This article was downloaded by: [Narodna Biblioteka Srbije], [Vladimir Stevanovic]

On: 31 October 2014, At: 03:36

Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK

Heat Transfer Engineering

Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/uhte20

Prediction and Control of Steam Accumulation

a a b a
Vladimir D. Stevanovic , Milan M. Petrovic , Sanja Milivojevic & Blazenka Maslovaric
a
Faculty of Mechanical Engineering, University of Belgrade, Belgrade, Serbia
b
Innovation Center of the Faculty of Mechanical Engineering, University of Belgrade,
Belgrade, Serbia
Accepted author version posted online: 26 Jun 2014.Published online: 24 Oct 2015.

To cite this article: Vladimir D. Stevanovic, Milan M. Petrovic, Sanja Milivojevic & Blazenka Maslovaric (2015) Prediction and
Control of Steam Accumulation, Heat Transfer Engineering, 36:5, 498-510, DOI: 10.1080/01457632.2014.935226

To link to this article: http://dx.doi.org/10.1080/01457632.2014.935226

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of
the Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions
 Heat Transfer Engineering, 36:498–510, 2015
Copyright 
C Taylor and Francis Group, LLC
ISSN: 0145-7632 print / 1521-0537 online
DOI: 10.1080/01457632.2014.935226

Prediction and Control of Steam

Accumulation

VLADIMIR D. STEVANOVIC,1 MILAN M. PETROVIC,1 SANJA

Downloaded by [Narodna Biblioteka Srbije], [Vladimir Stevanovic] at 03:36 31 October 2014

MILIVOJEVIC,2 and BLAZENKA MASLOVARIC1

1
Faculty of Mechanical Engineering, University of Belgrade, Belgrade, Serbia
2
Innovation Center of the Faculty of Mechanical Engineering, University of Belgrade, Belgrade, Serbia

Steam accumulators are applied as buffers between steam generators and consumers in cases of different steam production
and consumption rates. The application of the steam accumulator saves energy, reduces pressure fluctuations, and prevents
aging of tubes and pressurized vessels in steam generators. In this paper, modes of the steam accumulator operation are
analyzed and the general design of the steam accumulator control system is defined. Equilibrium and nonequilibrium
thermodynamic models of the steam accumulator are presented with the aim of predicting the steam accumulator capacity
and as support to the design of the control system. The equilibrium model is based on the mass and energy balance equations
of the total water and steam content in the accumulator, while the nonequilibrium model is based on the mass and energy
balance equations for each phase and closure laws of nonequilibrium evaporation and condensation rates. The steam
accumulator pressure transients are simulated for constant steam charging and discharging flow rates, and the influence of
the nonequilibrium condensation and evaporation rates on the steam accumulator capacity is shown. It is concluded that
the commonly used equilibrium thermodynamic approach to the steam accumulator design does not provide accurate results
in cases of rapid charging and discharging transients; therefore, there is a need for the application of the nonequilibrium
approach.

INTRODUCTION cannot be followed by decreases of the steam generator’s load,

steam should be discharged to the atmosphere or to low-pressure
Steam accumulators are applied as buffers between steam condensers, which is the pure loss of energy. The steam genera-
generators and consumers in cases of different steam production tors cannot follow the rapid steam consumption changes because
and consumption rates or staggered time periods of steam gener- of the inertia of the fuel combustion in the furnace and the inertia
ation and consumption. The steam accumulator is a pressurized of heat transfer from the furnace to the boiling water. Also, safety
vessel filled with water and steam (Figure 1). The accumulator is requirements imposed on pressure and temperature changes in
being charged in periods of lower steam consumption or surplus the steam generator’s tubes and the steam drum do not allow
of steam production, where the pressure in the accumulator in- rapid load changes. However, dynamic changes of the steam
creases and the steam condenses. In periods of increased steam consumption do cause pressure changes in steam pipelines, the
consumption that is not covered with the production rate, steam steam generator’s tube system, and pressurized vessels, and they
is discharged from the accumulator to the consumer, the pressure lead to dynamic thermomechanical loads and endanger integrity
in the accumulator decreases, and water in the accumulator adia- of energy equipment. These negative effects can be reduced sig-
batically evaporates. Therefore, the steam accumulator operates nificantly by the application of steam accumulators.
with sliding pressure. Steam accumulator applications enable Steam accumulators are applied in the textile industry for dye
considerable energy savings. Without the steam accumulator the works [1], for glassmaking, rubber vulcanization, and tobacco
steam generation rate must equal the maximum steam consump- processing [2], in metal manufacturing [3], and in thermal power
tion rate. During shorter periods of lower consumption, which plants for electricity production during peak loads [4]. Steam
generated in cooling systems by water vaporization in some
technological processes has variable pressure and temperature.
Address correspondence to Professor Vladimir D. Stevanovic, Faculty of
Mechanical Engineering, University of Belgrade, Kraljice Marije 16, 11120
In these cases the steam accumulator can be applied to provide
Belgrade, Serbia. E-mail: [email protected] stable steam supply to the steam turbine plant, which is built

498
 V. D. STEVANOVIC ET AL. 499

rameters, which might have different temperatures although the

two phases are in contact, and the evaporation and condensation
rates are finite. Application of the nonequilibrium model leads
to practical implications that are important for the design of the
steam accumulator and its control system.
In this paper possible modes of the steam accumulator oper-
ation are analyzed and a general design of the steam accumu-
lator control system is presented. The measurement and control
circuits are defined for the regulation of steam accumulator
charging and discharging, the steam flow rate in the accumula-
tor by-pass steam pipeline with pressure reduction station, the
regulation of steam temperatures at the steam accumulation in-
Downloaded by [Narodna Biblioteka Srbije], [Vladimir Stevanovic] at 03:36 31 October 2014

let, and the regulation of water level in the accumulator vessel.

