Structural Analysis of The U.S. Coast Guard Island Class Patrol Boat

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 26

SNAME Transactions, Vol. 96, 1988, pp.

221-246

WPB-1307 at flank speed in calm seas

Structural Analysis of the U.S. Coast Guard


Island Class Patrol Boat
Edward S. Purcell, 1 Member, Stephen J. Allen, 1 Member, and Richard T. Walker, 1 Visitor

A structural analysis of the forward bottom plating of the Island Class patrol boat was performed
using a variety of techniques, including traditional calculations, finite-element modeling, reliability
analysis, tank testing, field testing, and pressure calculations. A test boat was driven at high speeds
and sea states, causing slamming and some dishing of the plating. Strain gages were used to
record the deformation, and accelerations were also recorded. These are compared with
accelerations and pressures measured in the towing tank. Measurements of the dished plating
were used to calculate an estimated pressure that would produce such a deformation. Comparisons
of model pressures and accelerations with field measured stresses and accelerations proved to
be quite difficult to make, due to the differences in data processing techniques. Recommendations
are made for structural improvements to the vessel, including plating thickness increases and
intercostal stiffening.

IN SUPPORT of the Vice President's D r u g Interdiction w h e n necessary, search and rescue. Many of these craft
Program, the U. S. Coast Guard initiated the construction o p er at e in the harsh Gulf Stream e n v i r o n m e n t off the
of 16 Island Class patrol boats in 1985 for use in offshore nation's southeast coast, w h e r e drug traffic into the U n i t e d
patrol work. T h e missions of this craft, which is a devel- States is most prevalent. Because of this severe operating
o p m e n t of the existing Vosper-Thornycroft 33-m (110 ft) e n v i r o n m e n t , possible structural problems w e r e antici-
design, are law e n f o r c e m e n t , surveillance, boardings and, p at ed by the U.S. Coast Guard if the vessel w e r e d r i v e n
too hard in higher sea states. Thus, the i n t e n d e d service
1 Ocean Engineering Branch, United States Coast Guard Re- of these n e w craft r e q u i r e d an analysis of the effects of
search and Development Center, Groton, Connecticut.
speed and sea state upon the hull structure.
The views expressed herein are the opinions of the authors and
not necessarily those of the Department of Transportation or the Background
U.S. Coast Guard. P r o c u r e m e n t of the Island Class Patrol Boat was initiated
Presented at the Annual Meeting, New York, N.Y., November
9-12, 1988, of THE SOCIETYOF NAVALARCHITECTSAND MARINE in 1982 and is detailed by Latas [1]. 2 Th e existing Vosper-
ENGINEERS. 2 Numbers in brackets designate References at end of paper.
221
Thornycroft 33-m patrol boat, which had seen extensive Table 1 Vessel characteristlcsDIsland Class patrol boat
service in several navies, was chosen as the p a r e n t erai:t.
Vessel characteristics of the Island Class WPB are listed Length overall . . . . . . . . . . . . . . . . . . . . . . . 110 ft-0 in.
in Table 1, and Fig. 1 shows its profile view. In these, the Length between perpendiculars . . . . . . . 104 ft-0½ in.
m a x i m u m speed and range have b e e n intentionally omit- Beam, molded at deck amidships . . . . . . 21 ft-1 in.
ted. Since this was an existing design, the design loads and Depth, molded at deck amidships . . . . . . 10 ft-ll¼ in.
structural strength were u n k n o w n to the Coast Guard. In
Drag in 104 ft-~ in. LBP . . . . . . . . . . . . . 2 ft-0 in.
order to ensure reliable hull integrity over the 15-year
specified lifetime, a structural analysis p r o g r a m was un- Draft, mean to design water line . . . . . . 6 ft-5¼ in.
dertaken. Hull . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . steel
Superstructure . . . . . . . . . . . . . . . . . . . . . . . aluminum
The structural analysis program consisted of the follow- Framing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . longitudinal
ing elements: Design displacement
• F i n i t e - e l e m e n t model (FEM) of the hull shell plating (7 ft-3 in. baseline draft) . . . . . . . . . . . . 165.12 long tons
and associated structure, developed by U.S. Naval Displacement, light ship . . . . . . . . . . . . . . 117.3 long tons
U n d e r w a t e r Systems C e n t e r (NUSC) [2]. Complement . . . . . . . . . . . . . . . . . . . . . . . . . 18
Provisions for . . . . . . . . . . . . . . . . . . . . . . . . 5 days
• Structural calculations by traditional methods, per- Fresh water (100%). . . . . . . . . . . . . . . . . . . 1760 gal
formed by U. S. Naval C o m b a t a n t Craft E n g i n e e r i n g Fuel oil (95%) . . . . . . . . . . . . . . . . . . . . . . . . 10 382 gal
Station [3]. Main engines . . . . . . . . . . . . . . . . . . . . . . . . two Paxman Valenta 16
• Towing tank testing and analysis using a ~ scale model, RP200M V type; 3000
bhp at 1500 rpm {max)
conducted by the Davidson Laboratory of Stevens Shaft horsepower . . . . . . . . . . . . . . . . . . . . 2910 shp @ 802 rpm
Institute of Technology [4]. (max)
• Full-scale testing of Island Class patrol boat, con- Propellers . . . . . . . . . . . . . . . . . . . . . . . . . . . two 49.6-in. dia, 61-in.
pitch (0.7R), 5-bladed,
ducted by USCG R&D C e n t e r [5]. skewed
• Structural reliability analysis of a selected plating Generators . . . . . . . . . . . . . . . . . . . . . . . . . . two Caterpillar 3304T, 99
panel [6]. kW
• Review of all these efforts, with integi'ation of con- Maximum sustained cruising speed . . . . 26.0 knots (at half load)
clusions and recommendations.
as a result of operation at speed in severe sea conditions.
Additionally, results from this analysis could also provide
The finite-element model analysis guidance during later full-scale testing for selection of bot-
A NASTRAN (National Aeronautics and Space Admin- tom panels i n s t r u m e n t e d with strain gages and acceler-
istration Structural Analysis) finite-element model (FEM) ometers. If additional structural analyses were required,
of the complete shell plating was used to conduct struc- the FEM could focus on detailed areas. Using a fine mesh
tural analyses of the WPB subjected to h y d r o d y n a m i c plan- in these detail areas would provide more detailed analysis
ing loads, and is discussed in [2]. The FEM, which contains of specific geometries. Because the occurrence of yielding
over 5,000 elements, included all hull plating, five struc- in bottom plating or other hull structures was initially
tural bulkheads, longitudinal and transverse stiffeners, and considered a "failure," there was no n e e d to employ a
a coarse representation of the superstructure. Figure 2 n o n l i n e a r (beyond yield) analysis. NASTRAN, as described
shows a schematic of the plating elements used in the by Butler and Michel [7], was selected as the most exten-
FEM. The i n t e n t of this analysis was to identify areas of sively used and d o c u m e n t e d FEM program for linear anal-
the bottom structure which might exceed yield strength ysis.

Fig. 1 Profile view of Island Class patrol boat

222 USGG Island Class Patrol Boat


IHUUHJiUJJJJJJJlJJJJJJMJJJJJJJJiilJJIlUtlDaH

Fig. 2 Finite-element model of all hull platings, including transom [2]

Input to the FEM consisted of machinery and super- Other than the Heller-Jasper pressure distribution and
structure loads, tankage loads, and hull pressure loads as later work by Allen and Jones [10], there is little guidance
shown in Fig. 3. Machinery, superstructure and tankage as to what the exact pressure distribution should be. Data
loads were developed from design drawings and required used by Heller-Jasper and later investigators had been
only straightforward calculations of foundation loads, or obtained from hard-chine planing craft. One problem with
distributed loads in the case of tankage. However, the using the Heller-Jasper method for the Island Class boat
selection of the bottom pressure loading requires some is the fact that it has a rounded bilge and may be consid-
explanation. ered as a semi-displacement craft. One would expect the
At the time of the FEM analysis, little information was pressure distributions between hard-chine and round-bilge
available as to the bottom pressure load distribution, that planing craft to be different. Nevertheless, it was consid-
is, its magnitude and variation with time and location. It ered that the Heller-Jasper distribution provided a starting
was intended that model testing conducted at the Dav- point for our analysis and subsequent model testing would
idson Laboratory of Stevens Institute of Technology would allow its refinement.
provide some additional information. But to exercise the The actual pressure distribution used in the NASTRAN
NASTRAN FEM, an initial static load distribution whose FEM is shown in Figs. 4 and 5 from HelleroJasper. Design
effect upon the structure was equivalent to that of the pressure at a location on the vessel bottom is obtained
dynamic loading was needed. After a review of available from the product of the impact factor "'F'" from Fig. 4 and
literature, a design method for development of this equiv- the value of pressure from Fig. 5. Thus, the transverse and
alent static load distribution presented by Heller and Jas- longitudinal distributions of pressure are accounted for.
per [8] in 1960 was utilized. In this method an equivalent Initially, the value of the peak pressure Po was set equal
static load distribution is developed from knowledge of to 15 psi. Since the analysis was linear, stress values ob-
the heave acceleration at the center of gravity, vessel dis- tained could be factored up or down depending upon the
placement, waterline length, deadrise angle, and beam. final value of Po.
The equivalent static load formulation of Heller-Jasper The resulting stress and displacement data from the
provides for both a longitudinal and transverse hull pres- above analysis were post-processed to provide graphic rep-
sure distribution. A discussion of the Heller-Jasper method resentation of the NASTRAN output. Stress output is pre-
as well as modifications by later investigators is provided sented using the von Mises criterion, allowing comparison
by Silvia [9]. of output with tensile stress data. Details of the yon Mises

SUPERSTRUCTURE
LOADS

OF HULL PLATE LOADS


AND STIFFENERS

\
BUOYANCY LOAD
ON EACH ~ETTED
PLANING LOAD NODE
AS PER HELLER
JASPER

Fig. 3 Schematic representation of loads applied to WPB hull [2]

USCG Island Class Patrol Boat 223


1.0
N met. Based u p o n these results, the areas of most concern,
that is, b o t t o m plating b e t w e e n frames 13 and 17, w e r e
identified.

,/
u_ .50

--~'.25
Traditional structural analysis
Because the F E M analysis would not be c o m p l e t e d in
time to support the p l a n n e d full-scale testing, an analysis
0
utilizing traditional design p r o c e d u r e s was c o n t r a c t e d with
100 80 60 40 20 0 the Naval Sea C o m b a t Systems E n g i n e e r i n g Station in
Percent of Length from Bow Norfolk, Virginia. T h e objective of this analysis was to
p r o v i d e m o r e guidance for p l a c e m e n t of strain gages in
Fig. 4 Impact factor as a function of distance from bow [8] the u p c o m i n g full-scale testing. The scope of this analysis
was limited to major hull structural c o m p o n e n t s such as
z frames, longitudinals, deck, and b o t t o m plating. Structures
p :Po [l+cos 2n(-'~ - -~)] internal to the hull such as engine mounts or other foun-
2
dations w e r e not e v a l u a t e d for structural adequacy.
The analysis was divided among t h r e e categories. In the
first category, b o t t o m structure was analyzed for a d e q u a t e
thickness (plating), and a d e q u a t e section modulus (longi-
tudinals). In the second category, hull side structures (plat-
/ .~° \i
ing and longitudinals) w e r e also e x a m i n e d for a d e q u a t e
thickness and section modulus. A n d in the last analysis
°°o" category, the c o m b i n e d hull girder, longitudinal f r a m i n g
and plating b e n d i n g stresses w e r e c o m p u t e d .
B o t t o m structure
H y d r o d y n a m i c b o t t o m pressures w e r e d e v e l o p e d from
Fig. 5 Geometry and notation of transverse load distribution [8] the same design m e t h o d of Heller-Jasper used in the F E M
analysis. H o w e v e r , in the F E M analysis, the peak design
m e t h o d m a y be obtained in [11]. Again, the failure crite- pressure was assumed r a t h e r than calculated. Only the
rion is p r e s u m e d to b e yielding of any structural plating distributions of longitudinal and transverse pressure pre-
element. sented by Heller-Jasper had b e e n used. In the traditional
T h e hull steel is m a n u f a c t u r e d is accordance with British analysis, the actual value of pressure on a given hull panel
Specification BS4360, T y p e 43A, which has a nominal yield was calculated based upon an equation originally pre-
strength of 40 000 psi. T h e deck and superstructure are s e n t e d by Heller-Jasper [8] and later discussed by Silvia [9]
aluminum, but w e r e not r e p r e s e n t e d in the FEM. Bottom to take a d v a n t a g e of the results of m o r e r e c e n t investi-
plating r a n g e d from 7 to 10 l b / f t 2 and a nominal thickness gations.
of 0.16 in. was used for calculations involving the 7-1b plate In o r d e r to develop a pressure distribution using the
in the forward portion of the b o t t o m plating. modified Heller-Jasper m e t h o d , one must first calculate
Von Mises stress contours throughout the hull plating the accelerations at the vessel's c e n t e r of gravity (CG).
are shown in Fig. 6. Stresses in excess of 29 000 psi are H o g g a r d and Jones [12] r e v i e w e d several m e t h o d s for the
found in plating e l e m e n t s n e a r frames 13 and 17 (lightest p r e d i c t i o n of vessel accelerations. In their review, an equa-
shades), based upon the 15-psi p e a k pressure loading. Be- tion for the average of the one-tenth highest accelerations
cause several yon Mises stress values w e r e close to the at the CG was d e v e l o p e d from regression analysis of ex-
yield stress of the plating steel, the analysis was r e p e a t e d isting planing craft data. At the CG:
using a 25-psi p e a k pressure for input, with all o t h e r values
unchanged. F o r the Po = 25 input, yield values are ex- N x10C G = 7.0(H~lBpx)(1
~
+ r12) 1 (FNv)I(Lv/B vx )~2~
c e e d e d in several plating e l e m e n t s near frames 13 and 17. where
T h e results of the F E M analysis indicated that t h e r e was
i n d e e d the possibility of e x c e e d i n g plating yield strength N i C G = a v e r a g e of ~ highest accelerations at CG, g
10
d e p e n d i n g u p o n the value of the p e a k pressure Po. Al-
though the exact pressure distribution n e e d e d r e f i n e m e n t , HI = significant wave height, ft
the overall objective for conducting the F E M analysis was B,, = m a x i m u m chine b e a m , ft

