ENERGY & ENVIRONMENTAL EXPO 95

Offshore & Arctic Operations Symposium

American Society of Mechanical Engineers Petroleum Division
Houston, Texas, January 30, 1995

DEVELOPMENT & VERIFICATION OF A SIMPLIFIED METHOD

TO EVALUATE STORM LOADINGS ON AND CAPACITIES OF
STEEL, TEMPLATE-TYPE PLATFORMS

Robert G. Bea
Department of Civil Engineering
University of California
Berkeley, California

ABSTRACT INTRODUCTION
During the past three decades, an immense amount of Requalification of existing offshore platforms involves
effort has been devoted to development of sophisticated developing an understanding of the Ultimats Limit Stats
computer programs to enable the assessment of storm (ULS) lateral load capacity of the structure. Nonlinear
wind, wave, and current loadings and the ultimate limit analyses of offshore platforms are difficult and costly to
state capacity characteristics of conventional, pile-sup­ perform. Given a large number of structures to be re­
ported, template-type offshore platforms. These pro­ qualitied, it is desirable to have a simplified method to
grams require high degrees of expertise to operate prop­ estimate the ultimate limit state capacity of the platform.
erly, are expensive to purchase and maintain, and re­ A simplified procedure has been developed to estimate
quire large amounts of manpower and time to complete the ULS!ateral load capacity of the three primary com­
the analyses. Due to the sophistication of these programs ponents that comprise conventional [Link] template-type
and the expertise required to operate them, experience offshore platforms: the deck legs, the jacket, and the pile
has shown that it is easy to make mistakes that are diffi. foundation. In addition, a simplified procedure has also
cult to detect and that can have significant influences on been developed to estimate the wind, wave, and cutTent
the results. lateral loadings.
This paper summarizes the development of simplified With these results, the Reserve Strength Ratio (RSR)
procedures to evaluate storm loadings imposed on and can be dBtermined as:
induced in template-type platforms and to evaluate the

~~
ultimate limit state lateral loading capacities of such plat­
RSR= (1)
forms. Reasonable simplifications and high degrees of
user friendliness have been employed in development of
the software to reduce the engineering effort, expertise, Ru is the ultimate lateral load capacity. SR is the refer­
and costa associated with the analyses.
ence storm total maximum lateral loading. The refer­
Verification of these procedures has been accomplished
ence lateral loading is that specified in current platform
by comparing the results from the simplified analyses
design guidelines such as API RP 2A (AP!, 1993).
with the results from three-dimensional, linear and non­
The remainder of this paper will detail development,
linear analyses of a variety of template-type platforms.
verification, and applications of the simplified proce­
Good agreement between results from the two types of
dures to estimate the storm loadings, the ULS capacity of
analyses has been developed for the evaluations of both
template-type offshore platforms, and "fragility" curves
loadings and capacities.
that characterize the likelihoods of platform failure for a
given storm intensity.
APPROACH legs, each of the hays in the vertical truss system that
Fig.1 summarizes the analysis process. In these analy­ comprise the jacket, and the piles (axial and lateral) are
sss, an attempt is made to use "unbiased" estimates of the determined based on the ultimate limit state capacities of
parameters that determine both loadings and capacities. the elements that comprise these components. Brace ca­
A Personal Computer program identified as ULSLEA pacities depend on the direction of loading (tension,
(Ultimate limit S - limit Equilibrium Analyses) has compression), and the capacities of the jointe at the ends
been developed to perform the analyses. of the braces. Provisions are made for local loadings
from waves and currents, deck loading P-b. effecte, mo­
ENVIRONMENTAL ment induced shears at the top and bottom of the jacket,
and the shear resistance developed by the battered jacket
USER INPUT STRUCTURAL legs and piles.
Comparison of the storm shear profile with the platform
FOUNDATION sheer capacity profile identifies the "weak link" in the
platform system. The base shear or total lateral loading
WIND at which the capacity of this weak link is exceeded
defines the static lateral capacity of the platform (Rus).
WAVE

