Proceedings of the 2006 WSEAS/IASME International Conference on Heat and Mass Transfer, Miami, Florida, USA, January 18-20,

2006 (pp62-68)

Stability analysis of natural circulation systems

HEIMO WALTER and WLADIMIR LINZER
Institute for Thermodynamics and Energy Conversion
Vienna University of Technology
Getreidemarkt 9, A-1060 Vienna
AUSTRIA

Abstract: In natural circulation systems with unequally heated risers which are connected by a header the static
instability, namely reverse flow, can occur. This paper presents different possibilities to determine the mass flow
distribution in the tube network of such systems. To calculate the mass flow distribution a graphical and the numerical
analysis (stationary and dynamical) was used. The investigation was done for a two-pass boiler under variable
operating pressures and heat loads. The dependence of the critical heat absorption ratio Vcrit from the pressure is
discussed. A lower boundary for the part load operation of the analysed natural circulation steam generator was
determined. Modifications of the boiler geometry were analysed and the calculation results were presented and
discussed.

Key-Words: Natural circulation, Reverse flow, Stability analysis, Steam generator, Two-pass boiler

1 Introduction unheated.
Natural circulation steam generators have many In the present study two different boiler designs are
advantages compared to forced- and once-through analysed:
circulation boilers. The main advantage of the natural 1. The tubes of both riser systems are connected with
circulation system is the simple design which results in the drum (design case 1). No intermediate headers
an unproblematic operation. The part load operation of and riser relief tubes (displayed with broken lines in
such a boiler is possible without special auxiliary Fig. 1) are implemented in this case.
equipment. Compared with the forced circulation boilers 2. The unheated part of both riser systems is replaced
the power for the feed water pump is lower. by an intermediate header and riser relief tubes
The steam quality of the water/steam mixture in the (design case 2). The ratio of the cross section area of
evaporator tubes of natural circulation steam generators the riser tubes to the relief tubes is 3.05.
is in the normal operation case not higher than
approximately 30 %. Therefore film boiling respectively Steam to superheater
critical heat flux can only occur under unfavorable PD
conditions. 20 m
Steam generators with natural circulation consist of a Drum
Riser relief tube
more or less complicated tube network. To avoid
instable circulation certain design criteria must be taken Intermediate header
into consideration. Unstable conditions can be the result 15 m Feed
of static [1]-[4] (e. g. reverses flow) and/or dynamic water
instabilities [5]-[7] (e. g. density wave oscillation).
In this paper the static instability, namely the reverse .
m0
.
m1
.
m2
flow, will be illustrated. Flow reversal can occur in
systems with unequal heated tubes which are connected
by a header (see Fig. 1). The investigation to identify .
q1
.
q2
reverse flow in a natural circulation system can be done
Riser system 1

Riser system 2
Downcomer

with the aid of a graphical and a numerical analysis.

2 Investigated boiler 0m
Figure 1 shows a sketch of the investigated two-pass
natural circulation steam generator. The higher heated Lower header
riser system 2 corresponds to the furnace and the lower
heated riser system 1 represents the tubes of the wall in Fig. 1: Design of the two-pass boiler
the second pass. The large diameter downcomer is
Proceedings of the 2006 WSEAS/IASME International Conference on Heat and Mass Transfer, Miami, Florida, USA, January 18-20, 2006 (pp62-68)

