by Peter Yule
system. Norwood recalls:
The submarine was run across the sound range, power closed
off and the boat allowed to coast. The tones disappeared
immediately.
The tremendous turbulence created when Collins made way on
the surface had always been eye-catching and it seemed likely that
this was in some way interacting with the propeller to generate the
propulsion system noise. DSTO designed, developed and applied
a strain gauge system to measure the vibration of the propeller
blades. Tests with the submarine dived showed the vibrations of
the propeller matched the noise problems measured on the acous-
tic range. In tests with the submarine running on the surface the
level of propeller vibration was found to be so high that it exceeded
what should have been possible with the propulsion system.
It appeared that additional energy was being fed into the pro-
peller and the flow characteristics around the hull were the prime
suspect. DSTO did not have a testing tank but its Air Vehicles Divi-
sion had a research wind tunnel, so it was controversially decided
to use the aerodynamics research tool to evaluate the flow patterns
around the submarine’s hull and into the propeller. David Wyllie,
who was appointed chief of the Maritime Platforms Division in
1998, recalls strong opposition from parts of the navy to the idea
that a submarine could be tested in a wind tunnel, although air-
craft designs had been tested in water tunnels for years. Indeed,
utilising the wind tunnel may have had such a direct inspiration.
Wyllie recalls the story around Fishermans Bend at the time that a
friend of Bill Schofield, a professor in aerodynamics, having seen
pictures of the Collins, had rung to express his amazement that
the inefficiency of the hull shape should be so obvious.
The use of the wind tunnel was typical of how DSTO went
about its business, using initiative to make the best use of limited
resources. So the acoustic group used the knowledge of Dr Bruce
Fairlie of the Air Operations Division to set up tests that would
yield information on the interaction of flows around the hull and
into the propeller. Much of the experimental equipment had to
be designed on site by technical officer Peter Climas. Computer
238
T H E C O L L I N S C L A S S S U B M A R I N E S T O R Y
modelling was used, even though the available software was not
entirely suitable and was developed as the experiments progressed.
This procedure differed little from the design of the whole exper-
iment, however, as all the data yielded had to be interpreted to
allow for the differences between air and sea water in density, vis-
cosity and other parameters. The outcome was interesting enough
for Wyllie to remark on the number of senior US Navy officers
who began to visit DSTO’s Fishermans Bend facility and how this
promoted collaborative research between DSTO and its American
equivalents.
In 1993 the project office had sought contacts in the US Navy’s
engineering and research establishments. At first the Americans
resisted talking about their submarine technologies, but even-
tually an exchange developed over aspects of ‘technologies’ to
do with the sub-marine environment. As the Collins acoustic
signature became an increasing problem, Greg Stuart from the
project worked with diplomatic and scientific staff in the USA and
Australia to gain access to the US Navy’s expertise in acoustics.
The Americans closely guard this expertise, but eventually agreed
to model the performance of the Collins. The results correlated
closely with those of DSTO’s wind tunnel tests and confirmed
what the acoustics group had been telling the navy.
Several features of the Collins hull shape generated significant
turbulence: the cylindrical array knuckle, the fin, and the hull cas-
ing. This disturbed mass of water was tumbling over the abrupt
end of the casing and arriving at the propeller in two turbulent
streams that were nicknamed ‘rabbit’s ears’. Each propeller blade
hit turbulence twice in every revolution, increasing their natural
vibration and inducing cavitation. Kockums had tested a model
of the Collins propeller and was certain it did not cavitate, but
in the tank tests Kockums had trialled the propeller behind a per-
fectly cylindrical mount; there were no ‘rabbit’s ears’ to excite the
blades.
The findings from this research at DSTO and in America led to
the design of a series of modifications to the fibreglass casing and to
the fin. These succeeded in taming the ‘rabbit’s ears’ but the seven-
tonne propeller remained a problem. The propellers made for the
Collins class were cast from Sonoston and the first batch were
hand-chiselled from the casting – with the result that they were far
from the high-tolerance product of multi-axis machine tools that
T H E R O L E O F D E F E N C E S C I E N C E : N O I S E A N D D I E S E L S
239
drove nuclear-powered submarines in the north Atlantic. The
blades often did not line up in a flat plane, one behind the other,
but could be skewed cross-wise. Nor did the alloy perform as it
was supposed to. Ross Juniper of the acoustics group at DSTO
had Sonoston analysed. It did indeed have natural damping char-
acteristics, but they were engaged only at dynamic stresses much
higher than any likely to be encountered in the operational profile
of an Australian submarine. When not under these stress load-
ings, Sonoston was only a little better than conventional nickel-
alloy bronze. DSTO research also exposed the potential for this
material to crack when placed under load in seawater.
