You are on page 1of 14

Proceedings of ASME Turbo Expo 2018

Turbomachinery Technical Conference and Exposition


GT2018
June 11-15, 2018, Oslo, Norway

GT2018-75313

KONGSBERG GAS TURBINES THROUGH FIFTY YEARS – A REVIEW OF THE PRODUCTS


AND THE HISTORY

Tore Naess
Consultant
Kongsberg, Norway

ABSTRACT New arms technology based on electronics was emerging and


In 1964 Kongsberg Våpenfabrikk AS decided to develop a small the administration therefore actively engaged in NATO projects
gas turbine for power generation, primarily for stand-by and to secure offset deals, both for conventional artillery with
emergency power. The engine was called the KG2 and had a advanced guidance systems and advanced missiles. Volume
unique all radial rotor design which was to become the trade production technology had also been developed through the
mark for the later Kongsberg designs. The onset of the oil Volvo agreement in 1957 and manufacturing of the new AG3
exploration in the Norwegian sector of the North Sea in the rifle for the Norwegian Army. All this had given the company
1970’s gave the new business an opportunity to qualify for great know-how and capacity within electromechanical
continuous drive applications and to expand into the technology and design as well as advanced manufacturing.
international oil- and gas industry. In the following years a License manufacturing and offset deals were not considered
larger engine, the KG5, was launched and a third engine satisfactory in the long run and the Board of Directors was
program was initiated, but never completed. The gas turbine looking for a product of their own. The challenge given the
know-how that was established in Kongsberg in these years administration from the Board of Directors was to develop a
was of great significance to the overall Norwegian gas turbine product which would:
competence environment and was a deciding factor when  Utilize the electromechanical know-how and
Dresser-Rand first partnered with and later, in 1987, acquired capabilities
the business. Under the new ownership the company became  Be so genuine that it would secure a part of the world
able to offer compressor- and power generation packages market if successful
based on large aero-derivative gas turbines and it was soon  Be a part of a “young” product technology in order to
recognized as a significant supplier, both nationally and ensure a growing market and a product life which
internationally. The present paper provides a review of some of would give proper payback
the unique design features of the KG series of engines as well
as some of the typical applications. It also describes the 1. THE BEGINNING (1960-1968)
transformation of the company from a small industrial gas 1.1 The design Phase
turbine supplier to the recognized supplier of large, Beginning around 1960 followed a period of strategic
compressor- and power generation packages for the oil and gas discussions, market assessments and discussions with possible
industry. development partners. It should however; take almost four
years before the final go ahead was given. The 10 th of April
INTRODUCTION 1964 the Board of Directors formally approved the start-up of
Kongsberg Våpenfabrikk AS (KV), located in Kongsberg- the COPGAT, (Cargo Oil Pump GAs Turbine), later to be called
Norway and founded in 1814, was originally a small arms KG2 [1]. This is considered the official birth date of Kongsberg
manufacturer wholly owned by the Norwegian Government. gas turbines. Although not directly connected, it is interesting to
Rifles of own design, among others the famous Krag- note that Ægidius Elling, the first one to make a gas turbine
Jørgensen, and license manufacturing of larger artillery were with excess power, did much of his development work at KV
the main products up until the mid-1950s, but then the orders some 60 years before [2,3].
for conventional arms were not coming as regular as before.

1 Copyright © 2018 ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 09/05/2018 Terms of Use: http://www.asme.org/about-asme/terms-of-use


Under the leadership of R. Jan Mowill a project team was
established and went to Los Angeles and did the main part of
the design work with the layout in four months. The gas turbine
option satisfied all the previously established criteria. It
represented an advanced mechanical product with a market
growth potential. Another factor was that KV through its
military products had contacts within Boeing and Pratt
&Whitney Aircraft, (PWA).

The design philosophy was aiming at a robust and simple


design which was considered significant for market acceptance
[4]. A centrifugal compressor was then a natural choice, but it
took longer to decide upon combining it with a radial inflow
turbine. The contacts with PWA and their consultant on radial
turbines, Mr. Homer J. Wood, played a significant role in the
decision. PWA also assisted in the development of the
combustor, but the concept and the layout work was performed
by the team from KV. The detail design of the parallel shaft
reduction gear was successfully left to MAAG Gear AG. in
Switzerland. In a little over 3 years, two prototypes were
designed, manufactured and initially tested in Kongsberg. FIGURE 1. KG2 GAS TURBINE LAY-OUT

The first sale was made in 1968 to the Norwegian Water & 1. Speed/vibration pick ups
Electricity Board (NVE) for the island of Røst in Lofoten, 2. Hirth coupling
almost before the initial testing was completed. The normal 3. Heat shield
electricity supply to Røst was an underwater cable to the 4. Compressor diffuser
nearest island and this was frequently damaged in storms. They 5. Nozzle guide vanes
6. Volute (spiral housing)
were therefore in need of a stand-by generator set which could
7. Fuel nozzles (6 off)
be brought online quickly. 8. Exhaust diffuser

1.2 Unique technology KG2 Key Data


When KV decided to enter the gas turbine business in 1964, the  Single shaft design
market was dominated by axial machines, and radial  Single stage centrifugal compressor pressure ratio: 3.25
technology was associated with turbochargers. However, the  Single stage radial inflow turbine
compactness, simplicity and apparent robustness of the “all  Rotor speed: 17100 rpm
radial” approach were considered important parameters. A  Thermal Efficiency: ~16 %
single stage centrifugal compressor mounted “back-to-back”  ISO power: 1200 kW
with a single stage radial in-flow turbine, both having blades  Single tangential can combustor (6 nozzles)
integral with the disc, achieved all three targets. At the time
there was no experience available for an engine of this concept The overhung rotor caused quite a lot of concern in the
and size. design phase since it would be operating above twice the
critical speed. A special bearing test rig was therefore designed
A further challenge was presented by the design team’s and built in order to test various bearing parameters. Since the
decision to go for an overhung rotor suspension. This maritime market was considered essential, the rig was built to
cantilevering however, retained all the bearings in the cold end simulate various effects of wave movements under operation
of the engine and improved mechanical integrity. The single- and sleeve bearings were considered essential due to the long
shaft design avoided the additional cost, size and weight life requirements. A set of proprietary tilting pad journal
inherent with a two-shaft machine and resulted in an engine bearings were developed in the bearing rig and they showed
complete with integral reduction gear with an overall length of exceptional stiffness and damping characteristics. Tilting pad
less than two meters, Fig. 1. bearings were also selected for the thrust bearing due to its load
capability and ability to adjust to varying load conditions.

