A DYNAMIC MODEL OF ONCE-THROUGH STEAM GENERATOR

(OTSG) FOR PROTOTYPICAL ADVANCED REACTOR

Fan Zhang and Jamie Coble*
Department of Nuclear Engineering
University of Tennessee
Knoxville, TN 37996 USA
[email protected]; [email protected]*

ABSTRACT

A new design of Small Modular Reactor, the Prototypical Advanced Reactor (PAR), features
two sodium primary systems each connected to dedicated steam generators (SGs) and sharing a
balance of plant (BOP) system. PAR adopts once-through steam generators (OTSG) for its
compact structure. The existence of superheated steam in a OTSG improves turbine design by
eliminating the need of bulky moisture separator in SG. The two cores and SGs’ design enable
PAR to adjust power flexibly, as a result, introduce more challenges to the control. Therefore, a
fast simulation model with high accuracy OTSG model is needed. In this paper, a dynamic model
is built for OTSG of PAR based on the conservation of mass, momentum, and energy. The OTSG
is simulated as lumped pipe and divided into subcooled, evaporating, and superheated regions by
moving boundaries. SIMULINK is adopted to this OTSG simulation to be merged into existing
PAR system models easily. The steady state results show that key parameters are rational. The 5%
step increase of feedwater enthalpy transient demonstrates that the model is capable to respond to
the perturbations. Since the basic laws and algebraic method used in this paper are suitable for
other fluids, the model can be easily applied to any other OTSG. Future work will also focus on
model optimization, merging and improved control. The final aim is to perform full scope real-
time simulations for PAR under normal operational transients.
Key Words: OTSG, dynamic model, PAR, SIMULINK

1   INTRODUCTION

Small modular reactors (SMRs) are nuclear reactors with less power (300 MWe or less) than
traditional full scale plants. The compact designs and factory fabricated features enable them to be
transported by truck or rail to a nuclear power site, and used in parallel to get larger total power output.
SMRs are planned to be operating commercially within next two decades, and will play a significant role
in addressing energy, security, economic, and climate goals. One SMR design, the Prototypical Advanced
Reactor (PAR), is a liquid sodium fast breeder reactor design with a power of 40 MWe. There are three
loops in PAR: primary loop, intermediate sodium coolant loop, and the secondary loop. As shown in
figure 1, Prototypical Advanced Reactor (PAR) includes two independent reactor cores [1], each
connected to a dedicated intermediate Heat eXchanger (HX) and Steam Generator (SG). Steam from the
two modules mixes within one steam header, and is sent to the common balance of plant (BOP) system to
generate electricity.
PAR adopts Once-through steam generator (OTSG) for its configuration is compact enough to enable
it to be located in the pressure vessel. The existence of superheated steam in OTSG eliminates the bulky
moisture separator in SG and benefits for turbines. Those advantages make OTSGs are widely used in
SMRs’ design.

*
Corresponding author: Jamie Coble

NPIC&HMIT 2017, San Francisco, CA, June 11-15, 2017 1999

 The SG is one of the most important components in nuclear power plants, which always need special
attention. The dual-SG design in PAR introduces extra challenges for control than the traditional one SG
design, especially under the situations that the two reactor cores in PAR operate at different power levels.
The steam from a SG is an important variable in nuclear plant control, which is both needed in BOP
system control and reactor power control. Therefore, a good OTSG model is needed for whole system
simulation.

Figure 1. Prototypical advanced reactor power blocks [1]

