PART TWO: POWER AND ENGINEERING
292
Figure 5.6: Schematic arrangement of the basic steam turbine types.
(a) Simple impulse turbine: de Laval (c.1883). Key: 1: stationary nozzles; 2:
rotating blades; 3: wheel; 4: rotating shaft. The lower diagram shows the change
in pressure (P) and velocity (V) of the steam as it passes through the turbine.
(b) Reaction Turbine: Parsons (c.1884). Key: 1: Nozzles and guide vanes; 2:
rotating blades; 3: rotating drum; 4: rotating shaft. The conical drum counteracts
the decrease in density of the steam as it passes through the turbine. This
arrangement was not used by Parsons who employed a limited number of step
changes in drum diameter. The lower diagram shows the change in pressure (P)
and velocity (V) of the steam as it passes through the turbine.
STEAM AND INTERNAL COMBUSTION ENGINES
293
(c) Pressure-compounded impulse turbine: Rateau (c.1900). Key: 1: nozzles; 2:
rotating blades; 3: nozzles in diaphragm; 4: rotating wheels; 5: rotating shaft.
The lower diagram shows the change in pressure (P) and velocity (V) of the
steam as it passes through the turbine.
(d) Velocity-compounded impulse turbine: Curtis (1896). Key: 1: nozzles; 2:
rotating blades; 3: stationary guides; 4: rotating wheels; 5: rotating shaft. The
lower diagram shows the change in pressure (P) and velocity (V) of the steam as
it passes through the turbine.
Reproduced from E.F.C.Somerscales, ‘Two Historic Prime Movers’, ASME
Paper number 84-WA/HH-2 (1984).
impulse turbine that is now commonly known by his name. The principle of
this turbine is illustrated in Figure 5.6 (c). As in the Parsons turbine, the
pressure drop is divided between a number of stages, thereby limiting the
steam speed. This type of turbine avoids the leakage problem of the Parsons
turbine by only using each stage pressure drop in the stationary nozzles. The
Rateau turbine is classed as a pressure-compounded impulse turbine.
The Rateau turbine differed from the Parsons turbine in another important

respect. In the latter the moving blades are attached to the periphery of a
rotating drum, with the seals between the stages at its outer surface and at the
outer ends of the rotating blades. In the Rateau turbine the blades are mounted
on the periphery of a disc carried on the turbine shaft, so there is only one
interstage seal and it is where the rotating shaft passes through the diaphragm
separating the stages (see also Figure 5.6 (c)). This type of construction has
two advantages over the system used in the Parsons turbine: the leakage area,
and, hence the leakage flow, is smaller because the seal is at the shaft, which
has a smaller diameter than the Parsons drum; and a shaft seal involves parts
with heavier dimensions than a ring of stationary blades, so it can be made
more effective.
The disc construction of the Rateau turbine would, because of the pressure
drop across the moving blades, result in an unacceptably large axial thrust on
the shaft bearings if applied to the Parsons turbine. The axial thrust that does
exist in the latter is smaller because it only acts over the blade area, and is
accommodated by thrust balancing cylinders, connected externally to points of
higher steam pressure, at the high pressure end of the rotor.
Because the solid construction of the drum is better suited to the high
temperatures encountered in the high pressure stages of modern turbines, the disc-
on-shaft construction has been effectively abandoned in these sections and the
blades are mounted on vestigial discs machined out of a forged cylindrical rotor.
Although the Rateau turbine mitigated the leakage problem of the
Parsons turbine, it still required a long shaft to accommodate the large
number of pressure compounded stages. In 1896 the American C.G.Curtis
patented two concepts that resulted in a substantial reduction in the overall
length of the turbine. The first of these was the recovery of the velocity
PART TWO: POWER AND ENGINEERING
294
Figure 5.7: Cross section of the first Parsons steam turbine. This was supplied
with saturated steam at 5.5bar (80psig), ran at 18,000rpm and produced between

