Advantages of calculation methods in design

of highly loaded steam turbine stages

D. Obrzut , supervisor A. Gardzilewicz

1. Introduction

A tendency has been observed in recent years towards increasing load of turbine
stages. This tendency refers both to steam and gas turbines, with its basic goals to
reduce the number of turbine stages, shorten the turbine axial size, and, what is most
important, decrease turbine manufacture costs. These goals are a source of new
problems for turbine designers, who are forced to seek novel ways to minimise flow
losses in stator and rotor rows. The increased loads are equivalent to increased
enthalpy drops, higher flow velocities, and larger forces acting on rotor and stator
blades, as compared to traditional design solutions.
To execute higher stage load, the flow passages in stator and rotor rows are to be
modified, first of all by changing the geometry of blade profiles. The new profiles are
expected to execute changed kinetics of steam flow through the turbine to secure high
efficiency, meeting at the same time all strength and dynamic criteria.
The shapes of standard and highly loaded profiles used in rotor cascades are shown in
Fig. 1 according toYamamoto [1].

Fig. 1 Comparing rotor cascades used in: typical turbine stages (a)
and highly loaded stages (b) [1].

As seen in the figure, the rotor profiles of highly loaded stages create longer flow
passages with remarkably changing cross sections, which results from visibly larger

1
change in flow turning angle realised by the blade row. The flow is to approach and
leave the rotor cascade in a highly loaded impulse stage at a relatively small angles, as
compared to a traditional design. Such large flow deflection angles, amounting even to
as much as 160o, can be reached only thanks to visible in the shapes of both the
pressure (concave) side and the suction (convex) side of the rotor profile, see Fig. 1.
Obviously, higher flow turning is a source of higher losses. Seeking for possible ways of
their reduction was a primary goal of investigations of highly loaded profiles. It turned
out that the factor which exerted the highest impact on the level of losses was the shape
of the inner passage. As a consequence, the convex side of those profiles started to be
designed with an extremely long and slightly changing segment to lead smoothly the
flow at the cascade inlet, and even longer, almost rectilinear exit segment, the role of
which was to secure small flow angles at cascade exit. The investigations have
revealed that the use of thicker profiles in those cases limits the growth of secondary
and edge losses, see Amos [2].
An attempt to use highly loaded stages in the ALSTOM Elbląg Design Office was
connected with a new design of a series of 50 100 MW impulse heating turbines, the
entire flow system of which was to be contained in a single casing. A detailed analysis
has revealed that in order to meet this requirement the number of stages was to be
reduced by 10-20%. In that case the stator profiles had to operate at exit angles equal
to 10 12o, which was equivalent to the turning angles equal to 130 140o in rotor rows.
These values were smaller than those observed under extreme high load conditions,
examined, for instance, by Yamamoto. The increased load of the newly designed stage
was expected to take place at relatively high efficiency.

2. Analysis and selection of impulse stage profiles for higher loads

The profiles to be used under high-load conditions were selected from the bank
of impulse profiles, available in the ALSTOM Elbląg Design Office. The selected stator
profiles PKH3 and rotor profiles P2V had been developed in the concern and were
verified in tests performed by VKI, Belgium, in the late nineties of the last century.
A series of numerical analyses were performed to confirm applicability of the
selected profiles to new purposes. Relevant viscous flow calculations were done using
FlowER, specialised CFD code developed for 3D studies of flows through turbine
stages and sections. The calculations were performed for rows with cylindricall blades,

2
in which the inner row diameter was Dw 800 mm, and the blade length ranged
between l . 20 - 60 mm. The assumed stagger angles and blade pitches secured high
stage efficiency. The geometry of the rows with PKH-3 and P2V profiles is shown in Fig.
2, along with the structural grids used in CFD calculations [3]. The calculated flow
losses in the new PKH-3 and P2V rows were compared with those recorded in the
traditional impulse stages with PLK and P2 profiles (see fig.2).

Fig. 2 Comparison geometry of rows with PKH-3 ; P2V and PLK P2 profiles, complemented
by the grid used in CFD calculations [3].
It turned out that the losses in the stator rows with new PKH3 profiles were lower by
o
1% at small exit angles 1<11 and larger velocities, compared to those recorded in
typical impulse stage rows with PLK profiles, see Fig. 3,
a). b).

Fig. 3. Comparing energy losses as a function of exit angle 1g for

PKH3-40 and PLK-50 profiles [3].

3
This lower level of flow losses in the rows with PKH3 blades mainly resulted from the
efficiency based optimisation of profile geometry. The new stator profiles were designed
as much thicker than their standard equivalents, therefore their width, which is generally
assumed the strength requirements, could be reduced. At a comparable blade height,
this reduction in blade width led to the reduction of secondary and edge losses.
It is noteworthy that the relative blade pitches determined in the tests for the new
cascades exceeded 0.8, while the blade pitches used in conventional stator cascades
were at the level of 0.7.
Slightly lower losses, as compared to standard designs, were also recorded for
new rotor rows with blade profiles P2V. These loss reductions were mainly observed for
the flows at small inlet and exit angles, characteristic for highly loaded stages.
Comparing the results obtained for new and old designs has revealed that the loss
reduction was clearly visible for inlet angles 1< 27o . In those cases the differences
exceeded 2%, although the minimum loss level for nominal flow conditions was lower by
0.5%, in standard rows, see Fig. 4.

