You are on page 1of 21

Marine Technology, Vol. 17, No. 1, Jan. 1980, pp.

8-15

Shaft Alignment Methods with Strain Gages and Load Cells


Robert B. Grant 1

The paper presents methods of aligning propulsion shafting using load cells for direct bearing reaction mea-
surement and procedures for shafting alignment by measurement of shaft bending moments with strain
gages. Additionally, a unique means of measuring reduction gear and stern tube bearing reactions is de-
scribed.

CURREN T PROCEDURES for main propulsion shafting align- on the jack, is seldom as good as + 1.0 percent of full load. Second,
m e n t are well documented and in constant use at shipyards and and more importantly, the gage measures hydraulic pressure and
repair facilities. With smaller vessels the gap and offset method not the actual load on the piston. The piston seal in the jack
is convenient and yields satisfactory results. In large vessels, transmits some of the load from the piston directly to the cylinder
where disassembly of couplings is not practical, the practice is wall and thus bypasses the gage reading. This error can easily be
to measure the shaft weight supported by all accessible bearings l0 percent of the load on the jack and will make a considerable
on the shaft, using the "jack check" method for determining addition to the hysterisis curve when data are plotted to deter-
bearing reactions. mine the bearing reaction.
This paper introduces alternative methods of achieving the A straightforward method of eliminating these measurement
desired shafting alignment, utilizing electronic load cells for the problems is to incorporate a strain-gage load cell into the jacking
direct measurement of bearing loads. Additional methods in- setup as shown by Fig. 1. The load cell is a reliable rugged device
corporating strain gages for both lateral and vertical alignment which can be purchased in a wide range of configurations and load
are discussed. Finally, for measuring bearing reactions within a capacities. Accuracies of =t:0.25percent full load are standard with
reduction gear box or stern tube, a method based on external hysterisis specified at less than 0.05 percent of full load. The load
measurements is described. cell must be electrically connected to a readout device which
typically has a digital display calibrated in pounds.
Background The flexibility in load cell configuration will permit the as-
sembly of a small high-pressure jack and the load cell into an
The need for deviation from a straight-line alignment condition integrated unit for convenience of installation. An additional
and reference to the necessary backup calculations is contained benefit is the ease with which calibration can be accomplished
in the SNAME hook, Marine Engineering, Chapter XI. The using a universal testing machine in a testing laboratory. The load
shafting calculations referred to in this paper were prepared by cell is simply set up in the machine and connected to its readout.
a modified digital computer program based on the program As load is applied, the readout and the testing machine load are
originally developed at the Boston Naval Shipyard. Use of a recorded and compared for conformance with the specified ac-
computer program of this type is essential to complete the re- curacy.
quired calculations for alignment work. The program output
includes:
• Bearing reactions with all bearings in line Direct bearing reaction measurement
• Deflection and slope influence numbers
• Bearing reaction influence numbers Load cells under bearings. During new construction where
• Shear force at discrete points the bearing chocks have not been installed, load cells can be used
• Bending moment at discrete points to make direct measurements of bearing reactions. In this method
• Slope at discrete points the low-profile load cell is located between the bearing housing
• Deflection at discrete points and the foundation as shown by Fig. 2. One half of the weight is
• Weight at discrete points supported by the two jacking screws and the other half is on the
• Stiffness at discrete points load cell. The readout must be corrected for this as well as for the
Additional runs can be made to calculate bearing reactions with weight of the bearing housing since it may be an appreciable
specific bearing elevations above a straight line. The program can portion of the total load on the load cell. If the load cell location
also calculate the influence numbers which are needed to calcu- is to be offset from the shaft as shown in Fig. 2(a), it will be nec-
late actual bearing reactions when it is not possible to locate the essary to insure that the lateral alignment is very good because
jack adjacent to the bearing during the jack-check procedures. in this location the load cell will respond to lateral forces as well
as vertical forces due to shaft load on the bearing. This problem
can be overcome by locating the load cell on the shaft centerline
Improved jack-check procedure forward or aft of the bearing center so that lateral forces are
Common practice is to use a "calibrated" jack to measure reacted by the jacking screws and not sensed by the load cell.
bearing reactions. The calibrated jack as a measuring instrument With the load cell in place and the readout located within view
has two problems which bear on the accuracy of the measure- of the machinist, small bearing height adjustments can be made
ment. First the pressure gage, usually calibrated in pounds of load while the load is being monitored. If elevation changes greater
than 5 mils are necessary, best results are obtained by inserting
Diehl and Lundgaard, Inc,, Bainbridge Island, Washington. shims under the load cell in addition to the jackscrew adjust-
Presented at the October 14, 1978 meeting of the Pacific Northwest
Section of THE SOCIETYOF NAVALARCHITECTSAND MARINEEN- ments.
GINEERS. Using a load cell at each pedestal bearing on the shaft in this

8 0025-3316/80/1701°0008500.41/0 MARINE TECHNOLOGY


m a n n e r permits a substantial reduction in manpower and time
to complete the alignment procedures compared with the pro-
DG
IT
IALREADOUT
cedure using the jack-check method. An additional important
benefit is the ability to detect the effect of a bent shaft or out-
of-square coupling faces. To make these measurements the shaft
is rotated through three or four revolutions using the jacking gear
and stopped each 45 deg of rotation. A plot of the load cell data
will show a sinusoidal variation in the bearing reaction if the shaft
is bent or has a flange which is not square. This procedure, of
course, can be done by a series of jack checks but it is time-con-
suming and not as accurate.
A specific example where this type of measurement was made
involved a ship with an electric drive motor located at the forward LOAD
CELL
end of the lineshaft. In this case the motor manufacturer estab-
HYDRAULIC
lished a requirement to perform a "swing check" to verify that JACK
the lineshaft-to-motor coupling was square within specified
limits. This swing check was to be made by supporting the for-
ward end of the motor shaft at its normal elevation on a sling
attached to the overhead. With the sling rigged and the bearing
shell removed the forward end of the shaft would be free to move
in a horizontal plane as a consequence of shaft runout during
rotation by the jacking gear. In this case the specification limit
was 0.012- in. (0.304 ram) total indicator reading (TIR) as mea-
sured with a dial indicator at the bearing position.
As an alternative to this procedure the test was accomplished
using a load cell, which had been placed under the forward motor
bearing for alignment purposes, to measure the variations in load
as the shaft was rotated. A plot of the output data is shown in Fig.
ECEg
3. The peak-to-peak load variation was converted to an equiva-
lent runout using the bearing influence number in the relation-
ship: HAND PUMP
L
Runout'riR -
G × 15-5 Fig. 1 Hydraulic jack with load cell

LOAD CELL

JAB
CO
KN
LITG
~~--
S JACKING _ _
BOLTS

LO
CEA
LD
L

LOAO - /
[ CELL CELL
a
b
Fig. 2 Load cell installation below bearing

JANUARY 1980 9
WEIGHT ON LOAD
CELL - Ibs.

29,000

800

600
[------ EQUIVALENTRUNOUT 0.0072 in.