Fig. 6 von Mises stress distribution on entire hull plating for Po = 25 psi and
Po = 1 5 p s i [ 2 ]

224 USCG Island Class Patrol B o a t


r = smooth-water r u n n i n g trim, deg design panel to the Spencer reference area must be found.
FNv = volumetric Froude n u m b e r = V~ [gV½]} The Spencer reference area is found from
V= speed, knots
g = acceleration due to gravity (32.2 f t / s e c 2) An = (25 A r ) / d (ft 2)
V= displaced volume, ft 3 where Ar is the displacement in long tons (LT) and d is
L, = projected chine length, ft draft in feet.
To estimate NL CG u n d e r severe operating conditions, For the Island Class vessel, AR = [(25) (165.12)] / 6.48 =
]'0 637 ft 2. For a typical panel design area of 23 by 12 in., A
a significant wave height of 10 ft and a speed of 24 knots = 1.92 ft 2. Thus, A / A n = 0.003. F r o m Fig. 7 one finds
were assumed. W h e n these values, along with ships data t h a t f = 0.7 for the ratio A / A R = 0.003. Using this value
listed in Table 1, are e n t e r e d into the above equation, an o f f , the design pressure is then
acceleration value of 0.91 g (in excess of gravity) will result.
Based u p o n the above value of N ± C G "~ 1.0 g, hydro- Po = (23.9) (0.7) + 2.9 = 19.6 psi (135.1 kPa)
d y n a m i c p l a n i n g pressures were the~ developed using the To estimate the m i n i m u m plating thickness r e q u i r e d for
m e t h o d of Heller-Jasper [8] as discussed by Silvia [9]. In the above panel, it is assumed that the plate is subjected
Silvia's paper the following formula, which has b e e n de- to a uniform pressure load of Po = 19.6 psi. For the con-
rived from Heller-Jasper, is presented for the design pres- servative assumption that all edges are fixed, the m a x i m u m
sure for a bottom plating panel: stress may readily be found using a handbook formula.
Po = f [ ( A , c o s f i ) / ( 1 4 . 5 5 L B x ) ] ( 1 + N=) + Pj, Using Case 8 from Roark [13] the m a x i m u m stress is

where max = ( B , q b 2 ) / t ~
Po = bottom design pressure, psi where
f = load distribution factor
q = loading, psi
fl = deadrise angle, deg
h, = displacement, lb b = shorter plate dimension, in.
L = length on design waterline, ft t = plate thickness, in.
fl~ = handbook factor d e p e n d i n g upon plate aspect
Bx = m a x i m u m chine beam, ft
N, = heave acceleration at CG, g For the design plate 23 by 12 in., a = 23 in., b = 12
Ph = hydrostatic pressure, psi in., a / b = 1.92, and/3~ = 0.497 using the table for Case
Using fl = 12 deg and a value of N.. = 1 from the 8 in Roark. With a factor of safety = 1.1 r e c o m m e n d e d
Hoggard-Jones acceleration calculation for H~ = 10 ft and by Heller-Jasper, the m i n i m u m thickness required for the
bottom plating to resist yielding at 40 000 psi is
V = 24 knots, the above equation predicts a design pres-
sure Po = 2 3 . 9 f + 2.89 psi. The assumption is made that tR = [(1.1 × 0.497 × 19.6 × 122)/40000]-~
the Hoggard-Jones N1 CG value may replace the N~ heave tR = 0.197 in. (5.0 ram)
acceleration in the preceding equation.
The load distribution factor, f, takes into account the Thus the predicted hydrodynamic bottom pressure will
panel size and location, and is obtained from the Allen- result in stresses in excess of yield strength for a plate
Jones pressure reduction curve shown in Fig. 7. Before t h i n n e r than 0.197 in. Using the above method, all bottom
this curve can be used, the ratio of the actual area of the plating panels were reviewed. The results showed that,

1.00

I IIIII
\ IIIII
0.80 \
\ IIIII
o
0.60
IIIII
IIIII III
LI.
g
-.=

0.40
IIIII III
o,
III
0.20
III
IIIII ~" ~' ~ , . ~ ~ F; = 0.14

0 i IIIII
.001 .005 .01 .05 0.1 0.5 1.0
NAR
Fig. 7 Load distribution factor [9]

USCG Island Class Patrol Boat 225


based u p o n plating yield s t r e n g t h alone and assuming no Secondary stress, 0" 2
allowance for corrosion or r e d u c t i o n in thickness from
Secondary stress 0"2 results from the b e n d i n g of b o t t o m
sandblasting, m a n y b o t t o m plating panels w e r e too thin.
longitudinals d u e to h y d r o d y n a m i c pressure. To calculate
As a check on the above calculation, the A m e r i c a n Bu-
the secondary stress, the m e t h o d of Heller-Jasper was fol-
reau of Shipping (ABS) criterion was used [14]. The ABS
lowed. Details of the p r o c e d u r e m a y be found in that
equation for the m i n i m u m b o t t o m shell plating thickness
r e f e r e n c e and will not b e r e p e a t e d here. I n p u t to the
is
p r o c e d u r e included N ± C G , which again was assumed to
10
t = Sd~/362 + 0.08 b e equal to 1.0. T h e resulting 0-2-value is found to be 1017
where psi.
t = thickness of b o t t o m shell plating, in. Tertiary stress, 0"3
S = frame spacing, in.
d = draft, ft T e r t i a r y stress arises from the b e n d i n g of the plating
itself b e t w e e n stiffeners. C o m p u t a t i o n of 0"3 p r o c e e d s in
F o r the Island Class WPB a fashion similar to the m e t h o d for c o m p u t i n g m i n i m u m
plating thickness. Heller-Jasper provides the following for-
t = 24 × 7.2½/362 + 0.08 mula:
t = 0.26 in. (6.6 mm) o'3 = 5.46 k p ( b / h ) 2
Thus if the ABS criterion w e r e a p p l i e d to the Island where
Class WPB, the r e q u i r e d thickness would be 0.26 in., which b = shorter of p a n e l dimensions, in.
is considerably g r e a t e r than the nominal thickness of the p = plating design pressure, psi
actual 7-1b plating [0.16 in. (4.1 mm)]. h = plating thickness, in.
Bottom longitudinals w e r e e x a m i n e d using a similar pro- k = coefficient which d e p e n d s u p o n b o u n d a r y con-
c e d u r e and w e r e found to be a d e q u a t e throughout. See ditions, aspect ratio, and point of stress mea-
the Heller-Jasper p a p e r for sample calculations. surement
W h e n the above formula is used with the Island Class
Longitudinal strength data, the resulting value of o"3 is 27 687 psi.
Following the m e t h o d of superposition used in [8], the
L o n g i t u d i n a l strength is m e a s u r e d by c o m p u t i n g the t h r e e stresses are added, giving a total of 38 659 psi. F o r
longitudinal b e n d i n g stress 0-~: the plating yield s t r e n g t h of 40 000 psi, this gives a factor
0"~ = M b / S M of safety of 1.0, which is less than the value of 1.1 rec-
o m m e n d e d by Heller-Jasper.
w h e r e Mb is the longitudinal hull b e n d i n g m o m e n t (ft-lb)
and S M the m i n i m u m section modulus at midsection Conclusions from traditional analysis
(in.2-ft). Results from the traditional analysis show that the bot-
T h e longitudinal hull b e n d i n g m o m e n t was calculated tom plating scantlings are marginal w h e n based u p o n yield
using the formula from Heller-Jasper: strength. T h e i n a d e q u a t e thickness of several b o t t o m plat-
ing panels also results in a high value for tertiary stress,
Mb = [ ( W L ) / 1 9 2 0 ] ( 1 2 8 i j J g -- 178 g c c / g -- 50) o'3. E x c e p t for the b o t t o m plating, all other hull structural
where e l e m e n t s are a d e q u a t e to m e e t the p r e d i c t e d h y d r o d y -
namic loads. As expected, the conclusions r e a c h e d from
gB = vertical acceleration at bow, f t / s e c 2 this analysis a g r e e d with the F E M results in pointing to
~)cc = vertical acceleration at CG, f t / s e c 2 several plating e l e m e n t s which might t e n d to yield. These
W = displacement, lb plating e l e m e n t s b e c a m e the subject of full-scale testing
L = w a t e r l i n e length, ft with strain gages d e s c r i b e d later in this paper.
If the above accelerations are assumed to be the a v e r a g e
of the o n e - t e n t h highest, t h e n the N ± C G value of 1 g
10
previously calculated for the condition with a 10-ft signif- Model tests
icant wave height and 24 knots m a y b e used. A similar O n e of the major uncertainties in both the F E M and
calculation m a y also be d o n e for the acceleration at the traditional structural analyses was the m a g n i t u d e and dis-
bow, as shown by Silvia [9], to yield the value N x B O W = tribution of the h y d r o d y n a m i c pressures, especially d u e to
3.4 g. Thus 10 slamming, which occur on the vessel's b o t t o m structure.
Both the F E M and traditional analyses have utilized the
Mb = [(369869)(104)/1920](128 × 3.4 -- 178 -- 50)
m e t h o d of Heller-Jasper [8] and Allen-Jones [10] for bot-
Mb = 4 151 163 lb-ft t o m design pressure despite the fact that these m e t h o d s
T h e calculation of the section modulus at midsection w e r e d e v e l o p e d from h a r d chine planing craft data. T h e r e
follows the m e t h o d outlined in any standard naval archi- is not m u c h b o t t o m pressure data for round-bilge craft
t e c t u r e textbook. T h e m i n i m u m value of the section mod- such as the Island Class patrol boat.
ulus was found to b e about the keel and is equal to 417 Bailey [15] c o n d u c t e d limited m o d e l tests on round-
in.2-ft, which exceeds the m i n i m u m section modulus re- bilged planing craft. F i g u r e 8 from [ 15] presents the results
q u i r e d by ABS standards listed in [14]. T h e resulting of several tests in which the m a x i m u m pressures are 12
p r i m a r y stress is then psi in the b o t t o m plating region 30 p e r c e n t aft of the
forward p e r p e n d i c u l a r for a s p e e d of 30 knots and in a
o'1 = 4 151 1 6 3 / 4 1 7 = 9955 psi significant wave height of 4 ft.

226 USCG Island Class Patrol Boat


0.3 L .

Location of Pressure Measurements

14
Ship length - - 100 ft +/~
S~i;i~,t4~a91 tons ~ Position
121
Fig. 9 k-scale Island Class model showing pressure gage
10 locations [4]
,=.~ %

a c c e l e r o m e t e r readings, as well as heave, pitch and re-


sistance.
t~
Rough-water test conditions
Tests were p e r f o r m e d in Pierson-Moskowitz sea states
with significant wave heights of 6.0 and 10.0 ft. Loading
conditions of the m o d e l were both light (295 800 lb) and
heavy (361 500 lb). A m i n i m u m of 70 wave encounters
was necessary to get statistically meaningful results. Speed
r a n g e d from 12 to 36 knots.
Rough-water test results
0 I I I I i Results of the above m o d e l tests are s u m m a r i z e d in Figs.
10 20 30 40 50 11-16. In Fig. 11, the average of the ~ highest acceleration
Speed in Knots at the CG is p l o t t e d against s p e e d for the light- and heavy-
0 1.0 2~0 3.0 4.0 V/'~- load conditions. As expected, the acceleration values in-
I I I I I
t I I crease with s p e e d : For the condition of 24 knots and HI
0.5 0.75 1.0 FN
= 10 ft the N 1 C G = 0.6 g (heavy displacement).
Fig. 8 Slamming pressures experienced by round bilge hull in 10
irregular head seas [15] Figures 12-15 show the longitudinal distributions of ex-
t r e m e pressure along the 5-ft buttock line for combinations
of load and wave height. The pressure distribution is seen
In o r d e r to b e t t e r define the m a g n i t u d e and distribution to be the m a x i m u m for the heavy-load condition with a
of the b o t t o m pressure on the Island Class hull, and its 10-ft significant wave height as shown in Fig. 15. The p e a k
variation with vessel loading condition, speed, and sea pressures r e c o r d e d occur at a distance 100 ft forward of
state, a series of m o d e l tests was commissioned at the the transom, which is b e t w e e n stations 0 and 1 high up
Davidson L a b o r a t o r y of Stevens Institute of Technology. on the bow flare. The pressures c o r r e s p o n d i n g to the plat-
In these tests, a ~0th scale m o d e l of the Island Class patrol ing b e t w e e n frames 14.5 and 15 are r e c o r d e d by gage 4,
boat in which several pressure gages had b e e n installed 76 ft forward of the transom. T h e range of ~0th highest
was t o w e d u n d e r various loading, speed, and wave height pressures r e c o r d e d h e r e is a p p r o x i m a t e l y 2 to 5 psi for all
conditions. T h e following p a r a g r a p h s contain a discussion load and speed combinations shown.
of the m o d e l and instrumentation, test p r o c e d u r e , and F i g u r e 16 shows a graph of e x t r e m e pressures r e c o r d e d
results. Additional details of these m o d e l tests are docu- by the panel gage. As can be seen from this figure, the
m e n t e d by Burtness [4]. m a x i m u m pressure occurs for the condition of light load
and 10-ft significant wave height, with a peak value of 13.8
I n s t r u m e n t a t i o n o f model psi (95.1 kPa).
T h e ~0th scale m o d e l with installed pressure gages is
shown in Fig. 9. Each gage had a d i a m e t e r of ~ in. and C o n c l u s i o n s from model tests
was m o u n t e d flush with the b o t t o m of the model. Eight Tile 13.S-psi value of e x t r e m e pressure r e c o r d e d by the
gages w e r e installed along the 5-ft buttock line, as shown panel gage agrees with that r e p o r t e d earlier by Bailey [15]
in Fig. 10. In addition to these eight gages, an additional for the c o r r e s p o n d i n g location, that is, a p p r o x i m a t e l y one
" p a n e l gage" d e v e l o p e d by the Davidson L a b o r a t o r y was third of the hull length aft, but it should be n o t e d that the
installed b e t w e e n frames 14.5 and 15. This p a n e l gage tests w e r e at different speeds and sea states. Also, this value
m e a s u r e d the i m p a c t pressures on a one by 2-ft p a n e l (full is lower than that p r e d i c t e d by using traditional design
scale), 1.48 ft from the keel on the port side. p r o c e d u r e (19.6 psi). Note that this pressure was calculated
In addition to the above pressure gages, a c c e l e r o m e t e r s using N ~ C G , not e x t r e m e acceleration.
10
w e r e p l a c e d at the bow, CG, and stern area. Data collected
during tests included all pressure and p a n e l gage readings, T h e Nt CG acceleration from the m o d e l tests corre-