In order to overcome deficiencies of the present steam gener-
Figure 1 Steam accumulator layout.
ator design methods based on the equilibrium thermodynamic
models, the research performed included development of a new
to recover the energy of steam from the technological process. thermodynamic model of the steam accumulator, based on ther-
Such solutions are applicable in ferro-alloys, steel and iron, and mal nonequilibrium between the liquid and vapor phase. This
glass industries [5]. Fast reaction times and high discharge rates nonequilibrium model provides more accurate predictions of
make steam accumulators a promising option for the compen- the steam accumulator pressure, temperature, and water level
sation of fast transients in insolation for solar thermal power during accumulator charging and discharging transients than
plants [6, 7]. The applications of the steam accumulator in two the equilibrium model. In addition, this new model can be eas-
solar thermal power tower plants under construction in China ily programmed for computer simulations and it can simulate
are presented in references [8] and [9]. steam accumulator transients of any complexity. The nonequi-
Understanding steam accumulator transients, characterized librium model of the steam accumulator is based on the mass and
with sliding pressure and phase transitions, and a design of the energy balances for each phase (liquid water and steam). The
proper control system are crucial for reliable and efficient accu- evaporation and condensation rates are calculated with semiem-
mulator operation. The pressure control in the steam network of pirical correlations, which take into account the dependence of
an industrial power plant with the application of the steam ac- the phase transition rates on the water mass subcooling or su-
cumulator was tested in the simulator environment in reference perheating in regard to the saturation state and on the water
[10] and advantages of the model predictive control method inventory in the steam accumulator vessel. The heat transfer
were demonstrated. It was stated that the steam accumulator is rate from the steam towards the liquid water is calculated when
controlled by the charging and discharging logic [10], but no steam is at a higher temperature than water. The model balance
details were presented. equations form a set of differential equations that are solved
The steam accumulator operation with the sliding pressure numerically for specified initial conditions. The model is vali-
leads to pronounced thermal nonequilibrium between water and dated by comparing numerical results with measured data from
steam in the accumulator. The application of equilibrium ther- the accumulator charging test. The results presented in this paper
modynamic models to the design of the steam accumulator and show that the developed model is a tool for the design of steam
its automatic control system might lead to malfunctions, such accumulator vessel volume and the control system that should
as an inability to reach required capacity or difficulties in the govern the accumulator charging and discharging transients.
control of accumulator operation. Despite these drawbacks, the
equilibrium thermodynamic models are commonly applied for
steam accumulator design [2–4, 6, 7]. The published analytical STEAM ACCUMULATOR APPLICATION
models of steam accumulator are based on thermal equilibrium
between the liquid and steam phase, although a substantial de- A supply of high and low pressure steam to consumers is
viation from thermal equilibrium can exist in the accumulator observed, as presented in Figure 2. The generated high-pressure
during rapid charging or discharging transients. The equilib- steam flows from the steam boilers (SBs) to the high-pressure
rium model assumes that the liquid and vapor phase in contact header (HPH). A part of the steam flow rate is delivered from
have the same temperature, equal to the saturation temperature the HPH to the consumers that require steam at the high pressure
that is uniquely determined by the pressure. It also implies that (HPCs), while the rest is transferred to the low-pressure header
any thermodynamic change of the liquid and vapor state is fol- (LPH) by the pressure reduction station (PRS). From the LPH
lowed by evaporation or condensation with an infinite rate that the steam is delivered to the consumers at low pressure (LPCs).
promptly leads to a new equilibrium thermodynamic state. On The consumers at both high and low pressure may have variable
the contrary, the model based on the thermal nonequilibrium consumption of steam. On the other hand, the generation rates
calculates separately the liquid and vapor thermodynamic pa- of the steam boilers are not able to follow dynamic steam
heat transfer engineering vol. 36 no. 5 2015
 500 V. D. STEVANOVIC ET AL.

which enables the charging of the steam accumulator and the

pressure in the accumulator vessel increases. The pressure in-
crease in the accumulator leads to the steam condensation and
its compression in a volume above the liquid surface. Both pro-
cesses lead to steam accumulation in the vessel, but at pressures
not close to the critical pressure the effect of compression is
almost negligible compared to the accumulation by steam con-
densation. During the periods of higher steam consumption rates
than the steam generation rate, the steam pressure in the head-
ers that distribute the steam to the consumers decreases. The
pressure decrease in the HPH should actuate the closing of the
valve in the PRS, which leads to the consequent decrease of
Downloaded by [Narodna Biblioteka Srbije], [Vladimir Stevanovic] at 03:36 31 October 2014

the pressure in the LPH and actuates the accumulated steam

discharge from the steam accumulator to the LPH and further
to the low-pressure consumers (LPCs in Figure 2). The increase
of steam consumption by LPCs and consequent pressure de-
crease in the LPH should also actuate the steam accumulator
discharge. During the steam accumulator discharging the pres-
sure in the accumulator vessel decreases and steam is generated
by water evaporation. Thus, the steam delivery to the consumer
increases and the pressure decrease in the pipeline system of the
steam boilers is reduced or almost eliminated. In addition, the
steam accumulator enables significant saving of steam boilers’
Figure 2 Steam accumulator connection to the high-pressure and low-pressure fuel consumption. The steam accumulation enables the boilers
headers for steam supply to consumers together with control circuits. SBs, steam to operate at a lower load than should be necessary according
boilers, HPH, high-pressure header, LPH, low-pressure header, HPCs, high- to the maximum steam consumption, and eliminates the steam
pressure consumers, LPCs, low-pressure consumers, PRS, pressure reduction
station, PRS-AI, pressure reduction station at accumulator inlet, PRS-AO, pres-
discharge to the relief tank or to the atmosphere in periods of
sure reduction station at accumulator outlet. Symbols for measuring and control lower consumption.
circles, the first letter: F, flow rate, G, position, L, level, N, control circuit, P,
pressure, T, temperature; other letters: A, alarm, signal on boundary condition,
C, self-regulating continuous control, I, digital display, P, control measurement,
R, record. CONTROL OF THE STEAM ACCUMULATOR
OPERATION
consumption rates due to the inertia of the fuel combustion in the
furnace, as well as due to safety reasons, since dynamic changes Possible Modes of Operation
of steam generation lead to dynamic loads exerted on the walls
of pipelines and pressure vessels used in power installations. Steam consumption might be stochastic, but here the periodic
The steam generation and steam consumption might be even variation of consumption is analyzed, which is a characteristic
staggered in time; that is, during steam production there might of the well controlled technological process. Possible modes of
be no consumption and vice versa, as is the case with steam the steam accumulator operation are shown in Figure 3 for cases
generation in a solar thermal plant under variable insolation. when the variable steam consumption is delivered from the low
In order to suppress these dynamic disbalance between steam pressure header LPH and the steam consumption from the HPH
consumption and steam generation rates, the steam accumulator (Figure 2) is constant. The diagram in the top left corner shows
is connected in parallel to the pressure reduction station PRS, the assumed periodic steam consumption of the low-pressure
as shown in Figure 2. The steam accumulator is supplied with consumers (LPCs) (Figure 2). The vertically aligned diagrams
steam from the high-pressure header (HPH) in periods of higher in the left, middle, and right columns show the possible modes
steam production than consumption of both high-pressure and of steam accumulator operation, that is, the different coupling of
low-pressure consumers, and the steam is delivered from the the steam accumulator charging and discharging flow rates and
accumulator to the low-pressure header (LPH) in periods of the steam flow rate through the pressure reduction station PRS
higher steam consumption than generation rates. in the parallel connection to the steam accumulator (Figure 2).
The accumulator is filled with water and steam, where both A description of the possible steam accumulator operational
phases are saturated in steady-state conditions. During the pe- modes follows:
riods of lower steam consumption, the steam pressure in the
HPH increases. This pressure increase should actuate the open- a. Diagrams in the left column in Figure 3 show the mode of
ing of the control valve at the accumulator inlet steam pipeline, operation with the accumulator charging in the periods of
heat transfer engineering vol. 36 no. 5 2015
 V. D. STEVANOVIC ET AL. 501

the steam accumulator charging and discharging rates, while

the flow through the pressure reduction station is constant and
adjusted to the value that is lower or equal to the minimum steam
consumption rate. The operational mode described under (c)
requires only the control of the steam charging and discharging
rates, while the pressure reduction station is closed. But in this
case all steam to the LPCs flows through the accumulator and the
steam velocities in the steam accumulator inlet and outlet steam
pipelines might be too high. Also, the whole steam flow through
the accumulator might lead to high water droplets entrainment
from the water surface and their carryover with steam out of the
accumulator.
Downloaded by [Narodna Biblioteka Srbije], [Vladimir Stevanovic] at 03:36 31 October 2014