400

200

28,000

800

600

27,400
L
I I I 1
SHAFT REVOLUTIONS

Fig. 3 Shaft runout test data

where Moment method for shaft alignment


L = peak-to-peak load variations, lb There are several problems with the jack-check procedure for
G = bearing geometry constant shaft alignment which make the use of an alternative method
I5-5 = bearing influence number from lineshaft computer attractive. Some of these problems are as follows:
program • Numerous individual jack checks are necessary to complete
The geometry constant takes into consideration the height of the an alignment procedure particularly when several bearings must
bearing centerline above the load cell and the spacing between be adjusted.
the load cell and the jacking bolts. • Often some disassembly is necessary to gain access to a
L o a d cells in p l a c e of b e a r i n g shells. A shafting system forward gear bearing and to loosen gear bearing caps.
which has a gearbox presents a problem for bearing reaction • The procedure does not afford a means for verifying lateral
measurements using load cells since it is not practical to place the alignment of bearings.
entire weight of the gearbox on three or four points. As an alter- • Inherent inaccuracies in the jacking procedure are caused
native the feasibility of locating the load cells directly in the by:
bearings has been investigated and found practical. The proposed - - s h i f t i n g load centers in adjacent bearings,
method incorporates a pair of clevis pin load cells located within - - i n a c c u r a t e dial indicator readings due to motion of the in-
an assembly which temporarily replaces the bearing shell. W i t h dicator base caused by deflections in surrounding members
one load cell assembly in place in each main bearing as shown in as the load is transferred from the bearing to the jack,
Fig. 4, a direct measurement of bearing reactions is possible while - - h y s t e r i s i s in the shafting system,
final adjustments are made to the gearbox position. A bronze pad - - j a c k measurement errors if a load cell is not used, and
on the surface of the load cell arrangement would permit rotation - - m e a s u r e m e n t errors because the jack cannot be located at
of the shaft for other parts of the alignment procedure. Instal- the bearing centerline
lation of the load cells in the gearbox is straightforward and in- Most of the problems cited with the jack-check procedure can
cludes the following major steps: be overcome with the use of load cells fo r direct bearing reaction
• Remove bull gear bearing caps and upper bearing shells. measurements. For the situation where load cells cannot be in-
• Employ bridge gages to accurately establish the shaft posi- stalled because chocks are in place or because special-configu-
tion relative to the gear casing at the bearing split line. ration load cells are not available, the moment method has special
• Jack up the shaft and roll out the lower bearing shells. advantages. Although the moment method does overcome
• Locate the load cell assemblies in place of the bearing problems with the jack-check and direct bearing reaction pro-
shells. cedures, the reader will recognize t h a t it is more d e p e n d e n t on
• Remove the jacks and check the shaft position relative to the calculations and requires a somewhat more complex measure-
bridge gages. ment technique. In this method, measurements of bearing re-
• Place shims in the location provided in the load cell assembly actions are replaced by a set of bending moment measurements
to return the shaft to the same position it occupied when resting which are combined with computer calculations to find the
on the lower bearing shell. bearing reactions. For the initiai alignment of a propulsion shaft
A cross section of the clevis pin load cell is shown in Fig. 5. The the procedure is as follows:
design of the clevis pin load cells, which are sensitive to shear 1. Complete normal a l i g n m e n t calculations for the
strain, is such t h a t they are sensitive to loading in one axis only. straight-line condition and develop a table of moment influence
Because of this single-axis sensitivity, this measuring arrange- numbers.
ment will sense only vertical loads on the bearings and will not 2. Calculate the desired final bearing elevations using the
respond to horizontal loading caused by lateral misalignment. normal bearing influence numbers.

10 MARINE TECHNOLOGY
I

0
Ii
l ,I o.L
i

© 0
L u_z] LL~

SHEAR PIN LOAD CELLS

Fig. 4 Load cell installation CENTERLINE


in reduction gearbox
DIMENSIONEQUAL / ........ SHAFT
TO BEARINGSHELL . . . . ~ 4 ; r BRONZEPAD
THICKNESS "~I

- \

>

3. Select measuring stations on the shaft for the installation


of strain-gage bridges.
4. Using the desired final bearing elevations, calculate the
final bending moments using the moment influence numbers.
5. Attach strain-gage bridges to the selected measuring sta-
tions on the shaft.
6. Measure the bending moments in the initial condition.
7. Compute the difference between the initial and final mo-
ments at each station.
8. Using the moment influence numbers, calculate the mo-
ments which will exist as each bearing in sequence is positioned;
that is, moments at each station when Bearing 1 is in its final ELECTRICAL
position, moments when Bearings 1 and 2 are in the final position, X STRAINGAGELOCATIONS CONNECTION
moments when Bearings 1, 2, and 3 are in their final positions,
etc.
9. Move the bearings one at a time, observing the moment
readout for indication that the final position has been
achieved.
Steps 1 and 2 are the same as those for alignment by either
jack-check or load cell methods except that a table of moment
influence numbers must be developed. This is done by successive SHEARPIN LOADCELL
computer program runs with bearings raised one at a time above Fig. 5 Cross section of clevis pin load cell

JANUARY 1980 11
35 48 B1
42

600 700 900 IO00 1100

MA -49,000 i n - l b MA -52,000 i n - l b
--A
MB -66,000 in- Ib MB +3,000 i n - l b MA -74,000 i n - l b

MB ÷|76,000 i n -l b

® BEARING LOCATIONS

A BEARINGS ON STRAIGHT LINE MA -22,000 i n - l b

B BEARINGS ELEVATED MB -139,000 i n - l b


5 + 0.010 in
6 + 0.020 in
7 + 0.055 in
8 + 000
9 + 000
i:
--B
_m
Z
m
-4 Fig. 6 Shaft bending moment diagram
m
('}
Z
z
0
t-
O
.<
the straight-line condition. The differences in bending moments Table 1 Moment influence numbers
at the selected locations are then tabulated and divided by the
a m o u n t of bearing elevation used in the calculation. Table 1 is Bending Moment (in.-lb per O.O01-in.rise of bearings)
an example of a moment influence number table.
3 4 5 6 7 8
During the alignment procedure, shaft bending moments are Bearing N°°
Station N 20 24 35 42 48 57
measured with strain-gage bridges attached to the shaft at the Bearing N °
selected measuring stations. In order to avoid ambiguous results, 3 -7890 +3590 -820 +220 -60 +30
the number of these stations must be equal to the number of 4 +3520 -4120 +1990 -540 +140 -110
-810 -2730 -2730 +1720 -440 +300
bearings supporting the shaft. In selecting the stations, a moment 5 +200 -570 +1520 -2620 +1660 -1190
6
diagram constructed from the computer output data, such as 7 -70 +180 -360 +1630 -2640 +4890
shown in Fig. 6, will be helpful in choosing locations which have 8 +50 +140 -300 -1040 +2610 -20880
the greatest change in moment between the straight-line and final 9 -40 +100 -200 +700 -1310 +16900
alignment condition.
Locations for gages should be readily accessible for the actual no additional bearings forward of the gearbox or aft of the stern
gage installation and must not be close to discontinuities such tube, the load on these bearings may be determined by a
as couplings or shoulders, which cause stress concentrations and straightforward calculation based on a pair of moment mea-
resultant inaccuracies. surements. The measurements must be taken at stations which
Once the measurement locations are defined, Step 4 is to cal- lie between the gearbox and the first spring bearing or between
culate the bending moments at each of these stations for the final the stern tube and the last spring bearing. If the shafting ar-
aligned condition using the bending moment influence number rangement does not permit this, a third bearing may be included
information. in the shaft section under study; however, with three unknown
For strain-gage work on shipboard shafting, the weldable type bearing reactions it will be necessary to make three m o m e n t
of gages have proved to be very convenient. Temperature and measurements and to solve three simultaneous equations. Re-
cleanliness problems during installation are not severe as in the ferring to Fig. 7, the calculation method is as follows:
case of bonded gages. Also, the gages can be prewired in the lab- 1. Moments are measured at Stations A and B: Mam a n d
oratory prior to the start of onboard work. When the gages are Mbm
installed, it will be convenient to have sufficient wire length so 2. Moments about A and B due to the weight of the shaft and
that the readout of all bridges can be at one location. Gage wiring gear are calculated; M a c and Mbc
diagrams and suggestions for zeroing the readout are contained 3. Moment equations are formed:
in Appendix 1. It is worth noting at this point that it is essential
that the shaft be capable of rotation through 180 deg during the ZMa = 0
gage zero adjustment procedure. Mare + (R1 • S1) + (R2- $2) + Mac = 0
With initial bending moment data from each station, Step 6, ZMb = 0
the differences in bending moments at each station are calcu-
M b m + (R1 . $ 3 ) + (R2-$4) + Mb~ = 0
lated. This information combined with the moment influence
numbers yields a set of simultaneous linear equations, equal in 4. The solution to these equations gives the bearing reactions
n u m b e r to the number of bearings in the system, in the form directly. An example of this calculation is contained in Appendix
2. Note that hand calculations are possible a n d ' t h a t a computer
A M A = (I1.A X D t ) + (I2-A X D2) + (I~-n × D3) . . . calculation is not required.
+ IN-A X D N Two major uses for this procedure are evident: First in mea-
suring gear bearing reactions both hot and cold without the need
AMB = (I1-B X D1) + (I2-B X D 2 ) . . . to do any disassembly to get to the forward end of the shaft for
AMx = (Ii.x × D~). . . a jack check, and second in providing a means to measure stern
bearing weardown while the vessel is afloat. In the latter case the
where bearing influence number of the stern bearing on itself would be
used to convert the change in bearing reaction to a displace-
AM = measured moment. Subscripts A, B, C etc. refer to ment.
measurement stations
I = moment' influence number. Subscripts l-A, 2-A, etc.
Lateral alignment of shafting
refer to the effect of Bearing I on Station A, Bearing
2 on Station A, etc. The strain-gage bridges and readout equipment may also be
D = change in bearing elevation in mils; subscripts refer to used to perform a very accurate alignment of the entire shaft
bearing numbers
These equations can be solved in a rigorous fashion to yield the RI
reaction
required change in elevation for each bearing or an intuitive ap-
proach can be taken to give an approximate solution. Gearbearing
J
Step 8 is an aid to the process of changing the elevation of each
bearing. With knowledge of the moments at each station as each
bearing is brought to its final position, the actual movement of
R2
s
reaction
Dimensionsfrom
the shafting
arrangement
~

