Research Article Time Intervals for Maintenance of ...
Transcript of Research Article Time Intervals for Maintenance of ...
Hindawi Publishing CorporationMathematical Problems in EngineeringVolume 2013, Article ID 125856, 15 pageshttp://dx.doi.org/10.1155/2013/125856
Research ArticleTime Intervals for Maintenance of Offshore Structures Based onMultiobjective Optimization
Dante Tolentino and Sonia E. Ruiz
Instituto de Ingenierฤฑa, Universidad Nacional Autonoma de Mexico, 04510 Coyoacan, DF, Mexico
Correspondence should be addressed to Dante Tolentino; [email protected]
Received 22 March 2013; Revised 28 June 2013; Accepted 5 July 2013
Academic Editor: Dan Simon
Copyright ยฉ 2013 D. Tolentino and S. E. Ruiz. This is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.
With the aim of establishing adequate time intervals for maintenance of offshore structures, an approach based on multiobjectiveoptimization for making decisions is proposed.The formulation takes into account the degradation of the mechanical properties ofthe structures and its influence over time on both the structural capacity and the structural demand, given amaximumwave height.The set of time intervals for maintenance corresponds to a balance between three objectives: (a) structural reliability, (b) damageindex, and (c) expected cumulative total cost. Structural reliability is expressed in terms of confidence factors as functions of time bymeans of closed-formmathematical expressions which consider structural deterioration.Themultiobjective optimization is solvedusing an evolutionary genetic algorithm. The approach is applied to an offshore platform located at Campeche Bay in the Gulf ofMexico.The optimization criterion includes the reconstruction of the platform. Results indicate that if the first maintenance actionis made in 5 years after installing the structure, the second repair action should be made in the following 7 to 10 years; however, ifthe first maintenance action is made in 6 years after installing the structure, then the second should be made in the following 5 to8 years.
1. Introduction
Structures are continually exposed to different environmentalloads that modify both their mechanical properties and theirstructural performance over time. Due to these modifica-tions, it is useful to formulate maintenance plans for thestructural systems. Several approaches of maintenance planshave been proposed in the literature considering differentnumbers of optimization objectives: that is, (a) minimizingthe expected cumulative maintenance cost over a time inter-val (as a single objective) [1], (b) minimizing maintenancecost and maximizing load-carrying capacity and durability(as two independent objectives) [2], and (c) more powerfulapproaches that optimize several objectives simultaneously[3โ6].
In this paper a multiobjective optimization approach isproposed. Such approach considers a balance among thefollowing objectives: (1) reliability, expressed in terms of theconfidence factor that takes into account the structural dete-rioration over time by means of closed-form mathematicalexpressions, (2) damage index, expressed in terms of the ratio
between the structural capacity and the structural demand,and (3) expected cumulative total cost, which considers costsof inspection, repair, and failure.The optimization problem issolved by means of the elitist Nondominated Sorting GeneticAlgorithm (NSGA II) [7] which gives place to a set ofnondominated solutions that are used to recommend timeintervals for making decisions for maintenance of an offshorestructure.
For the particular case of offshore structures, the struc-tural deterioration phenomenon is mainly due to fatiguecaused by waves acting continuously on the steel elements.The decrease of the resistance of the structural capacity overa time interval is caused by the cracking of the tubularjoints. The inspection of an offshore structural system hastwo objectives: (a) to detect both the presence and the sizeincrease of cracks and (b) to perform the necessary main-tenance actions to the structure accordingly. Many authorshave proposed inspection andmaintenance plans for offshorestructures using different approaches such as the following:(a) the analysis of risk and reliability of welded connectionssubject to fatigue [8โ10]; (b) the use of methodologies that
2 Mathematical Problems in Engineering
take into account a fatigue sensitivity analysis in steel joints[11]; (c) the application of probabilistic detection methodsin order to study the influence of repetitive inspection forfatigue in joints [12]; (d) the implementation of simplifiedapproaches and the use of practical design parameters such asfatigue design factors [13] and/or reserve strength ratios [14];(e) the consideration of damage caused by fatigue, buckling,and dents on structural elements [15], and (f) the use ofBayesian techniques [16].
However, neither a multiobjective criterion for makingdecisions of maintenance actions on โjacketโ marine plat-forms, nor simplified closed-form mathematical expressionsfor taking into account the structural deterioration of bothstructural capacity and structural demand over a time inter-val are considered in the studies mentioned above, as it isdone in the present study. The advantage of using closed-formmathematical expressions is the possibility of evaluatingthe structural reliability without having to performnumericalintegrations.