4 kW and 7.5 kW at the generator terminals.
Steam was admitted through the inlet (A) and flowed axially to right and left
through alternating rows of moving and fixed blades (shown schematically). The
steam exhausted into the passages C and D which were cast in the turbine casing.
The exhaust pressure (about atmospheric) was maintained by a steam ejector (B).
The diagram also shows the governor (designed because available rotary governors
could not withstand the high shaft speed), which consisted of a small centrifugal
fan, located just to the left of the item lettered C in the diagram, rotating at the
turbine shaft speed, and drawing air from one side of a diaphragm (located just
above the item lettered C). The motion of the diaphragm, as the vacuum varied
with turbine speed, was conveyed by the link (E) to the governor valve controlling
the steam admission. The sensitivity of this governor was increased by sensing the
generator voltage (actually the magnetic field strength, which is proportional to the
voltage) and bleeding more or less air into the vacuum system. In the figure, F is
the soft-iron field sensor mounted on the dynamo pole pieces, and coupled to this
STEAM AND INTERNAL COMBUSTION ENGINES
295
is the spring loaded brass arm (G) that controls the flow of the bleed air. Forced
lubrication using a screw pump (H) was used. The arrows show the direction of oil
flow. The small diagram at the foot of the figure shows one of the bearings that
were designed to allow some transverse movement of the shaft, due to small
residual imbalances. The shaft ran in a long thin sleeve, surrounded by a large
number of washers alternately fitting the sleeve and the casing. The washers were
held in contact with each other by the pressure of a short helical spring
surrounding the sleeve near its end and tightened up by a nut threaded on the
sleeve. The bearing oil was supplied by the screw pump.
Reproduced with permission from W.G.S.Scaife, in First Parsons International
Turbine Conference (Parsons Press, Dublin and London, 1984).

energy in the steam jet leaving a converging-diverging nozzle in several rows of

moving blades with stationary turning vanes between each row (see Figure 5.6
(d)). This multi-row de Laval turbine is known as a velocity-compounded
impulse turbine. The second arrangement patented by Curtis, which was his
special contribution to turbine engineering, was to combine velocity- and
pressure-compounding (Figure 9).
To obtain the necessary funds to support the development of his turbine,
Curtis sold most of his patent rights to the General Electric Company of
Schenectady, New York, in 1897. However, a satisfactory working machine was
not produced until 1902.
The Curtis turbine, like the Rateau turbine, was a disc turbine. Eventually,
as discussed below, the Curtis and Rateau turbines were combined, so that
later forms of impulse turbines that have derived from these two types
generally tend to be of the disc construction, or at least have vestigial discs
machined from a drum.
By 1900 the four basic types of steam turbines had been developed into
practical machines. Although there was no general and immediate tendency to
adopt this prime mover, in spite of its obvious advantages, by the end of the
following decade the steam turbine had established itself in the electric power
generation industry, and had also provided a number of convincing
demonstrations of its usefulness as a marine power plant.
1900–1910: turbines take over
During the first ten years of this century the application of steam turbines to
both electric power generation and marine propulsion spread very rapidly in
consequence of a number of influential demonstrations of its capabilities. The
first of these was in 1891 when a measured steam consumption of 12.2kg/
kWhr (27lb/kWhr) was reported for one of the 100kW condensing (the first
such) machines installed at Cambridge. This was a record for steam turbines
and equalled the performance of the best simple expansion reciprocating steam
engines.
PART TWO: POWER AND ENGINEERING

296
Figure 5.8: Vertical section through the 5000kW vertical Curtis steam turbine
supplied by the General Electric Co. to the Commonwealth Electric Co. for
installation in the Fisk Street Station, Chicago in 1903. Steam inlet conditions:
11.9bar (175psig), 271°C (520°F); exhaust: 0.068bar (28in Hg vacuum);
minimum steam consumption 6.81kg/(kW/hr) (15.0lb mass/(kW/hr)). The
turbine had two pressure compounded stages consisting of a nozzle, three rows
of stationary guide and four rows of moving blades attached to one wheel. The
vertical arrangement was discontinued by the General Electric Co. built after
1913. Reproduced with permission from C.Feldman ‘Amerikanische
Dampfturbinen’, Zeitschrift des vereines deutscher Ingenieure, vol. 48 (1904), p. 1484.