,
~
S3
,
--]
S4 - - - - - - - ~ ' -

bearings can proceed quite fast. As each bearing is moved, the diagram
readout is switched to the adjacent strain-gage bridge so that the
effect of the elevation change can be observed as it occurs, thus
eliminating a cut-and-try approach.

A B
Gear bearing reactions R R1

An extension of the foregoing method, actually a partial solu- - - ~Gear


tion to the moment alignment procedure, may be applied to re-
duction gearbox and stern tube bearing loading. If the shaft has Fig. 7 Free body diagram

JANUARY 1980 13
G2-
G4 k I ] G3
~ Gl
G3

Fig. 8 Arrangement of gages on shaft


l SO1O8~L ~IY]O
.~c

]
g

b.
Sgl 0C6C

Sgl ZlZll
Sgl ~9 Old:) - - ~ ~ i u ~ < v
system in the lateral direction. When aligned so that all bearings
are in a single vertical plane, the shaft will have zero bending
moment in this plane. This condition is easily verified by the
strain-gage readout. As in the case of the vertical alignment to
a desired condition, the lateral alignment can be based on the
solution to the simultaneous equations or it can be accomplished
by trial. Experience has shown that it is possible to detect de- Sgl ~9E~
viations which are less than 0.001 in. (0.025 mm).
SBI ~£ OldD "~'-"~=:~
Summary
Applicability of the alignment methods discussed in this paper
will vary from vessel to vessel with the cost and availability of
special equipment a major consideration. Where sufficient work
is to be undertaken to warrant purchase of standard and special sgl ~£Z~
load cells, the direct load measurement method will yield the best
results. A relatively small investment will permit the addition
of load cell to the jack-check procedures with an a t t e n d e n t im-
"o
provement in accuracy. For an initial alignment, the m o m e n t O
method offers the possibility of improved accuracy with the ex- Sgl ~C 01d3 h E
penditure of fewer machinist man-hours to complete a satisfac- cD
tory alignment. The ability to conduct a rapid and accurate lat- E
.O)
eral alignment is a side benefit of this method. Where stern
bearing weardown or bull gear bearing loads are in question, the Sgl 9~Z~ .~
moment method presents a unique means of measurement which
can be accomplished in m i n i m u m time on board. Because of the
rapidity with which these measurements can be made, it is a (,9
practical means of determining bearing reactions in both the hot O
,i-,
and cold conditions.
Sgl ~£ OldD ----~

sg3 L~6~
Appendix 1
Application of strain gages
In making propulsion shaft bending moment measurements,
a strain-gage bridge configuration is recommended because
- - i t provides four times the output signal of a single gage, o Sg'l 9109
- - t h e circuit is insensitive to temperature changes in the shaft, P~
and O'IdD
- - t h e circuit cancels out the effect of longitudinal strains in
the shaft.
Sg] 69C~ -
A © Plus excitation

A output
Sgl og~6 dO~d

Minus excitation _o ~
B output
_J eo

Fig. g Wiring diagram

14 MARINE TECHNOLOGY
STATIONS 48 57 60 _ 69
$60_48 = 156.38 in.

I FT --I ,1 1 $69_48 = 191.13 in.

$60_57 = 10.38 in.

$69_57 = 45.13 in.

R60 _ _ ~ R69 "

BULL GEAR
Fig. 11 Free body diagram

The physical arrangement of gages on the shaft is shown in Fig. Combining these equations, the relationship of shaft bending
8 and the wiring of the gages and connections to the readout m o m e n t to bridge output is
equipment is shown in Fig. 9. V~.F.M.C.N
Positive bending defined as concave up ~ j results in com- Vo = 4EI
pression of Gages 1 and 3 and tension in Gages 2 and 4, resulting
in "A" output becoming positive relative to "B" output. T:Cpical bridge output voltages will be in the range of 100 to 1000
Since strain gages cannot measure absolute values of strain but microvolts. A good-quality general-purpose laboratory digital
do measure changes in strain, a special technique is required to voltmeter will provide readings of adequate accuracy for the
"zero" the readout equipment. The procedure is to rotate the purpose. Alternatively, special-purpose strain-gage instrumen-
shaft so that the bridge is symmetrical about the vertical plane tation can be used.
and to note the output reading. Next the shaft is rotated exactly
180 deg and the output again recorded. Physically this shaft
rotation has caused a bending stress reversal in the shaft and Appendix 2
would have resulted only in a sign change in the output if the
bridge had initially been balanced to zero output at zero strain.
With the two readings, R 1 and R 2, in hand, the correct zero point Example of gearbox moment calculations
is calculated as As a means of demonstrating the application of the m o m e n t
method to gear alignment, the reduction gear on a geared turbine
(R; - R2)
Z=R1 ship has been used as a model. The arrangement of this shaft and
2 the station numbering system used are shown in Fig. 10. Values
This value may be added to all subsequent readings or the of bending moments at two shaft stations, 48 and 57, were taken
readout may be offset by this amount so that the addition is from a lineshaft computer program output. Moments at Stations
eliminated. In this procedure it is necessary to correctly follow 48 and 57 (Fig. 11) due to the weight of the shaft and bull gear
the algebraic sign rules. were calculated using the tabulated shaft dimensional data which
Equations used in converting the observed strain-gage readout were input to the tineshaft program. A comparison of the bearing
voltages to shaft bending moments along with a definition of the reaction from the hand/pocket calculator computation and that
symbols used are as follows: obtained from the lineshaft program shows almost exact agree-
ment. The slight difference is attributed to round-off errors in
SYMBOL REPRESENTATION UNITS the calculator.
S bending stress in shaft lb/in. 2 Moments:
M bending moment in shaft !b-in. • Due to weight of shaft and gear
I moment of inertia of shaft section in.4 based on hand calculations: Mw4s = 2 613 097 in.-lb
c radius of shaft section in. Mw57 = 374 200 i,n.-lb
E modules of elasticity lb/in. 2 • Measured with gages: Mm4s = - 1 3 8 720 in.-lb
6 strain in./in. Mm57 = +174 963 in.-lb
V,, bridge output volts Mm+(S×R6o)+(S×R69)-Mw=0
Ve bridge excitation volts ~M4s = 0 - 1 3 8 720 - 156.38R6o- 191.13R69 + 2 613 097 =0
N n u m b e r of active gages ...
~M57 = 0- 174 963 -10.38R6o - 45.13/~69 + 574 200 = 0
AR/R
dimensionless, R69 = 11 865 lb
F gage factor A L / L usually ~ 2.0
R6o = 1 321 lb
where: • Results obtained by the lineshaft computer program:
R = gage resistance
L = gage length R69 = 11 841 lb
Mc R6o= I 3 0 4 1 b
Shaft bending stress: S -
I
Metric conversion factors
Relation of stress to strain: S = E~ 1 in. = 25.4 mm
Ve.F.E.N 1 lb = 0.45 kg
Strain gage output: Vo = 1 in.-lb = 0.113 N • m
1 lb/in.2 = 6.89 kPa