3 Graphical Analysis The characteristic curve for the downcomer is

The graphical analysis is used to study the mass flow
distribution under steady state conditions. For the
determination of the mass flow rate in the individual
tubes of the boiler it is necessary to combine the
different tube characteristics. The characteristic curve for
a tube can be obtained by plotting the total pressure
difference between the drum and the lower header pDH
against the mass flow rate (see Fig. 2). The total pressure
difference pDH can be calculated by
pDH = pFric + pH + pAccel. (1)
pFric includes the pressure loss to friction as well as
the pressure loss to e. g. in- and outflow and bends. The
pressure difference due to acceleration pAccel is identical
to zero in case of an unheated tube. pH describes the
pressure difference of the static head.
For the graphical analysis the total pressure difference
pDH for every individual tube must be equal (equation
(2)) and also the mass balance for the lower header
(equation (3)) must be fulfilled.
pDH1 = pDH2 = pDH3 (2)
m& 0 = m& 1 + m& 2 (3)
The negative sign in equation (3) is used if the mass
flow in the lower heated tube is directed downward. The
counterclockwise flow direction is counted positive.
Figure 2 should help to explain the procedure for the
graphical determination of the mass flow distribution in
the tubes of the boiler design 1. At the horizontal axis of
Fig. 2 the mass flow and at the vertical axis pDH is
outlined.
0 m1
illustrated in Fig. 2 by the full line and that for the riser
system 1 and 2 by the doted lines. Dependent on the
flow direction of riser system 1 the mass flow for the
unequally heated riser systems must be added or
subtracted at constant pressure difference pDH (In case
of reverse flow in the lower heated riser system the
characteristic curve for m & 1 < 0 must be subtracted (see
arrows in Fig. 2) from the characteristic m & 2 > 0). This
results in a circulation characteristic (broken lines)
dependent on the flow direction of riser system 1.
It can be seen that the circulation characteristic for
upward flow direction in both riser systems (signed with
m& 1 > 0 and m& 2 > 0) has only one intersection point with
the characteristic of the downcomer (point A). In this
intersection point equation (2) and (3) are fulfilled. The
associated circulation mass flow m & 0 , m& 1 and m& 2 as well
as pDH can be read from the axis. In case of reverse
flow in riser system 1 the circulation characteristic
(signed with m & 1 < 0 and m& 2 > 0) has two intersection
points with the graph for the downcomer (point B and
C). Only solution B is stable.
Higher heat absorption at the lower heated system 1
moves the circulation characteristic for reverse flow in
direction of lower mass flow rate and pDH. This leads to
the consequence that the solutions B and C move
together. The point where B and C are merged represents
the limit of univocal flow direction in the riser system
with the lower heat absorption. For further increase of
the heat flow q&1 the downcomer characteristic does not
cross the graph for the circulation characteristic with the

mass flow m & 1 < 0 and m& 2 > 0. In this case only upward
e
m. s ys t
DH

Do
wn flow is possible in both riser tubes.
p

1 >
se r

co The graphical solution is very illustrative but also

Ri

m
m > em 2

er
very time-consuming. Therefore a numerical simulation
0

is preferable to find design criteria to prevent reverse

. t
r s ys

flow. In the following paragraphs the results of such

2
R i se

numerical analysis (stationary and dynamical) will be

B presented.
C
0, . is ti c
m. ra c t e r o n

2 >
0
ti
ch a rc u l a

4 Stationary Analysis
e r is tic
0

The evaluation of the stationary behavior of steam

Ci
. s te m 1

1 >
c h a r a ct o n
1 < 0, m >

generators is a standard procedure for the design of

2
ti
.
Ci r c u l a
m1 < 0

heating surfaces. Many computer programs e. g. NOWA

R ise r s y

A [8] were developed in the past to calculate the stationary

operation conditions of steam generators. The data for
.m

the geometry (e.g. tube diameter and lengths, wall

thickness, etc.), the system pressure as well as the heat
absorption profile for the heated tubes are necessary for
Mass flow the input of such a code. The temperature-, density-,
pressure- or mass flow distributions in the tube network
Fig. 2: Graphical solution of the boiler are the result of such a calculation.
Proceedings of the 2006 WSEAS/IASME International Conference on Heat and Mass Transfer, Miami, Florida, USA, January 18-20, 2006 (pp62-68)

2
q 2= 75 kW/m
50 .
m2 .
p = 80 [bar] m1[kg/s]
D
40 . 10
m2

320
3 00

20 0

1 00
30 0 0
Solution A
.
q [kW/m2]
M ass flow [kg/s]

1 D
20 Solution B -10 75
100
Solution C
10 . -20
m1 q 2 = 320 kW/m2
200
0
1 5 10 20 30 40 50
. .
V= q2/q1
-10 C 300
.
m0
A V=1 V=2 V=5 V=10
320