CURRENT
Storm
Shear Profiles
DECK

JACKET

PILES

STATIC

DYNAMIC

LOAD UNCERTAINTIES FIGURE 2 - ULTIMATE LIMIT STATE LIMIT EQUILIBRIUM

ANALYSIS TO DEFINE PLATFORM STATIC CAPACITY
Pf f H
CAPACITY UNCERTAINTIES

The ststic lateral loading capacity is corrected with a

FIGURE 1 - PLATFORM LOADING, CAPACITY AND '1oading effecte" modifier (Fv) [Bea, 1991; Bea, Young,
PROBABILITY OF FAILURE ANALYSES 1993] to recognize the interactive effecte of transient
wave loadings and nonlinear hysteretic platform re­
sponse; thus
Fig. 2 illustrates the approach used to compute the static
capacity of the platform system. A series of storm load­ Ru =Rus(Fv) (2)
ing profiles of horizontal shear are developed. The plat­
form elemente and appurtenances are modeled as a se­ Once the best estimate capacity has been determined,
ries of equivalent vertical cylinders located at the wave the platform probability of failure (pt) is determined
crest position. Block volumes are used to model the deck conditional on the occurrence of a given storm intensity.
elemente. Simplified methods are used to estimate the Storm intensity is based on 1he expected maximum
wave kinematics. Standard API methods 919930 are wave height (H) with the other storm parameters (wind
used to estimate the wind, wave, and current loadings speed, current velocities) conditional on the time and di­
hased on user specified wind epeed, wave height and rection of occurrence of H~ The storm intensity is ex­
period, and current velocities. Loadings from inundation pressed with the total maximum lateral force developed
of the lower decks by the wave crest also may be deter­ on the platform by the storm (S I H).
mined.
Uncertainties in the loadings (<JS) and capacities (<>Ru)
A profile of horizontal shear capacity of the platform is
developed. The horizontal shear capacities of the deck and the correlation (p) between Ru and S are specified
and the conditional probabilities Gf failure (Pfl H) de­ cumulations on all underwater elements (10 cm). The
termined (Bea, 1990). To determine the probability of design storm maximum total lateral force was Sn =4.5
failure for all storm intensities (Pf), the conditional prob­ MN.
abilities Gf failure are multiplied by the probabilities of To perform the ULS capacity analyses, the storm condi­
experiencing a given storm intensity (PI H) and summed tions were specified as combinations of wind speeds and
over all storm intensities (Bea, et al, 1994). currents that were conditional on the time and direction
of occurrence of a given expected maximum wave
height (H). H was ranged from H = 10 to 25 m. The
EXAMPLE variation Gf the maximum lateral force components with
The analyses outlined in Fig. 1 will be illustrated with H are summarized in Fig. 4. The wave loadings acting
application to a four-leg platform (Fig. 3) that was in­ on the lower decks of the platform exceed the wind
stalled in a water depth of 73 m in 1984. The platform loadings for H ;;, 15 m and become equal to the wave
supports a 34 MN deck weight and has two boat land ­ loadings acting on the rest of the structure at H =22 m.
ings. The cellar deck is located at +10.6 m;. The jacket
legs are 1.2 m in diameter and the 1.0 m diameter piles
are grouted inside the legs. The jacket braces range from "
41 cm to 61 cm in diameter. The sea floor is covered with
al m thick layer of soft clay. Below this layer, the foun­
....
<
a:
w .. -Total For<*

dation soils can be classified as stiff, over-consolidated

cohesive sediments that have an average shear strength '.3i - -W•V., Jto0bl

of 100 k:Pa to a depth of 10 m. lE '

::>en "
lEW
;;co 10
<"'
:::El!
....
~
~
•
10
" ..
EXPECTED MAXIMUM WAVE HEIGHT • H • m
"

FIGURE 4 - STORM LATERAL FORCE COMPONENTS

These results were verified with results from detailed

three dimensional wind - wave - current loading analy­
ses for comparable wind, wave height, and current con­
ditions. In general, the simplified procedure under-pre­
dicted the total maximum lateral loadings by about 20 %.
The principal differencas were tracad to the directional
spreading, shielding, and current blockage kinematics
factors that were introduced into the simplified analyses.
These factors had not been included in the original de­
sign loading analyses. Once these same kinematics cor­
rection factors were introduced into the detailed analy­
ses, the simplified procedure tended to slightly overesti­
FIGURE 3 - EXAMPLE PLATFORM mate the forces. This difference wae traced to the lack of
horizontal spatial distribution in the platform elements in
the simplified analyses. All of the platform elements are
The platform was designed according to 1983 API RP concentrated at a single vertical position in the wave
2A guidelines with a 100-year design wave height Gf HD crest in the simplified analyses.
= 12 m. The design criteria included storm associated The vertical profile of static shear capacity of the exam­
currents (1.2 m /sat surface). The hydrodynamic forces ple platform from the simplified ULS analyses is sum­
were computed with the Morison formulation based on a marized in Fig. 5. The vertical profiles of shears for three
drag coefficient of Cd= 0.75 and an inertia coefficient of storm conditions acting on the platform are also shown.
Cm;: 2.0. Provisions were made for marine growth ac­ The lateral shear capacity in the jacket was generally de­
termined by the tensile - compressive brace capacity in based on a 100-year return period wave height condition
each bay. Due to the grouted joints and heavy-wall joint (H = 12 m). The lateral loading pattern was proportion­
cans, the joints did not control the diagonal brace lateral ally increased to push the stnreture to ULS. The results
capacities. are summarized in Fig. 6. The results indicate that the
platform has a total lateral capacity of Rus = 12.5 MN.
1•,.....~,~u,._•~T,_~,_..,,,.....~T,_~....,J .......~....-~~
.fl \ :
"
,: ::::::·:··E~~·J:::: :1:::::::r:. ::J . u·­ '
PLATFORM SHEAR CAPACITY
z " /
~

I
................... .,. :E 10
,/;/
--·H • 18 m Sh•V a:
45 ~- ........ -- ··H • 20,I m She•
<
w
:x:
• v
I " . '
"' •
:: ~ IJ~j-~}1:···::;·-[Link]..====r ,, /
w
. . . . ,........-1 "'<
Ill
4

.-:'1.:----1 I......1 • .. / .
••L.....G..._.~....._~ c....~.L..~...i...~....i.~....._
v
.. . ..
0 5 10 115 20 21 30 35