Before building a new type of nuclear system, simulations are used to assess the performance and the
consequences of possible configuration changes to provide information and instruction. Two kinds of
dynamic models are widely used for single device in nuclear system such as steam generators. One is fine
mesh model, which typically involves the distributed parameter described by partial differential equations
and its discretization equations. It is mainly used for detailed thermal hydraulic analysis and has high
accuracy which need much calculation time [2]. The other model is the lumped parameter model
described by ordinary differential equations (ODEs), which is mainly used for fast simulation, compound
system simulation and control system encompassed simulation. For this kind of simulation, the models
should be high fidelity and faster than the former kind, which is able to simulate the perturbation
responses of the nuclear power plant fast. This kind of model is suitable to assess the full scope
compound system performance and train operators. In addition, it can be used to define relationships
between a set of plant signals, which is useful for equipment health monitoring, diagnostics, and
prognostics. A SIMULINK model of the PAR system of was previously developed [3]. This model used a
linearized perturbation model of the SG; to improve the performance of whole system model, a dynamic
nonlinear model for OTSG of PAR is built in this paper, which will be integrated in the full PAR model in
the future.
Previous research has been completed for OTSG simulation. Li et al. has developed a lumped
parameter model of the helical coiled OTSG of the 10 MW high-temperature gas-cooled reactor (HTR-
10) [4]. Dong has developed a model for the OTSG of modular high temperature gas-cooled reactor
(MHTGR), which has been applied to the real-time simulation software for the operation and regulation
features of MHTGR-based commercial nuclear plant [5]. The fluid in primary side of OTSG in the
research of both Li and Dong is helium. Zhu et al. has developed a dynamic model for OTSG with

NPIC&HMIT 2017, San Francisco, CA, June 11-15, 2017 2000

 concentric annuli tube using Gear method. The fluid in primary side is water [6]. Berry developed a
model of OTSG with sodium as the primary fluid, however, the method is not convenient to be modeled
by SIMULINK [7]. Therefore, a dynamic model for OTSG with sodium as primary fluid is developed in
this paper.
All of these researches apply moving boundary method into OTSG simulation. Since moving
boundary method describes the heat transfer character appropriately, it also be adopted in this paper.
Based on conservation laws of mass, momentum, and energy, this paper gives the algebraic model for
OTSG with sodium in the primary side. Two moving boundaries divide OTSG into subcooled region,
evaporating region and superheated region. Steady state for the OTSG model is conducted by
SIMULINK, which shows that the steady-state calculation results are rational. The transient of 5% step
increase of feedwater enthalpy is conducted and results show the key parameters changes are rational and
reach the new steady state. Future work will focus on the model optimizing, model merging and controller
improving. The real-time full-scope simulation and control for PAR under operational steady state and
transient will be achieved in the future.

2   METHODOLOGY

The dynamic model for OTSG is developed in this part. Figure 2 shows the flow chart of method
solution. The OTSG model involves solving conservation equations for both primary side and secondary
side, tube wall equations, fluids properties calculation and heat transfer models. The details of these
blocks are described in this section.
OTSG

Heat transfer
Conservation equations Tube wall Fluids properties
models

Sodium in Water and steam in Water and

Sodium
primary side secondary side steam

Figure 2. the flow chart of method solution

2.1   Description of OTSG Nodes in PAR and Assumptions

The OTSG of PAR is a vertical, shell-and-tube type SG. The counter flow design is used to obtain
effective heat transfer between two sides of OTSG. Fluid in the primary side of OTSG is the single-phase
sodium, which is the sodium in intermediate coolant loop of PAR. The fluids in the secondary side are
water and steam, which are in secondary loop of PAR. To model the OTSG performance better, the OTSG
fluid channel are divided into three regions, which are subcooled region, evaporating region, and
superheated region, with two moving boundaries, which are defined according to the specific enthalpy of
water in the secondary side. Subcooled region is the region where the fluid enthalpy is less than saturated
water enthalpy; nucleate boiling region is the region where the fluid enthalpy is equal to or greater than
saturated water enthalpy but less than saturated steam enthalpy; the superheated region is the region
where the fluid enthalpy is greater than saturated steam enthalpy. In this paper, two phase flow in all the
tubes of the OTSG are assumed to be homogenous, which is a common assumption in fast fluid modeling
[4][5][6]. Therefore, both sides of OTSG are lumped and treated as single straight flow channel.

