A variable buoyancy system for deep ocean vehicles

Abstract

A variable buoyancy system has been developed for underwater vehicles operating deep in the ocean. This paper reports on the design, testing and development of the system. The system was designed to change buoyancy at up to 1 l/min at a depth down to 6000 m. The results showed that the system worked at its design specifications after modification but that friction losses resulted in a relatively low efficiency of around 35 % at low working depth, but efficiency increased with increasing depth to about 70% at 6000 m. Efficiency could be increased further with redesign or with changes in specification.

Full text

061215-063
A
variable
buoyancy
system
for
deep
ocean
vehicles
M.
Worall
BEng,
PhD,
AMIMechE,
A.
J.
Jamieson
BSc,
PhD,
A.
Holford
BSc,
R.
D.
Neilson
BSc,
MSc,
PhD,
M
Player
MA,
DPhil(Oxon),
CEng,
MIEE,
P.
M.
Bagley
BSc,
PhD,
CEng,
MIEE.
Abstract-
A
variable
buoyancy
system
has
been
developed
for
underwater
vehicles
operating
deep
in
the
ocean.
This
paper
reports
on
the
design,
testing
and
development
of
the
system.
The
system
was
designed
to
change
buoyancy
at
up
to
1
1/min
at
a
depth
down
to
6000m.
The
results
showed
that
the
system
worked
at
its
design
specifications
after
modification
but
that
friction
losses
resulted
in
a
relatively
low
efficiency
of
around
35
%
at
low
working
depth,
but
efficiency
increased
with
increasing
depth
to
about
70%
at
6000m.
Efficiency
could
be
increased
further
with
redesign
or
with
changes
in
specification.
Index
Terms-
AUV,
ROV,
variable
buoyancy,
variable
ballast
I.
INTRODUCTION
A
variable
buoyancy
system
has
been
developed
at
Oceanlab,
to
provide
low
cost,
low
power
consumption,
regenerative
variable
buoyancy
capability
for
underwater
vehicles
operating
deep
in
the
ocean.
A
prototype
was
commissioned
and
built.
This
was
tested
and
a
large-scale
system
was
then
designed.
This
paper
describes
the
design
of
the
large
scale
system,
results
of
tests
carried
out
and
modifications
made
to
improve
performance.
II.
BACKGROUND
Oceanlab
use
autonomous
scientific
instrument
packages
This
work
has
been
supported
by
the
University
of
Aberdeen
through
their
knowledge
transfer
fund
and
Scottish
Enterprise
through
their
proof
of
concept
scheme.
M.
Worall
is
with
Oceanlab,
The
University
of
Aberdeen,
Newburgh,
Aberdeenshire,
Scotland,
AB41
6AA,
(phone:
+44(0)1224
274412;
fax:
+44(0)1224
274402;
e-mail:
m.worall@
abdn.ac.uk).
A.
J.
Jamieson
is
with
Oceanlab,
The
University
of
Aberdeen,
Newburgh,
Aberdeenshire,
Scotland,
AB41
6AA
(phone:
+44(0)1224
274447;
fax:
+44(0)1224
274402;
e-mail:
ajamieson@abdn.ac.uk).
A.
Holford
is
with
Oceanlab,
The
University
of
Aberdeen,
Newburgh,
Aberdeenshire,
Scotland,
AB41
6AA
(phone:
+44(0)1224
274412;
fax:
+44(0)1224
274402;
e-mail:
a.holford@abdn.ac.uk).
R.
D.
Neilson
is
with
the
Engineering
Department,
The
University
of
Aberdeen,
Kings
College,
Aberdeen,
Scotland,
AB24
3FX
(phone:
+44(0)1224
272797;
fax:
+44(0)1224
272497;
e-mail:
r.d.neilson@abdn.ac.uk).
M.
Player
is
with
the
Engineering
Department,
The
University
of
Aberdeen,
Kings
College,
Aberdeen,
Scotland,
AB24
3FX
(e-mail:
m.player@
abdn.ac.uk).
P.
M.