A possible mode of the steam accumulator operation is shown

in Figure 4 for the case when the variable steam consumption is
delivered from the high pressure header HPH (Figure 2), while

Figure 3 Modes of steam accumulator charging and discharging in case of

variable steam consumption from the low-pressure header (top left diagram)
and constant consumption from the hig-pressure header: Steam charging and
discharging rates and steam bypass flow through the pressure reduction station
are variable (left column), the steam charging rate and the steam by-pass flow
through the pressure reduction station are constant, and the steam discharging
rate is variable (middle column); the whole steam consumption flows through
the accumulator (right column).

lower steam consumption and the accumulator discharging

in the periods of higher consumption. In addition, the flow
rate through the pressure reduction station in parallel con-
nection to the accumulator is also reduced in the periods of
lower consumption, while this flow rate is constant in peri-
ods of higher steam consumption and the necessary excess
steam flow is supplied from the accumulator. This mode of
operation provides the minimal accumulator charging and
discharging flow rates.
b. In the mode of operation presented in diagrams in the middle
column of Figure 3 the flow rate through the reduction station
in parallel to the accumulator is constant and equal to the
minimal steam consumption rate. The steam accumulator
charging rate is also constant. Only the steam accumulator
discharging rate changes according to the consumer’s needs.
c. Diagrams in the right column of Figure 3 show the mode
in which the whole steam consumption flows through the
accumulator, while there is no flow in the pressure reduction
station in parallel to the accumulator. The steam accumulator
charging rate is constant, while the steam discharging rate
from the accumulator is variable according to the consumer
needs.

The operational mode just described under (a) is provided

by the control of the accumulator charging and discharging
steam rates, as is described in the following section, together Figure 4 Steam accumulator charging and discharging in the case of variable
with the control of the steam flow rate through the pressure steam consumption from the high-pressure header and constant consumption
reduction station connected in parallel to the accumulator. The from the low-pressure header: Steam charging and discharging rates and steam
mode described under (b) is also achieved by the control of bypass flow through the pressure reduction station are variable.

heat transfer engineering vol. 36 no. 5 2015

 502 V. D. STEVANOVIC ET AL.

the steam consumption from the LPH (Figure 2) is assumed con- reduction station PRS in order to keep the pressure in the low-
stant. Diagrams show the accumulator charging in the periods pressure header (LPH) within the prescribed pressure interval.
of lower steam consumption and the accumulator discharging If the pressure in the low-pressure header (LPH) decreases
in the periods of higher consumption. In addition, the flow rate and the control valve in the PRS-AO is fully opened, the control
through the pressure reduction station in parallel connection circuit 1NIRC sends a signal for the opening of the control valve
to the accumulator is variable, where the minimal flow rates in the pressure reduction station PRS.
through the pressure reduction station coincide with instants of
steam accumulator maximum charging and discharging rates.
In the general case, consumptions from both the high- Control of the Steam Accumulator Charging
pressure header and the low-pressure header can be variable
and stochastic (i.e., not periodic as it is shown in Figures 3 and The control circuit 2NIRC (Figure 2) adjusts the control valve
4). For such cases, the only possible mode of operation that in the pressure reduction station PRS-AI, in the accumulator
Downloaded by [Narodna Biblioteka Srbije], [Vladimir Stevanovic] at 03:36 31 October 2014

provides utilization of the total accumulation capacity of the inlet steam pipeline, on the basis of the measured pressure in
steam accumulator is the variable steam accumulator charging the high-pressure header (HPH) in front of the accumulator and
and discharging rates, as well as variable flow in the steam ac- the measured pressure in the accumulator vessel.
cumulator bypass line with the pressure reduction station (PRS If the pressure in the high-pressure header (HPH) increases
in Figure 2). The control for such a general case of operation is above prescribed setpoint and the pressure in the accumulator
presented in the next subsection. is lower than the pressure in the HPH, the circuit 2NIRC sends
the signal for the opening of the valve in the PRS-AI.
If the pressure in the steam accumulator is equal or higher
Control System for the Steam Accumulator Operation than the pressure in the HPH or the pressure in the HPH is lower
than prescribed lower setpoint, the control valve in the PRS-AI
Measurement and control circuits for the control of the steam is being closed or it remains closed.
accumulator operation are shown in Figure 2. The charging of
the steam accumulator is controlled with the throttling valve in
the pressure reduction station PRS-AI, which is being positioned Control of the Steam Accumulator Discharging
with the control circuit 2NIRC. The accumulator discharging is
controlled with the control throttling valve in the pressure reduc- The control circuit 3NIRC (Figure 2) adjusts the control
tion station PRS-AO, which is being positioned with the control valve in the pressure reduction station PRS-AO in the accumu-
circuit 3NIRC. The position of the control throttling valve in the lator outlet steam pipeline based on the measured pressure in
pressure reduction station PRS, which is in parallel connection the header LPH after the accumulator and the pressure in the
to the steam accumulator, is determined by the control circuit accumulator vessel.
1NIRC. A description of these control circuits follows. If the pressure in the low-pressure header (LPH) decreases
below minimum setpoint and the pressure in the accumulator is
higher than the pressure in the LPH, the control circuit 3NIRC
Control of the Flow Through the Pressure Reduction Station sends a signal for the opening of the control valve in the PRS-
(PRS) AO.
If the pressure in the LPH increases above the maximum
The control circuit 1NIRC (Figure 2) adjusts the valve in setpoint, the circuit 3NIRC sends a signal for closing of the
the pressure reduction station (PRS) based on the measured valve in PRS-AO.
pressure in the high-pressure header (HPH) in front of the accu- If the pressure in the accumulator is lower than the pressure
mulator and the pressure in the low-pressure header (LPH) after in the LPH, the control valve in the PRS-AO does not open.
the accumulator toward the consumer. In addition, the signal
of the fully closed or opened position of the control valve in
the pressure reduction station PRS-AO after the accumulator is Control of the Steam Temperature at the Steam Accumulator
used. Inlet
If the pressure in the high-pressure header (HPH) decreases
below the specified setpoint value, the circuit 1NIRC sends The control circuit 2TICA (Figure 2) adjusts the valve for
a signal for the closing of the control valve in the pressure the control of the cold feedwater flow in the attemperator (valve
reduction station PRS in order to keep the pressure in the HPH 3 in Figure 2) based on the measured steam temperature after
within the prescribed interval. This signal has a priority over the attemperator, the measured steam accumulator pressure, and
other signals sent to the circuit 1NIRC. the signal on the position of the control valve in the PRS-AI.
If the pressure in the header LPH increases above the setpoint The measured temperature after the attemperator is compared
value and the valve in PRS-AO is fully closed, the circuit 1NIRC with the saturation temperature determined by the measured
sends a signal for the closing of the control valve in the pressure accumulator pressure. If the measured temperature is higher than
heat transfer engineering vol. 36 no. 5 2015
 V. D. STEVANOVIC ET AL. 503