JANUARY 1980 15
Marine Technology, Vol. 17, No. 1, Jan. 1980, pp. 16-28

U. S. Navy's 3KSES--artist's conception

Noise and Vibration Control Techniques for the U. S. Navy


3000-LT Surface-Effect Ship
Gary L. Fox 1

The Navy 3000-LT Surface-Effect Ship (3KSES) Program has provided a major advance in the development
of high-performance ship technology in many areas. One such area is noise and vibration control tech-
niques applicable to lightweight/high-power vehicles. New analytical methods have been evolved and sub-
stan{ial te~sting accomplished to support certain theoretical aspects of the analysis or to provide the neces-
sary empirical data. This paper presents a summary of the technical approach used in the 3KSES Noise
and Vibration Analysis, a description of the systems installed to achieve the desired acoustical control, and
comparison of the predicted ship noise and vibration environment with the Navy specification. The noise
sources considered are those related to the major machinery, that is, the gas turbine engines, the large lift
fans, and the waterjet propulsors.

l Supervisor: Ship Noise and Vibration, Rohr Marine, Inc., Chula Vista, Presented at the March 28, 1979 meeting of the San Diego Section of
California. THE SOCIETY OF NAVAL ARCHITECTS AND MARINE ENGINEERS.

16 0025-331618011701-0016500.51/0 MARINE TECHNOLOGY


LIFT ENGINE ~ ~ I ~----LIFT FAN
EXHAUST ~ INLETS

LIFT FANS
COMBUSTION AIR k - - L I F T SYSTEM
~---PROPULSION INLET ~ _ _ REDUCTION GEARBOX
GAS TURBINE
LIFT SYSTEM GAS TURBINE
WATERJET
PROPULSOR
& REDUCTION GEARBOX
Fig. 1 3KSES general arrangement

Introduction relative dominance of the airborne and structureborne noise


components, hull vibration levels, and the amount of attenuation
SELECTION of the noise and vibration control suit begins im- required by the various ship system requirements. Because range
plicitly during the first stage of ship system design. The general is a top-level requirement, optimization of the acoustical control
ship arrangement plays a major role in dictating the required suit is based on attenuation effectiveness versus treatment
effectiveness of an acoustical control system by determining the weight.
proximity ,of the noise- and vibration-producing equipment to The effect of the various acoustical control elements is included
the habitability areas and critical operational spaces. The type in subsequent calculations to evolve an acceptable treatment suit.
of propulsion system selected, with the implied noise and vi- A computer program was written based on the acoustical analysis
bration levels and operating frequencies, together with the flow to aid in the iterations of calculations. This program calcu-
characteristics of the hull structure, also tends to predetermine lates the SPL for the nine octave bands in each space of the
the unsuppressed ship noise and vibration environment. ship.
Once the general arrangement and major machinery are se-
lected, the stage is set for the acoustical analysis. The first step
is to calculate unsuppressed ship noise levels. Aeronoise power
levels for the major machinery are used directly to determine the
General arrangement and main machinery
resultant noise power level transmitted to the various spaces of The general arrangement and major machinery of the 3000-LT
the ship due to the airborne noise path. (long ton) surface-effect ship (3KSES) is shown in Fig. 1. The
The intrinsic structural noise power for each machine is cal- major machinery is located principally in the extreme port and
culated based on the aeronoise power levels and the physical starboard spaces of the ship. In the forward spaces are the lift fans
characteristics of each machine. The structureborne noise, used which supply air to the cushion and seals. Near the center is the
in conjunction with structural path attenuation and hull plating lift fan gas turbines. The combustion air inlet, also located
radiation efficiency, provides the basis for calculating the aero- amidships, supplies air to lift engines and to the propulsion gas
noise power transmitted into a space due to the structureborne turbines located farther aft. The waterjet propulsors, two port
path. The structureborne noise-level is also used to calculate the and two starboard, are located at the stern of the ship:
structural vibration acceleration spectral density at various The topside arrangement shows the combustion air inlets, the
points in the ship. lift engine exhaust stack, and the lift fan inlets located remote
The resulting noise and vibration level predictions of the un- from the superstructure spaces.
suppressed ship are then compared with the specifications. Each The main deck general arrangement features passageways
space is categorized with a required maximum sound pressure running the entire length of the ship. The lift fan spaces, com-
level (SPL) for octave bands from 32 to 8000 Hz. The applicable bustion air inlet, lift engine exhaust, and electrical generators are
U. S. Navy specification is modified OPNAVINST 9330.7A thereby separated from the interior working spaces such as the
(proposed). The modification included the addition of an " H " noise-critical Combat Information Center (CIC). The lift fan
category space. These spaces are normally unmanned and have rooms are two decks high and therefore are noise sources for both
no SPL requirements--e.g., remotely monitored engine rooms. the main deck and second deck.
The acoustical treatment suit is derived by considering the The second deck is similar to the main deck in that the major

' JANUARY 1980 . 17


Table 1 Value of Cr for lift fan Table 2 Value of Cpfor waterjet propulsor

Frequency, Frequency,
Hz 32 63 125 250 500 1000 2000 4000 8000 BFI a Hz 32 63 125 250 500 1000 2000 4000 8000
Cf 35 35 35 34 32 31 26 18 10 3 Cp 61 61 62 62 63 65 62 59 54

a Blade frequency increment.