Bagley
is
with
Oceanlab,
The
University
of
Aberdeen,
Newburgh,
Aberdeenshire,
Scotland,
AB41
6AA
(phone:
+44(0)1224
274408;
fax:
+44(0)1224
274402;
e-mail:
p.bagley@abdn.ac.uk).
called
landers
that
are
deployed
from
a
surface
vessel
and
descend
to
the
sea
floor.
Experiments
are
carried
out
and
then
the
lander
is
retrieved
by
shedding
ballast
[1].
A
VBS
was
first
considered
so
that
a
lander
could
be
retrieved
without
having
to
shed
ballast,
could
hover
in
mid-water
and
soft
land
on
the
seabed,
and
so
minimize
disturbance.
Jamieson
[2]
describes
the
development
of
the
concept
and
its
application
to
lander
technology.
A
low
power,
compact
variable
buoyancy
system
could
also
benefit
underwater
vehicles,
especially
unmanned
underwater
vehicles.
Patent
No
W02005019021
[3]
reveals
a
variable
buoyancy
device
for
controlling
the
buoyancy
of
unmanned
underwater
vehicles
such
as
remotely
operated
vehicles
(ROVs)
and
autonomous
underwater
vehicles
(AUVs).
ROVs
are
used
extensively
for
the
exploration
of
the
ocean
in
the
scientific
sector,
for
inspection,
maintenance
and
construction
in
the
oil
and
gas
industry
and
for
remote
intervention
and
surveillance
in
the
military
sector.
ROVs
are
controlled
and
powered
from
the
surface
through
an
umbilical
cable
and
need
extensive
on-board
logistical
support
from
a
field
service
vessel
(FSV).
A
ROV
generally
controls
its
position
in
the
water
column
and
lifts
or
lowers
payloads
by
using
vertical
thrusters.
When
seawater
properties
vary
from
site
to
site,
or
a
change
of
tooling
is
required,
the
ROV
may
require
trimming
to
maintain
slight
positive
buoyancy
and
level
attitude.
This
can
be
time
consuming
and
is
usually
done
manually
by
lifting
the
ROV
out
of
the
water
and
changing
or
shifting
ballast.
A
variable
buoyancy
capability
could
enable
a
ROV
to
control
its
position,
manipulate
payloads
and
trim
a
vehicle
in
the
water.
An
AUV
is
a
robotic
vehicle
that
is
powered
from
on-board
sources
and
controlled
using
pre-
programmed
mission
profiles.
The
vehicle
does
not
require
logistical
support
except
for
deployment
and
retrieval.
The
energy
source
is
limited
[4]
and
so
its
functions
are
restricted
mainly
to
observation,
inspection,
and
surveying.
An
AUV
is
generally
neutrally
buoyant
and
moves
vertically
by
lift
generated
from
control
surfaces
as
it
moves
forward.
The
limited
energy
source
constrains
the
range
of
the
vehicle
and
its
data
gathering
capacity.
Energy
that
is
consumed
in
propelling
and
steering
the
AUV
reduces
the
data
that
can
be
obtained.
A
variable
buoyancy
capability
could
enable
AUVs
to
use
energy
more
efficiently
and
enhance
its
functionality.
A
prototype
VBS
was
designed,
built
and
tested
at
Oceanlab.
The
results
and
conclusions
were
described
by
Worall
et
al
[5].
Fig.
1
shows
a
photograph
of
the
prototype
1-4244-0635-8/07/$20.00
©2007
IEEE
I

061215-063
Pressure
housling
Fig.
1.
Prototype
VBS
VBS.
It
consisted
of
an
axial
piston
hydraulic
pump
driven
by
a
120W
dc
motor
and
actuating
a
single
acting
pressure
intensifier.
It
was
designed
to
operate
at
an
ambient
pressure
of
up
to
300bar.
The
system
was
tested
using
a
flow
control
valve
to
produce
back-pressure
and
so
simulate
ambient
pressure.
Tests
showed
that
the efficiency
of
the
system
was
approximately
80%
and
flow
was
80
cm3/min
at
a
maximum
back-pressure
of
300bar.
It
was
concluded
that
the
system
worked
efficiently
but
the
flow
was
only
be
suitable
for
small
changes
required
for
systems such
as
buoyancy
engines
that
are
used
in
profiling
drifters
and
gliders
[6].