The manufacturing technology was such in the early 1960s


that both the compressor and turbine stage had to be split in two
pieces or “wheels”, Fig. 2.

2 Copyright © 2018 ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 09/05/2018 Terms of Use: http://www.asme.org/about-asme/terms-of-use


and application engineers with naval engineering background
that had been recruited started looking for other opportunities in
the maritime sector. The first attempt was the TURBOSAFE™
application [6].

TURBOSAFE™ was a KG2 emergency/stand-by generator


set in a container-like enclosure, complete with controls. The
TURBOSAFE™ container was light and essentially vibration
free, so it could be placed on top of the ship superstructure and
be a back-up unit if one of the main auxiliary engines failed.
During the first two years more than twenty engines were sold
and installed. The TURBOSAFE™ concept was later to be
combined with the TURBINERT™ concept [7]. This was
producing inert gas from the exhaust gas in an afterburner in
order to prevent tank explosions and ventilating the tanks
afterwards with bleed air from the gas turbine compressor.
FIGURE 2. KG2 ROTOR (PARTLY CUT-AWAY)

1. Forward journal bearing/thrust bearing The team had not forgotten its first customer NVE either. In
2. Aft journal bearing 1970 there were several small fishing communities in Northern
3. Labyrinth seal Norway which had inadequate power supply. The introduction
4. Compressor inducer of refrigeration plants made the situation difficult and
5. Compressor impeller potentially very costly in case of a power failure. NVE wanted
6. Turbine impeller a unit that they could have stand-by for several of these
7. Turbine exducer communities and the first mobile unit, Fig. 3 was developed
and ready in 1971. NVE bought three of these initially and later
The compressor inducer was investment cast and the
two more to be placed in fishing communities in Northern
impeller section with straight radial blades made from a
Norway as standby engines.
forging, both in a precipitation hardening stainless steel. The
turbine impeller was made from a Nimonic 90 forging whereas
the exducer was investment cast in 713LC. The two sections
were held together with a patented pre-stretched center
bolt/sleeve arrangement [5]. Differences in thermal growth
were allowed by a Hirth coupling (radial tooth coupling). The
turbine wheel and the exducer was also centered with a Hirth
coupling and held together with a sleeve arrangement. The tip
speed of the turbine wheel was in the excess of 500 m/s which
required a material with an optimum combination of creep and
tensile strength.

The compressor diffuser had three stages of precision cast


vanes. The combustor was of the reverse flow can type with a
center tube to achieve even temperature distribution. The fuel
nozzles could be pure liquid or gas, but a dual fuel version was
early introduced. The combustor was tangentially mounted to
the centerline and hot gas from the combustor was directed into
the nozzle guide vanes via a scroll or volute. The nozzle guide FIGURE 3. FIRST KG2 MOBILE UNIT
vanes were made from a precision cast cobalt alloy (X40) and
were un-cooled.
3. THE GROWTH PERIOD (1971-1975)
2. THE ESTABLISHING PERIOD (1968-1971) 3.1 Qualifying for oil & gas
The original market study which concluded in a potential The first half of the 1970s was a period with high activity and
market for oil pumping gas turbines, had presumed an oil many significant milestones. The discovery of oil by Phillips
tanker size in the order of 40,000 tons deadweight (TDW). The Petroleum on the Ekofisk field in the North Sea in 1969 meant
tanker size however, had grown steadily since the Suez crisis that offshore oil & gas was to become a significant market in
and was now in the order of 300,000 TDW. Cargo oil pumping the years to come. It was a difficult market to get access to for a
was no longer an option. It was therefore natural that the sales newcomer in the market however.

3 Copyright © 2018 ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 09/05/2018 Terms of Use: http://www.asme.org/about-asme/terms-of-use


KG2 was probably not considered initially when Phillips
Petroleum was evaluating power generation for the Ekofisk
field. Nevertheless, already in 1971 KV had orders for ten KG2
engines for base load power generation on the Ekofisk field.
The number of engines installed grew steadily and by 1977
Phillips Petroleum had about 60 KG2 engines in total on the
various platforms in the Ekofisk complex.

It did not take long before the base load duty in the rough
North Sea environment revealed the first technical challenges.
In the most hectic period more than 500 change notices were
registered in one year and these had to be implemented in the
whole cycle from parts on order to engines in the field.
Engineers would work day and night trying to understand the
root cause of the various incidents and service people in the
other end would spend weeks offshore in implementing
modifications.