the saturation temperature, the control valve for the feedwater are calculated as ṁ P T 1 = ṁ c − ṁ e and ṁ P T 2 = ṁ e − ṁ c . The
flow control is being opened; otherwise it is being closed. volume balance is also applied:
If the control valve in the PRS-AI in front of the accumulator
is fully closed, the circuit 2TICA sends a signal for the complete V1 + V2 = V (5)
closure of the valve 3 in order to prevent feedwater injection in The evaporation rate is calculated as
the steam line when there is no steam flow.  
The control of the steam temperature in the PRS by the con- ρ1 V1 h 1 − h 
ṁ e = , for h 1 > h  (6)
trol circuit 1TICA is similar to the already-described control by τe r
2TICA. The control of the water level in the steam accumulator and ṁ e = 0 if water is saturated or subcooled, that is, h 1 ≤ h  .
vessel is performed by the circuit LICA (Figure 2). Similarly, the condensation rate is predicted as
 
ρ1 V1 h  − h 1
Downloaded by [Narodna Biblioteka Srbije], [Vladimir Stevanovic] at 03:36 31 October 2014

ṁ c = (7)
MODELING OF THE STEAM ACCUMULATOR τc r
OPERATION
and ṁ c = 0 if water is saturated or superheated, that is, h 1 ≥ h  .
A closure law for the calculation of evaporation and con-
Simulation of the steam accumulator charging and discharg-
densation relaxation times τe and τc is derived starting from
ing tests is performed with both nonequilibrium and equilibrium
the statement that the mass transfer rate of phase transition is
models.
determined by the product of the interfacial area concentration
ai and the phase transition mass flux ji ,
Nonequilibrium Model i = ai ji (8)

The steam accumulator model is based on the following mass where index i equals e or c. From Eqs. (6) and (7) it follows that
and energy balance equations. the phase transition rate is
 
Liquid mass balance: ρ1 h 1 − h  
i = (9)
d M1 τi r
= ṁ 1B + ṁ P T 1 (1)
dt The expression for the relaxation time is derived from the
Steam mass balance: equality of the right-hand sides of Eqs. (8) and (9):
d M2  
= ṁ 2B + ṁ P T 2 (2) ρ1 h 1 − h  
dt τi = (10)
ji ai r
Liquid energy balance:
The phase transition mass flux is calculated by applying the
d H1 dp
= (ṁh)1B + ṁ P T 1 h  + Q̇ 21 + V1 (3) coefficient of heat transfer k1i between the steam–water interface
dt dt and the surrounding water mass,
Steam energy balance: k1i |T1 − Tsat |
ji = (11)
d H2 dp r
= (ṁh)2B + ṁ P T 2 h  − Q̇ 21 + V2 (4)
dt dt where the numerator in Eq. (11) represents the heat flux at the
In the preceding equations the liquid water inlet and outlet interface due to phase transition, T1 is the water temperature,
mass flow rates are calculated as ṁ 1B = ṁ 1,in − ṁ 1,out , and and Tsat is the saturation temperature determined by the pressure
consequently for the steam inlet and outlet mass flow rates are in the accumulator. Evaporation and condensation take place at
ṁ 2B = ṁ 2,in − ṁ 2,out . These mass flow rates are prescribed time the surface of the water pool and at the interface of bubbles
functions or depend on the difference between a calculated pres- within the water volume. The bubbles are formed during accu-
sure in the steam accumulator and a prescribed pressure in the mulator discharging due to adiabatic nucleation in superheated
volume connected with the steam accumulator (upstream and water. During accumulator charging, the bubbles are formed
downstream steam header, water storage reservoir, etc.). The in- due to steam injection in the water volume through perforated
let and outlet energy flow rates at the steam accumulator bound- tubes. Although these mechanisms of bubble formation are dif-
aries are calculated as (ṁh)1B = ṁ 1,in h 1,in − ṁ 1,out h 1,out for ferent, it is assumed that at a certain distance from the location
liquid water and (ṁh)2B = ṁ 2,in h 2,in − ṁ 2,out h 2,out for steam. of bubbles formation in a large water volume, their number will
The water and steam inlet enthalpies are specified according to be the same for the same steam volume fraction. Therefore, the
the operating conditions, while the outlet enthalpies are equal following derivation of the phase transition model holds both
to water and steam enthalpies in the accumulator vessel that for the evaporation and condensation. An approximate analysis
are calculated by the model balance equations. The liquid and of the amounts of phase transitions on the water pool surface
steam mass changes due to evaporation and condensation rates and at the bubbles interface shows that the former is negligible.
heat transfer engineering vol. 36 no. 5 2015
 504 V. D. STEVANOVIC ET AL.

That is, the results presented in the Results section and discus- Interfacial area concentration is calculated as the product of
sions with Figure 10 (shown later) show that the total steam the concentration of bubbles n and the area of the spherical
evaporation or condensation rate is in the range between 1 kg/s bubble surface:
and 10 kg/s in a horizontal cylindrical accumulator of length

Db 2
L = 11.9 m and diameter D = 2.9 m. The corresponding heat ai = n 4π (17)
transfer rate for the phase transition between the water volume 2
and all steam–water interface surfaces Q̇ P H is in the range be-
The bubble diameter is calculated on the basis of the critical
tween 1.7 and 17 MJ/s (for the accumulator pressure of 4 MPa
Weber number Wecr for bubbly flow [11]:
and corresponding latent heat of evaporation/condensation 1.7
MJ/kg). First it is assumed that the phase transition takes place W ecr σ
Db = (18)
only on the surface of the water pool, that is, the presence of ρ1 (u 2 − u 1 )2
bubbles in the water volume is neglected. For the stagnant water
Downloaded by [Narodna Biblioteka Srbije], [Vladimir Stevanovic] at 03:36 31 October 2014

where Wecr = 1.24. The concentration of bubbles is predicted

volume in the accumulator, for instance, at 4 MPa steam pres-
in the same analytical form as developed for the adiabatic evap-
sure, the temperature gradient on the laminar water side of the
oration in nozzle flow [11,13],
water pool surface is calculated from the equality of the heat
flux calculated with the well-known Fourier’s law and the heat n = 106.8−0.15 log Gi (19)
flux of phase transition as follows:
  where the parameter Gi is defined as
 dT1  Q̇ P H /(L D)
 