2. Basic Definitions of Structural Reliability,Damage Index, andExpected Cumulative Total Cost
2.1. Structural Reliability. The structural reliability can beexpressed bymeans of different indicators, such as probabilityof failure in a time interval, expected number of failures perunit time, confidence factors, and confidence levelss. In thispaper, the confidence factor, ๐conf, is used as a measure ofthe reliability implicit in the structure. The confidence factorwas originally introduced in order to evaluate the reliability ofsteel building structures [17โ20]. Later, the original method-ology was extended to the case in which the mechanicalproperties of the structural elements present degradation (i.e.,due to corrosion, fatigue, etc.), and as a consequence, eitherthe structural capacity or the structural demand, or even bothchange over time [21โ24]. In the present study the confidencefactor is evaluated under the assumption that the struc-tural capacity deteriorates linearly over time. For this case,the confidence factor, ๐conf, is calculated as mentioned inAppendix A.
2.2. Damage Index. The damage index (DI) is defined asthe ratio of the structural capacity, ๐ถ, and the structuraldemand,๐ท. The damage index varies between 0 and 1, where0 indicates that the structure does not have any damage, and 1refers to total damage, whichmeans that the structural systemhas reached its failure condition.
2.3. Expected Cumulative Total Cost. The expected cumu-lative total cost function is defined as the sum of theexpected cost of inspection, repair, and failure, at the end ofa time interval [0, ฮ๐ก), as derived in [25] and presented inAppendix B.
3. Basic Definitions of Genetic Algorithms
Genetic algorithms (GAs) are adaptive methods used tosolve, search, and optimize problems. Based on the principle
Objective function 1
Pareto front
Dominated solutions
Nondominated solutionsO
bjec
tive f
unct
ion
2
a
b
Figure 1: Dominated and nondominated solutions.
of the survival of the fittest [26], GAs have been appliedsuccessfully in many problems and different areas (i.e.,economics, medicine, engineering, etc.). GAs are commonlycomposed of three basic operators: (1) selection, (2) crossover,and (3) mutation. Recently, GAs have been used by civilengineers due to the ease of handling multiple goals directlyand simultaneously [27, 28]. In this study, GAs are selectedas a tool of optimization in order to solve a problem inwhich three objectives must be satisfied. Commonly, inmultiobjective optimization a set of nondominated solutionsis obtained, all of which form a Pareto front. The concept ofdominance is as follows [29].
Assuming that all objective functions are to bemaximized(see Figure 1), it is said that the solution ๐ is dominated bysolution ๐ if conditions 1 and 2 are true.
(1) Solution ๐ is not worse than ๐ considering all theobjectives.
(2) Solution ๐ is strictly better than ๐ in at least oneobjective.
In order to find a Pareto front that helps the decisi-ons makers to formulate plans of maintenance of offshorestructures which present structural degradation over a timeinterval, the elitist NSGA II method [7] was used in thepresent study. The approach consists of the following steps:
(1) Randomly create an initial population ๐๐of size ๐,
where ๐ = generation index (๐ = 0, 1, 2, . . . , ๐).(2) Selection, crossover, and mutation are made to create
the offspring population ๐๐of size๐.
(3) Combine parent, ๐๐, and offspring, ๐
๐, to create the
population ๐ ๐of size 2๐.
Mathematical Problems in Engineering 3
(4) A fast nondominated sort of size ๐ ๐is made to create
the population ๐๐+1
of size๐.(5) Repeat the process from the second step using the
population selected above to create the offspringpopulation, ๐
๐+1.
The offspring population ๐๐is obtained using binary
tournament selection, intermediate crossover, and Gaussianmutation. These concepts are explained below.
Binary Tournament Selection. ๐ individuals (commonly 2)are selected randomly from the size population. The selectedindividuals compete against each other. The individual withthe highest fitness win, and it will be included for crossoverand mutation [30].
Intermediate Crossover. It is the process of creating offspringsby a weighted average of the parents, and it is controlled bythe following ratio [31, 32]:
Offspring = parent1 + ratio (parent2 โ parent1) . (1)
The value of โratioโ is commonly assumed equal to 1.2 [32].
Gaussian Mutation. It consists in adding a random numbertaken from aGaussian distribution withmean 0 to each entryof the parent vector.The standard deviation (๐) is determinedby the parameters โscaleโ and โshrink,โ as follows:
๐ = scale(1 โ shrink(๐๐)) , (2)
where the โscaleโ parameter determines the standard devia-tion at the first generation; the โshrinkโ parameter controlshow the standard deviation shrinks as generations go by; ๐ isthe current generation, and ๐ represents the number of totalgenerations under consideration. The values of โscaleโ andโshrinkโ are commonly assumed equal to 1.0 [32].