NPIC&HMIT 2017, San Francisco, CA, June 11-15, 2017 2001

 Sodium
in x Steam

7 7

6 6
Superheated

5 5

Primary Secondary
side Evaporating side

4 4

3 3

2 Subcooled 2

1 1

Sodium
Feedwater
out

Figure 3. Nodes scheme of OTSG model

Figure 3 shows the nodal scheme of OTSG model. Point 1 to 3 is the subcooled region and point 2 is
the central point of this region. Point 3 to 5 is the evaporating region and point 4 is the central point of
this region. Point 5 to 7 is the superheated region and point 6 is the central point of this region. Since the
boundaries between regions are movable, so the length of subcooled region L13 and length of superheated
region L15 vary by time, which are set as state variables. The nodes in primary side are the same with the
secondary side for easy calculation.
Since the SIMULINK model in this paper is for future whole PAR system model simulation to
facilitate perturbation assessment, health monitoring, diagnostics and prognostics, the calculation speed is
an important evaluation. To enable model high speed while little sacrifices of accuracy, two reasonable
assumptions are made: Axial heat conduction is ignored; Enthalpy and pressure in the same heat transfer
region varies linearly, which means these values of central point can be calculated by the nearby boundary
points.

2.2   Conservation Equations and Algebraic Method

Sodium in primary side, water and steam in secondary side all follow the conservation laws of mass
momentum and energy describes by Equation (1) (2) (3) with the coordinate frame given in figure 3.

¶r ¶ (ru )
= (1)
¶t ¶x

¶(ru ) ¶(ru 2 ) ¶P ru 2 Ff S

= - - - - rg (2)
¶t ¶x ¶x 2 A

¶(rh) ¶(rhu ) q ¶P

= - + + (3)
¶t ¶x A ¶t

NPIC&HMIT 2017, San Francisco, CA, June 11-15, 2017 2002

 where 𝜌 is the fluid density (kg/m3), u is the fluid velocity (m/s), 𝑃 is the fluid pressure (Pa), F𝑓 is the
friction factor, 𝑆 is the wetted perimeter (m), 𝐴 is cross sectional area (m2), 𝑔 is the gravitational
acceleration (N⋅m/s2), ℎ is the fluid specific enthalpy (J/kg), and 𝑞 is the heat flux per unit length (W/m2).

Apply equation (1)(2) (3) for both primary side and secondary side for the OTSG based on the nodes
in Figure 2. By integrating the mass conservation equation and energy conservation equation in different
regions respectively along x-axis direction using the Leibniz theorem [8], the ODEs can be obtained. For
example, equation (4) and (5) shows the ODEs for fluid in secondary side in subcooled region.

d ( r s 2l13 ) dl (4)
- r s 2 13 + Ds 3 - Ds1 = 0
dt dt

d (hs 2 r s 2l13 ) dl Q dp l dl
- r s 3hs 3 13 + Ds 3hs 3 - Ds1hs1 = 2 + s 2 13 - ps 3 13 (5)
dt ¶t As 2 dt dt

where 𝑙13 is the length of subcooled region (m), 𝐷 is the fluid mass velocity (kg/(s*m2)), and Q2 is the heat
flux in the subcooled region (W). The explanation of subscripts is shown in Table 1.

Table I. Explanation of subscription

Subscription s m fw 1,2,3,4,5,6,7
Secondary Metal tube Point
Explanation Feedwater
side wall 1,2,3,4,5,6,7

The enthalpy in node 2 is equal to the average of enthalpy in node 1 and 3. Since the fluid in node 3
is saturated, the enthalpy is just the function of pressure. So equation (5) can be described as equation (6).

é rs 3 (hs1 - hs 3 ) ù dl r l dh l æ ¶h ö dP Q (6)

ê + ( Ps 3 - Ps1 ) ú 13 + s 3 13 s1 = 13 ç1 - rs 2 s 3 ( P = Ps 3 ) ÷ s 3 + (Gs1 As1 + Gs 3 As 3 )(hs1 - hs 3 ) / 2 + 2
ë 2 û dt 2 dt 2 è ¶Ps 3 ø dt As 2

Taking time-delay of secondary inlet the subcooled region into consideration, the fluid flow rate in
node 3 and the enthalpy in node 1 can be presented by first-order delay as shown in equation (7) and (8)
[8].

dhs1 h fw - hs1
= (7)
dt t s

dGs 3 Gs1 - Gs 3
= (8)
dt t s

Since the pressure-flow process dynamic behavior is much faster than the enthalpy–temperature
dynamic behavior, the time-dependent term in the momentum equation can be ignored to simplify the
equation (2) as shown below [4].