3.2 From KG2 to KG2-3C


In the beginning of this busy build-up period, steps were also
taken to uprate the engine. A change of turbine wheel material
from Nimonic 90 to Waspaloy made it possible both to raise the
turbine speed to 18,000 rpm and increase the turbine inlet
temperature. This gave more power and slightly better
efficiency. The configuration was given the designation KG2- FIGURE 4. THE NEW MOBILE UNIT
3A. The initial turbine wheel in the new material was designed
Up until the early 1980s altogether 86 mobile units had been
for optimum life according to the new material data, but soon it
sold in twenty-two different countries around the world. The
was apparent that there was a turbine blade resonance with the
primary market was in the Middle East region, but generally the
nozzle guide vane passing frequency. The solution was to
units went to remote areas where the electrical infrastructure
increase the blade thickness especially at the inlet tip. This
was missing, but fuel supply and prices were acceptable.
shifted the natural frequency sufficiently above the passing
frequency and it is this design that has survived to this day.
3.4 The KG5 gas turbine
Already before the KG2 was well established in the market, the
In parallel with the uprated turbine wheel there was an
idea of a second and larger gas turbine engine was launched.
attempt to make an improved compressor stage. This
Preliminary studies had been made over several years and these
configuration was called KG2-3B, but testing revealed that the
concluded with a 2-shaft machine with a power of nominally 3
compressor version was not stable and it was never introduced
MW, the KG5.
to the market. A little later however, in 1973, the KG2-3C
version was introduced. It was basically the -3A with some
The “all radial” concept was retained for the gas generator,
minor improvements of mechanical nature. The KG2-3C is the
but it had to be a 2-shaft machine as there was a practical and
version1 that accounts for the majority of the KG2 sales over
economical limit to how large the turbine impeller could be
the years.
manufactured. The radial turbine impeller size was chosen so
that one could use the same forging as for the KG2, but there
3.3 Mobile unit improved
had to be a single axial stage behind it. The overall pressure
The mobile units which were originally developed for NVE had
ratio was increased and a special compressor test rig was built
raised some interest internationally, but the initial design soon
in order to optimize the compressor geometry. The compressor
proved to have a weakness. It had too high axle load on the rear
impeller had to be made in a titanium alloy in order to achieve
axle, requiring a special permit to be towed on the road. The
acceptable centrifugal stresses and the “back to back” with the
next version, which was introduced in 1974, had twin rear axles
turbine feature from the KG2 was abandoned in favor of
and was sturdier in general, Fig. 4.
simpler rotor dynamics. Much of the other mechanical design
and the combustion system were derived from the KG2. The
reduction gear was of a “compound star” epicyclic type, again
made by MAAG, Fig. 5 and 6.

1
See Annex A for data on the various KG2 versions.

4 Copyright © 2018 ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 09/05/2018 Terms of Use: http://www.asme.org/about-asme/terms-of-use


consisted of changing the number of NGVs and increasing the
pre-tightening of the damping joint.

In addition to the delay, the KG5 also had to face


competition from Solar Centaur and Ruston TB5000, which
were both larger in power and more efficient machines. For
these reasons KG5 never became a great commercial success,
altogether 56 engines were sold.

FIGURE 5. KG5 GAS TURBINE LAYOUT

1. Air Inlet
2. Centrifugal Compressor, Inducer
3. Centrifugal Compressor, Impeller
4. Compressor Diffuser
5. Outer Casing
6. Combustor
7. Volute
8. Nozzle Guide Vanes FIGURE 6. THE KG5 GAS TURBINE
9. Radial Inflow Turbine & Exducer
10. Axial Turbine, single stage KG5 Key Data
11. Exhaust Diffuser  2-shaft design
 All radial gas generator
The primary purpose of the KG5 engine development was to o Single stage centrifugal compressor pressure ratio: 6.5
o Radial inflow turbine
have more than one gas turbine model in the market to be able
o Max speed: 17400 rpm
to offer a wider power range. Due to the higher pressure ratio  Single stage axial power turbine
and the 2-shaft configuration it would also have a better  Power turbine speed: 12000 rpm
thermal efficiency than the KG2 and be suitable for mechanical  Thermal efficiency: ~21.5 %
drive. Four engines were sold before the prototype was ready,  ISO power: 3110 kW
but the development testing which started in August 1975,  Single tangential can combustor (6 nozzles)
revealed a serious technical issue; exducer blades would break
off in certain test conditions [8]. It took many nights of testing
and analysis before one was finally able to conclude that there 4. PEAK YEARS FOR KG (1976-1979)
was a blade resonance phenomenon caused by excessive 4.1 Worldwide expansion
centrifugal stiffening of the exducer blade. A new casting tool The years from 1976 to 1979 later proved to the peak years in
and new parts had to be made before the development program terms of number of engines sold and delivered. In 1976
could continue. Kongsberg also for the first time became a packager of a gas
turbine from another company. Garrett AiResearch (now part of
The first ordinary shipment took place in 1977 and the KG5 Honeywell International) had developed an industrial gas
fleet had accumulated more than 100 000 hours when in 1982 turbine based on their successful turbo-prop engine TPE331.
another technical issue revealed itself. During witness testing in The engine was called IE831 and had a nominal rating of 500
Kongsberg one engine experienced a high cycle fatigue failure kW. In the Kongsberg product range it was named KG831.
of a single turbine blade. It took more than a year and two more
similar incidents in the field to determine the root cause and Annual number of engines, KG2 and KG831 combined, was
implement corrective actions [9]. It appeared that during between 100 and 120 in this period. In the first half of 1978
unloading the dampening joint between the turbine impeller orders for 109 engines had been received. Sales and Service
and the exducer would loosen due to thermal transients. This were represented “all over the world” ranging from the now
would allow a resonance with the NGV passing frequency to (since 1974) wholly owned subsidiary NATCO (North
become strong enough to cause failure. The corrective action American Turbine Company) in Houston Texas, via OEM

5 Copyright © 2018 ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 09/05/2018 Terms of Use: http://www.asme.org/about-asme/terms-of-use


agreement with Kobe Steel in the Far East and use of KV sales On the 22nd of April 1977 there was a blow-out on the
offices where they were present, to local agents in virtually Ekofisk-Bravo platform in the North Sea. The oil and gas
“all” other countries with a market potential. There were mixture fortunately did not catch fire, but it was important to
already installations in more than 50 different countries around keep that from happening. Seaway Falcon happened to be
the world and service engineers from Kongsberg were global nearby and was chartered to keep spraying water over the
travelers, well recognized at Oslo Airport. platform. For seven days one could follow Seaway Falcon on
TV in the headline news with the dual water jets spraying water
The high number of engines per year2 put a lot of strain on over the platform, Fig. 8.
the operational capabilities of all disciplines. This was
particularly noticed when the large orders for Saudi Arabia
were executed. There had been some sales in Saudi Arabia
already, both mobile units and some generator sets through a
German contractor, but in 1979 the first order with RSAF,
(Royal Saudi Air Force), as end customer was secured, 30
indoor generator sets. The delivery was completed to
everybody’s satisfaction, which led to successive orders for the
same end customer. The largest single order received came in
1983 with 60 engines. Altogether 128 engines were delivered,
the last twenty as late as 1994. These engines were all running
semi-continuously on liquid fuel.