power levels are used to predict the intrinsic structureborne noise
machinery and ducting are again separated from the hotel by a as well as the spectral density of the machinery random vibration.
passageway. The lift fan gas turbines are located amidships on This is a recent development supported both analytically and by
this deck. The central location of these engines makes it the experiment [5].
d o m i n a n t noise source for a large portion of the ship. The com- The ratio, R, of aeronoise power to intrinsic structureborne
bustion air inlet influences both 2nd and 3rd deck spaces. power is derived from the following semi-empirical formula:
The third deck houses much of the auxiliary equipment as well
as the stern seal air duct and the propulsion system. Note in Fig. R ~-- 1.5 (p.~)(t)(v + 300) (5)
pc
1 that the waterjet propulsors are located at the stern of the ship where
on the third deck, relatively remote from the living quarters. p = density of air
Working areas are again located inboard of the major ma- c = velocity of sound in air
chinery. (p.~) = average density of machine casing material
(t) = average thickness of machine casing
M a c h i n e r y source noise and vibration levels v = frequency, Hz
For purposes of analysis, the radiated noise from the large Any consistent set of units may be used in equation (5) except
airmoving machines is separated into three components: the inlet, that c must be in distance/seconds.
the machinery casing, and the exhaust. In some cases these data
have been measured directly or indirectly; in other cases, only U n s u p p r e s s e d ship c a l c u l a t i o n s
the total noise power is known or can be calculated. For example,
the gas turbine noise levels have been measured separately but Acoustical control systems are installed only when analysis
the lift fan noise components are calculated based on the com- shows that noise or vibration levels or both exceed criteria lim-
bined noise level. i t s - s o m e w h a t like zero-base budgeting. The unsuppressed ship
L i f t fan noise prediction. The total noise power level, Lt, is noise calculations are useful beyond mere determination of ex-
calculated as [1] 2 cessive levels. If the proper analytical techniques are adopted at
the start, these calculations serve as diagnostic inh)rmation to
Lt = C/+ BFI + 10 logQ + 20 logP (1) help select the most effective control system.
where Q is the air volume flow rate, ft3/min, P is the differential The technique adopted separates the receiving space noise into
pressure in inches of H20, and C/is given in Table 1. The value three source components: the outside airborne path, the inside
of BFI, the blade frequency increment, is added only in the octave airborne path, and the structureborne path. Interrogation of the
band into which the blade passing frequency falls. relative influence of the various components gives the necessary
In order to calculate the relative magnitudes of the compo- insights to select the type of t r e a t m e n t required. A schematic
nents, assume that the known total noise power level, Lt, is made diagram showing the various noise paths for a typical section of
up of the casing, Lc, inlet, Li, and exhaust, Le, such that the 3KSES is shown in Fig. 2. A flow chart, Fig. 3, shows the
calculation steps required for each receiving space in the ship.
Lt = Lc + Li + Le (logarithmic addition) (2) The flow chart allows for acoustical control elements to be added
during design iterations.
It was judged that the casing noise is about 3 db less than the
exhaust noise, and that the inlet noise is about 2 dB less than the
Airborne noise path. Noise intensities from both the inside
and outside paths are influenced by bulkhead transmission loss.
exhaust noise. It follows from equation (2) t h a t the following
The outside airborne path is additionally influenced by the source
component noise levels are
proximity, mufflers, and shell plating transmission loss. Trans-
Le = Lt - 3 mission loss (TL) of a simple panel is primarily related to the
Li = Lt -- 5 (dB) surface density, M~, and is approximately [6]
(3)
Lc = Lt - 6 T L = 23 + 15 logM~ (6)
Waterjet propulsor noise prediction. Noise radiated from for the frequency bands of 125 through 2000 Hz.
the waterjet propulsor is due only to the casing. Three formulas Because of the importance played by the T L of panels in the
considered for predicting the propulsor noise level showed large noise analysis, substantial testing was performed on complex
differences in the higher-horsepower region [2-4]. The general panels characteristic of the 3KSES. Empirical values were then
form of predictive airborne noise level, Lp, is used in the calculations. An example of measured T L is shown
Lp = Cp + A logHp (dB) (4) in Fig. 4.
Structureborne noise path. The structureborne noise
where Cp is a function of frequency octave band given in Table analysis is much more complicated and less well understood than
2 and Pip is the horsepower of the pump. The large differences the airborne counterpart. The ability to predict structurenoise
at high-horsepower level are due to the different value of the power based on the aeronoise power of a machine is a major step
coefficient, A, which varied between 7 and 20. Because of the in the understanding of this important area.
effort expended to develop a cavitation-free, efficient p u m p de- In addition to the intrinsic structurenoise power, three addi-
sign, a numerical average of 12 was used as a conservative esti- tional calculations must be completed to relate machinery noise
mate for the horsepower dependency of p u m p noise. to sound power radiated into a receiving space. These are (a) the
Intrinsic structurenoise prediction. Casing-radiated noise power insertion loss of the machinery supporting structure, in-
Numbers in brackets designate References at end of paper. cluding the foundation and any isolation system, (b) acoustical

18 MARINE TECHNOLOGY
STRUCTUREBORNE NOISE PATH

noise paths

' %

power flow through the ship structure, and (c) the radiation ef- lations. Vibration levels can, however, be derived from the power
ficiency of a structural bulkhead panel. flow equations at any step in the analysis.
For the purpose of unsuppressed ship calculation the power The unsuppressed ship sound pressure levels calculated for
insertion loss, Ip, due to machinery supported structure can be the CIC is shown in Fig. 7. The various components as well as the
considered simply a structural discontinuity. At this stage in the combined sound pressure levels are compared with the applicable
analysis, there is little difference between Steps (a) and (b) of the Navy specification. It is clear that acoustical control techniques
foregoing. Noise travels relatively u n a t t e n u a t e d in a structure will be needed to meet the specification. The general character-
except for discontinuities. Depending on the type of structural istics of the control system can be derived by inspection of the
junction, the acoustical power is partly reflected or split into relative levels of the components. It is clear that, because of the
multiple paths. An example of this, taken from reference [7], is dominance of the structureborne noise, vibration isolation sys-
shown in Fig. 5. tems should be considered. The excessive levels due t~ outside
The final step in the structureborne noise analysis is to cal- airborne path imply muffling requirements or bulkhead trans-
culate the radiation efficiency of a vibrating panel. The theo- mission loss augmentation or both. The inside airborne path
retical expression of the radiated power, Brad, of a plate radiating implies the need for source room transmission loss augmenta-
on both sides divided by structureborne noise power, PSB, is tion.
Is] By analyzing a number of selected spaces in the ship, a basis
is formed for the selection of a trial system.
Brad
-- 2 p C f f r a d / ( p s 2 7 r f ~ t + 2pCffrad) (7)
PSB
where Acoustical control elements
p, = surface density of plate The three general types of control elements used on the 3KSES
~t = damping factor are described as well as the predictive techniques used to repre-
Grad = radiation ratio sent each element in the acoustical analysis. It is the domain of
p = density of surrounding medium the systems analyst to decide the relative worth and use of the
c = speed of sound in surrounding medium elements on a ship system basis. For example, if the number of
f = frequency source noise spaces is less than receiving spaces, the most cost-
effective and weight-effective means to control noise is to apply
Because equation (7) is dimensionless, any consistent set of units the acoustical control treatment as near to the source as possible.
applies. The value of arad, obtained from reference [9], is pre- Due to practical limitations in source noise treatment, it may be
sented in Fig. 6. required to install distributive noise control systems as well.
Unsuppressed ship analysis. It is worthy of note that ship Vibration isolation--technical a p p r o a c h . The model, Fig.
system noise analysis uses power flow equations exclusively. This 3, used to predict the structureborne noise power inserted into
technique was inspired by the Statistical Energy Analysis method the ship hull via the machinery mountings and foundations in-
(for example, reference [10]). Using the machinery airborne cludes the use of vibration isolators. Because of the contribution
power levels, the most common source noise measurement, the of structureborne noise, it would be futile to treat only the walls
power levels in a receiving space can be calculated directly. There of the machinery spaces. For the purpose of noise and vibration
is no need to calculate vibration levels at any step in the calcu- control, it is essential to minimize the transmission of vibrational

JANUARY 1980 19
i DETERMINE
SOURCE AIRBORNE (A/~)
I
DETERMINE
SOURCE STRUCTUREBORNE (S/B)
]
NOISE POWER SPECTRb~ NOISE POWER SPECTRUM

CALCULATE I
EQUIPMENT MOUNTING SYSTEM
INSERTION LOSS

I
i
DETERMINE MOST

i
' DETERMINE MOST
DIRECT INSIDE PATH DIRECT OUTSIDE PATH FOUNDATION
TO RECEIVER SPACE TO RECEIVER SPACE INSERTION LOSS

I
LOSS BASED ON NUMBER
I
1,
CALCULATE NOISE
TRANSMITTED THRU
I
1
DIRECT STRUCTURAL PATH
AND TYPE OF PARTITIONS SHELL PLATING TO RECEIVER SPACE
SEPARATING SOURCE ROOM