SHEAR CAPACITY OR STORM SHEAR • MN

'
• " "
UPPER DECK HORIZONTAL DISPLACEMENT· cm

FIGURE 5 - PLATFORM STORM LOADING SHEAR

PROFILE COMPARED WITH STRUCTURE SHEAR FIGURE 6- NONLINEAR STATIC PUSH OVER ANALYSIS
CAPACITIES RESULTS

Based on the simplified loading and capacity analyses, The dats points shown in Fig. 6 identify the nonlinear
the static shear capacity of the platform is determined by events that developed in the platform. All of these non­
the deck legs. The best estimate deck leg shear capacity linear events were confined to the piles.
occurred when H ~ 17 m. For this condition, the total lat­ These results gave concern for the verification of the
eral static loading and capacity is Rus =9 MN. simplified capacity analysis summarized in Fig. 4. The
However, if the deck loadings on the example platform difference between the detaHed nonlinear lateral capac­
are somewhat lower than estimated by these analyses, ity (Rus = 12.5 MN) and that estimated by the simplified
the static lateral capacity of the platform would be con­ method (Rus = 10 MN) was traced to the lateral loading
trolled by the foundation piles. In this case, the lateral pattern that had been used to perform the nonlinear
shear capacity of the platform would occur when H ~ analyses.
20.5 m. For this condition, the total lateral stetic loading In the case of the detailed nonlinear analysis, the loading
and capacity is Rus =13 MN. pattern had not been changed as the wave loadings in·
There are large differences in the reserve strength of the creased. In the case of the simplified analysis, the wave
components that comprise this platform. The jacket has loading pattern was changed as a function of the wave
a much larger reserve strength than either the deck legs heights. This resulted in more force at the top of the
or the foundation piles. The insights provided by these jacket due to wave crest loadings on the lower cellar
simplified loading and capacity analyses could be used to deck. The fmlure mode was shifted from the pi1es to the
develop a more balanced design in which there would be deck legs. The simplified method identified an error in
comparable levels of reserve strength in the deck legs, the nonlinear analysis. Once the loading pattern was ad­
jacket, and foundation piles. The simplified analyses also justed as a function of the wave heights in the nonlinear
provide an expedient way to examine the potential analysis, the static push-over results indicated a lateral
effects of damage and defecte on the capacity of the capacity of Rus = 9 MN.
platform. High defect - damage probability and high The difference between the nonlinear stetic push-over
lateral capacity consequence members can be defined Rus = 9 MN and the simplified ULS capacity of Rus = 10
and alternative load paths provided to minimize MN was traced to neglect of the vertical loading - lateral
excessive loss of capacity. displacement (P. A) moments in the deck legs in the
Detailed results from nonlinear static push over analy­ simplified analyses. An approximate analysis of the rela­
ses were developed for this platform. To perform these tive lateral displacement between the bottom of the deck
analyses, a nodal lateral loading pattern was developed and the top of the jacket based on four fixed-fixed end
columns free to displace at their top ends and loaded year return period storm conditions. The design hydro­
with the estimated ULS wind and wave deck loads pro­ dynamic drag and inertia coefficients were in the range
duced a moment that was used to reduce the ULS capac­ of Cd= 0.5 to 0.7 and Cm= 1.5 to 2.0. Design criteria for
ity of the deck legs. This correction brought the results only one of the platforms (C) included storm associated
into good agreement. currents. Joint designs included gusseted, heavy wall
Ambient vibration measurements performed onboard joint cans, and groutad leg - pile annuli.
this platform indicated a natural period of Tn = 1.5 see. Several of these platforms experienced severe loadings
The ULS wave H = 17 m had a period of Tw = 12 sec. from tropical cyclones. Two of the platforms (A, B) less
=
Thus, Tw I Tn 8. The platform was capable of develop­ than one year old were located close to the path of hurri­
ing a system ductility of µ = 3. The results developed by cane Hilda (1964). One of the platforms failad (A) and
Bea and Young (1993) indicate a wave transient loading the other axperienced significant damage (B). In a pe­
~nonlinear response correction factor of Fv = 1.2 . Thus,
riod of 15 years, platform F experienced four storms that
the best estimate capacity was Ru 10.8 MN. = generated wave heights that were approximately equal
Fragility curves were developed that expressed the to or greater than its design wave height. One of the
probability of platform failure conditional on a specified platforms (C) was in the immediate path of hurricane
wave height (Pl ff). The uncertainties in the platform Camille (1969) and experienced green water in the lower
decks without significant damage. There were seven
capacity and the storm loadings were evaluatad as out­ other almost identical platforms in the same vicinity that
lined in Bea (1990). Both inherent or "natural" random­
also survived without significant structural damage.
ness (Type I) and modeling - parameter (Type II) uncer­ Table 2 summarizes the results of the verification analy­
tainties were included in the analyses (Bea, 1993). The ses. The results are keyed to the figure numbers in this
results of the analyses are summarizad in Fig. 7. paper. The remainder of this section will discuss the
Inclusion of Type II uncertainties has important effects analyses.
on theresults.