Although the unsatisfactory propeller characteristics that
bedevilled the submarines’ acoustic signature were largely due to
hull-generated turbulence, DSTO’s research suggested there were
serious flaws in the original propeller design and manufacture and
provided the navy with the basis to investigate alternative designs.
Further research on an increasingly quiet submarine found
that, although propeller blade vibration contributed to noise gen-
eration, it was not the sole emitter of sound tones. The entire drive
train was involved and it was the hull that radiated any propeller-
induced vibration. Further investigations headed off to the new
problem of how to damp the entire hull of a submarine.
While making progress towards producing a submarine that
operated quietly on battery power, Norwood and his group had
other urgent problems with vibration. The Collins submarines
were noisy when they snorted. The diesels – or, more accurately,
the generator sets – vibrated excessively, and Norwood’s col-
league, Geoff Goodwin, a propulsion engineer from the Mar-
itime Platforms Division, was working on the problem. However,
although engine vibration may have been the noise source, it was
not the means of transmission. DSTO saw that the intermediate
masses in the two-s
tage mounts on which the diesel generator sets
were placed were too light. That meant that the masses could be
vibrating at the same frequencies as the generator set, with a dan-
ger that the mounts could be ineffective at those frequencies and
could actually amplify the noise transmission.
Traditionally, an intermediate mass system for two-stage
mounts for heavy industrial engines should have the same mass
as the machinery mounted on it. Yet, when the Type 471 design
was selected it was recognised as weight limited, with the small
240
T H E C O L L I N S C L A S S S U B M A R I N E S T O R Y
margin of only 20 tonnes to accommodate changing requirements
over the life of the boat. One of the weight-saving measures used
by Kockums was to lighten the generator mounting structure,
and the designers attempted to compensate by fitting the engine
mounts with absorbers tuned to cancel out critical vibration fre-
quencies. This was ineffective. The tuned absorbers used a split
mass, and when analysed they had different characteristics in the
way they damped vibration depending on whether the split was
aligned with, or at right angles to, the vibration source. The design
made it possible for them to be installed either way and it was pos-
sible that they could be wrongly installed. DSTO redesigned the
absorbers to ensure better tuning and uniformity in installation.
The diesel exhaust was another noise source which DSTO anal-
ysis showed was essentially a design issue. The space within the
engine room and fin is limited, and cramming the mufflers and
exhaust pipes into the confined area had been achieved by using
mufflers that were too small for good acoustic performance. For
example, the uneven lengths of the pipes in the exhaust system pro-
duced undesirable tones. The acoustic group’s analysis prompted
a successful redesign of this part of the exhaust system. They then
examined the exhaust outlet at the top of the fin and conducted
experiments, including water injection into the exhaust, but such
is the complexity of underwater acoustics that many of the results
proved inconclusive.
It was the nature of the acoustic group’s work that experi-
mentation did not always find the solution. As the analysis of
Collins’ acoustic signature developed, two distinct tones could
be heard when it was on battery power. One proved to be from
a redundant system that could simply be switched off, but the
other was more difficult. ‘Malice’, a near-field acoustic hologra-
phy research tool, helped to indicate that the tone emanated from
the battery stack exhaust fans. John Dickens, an electrical engineer
with the group, found the fans had been designed for two-speed
operation and, although they were on slow for 90 per cent of the
time, the design had been optimised to the high-speed requirement.
Dickens suggested that the cause could lie here and the problem
was resolved by altering the voltage supplied to the fans.
While the Ship Noise and Vibration Group was analysing the
submarines’ acoustic characteristics, another section of DSTO was
investigating another problem area. In 1997 Dr Geoff Goodwin
T H E R O L E O F D E F E N C E S C I E N C E : N O I S E A N D D I E S E L S
241
received a diesel engine piston at his Maribyrnong laboratory.4 A
hole had been opened in its side. A few months later, a second
piston arrived in a similar but less advanced state, and there were
stories that the pistons of the new submarines were melting. This
was the beginning of research by the Propulsion and Energy Man-
agement Technologies Group into the Hedemora diesel engines.
Goodwin was an expert on marine diesel engines. He had
gained his PhD from Sussex University in the 1970s for his thesis
on engine problems with the Royal Navy’s Oberon submarines.