1
EACH FROM RECEIVING ROOM
SOLACE
FOR
EACH
RECEIVING I EALCULATE INTERSECTION LOSS I
LOSS BASED ON S~BER DUE TO NL~BER OF
SPACE
AND TYPE OF PARTITIONS STRUCTURAL INTERSECTIONS
TRANSMISSION & REVERBERATION LOSS SEPARATING SOURCE BETWEEN SOURCE AND
DUE TO RECEIVING FROM RECEIVING ROOM RECEIVING ROOM
ROOM TREATMENT

I
T
CALCULATE
NOISE LEVEL IN
I CALCULATE
TRANSMISSION & REVERBERATION LOSS
DUE TO RECEIVINU
I
1
EFFICIENCY OF RECEIVING
ROOM STRUCTURE
COMPARTMENT DUE TO
ROOM TREATMENT
A/B INSIDE PATH

1
CALCULATE

I I I I
ADD 3dB TO NOISE LEVEL IN TRANSMISSION & REVERBERATION LOSS
COMPENSATE FOR COMPARTMENT DUE TO DUE TO RECEIVING
FLANKING PATHS A/B OUTSIDE PATH ROOM TREATMENT

NOISE LEVEL IN
S/B
COMPARTMENT DUE TO
PATH
I

ADD 3dB TO
COMPENSATE FOR
FLANKING PATHS
ALL SUM LEVELS FOR ALL
SOURCES SOURCES AND PATH TYPES

Fig. 3 Compartment noise prediction flow diagram

40--

TL
f
o

lO0
FREQUENCY (llz) I000 lO,O00

Fig. 4 Transmission loss of basic structural panel

20 MARINE TECHNOLOGY
30

25

20

"1o
15
<I

10

0.1 l.o lU iuo

a = t21t 1
Fig. 5 Attenuation offlexuralwaves atplatejunctures(in absence oflongitudinalwaveinteractions),asfunction ofthicknessratio [7]

energy from machinery to ship structure as far as can be accom- equations, however, must be in the form of p o w e r ratios to be
plished within the practical limits of machinery primary func- consistent with equation (5). The equations derived included the
tional requirements. impedances of the following components, shown schematically
The power flow equations for the foundation insertion loss in Fig. 8, of the system: 1) the machine, 2) machinery mounts, 3)
were developed in a manner similar to the earlier impedance intermediate subbase, 4) isolation mounts, and 5) ship foundation
techniques (see for example, references [11] and [12]). The and hull. The power ratio is calculated:

/ )--.--'.~'O

~" I0
,/
0 20

,,.,(
/
-30

40
32 63 125 250 500 1000 2000 4000 8000

CENTER FREQUENCIES OF STANDARD OCTAVE BANDS (llz)

Fig. 6 Radiation ratio (dB) versus frequency for 1/B-in. (3.175 mm) aluminum panel

JANUARY 1980 21
130

120

110

d" 100

P~
90
Fig. 7 Sound pressure level in
v
,J
the Combat Information Center
(unsuppressed ship)
80

~0

60

50
10 100 1000 I0,000
FREQUENCY (llz)

. . . . COHBINED S P L ~ O U T S I D E AIRBORNE PATlt ---O--INSIDE AIRBORNE PATII


---[]--STRUCTUREBORNE PATH --NOISE LEVELCOAL

E @, deflections give rise to large misalignmenl~s between machines


P5 _ 2 ~ , i i .2
R:~IZ1Z2 J
x connected by shafting that are mounted separately and between
gl coM1 z~• ZIZ2
i l + z i z1 :~
i .Jr zi2z~] 2 the machines and interfacing subsystems. Large relative de-
~a + E ~J44+ ~ rl~5 flections also occur between the machines and interfacing sub-
J )
(8) systems and hull structure. A common method used to keep the
machinery aligned when soft mountings are required is to rigidly
where the loss factors are defined as mount the equipment that is connected by shafting to a common
base and soft mount the base. This design approach is consistent
• 2R~ IZ~Z~l ~ with the General Specifications for Ships of the United States
,7!~,~= ~M3 Iz~z~ + z~z~ + z~z~l ~ Navy, Section 073e. However, the problem of relative motion
~/i4 = 2R~ IzAz~l 2 between the machine and the interfacing subsystems remains.
wM3 Iz~zJ4 + Z~Z~ + z~z~l ~ Based on the relative motion between the machinery and ship
hull as the limiting criteria for soft mounting, a nominal natural
and frequency of 5 Hz was selected. The calculated power insertion
Mi = mass of ith component o f j t h finite element loss, ILp, based on equation (8), is enveloped by the following
= frequency (radians/unit time) values:
Z} = impedance of ith component o f j t h finite element 19 + 5 log/ 22 < f < 707
R~ = viscous damping coefficient of ith component o f j t h finite ILp = 22+51ogf 708<f<2828 (9)
element
25+51og/ 2829<f<11313
Any consistent set of units may be used in the preceding equa-
tions. Depending on the level of sophistication needed in the where [ is frequency in Hz.
analysis, the impedance may be a lumped parameter, that is, Zi V i b r a t i o n i s o l a t i o n - - s y s t e m d e s c r i p t i o n . Particular design
= Ri + j ( w M i - K i / w ) , or complex infinite system impedances, considerations vary between the propulsion machinery and the
Zi = 2pici(1 + j ) (reference [70, where Ki is the lumped p a r a m - lift machinery due to basic differences in the nature of the com-
eter element stiffness and Pi and ci are the infinite beam density ponents, arrangement and functional constraints.
and speed of sound in the beam. A more accurate impedance can The lift system subbase, Fig. 9, is a truss-type structural
be derived based on finite, continuous-element theory similar foundation on which one LM2500 engine, one reduction gear set,
to those found in references [11] and [13]. The continuous-ele- and three lift fans are installed. The subbase is installed on
ment theory was used in the 3KSES analysis. elastomeric vibration isolation mounts which are supported by
In general, the lower the natural frequency of the machinery the primary hull structure.
mounting system, the lower the insertion of vibrational energy The subbase is constructed of steel in way of the gas turbine
from the machine to the ship structure. However, the lower the engine and gear gearbox and aluminum in way of the lift fan in-
natural frequency of the machinery mounting system, the larger stallations. The interface bet~veen the two materials is mechan-
the deflections due to internal loads and the higher the dis- ically fastened. Steel was selected for the structure in the engine
placement amplitudes due to external excitation. These large area in view of the complexity and weight required for the pro-

22 MARINE TECHNOLOGY
1.

2.
Engines (2)

Secondary
/ T o t a l Hass, H1
Space-Time Average
Free V e l o c i t y , v,
) _.• F12' V12

Isolators (I)
I ~'2
.i = R2i . . i
IK2/~

( ( 'I _._•'F•3,Vi 2.3

Fig. 8 Structureborne noise pow- \


Total~ss, M3 Space-Time Average Velocity, v3
Loss Factor, rl3
er flow diagram for force, velocity,
and impedance definitions
)/
3. Subbase