 dy  = λ1
(12)
Gi =
16πσ3
 2 (20)
3k B r 2 T1 h 1 h  − 1 ρ2
where y is the normal distance from the interface surface; λ1
is the thermal conductivity of subcooled or superheated water The first empirical constant in the exponent of Eq. (19) has
and its value is slightly greater or lower than 0.62 W/(m-K) at 4 a value of 12.5 in the original correlation [13], while here the
MPa. The numerator on the right-hand side represents the heat lower value 6.8 is applied in order to obtain good agreement
flux from (in the case of condensation) or toward (in the case of the model pressure transient predictions with measured data.
of evaporation) the interface surface due to phase transition. This lower value implies a much lower concentration of bubbles
According to Eq. (12), the temperature gradient is in the range under adiabatic evaporation in the accumulator stagnant water
  than in the high-velocity water nozzle flow (the reduction of the
 dT1  (1.7 × 106 to 17 × 106 )/(11.9 × 2.9) K
  bubbles concentration is approximately of the order 105), which
 dy  = 0.62
= 79 to 790
mm is physically plausible behavior.
(13)
Besides the presented analytical approach for the calculation
Assuming that the water surface temperature equals the tem-
of the evaporation and condensation relaxation times τe and τc ,
perature of saturation determined by the steam pressure in the
the mean values of these parameters are determined by com-
accumulator, the result from Eq. (13) implies water subcool-
parison of calculated pressure transients with measured values
ing or superheating in the range from 79 K to 790 K, which is
in benchmark tests. Comparison of analytically and empirically
obviously impossible. Therefore, it is concluded that the contri-
obtained values of the relaxation time is presented and discussed
bution of the water pool surface to the phase transition rate is
in the Results and Discussion sections.
negligible, which means that the phase transition takes place on
The heat transfer rate from the superheated steam to the liquid
the steam–water interfaces of steam bubbles. According to this
and vice versa in Eqs. (3) and (4) is calculated as
conclusion, the heat transfer coefficient and the interfacial area
concentration in Eqs. (10) and (11) are calculated for the condi- Q̇ 21 = k2i ai (T2 − T1 ) V1 (21)
tions of steam bubbly flow in stagnant water. The heat transfer
coefficient between the bubble interface and the surrounding where the steam–water interface area concentration ai is cal-
continuous water is calculated as [11] culated by Eq. (17) and the convective heat transfer coefficient
for the heat transfer between the bubble interface and the steam
λ1  1/2 1/3

in the bubble k2i is approximately determined by the theoretical
k1i = 2 + 0.74Reb Pr1 (14)
Db solution for the transient conduction in the sphere [11],
where the Reynolds number for bubble flow is calculated as λ2
k2i = 16 (22)
Db ρ1 (u 2 − u 1 ) Db
Reb = (15)
μ1 The system of balance equations (1–5) is transformed in a
and the relative velocity between the steam bubble and the stag- set of first–order differential equations as follows. The steam
nant water mass (u1 = 0) is calculated as [12] and liquid water volumes in the volume balance (5) are written
 1 as products of the corresponding mass and specific volume.
σg (ρ1 − ρ2 ) 4 Further, the specific volumes of liquid water and steam are
u 2 = 1.4 (16)
ρ21 written as functions of pressure and corresponding enthalpies,
heat transfer engineering vol. 36 no. 5 2015
 V. D. STEVANOVIC ET AL. 505

that is, v1 = v1 ( p, h 1 ) and v2 = v2 ( p, h 2 ). The volume balance volume is v = V M, where V is the total internal volume of the
is differentiated by time and the following equation is obtained: steam accumulator. Differentiation of the total enthalpy leads to
   
d M1 d M2 ∂v1  dp ∂v1  dh 1 dH
=M
dh
+h
dM
(29)
v1 + v2 + M1 +
dt dt ∂ p h dt ∂h  p dt dt dt dt
The specific enthalpy derivative is
   
 
∂v2  dp ∂v2  dh 2 dh dh dr dp dx
+M2 + =0 (23) = +x +r (30)
∂ p h dt ∂h  p dt dt dp dp dt dt

In energy balances (3) and (4) the total enthalpies H1 and H2 and the quality derivative is
are replaced with corresponding products of masses and specific dx 1 v dM
Downloaded by [Narodna Biblioteka Srbije], [Vladimir Stevanovic] at 03:36 31 October 2014

enthalpies, and after differentiation this is obtained: =−  

dt M (v − v ) dt
 
dh 1 1  dp d M1   
= (ṁh)1B + ṁ P T 1 h + Q̇ 21 + M1 v1 − h1
dt M1 dt dt 1 dv  v − v  d v  − v  dp
− +  (31)
(v  − v  ) dp (v − v ) 2 d p dt
(24)
Due to the thermodynamic equilibrium condition, the specific
volumes of the saturated water v’ and steam v”, the saturated
  water enthalpy h’, and the latent heat of phase transition r are
dh 2 1  dp d M2
= (ṁh)2B + ṁ P T 2 h − Q̇ 21 + M2 v2 − h2 only functions of pressure. By the introduction of Eqs. (29),
dt M2 dt dt
(30), and (31) into (28) the differential equation for the pressure
calculation is obtained in the following form:
(25)
Substitution of Eqs. (24) and (25) in (23) leads to
           
∂v1  ∂v2  1 ∂v2 
dp h 1 ∂h p
− v1
d M1
dt
+ h 2 ∂h p
− v 2
d M2
dt
− ∂v
∂h p
(ṁh)1B + ṁ P T 1 h  + Q̇ 21 − ∂h p
(ṁh)2B + ṁ P T 2 h  − Q̇ 21
=         (26)
dt ∂v1  ∂v1  ∂v2  ∂v2 
∂p  + v1 ∂h p M 1 + ∂p  + v 2 ∂h p M2
h h

Equations (1), (2), (24), (25), and (26) provide a set of five
first-order ordinary differential equations for the prediction of  rV 
water and steam masses, enthalpies, and steam accumulator dp (ṁh)1B (ṁh)2B + v M−v − h (ṁ 1B + ṁ 2B )
=    

M −v dr M −v
V V
pressure, respectively. These equations are solved numerically dt M dh + − r dv 
− r d(v  −v)
−V
 
v −v dp  
v −v dp
by using the Runge–Kutta method [14] for specified initial val- dp (v  −v  ) 2 dp
(32)
ues of dependant variables, that is, water and steam masses and
Equations (27) and (32) are solved numerically by using
enthalpies and initial steam accumulator pressure.
the Runge–Kutta method for specified initial values of the total
mass of water and steam and for the initial pressure in the steam
accumulator.
Equilibrium Model