4.2 Pump sets for fire fighting


Following the tradition of developing solutions for the shipping
industry, a “fire-fighting” unit was developed together with the
pump manufacturer Thune-Eureka for the new fleet of supply- FIGURE 8. THE EKOFISK-BRAVO BLOW-OUT
and support vessels for the offshore oil & gas industry. The unit
consisted of a KG2 driving a centrifugal pump via a clutch The blow-out was capped after seven days and the KG2s
arrangement for easier starting, Fig. 7. kept running all the time. Quite a few new sales were booked
after this incident. Altogether twenty-two pump sets were
delivered over the next years, the last one in 1982.

5. THE KG3 DEVELOPMENT PROJECT (1980-1986)


It had been clear for some time that a new and more efficient
gas turbine to replace the KG2 was needed, but there was a
great concern that the development cost would be difficult to
carry. Finally in 1980 the Board of Directors approved the
project called the KG3 on the expressed condition that KV
should find an industrial partner to share the risk.

A separate project organization was established with Jan


Mowill as leader and staffed with key people from the Gas
Turbine division. As with the original KG2 development,
assistance had been solicited from PWA. The relationship with
PWA had been strengthened in the meantime during the parts
FIGURE 7. KG2 PUMP SET manufacturing program for the F100 engine for the F16
aircraft. A design office was quickly established in the vicinity
The pump would deliver sea-water to jet nozzles on top of of the PWA plant in Florida and a number of experienced PWA
the ship. The first two units were sold to Stolt-Nielsen for designers were assigned to the team. During the stay in Florida
installation on board the combined supply and fire-fighting ship a complete conceptual design was developed and drafted.
M/S Seaway Falcon. Tests showed impressive characteristics.
The water jet could reach 170 m away with a height of more The concept was based on the “all radial” design from the KG2,
than 50 m. The maximum flow was about 4000 m3/hr. but with some distinct differences. [10] There was a bearing
between the compressor and the turbine and there was a multi-
can combustor of a low emissions type, (DLE) [11]. Hot gas
from the combustors was flowing into a double-walled torus
2
See Annex B for KG2 unit sales. [12] instead of a volute, before entering the nozzle guide vanes.

6 Copyright © 2018 ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 09/05/2018 Terms of Use: http://www.asme.org/about-asme/terms-of-use


Also the centrifugal compressor and the radial inflow turbine excessive unbalance at lower speeds. A more serious issue
were both machined from solid forgings of titanium- and appeared when full speed was reached. Turbine blade resonance
nickel-based alloys respectively. The reduction gear was a with the number of nozzle guide vanes with subsequent fatigue
“compound star” epicyclic type like the KG5, again failure destroyed most of the hot section of the engine. This
manufactured by MAAG, Fig. 9 and 10. caused severe delays to the program as new analyses had to be
made and new or modified hardware had to be provided.

There were two main reasons for the many technical


challenges with the KG3 project. First of all the technical
specifications had been set too ambitious. Both pressure ratio
and the corresponding tip speed were far beyond the state of the
art. Especially the ultimate goal of 1300 °C firing temperature
proved to be a mistake. This was based on an assumption that
one would eventually develop a turbine wheel that could
withstand such a temperature.

Although the initial firing temperature was to be much


lower, the other hot section parts had to be designed to meet the
same temperature level, leading to, in some cases like the
nozzle guide vanes, an unnecessary complex and costly design.
FIGURE 9. KG3 GAS TURBINE LAY-OUT The “stretched” design was a strong contributing factor in the
failures that occurred in the test phase.
1. Front journal bearing
2. Compressor
Another mistake was making the turbine wheel in one piece
3. Compressor diffuser
4. Aft journal bearing/thrust bearing instead of two, impeller and exducer, with a damping joint
5. Combustor (6 off) between them, as in the KG2 and KG5. This contributed
6. Tension bolt strongly to the blade resonance failures. The single piece
7. Inlet manifold (torus) turbine wheel proved to have a number of natural frequencies
8. Nozzle guide vane assembly due to its complex geometry. It was virtually impossible to find
a “window” in the Campbell diagram for the nozzle guide vane
passing frequency.

By separating the two as in the KG2 and KG5, the turbine


impeller, which is exposed to the highest excitation force and
has a relatively simple geometry at the inlet, could be designed
to be free of any resonance condition. The excitation forces on
the exducer with its curved, relatively thin blade are much
lower and although there are possible resonances, the excitation
isn’t strong enough to cause failure.

In the end however, the starting point had been so extreme


that the project could not be saved. At the end of the testing in
1986 one had a compressor with acceptable performance, but a
turbine wheel with less than optimal diameter and a more
conventional material than originally presumed. The
combination could theoretically give an overall efficiency of
27.4% and a power of 1370 kW, but this was never proved in
test, Fig. 11.
FIGURE 10. KG3 ROTOR ASSEMBLY BEING MOUNTED

The prototype testing soon ran into problems however.


Initially there was a rotor instability that prevented running up
to full speed. It took a number of tests with modification of
hardware to find the root cause. It appeared that the bending
stiffness of the center structure holding the compressor and
turbine sections together was too low, resulting in a varying and

7 Copyright © 2018 ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 09/05/2018 Terms of Use: http://www.asme.org/about-asme/terms-of-use


DI’s motives for entering into a partnership were two-fold.
The most important reason for DI, primarily being a
compressor manufacturer, was to have a company in Norway
which could help marketing their compressors into the
Norwegian sector of the North Sea. Of secondary importance
was the fact that they had just acquired the IE-990 gas turbine
from Garrett AiResearch.