¶(ru 2 ) ¶P ru 2 Ff S

+ = - - rg
¶x ¶x 2 A (9)

NPIC&HMIT 2017, San Francisco, CA, June 11-15, 2017 2003

 Integrate equation 6 in the subcooled region, we can get the following equation:

Ps3 - Ps1 = -rs 2 gl13 - Ffs 2l13 Ds22 (10)

Derive equation (10) to obtain equation (11)

dPs 3 dl Ffs 2Gs 3 dGs 3

= -( r s 2 g + Ffs 2Gs22 / As22 ) 13 - (11)
dt dt As22 dt

The mass flow of point 3 to point 7 are treated as equal since the density in these two regions are
much less than in the subcooled region. Applying the same method for evaporating region and
superheated region to get the ODEs.

2.3   Equations for Tube Wall

General heat conduction equation is

dT
r c = f (12)
dt
where 𝜌 is the density (kg/m3), c is the specific heat (J/kg*K), T is the temperature (K), f is the heat
production per volume (W/m3)
Integrate equation (12) using Leibnitz theorem in each region of tube wall, we can obtain equations
(13) (14) and (15).

dTm 2 Qm 2 T - T dl
= + m 4 m 2 13 (13)
dt Am 2 r m 2 cm 2l13 l15 dt

dTm 4 Qm 4 T - T dl T - T dl
= + m 6 m 4 15 + m 4 m 2 13 (14)
dt Am 4 rm 4 cm 4 (l15 - l13 ) l17 - l13 dt l15 dt

dTm 6 Qm 6 T - T dl
= + m 6 m 4 15 (15)
dt Am 6 r m 6 cm 6 (l17 - l15 ) l17 - l13 dt

2.4   Heat Transfer Models

Convection heat transfer between primary and secondary side is convection heat transfer. The
general heat flux of convection heat transfer is given by equation (16).

q = ht (T1 - T2 ) (16)

where q is the heat flux, ht is the heat transfer coefficient, T1 is the temperature of subject 1, and T2 is the
temperature of subject 2.
The heat transfer model is described by nine equations: convection heat transfer equations between
sodium in primary side and tube wall surface in primary side in three regions; conductivity heat transfer
equations within tube wall in different regions; convection heat transfer equations between tube wall

NPIC&HMIT 2017, San Francisco, CA, June 11-15, 2017 2004

 surface in secondary side and fluid in secondary side in different regions. The heat transfer coefficient
used for sodium in primary side to metal tube wall is 7046.77 W/m2*K according to Reference [9].
The fluid status in the secondary side of OTSG are different. The Dittuse-Bolter correlation is used
for single-phase liquid in subcooled region and single-phase steam in superheat region [10].

Nu = 0.023Re0.8 Pr 0.33 (17)

where Re and Pr is the Reynolds number and the Prandtl number of fluid, respectively.
Chen correlation is used in evaporating region [11]. In Chen correlation, two basic mechanisms in
the heat transfer process for the fluid flow boiling are taken in to consideration. They are the ordinary
macro-convective mechanism of heat transfer in flowing fluids, and the micro-convective mechanism
related to bubble-nucleation and growth. So the total heat transfer is contributed by these two mechanisms
additively as equation (18).

h = hmic + hmac (18)

where the h is total heat transfer coefficient, hmic is the micro convective mechanism heat transfer
coefficient, and hmac is the macro convective mechanism heat transfer coefficient. Forster and Zuber’s
formulation is used to calculate hmic [11].