A second significant installation was the two 1000kW turbines, the largest
built to date, that were delivered in 1900 by C.A.Parsons & Co. to the city of
Elberfeld in Germany. These had a steam consumption of 9.12kg/kWhr
(20.1lb/kWhr). These turbines had many of the features of the modern unit:
they were two cylinder tandem compound (see below), and the condenser was
placed below the level of the operating floor, immediately under the exhaust
from the low pressure cylinder.
Another very influential steam turbine, which was placed in service in 1903
at Chicago’s Fisk Street station (see Figure 5.8), was the third Curtis steam
turbine sold by the General Electric (GE) Company. It had a power output of
5MW and was the most powerful steam turbine built up to that time. It was
unusual in being arranged with its axis vertical (a type of construction
abandoned by GE in 1913) and with the alternator above the turbine.
Pressure-andvelocity-compounding were used, and each of the two pressure-
compounded stages consisted of a row of nozzles, three rows of stationary
turning vanes, and four rows of moving blades attached to one wheel.
STEAM AND INTERNAL COMBUSTION ENGINES
297

1910–1920: blending of types
As the patent protection on various types of turbines expired the
manufacturers devised hybrid machines. One of the most significant of these
combinations was the addition to the Parsons turbine at the inlet of a single
velocity-compounded stage, sometimes called a control stage, in place of a
number of the high pressure reaction stages. The rotor of this combined
Curtis-Parsons turbine is much shorter than the rotor of the pure Parsons type
of the same power output. This arrangement confines high temperatures and
pressures to a shorter portion of the turbine, minimizing expansion effects due
to temperature gradients, which can lead to eventual failure of the machine.
Although the Curtis stage is not as efficient as the reaction stage it replaces, the
minimization of leakage and improvement in reliability result in a net gain to
the user of the turbine.
Another technique for minimizing the turbine shaft length, which was
probably first introduced into regular steam turbine design practice with a
25MW turbine built in 1913 by C.A.Parsons & Co. for the Fisk Street station
in Chicago of the Commonwealth Edison Company, was the tandem
compound. In this machine two rotors are arranged in two separate cylinders,
with bearings in each cylinder. The steam passes in succession through the two
cylinders. The rotors are coupled between the cylinders and the load is
connected to one end of the shaft, usually at the low pressure cylinder.
1920–1930: increasing size
The period between 1920 and 1930 was characterized by a very rapid growth
in steam turbine power output; preliminary attempts to use very high steam
pressures; and a large number of serious mechanical failures.
Increasing power output implies increasing mass rate of flow of steam, and
increasing dimensions. The largest dimensions are encountered at the exhaust
from the turbine, where the steam has the largest volume, and the critical
dimension in this region is the length of the last row blades, which are the
longest in the turbine and, therefore, are subjected to the highest centrifugal

stress. Consequently, very careful consideration must be given to the
mechanical design of the rotor and the last row blades, their manufacture and
the materials used, and the means of attaching the blades to the rotor. Since
sound forgings for the rotor can be assured with greater certainty the smaller
their size, there was a tendency during the period from 1920 to 1960 to use the
disc construction in the low pressure sections of the turbines.
Where the length of the last row blades could not be increased a number of
alternative techniques were developed in the 1920s including Baumann
multiple exhaust, due to K.Baumann of the Metropolitan-Vickers Co., multiple
PART TWO: POWER AND ENGINEERING
298
exhaust flows, and multiple low pressure turbines (not used until the 1930s).
Extreme forms of this arrangement, involving three or four low pressure
turbines, each with a double exhaust flow, have been used on the very high
output machines constructed from 1960 onward.
In order to accommodate the largest outputs a combination of decreasing
the turbine speed and multiple exhaust flows was used in the 1920s. This was
in the form of a cross-compound turbine in which the high pressure cylinder
operated at 3000 or 3600rpm and the low pressure cylinder ran at 1500 or
1800rpm, the two cylinders being coupled to separate alternators. This
arrangement appears to have been first used for three 30MW turbines installed
by Westinghouse in 1914 at the 74th Street station in Manhattan of the
Interborough Rapid Transit Co.
The cross-compound tends to be an expensive solution to the problem of
building turbines with large outputs, and European machines, with their lower
speeds (1500 and 3000rpm), did not employ it as extensively as turbines built
in the United States. The two-speed cross-compound has not been used since
the 1970s (except to accommodate very large turbines with outputs of
1000MW or more), because improved materials of construction removed the
incentive for its use.

