The Swedish Development of Turbogenerators with directly

water-cooled Rotors
Sture Eriksson, Member IEEE
Royal Institute of Technology
Electrical Machines and Power Electronics
Teknikringen 33, SE 100 44 Stockholm, Sweden
sture.ja.eriksson@telia.com

were combined with the radial turbine as figure 1 shows.

Abstract – Large steam turbine-driven generators rated at a
The advantages with the STAL-turbine were that they were
few hundred MW and higher constitute, in many respects, a
big engineering challenge. The Swedish manufacturer of
very compact, had a good efficiency, and thus were cost-
heavy electrical equipment, Asea was faced with this effective. This type of turbine proved to be very suitable
challenge in the late sixties, when the company started to for industrial backpressure applications and STAL
develop such generators for nuclear power plants. Due to the manufactured large numbers of such units for installation
company’s background, it was necessary to choose new all over the world. Usually, the industrial turbines were
design concepts, and Asea decided on a very unique concept, rated below 50 MW, consequently the two generators less
turbogenerators with directly water-cooled rotors. The than 25 MW each. Asea therefore became an important
development led to difficult teething problems before the manufacturer of smaller turbogenerators. The delivery in
generators could be delivered and operate satisfactory; a 1965 of four 76.4 MVA, 3000 rpm generators for the
process, which took around a decade to complete. Since then, Swedish State Power Board “Vattenfall” were the largest
the operation records have been very good. These turbogenerators with which Asea had real experience when
turbogenerators constitute the only existing group with a the rapid development of the much larger generators for
significant number of two-pole, directly water-cooled rotors. the nuclear power plants started.
The background, the development and results are
summarized in this paper.

I. THE START POSITION

Asea had, in the early 1960’s, a position as one of the

world’s leading manufacturers of hydropower generators.
An important reason for this was a large domestic market
for such generators. The harnessing of the abundant energy
from the Swedish waterfalls for electricity production
started already in the late 19th century and the construction G
of new hydropower plants continued then for 70 – 80
years. During this long period, there was a stable growth in
generator size and several Asea generators have been
milestones also in the international development; some Figure 1. A radial flow high-pressure turbine combined
were the largest in the world at the time they were put into with two axial flow low-pressure turbines requires two
operation. generators while a common axial flow turbine drives one
generator
The situation for Asea as manufacturer of turbogenerators The diagram in figure 2 shows how the size of Asea’s
was very different, even if the company had delivered its turbogenerators had developed until the mid 1960’s. The
first directly steam turbine-driven generator already in corresponding international development has been
1903. Steam turbines had been used for driving dynamos included for comparison and also the hydropower
since the late 1880’s, but the directly coupled turbo- generators. It is evident that Asea, at that time, had a
generator with cylindrical rotor was first introduced around pronounced profile as manufacturer of hydropower
the turn of century. It had been invented by Charles E. L. generators [1].
Brown, who was one of the founders of Brown Boveri Co.
(BBC). Some years later, in 1908, two Swedes, the
brothers Birger and Fredrik Ljungström invented and Turbogenerators were always delivered together with the
developed a special type of steam turbine, the so-called steam turbines. All the large electrical companies built
double rotation, radial flow turbine. They established a complete units consisting of both turbines and generators.
company named Svenska Turbinfabriks AB Ljungström There was hardly any separate market for turbogenerators
(STAL). The STAL-turbine was a reaction type turbine, in and Asea had, in reality, only one customer for these
which the steam expands in radial direction from the steam generators, its own daughter company STAL.
inlet through two counter rotating disks Each disk was
directly coupled to a generator rotor, so this concept
implied that two identical generators shared the turbine
power. For somewhat larger outputs, axial flow turbines
1-4244-1344-3/07/$25.00 ©2007 IEEE 109
 divided in a few variants. Many manufacturers used axial
MVA cooling (figure 3a). The rotor conductors are made from
hollow copper and the hydrogen enters into these
600 conductors through special openings at both end-regions
and is discharged to the airgap through radial holes in the
International conductors and the slot wedges in the axially central part
development of the rotor. A variant of the axial cooling is the so-called
sub-slot system, which also could be referred to as radial
cooling (figure 3b).
400
Hydropower generators A very different approach to provide the rotor winding
with hydrogen for cooling is the gap-pickup principle. In
Turbogenerators this case, the slot wedges are provided with inlet and outlet
holes so that gas could be taken from the airgap and forced
through channels in the conductors before it is discharged
back to the airgap (figure 3c).
200