A
higher
specification
system
was
then
developed
from
the
prototype
design
that
would
be
suitable
for
underwater
vehicles.
III.
SYSTEM
DESCRIPTION
A
list
of
capabilities
was
drawn
up
from
the
conclusions
from
the
tests
on
the
prototype
and
market
surveys
carried
out
in
the
ROV,
AUV
industry.
Key
capabilities:
*
Compensate
for
picking
up
and
dropping
off
loads
*
Control
descent
and
ascent
*
Anchoring
on
seafloor
*
Compensating
for
salinity,
density
and
pressure
variations
U
U
U
Emergency
release
and
recovery
Docking
and
construction
manoeuvres
Bolting
on
to
existing
submersibles
Autonomous
of
user
controlled
System
specification:
6000m
depth
rated
30kg
buoyancy
capability
1
kg/min
buoyancy
change
1.5
kW
power
consumption
75
Ah
capacity
for
3
cycles
Regenerative
Off-the-shelf
components
Fig.
2.
Regenerative
VBS
circuit
diagram
Fig.
2
shows
a
circuit
diagram
of
the
VBS
and
fig.
3
shows
a
photograph
of
the
system
in
its
pressure
housing.
The
system
has
two
modes
of
operation.
In
pumping
mode,
the
VBS
pumps
seawater
from
a
buoyancy
vessel
to
the
ambient
so
that
buoyancy
can
be
increased.
In
regeneration
mode,
a
buoyancy
vessel
is
filled
from
the
ambient
via
the
VBS,
decreasing
buoyancy
but
extracting
some
of
the
energy.
An
embedded
microcontroller
is
used
to
control
pump
speed
and
valve
sequencing.
The
microcontroller
is
also
used
to
log
data
from
transducers
that
measure
motor
voltage,
battery
voltage,
motor
current,
hydraulic
pressure,
ambient
pressure
and
pump
speed.
The
VBD
can
be
run
autonomously
by
pre-programming
the
microcontroller
with
a
mission
profile
or
it
can
be
controlled
in
real-time
through
an
RS232
connection.
All
of
the
components
are
pressure
sensitive
and
so
are
enclosed
in
a
single
pressure
housing.
The
pressure
housing
is
0.4m
inside
diameter,
0.46m
outside
diameter,
0.25m
in
length
and
the
material
of
construction
is
grade
5
titanium.
Two
6000m
rated
glass
hemispheres
are
used
as
end-caps
to
minimize
weight
and
cost.
A.
Pumping
mode
When
an
increase
in
buoyancy
is
required,
hydraulic
fluid
is
pumped
around
a
circuit
by
an
axial
piston
pump/motor.
The
pump/motor
is
driven
by
24v
dc
permanent
magnet
motor
and
its
speed
controlled
by
a
four
quadrant
controller.
Motor
speed
Fig.
3.
Regenerative
VBS
photograph
2

061215-063
is
proportional
to
the
voltage
across
it
and
is
controlled
by
chopping
the
frequency
of
the
supply.
The
four
quadrant
controller
also
allows
a
dc
motor
to
be
controlled
in
four
modes,
two
modes
are
motor
forward
and
reverse,
and
the
other
two
are
generator
forward
and
reverse.
The
microprocessor
is
programmed
to
allow
a
selection
of
motor
speeds
from
seven
discrete
speed
settings
that
are
in
proportion
to
the
supply
voltage.
Valve
A
is
a
four
way
three
position
valve
that
allows
fluid
to
flow
back
to
the
return
side
in
its
neutral
position
and
has
two
positions
to
actuate
the
pressure
intensifier.
Solenoids
in
valve
A
are
energized
alternately
to
direct
flow
to
the
pressure
intensifier.
The
pressure
intensifier
is
the
interface
between
hydraulic
and
seawater
circuits
and
multiplies
the
pressure
in
a
4:1
ratio
so
that
pump/motor
and
VBS
specification
are
matched
and
standard
hydraulic
components
can
be
used.
If
solenoid
Al
is
energized,
the
flow
path
is
from
the
pump/motor
to
port
P
and
through
port
A
to
service
port
SI
in
the
pressure
intensifier.