USCG Island Class Patrol Boat 227


~--Knuckle
A c c e l e r e m e t e r 14 ~y--Sprey Chine
Aecelerometer Aecelerometer Aecelerometer ~ 5' Buttock Line
( A f t Berthlns- F 13 (o;) -1 12 ( B r l d s e ) Gages

12'WL Deck at side


iO'
S'
_~._ !
6' ?
4' ===::e==== R.
v
2'
Base- (
l i n e 10 9 8 3 \ 2 1 0
PROFILE Panel Gage

VESSEL PARTICULARs

Lensth O v e r a l l - l l O ' O "


LenBth B.P. 104'1/2"
14'WL~
S t a t i o n SpaelnE 10.4042' (nu( ¢1e [
Beam (max) 21 ' 3 / 4 " 12'
~.. /
Scale Ratio 1:20
--" Spl ay 7"_~
10: !
eI

6g
,,

2I

Baseline
8' 6' 4' 2' ~ 2' ,' 6' 8'
BODY PLAN

Fig. 10 Pressure gage locations on Island Class WPB model [4]

sponding to the condition of heavy load, H ~ = 10 ft, and Based upon this limited series of model tests, a rough
speed of 24 knots, is 0.60 g, which is less~than the value longitudinal distribution of pressure over the bottom sur-
0.91 g predicted by the Hoggard-Jones formula for the face of the Island Class hull has been presented. This dis-
same conditions. However, if the acceleration formula de- tribution is limited, however, to extreme pressures along
veloped by Savitsky and Brown [16] is used, a value of the 5-ft buttock line. There were insufficient data to ac-
N~CG = 0.66 will result, which is in excellent agreement curately develop a transverse pressure distribution, due
to model physical constraints which prevented installation
with the model test results. of additional transverse gages.

3.0
Significant ]
Load wave height (ft)l
2.5 o - - Light 6 I
n=
n - - Light 10 I
2.0 " - - Heaw 6
• - - Heavy 10

o 1.5

z 1.0 -.-Light 10"


..--Heavy 10"
.--Light 6'
0.5- ~Heavy 6'

; lo 15 2'0 2s 30 35 ,o
Steed (kno=)
Fig. 11 CG acceleration versus speed for island Class WPB [4]

228 USCG Island Class Patrol Boat


ii
I" 4- 18kt , i
O 24 kt /'
• h 30 kt ...../
i.. <: 36kt ] !,/I
.e ,x. / V
/
i
III
, ~.~. /
, ".-<..?" "\,,-4,~-.... /

t, 0t ~ 2'0 ~ 4-'0 ~ 6'0 ' 8'0

Distance from Transom, ft


Fig. 12 Longitudinal pressure distribution along 5-ft buttock line, light load, 6-ft
wave [4]

In the following section on reliability analysis, methods transmit the information via V H F back to the vessel, where
are used to infer the full-scale pressure loading from the it was recorded on magnetic tape. Processing and analysis
full-scale stress data. of the data are reported in [17]. Preliminary calculations
and results from the finite-element model indicated that
the forward machinery c o m p a r t m e n t shell plating would
Full-scale testing have high stress and it was therefore decided to place
In addition to the various analytical methods employed, strain gages on this plating and some adjacent longitudinal
full-scale testing was u n d e r t a k e n to establish a relationship framing. Accelerometers were placed in this region and
b e t w e e n hull stress and acceleration measurements• also near the center of gravity to d e t e r m i n e the levels of
Two series of tests were conducted. The first was a brief acceleration associated with any high stress recordings.
test of the first boat of the class in March 1986 as described The p l a c e m e n t of strain gages is d e n o t e d in Figs. 17 and
in [5]. More extensive testing was conducted in March 18.
1987 aboard the n i n t h boat of the class, in the waters off Strain gage amplifiers and signal conditioners were se-
Cape Henry, Virginia and Cape Hatteras, North Carolina. cured in the bilge area of the forward m a c h i n e r y space
Many sea conditions were e n c o u n t e r e d , from fiat calm to and data were transmitted via shielded cable to a tape
a 10-ft significant wave height. A free-floating wave buoy recorder. Details of the i n s t r u m e n t a t i o n are contained in
was deployed each day of the testing to collect data and reference [5]. Accelerometers were placed both at bulk-

1.3

[ ] 12 kt ,:'/
ii •t- 18 kt i,A
io O 24kt i/,;,,
h 30kt ~,},'/
•~. ~ i1~t
X 36 kt ~t
~ 8
$ 7
o. e;
E
~ 5 --:'> "" .. . . . . . x ,,%, / / " / /
w . ~_.~ ... ....~ ',...)'~;/'
,,' j
4-

2 ,~-7------7---~a~'~- , . . . i . i i

0 20 4>0 ~7;0 8C, I O0

Distance from Transom, ft

Fig. 13 Longitudinal pressure distribution along 5-ft buttock line, light load,
10-ft wave [4]

USCG Island Class Patrol Boat 229


11
.4"

I0
[] 12 kt ,/"~1"
"1- 18kt ....,"' ,,',
]
O 24kt I ,''
A 30 kt ~ /" ,
"i 7 X 36 kt //' /'/ 7 ~ \~
t4
>------..//,;I////I,
I
13. ¢ / / / /
t
E
s a - - ~ 1-~--~-~:/- ~--" /
LU
2
1 I

,2
l
~ r i r I z I ~ i I
,J 20 4-0 £,0 ~;C, 1 O0

Distance from Transom, It


Fig. 14 Longitudinal pressure distribution along 5-It buttock line, heavy load,
6-ft wave [41

h e a d 13 (station 3) and near the c e n t e r of gravity. A port- e n c e d before the major slam and the p e r m a n e n t high
able a c c e l e r o m e t e r was also used to r e c o r d peak stress level following the slam, indicating that d e f o r m a t i o n
accelerations at various locations. T h e CG a c c e l e r o m e t e r of the b o t t o m plating has occurred. The d e f o r m a t i o n (Fig.
was m o u n t e d on a stabilized platform whereas the accel- 20) is well b e y o n d construction tolerances and later in-
e r o m e t e r at station 3 was fixed to b u l k h e a d 13. H o w e v e r , spection r e v e a l e d c h i p p e d paint and rusting a r o u n d the
the pitch angles that w e r e r e c o r d e d at the CG did not p a n e l boundary.
vary m o r e than + 5 deg, so t h e r e is no significant differ- T h e g r a p h of acceleration in Fig. 19 Was r e c o r d e d over
e n c e in the readings of the stabilized versus the fixed the last 30 sec of a five-minute test run. The a v e r a g e of
a c c e l e r o m e t e r d u e to pitch angle. T h e r e w e r e higher cor- the ten m a x i m u m values from each of the 30-sec records
r e s p o n d i n g values due to the distance forward of the CG, is 0.99 g. In all other heading, s p e e d and sea state com-
and less noise in the signal from the a c c e l e r o m e t e r at the binations, the average accelerations w e r e less than this
CG. All tests w e r e r e c o r d e d with the active fin roll sta- value and stresses w e r e m u c h less, as n o t e d in the following
bilization system in use and it is e s t i m a t e d that, as a result, section. Several a t t e m p t s have b e e n m a d e to analyze these
roll angles during test runs w e r e less than - 5 deg. d a t a with r e g a r d to N ~ C G , with varying results. T h e data
In this discussion, accelerations at the c e n t e r of gr~ivity are v e r y sensitive to analog filtering; for instance, the max-
will b e considered as those m e a s u r e d from an acceler- i m u m value of 1.19 g is r e d u c e d to 0.78 g w h e n filtered
o m e t e r r e a d i n g zero g at rest. T h e m a x i m u m acceleration to 5 Hz. T h e counting of peaks b e t w e e n zero crossings is
r e c o r d e d at the CG during h e a d sea slamming was 1.19 also v e r y sensitive to filtering and in turn affects the sta-
g, as shown in Fig. 19. Note the negative gravity experi- tistic N ~ C G to the point w h e r e it is v e r y difficult to corn-

I ,~ T

0 12kt ",'~i/
14-
-I- 18 kt , ,~ ,,,
1.3
O 24 kt ; ,' ',
12 i ] "
'~1,1)
X 36kt iI;l /~ '~k
9 i~ /
I,,"/ t II
k
I
O.
)/,/; .~ 1
.\ /,X.,, ,,'~, '~
I: ./---"-.~ .,",,"d .,' ..... 1
C
Ill :,,~_..___ ....._ ...~,.. '\~.... .,,/... ,. ,.. i~.
I,,i.I ,-.-, -~-~....~.._------~..y ~ :..".~;. / /
~-_ ,., .;;;.- _. . . . . . . ~... ..,, .. ,,. ,/ t'i
'l ,j
4-
, ~ - - - ~~___-SS~~__..~-.,.~.
?~._~$ - . . . . . . . .... ., . .__--~,j' /'/~
+_ ~ ~ o.... ~ .... t
G-~-'- ~ ~ , , ~ i i i i
.7:0 4- 0 ,S0 ~ 0 I 0 C,

Distance from Transom. It


Fig. 15 Longitudinal pressure distribution along 5-ft buttock line, heavy load,
10-ft wave [4]

230 USCG I s l a n d C l a s s P a t r o l Boat


I~" i .. - ~ _ ~

I
7~ ..~z / ./
• ~I" / ~., //
t/i ill I'

L./ ~ i -\ /,

i
12 ' 't~
' -~'0 24- 28 32 3,3
Speed (knots)

1- Light Load H1/3 - 10


A Heavy Load H1/3 ,, 10
O Heavy Load Hl/3 = 6
[] Light Load Hl/3 - 6

Fig. 16 Panel gage extreme pressure [4]

Fin Stab. Sewage


1

D
t Fr 17 Fr 13
T
Rn Stab.

I-I

Deck
t

'U L
6L
7L

. t_. Longitudinal=;

Potable Water v T Potable Water


/ -Strain Gauge Locations
7 lb. Shell Plating t
(.16"nora.)
S e c ~ n at Frame 15

Fig. 17 Forward machinery compartment

USCG Island Class Patrol Boat 231


F~D p a r e full-scale accelerations with those of the model. An
a t t e m p t is being m a d e at this t i m e by the Coast G u a r d
and the Navy to standardize the software used to process
BHD 13 full-scale acceleration data.
The strain gages w e r e zero-balanced at dockside before
i!iiiiiiiiiiiii!iiiiiiiiiii 10" I "2
6
ii] A~,. (CH4) each day of testing. The calculated stress due to hydrostatic
pressure was 6000 psi at the c e n t e r of the panel that
y i e l d e d during slamming. Generally, in all conditions
iii ~i
FR 131/2 w h e r e slamming did not occur, stresses w e r e less than
12000 psi at the c e n t e r of the plating panel and less than
9 8000 psi in the longitudinals (including the "stillwater'"
== == stress). D u r i n g slamming, however, several stresses over
!i~iiiiiiiiiiiii~iiiiiiiiiii~!iii!i!iiii~ii
20000 psi w e r e r e c o r d e d at the c e n t e r of the plating panel,
i !i i i i!i i!i i i i i!!!!i!!ii i! FR 14
with occasional impulses over 60000 psi. Graphs of the
i !i i i!i i i i i i i i i i ! 'i
stress before and after the major slam indicated that the
gage was o p e r a t i n g correctly. The d e f o r m a t i o n was mea-
!iiiW~ij~i!iiz!i sured using a dial m i c r o m e t e r after the boat was hauled;
ilililili!Ta~!~iiiii!ili FR 14112
;i~i;iii~i;ili~i~i!!!i!ili!i!ii!i!i!ii!i!il; the results are shown in Fig. 20.
i i li[i i i i i i i i i i i i i i i i i i i2i !!iy 1 ili Stress at the edge of the panel can theoretically be twice
~i~}ii~i~;ii:i!i!ili!iiiiiiiii!i!i!iiiiiiii:iii
iiiii!:!i!!i!~i:ii:iiiiiiiii::il
28i as great as at the center, and a comparison of Gage 5 with
~i;~is:ilisiiiii:isiiiisi;i;ilri:i:iri:i:i T i 27 the c e n t e r rosette indicates a factor of about 1.6 in this
FR 15
i!:ri~:~::~ii~i:iii::iii:iii: 22 i~ case. Note that Gage 5 (Fig. 18) is within 1 in. (25.4 ram)
of the edge but not at the theoretical edge.
W h e n "stress" is m e n t i o n e d herein, it is the equivalent
:::::::::::::::::::::::::::::
......... uniaxial stress as d e v e l o p e d in the following m a n n e r :
:~i i i i i!i i i!i!i~)iiTramldu¢~
i !i i i i ilili!i!i i FR 151/2 1. Voltages from each of the t h r e e e l e m e n t s of the strain
(in wayA~ gage rosette w e r e r e c o r d e d on m a g n e t i c tape.
2. The t h r e e voltages were simultaneously digitized,
c o n v e r t e d to strains by calibration factor, and inserted into
FR 16 the following equation to calculate the m a x i m u m and min-
i m u m principal stresses, o', and O'q:
O'p,q

.rS E2 -L-1- -[ s+eT] e3 1 +---~1x/(E, - c3) 2 + (2E2 - el - c3) 2]


BHD 17 m m
GL CH 8 CH 12 where
, \ \ E = modulus of elasticity
Acc~(C.G.) ~ (C.G.)
v = Poisson's ratio
Fig. 18 Placement of strain gages E = strain

7
x 104 6 -
Note: Estimated 6000 psi at rest,
5" (not included here)
4-

,= 3-
L T h r o t t l e s cut back
2-
1.