Asea

1920 1940 1960

Figure 2. Increase in turbogenerator and hydro-
power generator ratings for Asea and internationally

II. GENERATOR DESIGN

The rotor is usually the bottleneck in a turbogenerator,

especially in two-pole machines, and the design of such a b c
rotors is very crucial. The rotor diameter is limited by the
centrifugal stresses and the strength of the available Figure 3. Cross section of hydrogen-cooled turbo-
materials. The active length is limited by rotor dynamic generator rotor slots with a) axial cooling, b) subslot
properties. If a rotor becomes too slim, it would be very cooling, and c) gap-pickup cooling
difficult to avoid severe resonance vibrations induced at or Common for all these rotor cooling concepts was a fairly
near critical speeds. Furthermore, the total flux has to pass complicated manufacturing process, but also the necessity
the rotor centre and it is important to avoid that this region to build the generators as pressure vessels with efficient
becomes magnetically saturated. Cooling is always very sealings, and external systems for hydrogen supply and
essential for all types of electrical machines and it became control, required much attention too.
more critical when the machine size increased. Air-cooling
was clearly insufficient for large turbogenerators and the
introduction of more effective cooling has been the main Manufacturing of large turbogenerators was an important
road towards higher ratings. The chosen cooling concepts and also prestigious industrial activity during the sixties
have also had a big impact on the generator design in many and seventies. In those years, there were around 20
other respects and were a factor, which differentiated the companies, most of them in Europe, which more or less
manufacturers from each other. The three main parts independently developed such large generators. Most
requiring cooling were the stator core, the stator winding industrialized countries had been much more dependent on
and the rotor winding. In the early 1960’s, most fossil fuels for electric power production than Sweden. Oil
manufacturers used directly water-cooled stator windings and coal fired power plants generated most of the worlds
for large turbogenerators and directly hydrogen-cooled electricity and the power plants had become bigger and
rotor windings. Asea had introduced hydrogen for indirect bigger. To have fewer, but larger units in each plant was
cooling of turbogenerators above 50 MVA. The inner parts cost-effective and therefore, it had been a pressure on
were in principle similar to those in the air-cooled development of very large steam turbines and
generators with radial slot rotors, but they were contained turbogenerators. Asea had not been subject to this and was
in a tight pressure vessel. clearly behind its important competitors in this field.

Some different concepts were developed for direct Direct water-cooling of stator windings represented state of
hydrogen-cooling of the rotor windings and, in several the art in the late 1950’s and had received general
respects, these concepts characterized the whole generator acceptance as a very efficient solution. Therefore, it was a
natural question whether it would be advantageous to also
design. Two basic concepts could be identified, axial
use this method for cooling rotor windings. It was easy to
cooling and gap-pickup cooling, but both of these can then
figure out that theoretically water-cooling was superior to