6.2 DC-990 gas turbine


The engine, renamed DC-990, was a 4.4 MW, two-shaft
machine suitable for compressor drive, Fig. 12, and in order to
increase the production volume they wanted to go into the
power generation market. Combined with the KG3 which was
under development by KV, the plan was to constitute a
competitive small gas turbine power generation business.

FIGURE 11. KG3 PROTOTYPE ENGINE

KG3 Key Data


(First version-as originally specified)
 Single shaft design
 Single stage centrifugal compressor pressure ratio: 9
 Radial inflow turbine
 Rotor speed: 35115 rpm
 Thermal Eff.: 29 %
 ISO Power: 1650 kW
 6 axial can combustors, DLE FIGURE 12. DC-990 GAS TURBINE LAY-OUT

In the partner agreement with Dresser Industries (see below) DC-990 Key Data
they had reserved the right to stop the project if certain  2-shaft design
performance criteria were not met. [13] As described above the  Gas generator
o 2-stage centrifugal compressor pressure ratio: ~6
achieved data were lower than specified. Dresser Industries
o 2-stage axial turbine
decided not to cancel the project right away, but reserved the o 12000-18000 rpm
right to do so later, dependent on how the government financing  Power turbine
was resolved. This was not resolved to their satisfaction and the o 3-stage axial
project was finally shelved in 1987. o Nom. Speed: 7200 rpm
 Thermal Eff. : ~30%
6. CHANGES IN OWNERSHIP (1985-1987)  ISO power: 4400 kW
6.1 Partnering with Dresser Industries  Annular combustor
KV’s administration was under a lot of pressure from the Board
of Directors when the search for a partner for the gas turbine A new package was developed for the DC-990 and the KG5
activities started in 1981/82. Some twenty different companies test cell was modified such that both DC-990 and KG5 could be
were contacted and with some further negotiations were started tested with only minor re-arrangements. There was an
without result. Dresser Industries (DI) had early been pointed immediate success with the DC-990 with several sales in
out as the most interesting candidate based on their newly Europe, mainly for industrial total energy processes or
acquired IE-990 gas turbine (see below) and their license to municipal district heating, and 1986 became the best year ever
package General Electric (GE) aero-derivative engines. for the sale of the DC-990 with twenty-three engines sold in
total from DI in Olean NY, Houston and Kongsberg.
In January 1985 KV and DI signed a letter of intent and the Unfortunately the DC-990 experienced serious technical issues
21st of June the final “Joint Venture Agreement” was signed. in base load operation to such an extent that Dresser-Rand (see
The new company, Kongsberg Dresser Power Inc. (KDP), was below) gave up and sold the product to Volvo Aero in 1997.
a 50/50 joint venture with equal representation in the Board of
Directors. It was legally headquartered in the former NATCO
office in Houston and had a sales office in Singapore, but the
main part was still in Kongsberg.

8 Copyright © 2018 ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 09/05/2018 Terms of Use: http://www.asme.org/about-asme/terms-of-use


6.3 Dresser-Rand Power Inc. 7. DRESSER-RAND (1987-2005)
In August 1986 KV was informed that DI had entered into an 7.1 Transition to large engines
agreement with Ingersoll Rand (IR) about merging of their Ever since the formation of KDP, the joint venture company
respective compressor and gas turbine activities. The new joint had worked hard to win orders using the LM2500 engine from
venture was named Dresser-Rand (D-R). KV’s Board of GE as driver. The Houston branch was first out, but in 1987 it
Directors approved the transfer of the shares in KDP Inc. from was Kongsberg’s turn. The order for a LM2500 generator set
DI to D-R in January 1987. KV’s financial troubles continued for the Veslefrikk platform in the North Sea was received and
however and in the spring of 1987 it was forced into a the customer was Statoil. The contract called for a complete
liquidation phase. [13] The 5th of June 1987 KV signed an generating set with controls, full load tested and delivered in
agreement to sell their shares in KDP to D-R. The name of the twelve months, which was six months shorter than the delivery
company was changed to Dresser-Rand Power Inc. (DRP). time normally seen in the market!

On the product side work had been started to upgrade the A project team was quickly established and as there was no
KG2 with a new compressor stage, Fig. 13. The new direct, previous experience with this size of packages they
compressor was to be based on DI compressor know-how, with could start with a blank sheet of paper. This was utilized to
higher pressure ratio and better efficiency. The new version was come up with some new and innovative solutions based on
called the KG2-3E and had a “backward curved” titanium basic gas turbine knowledge and extensive packaging
impeller in one piece. This design improved the surge margin, experience. Fundamental was employing the “torque tube” base
which had been marginal in the original design. It was also plate that IR in UK had developed [14] instead of a more
equipped with a channel diffuser instead of the original plate conventional multi point skid.
diffuser which was cast in bronze segments and had been less
durable than desired. The torque tube, a central tube along the axis of the
machinery, had several advantages. It is a well-known fact that
the cross section that can take the most torque with the smallest
cross section is a tube. In this case it gave a weight reduction of
almost 25% compared to the multipoint skid. It was also
designed as a 3-point mount and rigid enough to make
alignment independent of the platform deck. It also allowed the
auxiliary systems to be placed along the edges of the baseplate
underneath the base plate deck which allowed for easier
maintenance, Fig. 14.