A
pressure
difference
builds
up
across
the
pressure
intensifier
and
hydraulic
fluid
flows
from
the
other
side
of
the
piston
through
service
port
S2,
to
port
B
and
port
T
and
back
to
the
return
side
of
the
pump/motor.
Motion
of
the
piston
causes
seawater
to
be
pumped
out
of
the
right
side
of
the
intensifier
and
through
valve
C,
valve
D
and
non-return
valve
E.
Seawater
is
drawn
from
the
buoyancy
vessel
through
valve
B
and
enters
the
pressure
intensifier
to
the
right.
When
the
piston
in
the
intensifier
reaches
the
end
of
its
stroke,
an
inductive
sensor
SW2
detects
it,
valve
Al
is
de-energized
and
valves
A2,
B
and
C
are
energized.
The
piston
moves
to
the
left
causing
water
to
be
pumped
out
to
ambient
from
the
left
hand
side
and
drawn
in
from
the
buoyancy
tank
to
the
right
hand
side.
When
the
piston
reaches
the
end
on
its
stroke,
inductive
sensor
SWI
detects
it,
solenoids
A2,
B
and
C
are
de-
energized,
and
solenoid
Al
is
energized,
repeating
the
cycle.
The
actuation
of
the
piston
in
the
intensifier
continues
until
the
required
volume
of
water
has
been
pumped
from
the
buoyancy
tank.
An
accumulator
stores
excess
fluid
so
that
leakage
can
be
compensated
for
and
enables
the
hydraulic
system
to
operate
in
a
closed
circuit.
The
pressure
relief
valve
protects
the
circuit
from
excessive
pressure.
A
non-return
valve
E
prevents
flow
through
valve
E
when
valve
D
is
off.
The
system
is
filled
through
a
quick-connect
coupling,
and
air
is
removed
from
the
system
by
a
bleed
nipple.
B.
Regeneration
mode
When
a
decrease
in
buoyancy
is
required,
valve
D
is
opened
and
pressure
from
the
ambient
flows
through
a
filter
and
metering
valve
F.
If
valve
A
is
in
its
neutral
position,
the
path
to
the
circuit
is
shut
off
and
the
piston
in
the
pressure
intensifier
will
not
move.
If
leakage
should
occur
then
the
piston
will
only
move
to
the
end
of
its
stroke,
and
so
there
will
be
no
flow
to
the
buoyancy
tank.
If
solenoid
Al
is
energized,
then
seawater
flows
past
valve
C
and
acts
on
the
right
side
of
the
pressure
intensifier,
building
a
pressure
difference
across
it.
Pressure
builds
up
in
the
hydraulic
circuit
and
hydraulic
fluid
flows
from
service
port
S1
and
on
to
port
A
and
P.
At
the
same
time
a
path
is
opened
to
service
port
S2
through
ports
B
and
T.
Seawater
flows
from
the
left
hand
side
of
the
pressure
intensifier
through
valve
B
and
on
to
the
buoyancy
vessel
due
to
the
motion
of
the
piston.
A
pressure
difference
across
the
pump/motor
and
flow
produces
a
torque,
which
drives
the
dc
motor
and
generates
an
electric
current.
The
four
quadrant
controller
allows
the
dc
motor
to
be
used
as
a
generator
and
to
charge
the
24v
dc
battery
pack.
When
the
piston
in
the
pressure
intensifier
reaches
the
end
of
its
stroke,
inductive
sensor
SWI
detects
it,
solenoid
Al
is
de-energized
and
solenoid
A2
is
energized.
Solenoids
in
valves
B
and
C
are
energized
so
that
paths
are
opened
from
ambient
to
buoyancy
tank.
Seawater
from
ambient
enters
the
left
hand
side
of
the
pressure
intensifier
through
valve
B.
Seawater
from
the
right
hand
side
of
the
pressure
intensifier
flows
through
valve
C
and
on
to
the
buoyancy
vessel.
With
solenoid
A2
energized,
hydraulic
fluid
flows
from
port
S2
and
through
port
B
and
P
and
on
to
the
pressure
side
of
the
hydraulic
circuit.