-20
.
m .
q [kW/m2]
1 0 1 2 2
-30
.
m0 B,C
-40

-50 0 1
Fig. 3: Mass flow distribution at different heat
absorption ratios V [2]
Figure 3 shows as a result of a static stability analysis
the mass flow distribution in the tubes of the steam
generator. The calculations are done for the boiler design
shown in Fig. 1 without the riser relief tubes and the
intermediate headers (design case 1). The heat flux for
riser system 2 with q& 2 = 320 kW/m2 was kept constant
while the heat flux for q&1 decreases. A negative sign in
Fig. 3 describes a flow direction from the drum to the
lower header.
Up to a heat absorption ratio of V = q& 2 / q&1 = 7 only
one solution (A) with upward flow direction in both riser
systems can be found. With a heat absorption ratio
higher than 7 two stable (A and B) and one unstable (C)
solutions are possible for the circulation mass flow. The
solution C in Fig. 3 was found with the help of the
graphical method presented in the paragraph above. The
heat absorption ratio of 7 represents the critical value
Vcrit between a stable upward flow in both riser systems
and a possible operation condition with reverse flow in
the lower heated riser system 1.
A complete solution for the mass flow in the lower
heated riser system 1 m & 1 for all possible heat absorption
ratios is presented in Fig. 4. The Fig. shows that reverse
flow is possible for a heat absorption ratio V > 7 (point
C) at the constant heat flux for riser system 2 with
q& 2 = 320 kW/m2. The critical heat absorption ratio Vcrit
decreases with lower total heat flux. This can be seen for
the curve for q& 2 = 75 kW/m2 (point D).
2
Fig. 4: Mass flow m& 1 at different heat absorption ratios
V as 3-dimensional surface [2]
The static analysis has shown that both stable
solutions (A and B) for the mass flow distribution can be
found. But a static analysis can only find the solution
under stationary conditions. As shown in [2] the correct
static results alone are not a guaranty for a safe operation
of a natural circulation system with unequally heated
risers under all conditions. Therefore the prediction of
the transient behavior of the boiler e. g. under hot start-
up conditions or heavy load changes must be done by a
dynamic analysis.
5 Dynamic Analysis
The one-dimensional mathematical model for the flow in
the tubes, which was used for all calculations in this
paragraph, includes a homogeneous equilibrium model
for the two phase flow and applies a correction factor for
the two phase pressure loss according to Friedel [10].
The collectors and distributors respectively are assumed
to be points. The steam, which enters the header, is
evenly distributed to the outlet tubes. A detailed
description of the computer code is given in [11].
5.1 Boundary and initial conditions
The following initial conditions are used for the dynamic
simulations:
the steam generators are filled with water near
boiling condition
the pressure distribution of the fluid is due to gravity
and the velocity of the working fluid in the tube
network of the evaporator is zero and
the fluid temperature in the evaporator of the boiler
is equivalent to the boiling temperature at drum
pressure.
 Proceedings of the 2006 WSEAS/IASME International Conference on Heat and Mass Transfer, Miami, Florida, USA, January 18-20, 2006 (pp62-68)

The drum pressure of the steam generator is constant
during the whole simulation. The heat fluxes to the two
riser systems of the two-pass boiler are boundary
conditions for the simulation. The water level was
controlled at the drum centerline.
The system pressure will be varied in steps between
80 bars and 12 bars (80 bar, 65 bar, 50 bar, 40 bar,
30 bar 20 bar and 12 bar).
The heating ramp for the simulated hot start-up was
selected according to [9] and is shown in Fig. 5. It can be
seen that in the first period of the simulation the heat
mass flow starts to slow down. Approximately 500
seconds after the start of the simulation the flow
direction of the riser system 1 changes rapidly to upward
flow. Following the change of the flow direction, the
mass flow tends after a short peak to a much lower mass
flow. After about 600 seconds the mass flows in the
downcomer and both riser systems have achieved the
steady state values.
60.0
Riser System 2
40.0

M ass flow [kg/s]

flow increases up to 30 % of the full load. Between 30 s
and 150 s after the ignition of the burner the load is Downcomer
20.0
constant. During the time period between 150 s and
400 s the ramp-like heat input reaches the full load. The 0.0
Riser System 1 p = 80 [bar]
total time for the simulation is 10000 s. D
Q = 4.8 MW
-20.0
6000.0 V=9
V = 9.5
-40.0
0.0 200.0 400.0 600.0 800.0
Time [s]
H eating [kW ]