! i l __,,......._ Pla!lorm A
This structure (fig. 8) was designed for a 25-year return
.. i ~---··"!'.........
,............ ;......... +······ +·········~ period wave height of HD = 14 m without any air gap.
;/i
/ ' Due to an error in determining the water depth, the plat­
,<]7/+··········,·;_····_··~··~··
..·;,· ,· u·.·~· ·,·;"· f;, · ·,· !:· · ·,.·,.· ;."· · ·,· · · · · ~ form was placed in a water depth 0.6 m greater than
originally intended.
The simplified ULS capacity analysis Rus = 7.1 MN
) (broadside). The critical mode of failure was in the deck
legs at the top of the jacket. The top row of diagonal
braces were also close to failure for this condition. The
I API reference lateral load of SR = 9.4 MN includes a
wave crest loading on the lower cellar deck of the plat­
form of 2.2 MN.
10
" 14
•• " •• " " The results of the nonlinear push-over analyses of this
EXPECTED MAXIMUM WAVE HEIGHT • H • meters platform are summarizad in Fig. 9. The ULS failure
mode involved failure of the deck legs and top bay of di­
agonal braces. This failure mode was obtained only
FIGURE 7 - FRAGILITY CURVE FOR EXAMPLE when the dynrunic stiffness and capacity characteristics
PLATFORM of the foundation were recognized and biases removed
from the evaluations of the pile lateral and axial capaci­
ties (Bea, 1992a). The use of conventional static capacity
GULF OF MEXICO PLATFORM VERIFICATIONS analysis methods under-predicted both foundation stiff­
Analytical and performance experience with six Gulf of ness and capacity and indicated that the failure mode
Mexico (GOM) platforms has been usad to verify the was initiated and constrainad to the foundation piles.
simplifiad analyses. The characteristics of these plat­
forms are summarized in Table L The verification cases
include four self-contained drilling and production plat­
forms and two tender assisted drilling platforms de­
signad and installed during the period 1959 through
1970. All were built from A-36 steel. All were founded on
good soils. Design criteria ranged from 25-year to 100­
 ..., fi2.2
=
(Fig. 9) then Fv 1.2 The best estimate dynamic capac­
' ~.' A
. / •14Am
ity would be Ru = 8.5 MN. The computed pask storm
loading exceeded the computed platform capacity.
•••• •
• The platform should have failed. It did, and in the way
;~:'':%~:=~::::~~!:t~\:~~~~t:~::~[Link]<~~::;-:;,,,:~:0f~:::i~nR~::~~~,~~:~::::~::::~t:x'~~:~:::::~;::::u·: predicted by the simplified loading and capacity analy­
ses.

FIGURE 8 ·ELEVATIONS OF PLATFORM A

z • i
;
::;
'
w
0 • , ""
a:
0
/ FIGURE 10 ·PLATFORM A AFTER HURRICANE HILDA
u.
_, ' j

Ci! ' I Platform B

w
IC_, • I
This is a nearby similar platform (Fig. 11) installed in a
_, ' ·7 water depth of 66 m in 1963. The platform was designed
< for a 100-year return period storm with a maximum
::;... I wave height of HD = 17 m with an air gap of 1.5 m. The
0
0 1O 20 30 40 so 10 70 ao design hydrodynamic forces were determined based on
Stokes Fifth Order Theory kinematics, no current, and a
DECK DISPLACEMENT • cm
drag coefficient Cd =0.5. The total lateral design loading
was SD = 8 MN [Marshall, Bea, 1976].
FIGURE 9- PLATFORM A STATIC PUSH OVER ANALYSIS Application of the simplified analyses indicated Rus = 20
RESULTS MN for the broadside loading condition. The critical
mode of failure involved compressive buckling in the top
This platform failed one year after it was installed (Fig. three levels of vertical and horizontal diagonals. The
10) [Hilda Meeting Transcript, 1964]. The failure was end-on loading condition indicated a capacity that was
apparently triggered in the deck legs and top bay of di­ 25 % larger than for the broadside condition (Rus =25
agonal braces. Based on the hindcast hurricane condi­ MN).
tions 1h at were present at the platform [Bea, 1974], the The results of the detailed nonlinear analyses of this
maximum total lateral loadings were estimated to be SM platform indicated Rus =19.1 MN (Fig. 12). As for the
platform A analyses, these lateral capacity analyses con­
=9.0 MN (maximum wave height Hm =18 m). The sidered the pile foundation dynamic loading character­
maximum wave resulted in significant inundation of the
istics and removal of biases from the axial and lateral ca­
lower deck. pacity evaluations. The failure mode involved the top
Recognizing that the platform had a natural period Tn = two bays of vertical diagonal braces.
1.8 sec and that the period of the maximum waves were
in the range of Tw = 10 to 11 sec, Tw I Tn =5.6 to 6.1.
Given that the platform could develop a ductility ofµ= 3
 TABLE 1 -VERIFICATION PLATFORM CHARACTERISTICS & STORM EXPERIENCE

NO. Lower
Plat Year Water Legs I Deck El Joints Soils Storm Experience
Depth Piles +mMGL
ID m
A 1964 52.4 818 9.2 gusset 1Om soft clay over Hilda (1964), Failed
stiff clav
B 1963 66.2 818 12.8 hvy wall, as A Hilda, Damaged
gusset
c 1968 98.2 8/16 13.1 heavy sands and stiff clay CamiUe (1969), No Damage
wall
0 1970 82.6 818 14.0 orout asC Frederic (1979), No Damaoe
E 1959 t 5.9 818 11.9 grout 1 m soft clay over Carta (1961 ), Beulah (1967),
stiff clav Alicia (19851, Damaoed
F 1963 42.7 515 10.4 grout 3 m soft clay over Carta (1961), Hilda (1964), Celia
stiff clay (1970), Carmen (1974), Betsy
(19791. Allen (19801, Oamaaed

