inner surface temperature,
heat transfer area and metal wall mass of convective evaporation pipes at the ralative position [ x, y] ,
Tdlxy is the saurated temperature of steam-water mixture under pressure p dlxy 2 , U mdl , Cmdl and Omdl are the
density, specific heat capacity and heat transfer coefficient of pipe walls.

3.5. decline pipes

According to assumption (1), decline pipes is a single-phase flow region of drum-boiler, the outlet
pressure of decline pipes can be calculated by the momentum conservation equation as follows:
2
pd pb  'H bw g X bw  [ dWd 2 (24)
Taking the heat dissipation effect into account, the outlet water enthalpy of decline pipes is:
840 Tao Hong et al. / Procedia Engineering 31 (2012) 834 – 843

hwd hwjq (1  k d ) (25)

Where: hwd and hwjq are the outlet water enthalpy of decline pipes and economizer, [ d , k d and 'H bw are the
pressure loss coefficient, thermal coefficient and elevation of decline pipes.

3.6. superheater

The working conditions of superheater is the worst in all heating surfaces of marine tube boiler. The
practice has proved that the inhomogeneity of steam flow and differences of flue gas temperature between
the bundles are main reason that cause the local overheat. By the analyzing the heat transfer and steam
flow distribution, the distributed parameter model of superheater module has been getten as follows.
In trems of marine natural recirculation drum-boiler, there is usually arranged many superheaters
along the flue gas flow direction. According to assumption (9), the flue gas temperature that enter the first
stage superheater was:
Tgrgxy 11 Tgxy  qdlxy 1 Adlxyi [(Wo  Wa )C pg ] (26)
The outlet gas temperature of every stage superheaters can be sequentially derived from formula(26):
Tgrgxyi 2 Tgrgxyi 1  q grxyi 1 Agrxyi [(Wo  Wa )C pg ] (27)
The convective heat transfer flux of superheater was:
0.6
qgrxyi N gri (Wo  Wa ) [(Tgrgxyi1  Tgrgxyi 2 ) 2  Tgrxyi1 ] (28)
Where: Tgrgxyi is the flue gas temperature distribution of superheater, subscript 1 and 2 respectively refer to
inflow and outflow, Agrxyi , q grxyi and Tgrxy 1 respectively are the heat transfer area, heat transfer flux heat and
outer surface temperature of superheater at the ralative position [ x, y] , N gri is the flue gas heat transfer
coefficient of superheater, C pg is the specific heat capacity at constant pressure of flue gas.
Influenced by the superheater heat load, superheater structures and superheater inlet static pressure
distribution, there is a deviation among the steam flow of each branch pipes. Assumed the outlet and inlet
pressure of each pranch pipes is the same, the steam flow distribution can be calculated by the momentum
conservation equation as follows:
Wgrij 2[( pb  pgr )X grij  'H grij g ] [ grij (29)
Where: Wgrij , [ grij , X grij and 'H grij respectively are steam flow, pressure loss coefficient, average specific
volume and relative height of branch pipe, p gr is the pressure of outlet superheated steam.
Dividing control bodies along the flow of fluid in superheater, and assumed the control body at
ralative position [ x, y] located on the j branch pipe in the i superheater, the energy conservation equation
of superheater was established as follows:
Vgrxy dhgrxy 2 dt X grij [Wgrij ( hgrxy 2  hgrxy 1 )  Qgrxy ] (30)
Formula (30) shows that the steam pressure in superheater approximately presents a linear variation
along the fluid flow direction:
p grxy p gr  ( pb  p gr ) 'H grxy 'H gr (31)
Where: Vgrxy , p grxy , hgrxy , 'H grxy and Qdlxy respectively are the volume, pressure, specific enthalpy, relative
height and heat adsoration of superheater at ralative position [ x, y] .
Heat adsoration of superheater and wall temperature distribution can be calculated as follows[17]:
dTgrmxy 2 dTgrmxy Omgr d dTgrmxy
˄r
0.8
Qgrxy K grxyWgrxy (Tgrmxy 2  Tgrxy )ˈ mgrxy Cmgr q grxyi Agrxy  Qgrxyˈ U mgr Cmgr ˅ (32)
dt dt r dr dr
Where: K grxy , Agrxy , m grxy , Tgrmxy 2 and Tgrxy respectively are the heat transfer coefficient, heat transfer area,
metal wall mass, inner surface temperature and superheated steam temperature, U mgr , Cmgr and Tgrxy are the
 Tao Hong et al. / Procedia Engineering 31 (2012) 834 – 843 841

density, specific heat capacity and heat transfer coefficient of pipe walls.