4000.0 Riser system 2

2000.0
Riser system 1
0.0
0.0 200.0 400.0
Time [s]
Fig. 5: Heating ramp
The total heat absorption (i.e. the heat absorption for
both riser systems) will be varied in steps between
8 MW and 0.8 MW (8 MW, 6.4 MW, 4.8 MW, 3.2 MW
and 1.6 MW). The value of 8 MW was chosen as upper
extreme. For industrial furnaces with oil firing the
corresponding heat absorption can be assumed between
5 MW - 6 MW. The value of 0.8 MW can represent the
Fig. 6: Mass flow distribution at a heat absorption ratio
of V = 9 and V = 9.5
The situation for the circulation mass flow after a
moderate change of the heat absorption ratio from V = 9
600.0 800.0 to V = 9.5 is different. The mass flow in the tubes of the
boiler at V = 9.5 starts similar to that shown for a heat
absorption ratio of V = 9. The mass flow of the working
fluid in riser system 1 starts in the downward direction too.
With the beginning of the steam production in the lower
heated riser system 1, the mass flow starts to slow down.
But, compared to the case with V = 9 the heat flux to the
riser system 1 is not high enough to change the flow
direction. Therefore the flow direction remains different in
both riser systems at steady state.
1.50
Pressure difference [bar]