é l 0.79c0.45
, r G ù
0.49 0.23

hmic = 0.00122 ê f0.5 p0.29f f0.24 0.24 ú DT 0.24 DP0.75 Su (19)
êë s µ f h fg rg úû

where 𝜆 is the thermal conductivity (Btu/h*ft* oF), Cp is the heat capacity (Btu/lb*oF), G is the
gravitational constant, σ is the vapor-liquid surface tension (lb/ft), hfg is the latent heat of vaporization
(Btu/lb), ∆T is the wall superheat (oR), and ∆P is the pressure difference corresponding to the wall
superheat (psf). Su is the suppression factor represents the ratio of the effective superheat to the total
superheat of the wall, subscript f stands for liquid water phase and g for steam phase.
The macro convective heat transfer coefficient hmac is calculated by

l f
hmac = 0.023Re0.8
f
Prf0.4 F (20)
D

where D is the hydraulic diameter (m) and F is the ratio of the two-phase Reynolds number to the liquid
Reynolds number, which is calculated by Martinelli parameter Xtt [11].

2.5   Fluids Properties Calculation

The sodium in primary side is the single-phase sodium. Therefore, the sodium properties involved in
this model uses lumped parameter from two references [9][12].
Lots of properties of water and steam are involved in this model state, including density, enthalpy,
heat conductivity coefficient, viscosity, surface tension, specific heat, latent heat, saturated pressure based
on temperature. After compared international water and steam calculation formulations [13],
“International Association for the Properties of Water and Steam Industrial Formulation 1997” (IAPWS-
IF97) is adopted in this model to calculate the properties of water and steam [14].

NPIC&HMIT 2017, San Francisco, CA, June 11-15, 2017 2005

 2.6! Inputs of SIMULINK Model
To solve the equations above, select primary sodium pressure, enthalpy, mass flow at the inlet point
and secondary feedwater pressure, enthalpy, mass flow at the inlet point as six input variables, which can
be provided by PAR IHX and BOP system calculation when merged into the PAR whole system
simulation.

3! RESULTS AND DISCUSSION

The steady-state simulation was performed and all the key parameters can reach to the steady. The
calculation results show that subcooled region length is 10.82m and the evaporation region length is
20.78m. Figure 4 shows the 500s steady state simulation of temperature in different regions of the two
sides of metal wall between the primary and secondary side of OTSG and the enthalpy of deferent regions
of the primary sodium and the water and steam in the secondary side. The surface temperature of metal
wall in primary side is higher than the surface temperature of metal wall in secondary side. The values of
these key parameters are rational to the OTSG design.

Figure 4. Steady state of key parameters

NPIC&HMIT 2017, San Francisco, CA, June 11-15, 2017 2006

 The 5% increase of the feedwater enthalpy transient is conducted. At first, the system is performing
the steady state 100s without any perturbation and a 5% feedwater enthalpy step increase is insert at 100s.
Figure 5(a) shows the subcooled region length and evaporation region length in 5% step increase of
feedwater enthalpy transient. The subcooled region decreases with the enthalpy increases because that the
heat needed for preheating the feedwater to saturated water decreases. Figure 5(b) shows the surface
temperature in different regions of the metal wall in secondary side. Figure 5(c) shows the surface
temperature in different regions of the metal wall in primary side. Figure 5(d) shows sodium enthalpy in
different regions. The parameters in figure 5(b-d) decrease slightly after the 5% step increase of feedwater
enthalpy perturbation, because the increased feedwater enthalpy makes heat transfer between the OTSG
primary side and secondary side decrease. Figure 5 shows that the parameters can reach new steady state.
The key parameters changes in transient are rational.

(a) (b)

(c) (d)

Figure 5. Response to a 5% step increase of feedwater enthalpy transient in the (a) subcooled region length
and evaporation region length, (b) surface temperature in different regions of the metal wall in secondary
side, (c) surface temperature in different regions of the metal wall in primary side, and (d) Sodium enthalpy
in different regions

NPIC&HMIT 2017, San Francisco, CA, June 11-15, 2017 2007

 4   CONCLUSIONS

A dynamic model of OTSG for PAR is developed in this paper. The OTSG is divided into subcooled,
evaporating and superheated region by two moving boundaries according to the enthalpy of the secondary
fluid. The model is based on the conservation laws of mass, momentum, and energy, and solved by
integrating using Leibnitz theorem. For heat transfer models, Chen correlation is used in evaporating
region and the Dittuse-Bolter correlation is used for both single-phase liquid in subcooled region and
single-phase steam in superheat region. IAPWS-IF97 is used for the properties of water and steam
calculated for its high accuracy.
Simulation of the OTSG model is carried out by SIMULINK, which is easy to be merged into the
existing PAR system model and has good graphic function. The steady-state results show that OTSG
model can reach steady-state and the results are rational. The 5% step increase of feedwater enthalpy
perturbation transients is conducted to demonstrate that the model is able to respond to perturbation and
build new steady state. More transient analysis will be done before the next stage of model optimizing,
merging and control improving. The high-performance real-time full-scope simulation for PAR will be
developed in the future.