The equilibrium model is based on the mass and energy

RESULTS AND DISCUSSIONS
balance for the whole steam accumulator internal volume.
Mass balance:
The developed models are applied to the simulation and anal-
dM yses of charging and discharging transients in the steam accu-
= ṁ 1B + ṁ 2B (27)
dt mulator, which is a horizontal cylindrical vessel with the total
Energy balance: internal volume of 64 m3, outer length of 11.9 m, and outer
diameter of 2.9 m. The accumulator operates between 25 bar
dH dp and 55 bar. Steam is introduced in the accumulator through the
= (ṁh)1B + (ṁh)2B + V (28)
dt dt perforated tubes that are submerged in the water volume from
where M is the mass of fluid in the accumulator (both water and the steam headers located at the top of the accumulator vessel
steam) and H is the total enthalpy of the fluid in the accumulator, interior, as shown in Figure 5.
calculated as H = Mh. The specific enthalpy of the water and In all performed simulations it is assumed that in the initial

   as h = h + xr . The
steam in the accumulator is state the water and steam are in thermal equilibrium determined
 calculated

quality is predicted as x = v − v v − v and the specific by the initial pressure.

heat transfer engineering vol. 36 no. 5 2015

 506 V. D. STEVANOVIC ET AL.
Downloaded by [Narodna Biblioteka Srbije], [Vladimir Stevanovic] at 03:36 31 October 2014

Figure 7 Measured and calculated pressure in the steam accumulator during

the test of steam accumulator charging (NEQ, nonequilibrium model results,
EQ, equilibrium model results).
Figure 5 Steam accumulator interior.
pressure increase, which is the result of the infinite condensation
Nonequilibrium Model Validation rate as its inherent characteristics.
The value of the condensation relaxation time coefficient,
The developed nonequilibrium model is validated for a test of calculated with Eq. (10) is shown in Figure 8 and compared
steam accumulator charging [15]. The measured charging mass with the adopted constant value of τc . The model prediction
flow rate is shown in Figure 6. There are no other inlet or outlet with Eq. (10) shows higher values of τc in the initial period of
flows. The liquid water initial volume is 86% of the total accu- transient and lower values during the remaining period. Since
mulator volume. The numerical predictions of pressure increase the constant empirical value of the relaxation time τc provides
with the nonequilibrium and equilibrium models are compared better results, this value is used in further calculations in this
with the measured values in Figure 7. The nonequilibrium model paper, and the equality of τc and τe is applied.
simulation is performed with the condensation relaxation time
calculated with Eq. (10) and with the constant empirical value
τc = 85 s. A rapid pressure increase up to 36 bar is measured Steam Accumulator Charging and Discharging Tests
at the beginning of the charging process until 25 s due to the
rise in the steam inlet flow rate as shown in Figure 6. Later on, Figure 9 shows calculated pressure changes in the steam ac-
a slower pressure increase is shown in Figure 7. The best agree- cumulator during the charging and discharging conditions with
ment of the calculated pressure transient with the measured data constant inlet and outlet steam mass flow rates 5 t/h, 10 t/h,
is obtained by the application of the nonequilibrium model and and 30 t/h. The initial pressure for the charging numerical ex-
the constant empirical value of τc . The results obtained with the periments is 25 bar and the final pressure is 55 bar. The inlet
prediction of τc with Eq. (10) show a slightly faster pressure steam is saturated according to the instantaneous pressure in
increase after 300 s. The equilibrium model shows the smaller the steam accumulator vessel. The initial pressure for the accu-
mulator discharging numerical experiment is 55 bar, while the
final pressure is 25 bar. In the initial state the water fills the half

Figure 6 Measured inlet steam mass flow rate during the test of steam accu- Figure 8 Condensation relaxation time calculated with Eq. (10) and the con-
mulator charging. stant empirical value 85 s.

heat transfer engineering vol. 36 no. 5 2015

 V. D. STEVANOVIC ET AL. 507

effects are taken into account with the nonequilibrium model

presented here.
The amounts of accumulated and discharged steam for here
simulated charging and discharging steam accumulator tran-
sients are shown in Figure 10. The circles indicate the ends
of the charging and discharging periods predicted with both
nonequilibrium and equilibrium modeling approach. The total
mass of steam that flows in and out of the accumulator during
charging and discharging is noted in parentheses together with
the time instants of the ends of charging and discharging periods.
For example, the nonequilibrium model prediction shows that
the maximum pressure of 55 bar during accumulator charging
Downloaded by [Narodna Biblioteka Srbije], [Vladimir Stevanovic] at 03:36 31 October 2014

with the steam mass flow rate of 30 t/h is reached after 408 s
and during this time 3400 kg of steam is accumulated. As shown
Figure 9 Pressure changes during accumulator charging and discharging with in Figure 10, the nonequilibrium model predicts lower capacity
steam mass flow rates 5 t/h, 10 t/h, and 30 t/h. and shorter charging and discharging periods than the equilib-
rium modeling approach. Hence, the equilibrium model gives
false information about the steam generator capacity, but this
of the accumulator volume in the case of accumulator charging divergence from the exact results obtained with the nonequi-
test, while in the case of discharging, test water initially fills librium model reduces with the decrease of steam accumulator
60% of the steam accumulator volume. The steam charging and charging and discharging flow rates, and vice versa—the di-
discharging flows are stopped when the accumulator pressure vergence of the equilibrium modeling results from the exact
reaches maximum (55 bar) and minimum (25 bar) values. The nonequilibrium results increases with the increase of the steam
results obtained with the nonequilibrium model presented here generator charging and discharging flow rates. In the case of
are shown together with the results of the equilibrium model. 30 t/h charging or discharging rate, the relative difference be-
The presented real pressure changes in Figure 9, predicted with tween calculated equilibrium and nonequilibrium charged or
the nonequilibrium model, show that in cases of charging the discharged steam mass, upon reaching the maximum/minimum
pressure decreases after reaching the maximum setpoint value setpoint pressure, is 14.5%; in the case of 10 t/h rate it is 6.2%,
and the stoppage of steam charging flow. After the relaxation and 4.3% in the case of 5 t/h charging and discharging rate.
time the pressure reaches the new steady-state equilibrium con- This observation is consistent with the basic statement that the
dition at a value lower than 55 bar. In cases of discharging, equilibrium thermodynamic changes exist only under indefi-
the pressure increases after reaching the minimum setpoint of nitely slow state changes of the thermodynamic system that are
25 bar and the stoppage of the steam discharging flow. Af- imposed by infinitesimal disturbances. Results presented here
ter the relaxation time the pressure reaches steady equilibrium
state at a pressure level higher than 25 bar. This nonequilib-
rium effect is pronounced the most in cases, with the highest
charging and discharging rates 30 t/h, while it decreases with
the reduction of the steam inflow and outflow rates. Hence, in
the case with 5 t/h in and out flow rates the final steady-state
pressure slightly decreases from the boundary setpoints. The
results obtained with the equilibrium modeling approach show
that after reaching the maximum and minimum setpoint pressure
and boundary steam flow stoppage there is no further change
of pressure. This is obviously not the real situation, since the
evaporation and condensation processes cannot be conducted
with infinite phase transition rates. During charging the steam is
compressed, the pressure increases, and the water is subcooled,
that is, the water temperature is lower than the saturation temper-
ature. After the charging stoppage, steam continues to condense
until the water temperature reaches the saturation temperature. Figure 10 Steam mass that flows in and out of the accumulator during charging
During discharging the pressure in the accumulator decreases and discharging with mass flow rates 5 t/h, 10 t/h, and 30 t/h predicted with the
nonequilibrium (NEQ) and the equilibrium (EQ) modeling approach (the first
and water becomes superheated, that is, the water temperature numbers in parentheses are time instants when the maximum operating pres-
is higher than the saturation temperature. After the discharging sure 55 bar during charging and minimum pressure 25 bar during discharging
stoppage, water continues to evaporate until the water temper- are reached, while the second number presents the accumulated or discharged
ature reaches the saturation temperature. These nonequilibrium amount of steam).