FIGURE 13. KG2-3E GAS TURBINE ROTOR

KG2-3E Key Data


 Single shaft design
 Single piece centrifugal compressor PR: 4.5
 Radial inflow turbine
 Rotor speed: 18880 rpm
 Thermal Efficiency: ~18.3 %
 ISO power: 1950 kW
 Single tangential can combustor

Otherwise there were no principal changes to the original


design, which had grown from 1200 kW in 1968 to 1950 kW
twenty years later. Most of the KG sales after 1987 have been
with KG2-3E. FIGURE 14. TORQUE TUBE BASEPLATE FEATURES

1. Baseplate
2. Gas Generator Turbine (LM2500/LM2500+)
3. VECTRA Power Turbine

9 Copyright © 2018 ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 09/05/2018 Terms of Use: http://www.asme.org/about-asme/terms-of-use


4. Exhaust Collector was turned back to the oil & gas sector and the LM2500 series
5. Turbine Handling Crane of drivers.
6. Fuel System
7. Turbine Lube Oil System 7.2 The MT250 microturbine
8. Turbine Lube Oil Cooler
9. Lube Oil Vent System
In February 2000 IR, in an agreement with Halliburton, the new
10. Instrumentation Panel owner of Dresser Industries, became the 100% owner of D-R.
11. Water Mist Fire Fighting In practical terms the Kongsberg operation was not affected
much by the change, but for one in this connection interesting
Figure 14 shows the later developed split version where the project; the MT250 microturbine.
driven equipment is mounted on a separate baseplate. The two
parts has precision machined mating flanges on the torque tube, The microturbine market was blooming in the US at that
which allow them to be mounted together after completion. time and IR was determined to take part in it. Northern
Another important feature with the design was the built-in Research and Engineering Corporation (NREC) in Portsmouth
crane for lifting the turbine in and out of the package. This NH, an IR subsidiary, had already developed recuperator
made the handling of the turbine much easier and opposed to technology and a 70 kW microturbine and was now challenged
the conventional A-frame and beam outside the enclosure, the with developing a 250 kW, recuperated gas turbine. IR had
crane was always available and there were no lifting parts that recognized that the KG2 gas turbine temperatures and pressure
could get lost in storage. ratios were a very good match for the recuperated microturbine
of this class and they liked the all radial configuration, as it
The package was delivered in 12 months as agreed, and both would make the adaptation of a recuperator easier.
the project execution and the engineering and manufacturing
capabilities of the organization had been proven to Statoil’s IR wanted Kongsberg to take responsibility for the
satisfaction and this placed the company on the bidders list in mechanical design of the core engine. The aerodynamics and
1988 for the next big project; Sleipner A. To Kongsberg the DLE-combustor would be IR responsibility and this was
standards this was a mega project: two generating sets and five already much developed. The design and analysis work started
compressor sets all driven by LM2500. The contract was won in September 2000 and in record time the initial layout was
and the packages were already delivered to the module yard ready. All the design work was done with 3D modeling, which
when in 1991 Statoil had the infamous accident with the made the analysis work simpler.
concrete platform which collapsed and sank. [15] The project
team, with the assistance from the test and service people, was The design heritage can clearly be seen from the 3D-model
able to turn around and modify the compressor sets temporarily of the rotor, Fig. 15 apart from the size. It was about half the
from gas only to liquid fuel so that they could be tested in the diameter of the KG2! Another distinct difference was the
module that Statoil had placed on a small reef while waiting for turbine wheel, which was investment cast to near net shape in
the new concrete platform to be ready. This secured the one piece. It was expected that the single piece design would
progress for Statoil such that the platform could be brought sustain the much lower excitation forces than had been the case
online as planned. on the KG3.

The Sleipner A project firmly established D-R Kongsberg as


a supplier of power generation and compressor sets for the oil
and gas industry. The next large projects came from Hydro on
the Norwegian shelf, but soon the effect of a world-wide sales
network appeared. Hibernia in Canada, Holland with a new
LM1600 package, Shell Rayong in Thailand and Maersk in the
Danish shelf all came on the client list in the early 90’s. New
milestones were reached in 97 with the Åsgard A project with
new package designs for the newly released LM2500+ and
LM6000 from GE. The LM2500+ also involved Dry Low
Emissions (DLE) combustors for the first time. When Dresser-
Rand in Olean had developed a lightweight industrial power
turbine named VECTRA for compressor drive with the
LM2500+, Kongsberg was quick to sell five units for pipeline
compression with British Gas in Scotland.
FIGURE 15. 3D-MODEL OF MT250 ROTOR
After a brief interval in 1999-2000 with some industrial
combined heat & power packages based on LM6000, emphasis IR started the sourcing process as soon as the first drawings
were completed and in only eighteen months after the start, in

10 Copyright © 2018 ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 09/05/2018 Terms of Use: http://www.asme.org/about-asme/terms-of-use


April 2002, the first prototype was ready for test in the manufacturing of the packages to the local shipyard. Typical
Portsmouth facility Fig. 16. Testing did not reveal any was the engineering for 14 LM2000 CO2 compressor packages
significant issues. to be built by D-R in Brazil for the Petrobras Pre-Salt project.

In order to minimize offshore installation work a “Single


Lift” mini module was developed for British Petroleum (BP),
Fig. 17. The package contained a control room which was
hooked up and tested before lifted on the platform.