0 ; , I
10 "13me (seconds) 20 30

Stress in Bottom Plating (von Mises) - - W P B 1309, 11 March 87, Head Seas, 24 kts, Sea State 4

; I

i-1 (at rest reference is at zero g)

i !
-2
10 Time (seconds) 20 30
Fig. 19 Acceleration at CG and stress in bottom plating

232 USCG Island Class Patrol Boat


t FWD

FR. 14 FR. 14

W a t e r Tank
0.133"

FR. 14 112 gJ tn
i- FR. 14 1/2

0.207" 0.232" 0.120"

W a t e r Tank 0.229" 0.308" 0.178"


(NominalPlating"rrddme~m.0.16'3

0.141" 0.209" 0.115"

FR. 15 FR. 15

Fig. 20 Bottom plating deformation measurements

3. T h e m a x i m u m and m i n i m u m principal stresses w e r e while u n d e r w a y , since the gage was z e r o e d with the still-
t h e n inserted into the von Mises [11 ] equation to p r o d u c e w a t e r b e n d i n g m o m e n t and hydrostatic pressure already
the equivalent uniaxial stress applied.
2

4. This stress was continuously p l o t t e d as a function of Structural reliability analysis


t i m e to p r o d u c e graphs such as Fig. 19.
R e f e r e n c e [6] contains the c o m p l e t e reliability analysis
T h e von Mises criterion used above neglects shear and
for a section of the shell plating in the forward m a c h i n e r y
assumes that the thin plating is subjected only to tensile
space, and the highlights of that study are p r e s e n t e d
stresses.
herein. P r e l i m i n a r y calculations had identified this as an
As an e x a m p l e of the relatively low plating stress re-
area of high stress and h e n c e it was selected for reliability
c o r d e d in sea state 3 (7.0 ft significant wave height), the
analysis.
stress values are a r r a n g e d in histographic form in Fig. 21.
Similarly, for sea state 4 (9.5 ft significant wave height),
see Fig. 22. Load profile development
Single strain gages w e r e located on the longitudinals Rather than use theoretical loads, the plating stress his-
n e a r the c e n t e r of the forward m a c h i n e r y c o m p a r t m e n t , tory from the field testing was c o n v e r t e d to a probability
as n o t e d in Fig. 18, Gages 27 and 30. Stress is directly density function similar to Figs. 21 and 22. Curve-fitting
p r o p o r t i o n a l to strain in this case and a sample plot is techniques w e r e used to d e t e r m i n e that these distributions
shown in Fig. 23. In general, stresses in the longitudinals w e r e mostly log-normal. T h e sea state associated with each
w e r e m u c h lower than plating stresses and within ac- probability distribution was c o n v e r t e d to wave e n c o u n t e r s
c e p t a b l e limits. at that p r o b a b l e stress level. This loading profile was t h e n
Stresses in the keel bar are also m u c h lower than plating a p p l i e d to an operational matrix consisting of eight com-
stresses as shown in Fig. 24. These stresses are directly binations of s p e e d and sea state in the h e a d sea condition.
p r o p o r t i o n a l to strain as m e a s u r e d by a single gage (No. Using an anticipated 15-year life at 2000 hr of operation
28). T h e n e g a t i v e values indicate that the top of the keel p e r year, a composite loading profile for the selected plat-
b a r is in compression and that the vessel is sagging slightly ing panel was developed.

USCG Island Class Patrol Boat 233


20
19
I p"] Test Data - - WPB 1309
18 /{ 13 MAR 87 - - Cape Henry
17
16 ]/1 Sea State 3 (H1/3 = 7 ft)
Head Sees - - 2 9 kts
15 r~ r~ 5-minute test
14 [ / I r/~ 30,000 data points
8c 13 r.~r~ Von Mises equivalent stress
I/~ J"/~ from strain gauge rosette at center
12 F./J ~ of Pace,, FR 14 1/2- 15, L1 - L2
11
, A i,/] m Note: Estimated 6000 psi at rest,
O
10 (not included here)
E
o~ 9
#. 8-
7
6
5
4
3 I/
2 k
1
0 . . . . . . . *'7- . . , ,
13 15 17 19 21 23 25 27 29 31 33 35 37 39

Stress (hundreds of PSI)


Fig. 21 Histograph--distribution of plating stress--state 3

Structural resistance
All these calculations were performed using loads from
An analysis of the structural resistance to this loading nonslamming conditions, since it was already known that
was then performed and a limit state equation was de- severe slamming could produce plate deformation. It can
veloped. F r o m the solutions to this equation, a probability be seen, therefore, that even without slamming, there is
of failure was determined where "failure," as before, refers a relatively high probability that yielding will occur over
to yield of the plating. For the panel of plating under the 15-year life if 7-1b plating is used.
consideration, the calculated probability of failure is 0.035 The resistance of the material, that is, the yield strength,
or 3.5 × 10 -~. In most structural engineering cases, it has is a r a n d o m variable. In order to realistically estimate the
been shown that a desirable value lies between 10 -3 and statistical characteristics of the yield stress, a n u m b e r of
10 -5, and that 0.035 is a relatively high value. For instance, carefully controlled tension tests were conducted, using
by using 9-1b plate in lieu of 7-1b plate, the value would samples of the plating provided by the shipbuilder. These
decrease to 3.1 × 10 -s, an acceptable level (see Fig. 25). samples were used to prepare ASTM (American Society

14
Test Data - - WPB 1309
13 17 MAR 87 - - Cape Hatteras
12 Sea State 4 (Ht/3 = 9.5 ft)
Head Seas - - 28 kts
11 5-minute test
10 30,000 data points
Von Mises equivalent stress
8 9- from strain gauge rosette at center
of panel. FR t4 1/2 - 15, L1 - L2
8-
Note: Estimated 6000 psi at rest,
7- (not included here)
c
E
6-
o.

5-

4-

i
1

0 m..----cTLr~
7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61
Stress (hundreds of PSI)
Fig. 22 Histograph--distribution of plating stress--state 4

234 USCG Island Class Patrol Boat


x 10 3 Strain Gauge No. 30

¢L

! r

0 10 "Ii me (seconds) 20 30

Note: Estimated 6000 psi at rest,


(not included hare)

Strain Gauge No. 27

xl0 3
O

-1

-2
0 10 Time (seconds) 20 30

Fig. 23 Stress on Iongitudinals

for Testing and Materials) standard specimens for the ten- Load effects
sion tests• A summary of the results of these tests is given
in T a b l e 2. The load effects are functions of the different types of
Note the exceptionally high ductility as reflected in the loads, the geometry and dimensions of the boat, material
ratio of ultimate strain to yield strain. Twelve samples were characteristics, and the boundary conditions for the plates.
tested, ranging from 4-1b to 20-1b plate. In addition, In order to make the reliability analysis as straightforward
hundreds of thickness measurements were obtained from as possible, a means of relating the load effects to the yield
the shipyard and a thickness distribution was analyzed for stress is required. Note, however, that yielding in a ma-
each plate size. terial is a three-dimensional p h e n o m e n o n and must be

x 10 3 5 (Slamming)
4
3
2
-- I
u)
o. 0
-I
-2

I0 Time (seconds) 20 30

Note: Estimated 6000 psi at rest,


(not included here)

(No Slamming)

x 10 3
O

-2 1 .;
10 Time (seconds) 20 30

Fig. 24 Stress on keel bar

USCG Island Class Patrol Boat 235


10 tensile in-plane forces. Mansour [18] presents a n u m e r i c a l
Note:
I solution to this equation for both compressive and tensile
Plating panel forces N~ A detailed description of that solution, the
Frame 14.5 - 15 __ b o u n d a r y conditions established, and the d e v e l o p m e n t of
Long. 1 - 2
the load effect terms are provided in [6].
The loads on the panel of plating can be generally di-
8 vided into those resulting from hydrostatics in stillwater,
and those resulting from ship's speed and motion. The
(n load information is provided principally by the strain-gage
m e a s u r e m e n t s taken during testing, where only the dy-
Q.
namic motions part of the load effect was measured. A
separate analysis p r e s e n t e d in [6] develops an estimate for
the stillwater loads. With this in mind, an algebraic ma-
nipulation of Mansour's solution into the form of the von
Mises yield criteria [11] allows the load effect terms to be
expressed as
10-s 10-4 10 .3 10 "2 10 "1 .99
= + 36p~k * ~-~ + 6Nxpk** -i5
Probability

Fig. 25 Probability of plate yielding in head seas

treated in a m a n n e r which accounts for that. For thin where


plates it is c o m m o n to assume a condition of plane stress,
L = m a x i m u m total c o m b i n e d stress at c e n t e r of
which allows the three-dimensional p r o b l e m to be r e d u c e d
plate
to two dimensions due to the very small thickness of the
N, = stillwater in-plane unit load, lb/in.
plate.
p = stillwater hydrostatic pressure
With regard to the case of yielding of the bottom plating
of a ship, the mathematical s t a t e m e n t of the p r o b l e m can t = plate thickness
be given by b = width of the plate (athwartship dimension)
Ap = effective dynamic pressure
74w = 1 ( p + N~-~x
32w2 + 2Nx~ ~32w + N~ ~W~oy2j k*, k** = algebraic combinations of Mansour's coeffi-
cients B and y

where The first t e r m in this equation represents the contri-


bution due to the stillwater condition. The third t e r m
74w = b i h a r m o n i c equation represents the c o n t r i b u t i o n of the d y n a m i c effects, and
w = plate deflection in z-direction the middle t e r m is an interaction t e r m resulting from the
p = u n i f o r m lateral pressure von Mises formulation of the problem.
D = plate flexural rigidity, given by Et3/12 ( 1 - v ) In order to make the limit state equation as simple as
E = modulus of elasticity possible and because longitudinal stress records from the
t = thickness of plate keel bar indicate that the variation in longitudinal b e n d i n g
v = Poisson's ratio stress is very small (relative to the stresses measured in
N~ = in-plane force on plate in x-direction the plate), all dynamic effects are c o m b i n e d into the Ap
N~ = in-plane force on plate in y-direction term. O n e should also note that the contributions of the
N~ = in-plane shearing force transverse force N~ and the shearing force Nx~ are consid-
ered to be small and are not included. This assumption is
Except in cases where the in-plane forces are explicitly
consistent with typical naval architecture design assump-
known, there are not m a n y closed-form solutions to this
tions.
equation. In ship structures, the principal application of
The terms N~, p, Ap, t, and b are all treated as r a n d o m
this equation is in the case of plating subjected to a large
variables. In order to make the solution of the limit state
compressive in-plane stress N , which brings with it the
equation more tractable, a slightly modified form is used
possibility of plate buckling. However, the case of interest
for the loading, L 2. The principal advantage is that it re-
in this paper presents a possibility of both compressive and
duces the n u m b e r of terms in the limit state equation:

Table 2 Summary of results of tension tests (BS-4360 steel, N~~ 6Nxb~k ** 36b4k *
Type 43A) Z = F~2 t2 t3 PI t4 P2

Coefficient of where
Property Mean Value Variationa
Fy = yield stress of material
Yield stress (ksi) 47.8 0.133 P, = ( p + Ap)
Yield strain 0.0020 0.361
Modulus of elasticity (ksi) 24,940 0.116 /°2 = (p2 + Ap2 + 2pAp) = /92
Ultimate strain
144.7 0.380 It should be noted here that e v e n though/'2 is the square
Yield strain of P~, the statistical distribution of each will not necessarily
Average percent reduction in cross-
sechonal area 61% 0.049 be the same. This results from taking the product of two
distributions. For this analysis it was found that P1 was
C0V = standard deviation divided by mean value. normally distributed while P~ was log-normally distributed.