110
all other cooling methods, but the practical problems in turbine/generator sets. Asea received an order for a 750
providing rotating parts with cooling water caused MWe BWR for Ringhals 1, while the turbines and
hesitation. Nevertheless, several manufacturers started generators should be supplied by English Electric.
studies, including experiments, regarding water-cooled Westinghouse received the order for an 800 MWe PWR
rotor windings. GE reports it had such a program from for Ringhals 2. STAL was chosen as supplier of the two
1957 until 1963, but abandoned it in order to focus on turbines with generators from Asea. These generators were
development of gap-pickup cooling. Other companies rated 504 MVA each. Less than a year later, Sydkraft,
pursued the water-cooling much further, notably BBC and which was the leading partner in OKG, placed an order for
Siemens but also the USSR manufacturer Electrosila, even a 600 MW unit with generator from Asea. Asea had thus,
if directly hydrogen-cooled rotors remained their main in addition to the Ringhals generators, an order for a 710
concept. The first turbogenerator with water-cooled rotor MVA turbogenerator, an enormous challenge taking into
was installed for regular operation in USSR in 1959. account that experiences from operation were still limited
to turbogenerators below 75MVA. In 1971, Asea received
firm orders from Vattenfall for four 577 MVA generators
III. THE NUCLEAR POWER PROGRAM and options for another four. Even Finland needed more
electricity and had also decided on using nuclear power.
The electric power supply in Sweden was almost entirely The private Finnish power company, Teollisuuden Voima
based on hydropower until the mid 1960’s. The power Oy (TVO), planned a power plant located in Olkiluoto at
consumption had shown a steady increase, since the end of the Finnish west coast. The contract for TVO 1 was signed
World War II, and it was projected to continue to rise. in 1973 and for TVO 2 in 1974. [42] STAL was chosen as
Many waterfalls had been harnessed and it had become the turbine supplier including generators from Asea. The
evident that the remaining hydropower resources would be generators should be rated 825 MVA each, thus the largest
insufficient for the future demand of electricity. In ever designed and built by Asea. [1]
addition, environmentalists had begun campaigning for the
preservation of the remaining rivers. It was hence
necessary to start developing other power sources. Sweden IV. STRATEGIC DECISIONS
didn’t have any fossil fuels and it was therefore natural that
the Swedish government initiated research aiming at the
development of nuclear power. The initial concept was a STAL’s old turbine concept had reached the end of the
heavy water moderated reactor with natural uranium as road. The combined radial-axial flow turbines could not
fuel. The government had given preference to this concept handle the large steam flow from really big nuclear
because of Sweden’s own uranium resources. reactors; in addition, the experience from the thermal
power plant in Stenungsund was discouraging. Therefore,
In parallel with the government’s efforts, some private and the company had started to design its own axial turbine.
municipal Swedish power companies formed a consortium, The order from Vattenfall for Ringhals 2 in July 1968 was,
OKG, for building a nuclear power plant, preferably with a in principle, based on this new design, but the matter was
not finally settled. In view of the problems in Stenungsund,
light-water reactor. In July 1965, OKG placed a turnkey
Vattenfall required that Stal-Laval should acquire a licence
order with Asea for a 440 MWe nuclear power plant with a
on an existing design. Therefore, a licence agreement was
BWR reactor. Many alternatives had been studied before
signed with BBC in April 1969 for steam turbines larger
this decision was reached, and Asea had even negotiated a
than 200 MW.
licence agreement with GE, but this was never signed.
Asea decided instead to develop its own light-water A very important question is now: “Why didn’t Asea also
reactor, which must be considered as a very brave and take a licence for the corresponding generators?” The
visionary step. Asea happened to be the only company in increase in size was the same. The generators also required
the world that developed light water reactors without new design concepts. The company had no experience with
licence from GE or Westinghouse. The steam turbine was really large turbogenerators. Looking from the outside, it
ordered from STAL. It was a double rotation radial/axial seemed like the prerequisites were more or less the same
turbine of the company’s traditional type, but much larger for the generators as for the steam turbines. Asea had a
than earlier units. The two generators were rated 271 MVA reputation as a successful supplier of generators, let be
each, by far the largest turbogenerators Asea had received mainly for hydropower, but was confident that it was also
order for at that time. It was decided to use the hydrogen- capable of developing large turbogenerators. Electrical
cooled design for these generators, but with directly water- machines of all kinds were core business for the company
cooled stator windings. The rotor windings were indirectly and acquiring licences had never been part of the strategy.
Therefore, according to well-informed sources, the
hydrogen-cooled.
alternative to take a licence also for the generators was
never investigated or seriously considered. Looking at
The State Power Board, Vattenfall had, during a few years,
Asea’s history, it is evident that the company had a long
looked into the possibilities to build a nuclear power plant tradition of developing the necessary technology in-house.
with a light-water reactor. The private power company
OKG’s decision to build Oskarshamn 1, pushed the
government and Vattenfall to go ahead and plan a large The single most important technical decision for the
nuclear power plant at Ringhals at the Swedish west coast. development of the large turbogenerators was to use direct
Orders for two units were placed in July 1968. Each unit water-cooling not only in the stator windings but also in
should consist of one reactor and two parallel the rotors. This was different from what other