The steady growth in turbine power output in the 1920s culminated in a
208MW turbine that was constructed by GE for installation in the State Line
station near Chicago in 1929. This was a three-cylinder cross-compound
turbine, and it remained the turbine with the world’s largest output until 1955.
The application of very high inlet pressures to steam turbines was first
attempted in 1925, in order to increase cycle efficiency. Three such turbines
were installed at the Edgar Station of the Edison Electric Illuminating
Company of Boston (now the Boston Edison Company). The first (3.15MW)
had an inlet pressure of 83bar (1200psig) and the other two (10MW) operated
at 97bar (1400psig). They exhausted at 24bar (350psig) into the steam line
connecting the low pressure boilers to two 32MW units.
In many instances in the 1920s and 1930s new high pressure turbines were
employed in connection with existing low pressure systems; a method known
as superposition. (In a few cases the high pressure turbine was actually
mounted on the low pressure turbine casing.) It allows existing, serviceable
plant to be increased in efficiency and power output for minimum cost, and for
this reason it was popular in the post-depression years of the 1930s.
The rapid advances in steam turbine technology in the 1920s were not
achieved without some cost. Early in this period the manufacturers of disc
turbines experienced an exceptional number of failures. It was discovered that
the discs were subject in service to vibrational oscillations leading to fatigue
failure. This was overcome by employing better design theories and also by
using heavier discs, including discs forged integral with the shaft. Important
contributions to the solution of this problem were made both in the United
STEAM AND INTERNAL COMBUSTION ENGINES
299
States and in Europe; particularly noteworthy was the work of Wilfred
Campbell of the General Electric Company, after whom one of the important
design tools developed at that time, the Campbell Diagram, is named.
1930–1940: refinement

Between 1930 and 1940 no really striking advances of the type seen in the
1920s occurred. The depression following the stock market crash in 1929 did
not provide a business climate in which large orders for steam turbines could
be expected. Consequently, turbine manufacturers turned to improving the
detailed design of steam turbines and a number of features that are now
common practice in steam turbine construction were introduced.
As a result of the desire to increase the inlet pressure the double shell
construction shown in Figure 5.9 was introduced. This decreases the load on
the turbine casing and fastenings by dividing the pressure difference from the
turbine inlet pressure to the ambient pressure between two casings, one inside
the other.
In the 1930s creep, the slow ‘plastic movement’ of steel subjected to high
temperatures and pressures, which was first identified in the 1920s, began to
be considered in steam turbine design. This phenomenon can significantly
alter stress distributions during the time of exposure of the turbine parts to
operating conditions. Its effects can be minimized by adding suitable alloying
materials to the steel that stabilize the material, and by developing extensive
empirical data on the material’s properties for use in design.
Warped, or twisted, low pressure blades, were introduced in the 1930s.
These compensate for the effect of variations in the blade tangential velocity
with radius so as to ensure that the steam impinging on the blade enters the
blade smoothly and with the minimum flow disturbance.
1940–1950: increasing speed
Steam turbine progress in the 1940s was constrained by the outbreak of the
Second World War. After the war ended turbine design showed a definite
trend away from the standard speed of 1500/1800rpm and the establishment
of the high speed turbine operating at 3000/3600rpm. Higher speed turbines
are smaller and lighter than comparable low speed machines. Smaller
turbines expand less when heated, so distortion is decreased, which improves
the turbine’s long-term reliability. Consequently, the high speed machine is

better adapted to increasing inlet pressures and temperatures and to the
application of reheat, which were features of steam turbine design from the
late 1940s onwards.
PART TWO: POWER AND ENGINEERING
300
In about 1948 there was a revival of interest in the reheat cycle (first
introduced in Britain at North Tees in 1920, and in the United States at Philo,
Ohio, in 1924) which was an added incentive for the introduction of the high
speed turbine. In this cycle the temperature of the partially expanded steam is
raised to about the original inlet temperature by passing it through a special heat
exchanger, the reheater, which is heated either by the boiler combustion gases or
by high temperature steam (the latter arrangement was only used in the 1920s),
and then returned to the turbine for further expansion to the condenser pressure.
The motivation for the earlier application in the 1920s had been to decrease the
moisture, which causes blade erosion, in the low pressure sections of the turbine.
Because the required additional valves and piping could not be economically
justified at that time, the construction of new reheat turbines ceased in the early
1930s. The revival of interest in the late 1940s was stimulated by a need to
increase plant efficiency in order to counteract rising fuel costs.
Reheat produces a 4–5 per cent improvement in cycle efficiency, and has a
number of other advantages compared to non-reheat operation. Thus, there is
a reduction in the mass rate of flow of the steam, which, in turn, leads to a
decrease in the size of the boiler feed pump, the boiler, the condenser, and of
the feed water heating equipment. This, together with the ability to reduce the
wetness of the steam at the exhaust makes reheat an attractive feature, and is
widely used in modern steam cycles.
1950–1960: very high pressures and temperatures
The first turbines handling steam at supercritical pressures (pressures in excess
of 221bar/3200psig) were built in the 1950s with the aim of improving cycle
efficiency, while avoiding the problems associated with increasing the turbine