236 USCG Island Class Patrol B o a t


The values of pressure are calculated using Mansour ([18], simulation with variance r e d u c t i o n techniques. The prob-
p. 324). F o r the assumption of an isotropic plate with all ability of yielding is d e t e r m i n e d according to the limit
edges fixed, a value of/3 = 0.004 and y = 0.041 results state equation. The statistical characteristics of the yield
w h e n using Curve No. 6 in that reference. These values stress F~, the p r i m a r y force Nx, the plate thickness t, and
are t h e n used to obtain values for k* and k**. The equation the plate length b are s u m m a r i z e d in Table 5. The statis-
then reduces to tical characteristics of the pressure terms P~ and P2 are
s u m m a r i z e d in Table 3.
N~2 + 0.0558 N~b~ P1 0.0993 b 4 Pz
Z F: Using RASCS, which is a simulation-based c o m p u t e r pro-
7 t3 t4 g r a m for reliability assessment [6], the probability of failure
This equation is solved r e p e a t e d l y for various load con- according to the limit state equation is d e t e r m i n e d . T h e
ditions. T h e resulting m e a n values of pressure are listed resulting probability of failure r e p r e s e n t s the probability
in Table 3. Note that P~ = p ÷ Ap and r e p r e s e n t s the of plate yielding due to one wave encounter. The resulting
sum of the hydrostatic and effective d y n a m i c pressures. probabilities of failure are designated as Pn, that is, prob-
T h e m e a n value c o r r e s p o n d i n g to the condition of 24 knots ability of plate yielding due to one wave encounter.
and 10-ft significant wave height (Test A108C) is 6.53 psi. The o c c u r r e n c e of waves is a stochastic process which
This is considerably lower than the design pressure pre- is without a m e m o r y . Therefore, it can be assumed to be
viously calculated using the Hoggard-Jones and Heller- a Poisson stochastic process. T h e rate of wave e n c o u n t e r s
Jasper equations, which resulted in a design pressure value r, in waves p e r minute, d e p e n d s on the speed of the ship
of 19.6 psi (135.1 kPa). T h e m e a n pressure value r e c o r d e d and wave characteristics. T h e n u m b e r of wave e n c o u n t e r s
on the p a n e l gage during the m o d e l tests at the same speed Nw in a life of 15 years and 2000 hr of operation p e r y e a r
and sea state was 1.74 psi (12.0 kPa). is given by
T h e main loads a p p l i e d to the boat are the stillwater Nw = 2000 (15)(60)(r)
load and the d y n a m i c loads. F o r each type of load t h e r e
are two types of load effects, namely, p r i m a r y and sec- T h e probability of failure P:,, due to n wave e n c o u n t e r s
o n d a r y load effects. The p r i m a r y load effect is d u e to hull can be d e t e r m i n e d assuming i n d e p e n d e n t events as fol-
bending. T h e secondary toad effect is due to plate action lows:
in the area of interest. F o r any type, the load effect can Pf,, = 1 - ( 1 - Ps,)"
be expressed as
Theoretically, the ship may encounter zero to i n f i n i t e
Load effect (e.g., stress) = [C(X)].[load] (e.g., pressure) waves during its life. H o w e v e r , the probability of any num-
T h e coefficient C(X) is a transfer function from loads to b e r of wave encounters during life follows a Poisson prob-
load effects. It d e p e n d s on loading, the g e o m e t r y and di- ability distribution of rate n. T h e probability of failure Pf
mension of the ship, b o u n d a r y conditions at the location during the life of the ship is given by
of interest, etc. Both C(X) and the w a t e r pressure are (~r)"
r a n d o m variables. T h e transfer function can be deter- e: = ,,=o
~, e:,,T exp (--/~r)
m i n e d using either analytical models or e x p e r i m e n t a l test-
ing. In this section the transfer function is d e t e r m i n e d Or it can be expressed as follows:
using naval a r c h i t e c t u r e p r e d i c t i o n models and the re-
c o r d e d stresses from full-scale testing.
In this study, only loads and load effects in head seas
are considered. No other h e a d i n g is considered because
the stress records indicate that they result in m u c h smaller w h e r e r is the life of the ship in minutes, which is 2000
stresses than the head seas condition. Eight combinations × 15 × 60 = 1 800 000 min. The probability of failure
of ship s p e e d and sea state for the head seas condition are can b e simplified to
considered, as s u m m a r i z e d in Table 4. The p e r c e n t a g e of Pr = 1 - exp ( - ~rP~ )
o p e r a t i o n in h e a d seas was estimated to be 20 p e r c e n t ,
the sum of the individual p e r c e n t a g e s listed in the table. The probabilities of failure PC and Pf are d e t e r m i n e d
P r u d e n t seamanship dictated that the ninth test not be for e v e r y plate thickness and ship s p e e d / s e a state com-
carried out. bination. F o r e v e r y thickness, the probability of failure in
head seas is d e t e r m i n e d as the n o r m a l i z e d w e i g h t e d av-
erage of the probabilities of failure in life that c o r r e s p o n d
Reliability assessment to that thickness. Each weight factor is the p e r c e n t usage
T h e probability of plate yielding at the location of in- of the ship for the corresponding ship s p e e d and sea state
terest was d e t e r m i n e d using the Monte Carlo c o m p u t e r combination. The normalizing factor is the same for all

Table 3 Statistical characteristics of PI and P2 (psi)

Test P1 Pz
Designation Mean COV Distribution Mean C0V Distribution
P401 3.09 0.185 normal 9.85 0.363 log-normal
P403 3.10 0.180 normal 9.92 0.348 log-normal
P404 3.31 0.202 normal 11.37 0.407 log-normal
C714 3.68 0.165 normal 13.90 0.321 log-normal
C715 3.63 0.188 normal 13.64 0.380 log-normal
P311 3.71 0.152 normal 14.06 0.301 log-normal
A107 7.03 0.080 normal 49.70 0.160 log-normal
A108C 6.53 0.098 normal 43.03 0.193 log-normal

USCG Island Class Patrol Boat 237


Table 4 Combinations of ship speed and sea state (head seas)

Ship Speed
Sea
State Low Medium High
Low Test P401 Test P403 Test P404
(1 and 2) (12 knots, 3 ft) (24 knots, 3 ft) (29 knots, 3 ft)
4% 1.7% 1%
Medium Test C714 Test C715 Test P311
(3 and 4) (12 knots, 8 ft) (24 knots, 8 ft) (29 knots, 7 ft)
4.7% 1.3% 0.7%
Hil~ Test A107 Test A108C No Test
h (12 knots, 10 ft) (24 knots, 10 ft)
5.3% 1% 0".3~
NOTES: 1. Wave height is H - t, the significant wave height.
2. Sum of all percentages is 20%, the estimate of time
spent in head seas.

thicknesses and is the total p e r c e n t usage of the ship in Conclusions and recommendations
h e a d seas, that is, 0.20. The resulting probabilities of plate
Certain portions of the b o t t o m plating of the USCG
yielding in h e a d seas as a function of thickness are shown
Island Class patrol boat have yielded (dished) during high-
in Fig. 25.
s p e e d s l a m m i n g events e x p e r i e n c e d during controlled
T h e full r a n g e of plate sizes is shown to d e m o n s t r a t e
tests. T h e d e p t h of the i n d e n t a t i o n is a p p r o x i m a t e l y twice
the m e t h o d , not to infer that lighter than 7-1b plate should
the plating thickness. A t t e m p t s w e r e m a d e to correlate
b e used. T h e gain of t h r e e orders of m a g n i t u d e b e t w e e n
the full-scale acceleration and stress m e a s u r e m e n t s with
7-1b and 9-1b plate cannot be ignored, h o w e v e r (3.499
the towing tank m o d e l acceleration and pressure mea-
E-02 to 3.096E-05). It changes the probability from a
surements, but difficulties w e r e e n c o u n t e r e d due to dif-
marginal condition to an a c c e p t a b l e level.
ferences in data collection and processing techniques. T h e
use of a finite-element m o d e l shows promise, but the pres-
sure profile used to load the m o d e l needs m o r e analysis
Pressure calculations from measured deflection and d e v e l o p m e n t .
The A p p e n d i x contains pressure calculations m a d e using T h e n e e d for discussion in several areas is apparent. T h e
the plating deformation from Fig. 20. These calculations use of the statistic, " a v e r a g e ~th highest acceleration," for
w e r e p e r f o r m e d at M.I.T. by T. Wierzbicki, u n d e r the a design criterion must be v i e w e d in the light of the ex-
direction of J. H. Evans. It is e s t i m a t e d that a d y n a m i c t r e m e values of stress r e c o r d e d d u r i n g s l a m m i n g excur-
pressure of 114 psi was r e q u i r e d to p r o d u c e that defor- sions. F o r example, towing tank data indicate that the ratio
m a t i o n and that the equivalent static pressure, uniformly of e x t r e m e acceleration to the average ~ t h highest is about
applied, would be 159 psi. This pressure far exceeds the 1.2, whereas the c o r r e s p o n d i n g ratio for stress in the plat-
design pressure values n o t e d in other portions of this pa- ing is about 10, from data taken during full-scale testing.
per. Note that this is a single m a x i m u m pressure value Estimates of the m a x i m u m i m p a c t pressure on the b o t t o m
and, as such, should not be used as an overall design cri- plating vary widely, including 13.8 psi from m o d e l testing
terion. A small a m o u n t of occasional d e f o r m a t i o n must be and 114 psi from calculations (Appendix).
a c c e p t e d if the vessel is to b e light e n o u g h to achieve As a result of the analysis and testing p r o g r a m , recom-
design s p e e d and p e r f o r m its mission. Problems arise, how- m e n d a t i o n s w e r e m a d e to increase the strength of the
ever, w h e n w i d e s p r e a d d e f o r m a t i o n occurs, leading to forward b o t t o m plating. F o r boats still u n d e r construction,
h y d r o d y n a m i c inefficiency and interior m a i n t e n a n c e an increase from 7-1b to 9-1b plate was r e c o m m e n d e d . F o r
problems. R e f e r e n c e is m a d e in the A p p e n d i x to the work existing boats, intercostal stiffening was a d v o c a t e d as a
of Loeser [20] w h e r e full-scale pressure testing of plating means of r e d u c i n g the aspect ratio of the b o t t o m panels
panels was conducted. In a p a n e l similar to the Island Class from 2:1 to 1:1, t h e r e b y r e d u c i n g the stress. O t h e r rec-
b o t t o m plating, a c e n t e r deflection of 2.4 in. was m e a s u r e d o m m e n d a t i o n s w e r e m a d e c o n c e r n i n g offsetting weight
u n d e r an a p p l i e d pressure of 1200 psi, without rupture. reductions in tankage and stores.
T h e r e is a n e e d to a g r e e on some standards for proc-
essing the data a c q u i r e d during full-scale testing. Selection
Table 5 Statistical characteristics of Fy, N., t, and b of transducers, signal conditioning devices, filters, and the
like, varies considerably throughout the testing c o m m u -
Random Distribution Mean nity. Comparison with m o d e l data is v e r y difficult due to
Variable Type Value COV Reference the variety of processing software and selection of statis-
/• normal
normal
47800 psi
-- 88 lb / in.
0.125
0.500
testing
Soares [19]
tics.
t, 4# normal 0.086 in. 0.0125 testing
t, 5# normal 0.112 in. 0.0125 testing
t, 6# normal 0.135 in. 0.0125 testing Acknowledgment
t, 7# normal 0.161 in. 0.0125 testing T h e authors wish to a c k n o w l e d g e the assistance of Mr.
t, 7.5# normal 0.177 in. 0.0125 testing
t, 9# normal 0.224 in. 0.0125 testing Orin Stark of the USCG Research and D e v e l o p m e n t Cen-
t, 10# normal 0.236 in. 0.0125 testing ter. His dedication to accuracy during his m a n y years of
t, 20# normal 0.486 in. 0.0125 testing field testing has e n a b l e d us to conduct our research with
b normal 11.75 in. 0.025 judgment confidence.