111
 manufacturers used to do and it had a profound impact on of the 1960’s, BBC and KWU built one 2-pole generator
the entire concept and the course of events that followed. each, and Electrosila also built a few 2-pole generators
with this type of cooling. Later, during the 1970’s, both
BBC and KWU also built some very large 4-pole
Direct water-cooling results in lighter and more compact generators with water-cooled rotors.
machines, which also are potentially more cost effective.
The drawback is that it is complicated to have the water The electric power consumption in Sweden increased year
circulating directly through the windings and other active after year during the 1960’s at a rate of 5 – 10 % annually.
parts of a machine. This cooling principle is therefore only Prognoses made indicated a need for around 20 large
used when necessary, mainly for very large machines. As nuclear reactors towards the end of the 1980’s. The
mentioned earlier, in case of large turbogenerators, the generator size had also grown and there were no reasons to
common solution was to have the stator winding directly believe that it would stop growing. Turbogenerators in the
water-cooled while the stator core, the rotor winding and 1000 – 2000 MVA size were anticipated. For such large
other parts were cooled by hydrogen. Asea decided to generators direct water-cooling of both stator and rotor was
avoid the hydrogen and apply direct water-cooling even in considered a necessity, at least within Asea. Therefore,
the rotor. A cross section of the rotor slot is shown in Asea was of the opinion that by developing generators
figure 4. with water-cooled rotors, the intermediate step with
directly hydrogen-cooled rotors could be omitted. This was
probably the most relevant and also most important reason
for the decision.
Managers at both executive and operative levels did not
question the direct water-cooling. On the contrary, it was
almost a policy to prioritize concepts, which would put
Asea in the technical forefront. As a conclusion, the
following reasons for choosing directly water-cooled rotors
have been identified:
• Water is the most efficient cooling medium
resulting in more compact and, for larger
units, more cost effective machines.
• Water-cooling is also applicable for very
large generators expected in the future when
hydrogen-cooling would be insufficient.
• The company had started to use water-cooled
rotors for salient pole machines, so this
technology was already familiar to the
organization and several synergies could be
Figure 4. Cross section of water-cooled turbo-generator expected.
rotor slot with stainless steel cooling tubes. • Hydrogen-cooling is not a realistic option for
hydropower generators, so by choosing direct
In 1964, Asea had received orders for two completely
water-cooling, it would be enough to develop
water-cooled machines from Vattenfall, a 125 MVAr
synchronous condenser and a 225 MVA hydropower only one technology.
generator.[2] The turbogenerators ordered in 1965 for • A few other leading manufacturers were also
OKG had only water-cooled stator windings, but the developing generators with this type of
intentions from the management were clear according to cooling.
the following short quote from a meeting in December
1966: “All efforts shall be made to receive an order as • It was possible to avoid costly development
soon as possible for a water-cooled turbogenerator, of an intermediate step with direct hydrogen
preferably for a peak power plant or a large gas turbine.” cooling.
The second part of the sentence indicates that the risk to go
• It was an advantage to avoid hydrogen due to
directly for a nuclear power application was considered too
the explosion risk, especially in nuclear
great. Asea’s next completely water-cooled machine was,
plants with sophisticated ventilation systems.
in spite of what has been said above, also a salient pole
machine; a 345 MVA, 900 rpm synchronous condenser • The stator housing did not have to be a
ordered by American Electric Power Company (AEP) for a pressure vessel with hydrogen sealings
transformer station in Dumont, Indiana [3]. around shaft ends, terminals etc.
• No external hydrogen system was required.
Asea was not the only manufacturer working with directly
water-cooled rotors. Some others also developed and built • The concept represented state-of-art, which
a few with water-cooled rotors. Manufacturers already emphasized Asea‘s high-tech profile and this
mentioned were BBC, KWU (Siemens + AEG) and was preferred by the management.
Electrosila. During the period of interest, the second half