FIGURE 16. MT250 MICRO TURBINE W/O ENCLOSURE

MT250 Key Data


 Single shaft design
 Single stage centrifugal compressor pressure ratio 4,5
 Radial inflow turbine
 Rotor speed: 44000 rpm FIGURE 17. SINGLE LIFT MINI MODULE
 Thermal Eff.: 31% (recuperated)
 ISO Power: 250 kW There were some tough years just after the formation of the
 Single radial can combustor, DLE new company and the financial crisis in 2008, but afterwards
the business grew steadily with the peak years3 in 2013-2014.
IR sold Dresser-Rand in 2004, see below, and in 2010 IR
sold the MT250 to Flex Energy, Portsmouth N.H., who is 8.2 KG2-3G
marketing it under the name GT250S as well as a scaled up Since the last upgrade of the KG2 with a new compressor in
version of 333kW, the GT333S. 1987, the KG2 design had been unchanged. A study was made
in 2003/2004 on a recuperated version based on the recuperator
8. DRESSER-RAND INC. (2005-2015) technology developed by the micro-turbine group in IR, but due
8.1 Growth in offshore oil & gas to lack of scalability and subsequent complexity of piping it
On August 26, 2004 First Reserve Corporation (FRC) a leading proved to be unrealistic.
private equity firm specializing in energy industry investments,
announced that it had agreed to acquire the Dresser-Rand The first presentation of the proposal of what was to become
Company from IR. The deal was closed on the 29th of October the KG2-3G was made in May 2007. The name KG2-3G was
and Dresser-Rand Company was from now on a standalone employed already in these first presentations since this could be
company. An initial public offering of common stock was made construed as being the next letter to be used in the KG2 line. 4
on the 16th of May 2005, and on the 5th of August the trading
started on the New York Stock Exchange under the symbol Some basic assumptions for a new engine had already been
DRC. Now D-R was an independent, publicly traded company made. In order to be able to tap into the retrofit market it should
for the first time in its history. build on the same design concept as the KG2 with an overhung
rotor with back-to back compressor and turbine, Fig. 18. It
In the following years D-R Kongsberg continued to offer should be in the 2 MW power range and have a pressure ratio
compressor and power generation packages based on the of about 7:1. The turbine inlet temperature was targeted to 1000
LM2500 series of engines, while at the same time supporting °C. The turbine should also have combustors with Dry Low
the installed fleet of KG engines with service and overhaul Emissions (DLE) as well as dual fuel (gas and liquid)
from its shop in Kongsberg. Following the tradition of capability.
developing solutions for the maritime sector, Kongsberg was
quick to respond to the needs of the growing fleet of FPSOs
(Floating Production Storage & Offloading) being built. Where 3
See Annex C for sales numbers per year.
appropriate one would partner with or outsource the actual 4
KG2-3E was the current version. 3F could be considered as the
recuperated version that was abandoned.

11 Copyright © 2018 ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 09/05/2018 Terms of Use: http://www.asme.org/about-asme/terms-of-use


Now followed a period of extensive testing, both on liquid
and gas fuel. There were no rotor dynamic or blade vibration
issues and the gear also proved to be running smoothly.
Towards the end of 2012 the initial engine qualification tests
were completed. During the spring of 2013 preparations for full
scale marketing were made and the product was finally released
in May.

In parallel with the qualification tests on the standard engine


there was worked on an application, based on an inquiry from
Germany, which involved external combustion of wood waste.
The outer casing was modified such that there are no
combustors and the pressurized air from the compressor is
FIGURE 18. KG2-3G ROTOR ducted from the casing on one side and is heated in a heat
exchanger by the exhaust from the wood burning. The heated
The project was finally approved in September 2008 and the air enters the engine again on the top of the casing. The engine
estimated time for the complete design phase was eighteen would thus only see clean, hot air from the heat exchanger, Fig.
months. In order to validate the initial combustor design it was 20.
invested in an atmospheric combustor test rig at the University
of Sheffield, UK. Tests were performed both with liquid fuel
and natural gas in order to reveal any problem areas.

In a little over 18 months, all the parts were defined and on


order and the assembly of the first prototype engine, Fig. 19,
could begin during the summer of 2010. Almost exactly two
years after the go-ahead was given, the 12th of October 2010 the
KG2-3G was cranked on the starter motor for the very first
time.

FIGURE 20. KG2-3G EXTERNALLY FIRED

Since there was no combustion development involved it was


considered that the technical solution could be derived to a
large extent from the micro-turbine project, (chapter 7.2 above).

The engine was named KG2-3G-EF for Externally Fired and


the development took place mostly in 2012. The outer casing
configuration for the externally fired unit, with the flanges for
air in and out, also lends itself to a recuperated version if that
becomes of interest at a later stage. Dresser-Rand is currently
FIGURE 19. KG2-3G GAS TURBINE PROTOTYPE marketing the KG2-3G-EF together with Ener-Core Inc. of
Irvine, CA using their patented Power Oxidizer as the heat
KG2-3G Key Data source.
 Single shaft design
 Single stage centrifugal compressor pressure ratio: 7 9. SUMMARY AND CONCLUSIONS
 Single stage radial inflow turbine Kongsberg Gas Turbines was started in 1964 as a way to fill a
 Rotor speed: 25500 rpm workshop after dwindling armament orders, and was very
 Thermal efficiency: ~26.5 % successful with its small, industrial gas turbines in the 1970s. A
 ISO power: 2150 kW strategic partnership with Dresser Industries in 1985 was turned
 4 angular can combustors, DLE into full ownership by the newly formed Dresser-Rand in 1987.

12 Copyright © 2018 ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 09/05/2018 Terms of Use: http://www.asme.org/about-asme/terms-of-use