238 USCG Island Class Patrol Boat


References Pulse Loading," International Journal of Impact Engineering,
1 Latas, R. M. et al, "U.S. Coast Guard 110 ft. Island Class Vol. 7, No. 2, 1988.
Patrol Boats--Concept to Completion," Patrol Boats '86, Trans. 28 Wierzbicki, T., Chryssostomidis, C., and Wiernicki, C.,
ASNE, March 1986. "Rupture Analysis of Ship Plating Due to Hydrodynamic Wave
2 Clark, J. H. et al, "A Structural Analysis of an Ocean-Going Impact" in Proceedings, Ship Structure Symposium, Arlington,
Patrol Boat Subjected to Planing Loads," 15th NASTRAN User's Va., Oct. 15-16, 1984.
Colloquium, NASA Conference Publication 2481, May 1987. 29 Nagai, T. and Chuang S.-L., "'Review of Structural Response
3 Russell, M. J., "Structural Analysis U.S. Coast Guard Island Aspect of Slamming," Journal of Ship Research, Vol. 21, No. 3,
Class 110-ft WPB," Naval Sea Combat Systems Engineering Sta- Sept. 1977, pp. 182-190.
tion Code 61, 15 Aug. 1986. 30 Evans, J. H., "Preliminary Design Estimation of Hull Gir-
4 Burtness, M. N., "Hydrodynamic Bottom Loading Distri- der Response to Slamming," TRANS. SNAME, Vol. 90, 1982.
bution on a Model of the 110 ft WPB Island Class Patrol Boat 31 Frankland, J. M., "Effects of Impact on Simple Elastic
Operating in Head Seas," Stevens Institute of Technology Tech- Structures," David Taylor Model Basin Report 481, April 1942.
nical Report SIT-DL-87-9-2584, Hoboken, N.J., Aug. 1987. 32 Hovgaard, W. "Structural Design of Warships," The United
5 "'Strain Gauge Testing of USCGC FARALLON (WPB- States Naval Institute, Annapolis, Md., 1940.
1301)," U.S. Coast Guard Resarch and Development Center Re- 33 Sellers, M. L. and Evans, J. H., "Bulkhead Plating," Journal
port (Interim), Groton, Conn., May 1986. of Ship Research, Vol. 3, No. 3, Dec. 1959.
6 White, G. J. and Ayyub, B. M., "Reliability Analysis of the
Island Class Patrol Boat," U.S. Coast Guard Research Publication, Appendix
Aug. 1987.
7 Butler, T. G. and Michel, D., "NASTRAN a Summary of An analysis of the permanent deformation of a
the Functions and Capabilities of the NASA Structural Analysis bottom panel of plating on the USCG Island
Computer System," NASA SP-260, National Aeronautics and
Space Administration, 1971. Class patrol boat 3
8 Heller, S. R. and Jasper, N. H., "On the Structural Design An analysis has been performed on the bottom plating of the
of Planing Craft," Quarterly Trans., RINA, July 1960. Island Class 110-ft patrol boat to estimate the hydrodynamic pres-
9 Silvia, P. A., "Structural Design of Planing Craft, A State sure which caused deformation and permanent set in a panel of
of the Art Survey," SNAME Chesapeake Section, March 1978. plating of one of the boats when undergoing tests at sea. For
10 Allen, R. and Jones, R., "A Simplified Method for Deter- purposes of the analysis, strain-gage information was provided for
mining Structural Design Limit Pressures on High Performance the panel of plating most severely affected, also the amounts of
Marine Vehicles," AIAA/SNAME, 1978. permanent set at nine points on the panel. The maximum per-
11 Hill, R., The Mathematical Theory of Plasticity, Oxford manent deflection, at the center of the panel, was 0.308 in. A
University Press, London, 1964. mean uniform plating thickness of 0.161 in. had been determined
12 Hoggard, M. and Jones, M., "Examining Pitch, Heave and and several tensile stress-strain curves for samples of the steel
Accelerations of Planing Craft Operating in a Seaway," High used were made available.
Speed Surface Craft Exhibition and Conference, 1980. The test panel was located between web frames at frames
13 Roark, R. J. and Young, W. C., Formulas for Stress and 14~ and 15 and between longitudinal Locations Nos. 1 and 2. At
Strain, 5th ed., McGraw-Hill, New York, 1982. Location 2 a longitudinal bulkhead provided the panel support
14 "Rules for Building and Classing Steel Vessels Under 61 instead of a light longitudinal as at Location 1. Thus, the panel
Meters (200 feet) in Length," American Bureau of Shipping, 1983. had dimensions of 23½ in. in the fore-and-aft direction and 12 in.
15 Bailey, D., "Performance Prediction-Fast Craft," Occa- in the athwartship direction. It was very slightly concave from
sional Publication No. 1, RINA, 1974. the exterior, was one panel removed from the bar keel, and about
16 Savitsky, D. and Brown, P. W., "Procedures for Hydro- 28 percent of the vessel's waterline length from the stem. The
dynamic Evaluations of Planing Hulls in Smooth and Rough space inboard at the time of the test was void.
Water," Marine Technology, Vol. 13, No. 4, Oct. 1976. According to the dynamic strain records, even a number of
17 "'Structural Analysis of the Island Class Patrol Boat," U.S. substantial instantaneous peak strains translating to stresses at the
Coast Guard Research and Development Center Report, Groton, center of the panel on the order of 20 000 psi were insufficient
Conn., Oct. 1987. to cause an automatic baseline shift of the strain recorder until
18 Mansour, A. E., "Ship Bottom Structure Under Uniform one very large slam was experienced. This resulted in a baseline
Lateral and In-Plane Loads," Schiff und Hafen, Vol. 5, 1967. shift of about 34 000 psi. Thereafter the throttles were cut and
19 Guedes Soares, C., "Probabilistic Models for Load Effects the vessel slowed.
in Ship Structures," Publication No. UR-84-38, Norwegian Insti- It has been assumed, therefore, that the plastic deformation of
tute of Technology, Trondheim, June 1984. the panel plating occurred during the one impact and within a
20 Loeser, H., "Investigation of the Plastic Deformation of
Flat Plates Under Lateral Load," Unpublished M. Sc. Thesis, time span of about 2½ to 3 sec as indicated by the strain and
M.I.T., Cambridge, Mass., 1954. acceleration records immediately preceding and following the
21 Sawczuk, A., "On the Initiation of the Membrane Action major slam. At the time of the slam the vessel was traveling into
in Rigid-Plastic Plates,"Journal de Mecanique, Vol. 3, No. 1, 1964, head seas at approximately 24 knots in a state 4 sea having a
pp. 15-23. significant wave height of 10 ft. None of the panel boundaries,
22 Jones, N., "Damage Estimates for Plating of Ships and Ma- including the No. 1 longitudinal, showed signs of permanent de-
rine Vehicles," PRADS--International Symposium on Practical formation.
Design in Shipbuilding, Tokyo, Oct. 1977, pp. 121-128. There were several rigorous analyses of the plate deflection
23 Jones, N., "A Theoretical Study of the Dynamic Plastic problem in the sixties and seventies. The basic approach of the
Behavior of Beams and Plates with Finite Deflections," Inter- so-called rigid-plastic hinge line method was developed by Wood
national Journal of Solids and Structures, Vol. 7, 1971, pp. 1007- and extended to large deflection problems by Sawczuk [21].4 Jones
1029. extended the solution even further by including inertia effects
24 Jones, N., "Slamming Damage," Journal of Ship Research, [22-24]. Hooke and Rawlings [25] performed a thorough exper-
Vol. 17, No. 2, June 1973, pp. 80-86. imental study which confirmed the validity of the theoretical
25 Hooke, R. and Rawlings, B., "An Experimental Investiga- predictions. The application of the above-mentioned methods to
tion of the Behavior of Clamped Rectangular Mild Steel Plates the damage prediction of ship's plating was discussed by Hughes
Subjected to Uniform Transverse Pressure" in Proceedings, In- in reference [26].
stitution of Civil Engineers, Vol. 42, 1969, pp. 15-23. The effect of pulse shape on the magnitude of permanent plastic
26 Hughes, O. F., "Design of Laterally Loaded Plating--Uni- deflection of structures was discussed by Symonds and Frye [27]
form Pressure Loads," Journal of Ship Research, Vol. 25, No. 2, and by Wierzbicki et al. [28].
June 1981, pp. 77-89.
27 Symonds, P. S. and Frye, C. W. G., "On the Relation Be- 3 Performed by J. H. Evans & T. Wierzbicki, M.I.T., May 1988.
tween Rigid-Plastic and Elastic-Plastic Predictions of Response to 4 References cited are included in regular listing above.

USCG Island Class Patrol Boat 239


The problem of slamming damage of ship hulls was analyzed
by, among others, Jones [24] and Nagai and Chuang [29].
Summary of calculations
Input parameters (see Fig. 26): P1
Plate length, a = 23.5 in.
Plate width, b = 12 in.
Plate thickness, h = 0.161 in.
Static yield stress, o"v = 47 800 psi
Measured maximum p e r m a n e n t deflection at the middle of the
plate:
~c,~
Fig. 27 Pressure pulse representation
(W/)dy n = 0.308 in.
Calculated values:
General procedure
Plate aspect ratio b/a = 0.5
Maximum dimensionless deflection (wJh)dy, = 1.91 The present calculations are based on the general theory of
dynamic response of a plate strip subjected to exponentially de-
Unknown: Calculate the most probable peak pressure ampli- caying pressure [28] and earlier results on moderately large de-
tude Pl which caused the plate in Fig. 26 to deflect p e r m a n e n t l y flections of statically loaded square and rectangular plates
by the a m o u n t (Wf)dy n. developed in references [21] and [23].
The amplitude of dynamic pressure can be expressed in the
Discussion of loading following way.
The hydrodynamic pressure of the affected area of the hull
resulted from the combination of the hydrodynamic impact of
the waves on a vessel traveling at 24 knots into a state 4 sea
(horizontal velocity) and the slamming impact of the hull (vertical where
velocity). In-plane loadings in the plating are believed to be very
small, so only the lateral loading (vertical) will be assumed to have Pl = peak load
caused the deformation. Pc = critical quasi-static pressure necessary to produce
The extent of damage to ship's plating from the hydrodynamic plastic bending load limit of plate
impact depends not only on the peak pressure (pressure ampli-
tude), P], but also on the shape of the entire pressure pulse. Beside
WJ)dr, = normalized p e r m a n e n t dynamic deflection
P], the transient pressure pulse can be characterized by the rise f ( - ) = function accounting for large deflections and mem-
time to and pulse duration t~. Previous studies on slamming pres- b r a n e resistance
sures for high-speed vehicles in waves have d e t e r m i n e d that the R = dynamic load factor
pulse shape can be well approximated by the exponential function
(see, references [24, 28]): Using the definition of the dynamic load factor [see equation (9)],
equation (2) can be rewritten in the equivalent form
t
p(t) = P,e-7,, (1) Wj 1
P, = pcf(~)wheref(--~WJ)= f[(--~)dy -~] (2a)
The rise time is very short and the meaning of the characteristic
time ta which defines the initial deeay rate of the pulse is shown where w//h is the equivalent normalized static deflection.
in Fig. 2 7 - - a two-parameter representation of the pressure pulse: Each of the factors appearing in equations (2) and (2a) will now
Pl = magnitude p a r a m e t e r be discussed separately.
ta = shape p a r a m e t e r
Critical quasi-static pressure Pc
F r o m the data provided by the Coast Guard it is not possible to Following Sawczuk [21] and Hughes [26], the formula for Pc is
find a unique pair [ Pl, td }. Therefore, assumptions will be made
regarding the pulse duration. 48 M0
F r o m the results published in the literature [29] and our own Pc = b2 [x/'3 + ~2 _ ~]2 (3)
experience [28] we assume that the most probable duration of
the pressure pulse is in the range where M, = (o-~h2/4) and is the fully plastic bending m o m e n t
td = 50 to 500 ms and ~ = b/a is the plate aspect ratio• For a square plate a = b,
~ = 1, and
Thus, calculations will be run for the lower bound ta = 50 ms
and u p p e r bound ta = 500 ms separately. pc- a2 - 12o-~
An additional assumption made in almost all analyses of similar
problems is that the dynamic pressure is uniformly distributed
over the surface of the plate. In the case of the Island Class l l 0 - f t patrol boat, ~k = 0.5 and
from equation (3) we find that

Pc = 12o'~ a 2.36 (4/

The coefficient 2.36 represents the increase in strength of the


rectangular plate with one side half as long as the other.
Effect of large deflections and membrane forces
The expression for the function f(W//h), derived by Sawczuk
[21] has the form

f(-~Z)=2~Z 1 ' ~ ' ) ( 1 - 3 ~ ) (5)


Fig. 26 Panel dimensions

240 USCG Island Class Patrol Boat


where

This formula is valid in the range W f / h > 1. In the case of a


square plate, t~ = 1 and equation (5) reduces to
1
f = -h- + 3 Ws (6)
h
In the case of plating under consideration, t~ = 0.5 and { = 0.65, o ± 2__ .L 4
SO n n n "~
= 1.252 w , 0.748 Fig. 29 Dynamic load factor (DLF)
~ + 3 Ws (7)
h
2
or, in expanded form R -- - - (10)
rr 1
! 0.748 1+----
2 7ta
S ~-h-j = 1.252 \ h )dyn = + 3(Wi _I
(7a)

k h )~. B where the parameter 7 is defined as


y2 _ 12o'~
Finally, an approximate expression for f ( W J h ) , derived in ref-
#=p (11)
erence [28] on the basis of a purely m e m b r a n e response of a plate
strip, is and p and t' denote respectively the mass density of the plate
(inertia effects) and the characteristic dimension of the structure.
f -- 5 h (8) In the case of a rectangular plate it is reasonable to assume
a+b
A comparison of predictions of equations (6), (7), and (8) is shown e - 2 ~ 18 in. (12)
in Fig. 28. This solution is valid for the plate fully restrained
against in-plane motion at the outer boundaries. We believe that Taking p = 7.45 X 10 -4 lb s e c 2 / i n / , the parameter ~/calculated
the panel which suffered the damage of Ws = 0.308 in. satisfied from (11) and (12) is
this condition of full end fixity.
From Fig. 28 [or from equation (7)] we find that f ( W s / h ) =
2.52 if W J h = 1.91. The coefficient 2.52 represents the increase ~-['l12X 47.8 × 103 = 1.54 × 103 1
"~]
in the plate resistance due to m e m b r a n e effects. The plate strip 31 = X 7.45 X 10 -4 sec
model [equation (8)] predicts a slightly higher value, f ( W e ~ h ) =
2.54. According to equation (10), the dynamic load factor depends on
the product ")ltd. A plot of the function R versus 1 / ~/td is shown
Dynamic load factor, R (DLF) in Fig. 29. The limiting case td ~ ~ corresponds to a suddenly
applied rectangular pressure pulse. As in elastic structures, R is
The concept of dynamic load factor is similar as in elasto-dy- then equal to 2. Another limiting case, td ~ O, represents the
namics; however, the formulas for R in the case of rigid-plastic ideal impulse loading.
structures are different. We define R as According to previous estimates for the pulse duration, two
values of td should be considered
R - (Wf)dy, under a particular pressure amplitude, Pl (9)
Wi 7td = 0.050 × 1.54 × 10 s = 77
yta = 0.500 × 1.54 × 103 = 770
A detailed analysis of the dynamic load factor under various load-
ing conditions was recently presented by Symonds and Frye in The range of variation of 7ta is shown in Fig. 29 by the shaded
reference [27]. We shall adopt here an approximate formulation, zone. It is clear that in all cases the corresponding value of the
developed in reference [28]. dynamic load factor effectively is R = 2. The reason for such a
It was found in a one-degree-of-freedom model, and with the high (actually maximum) value is that the dynamic pressure was
blast pulse shape assumed (to = 0), that assumed to be applied instantaneously with zero rise time
to = 0. In reality the rise time is always finite and the dynamic
load factor will be lower. See Fig. 30. In the absence of actual
measurements of rise time, to, it is difficult to give a precise
estimate of R.
/,,. Figure 31 is from [30] and is of elastic dynamic load factors
/ derived by Frankland [31]. Again tl is the pulse duration and T
q the natural period of vibration of the system. The figure illustrates
the fact that as h / Tgrows large, the DLF for steep fronted pulses

SY ~t ....
I / *
/ l

0
t ÷
I Z 3

Fig. 28 Comparison of predictions Fig. 30 Pulse with finite rise time

USCG Island Class Patrol Boat 241


PULSE SHAPES

.t_L____~_ .kL___k
~[ to ~ I--t,--q I--t,-.q ~.--t,-.q-" ~-t,---I

i,ot Y
g

_ _ _ 2_--

o.,..,.,,c <.o.o .cTo, s CO.F)

I I I I I l I I
.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0

',/r (..,> ,o/T)