TABLE 2 ·SUMMARY OF LOADING· CAPACITY VERIFICATIONS

Plat. So SR SM Aus Fn Ru RSA Rus nl BRus

I. 0. MN MN MN MN MN

A 5.6 9.4 9.0 7.1 1.2 8.5 0.90 7.4 1.04

Ilia 8\ ma 101 Ilia 91
21.4
B 8.0 12.0 8.6 20.0 1.2 24.0 2.0 19.0 1.07
Ilia 11l (fia 121 0.96

c 15.0 21.0 25.0 26.0 1.25 32.5 1.55 27.. 0 1.03

Ilia 131 Ilia 14\

0 11.3 10.3 9.0 14.0 1.2 16.8 1.63 12.5 0.89

(fia 15) (fia 16) 171
(liq

E 3.1 4.0 4.5 4.5 1.0 5.0 1.0 3.6 0.80

(fia 181 (fia 191 (fia 20\
5.0
F 3.6 9.0 8.1 (dam) 1.0 3.2 0.36 4.5 0.9
(fig 21) 6.0 (dam)
(rep) 1.0 5.9 0.66 5.4 (rep) 0.9
Ilia 22\ (fig 23)

So ·design lateral loading, SR - API Reference lateral loading, SM - maximum loading experienced, Rus - Simplified analysis
lateral capacity, Fn - transient loading - nonlinear response correction factor, Ru - best estimate lateral capacity, RSR ­
Reserve Strength Ratio, Rus nl - lateral capacity based on nonlinear analyses, BRus • Bias (Aus nl I Aus) in simplffied analy­
sis
 ~ ...--..··---t ..,
..""
'1--~1-t'.i.1..'-1---1~'.JI+ 111
.. H predicted Rus. The major pert of this difference is due to
the various assumptions that are made in the two analy.
ses regarding the nonlinear characteristics of the braces
and joints.
This platform survived the same storm that resulted in
the failure of Platform A The storm wave crests did not
reach the lower decks. The maximum lateral loadings
were estimated to be SM= 8.6 MN (Bea, 1974; Marshall,
Bea, 1976). The vertical diagonal braces in the platform
were extensively damaged; two of the jacket legs were
parted (Hilda Meeting Transcript, 1964). The platform
should have survived and it did. The platform damage
was repaired and this structure is in service today.

"·'
....
+tMm

'""
··;::·:·:::::::·:;>(;:O:;:;:;:::O:.<:::;:;:;:;:.:;:;:::::O:::::;:::;:;:;;::::;:::;:,:::'.~·:-:c-:-:-:--c.··

a•--..i

FIGURE 11 • ELEVATIONS OF [Link] B

z " '' ' ; _-.

::i;
vande0r..f&~ /
i
! ~.........
/ -
~Hl.5
0:
~~t:rc;;-·71·:-'-------;---------<-------~
It
.... 11

~ VA·
..J
i /!/
s.s '·······+····-·;·-······'·····-;·-········c····-~·-····1

~ ~ 0
0 20 . ..
DECK LATERAL DISPLACEMENT ·cm
80 FIGURE 13 ·ELEVATIONS OF PLATFORM C

FIGURE 12-STATICPUSHOVER RESULTS FOR PlatfonnC

PLATFORMS This platform (Fig. 13) is a more recent 8-leg, 12-pile
structure that was designed for a 100-year storm wave
height of HD = 18 m with an air gap of 1.5 m, the storm
These lateral capacity estimates were confirmed with associated currents (1.2 m I s at surface), and Cd = 0.5.
results from nonlinear push-over analyses published by The platform was installed in 1968 and was designed ac­
van de Graff and Tromans (1991) (Fig. 12). Their results cording to the draft guidelines of the first API RP 2A
indicated Rus ~ 21.4 MN for the analyses that suppressed Application of the simplified ULS capacity analysis
lateral pile failure. The failure mode for this condition gave Rus = 26 MN for broadside loading conditions. The
was concentrated in the vertical diagonal braces. The primary mode of failure involved compressive buckling
analyses that utilized conventional pile capacity charac­ in the third and fourth levels of vertical diagonal bracing.
terizations indicated Rus ~ 16.5 MN. The failure mode The broadside capacity was approximately 70 % of the
for this condition was concentrated in the piles. end-on loading capacity (Rus = 37 MN).
Note the differences produced by the two nonlinear These lateral capacity estimates are in reasonable
analyses (Fig. 12); there is ebout a 12 % difference in the agreement with those originally estimated for this plat.
form (Ma:rshall, Bea, 1976]) This earlier work indicated without substantial damage. The simplified loading and
Rus = 30.2 MN (ductile redundant with wave in deck) to ULS capacity analyses indicate that they should have.
Rus =36.7 MN (ductile redundant without wave in deck).
The Reserve Strength Ratios were estimated to be in the
range of RSRe = 2.1 to 2.5. The best estimate ca paeity
was taken as Rus = 30.2 MN. These resulte are
remarkable when it is remembered that they were devel.
oped using linear structure analyses.
Analyses published by Tromans and van de Graff
(1992])confirm thase static capacity resulte (Fig. 14).
Their work indicates Rus =38.6 MN (for end-on loading
condition) to Rus = 27.0 MN (for broadside loading con­
dition).

z
:E
40 . './\:
•
"
Cl
z lO
/ "'-­ ..;.;,:
15
< /
g " !/
...J 20 '
<
a: /
....w
j
15