112
 V. TWO GENERATIONS OF GENERATORS VI. COOPERATION

Asea built, during the 1970’s, 15 turbogenerators with During the most critical development period, the end of the
directly water-cooled stators and rotors (table 1). They can 1960’s and the first couple of years in the 70’s, Asea’s
be divided in two generations, notwithstanding the fact that engineers had hardly any direct contacts with external
they all were subject for continuous development and had experts and competitors. The only forum for
different ratings. The design of the first generation of communication with colleagues from leading competitor
represented a radical step in turbogenerator development companies was Cigré’s study committee for rotating
and it is not surprising that there was room for machines. It can be concluded that the input of external
improvements. There was first an almost fundamentalistic knowledge was limited. This was probably not due to an
approach to the use of water-cooling for all components, underestimation of the need, at least not from the engineers
even when this was neither technically nor economically concerned, but a lack of tradition. These engineers turned
the best alternative. to specialists in Asea’s Central Laboratory for help with
certain problems as they had usually done. Two important
Table 1. Asea’s originally delivered directly water-cooled factors could be part of the explanation why there was so
turbogenerators. little input of external know-how. One was that the
organization was overloaded with all the large orders and
Plant No. of Delivery Power Voltage Exciter simply did not have time for any outlook. Another was that
units years [MVA] [kV] Asea did not have sufficient experience from building
large turbogenerators to be able to approach the leading
Aroskraft 1 1973 294 17.5 Static manufacturers. You must have interesting information to
trade if you expect to obtain any.
Ringhals 2 1974 506.5 19.5 Static
Barsebäck 2 1975 - 77 710 17.5 Static
VII. MAJOR PROBLEMS
Ringhals 4 1977 - 79 577 21.5 Brushless
3 and 4
The manufacturing of the first water-cooled turbo-
Forsmark 4 1978 - 80 577 21.5 Brushless generators was problematic. The machines were
1 and 2 complicated. The workshop faced a lot of difficulties and
Olkiluoto, 2 1978 - 80 825 20 Brushless the operations took much longer than expected. The costs
Finland became very high. The risk for water leakages had been
discussed as a possibility; therefore it was no surprise
Some parts could preferably be air-cooled in a more when the first leakage was detected in the autumn of 1972.
traditional way. In addition, a number of faults had This was the first in a long series of leakages, which led to
occurred during manufacturing, testing and operation of a number of design modifications. Except for problems
the first generators. All this led to development and like these, the results from the performance tests of the
implementation of many new solutions in later generators first generators showed good agreement with predicted
and they became therefore considered a second generation. values, and losses as well as temperature rises met the
The most obvious differences between the first and the guarantees.
second generation were the stator core cooling, the cooling
circuits in the rotor and the excitation system. Figure 5
The development, manufacturing and operation of the
shows a sketch of a generator belonging to the first
GTD generators initially created many problems, both
generation.
technical and commercial. Many could probably have been
avoided through a slower development pace and more
comprehensive prototype tests, but several problems were
shared with other, even larger manufacturers. The 1960’s
and 70’s constituted a learning period for the generator
industry and the knowledge increased partly through some
generic failures, some of them very spectacular.

7.1 Water leakages in cooling tubes

The first water-cooled rotors suffered repeatedly from
water leakages, first in insulating hoses, but most of them
from small cracks in the cooling tubes in the end section of
the rotors. Analyses of different cracks showed that they
were usually caused by mechanical fatigue. The tubes were
subject to both rotational speed and start-stop frequent
deflections that initiated and propagated cracks. Asea
Figure 5. Directly water-cooled turbogenerator rated 710 modified the design of these cooling tubes in a number of
MVA, 3000 rpm steps increasing the flexibility and reducing the dynamic

113
 stresses to a safe level. The photo in figure 6 shows such
flexible water inlets in rotor coils.

Figure 6. Flexible water inlets in rotor coils.