Fig. 31 Pulse shapes and DLF

(to = 0) approaches 2 whereas for more gradually applied pulses Conclusion


(to > 0) the DLF approaches a value of one. [Xo is the static
Our best estimate for the dynamic pressure (rise time finite) is
displacement of a mass under a particular force and x(t) is the
dynamic displacement under a pulse whose peak value equals P~ = 114 psi. If the rise time is long enough and approaches a
the force statically applied.] static approximation, the pressure must be higher than PI = 159
Under the present circumstances three different solutions will psi to reach the same p e r m a n e n t set. If, on the other hand, the
loading is applied instantaneously (to = 0), the pressure amplitude
be given depending on the magnitude of rise time to:
to reach 0.308-in. p e r m a n e n t set would be 92 psi.
R = 2 for instantaneous pressure to = 0
R = 1.5 for a finite rise time of pressure pulse to Discussion
R = 1 for very slowly increasing pressure to ~ o0 The present findings are in full accord with the results published
in the open literature. For example, reference [29] reports on a
Determination of pressure amplitude
p e r m a n e n t set of about 0.275 in. (7 ram) on the hull of a high-
Having d e t e r m i n e d all components in equation (2), we are in speed aluminum craft. The estimated peak pressure of the pulse
a position to give estimates on the amplitude of the hydrodynamic of 10-ms duration was 92 psi. Hydrodynamic pressures as high as
pressure loading which caused the plate to deflect permanently 200 to 300 psi were quoted by Nagai as well as Kawakami and
by the given amount. Keil. One of the present authors conducted a scale-model test on
From equation (4) we calculate the effect of breaking waves. When converted to actual dimen-
sions, the hydrodynamic pressure was found to be 330 psi [28].
(0.161~ 2 Assuming realistically short rise time of pressure, our best es-
P c = 12 × 2.36 × 47.8 × 103 × \ 2 3 . 5 ) = 63 psi
timate for the amplitude of hydrodynamic loading is
The final calculations are given in Table 6. PI = 114 psi

Table 6

Static Dynamic Dynamic


to ~ oo to > 0 to = 0 Units

Pc 63 63 63 [psi]
from eq (4)

(W}),y. 1.91 1.91 1.91 "'"

R 1 1.5 2 ...
w~ w; 1
1.91 .1.27 0.955
h R

f 2.52 1.80 1.46


from eq (Ta)
[psi]
P, = pcf(W/~ 159 114 92
\n]

242 USCG Island Class Patrol Boat


versus the equivalent static pressure to cause the same defor- deflection [33]. This requirement can be expressed as:
mation of approximately 63 × 2.52 = 159 psi. The present anal-
ysis also gives additional information on the initiation of fully h = 0.001443 b
plastic response.
where
The so-called plastic collapse pressure necessary to develop the
fully plastic bending moment of the plate can be obtained by h = plating thickness, in.
dividing the dynamic pressure amplitude by the coefficient of b = lesser panel dimension, in.
membrane action H = pressure head, feet of seawater

114 For a 12-in. panel under an equivalent static pressure head of


m

Ppl..ti~ b~.d,.g 2.52


q

45.2 psi 358 ft (159 psi) as found above, this results in a plating thickness
of 0.33 in.
In spite of the fact that the final applied pressure heads w e r e
The end of the elastic range and the so-called first yield is reached at least 5~ times those when yielding first occurred in the thick
at the pressure plates and on the order of 17 times in the thin plates, nooe of
45.2 the plates ruptured nor "showed a crack."
P,,,t y,e~d- 1.5 - 30 psi Loeser [20] attempted to explore the upper limits of plate panel
strength experimentally by means of nine test pieces of differing
shape and aspect ratio. One panel in particular had quite similar
As one point of reference, a few experiments quoted by Hov- characteristics to those of the Island Class test panel. Loeser's
gaard [32] are of interest. Thin flat plates, both square and rec- panel was 26 by 13 by ~6 in. and was of ordinary medium steel.
tangular, were loaded with increasing hydrostatic pressure and At a test pressure of 1200 psi (2700 ft of seawater) the loaded
periodically the maximum deflection was measured, with the load deflection at the panel center was 2.412 in. At a pressure of 1270
on and after its removal. Initially, the deflection was "strictly psi (2857.5 ft of seawater) the seal of the test tank failed and the
proportional to the pressure" until the yield point was reached test had to be halted.
"somewhere in the plate." Thereafter, a permanent set began to
appear "increasing slowly until it amount(ed) to about 18% of the
total deflection." "After that the permanent set increas(ed) rapidly Metric Conversion Factors
and both curves (of deflection versus pressure) rose steeply, ap- 1 ft = 0.3048 m
proximating to straight lines." During that last phase the elastic 1 in. = 2.54 cm
deflection, represented by the separation between the two de- 1 ft 2 = 0.0929 m 2
flection curves, remained "fairly constant" and at extremely high 1 ft 3 = 0.0283 m 3
pressures it was "but a small fraction of the total deflection." 1 psi = 6.895 kPa
Evidently as a result of Hovgaard's influence, the U.S. Navy 1 lb = 0.45kg
adopted plating thicknesses for ordinary subdivision bulkheads 1 gal = 3.7.8.5 L
which permit a permanent set of 18 to 20 percent of the total 1 hp = 0.7457 kW

Discussion
Hung-Chi Lee, Member T h e structural design of high-speed boats has b e e n the
[The views expressed herein are the opinions of the discusser and subject of various papers over the years and each has
not necessarily those of the Department of Defense or the De- b r o u g h t an increased level of understanding of the hull
partment of the Navy.] loading conditions and structural reliability. T h e authors
in studying the Island Class structure have co n t r i b u t ed to
I would like to congratulate t e e authors for an excellent this effort and are to be congratulated.
p a p e r p r e s e n t i n g analysis, m o d e l testing and field testing T h e Island Class was p r o c u r e d based on a " p a r e n t c r a f t "
of the structure of the Island Class patrol boat. As a f o r m e r c o n c e p t and was driven by the n e e d to p r o cu r e drug in-
project manager, I appreciate the value of this c o m p re- terdiction vessels in a very short t i m e frame. T h e overall
hensive p a p e r to anyone pursuing the design of high- approach necessitated that specific craft features, includ-
speed, lightweight boats. ing the structure, follow the p a r e n t craft design. Th e struc-
This discusser would like to make some suggestions for tural design of the Island Class e v o l v e d from the designer's
future d e v e l o p m e n t : e x p e r i e n c e and had p r o v e n successful to m er i t the con-
In order to have a m o r e c o m p l e t e set of pressure dis- struction of 23 boats worldwide. H o w e v e r , the hull struc-
tribution data for the c u r v e d boat, it will be necessary to ture b e c a m e a m a t t e r of growing c o n c e r n as the Coast
h a v e m o r e pressure data from the existing model. This Guard's operational d em an d s increased and the fleet size
includes m o r e pressure gage locations and a g r e a t e r range e x p a n d e d beyond original plans. In two and a half years,
of both speed and wave height.
the Coast Guard has put m o r e operational hours on its
E n l a r g e d details of the higher stressed areas, including first four Island Class boats than all previous 23 boats have
labeled contours, should be provided. a c c u m u l a t e d in 20 years.
Additional figures should be p r o v i d e d showing details While t h er e are some variations in c u r r e n t m e t h o d o l o g y
of the results of the analyzed stress and the m e a s u r e d stress for structural analysis, each addresses t h r ee basic elements:
in both the full scale and m o d e l tests, respectively. load d e t e r m i n a t i o n , load r ed u ct i o n based on panel loca-
What problems w e r e e x p e r i e n c e d during both the full- tion, size and allowable stress. T h e authors' focus on the
scale and m o d e l tests, and what w e r e the m e th o d s used loads show how difficult they are to d e t e r m i n e and how
in o v e r c o m i n g t h em ?
they vary widely b e t w e e n full-scale m e a s u r e m e n t s , m o d e l
tests, and empirical predictions. In the context of this
Steven H. Cohen, Member, and Debabrata Ghosh, study, the full-scale trial m e a s u r e m e n t s as used by the
Member authors are certainly most applicable. These data used in
[The views expressed herein are the opinions of the discusser and conjunction with the assumed operational profile given in
not necessarily those of the Department of Transportation or the Table 4 have allowed the authors to project structural
U.S. Coast Guard.] reliability. H o w e v e r meaningful, this approach departs
USCG Island Class Patrol B o a t 243
somewhat from the general design philosophy for small As this ABS in-house guidance is being adopted in pro-
high-performance craft where the structural design is a posed ABS Rules for high-speed craft, shortly to be dis-
function of the crew's operational limits. tributed for industry review, this paper is a valuable source
Some designers joke that boats have two speeds: " o n of correlation that will certainly be considered.
and off." It is not surprising that boats are run to the crew One extremely important fact in the results is the ver-
limits where a punishing ride is the only good reason to ification that the plate panel maximum p e r m a n e n t set of
slow down. Though continuous slamming beyond 1 g can 0.308 in. was apparently due to one single impact in the
be minimized by p r u d e n t seamanship, continuous opera- order of 1.0 g. In previous work that ABS has done to
tion at more moderate levels of acceleration are very hard predict slamming pressures from set-in plating, there was
to ensure despite the designer's strident attempts to place very often the question of whether the set-in had been
sea state limits on the craft. Small craft designed under caused by one impact or by the cumulative effect of a
the general premise that the structural design limit is to n u m b e r of impacts. A very important fact is that none of
exceed the limits of the crew is taking both a practical the bottom longitudinals, even those adjacent to the set-
view and conservative approach. Designers will argue over in plate panel, showed signs of deformation. If it has not
how high the limit should be, but most agree that 1 g is already been done, it is suggested that the ultimate bend-
a lower limit. For structure in high-performance military- ing strength and stability of the longitudinal adjacent to
type craft m u c h higher accelerations are used. the set-in plate panel be predicted and the pressure cor-
In this regard, would the authors c o m m e n t on the sen- responding to the ultimate load be c o m p a r e d with the
sitivity of their results to the assumed operational profile pressures given in Table 6 in the Appendix. As the plate
in Table 4 and how it relates to the crew limits? Would panel area and shell area supported by the longitudinal
the authors also discuss the basis for using head-seas con- are the same, a direct comparison of pressures would seem
ditions only (which represent 20 percent of the operational appropriate.
time) especially since stillwater bending loads have been On a more general point, when considering high-speed
mentioned by the authors as one of the main loads applied vessels of this type, encountering extreme conditions and
to the hull? Also, the authors might check the midship extreme loads, it would seem conservative to presume that
section calculation noting that the deckhouse should not the occurrence of plate yielding be considered a "failure."
be considered to be effective structure. ABS criteria, for Presumably because of their use of quote marks, it is also
instance, include only the deckhouse structural elements the opinion of the authors that limitations of analytical
which are continuous over 40 percent of the vessel length. methods should not establish "failure" levels.
The equation referred to in the ABS Rules is essentially
Robert Curry, Member for normal-strength steel (34 000 psi yield) displacement
The authors are to be congratulated for an excellent hulls; some reduction in the thickness could have been
paper, extending established theory through practical ap- given for the use of 40 000-psi yield strength steel. It also
plication to full-scale tests and providing the slamming seems that there may have been a misinterpretation as far
pressure assessment from set-in plating in the Appendix. as the value of S is concerned. Should the authors wish to,
In that it has provided answers to some important ques- we would be pleased to provide more detail on the fore-
tions and directed attention to others, it is a milestone going.
paper that is now, and will be in the future, a very valuable
reference on the subject. Additional references
For in-house guidance for planing craft, we at ABS use 34 Heller, S. R. and Jasper, N. H., "On Structural Design of
a m e t h o d of predicting slamming load developed from Planing Craft," RINA July 1960.
Heller and Jasper [ 34 ] and Savitsky and Brown [ 35 ] (ad- 35 Savitsky, D. and Brown, P. W., "Procedures for Hydro-
ditional references follow some discussions). For the sub- dynamic Evaluations of Planing Hulls in Smooth and Rough
Water," Marine Technology, Vol. 13, No. 4, Oct. 1976.
ject boat running at 26 knots in 12-ft seas this m e t h o d 36 Allen, R. and Jones, R., "A Simplified Method of Deter-
gives a vertical wave-induced acceleration at LCG of 0.63 mining Structural Design Limit Pressures on High Performance
g and a corresponding design pressure of 18.95 psi, which Marine Vehicles," AIAA / SNAME.
for 40 000-psi yield strength steel gives a required bottom
shell thickness, in the critical area forward of amidships, Martin Burtness, Member
of 0.19 in. A second pressure prediction method that we
also use is that given by Allen and Jones [ 36 ]. This m e t h o d I wish to congratulate the authors on an excellent paper
requires the specification of a design wave-induced impact combining numerical predictions and structural analysis
load factor (wave-induced vertical acceleration at LCG) with model and full-scale testing. As a m e m b e r of the
often provided by designers or in specifications as an op- testing community, I always appreciate an opportunity to
erational limit. For general application we normally use a compare model test results with full-scale trials.
m i n i m u m of 1.0; however, for patrol boats a value of at Would the authors c o m m e n t on the applicability of the
least 2.0 would seem appropriate. In this case the design various prediction techniques used for estimating accel-
pressure is about 42 psi, probably a bit high for a round- erations? The Hoggard-Jones formula, for instance, has no
chine boat, and resulting thickness for 40 000 psi yield d e p e n d e n c y on deadrise angle, which would seem to be
plating with 12-in. spacing of longitudinals is 0.27 in. if a dominant factor. In addition, most prediction techniques
yield is limited to a point at the middle of the long edge are based on hard-chine planing craft, while the Island
of plate panel. A thickness of 0.22 in. would, with this Class has rounded bilges.
design pressure, indicate some local plastic response of the The full-scale testing analysis performed on this vessel
plate panel at the center of the long edges of the panel. brings to light a problem facing the testing community,
The acceptance criteria of yield stress at the midpoint of that of standardized data processing techniques. The au-
a long plate edge with these design pressures were de- thors note that signal filtering drastically affects the ex-
veloped from a n u m b e r of existing patrol boats of U.S. and treme and N~l~0accelerations and makes comparisons with
E u r o p e a n design. model test results very difficult. A 5-Hz low-pass filter was