heat transfer engineering vol. 36 no. 5 2015

 508 V. D. STEVANOVIC ET AL.

show that in order to charge or discharge the accumulator with

the prescribed amount of steam, it is necessary to reach a higher
or lower pressure in the accumulator than a pressure predicted
with the equilibrium model. Therefore, the equilibrium model
overestimates the actual steam accumulator capacity. In other
words, the equilibrium thermodynamic model provides false re-
sults if applied to the steam accumulator design, and the error of
prediction increases with the increase of the steam accumulator
charging and discharging flow rates.
The results obtained with the nonequilibrium model in Fig-
ure 10 also show that the amount of accumulated or discharged
steam decreases with the increase of the accumulator inlet or
Downloaded by [Narodna Biblioteka Srbije], [Vladimir Stevanovic] at 03:36 31 October 2014

outlet steam mass flow rate. In the case of 5 t/h charging or

discharging rate the accumulated or discharged mass of steam
is 10% higher than the mass in the case of 30 t/h rate. This
effect is also conditioned by the nonequilibrium nature of evap-
oration and condensation. The thermal nonequilibrium between
water and steam increases with the increase of the inlet or outlet Figure 12 Transient temperature changes across the accumulator wall pre-
steam mass flow rate, and the delay of condensation or evapo- sented in Figure 11 in the case of 30 t/h steam charging from 25 bar until 55 bar.
ration increases, which results in the decrease of the amount of The wetted surface wall temperature changes linearly from 224◦ C until 270◦ C
accumulated or discharged steam. for 408 s.
Results presented here are obtained under the assumption
that there is no accumulation of heat in the accumulator vessel
metal walls and internals. The influence of the heat accumulation 10. The transient temperature in the wall is calculated numeri-
in metal parts of the vessel on the accumulated or discharged cally by solving the one-dimensional heat conduction equation
steam mass is estimated under the following conditions. The ∂ Tw /∂t =αw (∂ Tw /∂ y), where the thermal diffusivity of the wall
accumulator vessel is designed for a pressure of 67 bar, with the is defined as αw = λw /(ρw c p,w ). The control volume method
corresponding steel wall thickness of 82.3 mm. The total vessel is applied with implicit time integration as described in refer-
mass is approximately 40000 kg. The thermal conductivity of ence [16]. The grid refinement test has shown that the sufficient
the steel wall is λw = 45 W/(m-K), the specific heat is c p,w = 570 number of control volumes along the wall thickness is 40, while
J/(kg-K), and the density is ρw = 7850 kg/m3. It is assumed that the time step of integration is 10−3 s. The temperature changes
the temperature at the inner wall surface equals water tempera- across the wall thickness at several time instants are presented
ture defined with the accumulator instantaneous pressure, while in Figure 12 together with the mean temperature across the wall.
the outer wall surface is adiabatically insulated. The plane wall At 408 s the mean wall temperature is 251◦ C. The amount of
of thickness δ = 82.3 mm is observed as presented in Figure 11. heat accumulated in the metal mass of the accumulator vessel
The left wall surface is adiabatically insulated, while the temper- for the temperature change from the initial 224◦ C to 251◦ C is
ature on the right wall surface linearly changes for 408 s from approximately [251 (◦ C) – 224 (◦ C)] × 570 J/(kg-K) × 40000
224◦ C until 270◦ C (water saturation temperatures for 25 bar (kg) ≈ 616 MJ. This amount of heat corresponds to the latent
and 55 bar, respectively). The initial wall temperature is 224◦ C. heat of condensation of approximately 616 (MJ)/1.775 (MJ/kg)
These conditions correspond to the accumulator charging with ≈ 350 kg of steam. This steam mass is 10% of the accumulated
the steam mass flow rate 30 t/h as presented in Figures 9 and steam mass of 3400 kg presented in Figure 10 for the case with
the steam accumulation rate of 30 t/h. For a low accumulator
charging and discharging rates, such as 5 t/h presented in Fig-
ures 9 and 10, the time period of transient is longer and the
mean temperature of the accumulator vessel walls equals the
steam–water temperature inside the vessel. Hence, the maxi-
mum accumulated heat is [270 (◦ C) – 224 (◦ C)] × 570 J/(kg-K)
× 40000 (kg) ≈ 1050 MJ. This amount of heat corresponds
to the latent heat of approximately 1050 (MJ)/1.72 (MJ/kg) =
613 kg of steam, where 1.72 MJ/kg is the mean value of the
latent heat of water/steam between 25 bar and 55 bar. Hence,
in the case of the accumulator slowly charging and discharging
Figure 11 Transient temperature changes in the accumulator wall. The left
with 5 t/h, the accumulation capacity increase due to the heat
surface is adiabatically insulated, while the right surface is at the saturation accumulation in the vessel walls is 613 (kg)/3740 (kg) = 0.16
temperature determined by the accumulator instantaneous pressure. (16%).
heat transfer engineering vol. 36 no. 5 2015
 V. D. STEVANOVIC ET AL. 509

CONCLUSIONS j mass flux of phase transition at the interface, kg/(m2-s)

k heat transfer coefficient, W/(m2K)
A design of the control system for the steam accumulator kB Boltzmann constant = 1.38 × 10−23 J/K
operation is presented. It defines the measurement and control L length, m
circuits for the regulation of steam accumulator charging and M mass, kg
discharging, the steam flow rate in the accumulator bypass steam ṁ mass flow rate, kg/s
pipeline with pressure reduction station, the regulation of steam n bubble concentration, m−3
temperatures at the steam accumulation inlet, and the regulation p pressure, Pa
of water level in the accumulator vessel. The steam accumu- Pr Prandtl number
lator with implemented control system enables variable steam Q̇ heat transfer rate, W
consumption from two pressure levels under constant steam r latent heat of evaporation/condensation, J/kg
Downloaded by [Narodna Biblioteka Srbije], [Vladimir Stevanovic] at 03:36 31 October 2014