inlet temperature, which requires the development of new materials, an
expensive and time-consuming process.
The first supercritical turbine was installed in 1957 at the Philo station of
the Ohio Power Company. It had an output of 125MW and used steam at
310bar (4500psig) and 621°C (1150°F). Double shell construction was used,
and, because of the exceptionally high pressure, the outer casing is almost
spherical in form.
A significant feature of these very high pressure turbines is the small blade
lengths in the high pressure stages (0.95cm in the Philo turbine), which results
from the high density of the steam. The leakage space is, in consequence, a
large fraction of the blade length, so turbines operating at very high pressures
should, to offset the leakage loss, be designed for large outputs, e.g Philo
(1957), 125MW; Philip Sporn (1959), 450MW; Bull Run (1965) 900MW.
In about 1955 the ‘average’ steam temperature reached its present plateau of
566°C (1050°F). The attainment of this temperature was made possible by the
STEAM AND INTERNAL COMBUSTION ENGINES
301
introduction of the ferritic stainless steels (11–13 Per cent chromium), and required
the adoption of special design features. In the high pressure and intermediate
pressure sections all parts were designed to allow free expansion and contraction.
For example, the steam chests and valves were mounted separately from the
turbine casing and connected to the inlet nozzles by flexible piping. To reduce
temperature gradients in the turbine, the partially expanded steam was arranged to
flow through the outer space of the double shell construction.
1960–1980: increasing size
During the 1960s a number of extremely large output machines were placed in
service. The mass rate of flow steam is so great (for a typical 660MW turbine
about 2.1×10
6
kg/hr or 4.7×10

6
lb mass/hr) that multiple low pressure sections
have to be provided.
The 1960s was a period when many nuclear power stations commenced
operation, and often these used either the boiling water or pressurized water
cycles in which steam is supplied at pressures ranging from 31 bar (450psig) to
69bar (1000psig), with the steam dry and saturated at about 260°C (500°F). To
compensate for the relatively low energy content of steam, the turbines have
large outputs (1300MW), with correspondingly large steam mass flow rates (e.g.,
a 1300MW ‘nuclear turbine’ handles about 7.3×10
6
kg/hr (16.0×106lb mass/hr)).
The resulting long (1.143m, 3.75ft) last row blades in the low pressure sections
have forced American practice to adopt the 1800rpm tandem-compound design.
Because of the lower speed, European designs of ‘nuclear turbines’ have
sometimes been able to employ 3000rpm machines, but most examples of this
type of turbine have been tandem-compounds operating at 1500rpm.
The saturated inlet conditions result in a high moisture content in the
turbine low pressure sections, leading to blade erosion unless reheating is used.
Because the water cooled reactor provides only a low temperature heat source,
reheating is not as effective as it is in fossil fuel-fired plants, so mechanical
moisture separation must be used in addition to reheating.
In the 1970s some even larger steam turbines, with power outputs in excess
of 1000MW, came into service. To handle the large quantities of steam
required by these machines, multiple low pressure stages were arranged in
parallel. Figure 5.9 shows the section of one of the two 1300MW turbines
completed by Brown Boveri in 1974 for the Gavin station of the American
Electric Power Company. Machines of this type are the largest ever to be used
on a fossil fuel-fired cycle.
The historical development of the steam turbine can be summarized in a

number of ways. Figure 5.10 shows the progress in power output, inlet
temperature and inlet pressure from 1884 to 1984.