4. Experiment and result analysis

4.1. object description

Take a certain type marine natural ecirculation boiler as an instance, and adopted the modle in this
paper to do the simulation research on its dynamic characteristics, the simulation program was writeen by
Visual Fortran 6.5. The main parameters of this type boiler is shown in Table 1.

Tab. 1 main structure and capability parameters of simulative object

Parameter names Parameter values Parameter names Parameter values

Superheated steam production 80 t h Boiler water mass 5.5 tons

Saturated steam production 5t h Drum inner dimensions Diameter 1.1 mˈlength 3.0 m
Superheated steam pressure 6.0 MPa Water wall I 38×3 mm
Superheated steam temperature 450±10ć 9 rows in all: the 1st row I 38×3mm, the 2nd ~ 9th
Convective evaporation
Drum pressure 6.175 MPa rows I 30×3mm, 1st~3th rows staggered, 4st~9th
pipe bundles
Supply water temperature 104ć rows regularly arranged

Startup time 30min

Decline pipes 4 rows in all: I 38×3mm, staggered
Fuel consumption 6.95 t h
3
gross volume of drum 2.89 m
Superheater 10 rows in all: I 20×2.5mm, staggered
3
Excess air coefficient 5.9 m

4.2. Results and Analysis

According to the parameters in Section 4.1, we carried out simulation experiment in the support of
Minis simulation platform, the results were shown in Figure 2-5. The curves in figures are model output,
and the marked points are actual measured values.
Fig.2 is the changes of drum pressure, superheated steam temperature, water temperature and water
level over time during the process of boiler startup. Figure shows that, the initial statues were temperature
60ć, drum pressure 0.2 MPa , water level 220mm. Boiler water expansion after heated, water level began to
rise, and water temperature began to rise at the same time. About one minute later, the heating process was
completed, boiler water began to vaporize, drum pressure rise. About 5 minute later, drum pressure rose to
0.8 MPa , we increased the fuel injection and slightly opened the saturated and superheated globe valves to
warm up auxiliaries. Then the water level began to decline, while the drum pressure, water temperature and
superheated steam temperature continues to rise. The rising speed is much faster than before because fuel
flow increased. About 6 minute later, boiler began to supply saturated steam, then water level continues to
drop, while the rising trend of drum pressure and water temperature slowed down. About 14 minute later,
drum pressure rose to 2 MPa and superheated steam temperature reached 360f10ć, boiler began to supply
superheated steam and automatic adjustments of boiler worked at the same time. The whole startup process
is about 30 minutes, that match the real performance of this type boiler.
Fig.3 and Fig.4 respectively are the response curves of drum pressure, water level, superheated
steam temperature and water temperature when fuel regulative valve and when boiler load stepup 10%
under operating conditions. Figures show that the simulation results and measured values agreed well
 842 Tao Hong et al. / Procedia Engineering 31 (2012) 834 – 843

with each other, it proved that the model in this paper has a satisfactory accuracy, can correctly reflect
object¶s static and dynamic characteristics under normal working conditions.
One of the purposes of model-based simulation study is to conduct the fault simulation experiments
in ensuring the safety and economy of experiments, and then get the more complete fault characteristics
of objects by reproduced various fault symptoms and fault phenomenas, thus provides the experimental
basis for fault diagnosis and treatment. Paper took the boiler exhausting valve leakage as an example, the
fault simulation results is shown in Fig.5. Figure shows that the drum pressure, boiler water level and
temperature were all decreased after failure occurred. Recause the decline of drum pressure, the steam
flow througt the superheater will reduce, therefore the superheated steam temperature will rise, and then
slowly drop and finally tends to stable along with the redistribution of steam flow. With te boiler load re-
balancing, drum pressure and water temperature will stabilize at a new balance point. As for the water
level, because it doesn¶t have the self-balancing capabilityes, will continue to plummet. The
phenomenons that reflected in the curves of Fig.5 was consistent with the theoretical derivations, it
proved the correctness of this model from another side.
MPa mm MPa mm

pb lb pb lb

Drum pressure Drum pressure Water level

Water level

t ( min ) t ( min ) t (s) t (s)

ū ū ū ū

Tbw Ts Tbw Ts

superheated steam superheated steam

water temperature temperature water temperature temperature

t ( min ) t ( min ) t (s) t (s)