generation. Steam is delivered from the high-pressure header to Re Reynolds number, Eq. (15)
the consumers and in periods of lower consumption the steam T temperature, K
accumulator is charged from this high-pressure header. Steam t time, s
is also delivered to consumers from the low-pressure header u velocity, m/s
and the steam accumulator discharges steam to this header in V volume, m3
periods of increased consumption. The possible modes of the Wecr critical Weber number, Eq. (18)
steam accumulator operation are presented and the proposed v specific volume, m3/kg
control system can satisfy conditions for the most general case x quality
of stochastic steam consumption. y normal distance form the interface, m
The steam accumulator charging and discharging transients
are simulated with both the nonequilibrium and equilibrium Greek Symbols
model. The nonequilibrium model is based on the mass and en-
ergy balance equations for each phase (liquid water and steam) α thermal diffusivity, m2/s
and nonequilibrium correlations for condensation and evapo-  phase transition rate, kg/(m3-s)
ration rates. The equilibrium model is based on the mass and δ thickness, m
energy balance equations for the fluid inventory (steam–water λ thermal conductivity, W/(K)
mixture) in the accumulator. The nonequilibrium model is vali- μ dynamic viscosity, kg/(m-s)
dated against measured data. Performed numerical experiments ρ density, kg/m3
of the steam accumulator charging and discharging transients σ surface tension, N/m
show a divergence of the equilibrium model results from the ex- τ phase change relaxation time, s
act nonequilibrium results. The error of the equilibrium model
prediction increases with the increase of the steam accumulator Subscripts
charging and discharging flow rates. The increase of the accumu-
lation capacity due to the heat accumulation in the accumulator
B boundary parameter
vessel steel walls is estimated in the range between 10% and
b bubble
16%, where this effect is more dominant during slow transients.
c condensation
e evaporation
i interface
FUNDING
in inlet
PT phase-change parameter
This research was supported by the Ministry of Education, out outlet
Science, and Technological Development of the Republic of sat saturation
Serbia (grant 174014). w wall
1 water
NOMENCLATURE 2 steam
1i transfer between interface and water
2i transfer between steam and interface
a steam–water interface concentration, m2/m3
21 interficial transfer from steam to water, Eq. (21)
cp specific heat at constant pressure, J/(kg-K)
D diameter, m
g gravity, m/s2 Superscripts
Gi parameter, Eq. (20)

H total enthalpy, J saturated water
h specific enthalpy, J/kg  saturated steam

heat transfer engineering vol. 36 no. 5 2015

 510 V. D. STEVANOVIC ET AL.

REFERENCES [14] Hamming, R. W., Numerical Methods for Scientists and

Engineers, Dover Publications, Inc., New York, NY,
[1] Goldstern, W., Thermal Energy Storage in Industry and 1986.
Power Stations, Proc. International Conference on En- [15] Stevanovic, V., Maslovaric, B., and Prica, S., Dynamics of
ergy Storage, Brighton, UK, BHRA Fluid Engineering, Steam Accumulation, Applied Thermal Engineering, vol.
pp. 113–120, 1981. 37, pp. 73–79, 2012.
[2] Price, N., Steam Accumulators Provide Uniform Loads [16] Patankar, S. V., Numerical Heat Transfer and Fluid Flow,
on Boilers, Chemical Engineering, vol. 89, no. 23 pp. Hemisphere, New York, NY, 1980.
131–135, 1982.
[3] Shnaider, D. A., Divnich, P. N., and Vakhromeev, I. E., Vladimir D. Stevanovic is a full professor of thermal
and nuclear power engineering at the Faculty of Me-
Modeling the Dynamic Mode of Steam Accumulator, Au-
chanical Engineering, University of Belgrade, Bel-
tomation and Remote Control, vol. 71, pp. 1994–1998,
Downloaded by [Narodna Biblioteka Srbije], [Vladimir Stevanovic] at 03:36 31 October 2014

garde, Serbia. He received his B.Sc., M.Sc., and Ph.D.

2010. in mechanical engineering in 1983, 1986, and 1992
[4] Beckmann, G., and Gilli, P. V., Thermal Energy Storage, from the University of Belgrade. His research work
Springer-Verlag, Berlin, Germany, 1984. is in the field of thermal-hydraulics, computational
multi-fluid dynamics, thermal and nuclear power
[5] Cardu, M., About the Low Parameters Steam Power Re-
plants safety, energy efficiency of power plants,
covery, Energy Conversion and Management, vol. 36, no. thermal and hydraulic transients in complex pipe
10, pp. 963–967, 1995. networks, and energy planning and policy.
[6] Steinmann, W. D., and Eck, M., Buffer Storage for Direct
Steam Generation, Solar Energy, vol. 80, pp. 1277–1282, Milan M. Petrovic is a research assistant and a Ph.D.
2006. student at the Faculty of Mechanical Engineering,
[7] Bai, F., and Xu, C., Performance Analysis of a Two- University of Belgrade. He received his M.Sc. degree
in mechanical engineering at the University of Bel-
Stage Thermal Energy Storage System Using Concrete
grade in 2012. His research interest includes compu-
and Steam Accumulator, Applied Thermal Engineering, tational fluid dynamics and multiphase phenomena.
vol. 31, pp. 2764–2771, 2011.
[8] Xu, E., Yu, Q., Wang, Z., and Yang, C., Modeling and
Simulation of 1 MW DAHAN Solar Thermal Power
Tower Plant, Renewable Energy, vol. 36, pp. 848–857,
2011. Sanja Milivojevic is a research assistant at the In-
[9] Chen, G. Q., Yang, Q., Zhao, Y. H., and Wang, Z. F., novation Centre of Faculty of Mechanical Engineer-
ing, University of Belgrade, and a Ph.D. student at
Nonrenewable Energy Cost and Greenhouse Emissions of
the Faculty of Mechanical Engineering, University
a 1.5 MW Solar Power Tower Plant in China, Renewable of Belgrade, where she received her M.Sc. degree
and Sustainable Energy Reviews, vol. 15, pp. 1961–1967, in 2006. Her research interest includes thermal hy-
2011. draulics in components of thermal power plants and
[10] Majanne, Y., Model Predictive Pressure Control of Steam district heating systems, such as temperature wave
propagation in pipelines and water hammer induced
Networks, Control Engineering Practice, vol. 13, pp.
by intensive steam condensation.
1499–1505, 2005.
[11] Kuznecov, Yu. N., Heat Transfer in Safety Problems of Nu-
Blazenka Maslovaric is a teaching assistant and a
clear Reactors, Energoatomizdat, Moskva, Russia, 1989 Ph.D. student at the Faculty of Mechanical Engineer-
(in Russian). ing, University of Belgrade. She received her M.Sc.
[12] Zuber, N., and Findlay, J. A., Average Volumetric Con- degree in mechanical engineering at the University
centration in Two-Phase Flow Systems, Journal of Heat of Belgrade in 2006. Her field of work is numeri-
Transfer, vol. 87, pp. 453–468, 1965. cal investigation of two-phase flow including phase-
change processes and separation of phases. She has
[13] Blinkov, V. N., and Frolov, S. D., Modeling of Flashing been involved in thermal hydraulics analyses within
Flow in Nozzles, Engineering and Physical Journal, vol. essential parts of thermal power plants and district
42, pp. 741–746, 1982 (in Russian). heating system.

heat transfer engineering vol. 36 no. 5 2015