Hydraulic
fluid
from
the
return
side
of
the
pump/motor
flows
through
port
T
and
A
and
on
to
service
port
S2.
When
the
piston
reaches
the
end
of
its
stroke
SW2
detects
it,
Solenoid
A2,
and
valves
B
and
C
are
de-energized
and
solenoid
Al
is
energized,
thus
repeating
the
cycle
until
the
required
amount
of
seawater
has
entered
the
buoyancy
tank.
IV.
RESULTS
Tests
were
carried
out
to
determine
the
performance
of
the
VBS
A.
Test
schedule
Preliminary
tests
revealed
that
a
compressed
air
supply
would
be
needed
to
pressurize
the
buoyancy
vessel
in
order
to
overcome
suction
losses,
and
a
time
delay
would
be
needed
to
enable
the
intensifier
to
enhance
the
filling
at
the
end
of
each
stroke.
The
system
was
operated
at
varying
ambient
pressures,
simulated
by
creating
back-pressure
with
a
high
pressure
flow
control
valve.
Each
test
was
run
for
20
minutes
during
which
the
flow
was
measured
in
a
number
of
ways.
The
volume
pumped
on
each
stroke
was
measured
by
a
100cm3
measuring
vessel.
The
time
taken
to
pump
1000cm3
was
recorded
every
5
minutes
so
that
any
variation
in
flow
could
be
determined
during
the
test
due
to
a
reduction
in
battery
voltage.
The
total
volume
pumped
was
recorded
after
each
test.
During
each
test,
measurements
of
battery
voltage,
motor
voltage,
motor
speed,
hydraulic
pressure
and
ambient
pressure
were
logged
at
regular
intervals
so
that
variations
in
values
could
be
observed
over
the
test
period.
After
each
test
the
battery
pack
was
recharged
so
that
all
of
the
tests
started
at
approximately
the
same
battery
voltage.
The
ambient
pressures
simulated
ranged
from
Obar
to
600bar
in
lOObar
intervals.
B.
Test
results]
Fig.
4
shows
how
work
in
(Win),
work
out
(Wout)
and
efficiency
(q)
change
with
back-pressure.
At
Obar
back-pressure,
the
pressure
required
to
actuate
the
3

061215-063
1200
1000
§~800
o
600
1-o
0
400
200
oQI
Qo
0
100
200
300
400 500
600
700
Back-pressure
(bar)
Fig.
4.
Variation
in
work
in,
work
out
and
efficiency
with
back-pressure.
Motor
speed,
time
delay
and
air
pressure
were
10OOrpm,
2s
and
2bar,
respectively.
pressure
intensifier
and
overcome
friction
was
approximately
50bar
and
the
work
done
was
approximately
600W.
There
was
a
loss
of
approximately
20bar
across
valve
A
and
losses
of
about
5
bar
each
across
valves
B,
C
and
D.
The
pressure
ratio
of
the
pressure
intensifier
also
results
in
a
four
to
one
flow
ratio.
At
a
pump
speed
of
10OOrpm,
a
pump
capacity
of
4.9cm3/rev,
and
a
pump
efficiency
of
95%,
volume
flow
was
approximately
4.71/min.
The
nominal
flow
from
the
VBS,
assuming
full
swept
volume
and
zero
time
delay
should
be
1.161/min.
However,
the
average
volume
flow
was
measured
at
0.31/min.
The
volume
discharged
during
each
stroke
was
approximately
31cm3
compared
with
a
swept
volume
of
72cm3,
therefore
there
was
a
57%
reduction
in
discharge
volume
per
stroke.
A
time
delay
of
2
seconds,
necessary
to
fill
the
cylinder
to
31cm3
also
reduced
the
average
flow.
The
57%
reduction
in
volume
discharged
was
due
mainly
to
restrictions
in
flow
on
the
suction
side
by
valves
B
and
C.
Valves
B,
C
and
D
were
chosen
because
they
were
suitable
for
use
in
seawater
systems
and
were
rated
to
over
600bar.
Tests
were
carried
out
at
higher
air
pressures
and
with
longer
time
delays.