The Kongsberg operation quickly adapted to the new http://proceedings.asmedigitalcollection.asme.org/proceed
product portfolio and has since been an internationally ing.aspx?articleid=2278772
recognized supplier of compressor driver- and generator [7] Westrum, Thorbjørn, “Apparatus for filling and, if
packages based on aero-derivative gas turbines for the last 25 desired, emptying inert gas poor in oxygen into or from
years. In 2015 Dresser-Rand Inc. was acquired by Siemens AG holds in ships, respectively”. US Patent 3776164 A, 1970.
and Kongsberg was again challenged with a new product https://www.google.com/patents/US3776164
portfolio. [8] Cotton John. L. and Singh, Riti, “Development and
Qualification of a Radial Flow 3000 kW Gas Turbine”,
Fifty years after start-up and through a number of gas CIMAC 14th International Congress on Combustion
turbine developments, product adaptations changes of Engines, GT26, pp. 1-24. Helsinki, Finland, 1981.
ownership and ups and downs, gas turbine know-how and [9] Naess, Tore and Teien, Kjell Olav, “The Effect of
experience is still alive and well in Kongsberg, for the benefit Thermal Transients on Radial Turbine Blade Vibration
of its owner and national as well as international clients. Damping”, ASME 1984 International Gas Turbine
Conference and Exhibit. 84-GT-110 pp. V005T13A013.
Amsterdam, The Netherlands, June 4-7, 1984.
ACKNOWLEDGMENTS DOI:10.1115/84-GT-110.
This paper would not have been possible without the support http://proceedings.asmedigitalcollection.asme.org/proceed
and encouragement from Odd Guldsten, VP & General ing.aspx?articleid=2276049
Manager of Dresser-Rand Kongsberg 2006-2014, in [10] Mowill, Jan and Strøm, Sigmunn, “An Advanced Radial-
researching and writing the company history. It would not have Component Industrial Turbine Engine”, ASME Journal of
become an ASME paper without the initiative and gentle Engineering for Power. Vol. 105 No. 4 (1983) pp. 947-
persuasion from Arne Lynghjem and Lars E. Bakken from the 952. DOI:10.1115/1.3227505
Norwegian Society for Thermal Turbomachinery and the http://gasturbinespower.asmedigitalcollection.asme.org/ar
Norwegian University of Technology and Science respectively. ticle.aspx?articleID=1420837
[11] Strøm, Sigmunn, “Gas turbine engine combustor”, US
Patent 4651534 A, 1984.
REFERENCES http://www.google.com/patents/US4651534
[1] Kongsberg Våpenfabrikk AS, Board of Directors Minutes [12] Strøm, Sigmunn, “Low pressure loss, convectively gas-
of Meeting No. 4 (1960-1969): pp. 410-416. The National cooled inlet manifold for high temperature radial turbine”,
Archives of Norway/The Regional State Archive of US Patent 4573315 A, 1986.
Kongsberg. http://google.com/patents/us4573315
[2] Hunt, Ronald J. “The History of the Industrial Gas [13] Norwegian Official Report. NOU 1989:2. “Kongsberg
Turbine - Part 1: The First Fifty Years 1940-1990” Power Våpenfabrikk: pp. 108-138. The National Library of
Engineer-Journal of the IDGTE Vol. 15 No. 2 (2011): pp. Norway. https://www.nb.no/items/URN:NBN:no-
6-36. http://www.idgte.org/synopses1997- nb_digibok_2013081306050
2015.html#P582 . [14] Potter, Trevor, “A Stiff, Lightweight Machinery Baseplate
[3] Johnson, Dag and Mowill, R. Jan, “Ægidius Elling-a on Three Spherical Supports”, ASME 1983 International
Norwegian Gas Turbine Pioneer”. Norwegian Museum of Gas Turbine Conference and Exhibit. 83-GT-146 pp.
Science and Technology, Oslo, Norway. March 1968. V003T07A008. Phoenix, Arizona, USA, March 27–31,
[4] Mowill, R. Jan, “Justifications for an all-radial 1600-hp 1983.
gas turbine engine”, ASME 1970 International Gas http://proceedings.asmedigitalcollection.asme.org/proceed
Turbine Conference and Products Show. 70-GT-33 pp. ing.aspx?articleid=2244617
V01AT01A034. Brussels, Belgium, May 24-28, 1970. [15] Jakobsen, Bernt and Rosendahl, Finn. “The Sleipner
DOI:10.1115/70-GT-33. Platform Accident”, Structural Engineering International
http://proceedings.asmedigitalcollection.asme.org/proceed Vol. 4 No. 3 (1994): pp. 190-193.
ing.aspx?articleID=2288368 https://www5.in.tum.de/~huckle/sleipner_Jakobsen.pdf
[5] Andvig, Tore Anton, “Arrangement for holding together a
turbine rotor and other aligned members of a gas turbine”.
US Patent 3642383 A. 1968.
https://www.google.com/patents/US3642383
[6] Westrum, Thorbjørn, “Operating Experience of
Kongsberg Gas Turbines and Marine Systems on Board
Merchant Ships”, ASME 1974 International Gas Turbine
Conference and Products Show. 74-GT-164 pp.
V01BT02A081. Zurich, Switzerland, March 30–April 4,
1974. DOI:10.1115/74-GT-164.

13 Copyright © 2018 ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 09/05/2018 Terms of Use: http://www.asme.org/about-asme/terms-of-use


ANNEX A
KG2 VERSIONS

Model Year of Base load Efficiency Turbine Pressure Exhaust Turbine Exhaust
Introduction rating speed Ratio mass flow inlet temp. Temp.
(ISO - kW) (LHV- %) (rpm) (kg/s) (Deg. C) (Deg C)
KG2-3 1968 1200 15,4 17 100 3,5 11,4 780 535
KG2-3A 1972 1400 16,1 18 000 3,9 12,8 810 562
1)
KG2-3B - - - 18 000 - - - -
KG2-3C 1973 1530 16,4 18 000 3,9 12,8 830 575
KG2-3D2) ~1980 1890 16,9 18 800 - 13,7 - -
KG2-3E 1987 1950 17,4 18 880 4,5 14,9 830 550
3)
KG2-3F - 1930 28,0 18 880 4,5 14,9 830 550
KG2-3G4) 2013 2100 27,4 25 500 7,0 9,0 1000 580
Notes:
1): KG2-3B had a "wide exit" compressor - never introduced to the market
2): KG2-3D was later designated KG2-SL3 and was a stand-by engine only
3): KG2-3F was meant to be a recuperated version - development never completed
4): KG2-3G is a completely new machine from 2012

ANNEX B. ANNEX C.
KG2 ANNUAL SALES AND ACCUMULATED LARGE GAS TURBINE PACKAGES
ANNUAL SALES AND ACCUMULATED

14 Copyright © 2018 ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 09/05/2018 Terms of Use: http://www.asme.org/about-asme/terms-of-use

You might also like