4. Primary
Isolators (J)
/ _ _ _ ~ F ~ s ' V~S

Loss Factor, nS

/
$. Deck

tection of aluminum from radiant heat and fire. The subbase is commonly used in aircraft engine nacelles, was adopted for all
largely of welded construction; however, certain elements are muffling treatment.
mechanically fastened to permit removal of machinery compo- Extensive basic theory on sound transmission through uniform
nents. ducts, lined and unlined, can be found in the literature. For the
Each propulsion subbase, Fig. 10, supports a pair of propulsion principal wave only, an approximate solution is obtained. It is
engines. The allowable misalignment in the shaft coupling to- shown that this virtually plane wave is the least-attenuated wave.
gether With the relatively long shaft will permit the required T h a t is, waves other than the one whose wave front is normal to
amount of relative motions of the subbase. Vibration isolation the duct axis (the principal wave) are attenuated more. The
of the pump for noise purposes was not necessary. least-attenuated wave usually is the most i m p o r t a n t acousti-
The propulsion engines subbase consists of a welded steel beam cally; the relation for duct attenuation, Ad, is [14]
system designed to provide a structural foundation for one pair Ad = 4.3 + KPL/S (dB) (10)
of propulsion engines and the power shaft bearing pedestals. The
where
structural elements are arranged and sized to provide a satis-
factory balance of maximum stiffness with minimum weight K= acoustic conductance ratio of duct walls
within the constraints of the engine mounting, primary ship P = perimeter of the duct cross-section
structure, and accessibility requirements. Steel was selected for S = area of the duct cross-section
the subbase structure in view of the complexity and weight in- L = length of duct
curred in providing adequate protection for an aluminum
structure from radiated heat and fire hazards. Since K is dimensionless, any consistent set of units for the ele-
The subbases are installed on an isolation system consisting ments of the ratio (PL/S) may be used. Conductance ratio, g, is
of elastomeric mounts generally similar to the standard Navy a property of the duct walls which often can be calculated ap-
"Series E" type but specifically designed to provide the required proximately and is measurable. For the important case of opti-
resiliency characteristics. The mounts are designed to function mally tuned linings, g is simply (pc/r), the inverse of the acoustic
satisfactorily in the high ambient temperature. resistance ratio at the the resonance frequency of the absorptive
I n t a k e and exhaust m u f f l i n g - - t e c h n i c a l approach. The liner. A theoretical restriction on this relation is that the acoustic
noise power transmitted via the outside airborne path is domi- admittance be small.
nant in the superstructure spaces and external stations. The Intake and exhaust m u f f l i n g - - s y s t e m description. The
muffling treatment must produce the required attenuation with arrangement of the combustion air inlet treatment, shown in Fig.
the minimum flow restriction and maintenance requirements. 11, consists of three sets of acoustically absorptive treatments.
As in the case of the structurenoise previously discussed, the One set is installed at the inlet entrance and one set each dedi-
approach was to accomplish maximum suppression of the nero- cated to the propulsion engines and lift engine, respectively.
noise as close to the source as possible. The effective design of Additionally, an acoustically absorptive lining is installed on the
absorption installations of the Helmholtz resonator array type, inner faces of the inlet structure.

JANUARY 1980 23
LIFT (

rBBASE
,UMINUM
I WAY OF Fig. 9 Lift system and subbase
,NS) (starboard side shown, port
similar)

:RUCTURE

su
(s
WA
ENGINE)
WATERTIGHT
BULKHEAD P E N E T R A T I O N

The splitters provided at the inlet entrance are arranged in Bulkhead transmission loss augmentation--technical
"eggcrate" fashion. The splitter panels are constructed of alu- approach. Vibration isolation was not expected to entirely solve
minum honeycomb sandwiched over a solid aluminum center- the noise problem in remote spaces. Acoustical barrier panels are
plate and between perforated skin panels. The absorptive panel required in the receiving spaces as well as in the machinery
sets dedicated to the particular engine installations are similarly spaces.
constructed but arranged in simple array. The structural panel t r e a t m e n t system consists of the insula-
T r e a t m e n t of the lift fan room inlets consists of splitters ab- tion panels installed on bulkhead, overhead, and deck areas for
sorptive on two sides, as shown in Fig. 12. The splitter design is transmission loss augmentation and reverberation control.
similar to the combination air inlet. Panel construction is designed to satisfy both acoustical and
Lift fan exhaust t r e a t m e n t is limited to the seal system air fire protection requirements. The noise control aspect consists
ducting running to the aft seal. T r e a t m e n t consists of the in- of the integration of noise barrier and reverberation control
stallation of internal splitters at the fan exhaust and a wralSping materials. The system is designed to be sufficiently robust to
of the outside of the duct with damping material, as shown in Fig. facilitate installation, resist impacts resulting from installation
13. The two splitter rings and outer wall lining are similar to the and normal usage, and to facilitate the penetration of distributive
acoustical t r e a t m e n t of the air inlets. system elements.
The internal acoustical t r e a t m e n t in the propulsion exhaust Due to variations in the acoustical and fire protection re-
consists of an absorptive splitter, a small centerbody, and wall quirements with respect to local shipboard areas, a variety of
t r e a t m e n t as illustrated in Fig. 14. The absorptive splitter ma- panel designs is necessary. Five types of panel have been used for
terial is Inconel 625 brazed honeycomb. An unperforated steel the noise control system in the 3KSES. The utilization of these
center sheet is between the two absorptive faces. External panels is as follows:
treatment of the ducting is limited to thermal requirements but Type A panel--is designed to satisfy the requirement for noise
will provide some augmentation of duct wall transmission loss. transmission loss augmentation and reverberation control in
The acoustical treatment of the lift engine exhaust consists of living spaces which do not require extensive fire protection.
one set of acoustical splitters; Fig. 15 shows the arrangement. Type AR panel--is identical in construction to the T y p e A

~ .,--FT9A-2A P R O P U L S I O N ENGINE
(OR LM2500)

Fig. 10 Propulsion engine subbase


(starboard side shown, port similar)

RUCTURE

24 MARINE TECHNOLOGY
ACOUSTIC SPLITTER
ABSORPT~fE PANELS
~O~/ LINING~ /IA ~ ~IRING

DLMISTER--/
.a... /""-.
SILENCERPANELS

Fig. 12 Lift fan inlets


DECK PLATING

Fig. 11 Combustion air inlet

ACOUSTICAL DUCT WALL

ACOUSTICAL SPLITTER RING S ACOUSTICALDUCT WALL

.. AMP ING TREATMENT /

/
liilf!:it; i£ " "'~ ACOUS S TT RING

Fig. 13 Lift system stern seal air duct Fig. 14 Propulsion engine exhaust

JANUARY 1980 25
panel but is resiliently mounted to enhance the suppression of
reradiated structurenoise in spaces bounded by ship structure
in which a high flux density of structure noise is present.
Type B panel--is designed for use on the structural bounda-
ries of living spaces which require a large measure of fire pro-
tection in addition to noise transmission loss augmentation and
reverberation control.
Type C panel--is designed for use on structural boundaries
requiring a high measure of noise transmission loss augmentation
but having no requirement for structural fire protection or rev-
erberation control in the space.
Type D panel--is designed specifically for installation on the
inboard structural boundaries of the lift tan spaces. The panel
HSTIC.AL ° SPLITTER RING is intended to provide both intensive transmission loss aug-
mentation and reverberation control in a major noise source
area.
Type Epanel--is simply standard acoustic board lining ma-
ACOUSTICAL DUCT WALL terial and is used to provide a measure of reverberation control
where no other structural t r e a t m e n t is required.
The construction of a typical panel, designated Type A, is
shown in Fig. 16. The Type A panel is designed to satisfy the re-
quirement for high noise transmission loss augmentation and
reverberation control in living spaces which do not require ex-
tensive fire protection. High T L can be achieved only by resil-
iently mounting the panel to the vibrating bulkhead. The
i'] transmission loss of typical bulkhead panel (AR) t r e a t m e n t is
.I
shown in Fig. 17. The effect of noise "leaks" and "shorts" should
be taken into account when the T L of a space boundary is cal-
culated. In addition, data in the 32-Hz and 63-Hz octave bands
are not available and must, therefore, be extrapolated based on
theoretical considerations.
Bulkhead transmission loss augmentation--system d e -
s c r i p t i o n . The distributed panel system for the main deck, Fig.
18(a), and for the second deck, Fig. 18(b), indicates the placement
Fig. 15 Lift engine exhaust and relative use of the various panels.
The main deck treatment is primarily to control the lift fan
airborne noise. The contribution t h a t the general arrangement
makes to the acoustical control system is clear. Panels installed
in the CIC are necessary both for airborne and structureborne
noise attenuation.
r'- The bulk of the panel treatment is located on the second deck.
The machinery density together with the living quarters prox-
imity provides the requirement of an effective acoustical control
system. AS in the main deck application, the panels are required
for both airborne and structureborne noise control.