10 ~
v
...J 7
g.... • /
0
FIGURE 15 ·ELEVATIONS OF PLATFORM D
0 10 20 >O 40 •o
DECK DISPLACEMENT • cm Platform D
This structure (Fig. 15) is the most recent of the 8-leg
GOM platforms studied. Like platform C, the platform
FIGURE 14 ·STATIC PUSH OVER RESULTS FOR was designed according to the first edition API RP 2A
PLATFORMC gnidelines for 100-year storm conditions that included a
design wave height Hp = 17.7 m, currents (1.1 m Is at
Ambient vibration measuremente performed onboard surface), an allowance for marine growth, and Cd = 0.6
this platform indicated that the natural period of this to 0.7 (function of member diameter).
platform is approximately Tn = 1.5 sec (Ruhl, 1976]) The The resulte of the simplified loading and capacity anal·
maximum wave heights in hurricane Camille had peri. yses are summarized in Fig. 16 (Bea, 1992b, 1992c). The
ods in the range of Tw = 11 to 12 sec. The ratio Tw / Tn = 100-year storm lateral loading shears as a function of el·
4.4 to 4.8. Recognition of transient loading • nonlinear evation are compared with the shear capacities of each
capacity performance effects indicates a loading effect of the bays in the jacket and the deck legs. In this case,
because of the vertical diagonal brace framing patterns,
factor of Fv = 1.25 (Bea, Young, 1993). This evaluation
the end-on loading capacity is less than the broadside
agrees well with results recently published by Stewart
l~ading capacity. Both capacities are governed by the
(1992])for a comparable platform, transient wave load­
diagonal brace and leg shear capacities in the fourth level
ing conditions, and nonlinear - dynamic response.
of bracing below the jacket top.
This platform, and seven other similar structures sur.
Results from nonlinear push-over analyses of this plat­
vived the intense portion of hurricane Camille. One of
'.on:' are summarized in Fig.17. The push-over analyses
the platforms recorded a 22 m wave height before the
mdicate that the vertical diagonals and several of the
wave stafffailed. Ssveral of the platforms indicated sub­
horizontal members in the top four bays of the jacket are
stantial wave crest damage in the lower decks (Bea,
involved in the failure mode. In this case, the simplified
1974). The maximum wave heights in the storm were
method over-estimated the capacities.
estimated to be approximately HM = 24 m. The maxi­
The tendency to over...estimate the capacities was traced
mum total lateral loading estimated on this group of plat­ to neglect of the local wave pressure induced moments in
forms ranged from SM =20 to 25 MN (Marshall, Bea, the upper levels of bracing. Corrections were introduced
1976; Stewart, et al., 1988]) All of the platforms survived
t.o the brace compressive capacities t.o rooognize the local
wave pressures (reduced capacities by 10 t.o 20 %). Once
this [Link] was introduced into the caleulation of the ver­ 27.4 m
tical diagonal brace compressive cspacities, the results
agreed very closely.

FIGURE 18- REVATIONS OF PLATFORM E

PlatformE
This platform (Fig. 18) was installed in 1959 in 15.9 m of
water. The 8-leg tender drilling assisted platform had
83.8 cm diameter, 1.3 cm wall thickness legs inside of
which were grouted 76 cm diameter piles. ft was single
SHEAR CAPACITY OR STORM SHEAR - MN diagonal braced with 32.4 cm diameter members that
were battered in the same direction. The jacket was
FIGURE 16-PLATFORMD100-YEAR STORM CONDITION placed in a water depth that was 4.6 m deeper than in­
SHEAR PROFILE AND STRUCTURE SHEAR CAPACITY tended. For that reason, the [Link] of the jacket was at ele­
vation -1.5 m. The platform was designed for 25-year
wave height criteria with HD =11.6 m. The present API
reference level wave height is HR = 13.4 m.
z
::E " [;~~
. __,,, r---­
:..­
The simplified loading analysis (Fig 19) indicsted a [Link]
lateral loading associated with the 100-year API wave
w
0
a:
"
10 ,
/
,,.,..
v height and forces condition t.o be SR = 4.5 MN. ft was this
condition that brought the platform to ULS for broadside
0
u. //' loading. The failure mode was concentrated in the deck
•
....
<
a:
w •
/.,
. 'l
/'
,___,, ,. ., I··--····· legs for both the end-on and broadside loading condi­
tions. There is a marked difference in the horizontal
~
1-..-eroaihide
/, loading cspacities of the structural components that
.... • ~ comprise the jacket. The deck legs are the weak link in
~ this platform structure system. Diagonal bracing of the
~ / deck legs could be very effective in rehabilitating this
• structure.
Detru1ed nonlinear analysis results have been developed
• S 10 15 20 2S 30 35 lO
for broadside and end-on loading conditions (Fig. 20)
DECK DISPLACEMENT • cm
(Bea, 1992c). These results indicate Rus = 3.6 MN for the
broadside loading and Rus = 5.0 MN for the end-on load­
FIGURE 17 PLATFORM D STATIC PUSH OVER ing. The nonlinear analysis did not indicate that the deck
ANALYSIS RESULTS legs had comparable shear capacities for end-on and
broadside loadings.
'This platform was located on the east side of the
Mississippi River delta. In 1988, this structure experi­
enced maximum wave hejghts during hurricane
Frederic of HM = 15 m. The maximum [Link] lateral
storm force was estimated t.o be S ~ 9 MN. The platform
survived without damage. The simplified analyses indi­
cate that it should have.
 ,, .-~~~-.-~~~~~~~~~~~~~
~+1Um
10 ··---~·-·· •••...•l....-.................
1.•..........•...•..••.•.j.••••..•..••.•..•.•.••

J:::::=t. ,...
• l : :

5 ·····-\· ~--·····l···-·········-···-·-t·-······-···········
'''''"''''''"'''''''''''''''@••••·'lffi~
.... •. ~
; ~· : : .·. •::~::!::::~:::::::::~~f=~::':~'.:-~·:~:::::::
:: 100-yr Sto~ !M•r.-

:::::~·::1:.:.:~~:r···················
i j
.....
... ~~_.