It
was
found
that
for
the
same
air
pressure,
the
average
flow
was
approximately
the
same,
whatever
the
time
delay.
For
the
same
time
delay,
volume
discharged
increased
80
70
E
60
60
-o
w
0
-
a)
10
20
0
0.5
1
1.5
2
Air
pressure
(bar)
with
increasing
air
pressure.
Fig.
5
shows
the
change
in
volume
discharged with
air
pressure.
It
can be
seen
that
almost
full
cylinder
volume
is
discharged
when
air
pressure
of
3bar
is
applied.
However,
it
was
thought
that
a
pressurized
tank
on-
board
an
underwater
vehicle
would
be
undesirable
or
would
compromise
safety.
Tests
were
completed
at
an
air
pressure
of
2bar
and
a
time
delay
of
two
seconds.
At
a
back-pressure
of
100
bar,
the
work
in
was
approximately
670W,
the
work
out
was
approximately
76W
and
efficiency
was
11%.
At
600bar
back-pressure,
work
in
was
1OOOW,
work
out
450W
and
efficiency
45%.
At
600bar
back-pressure,
the
average
volume
flow
was
approximately
0.3
1/min
showing
that
the
capacity
was
the
same
for
varying
back-pressures.
The
low
efficiency
was
mainly
due
to
the
reduced
stroke
volume,
time
delay,
and
pressure
losses
in
the
system.
Modifications
were
made
to
address
the
pressure
losses
on
the
suction
side
of
the
seawater
system
so
that
efficiency
could
be
increased,
and
time
delays
and
the
need
for
a
compressed
air
supply
could
be
eliminated.
C.
Modified
design
Fig.
6
shows
a
circuit
diagram
of
the
modified
VBS.
A
glossary
of
the
graphic
symbols
used
is
given
in
Table
I
of
Appendix
A.
The
main
difference
with
the
previous
design
is
the
elimination
of
valves
B
and
C
in
fig.
2
and
their
replacement
with
a
number
of
check
valves.
Unfortunately,
the
replacement
of
valves
B
and
C
meant
that
there
was
no
active
direction
control
on
the
seawater
side
and
so
the
regenerative
function
was
eliminated.
It
was
concluded
that
simple
modifications
could
increase
efficiency,
but
more
complex
modifications
would
be
required
to
include
a
regenerative
capability.
When
negative
buoyancy
is
required,
water
is
pumped
from
the
buoyancy
tank
to
the
ambient.
As
in
the
previous
version,
a
24v
dc
motor
drives
a
hydraulic
pump.
Fluid
is
pumped
around
the
circuit
and
when
the
pressure
intensifier
is
required
to
be
actuated,
solenoids
Al
or
A2
are
energized.
If
Al
is
energized,
fluid
flows
through
port
P
to
A
and
on
the
port
SI.
Fluid
on
the
return
side
flows
from
port
S2
to
port
B
and
T.
The
piston
moves
to
the
right,
drawing
in
water
from
the
Pressure
housing
-----------------------------------------------
2.5
3
3.5
Fig.
5.
Variation
in
volume
discharged
from
single
stroke
of
intensifier
with
increasing
air
pressure.
Time
delay
two
seconds
at
end
of
each
stroke.
Fig.
6.
Modified
VBS
circuit
diagram
4
0

061215-063
buoyancy
tank
on
the
left
hand
side
and
pumping
water
out
to
ambient
on
the
right
hand
side.
Four
check
valves
are
arranged
in
a
circuit
to
enable
flow
to
be
directed
to
suction
and
pressure
sides
whichever
direction
the
piston
is
traveling.
As
the
piston
reaches
the
end
of
its
stroke,
an
inductive
sensor
detects
it,
Al
is
de-energized
and
A2
is
energized.
Pumping
continues
as
the
piston
travels in
the
opposite
direction.
When
positive
buoyancy
is
required,
valve
B
is
energized,
and
flow
from
ambient
is
diverted
past
non-return
valve
C
to
the
suction
side
of
the
check
valve
circuit.
Water
flows
to
the
buoyancy
tank,
bypassing
the
pressure
intensifier,
thus
free
flooding
the
tank.
D.