7.2 Stress corrosion in retaining rings Figure 7. Barsebäck generator hall after retaining ring
One of the units in Barsebäck was running with full load in fracture in April 1979.
the morning hours on Good Friday April 13, 1979. The Many manufacturers of large turbogenerators used the
operators on duty had, during 15 minutes, noticed a slight same retaining ring material and Asea was not the only
increase of the vibration level on the slipring side turbogenerator manufacturer which experienced a retaining
generator bearing, when the vibrations suddenly increased ring failure.
drastically. The unit tripped and a fire alarm was received
from the turbine hall. An inspection could soon verify what
had happened. It was a matter of a retaining ring explosion. A study presented at an EPRI workshop on “generator
The slipring-side retaining ring had broken into three retaining rings” in October of 1982 reported 38 fractured
pieces, which were thrown out through the stator end- rings [4]. The matter of cracked retaining rings had, up till
winding and the generator end-cover. One of the heavy the Barsebäck accident, not received much public attention
pieces hit the pedestal bearing so lubrication oil was in the industry. The manufacturers tended to keep most
sprayed around. The short-circuit of the winding caused information to themselves. Since then, a replacement
arcing that ignited the oil and created the fire. The material has been developed, which is stress corrosion
investigations would very soon focus on the reasons for the resistant in water and humid atmosphere.
retaining ring failure. Figure 7 shows the machine hall with
the destroyed generator.
7.3 Cracks in rotor bodies
The examination of the fractured surfaces revealed a Cracks in electrical machine rotors can be disastrous,
primary crack caused by stress corrosion and secondary especially in the case of large machines and high-speed
ductile fractures due to sudden overload. The retaining machines. Asea experienced some serious rotor body
rings were made of a special, high strength, non-magnetic, cracks in the generators supplied to TVO in Finland, a
austenitic steel and it was known that it could be sensitive situation which required very special measures before it
to stress corrosion if it was exposed to water in was solved. The solution involved the use of advanced and
combination with high stresses. The generator had been in partly new theoretical tools as well as methods for
continuous operation for almost one year when the failure monitoring and inspection.
occurred. No water leakages had been detected, but were
nevertheless a matter for further investigations. The
In connection with a minor repair, inspections revealed
metallographic analyses indicated that the stress corrosion
cracks located at the bottom of winding slots right at the
crack had grown over 6 – 9 months and the area where the
end of a rotor, as shown in figure 8. A method was quickly
crack started, on the inside of the ring, was not ventilated.
developed for ultrasonic crack inspection from the rotor
A non-detectable micro leakage could have moistened the
surface. The inspections showed similar cracks in the rest
insulation material in contact with the ring.
of the teeth and also at a rotor still in use in the power
plant. Both rotors had been in operation for a few
thousands hours and had been subject to more than 100
start-stop cycles. It was decided to repair the faulty rotors
simply by removing the cracked zone through machining
and modifying the design to get rid of all stress-rising
notches etc. This shortened the active length by less than
three percent and would not reduce the generator
114
performance. The repairs were quite time-consuming. In note that the new rotors, which were installed later in 1981,
the mean time, one of the TVO units could be kept in have performed without any problems. The design was
operation with a third rotor. The same type of cracks could, improved at a number of points resulting in much higher
however, be expected to occur in this rotor and it was safety factors with respect to all types of fatigue stresses.
therefore important to carry out ultrasonic inspections at The cracks in the original rotors had practically nothing to
regular intervals, and to carefully monitor the rotor do with the water-cooling of the rotors, but rather with
vibrations. traditional extrapolation difficulties.

VIII. OPERATION RECORDS

It is evident, from what has been written that the first

generators as well as those in Olkiluoto had a difficult time
with serious teething problems. The second generation,
which comprises the eight machines for Ringhals 3 and 4
and for Forsmark 1 and 2, has performed reliably from the
very beginning. The others have done so after they have
been modified or provided with new rotors. This means
that the operation records from the early 1980’s up until
now have been very good. This is proved by statistics but
perhaps more important by the fact that the power
companies have chosen the same technology and supplier
also in the case of new, machines upgraded to higher
outputs. The diagram in figure 9a shows how the
unavailability for the water-cooled GTD generators has
varied over the years. The result has become very good,
Figure 8. Fatigue cracks initiated at the slot bottom at the also in comparison with other large turbogenerators as
rotor end. The cracks propagated slowly and turned shown in the availability diagram in figure 9b.
radially inwards to the rotor centre.
Why had the cracks in the TVO rotors occurred? It was
obvious that they had started in a sharp notch, but no
tensile stresses had been anticipated right there.
Theoretical and experimental investigations revealed that
this was wrong. A combination of low and high frequent
dynamic stresses had caused the cracks. Fracture
mechanics was used to calculate the threshold crack depth,
the propagation growth rate, and the critical, instable crack
size. The stress pattern was complicated but the analysis
indicated that it would take in the order of 10 months for a
crack to propagate to critical size, and this size was slightly
more than half of the rotor cross section. The rotor Figure 9. a) Unavailability for Asea/ABB’s water-cooled
stiffness would change significantly, long before a crack turbogenerators, b) accumulated generator unavailability
could become critical, so accurate vibration monitoring in nuclear power plants during 1980 – 2000.
should prevent a dangerous situation from occurring as
long as the old rotors were used [5].
IX. CONCLUSIONS
The turbogenerators in Olkiluoto were not the first that had The development of the water-cooled generators must be
this kind of problem. Both British and French considered as a major technical achievement for a country
turbogenerators had been subject to similar or even more like Sweden. The concept was at the beginning seen as
severe rotor cracks. However, comprehensive daring and pioneering. It was different from what was
investigations were started in order to find a correct usual in the industry and some viewed it as too risky.
explanation for the initiation and propagation of the Many problems occurred and the these generators were, for
discovered cracks. Extensive know-how was built up, the a number of years, not only questioned but even regarded
rotor designs were improved and new methods for as a serious failure. Extensive development efforts solved
inspections and monitoring were developed. It can be the problems and the generators have for decades had a
argued that Asea had been too ignorant before, but it seems very good reputation. They have during the last 25 years
as if other manufacturers had acted in similar ways. generated around 30 percent of Sweden’s electric power. A
corresponding figure for Finland is 20 percent. The
concept has been maintained and the technology is in no
The vibration monitoring, introduced at TVO during this respect obsolete.
critical period, has later become a standard praxis for large
power plant turbines and generators. It is of importance to