5   REFERENCES

1.   Coble, J.B., et al., “Technical Needs for Enhancing Risk Monitors with Equipment Condition
Assessment for Advanced Small Modular Reactors,” PNNL-22377, SMR/ICHMI/PNNL/TR-
2013/02, Pacific Northwest National Laboratory, Richland & USA (2013).
2.   Daogang Lu, Yu Wang, Bo Yuan, Danting Sui, Fan Zhang et al.,“Development of three-
dimensional thermal-hydraulic analysis code for steam generator with two-fluid model and
porous media approach,” Applied Thermal Engineering, Vol. 116, pp. 663-676 (2017).
3.   Xiaotong Liu, “Modeling and Simulation of a Prototypical Advanced Reactor,” M.S. thesis,
Department of Nuclear Engineering., The University of Tennessee, Knoxville & USA (2016).
4.   Li, Haipeng, Huang, Xiaojin, Zhang, Liangju, “A lumped parameter dynamic model of the helical
coiled once-through steam generator with movable boundaries,” Nuclear Engineering and
Design, Vol. 238, pp.1657-1663 (2008).
5.   Zhe Dong, “A Differential-Algebraic Model for the Once-Through Steam Generator of MHTGR-
Based Multimodular Nuclear Plants,” Mathematical Problems in Engineering, Vol. 2015, 12
pages (2015).
6.   Zhu Jingyan, Guo Yun, Zhang Zhijian, “Dynamic simulation of once-through steam generator
with concentric annuli tube,” Annals of Nuclear Energy, Vol. 238, pp.1657-1663 (2008).
7.   Greg Berry, “Model Of a Once-Through Steam Generator With Moving Boundaries and a
Variable Number Of Nodes,”  ASME Winter Annual Meeting, Boston, MA, USA (1983).
8.   Leibniz, “Supplementum Geometriae Dimensoriae, Seu Generalissima Omnium
Tetragonismorum Effectio Permotum: Similiterque Multiplex Constructio Lineae ex Data
Tangentium conditione,” Acta Eruditorum, pp. 385–392 (1693).
9.   Berkan, R.C. and B.R. Upadhyaya, “Dynamic Modeling of EBR-II for Simulation and Control,”
Annual Report, The University of Tennessee, Knoxville & USA (1988).
10.   Dittus FW, Boelter LMK, “Heat Transfer in Automobile Radiators of the Tubular Type,”
International Communications in Heat and Mass Transfer, Vol. 12, pp. 3-22 (1985).

NPIC&HMIT 2017, San Francisco, CA, June 11-15, 2017 2008

 11.   Chen JC, “A Correlation for Boiling Heat Transfer to Saturated Fluids in Convective Flow,”
Process Design and Development, Vol. 5, pp. 322-327 (1966).
12.   Zengyuan Qian, Thermal Physical Properties of Metal with Low Melting Point, Science Press.,
Shuang Fu & China (1983).
13.   Fan Zhang, Dao-gang Lu, Dan-ting Sui, Bo Yuan, and Chao Guo, “Preliminary Development of
Thermal Power Calculation Code H-Power for a Supercritical Water Reactor,” Science and
Technology of Nuclear Installations, Vol. 2014, 10 pages (2014).
14.   Cooper JR, Revised Release on the IAPWS Industrial Formulation 1997 for the Thermodynamic
Properties of Water and Steam, International Association for the Properties of Water and Steam,
Lucerne & Switzerland (2007).

NPIC&HMIT 2017, San Francisco, CA, June 11-15, 2017 2009