condition for low load operation or special equipment in 1.40 Riser System 1
process engineering (e. g. refuse incineration plant). 1.30
pDH
1.20 pDH
5.2 Results of the dynamic analysis
1.10
Figure 6 shows as a result of the dynamic simulation the V=9
V = 9.5
mass flow in the different tubes of the boiler. The 1.00
simulation refers to the boiler design without the p = 80 [bar]
0.90 D
Q = 4.8 MW Riser System 2
intermediate headers and riser relief tubes at 80 bars, a 0.80
total heat flow of 4.8 MW and a heat absorption ratio of 0.0 200.0 400.0 600.0 800.0
V = 9 (full line) and V = 9.5 (broken line). Time [s]
In the following the results for the heat absorption ratio
of V = 9 will be described in detail: Fig. 7: Pressure difference distribution at a heat
It can be seen that the lower heated riser system 1 starts absorption ratio of V = 9 and V = 9.5
in the downward direction with a relatively high mass
flow rate. The higher heated riser system 2 starts in the Figure 7 shows the pressure difference of the static head
upward direction with approximately the double mass pH for both riser systems as well as the total pressure
flow rate as the riser system 1. With the beginning of the difference between the drum and the lower header pDH.
steam production in the lower heated riser system 1, the
Proceedings of the 2006 WSEAS/IASME International Conference on Heat and Mass Transfer, Miami, Florida, USA, January 18-20, 2006 (pp62-68)
At first the results for the heat absorption ratio of V = 9
will be described:
During the first phase of the hot start-up pDH is smaller
than the pressure difference of the static head and
therefore the mass flow in the lower heated riser system
1 is directed downward. With the beginning of the steam
production pH starts to decrease. Between 480 and 520
seconds after the start of the simulation the increasing
steam production in the riser system 1 reduces the static
head pH and crosses the graph of the pressure difference
pDH. In this period, the mass flow changes from
downward to upward direction.
The pressure difference due to height pH to the case of
a heat absorption ratio of V = 9.5 is always higher than
pDH (see Fig. 7). Therefore the mass flow is always
directed downward in the lower heated riser system 1
also at steady state.
Fig. 8: Critical heat absorption ratios
Figure 8 shows the critical heat absorption ratio Vcrit for
the boiler design 1 at different system pressures and heat
loads.
A heat absorption ratio of Vcrit 5 is assumed to be a
design criterion for the two-pass boiler for stable
operation at all heat loads. A curve with the constant
heat absorption ratio of Vcrit = 5 (broken line) is included
in Fig. 8. It can be seen that the system tends to flow
reversal for low heat loads even at 80 bars.
Beside the overall circulation ratio the fluid velocities
are also of important interest for the boiler design. In a
certain temperature (pressure) range high velocities can
be the reason for erosion-corrosion ([13] [14]).
Figure 9 shows the velocity in the higher heated tubes. It
can be seen that the velocity corresponds over a wide
range of pressures and heat loads with the usual design
rules (The velocity of the steam-water mixture should be
approximately 12 m/s and the velocity in the
downcomer 4m/s [12]). A curve with the constant
velocity of 12 m/s is included in Fig. 9.
Fig. 9: Velocity in the tubes of riser system 2
The analysis of the basic design has shown that the
boiler should not operate in the pressure range < 30 bars
and in the very low load region, even at 80 bars.
Therefore modifications on the system geometry are
necessary. One of the most successful methods to get a
higher stability is to modify the thermo-hydraulic
behaviour of the circulation system. This can be done by
reducing the pressure loss in the downcomer or
increasing the pressure loss in the riser systems [4], [12].
The investigation on the design case 1 has also shown
that the mass flow in the higher heated tubes is higher
than necessary for cooling the tubes. Therefore a
redesigning of the unheated section of the riser systems
should be done with the aim to reduce the mass flow in
the higher heated riser system. This can be done by
replacing the unheated part of both riser systems by an
intermediate header and riser relief tubes (see Fig. 1) [4],
[12].
In the following the results for design case 2 will be
presented.
The investigation for this test case is subdivided into two
parts:
1. The number of unheated riser relief tubes for the
lower heated system 1 is two and for the higher
heated system 2 is four.
2. The number of unheated riser relief tubes for the
lower heated system 1 is four and for the higher
heated system 2 is two.
At first the design case with the four riser relief tubes
implemented in the higher heated riser system 2 will be
described.
Figure 10 shows the critical heat absorption ratios Vcrit
for the first analysed configuration of design case 2. The
modifications at the unheated part of the riser systems
has moved the lower limit for stable upward flow of
Vcrit = 5 in direction to the lower heat loads. Therefore no
operation restriction at lower heat loads is given for the
pressure range down to approximately 60 bars.
Proceedings of the 2006 WSEAS/IASME International Conference on Heat and Mass Transfer, Miami, Florida, USA, January 18-20, 2006 (pp62-68)
Fig. 10: Critical heat absorption ratios for design case 2
with four riser relief tubes at the higher heated riser
system 2
The simulation result for the velocity in the riser relief
tubes of the higher heated riser system 2 is shown in Fig.
11. It can be seen that the curve with the constant
velocity of 12 m/s is shifted in direction of higher
pressures and lower heat loads. A detailed analysis of the
data in the pressure region between 30 bars and 50 bars
and higher heat loads has shown that the velocities are
close to 12 m/s. Therefore a safe operation is also given
for this boiler design down to 30 bars.
Fig. 11: Velocity in the riser relief tubes at the higher
heated riser system 2