0
..........s ~~~ .....
10
~~~~ ..........
15
~~~

20
.01.3
Pion

SHEAR CAPACITY OR STORM SHEAR • MN

FIGURE 19 - PLATFORM E 100-YEAR STORM SHEARS

AND BROADSIDE AND END-ON SHEAR CAPACITIES FIGURE 21 • PLATFORM F ELEVATION AND PLAN

The platform was extensively damaged during four

z • hurricanes that produced wave heights at the location
::E that equaled or exceeded the 25-year design wave height
• of HD = 14 m (Bea, et al., 1988). The damage included
4 missing diagonal braces and cracked joints. There was
extensive damage to one leg of the platform that was de­
' veloped when acid from the well workovers was repeat­
2 edly allowed to leak onto the leg. One of the broken di­
agonal braces was attributed to a compressor that had
been accidentally dropped overboard during its installa­
tion.
0
The present API 100-year wave height is HR = 20 m.
0 20 40 60 80 100
Fig. 22 shows the 100-year storm loading shears and the
DECK DISPLACEMENT • cm platform shear capacities. The platform shear capacities
for two conditions are shown; one for the as-is condition
FIGURE 20 ·STATIC PUSH OVER RESULTS FOR and one for the damage to the joints and braces repaired
PLATFORME and the leg-pile annulus grouted. The effects of the re­
pairs and grouting are relatively small. This is because
the failure mode is concentrated in the deck legs.
Given the low natural period for tlris platform (Tn < 1
sec), the transient loading ~ nonlinear response correc~
tions to Rus are insignificant (Fv = 1.0). Thus, Ru= Rus.
"
Hurricane Carla subjected tlris platform to waves that
developed maximum lateral forces of approximately SM '
z'i
Q>~
._.! e
10

•
= 4.5 MN. It is likely that these waves approached this <• ....
platform end-on. The storm forces were less than the
platform capacity. The platform was damaged but did ",.
i:i : cl
....w:g
·10