Test
results
2
Fig.
7
describes
the
performance
of
the
system
as
back-
pressure
is
varied.
At
Obar
back-pressure,
the
hydraulic
system
operated
at
approximately
50bar
in
order
to
overcome
friction
losses
in
the
system
and
the
average
flow
was
measured
at
1.1
1/min.
This
compares
with
a
50bar
pressure
loss
and
an
average
flow
of
0.31/min
for
the
previous
version.
Time
delays
or
compressed
air
were
not
required
and
full
swept
volume
was
achieved
on
each
stroke.
This
showed
that
suction
side
pressure
losses
were
negligible
compared
with
the
previous
design.
Friction
losses
on
the
pressure
side
were
found
to
be
similar
to
the
results
described
in
fig.
4
despite
the
elimination
of
valves
B
and
C
because
volume
flow
had
more
than
trebled.
Pressure
losses
remained
approximately
constant
as
ambient
pressure
increased,
so
that
efficiency
was
low
because
of
the
proportionally
high
losses.
Efficiency
increases
as
ambient
pressure
increases,
from
32%
at
lOObar
to
71%
at
500bar.
The
work
done
by
the
hydraulic
system
increased
from
520W
at
Obar
to
1300W
at
500bar.
Low
efficiency
is
due
mainly
to
friction
losses
in
both
the
hydraulic
and
seawater
circuits.
The
system
was
tested
without
valve
B
to
evaluate
individual
pressure
drops.
The
pressure
required
by
the
hydraulic
system
to
pump
water
at
Obar
ambient
pressure
was
approximately
30bar,
and
so
valve
B
was
causing
a
20bar
pressure
drop.
Further
modifications
to
the
design
could
increase
the
efficiency
of
the
system.
A
direction
control
valve
with
lower
friction
losses
could
reduce
pressure
drop
in
the
hydraulic
circuit
and
increasing
the
tube
bore
by
upgrading
the
piping
system
could
further
reduce
pressure
drop.
Fig.
8
1400
1200
-o
800
Ih
600
o
400
2
400
200
0
1.5
14
.
1.3
2
1.2
O>
1)
1.1
X.
0
100
200
300 400
500 600
Back-pressure
(bar)
Fig.
8.
Variation
in
motor
speed
and
average
volume
flow with
back-
pressure.
shows
how
motor
speed
(n)
and
average
volume
flow
(Vave)
vary
with
back-pressure.
As
back-pressure
increases,
the
current
draw
increases
causing
a
voltage
drop
across
the
VBS.
Pump
speed
is
controlled
by
varying
the
voltage
across
the
electric
motor
in
discrete
proportions
of
the
total
voltage
and
so
speed
decreases,
causing
the
average
volume
flow
to
decrease.
Fig
8
shows
that
average
volume
flow
and
speed
decrease
from
1.421/min
and
1160rpm
at
Obar
back-pressure
to
1.161/min
and
920rpm
at
500
bar
back-pressure.
This
shows
that
a
flow
of
over
1
1/min
is
achieved
over
the
design
pressure
range
and
this
is
because
of
the
control
of
motor
speed.
V.
DIscussIoN
The
system
has
been
specified
to
operate
at
a
depth
of
up
to
6000m
and
to
pump
water
out
of
a
ballast
tank
at
up
to
1
1/min.
The
specifications
led
to
the
selection
of
the
hydraulic
pump/motor,
dc
motor
and
the
design
of
the
pressure
intensifier.
The
hydraulic
pump/motor
had
the
lowest
capacity
that
could
be
sourced
and
so
the
minimum
pump
speed
and
intensifier
area
ratio
determined
the
pressure
and
flow
characteristics
of
the
system.
If
regeneration
was
not
required
then
a
lower
capacity
pump
could
be
sourced
with
reduced
flow
for
a
given
speed,
enabling
the
intensifier
to
be
redesigned
with
a
lower
area
ratio,
so
reducing
its
size
and
weight.
A
reduction
in
flow
would
also
decrease
pressure
losses
in
valves
and
pipework.