115
 Some explicit conclusions are: [1] Sture Eriksson, Electrical Machine Development
– a study of four different machine types from a
- Asea chose the fully water-cooled concept:
Swedish perspective pp. 211-326, ISBN 978-91-
- to avoid development of directly hydrogen-
7178-617-3, PhD thesis at the Royal Institute of
cooled rotors as an intermediate step,
Technology, Stockholm, 2007
- because it was the most efficient cooling
method, [2] Tengstrand, Claes and Rönnevig, Carl Direct
- and also due to synergies with large salient pole cooling of water-wheel generators, influence on
synchronous machines. dimensions and generator parameters pp. 1-8,
Cigrè, Report 11-03, Paris, 1968
- Asea’s immediate challenge Asea became too [3] Landhult, Hans and Nordberg, Birger 345 MVA
large due to the drastic increase in machine size in fully water-cooled synchronous condenser for
combination with simultaneous orders for Dumont station, Part II. Design, Construction and
generators with different ratings. Testing pp.2765-
- Asea underestimated the difficulties when they 2772, IEEE Transactions on Power Apparatus and
decided to make most of the development on Systems, Vol. PAS-90 No 6, Nov/Dec 1971
customer orders. Manufacturing and testing of a [4] Viswanathan R. Retaining ring failures Paper
full-size prototype as basis for the order design presented at EPRI Workshop on Generator
had eliminated several problems and saved a lot retaining Rings, Paulo Alto, Oct.1982
of costs. [5] Carlsson, Janne, Eriksson, Sture and Sundstrand,
Arne Fatigue cracks in electric generator rotors –
- Mutual exchange of technical know-how with a case study pp. 989-1004, Proceedings of
other manufacturers increased substantially, as International Conference on Fatigue Thresholds,
expected, when Asea had gained certain Stockholm, June 1981
experience

Sture Eriksson received his M.Sc. E.E. from Chalmers

No generators of this particular design have been built for University of Technology in Gothenburg and his PhD
other countries than Sweden and Finland. Very little of the degree from the Royal Institute of Technology,
technology has been taken over by others. This could Stockholm. He worked for Asea/ABB from 1963 until
indicate that the concept is not competitive enough, but on 2000, mainly with management of electrical machine
the other hand, it has been preferred in open competition development. He was appointed adjunct professor of
for recently upgraded units. The cost-effectiveness is electrical machines at the Royal Institute of Technology
therefore probably not much different from other designs in 1987. At present, he works as R&D advisor for
Swedish national research programs related to electrical
of large 2-pole turbogenerators. Each manufacturer tends
machines and drive systems..Member of IEEE.
to hold on to their existing designs. The explanation is
more likely to be found in the market and industry
structure. The market for large steam turbines and
generators had just started to shrink when Asea had solved
the technical problems. There were simply no possibilities
for Asea to sell large turbogenerators when the
international situation was characterized by large
overcapacity. The conditions became different when Asea
became part of ABB. This company had access to the
global market and could have decided to include the
completely water-cooled generators in deliveries of large
steam power units, if advantageous. This has not been
done, the concept with directly hydrogen-cooled rotors,
emanating from BBC, has instead been chosen. Why, is it
better? This question is difficult to answer. The uniqueness
of the Asea generators is the water-cooled rotors and
absence of hydrogen. The management of ABB’s, and later
Alstom’s, turbogenerator operations located in
Switzerland, has obviously come to the conclusion that the
customers prefer more conventional, hydrogen-cooled
generators and the possible incentive for a shift of concept
will not outweigh the risks. It is, of course, difficult to
motivate radical design changes when the market for really
large turbogenerators has become limited.

.
X. REFERENCES

116