Fig. 4. Comparing energy losses for cascades with P2-35 and P2V-35 profiles, at different inlet
flow angles beta 1 [3].

Positive results of analyses of steam flow through stator and rotor rows with PKH3/P2V
profiles were the motivation for performing further optimisation tests, this time referring
to the entire stages. In that case the efficiency of a group of 3 stages with new and old
cylindrical blades operating in conditions typical for a heating turbine was examined. It
turned out that for the assumed inner diameter Dw and angle 1= 11o, the stages with

4
PKH3/P2V profiles were more efficient than those with PLK/P2 profiles within the
practically entire load range, defined by the u/c0 ratio in the diagram shown in Fig. 5.
Moreover, the efficiency characteristic of new stages as a function of their load turned
out more “flat” than the corresponding curve for traditional impulse stages.
a). b).

Fig. 5. Efficiency change vs. load for a section of 3 stages with radial PKH-3/P2V and PLK/P2
blades [3] .

When the stator exit angles were assumed approximately equal to 1 13.5o , higher
efficiency was recorded for standard stages.
The calculated optimal relative blade pitches for the stages revealing highest efficiency
were at the level of 0.8 for new PKH3 stators, and at the level of 0.7 for old PLK
designs. The blade pitches determined for the both rotors were similar to each other
and approximately equal to 0.7. In those cases the assumed width of the new stator
profiles also affected their efficiency. Taking into account strength conditions, the profile
width was reduced by 20% with respect to standard designs, favourably changing the
length of the turbine flow passage.

3. Design calculations of the turbine flow system.

Turbine stages with new highly loaded profiles were manufactured for the turbine 7CK-
65 to be operated in the Zielona Góra thermal-electric power station. The design
calculations were performed making use of the computer aided design system (CAD), in
operation in ALSTOM, which is schematically shown in Fig. 6.

5
 DESIGN OF TURBINE NEW
STAGES UNTYCYPICAL TURBINE
CONSTRUCTION DESIGNS

DATA OF TURBINE DATA AND PROJECT,

STAGES ASSUMPTIONS ASSUMPTIONS

-steam parameters
inlet(p0 ,t0 ,c0)
-enthalpy drop (h)
-mass flow rate(G)
-electrical and heat power DATABANK OF CALCULATION CFD
-imposed restrictions EXPERIMENTAL TURBINE ELEMENTS
RESULTS 3D&2D

HEAT CYCLE
CALCULATON

DATA FOR THE DATABANK OF THE

CALCULATION OF STANDARD
STANDARD STAGES TURBINE ELEMENTS

-steam parameters at stage - construction type of

inlet(p0 ,t0 ,c0) leakage
-enthalpy drop (h) in stage
-standard profiles
-hub diameter

SIMPLIFIED THERMAL
CALCULATION OF
TURBINE FLOW PATH
VERIFICATION OF
(0D,1D,2D)
TERMAL CALCULATION
3D
PRELIMINARY

SYNTHESIS WITH
MECHANICAL OPTIMIZATION
ANALYSIS DRAWINGS

ANALYSIS

GEOMETRY OF TURBINE PROJECT OF

AND PARAMETERS NEW SOLUTION

Fig. 6 Computer Aided Design System for highly loaded stages.

Thermodynamic parameters of the stages were determined from the turboset thermal
cycle calculations, which took into account parameters of the boiler, cooling water
system, and the assumed thermal and electric power output. It resulted from preliminary
calculations that stages with radial blades could be used in three first stage groups. For

6
these groups main dimensions were calculated, including diameters, blade heights, and
numbers of stages in each group. These data made the basis for detailed design and
check calculations, done using simplified 1D codes in which the results of inviscid flow
solutions were corrected by coefficients determined from 3D CFD calculations
performed for cascades with new profiles, according to the concept presented in [4].
The efficiency optimisation procedures, incorporated to the design codes, made it
possible to determine the numbers of blades and their most favourable stagger angles
in successive stage rows. The calculations took into account standardisation of
company designs with respect to diameters, overlaps, and sealing systems.
Increased stage loads, resulting from the assumed smaller angles and blade exits,
provide opportunities for reducing the number of stages with radial blades from 20 to 16
in the turbine, at slight efficiency decrease which, as relevant tests have proved, was
different by about <2% than the maximum level which could be obtained in
conventional impulse stages.
A computer generated view of the flow system in the first group of stages with cylindrical
blades is shown in Fig. 7.