Fig. 2. curves of main parameters when boiler cold-startup Fig. 3. the dynamic response of boiler when fuel stepup 10%
under rating running situations
MPa mm MPa mm

Drum pressure Water level Drum pressure Water level

pb lb lb
pb

t (s) t (s) t (s) t (s)

ū ū ū ū

superheated steam
water temperature temperature water temperature
Tbw Ts Tbw Ts
superheated steam
temperature

t (s) t (s) t (s) t (s)

 Tao Hong et al. / Procedia Engineering 31 (2012) 834 – 843 843

Fig. 4. the dynamic response of boiler when load stepup 10% Fig. 5. the dynamic response of boiler when steam exhaust valve
under rating running situations leak under rating running situations

5. Conclusion

From the perspective of improving the model accuracy and real-time capabilities, paper adopted
modular modeling method to establish the mathematical model of marine natural recirculation drum-
boiler. Paper focused on the modular decomposition of the marine boiler plant based on the analysis of
the interactions between every components of object, then detailed expounding the construction of every
module models, and validated the model with specific object. The model present in this paper is a kind of
full-scale model for marine natural recirculation drum-boiler, that have a good simulation accuracy and
real-time capabilities, can play some reference value on the performance analysis, operation optimization,
fault diagnosis and control strategy development of various natural recirculation drum-boilers.

References

Yao S G, Marine Thermal System [M]. Beijing: Science Press, 2003.

Astrom K J, Bell R D. Drum-boiler dynamics. Automatic, 2000, 36(3): 363-378.
Adam E J, Marchetti J K. Dynamic simulation of large boilers with natural recirculation [J]. Computers and Chemical
Engineering, 1999, 23(3): 1031-1040.
Kim J H, Choi S. A model on water level dynamics in natural circulation drum-type boiler [J]. International communication in
heat and mass transfer, 2005, 32(6): 786-796.
Tian L, Zheng D L, Liu J P. A Simplified Non-linear Dynamic Model of 330MW Unit [J]. Proceedings of the CSEE, 2004,
24(8): 180-185.
Tian L, Zheng D L, Liu J P. A Simplified Non-Linear Dynamic Model of 500MW Unit [J]. Power Engineering, 2004, 24(4):
522-525.
Wang G J, Li H Y, Zhang C Y. A Model for Analysing the Operating Characteristics of a Natu-ral Circulation Boiling System
[J]. Journal of Engineering for Thermal Energy &Powe, 1998, 13(73): 57-60.
Flynn K J, Malley M J. A drum-boiler model for long term power system dynamic simulation[J]. IEEE Transaction on Power
System, 199, 14(1): 209-217.
Bell R D, Astrom K J. Dynamic models for drum boiler-turbine-alternator units: data logs and parameter estimation for a
160MW unit [R]. Report TFRT-3192, Lund institute of technology, 1987.
Feng J K, Shen Y T. Boiler Principle and Calculation [M]. Beijing: Science Press, 1992.
Chu Y T. Research on Distributed Parameter Modeling and Simulation for Utility Boiler Based on Combustion Monitoring[D].
Wuhan: Huazhong University of Science and Technology, 2007.
Yuan Y C, Liu J C, Chen Z H. Theoretical Analysis and Experimental Research on Temperature Distribution of FlueGAS and
Steam in Large Capacity Utility Boilers [J]. Proceedings of the CSEE, 2002, 22(12): 56-60.
Yin C G, Rosendahl L, Condra T J. Further study of the gas temperature deviation in large-scale tangentially coal-fired boilers
[J]. Fuel, 2003, 82(9): 1127-1131.
Chen G. The Modular Modeling and Simulation Research for Marine Thermal System [D]. ℺∝: Huazhong University of
Science and Technology, 1999.
Reid R C. Prausnitz J M, Poling B F. The properties of gases and liquids [M]. New York: Mcgraw-hill, 1987.
Yang D, Chen T K, Li H X, et al. Nonlinear Mathematical Model for Flow Rate Dstribution in The Parallar Tubes of Boiler
Superheater/Rehheater and The Calculation Method of Tube Surface Temperature [J]. Proceedings of the CSEE, 2001, 21(5): 12-15.
Yang B, Li Z. Dominant Factor Modelling Method for The Thermal Syatem of Power Station [J]. Proceedings of the CSEE,
2005, 25(24): 96-102ˊ