·20
not fail during this storm. The analytical results are in :::e e.,
~g:
...
·30
conformance with this observation.
t=.::•
Platform F
_,_
<~
Q. ... • • • • 10
This platform (Fig. 21) is a 5-leg (4 corner, 1 center), SHEAR CAPACITY OR STORM SHEAR • MN
tender drilling assisted platform that was located in a
water depth of 45.7 min 1963. The leg. pile annulus was
ungrouted and the piles attached to the jacket with FIGURE 22 • PLATFORM F 100 YEAR STORM SHEARS
welded shimmed connections at the top of the jacket. AND SHEAR CAPACITIES
 Nonlinear analyees have been performed on this plat­ assessment of existing platforms. Results from the sim­
form (Fig. 23) (Bea, et al., 1988) for the as-is condition plified analyses can be ueed to help validate results from
and the repaired I grouted condition. The simplified the complex nonlinear analyses. In addition, the ap­
analyses tend to slightly over-predict Rus for this plat­ proaches outlined in this paper offer significant promise
form. The simplified analyses do a good job of predict­ as prsliminary design tools to help engineers better pro­
ing the capacities for both the damaged and the repaired portion lateral load capacity and damage • defect toler­
conditions. ance (robustness) in offshore platforms.
At the present time, a joint industry-government spon­
sored research project is underway to further develop
• and verify the simplified ultimate limit state limit equi ·
• librium analysis procedures. Additional verifications are
being performed on platforms, well protectors, and cais­
4 sons that failed and survived during hurricane Andrew.
Results from these verifications will be reported in the
3
near future.
2 --···/··t·
-a •Repaired & Grouted Legs
I ACKNOWLEDGMENTS
1·······
This paper is the result of a research project conducted
0
0 • 10 . " •m
DECK DISPLACEMENT
30 . 40
at the University of California at Berkeley under the
auspices ofthe Marine Technology Development Group.
Funding for this work has been provided by the National
and California Sea Grant College Program, Arco
Exploration and Production Technology, the California
State Lands Commission, Exxon Production Research
FIGURE 23 ·STATIC PUSH OVER RESULTS FOR Co., Phillips Petroleum Co., Chevron Corp., Shell Oil Co.,
PLATFORMF the Minerals Management Service, Mobil Research and
Development Co., and UNOCAL Corp. Without this
support, this work could not have been undertaken and
completed.
SUMMARY AND CONCLUSIONS
This paper is funded in part by a grant from the
The results summarized in Table 2 indicate that the
simplified analyses can develop evaluations of platform National Sea Grant College Program, National Oceanic
static lateral capacities that are good approximations of and Atmospheric Administration, U. S. Department of
those derived from detailed nonlinear analyses. The Commerce, under grant number NA89AA-D-SG138,
simplified static capacity bias (BRus = nonlinear Rus I project numbers R/OE-11 and R/OE-19 through the
California Sea Grant College, and in part by the
simplified Rus) for the seven verification cases discussed
California State Resources Agency. The views expressed
=
in this paper ranges from B Rus 0.80 to 1.03 with a
herein are those of the author and do not necessarily re·
mean value of B Rus = 0.95. fleet the views of NOAA or any of its sub-agencies. The
Although not discussed in detail in this paper, the sim­ U. S. Government is authorized to reproduce and dis­
plified analyses of storm wind, wave, and cutTent load­ tribute for governmental purposes.
ings are in good agreement with results from detailed
analyses. The simplified analyses are generally within ±
10 % of the detailed results as long as the same input is REFERENCES
used for the structure characteristics, environmental American Petroleum Institute, API, 1993,
conditions, and force computations. Recommended Practice for Planning, Designing and
Comparisons of the estimated lateral load capacities Constructing Fixed Offshore Platforms. AP!
with the estimated maximum loadings that these plat­ Recommended Practice 2A (RP2A), Twentieth Edition,
forms have experienced and with the observed perfor­ August, Dallas, TX.
mance characteristics of these platforms indicates that Bea, R. G., 1974, "Gulf of Mexico Hurricane Wave
the analytical evaluations of both storm loadings and Heights," Proc. of Offshore Technology Conf., OTC No.
platform capacities are in good agreement with the ex­ 2110, Houston, TX.
perience. Bea, R. G., Puskar, F. J., Smith, C., and Spencer, J. S.
The use of the simplified analytical procedures to esti­ "Development of AIM (Assessment, Inspection,
mate reference storm lateral loadings, platform capaci­ Maintenance) Programs for Fixed and Mobile
ties, and Reserve Strength Ratios are indicated to result Platforms," Proceedings Offshore Technology
in reasonable estimates that can be used for the rew Conference, OTC 5703, Houston, Texas, 1988.
 Bea, R. G., 1990, Reliability Based Design Criteria for Proceedings of the Offshore Technology Conference,
Caastal and Ocean Structures, National Committee on OTC No. 7075, Houston, Texas, 1992.
Coastal and Ocean Engineering, The Institution of van de Graaf, J. W., and Tromans, P.S. "Statistical
Engineers, Australia, Barton, ACT. Verification of Predicted Loading and tntimate Strength
Bea, R. G., 1991, Loading and Load Effects Against Observed Storm Damage for an Offshore
Uncertainties. Report to Canadian Standards Structure." Proceedings of the Offshore Technology
Association, Verification Program for CSA Code for the Conference, OTC No. 6573, Houston, TX.
Design, Construction and Installation of Fixed Offshore
Structures Project No. D-3, October, 1991.
Bea, R. G. , 1992a, "Pile Capacity for Axial Cyclic
Loading." J. of Geotechnical Engineering, American
Society of Civil Engineers, Vol. 118, No. 1.
Bea, R. G., 1992b, "Structural Reliability: Design and Re­
qualification of Offshore Platforms." Proceedings of the
International Workshop on Reliability of Offshore
Operations, National Institute of Standards and
Technology, Gaithersburg, Maryland.
Bea, R. G., 1992c, "Re-Qualification of Offshore
Platforms." Proc. Fifth Civil Engineering in the Oceans
Conference, American Society of Civil Engineers, College
Station, TX.
Bea, R. G., 1993, "Evaluation of Uncertainties in
Loadings on Offshore Structures Due to Extreme
Environmental Conditions." J.. Offshore Mechanics and
Arctic Engineering, Vol. 115.
Bea, R. G., and Young, C. N., 1993, "Loading and
Capacity Effects on Platform Performance in Extreme
Storm Waves and Earthquakes," Proceedings of the
Offshore Technology Conference, OTC No. 7140,
Houston, TX.
Bea, R. G., Cornell, C. A., Vinnem, J.E., Geyer, J. F.,
Shoup, G. J., and Stahl, B., 1994, "Comparative Risk
Assessment of Alternative TLP Systems: Structure and
Foundation Aspects," Jl. of Offshore Mechanics and
Arctic Engineering, Vol. 116.
Hilda Meeting Transcript., 1964, "Hurricane Hilda
Damage Conference." New Orleans, Louisiana, Nov. 23­
24, 83 p.
Marshall, P. W. and Bea, R. G., 1976, "Failure Modes of
Offshore Platforms," Proceedings, Behavior of Offshore
Structures, BOSS 76, Trondheim, Norway.
0

Ruhl, J. A., 1976, "Offshore Platforms: Observed

Behavior and Comparisons with Theory," Proceedings
Offshore Technology Conference, OTC No. 2553,
Houston, TX.
Stewart, G., Efthymiou, M., and Vugts, J. H., 1988,
"tntimate Strength and Integrity Assessment of Fixed
Offshore Platforms." Proc. of Int. Con{. on Behaviour of
Offshore Structures, BOSS'88, Trondheim, Norway, Vol.
3.
Stewert, G., 1992, "Non-Linear Structural Dynamics by
the Pseudo-Force Influence Method, Part II: Application
to Offshore Platform Collapse." Proc. Int. Offshore and
Polar Eng ([SOPE) Con{., San Francisco, CA.
Tromans, P. S., and van de Graaf, J. W., 1991, "A
Substantiated Risk Assessment of A Jacket Structure."