If
a
regenerative
function
were
1400
1200
A4000
w
800
0
.:-
600
400
200
0
0
100
200
300
Back-pressure
(b,
0.8
0.7
0.6
0.5
>,
0
0.4
^D
._
-
0.3
0.2
-
0.1
.
0
400
500
600
)ar)
Fig.
7.
Variation
in
work
in,
work
out
and
efficiency
with
back-pressure.
-I
i
w
5
Fig.
9.
Circuit
diagram
of
modified
VBS
with
regeneration

061215-063
to
be
required
then
a
more
complex
seawater
valve
arrangement
would
be
required.
Fig
9
shows
how
this
could
be
implemented
with simple
off-the-shelf
components.
The
check
valve
arrangement
is
replaced
with
four
two
way
two
position
normally
closed
solenoid
actuated
spring
return
valves
so
that
flow
between
the
pressure
intensifier,
buoyancy
vessel
and
ambient
can be
controlled.
The
microprocessor
will
control
the
valve
sequencing
so
that
seawater
can
be
pumped
from
buoyancy
tank
to
ambient,
or
from
ambient
to
buoyancy
tank
via
the
pressure
intensifier.
Standard
off-the-shelf
valves
can
be
sourced
for
ambient
pressures
of
up
to
300bar,
but
higher
pressures
would
require
specially
designed
and
built
valves.
[6]
D.
C.
Webb,
P.
J.
Simonetti,
and
C.
P.
Jones,
"SLOCUM,
An
Underwater
Glider
Propelled
by
Environmental
Energy",
IEEE
Journal
of
Oceanic
Engineering,
Vol.
26,
No.
4,
pp.
447-452
October,
2001.
VI.
CONCLUSIONS
The
system
is
simple
in
construction
and
operation,
and
it
is
easily
modified
to
meet
different
user
requirements.
The
results
showed
that
the
system
worked
at
its
design
flow,
but
with
excessive
friction
losses
resulting
in
a
relatively
low
efficiency.
Efficiency
was
increased
by
modifying
the
system,
but
could
be
increased
further
by
redesign
and
change
in
system
specification.
APPENDIX
A
Table
I.
Graphic
symbols
Description
Four
way-three
position
direction
cotrol
valve
Three
way-two
position
direction
contro
valve
Tvo
way-two
postior
direction
control
valve
Pressure
relief
valve
Symbol
D
.
escription
<
Filterlstrain
er
Flow
control
valve
ELEB
J
Pressure
ntensifier
I
Bleed
point
I
iNor
returnfchec
kvalve
HydrauiTc
pumnp/motor
Enc
losure
bounar
CI
jT,1
Gas
charged
accumulator
Self
seating
coupling
mi
Solenoid
operator
AA
Spring
operator
Pres
ure
gauge
REFERENCES
[1]
P.
M.
Bagley,
I.
G.
Priede,
A.
J.
Jamieson,
D.
M.
Bailey,
E.
J.
V.
Battle,
and
C.
Henriques
C2004.
"Lander
techniques
for
deep-ocean
biological
research"
Underwater
Technology,
vol
26
no.1,
pp3-11,
2004.
[2]
A.
Jamieson,
"Autonomous
lander
technology
for
biological
research
at
mid-water,
abyssal
and
hadal
depths",
Ch
5,
PhD
Thesis,
University
of
Aberdeen,
Scotland,
2004.
[3]
P.
M.
Bagley,
M.
A.
Player,
A.
J.
Jamieson,
"A
buoyancy
control
system",
Patent
No
W02005019021,
2005-03-03,
2005.
[4]
G.
Griffiths,
J.
Jamieson,
S.
Mitchell,
K.
Rutherford,
"Energy
storage
for
long
endurance
AUVs",
Proceedings
of
ATUV
Conference,
IMarEST,
London,
16-17
March,
pp.
8-16,
2004.
[5]
M.
Worall,
P.
M.
Bagley,
A.
Jamieson,
A.
Holford,
R.
D.
Neilson,
M.
Player,
"the
development
of
a
variable
buoyancy
system",
Proceedings
of
WMTC,
8-10
March
2006,
London,
UK.
Symbol
It
1--
I
-I
Fluid
lines
P
ot.
lines
1
.
t
'r
6
Electric
mO