In late 1987 he arrived in Adelaide as the logistics engineer-
ing manager for the newly established ASC, but he returned to
research in 1992 when Janis Cocking recruited him into DSTO’s
propulsion group as a combustion engineer.5 At the time the main
focus of this group was an examination of the technical risks
associated with fitting the new submarines with air-independent
propulsion.
The pistons in Goodwin’s laboratory suffered from ring carrier
failure: the piston ring carrier had broken away and iron debris
had beaten a cavity in the side of the aluminium piston. Good-
win and the failure investigators at DSTO deduced that this was
a manufacturing fault in a couple of pistons and a lack of any
similar incidents supported a view that it was an isolated event.
However, the submarines’ diesel engines were earning a surpris-
ing reputation for unreliability. Since one of the central design
concepts of the class had been to reduce the snorting time by hav-
ing sufficient generating capacity to quickly recharge the batteries,
the emerging unreliability of the engines significantly reduced their
capability.
Kockums knew the engines had vibration problems. The V18
diesel-powered generator sets to be used for the Collins class were
derived from Hedemora’s best-selling line of power generators.
These normally sat on the heavy chassis of a locomotive or on
the bed plate on a drilling rig. The units for the submarines were
unique monolithic generator sets, with a stiffer engine crankcase
together with the generator and bell housing forming its struc-
ture. There was no substructure, the complete generator set being
mounted on rubber isolators. Hedemora adopted this arrange-
ment for earlier Swedish submarines but these were smaller-bore
V12 units. The V18 generator set was nearly seven metres long
and weighed 23 tonnes, sufficient to allow the structure to bend
242
T H E C O L L I N S C L A S S S U B M A R I N E S T O R Y
and twist along its length, especially in the region of the bell hous-
ing. In the same way that a ruler can be ‘twanged’, this could
generate vibration at a natural frequency. This was the vibration
that Kockums sought to isolate with the tuned absorbers and that
DSTO’s acoustic group was seeking to correct.
From Geoff Goodwin’s viewpoint, this natural frequency of
the generator set was too close to others generated by the engine
and could start enough vibration to damage other components.
Already, there were reports of cracked and leaking auxiliary pip-
ing. The navy tried to manage the problem by slightly increasing
the diesel’s running speed, but the situation worsened, and so the
diesel speed was reduced. This reduced the vibration but Jeumont
Schneider warned that this would cook their generators. A com-
promise engine speed was agreed, but the consequence was that
the generators produced slightly less power and recharging the
batteries took a little longer.
The propulsion group also developed a computer model of
the engine to determine its natural frequencies. Validation of the
/> model was achieved by placing 14 accelerometers around a gen-
erator set on Dechaineux and then striking the end of the diesel
with a 15-pound plastic-faced mallet. This enabled a full struc-
tural model of the generator set to be completed by a team of
four scientists. The model was ‘fitted’ with stiffeners and bracing
bars that joined the engine block and generator across the bell
housing, and increased the lateral bending stiffness of the unit.
Results showed that the natural frequency of the stiffened unit
should have changed by several Hertz. A slightly lesser result was
achieved in practice; nevertheless, this simple modification was
good enough to halve the vibration. The mass of the bracing units
totalled less than 100 kilograms.
The propulsion group later received broken parts from auxil-
iary gearboxes that had failed. This was not unusual for DSTO,
as one of its standard roles is to improve the operation of defence
equipment by analysing recurring faults. Goodwin’s preferred
approach was to consult with the manufacturer and encourage
them to accept a need for change and to supply a modified part as
manufacturer’s components, carrying a stock number and a guar-
antee. The gearbox in the submarines is fitted with two starter
motors because the engine starts as an air pump, to blow water out
of the exhaust system when the submarine begins to snort. It was
T H E R O L E O F D E F E N C E S C I E N C E : N O I S E A N D D I E S E L S
243
the first Hedemora engine where two starter motors were needed,
and consequently there was more starting torque than in any other
Hedemora engine. The propulsion group put a failed interme-
diate gear through stress analysis and the part was redesigned.
The results were turned over to Hedemora, who did their own
checks and came back with a part that was even stronger than
had been recommended by DSTO. The problem was solved. For
his part, Goodwin enjoyed working with Hedemora and found
them responsive to his suggestions.
DSTO’s Propulsion and Energy Management Technologies
Group continues its work on the Hedemora generator sets and is
conducting thermal modelling and other experiments to improve
the consistency of operation of the turbocharger turbines. It has
continued to study problems with some nozzle failures, provid-