Suppressed ship results


.020 ALUMINUM CHANNEL
With each revision of the assumed acoustical control systems,
the sound pressure level is calculated again (and again and
78" again!). Eventually the optimum system emerges. To return to
the previous example of the CIC, Fig. 19 shows the suppressed
ship noise levels.

Acoustical treatment w e i g h t

!
~ ~ .iIO" OA~IPING MATERIAL
A summary of the weight associated with the noise reduction
.o2o BACK SHEE design previously described is presented in Table 3. The major
l " FIBROUS GLASS items of the system are related to the appropriate Ship Work
Breakdown Structure (SWBS) groups used fl)r weight accounting

~
[/4" ALLLMINL~'M HONEYCOMB PANEL purposes. The total acoustical control weight corresponds to
.006 ALUMINUM SHEET
aplorbximately 5.5 percent of the light ship weight.
1" FIBERBOARD - M~L-A-2305&
OR EQUIVALENT

Conclusions
It has been demonstrated that noise levels consistent with
U. S. Navy requirements can be obtained in advanced, hfgh-
performance ships. In order to achieve this goal, however, careful
Fig. 16 Type A panel attention must be paid to the general arrangement and.ship

26 MARINE TECHNOLOGY
80 ?
/
/
/
/
70 . . . . . . . . . . . . . :. . . . . . . . . . . /
/
/
_/
60 . . . . /

H
5O /
Z
/
/
/
4o
/
I
X

30

20

I0 I IIII I I I IIII1, I I I I Illl


32 63 100 1000
FREQUENCY (Hz)

Fig. 17 Transmission loss of " A R " panel

.FT t L,FT~I~I ~-~


L~Z I' "JJ'b F °
,; ,; : : u': -~ ~ ~- e

A
I GEN 4 If',,-. ?
//' 7~,

L MK-46 ..COM, CT.~, I


J---C.~ 1 ----.JL ._ I
I
TOI/P ,,..,/~ A
MAG / / / ./"/"/XZARIll, ClC• RADAR (OVHDI
LANDING HANGAR E.._~ ~/ / /

PLATFORM AREA PASS ,' I/ /:/ '", , , ~ , , I


- ~ . E~ EQPT RM
A~ ~
~,..
F ,gNdP
I/i"

IrZ~ I ELEC ,-.,E-

~ E GEN :3 1'~..'1D- UNASGN

Fig. 1 8 ( a ) Panel distribution--main deck

m ~ a m m ~ . u D

E-/ COMB All] INTAKE


C
¥1J PUMP RM ~,~ - A

¢;;¢=ccc E

IAH ~ ~ ~ ~'IL'~ I
I
I ......
/
~o~
E

I
' - 1 ~ ~ , ~ ] I -/- c

I1 ~ 7 2~" ~ _ / I r T " Y T . I x l _"-,1~~


.
II
LN ENG
T IIM

B AIR iNTAKE

Fig. 18(b) P a n e l distributionlsecond deck

JANUARY 1980 27
130

120

II0
I\\ I

HI ii]
100 I
v
9O Ill
NOISE REI)UCTION
ItEllU IRED

CIC GOAL-~
80 " Ill

70

60

50
I0
..... I
100 1000 I0,000
FREQUENCY (|lz)
Fig. 19 Sound pressure level in the Combat Information Center

Table 3 Noise suppression system weight summary The effort reported in this paper was supported by contracts
released through the Surface Effect Ship Program Office, United
SWBS ~ Weight, States Navy, Code PMS304, Washington, D. C.
No. Description long tons
182 propulsion engine subbases 23.45 References
182 propulsion engine resilient mounts 1.43
185 lift fan subbases 9.25 1 Graham, Barrio, "How to Estimate Fan Noise," Sound and Vi-
185 lift engine subbases 8.61 bration, May, 1972.
185 lift engine and fan resilient mounts 0.54 2 Feldman, Samuel, "Correlation of Predicted and Measured Sound
251 propulsion engine inlet sound suppression 12.32 Levels in Naval Surt~ace Ships," Paper Q7, 92nd Meeting of Acoustical
panels Society of America, 17 Nov. 1976.
259 propulsion engine exhaust sound attenuation 2.76 3 Thorpe, H. A., "Compartment Noise Due to Rotary Machinery;
567 lift engine exhaust sound attenuation 0.19 Prediction of," Bureau of Ships Memorandum Report, Code 371, Series
567 lift fan air inlet splitters 1.12 371-N9, Washington, D. C., 16 Jan. 1956.
567 stern seal air duct silencers .42 4 Bolt, Beranekand Newman, Inc., Handbook [or Shipboard Noise
635 acoustical panel treatment 30.65 Control, Prepared for Naval Ship Engineering Center, Contract No.
TOTAL 90.74 DOT-CG-20756A, Washington, D. C., Feb. 1974.
5 Thorpe, H. A., "Structure/Air Noise Power Ratio Spectra of
Ship Work Breakdown Structure. Machine Casings," Paper S2/FFF15, 96th Meeting of the Acoustical
Society of America, Nov. 1978.
6 Knudsen, V. O. and Harris, C. M., Acoustical Designing in Ar-
subsystem selection at the preliminary design stage, an acoustical chitecture, McGraw-Hill, New York, 1957.
control system must be selected on a ship system basis that is free 7 Cromer, L., Heckl, M., and Ungar, E. E., Structure-borne Sound,
from parochial interests, and tenacious implementation of the Springer-Vertag, Berlin, 1973.
acoustical control suit must be carried through to the detail level 8 Noise and Vibration Control, L. L. Beranek, Ed., McGraw-Hill,
New York, 1973.
of design. 9 Crocker, M. J. and Price, A. J., "Sound Transmission Loss Using
Statistical Energy Analysis," Journal of Sound and Vibration, Vol. 9,
No. 3, 1969.
Acknowledgments 10 Lyon, Richard H., Statistical Enert,,y Analysis of Dynamical
Systems: Theory and Applications, The MIT Press, Cambridge, Mass.,
The able direction of C. M. Lee, group manager and naval ar- l§75.
chitect at Rohr Marine, is acknowledged. The technical leader- 11 Derby, Thomas, "Evaluation of Isolation Mounts in Reducing
ship of Howard Thorpe provided the foundation of the acoustical Structureborne Noise," The Shock and Vibration Bulletin, No. 46,
1976.
control system described herein. The staff of ANCO Engineers, 12 Sharton, Terry, "Frequency-Averaged Power Flow Into a One-
in particular Dr. Terry Sharton, are recognized for their contri- I)imensional Acoustic System," Journal of the Acoustical Soeiety of
bution to the understanding of mounting insertion loss. The draft America, Vol. 50, 1971, pp. 373-381.
preparation by Grace Valdez, the technical illustration by Norm 13 Snowden, J. C., Vibration and Shock in Damped Mechanieal
Systems, Wiley, New York, 1968.
James,.and the critical review of T. (Jake) Jakubowski are greatly 14 Morse,P. M. and lngard, U., Theoretical Acoustics, McGraw-Hill,
appreciated. New York, 1968.

28 MARINE TECHNOLOGY

You might also like