244 USCG Island Class Patrol Boat


a t t e m p t e d , but a p p a r e n t l y filtered out the high-frequency pressure resulting in 0.308 in. of p e r m a n e n t set in the
slamming events which are of p r i m a r y concern. The Island Class boat was calculated to be about 159 psi. While
Ocean E n g i n e e r i n g C o m m i t t e e of the 18th ITTC con- a somewhat m o r e refined calculation can b e m a d e (taking
d u c t e d a " D a t a Analysis C o m p a r a t i v e Exercise" which into account the vibration frequency of the panel), the
investigated the different data analysis techniques being r e l a t e d peak d y n a m i c pressure has b e e n e s t i m a t e d as 114
used to h e l p standardize data filtering and processing tech- psi as noted.
niques or p r o v i d e some guidance for the e n g i n e e r in ana- In o r d e r to further increase our a p p r e c i a t i o n for the
lyzing data. strength of such thin panels, perhaps the authors would
W h e n c o m p a r i n g e x t r e m e values b e t w e e n m o d e l and be willing to add information on subsequent d a m a g e ex-
full-scale test results, caution must be exercised. Generally, perience.
70 wave e n c o u n t e r s are considered necessary to p r o v i d e In any case, m y thanks for a v e r y worthwhile p a p e r and
statistically meaningful results in an irregular seaway. This the o p p o r t u n i t y to c o m m e n t on it.
value is a trade-off b e t w e e n statistical accuracy and bud-
geting. Therefore, an e x t r e m e value m e a s u r e d from m o d e l Pin Yu Chang, Member
tests is only 1 1 / 70th highest o c c u r r e n c e and should not
be confused with the e x t r e m e value e x p e r i e n c e d by the We would like to congratulate the authors for their ex-
full scale w h e r e thousands of waves are e n c o u n t e r e d . cellent p a p e r about an i m p o r t a n t subject. F o r all high-
In light of the results and e x p e r i e n c e obtained from this s p e e d boats with.light structure, the structural strength is
p a p e r , I hope furthdr progress can be m a d e toward stan- m o r e critical and must be designed by m o r e accurate
d a r d i z i n g data processing techniques. structural analysis. Using a finite-element m o d e l ( F E M ) is
the best m e t h o d available.
J. Harvey Evans, Member W e agree with the authors that linear structural analysis
is a d e q u a t e by the failure m o d e they use. But the sea loads
P e r h a p s it is all too obvious, but I feel the great value u p o n the hull structure at high s p e e d and in heavy w e a t h e r
of this p a p e r overall is, first, in its thorough coverage of are mostly nonlinear, especially the loads due to slamming.
the m a n y analysis aspects of the complex p r o b l e m to be W e strongly r e c o m m e n d that after t h e linear analysis has
solved. Then, the p r o c e d u r e s followed and their relation- b e e n done, a nonlinear analysis should be carried out. T h e
ship to each other have b e e n carefully explained. Best of nonlinear sea loads can b e d e t e r m i n e d by the nonlinear
all, the j u d g m e n t s m a d e en route to quantitative solutions time domain program, ROSAS4, which we r e c e n t l y h e l p e d
have b e e n clearly stated and the numerical results from to d e v e l o p for the initial version originated by the D a v i d
each of several approaches have b e e n c o m p a r e d so that, Taylor Research C e n t e r (DTRC).
in the future, all m a y benefit and m o r e refined design In addition, a d y n a m i c analysis should also be c o n d u c t e d
estimates m a y be made. At the v e r y least, it has b e e n m a d e to take the d y n a m i c responses into consideration. Such an
m o r e clear w h e r e our k n o w l e d g e is most tentative. All this analysis can be h a n d l e d by NASTRAN with a d y n a m i c
is something of a r e c i p e for a good design p a p e r and this finite-element model. We have c o n d u c t e d such d y n a m i c
one fills the bill so very well. analyses for the Canadian patrol frigate, with results show-
T h e i n d e n t a t i o n and p e r m a n e n t set of bow plating turns ing significant i m p r o v e m e n t s over the static, or the so-
out to b e the m a t t e r of greatest c o n c e r n and perhaps, called quasi-static, approach.
without overstating the case significantly, it should be said
again that first yielding is not failure. In fact, criterion
for the u p p e r limit of permissible deformation still needs
Authors' Closure
to be defined. And, in specific terms, the criterion m a y T h e authors a p p r e c i a t e the interest shown by the dis-
vary with the vessel and the service. Will it be the like- cussers and offer the following response.
lihood of fatigue in the panel? Should the fatigue criterion Dr. Lee has p o i n t e d out the necessity for m o r e pressure
not take into account w h e t h e r or not the panel is in way taps in the m o d e l and we must note that the w o o d e n m o d e l
of a tank space? Or is the deterioration of h y d r o d y n a m i c is s o m e w h a t restricted in that respect. Plastic models have
p e r f o r m a n c e or loss of s p e e d a m o r e critical concern? Or, b e e n constructed with m a n y m o r e taps and n e w pressure-
also as n o t e d in the p a p e r , are subsequent m a i n t e n a n c e sensitive mesh products show promise in this area. In hind-
p r o b l e m s of greatest i m p o r t a n c e and a limiting factor? sight, we p r o b a b l y should have collected m o r e pressure
Finally, to e m p h a s i z e the reserve of panel strength de- data than we did.
v e l o p e d as m e m b r a n e stresses take over from b e n d i n g in E n l a r g e d details of the higher stressed portions of the
this highly nonlinear behavior, a few n u m b e r s from the hull w e r e obtained, and the n u m e r i c a l o u t p u t was of value,
p a p e r can now b e m o r e closely related. First yielding of but the color p a t t e r n s broke down u n d e r the fine mesh
the 23 X 12 X 0.16-in. plating panel has b e e n e s t i m a t e d detail and w e r e very h a r d to read.
from [13] to occur at a static pressure of 14.3 psi w h e n With respect to p r o b l e m s associated with full-scale test-
the m a x i m u m deflection (at the c e n t e r ) has r e a c h e d 42 ing, we can offer p l e n t y of insight. W e have o v e r c o m e
p e r c e n t of the thickness. several difficulties in the strain m e a s u r e m e n t field, such
T h e limit p r e s u m e d to be the m a x i m u m allowable for as hull p r e p a r a t i o n , space humidity, cable capacitance,
U.S. Navy transverse bulkheads (from the last formulation t e m p e r a t u r e compensation, gage b o n d i n g and electro-
of the A p p e n d i x ) will be r e a c h e d at a static pressure of 38 m a g n e t i c interference. W e e x p e r i m e n t e d in o t h e r areas,
psi w h e n the total m a x i m u m deflection will be about half such as t h e r m a l l y i n d u c e d hull stress, w h e r e data are
the thickness and about 20 p e r c e n t of this will r e m a i n sparse. O n e evening, after testing, we w a t c h e d the boat
after the pressure is r e m o v e d . "cool d o w n " b e t w e e n the hours of 6 p.m. and 10 p.m., as
Although Loeser's panel, as cited in the Appendix, was reflected by the exponential g r a p h of stress in the plating.
slightly heavier, it had not r u p t u r e d u n d e r pressure of 1200 In a later test, a r e f e r e n c e gage, t a p e d to the hull, but not
psi and the m a x i m u m deflection had r e a c h e d almost 13 b o n d e d to the steel, showed no such m o v e m e n t . W e oc-
times the thickness. casionally e x p e r i m e n t e d with re-zeroing the gages while
C o r r e s p o n d i n g to these values, the static equivalent underway, holding the boat on a single wave (stern sea)

USCG Island Class Patrol Boat 245


d u r i n g the process. T h r o u g h all of this testing, w e p u t a the n e w p a p e r by P a y n e [37], w h e r e our sentiments about
p r e m i u m on personnel who could m e e t the physical chal- the variation in pressure predictions are reinforced. In
lenge and come back with reliable data. C o n d u c t i n g trials r e g a r d to the 5-Hz low-pass filter, w e feel that it is nec-
in a heavy sea can b e stressful for people, as well as for essary to omit the high-frequency deck vibrations that do
the boat. not directly c o n t r i b u t e to the whole b o d y CG acceleration
Mr. Ghosh has p r o v i d e d some useful c o m m e n t s and during slamming. W i t h r e g a r d to the use of e x t r e m e val-
b a c k g r o u n d on the o p e r a t i o n of the boats. With r e g a r d to ues, our test data w e r e g a t h e r e d for each cell in the matrix
the structural reliability analysis, w e a p p l i e d m e a s u r e d (Table 4) for about 100 wave e n c o u n t e r s (5 minutes). If
stresses and pressures ( d e r i v e d from those stresses) in con- w e had the ability to c o m p a r e one-tenth highest accel-
junction with Table 4 to project structural reliability. At eration values before publication, we would have done so,
no time w e r e we c o n c e r n e d with a design approach; only but only r e c e n t l y did we acquire the p r o p e r software. W e
with the application of observed data. The results are very h o p e that we can soon publish the comparison of the m o d e l
sensitive to the o p e r a t i n g profile (Table 4), especially to data with the full scale.
the assumption of 1 p e r c e n t o p e r a t i n g time at 24 knots in Professor Evans has p o i n t e d out the n e e d to establish
a 10-ft significant wave height (head sea). Stress levels are an u p p e r lirrlit for permissible d e f o r m a t i o n and a r e q u e s t
highest in this cell of the matrix and we feel it is a con- to add information on d a m a g e experience. O n e of the
servative assumption that the average boat will spend 30 patrol boats was g r o u n d e d at high s p e e d during a chase
hr p e r year in those sea conditions. This condition is close and the b o t t o m plating in the forward m a c h i n e r y space
to the " c r e w l i m i t " that Mr. Ghosh suggests. H e a d seas was i n d e n t e d a p p r o x i m a t e l y one inch just f o r w a r d of the
w e r e chosen for the analysis because the r e c o r d e d stresses transverse frames, without rupture. About 20 ft ~ of 7-1b
at all other angles (bow, b e a m and following) w e r e v e r y plating (0.16 in. nom.) was c r o p p e d out and n e w plating
low i a b o u t 10 to 20 p e r c e n t of the h e a d seas values). The w e l d e d to the existing framing, which had not suffered
midship section calculation was c h e c k e d and found to be any visible damage. Conversely, after our testing, which
a d e q u a t e (344 in.2-ft versus 252 r e q u i r e d by ABS). The p r o d u c e d indentations of 0.308 in., no repairs w e r e made.
assumptions c o n c e r n i n g this calculation are n o t e d in ref- S o m e w h e r e in b e t w e e n , t h e r e is a n e e d for a definition of
e r e n c e [3] of the paper. No deckhouse e l e m e n t s are in- "a_cceptable d a m a g e . "
c l u d e d in this calculation. Dr. Chang notes that d y n a m i c finite-element analysis
Mr. Curry has p o i n t e d out some correlation with other would be desirable in this case. W e agree, because any
m e t h o d s of calculation of b o t t o m pressure. Again we have m e t h o d that can m o r e c o m p l e t e l y simulate reality is going
the p r o b l e m of d e t e r m i n i n g which " a c c e l e r a t i o n " is re- to help the designer of such a complex structure. O u r
q u i r e d - m a x i m u m , average or o n e - t e n t h highest. A n o t h e r e x p e r i e n c e with the static finite-element analysis has
p a p e r , by P a y n e [ 37 ], has r e c e n t l y b e e n published on this shown that the selection of p r o p e r stiffness in the b u l k h e a d
subject, noting the variation in pressures p r e d i c t e d by cur- e l e m e n t s is a very difficult task. With m o r e t i m e and
r e n t formulae. W e i n t e n d to analyze our data further, now money, we m a y have b e e n able to refine those calculations
that w e have the same algorithms that w e r e used to cal- and a g r e e m o r e closely with the r e c o r d e d stresses.
culate o n e - t e n t h highest accelerations in the towing tank. In conclusion, we wish to point out that several portions
W e had, indeed, m i s i n t e r p r e t e d the value of S in the for- of the data collected during the full-scale testing are still
m u l a as Mr. Curry has p o i n t e d out. After further consul- u n d e r g o i n g analysis. W e have a n o t h e r d e f o r m a t i o n event,
tation, we have a p p l i e d the formu!a, t = 0.0028 (S) x/-V, a comparison of the one-tenth highest accelerations in the
for planing hulls, w h e r e S is the smaller of the p a n e l di- towing tank with the full-scale data, a look at the variation
mensions. Correcting for the difference in steels (30 ksi of the angle of principal stress in the plating and a com-
versus 39 ksi) and using S = 12 in., V = 26 knots, we get parison of the stress at the e d g e of the panel versus the
a r e q u i r e d t = 0.15 in. This, we feel, is still too thin, based center. W e h o p e that some of these subjects will be the
on testing e x p e r i e n c e and the lack of any corrosion allow- topics of future papers.
ance. W e feel that 9-1b plate (0.22 in. nom.) is a p p r o p r i a t e
for the region subjected to slamming pressure.
Mr. Burtness points to the p r o b l e m of h a r d - c h i n e d ver- Additional references
sus round-bilge hulls and the applicability of various for- 37 Payne, P. R., "A Discussion of the Design Pressures Ap-
mulas. W e a g r e e that deadrise angle must b e a c c o u n t e d propriate to the Bottom of a Planing Boat," Ocean Engineering,
for in a r e - e n t r y slamming calculation and refer him to Vol. 15, No. 5, 1988.

246 USCG Island Class Patrol Boat

You might also like