The installation of four riser relief tubes in the lower
heated system 1 increases the critical heat absorption
ratio especially in the high pressure/high heat load
region (see Fig. 12). The limit for stable upward flow of
Vcrit = 5 is also moved to lower heat loads.
It can be seen that this boiler design has achieved the
highest values for Vcrit from all analysed configurations.
A comparison of the data has shown that the pressure in
the lower header for the design case 2 with four riser
relief tubes at the lower heated riser system 1 was higher
than in the other cases while the pressure in the lower
header of the basic design was the lowest. The higher
pressure in the lower header moves the critical heat
absorption ratio to higher values. This was demonstrated
also in [12].
Fig. 12: Critical heat absorption ratios for design case 2
with four riser relief tubes at the lower heated riser
system 1
Figure 13 shows the velocity in the riser relief tubes of
the higher heated riser system 2. In this case the curve
for the constant velocity of 12 m/s is moved into the low
load region for all operation pressures. This is a result of
the lower number of riser relief tubes at the riser system
2.
Fig. 13: Velocity in the riser relief tubes at the higher
heated riser system 2
On the basis of this result the current design case can be
used only in a very restricted operation range. Therefore
a further redesign (e. g. larger dimension for the riser
relief tubes and intermediate headers) of the boiler is
necessary.
6 Conclusion
A stability analysis for a natural circulation system with
two unequally heated riser systems which are connected
by a header with an unheated downcomer was
performed. For the determination of the static instability,
namely reverse flow, different methods have been
presented. The investigation was done for different
operation pressures and heat loads.
Proceedings of the 2006 WSEAS/IASME International Conference on Heat and Mass Transfer, Miami, Florida, USA, January 18-20, 2006 (pp62-68)
The study has shown that for every system pressure,
heat load and boiler configuration a special critical heat
absorption ratio Vcrit exists. This critical heat absorption
ratio must be determined for every configuration of such
a boiler. The critical heat absorption ratio depends also on
the thermo-hydraulic behaviour of the circulation system.
Heating conditions which exceed this critical value can
lead to unstable conditions including the possibility of
flow stagnation and or reverse flow.
The analysis of the data has shown that all
modifications which lead to a higher pressure in the
lower header results in a higher stability.
In boiler configurations with unequally heated tube
systems the number of riser relief tubes which are
connected with the lower heated tubes should be smaller
or equal to the number of riser relief tubes which are
connected with the higher heated tubes. Otherwise the
velocity in the relief tubes can exceed the upper limit.
7 Nomenclature
m& 0 Mass flow in the downcomer [kg/s]
m& 1 Mass flow in riser system 1 [kg/s]
m& 2 Mass flow in riser system 2 [kg/s]
q&1 Heat flow to riser system 1 [kW/m2]
q& 2 Heat flow to riser system 2 [kW/m2]
V Heat absorption ratio [-]
Vcrit Critical heat absorption ratio [-]
pAccel Pressure difference due to acceleration [bar]
pDH Total pressure difference between the drum and
the lower header [bar]
pFric Pressure difference due to friction [bar]
p H Pressure difference of the static head [bar]
References:
[1] Ledinegg, M., Instability of flow during natural and
forced circulation, Die Wrme, Vol.61, No.48, 1938,
pp. 891-898
[2] Linzer, W. and Walter, H., Flow reversal in natural
circulation systems, Applied Thermal Engineering,
Vol.23, No.18, 2003, pp. 2363-2372.
[3] Walter, H., and Linzer, W., Flow Stability of Heat
Recovery Steam Generators, Proceedings of the
ASME Turbo Expo 2004, Power for Land, Sea and
Air, Vienna, Austria 14.-17. June 2004, Paper Nr.
GT2004-53040, pp. 1-9.
[4] Walter, H. and Linzer, W., Investigations to the
Stability of a Natural Circulation Two-Pass Boiler,
Proceedings of the ASME - ZSIS International
Thermal Science Seminar II, Bled, Slowenia 13.-16.
June 2004, pp. 469-474.
[5] Walter, H., and Linzer, W., Dynamic Flow
Instability of Natural Circulation Heat Recovery
Steam Generators, Proceedings of the 16th
International Symposium on Transport Phenomena
(ISTP-16), Prague, Czech Republic, 29. August 1.
September 2005, pp. 1-11.
[6] Nayak A. K., Vijayan P. K., Saha D., Raj, V. V. and
Aritomi M., Analytical Study of Nuclear-Coupled
Density-Wave Instability in a Natural Circulation
Pressure Tube Type Boiling Water Reactor, Nuclear
Engineering and Design, Vol.195, 2000, pp. 27-44.
[7] Guanghui S., Dounan J., Fukuda K. and Yujun G.,
Theoretical and Experimental Study on Density
Wave Oscillation of Two-Phase Natural Circulation
of Low Equilibrium Quality, Nuclear Engineering
and Design, Vol.215, 2002, pp. 187-198.
[8] Nowotny, P., A contribution to the calculation of the
mass flow distribution in tube networks, Fortschritt-
Berichte VDI, Series 6, No.102, VDI-Verlag,
Dsseldorf, 1982
[9] Linzer, W., Das Ausstrmen von Siedewasser und
Sattdampf aus Behltern, Brennstoff-Wrme-Kraft,
Vol.22, No.10, 1970, pp. 470-476.
[10] Friedl, L., Improved Friction Pressure Drop
Correlation for Horizontal and Vertical Two-Phase
Pipe Flow. European Two-Phase Group Meeting,
Ispra, Italy, 5. 8. June 1979, Paper E 2, pp. 1 - 25.
[11] Walter, H., Modeling and Numerical Simulation of
Natural Circulation Steam Generators, Fortschritt-
Berichte VDI, Series 6, No.457, VDI-Verlag,
Dsseldorf, 2001
[12] Walter, H. and Linzer, W., The influence of the
operating pressure on the stability of natural
circulation systems, Applied Thermal Engineering,
(accepted for publication).
[13] Loos C. and Heitz E., The mechanism of the
erosion corrosion in fast fluid flows. Werkstoff
und Korrosion, Vol.24, No.1, 1973, pp. 38-48
[14] Kastner W., Riedle K. and Tratz H.,
Experimental investigation to the material
abrasion through erosion corrosion. VGB
Kraftwerkstechnik, Vol.64, No.5, 1984, 452-465