Fig.7 Section of the stage first group of blade system in the heating turbine 7CK-65.

7
The above figure was generated automatically after checking the stage construction in
stress and vibration calculation tests. Reducing the number of stages with radial blades
made it possible to resign from an additional turbine casing, which had been planned in
the preliminary design. The photo of the turboset and the axial cross section of the
turbine, including two stages controlling the steam flow to heating extraction points, are
shown in Fig.8.

8
Fig. 8. Axial cross section of steam turbine (up) and photo of 7CK-65 turboset (down).

9
 4. Experimental verifications

The results of calculations of the presented turbine flow system design were
verified experimentally. Routine warranty tests, done for the entire turboset, were
extended by thermal measurements on highly loaded cylindrical stages. A scheme of
the measuring system for these stages is given in Fig. 9.

P01

0 P1
P2
P3

T01

Pdol
1 2 3
T1
Tdol T2

P02 T1up
T2up
T3up

T02

Blok przyłączeniowy
NI CB-27T

Karta pomiarowa
NI PCIMCIA 3450

Komputer pomiarowy

Fig. 9 Schematic of the measuring system for testing radial stages in 7CK-65 turbine.

The measured pressures and temperatures have made the basis for calculating
enthalpies and determining the steam expansion line for this part of turbine, see Fig.10.
The tests had the form of a series of four measurements performed under different
conditions of turboset operation. The analysis has revealed that the uncertainly in the
efficiency estimation for the two first groups of stages did not exceed 1%. The third
stage group was excluded from the evaluation, as the operating parameters changed
remarkably there during the tests, due to the introduction of steam with different
parameters from seals to this group.

10
 i

bar
0
pomiar 1
pomiar 2
pomiar 3

1 grupa pomiar 4

2 grupa

3 grupa

x=1

3
1

Fig. 10. Expansion line in the 7 CK-65 steam turbine, according to measurements 1-4.

In Table 1 the calculated results of thermal parameters are compared with those
measured at the turbine. They are complemented by the resultant efficiency levels and
mass flow rates for the two first stage groups, treated as a whole.
Table 1
Measured Calculated
Inlet pressure [MPa] 6.287 6.287
o
Inlet temperature [ C] 498 498
Exit pressure [MPa] 1.553 1.553
Exit temperature [oC] 310.1 307.5
Stage group efficiency [-] 0.879 0.892
Mass flow rate [kg/s] 51.05 51.35

For the above very high agreement between the calculated and measured (using a
reducer) mass flow rate, the recorded efficiency differences exceeded 1.5%. They were
larger than the measurement errors. As the relevant analysis has revealed, this
efficiency difference was caused by not taking into account the unsteady flow effects in
the correction coefficients, calculated using the CFD codes and incorporated into the
design codes.

11
 5. Final conclusions

Designing the flow system of an impulse heating turbine of 65 MW power output in a

single casing required higher load to be carried by particular stages with cylindrical
blades. This goal was achieved by modernisation of the existing stator and rotor
profiles, done with using the of numerical analysis techniques.
The performed CFD calculations have made it possible to tune the correction
coefficients in the conservation equations used in simplified thermal and flow calculation
codes, which are included to the computer aided system for designing steam turbine
flow systems (CAD). These codes were used for designing the steam turbine intended
for operation in the Zielona Góra thermal-electric power station.
The performed experimental measurements have confirmed high efficiency and
operating potential of the newly designed machine. A single-body turbine 7CK65 with
16 stages with new profiles has been in faultless operation over 6 years now.

6. References

[1] A. Yamamoto , T. Matsunama , E. Outa

Three Dimensional Flows and Losses in an Ultra-Highly Loaded Turbine
ASME Turbomachinery, Conf. Hawaji, 2003

[2] Amos I.G. et al.,

Design and off-design optimisation of highly loaded industrial gas turbine stages,
Applied Thermal Engineering 24 (2004)

[3] D. Obrzut,
Designing highly loaded impulse turbine stages with the aid of modern computing
techniques .
Ph.D. Thesis, Gdansk University of Technology, Gdańsk 2008 (in Polish)

[4] A. Gardzilewicz
Current problems in design, construction and diagnostics of steam turbines.
IF-FM Scientific Reports, 502/1461/99, Gdańsk 1999 (in Polish)

[5] A. Gardzilewicz, P. Lampart, J. Świrydczuk, B. Łuniewicz, D. Obrzut;

A mixed 1D/3D Approach to design of turbomachinery blading system;
Seminer/Summer School CFD for Turbomachinery Applications. Sep. 2001, Gdańsk.

[6] B.Łuniewicz, K. Kietliński, D. Obrzut, Cz. Szyrejko, A. Gardzilewicz;

Turbine 7CK for a gas-steam block in EC Zielona Góra.
Proc. of the conference on „Technical Economic, and Environmental Aspects of
combined cycle power plant”, Gdańsk University of Technology, Gdańsk, 2004 (in
Polish).

12