Fast Nondominated Sort. A โnaive and slowโ procedure [29]is applied in order to sort a population into different nondo-mination levels. In this procedure each solution is comparedwith every other solution in order to find out if it is dominatedby any other solution in the population. When the compar-isons of each solution are completed, all individuals in thefirst nondominated front are found. All individuals obtainedin the first nondominated front are set apart temporarilyin order to find the second nondominated front, and theprocedure is repeated. The third nondominated front isobtained by setting apart temporarily all the individuals in thefirst and second front, and the procedure mentioned above isrepeated. In summary, different fronts๐น = (๐น
1, ๐น2, . . . , ๐น
๐) are
obtained in this approach, where๐น1is the best nondominated
set, ๐น2is the next best nondominated set, and so on. In case
of ๐น1< ๐, all members of the set ๐น
1are chosen to create
the new population ๐๐+1
; the remaining members are chosenfrom the subsequent nondominated sets in the order of theirranking (๐น
2, ๐น3, . . . , ๐น
๐). If ๐น
1> ๐, the members are chosen
using a crowded-comparison operator [29], in descendingorder and the best solutions needed to fill the population slotsare selected.
4. General Approach of the MultiobjectiveOptimization for Structural Maintenance
Themultiobjective optimization criterion for maintenance ofoffshore structures formulated in the present study considersthree objectives: confidence factor (๐conf), damage index(DI), and expected cumulative total cost (E[CTC]). The firsttwo objectives are subjected to the following constraints:
(i) the confidence factor is subject to ๐conf โฅ 1,(ii) the damage index is subject to DI โค ๐
0.
That is, the structural reliability expressed in terms ofthe confidence factor, ๐conf, must be greater than unity [18](values less than 1 consider that the structure is in the unsafeside). The damage index (DI) should be smaller or equal to agiven threshold (๐
0) which is related to a permissible level of
structural damage.The procedure to obtain the set of nondominated solu-
tions is divided here in two general steps: (a) the simulationof values of the three objectives (๐conf, DI, and E[CTC]) overa time interval, as indicated in the flowchart of Figure 2,and (b) the codification of the simulations, to perform themultiobjective optimization by means of NSGA II.
The simulation profiles are codified in order to performthe multiobjective optimization by using the GA mentionedin the previous section. Then, the optimization problem issolved as follows:
(1) maximize the minimum confidence factor;(2) minimize the maximum damage index;(3) minimize the expected cumulative total cost.
As a result, a set of nondominated solutions is found formaking decisions and for establishing different time intervalfor maintenance. In this study, the criterion for selectingthe maintenance time interval is based on a threshold valueof the confidence factor. When the confidence factor takesvalues between a certain range of values of the nondominatedsolutions, maintenance actions should be performed. Therange of values ๐conf to select the repairing time interval isas illustrated in Figure 3, where the nondominated solutionsin the bracket are those selected for repairing actions.
Here it is assumed that after several nondominatedsolutions are selected for maintenance actions, the criticalstructural elements are repaired to recover their initialcapacities, and a new set of nondominated solutions canbe obtained after reconstruction. Then, a second set ofnondominated solutions is obtained following the flowchartshown in Figure 2 (as done previously); however, an initialcumulative damage (ICD) is now taken into account for thesecond set of critical structural elements to be analyzed. Itis noticed that in the example presented in Section 5 thesecond set of critical structural elements is different from thefirst set analyzed in the first step. For illustrative purposes,Figures 4(a), 4(b), and 4(c) show projections of one selectedsolution (indicated with circles) corresponding to the firstmaintenance time interval, as well as one solution (indicatedwith triangles) corresponding to the second time interval forstructural maintenance.
4 Mathematical Problems in Engineering
For each simulation
t = 0, Random variables are generated for structural capacity, C, andstructural demand, D, based on their probability density functions
t = t + ฮt
t < service life?
All simulations are done?
End
Start
Yes
Yes
Yes
Yes
No
No
No
No
DI โค d0
๐conf โฅ 1๐conf
Calculate E[CTC]Calculate E[CTC]
Figure 2: Flowchart corresponding to the three objectives simulation.
Dominated solutions
Pareto front
Repair
๐(conf)
DI
E[CT
C]
Figure 3: Criterion to select repairing time intervals.
5. Illustrative Example
Themultiobjective approach proposed above is applied to anoffshore โjacketโ platform with the aim of finding its first
and its second time interval for maintenance. The platformis supposed to be located in Campeche Bay. The structureis 48m high and the water depth at the site is 45.11m (seeFigure 5). The expected weight of the deck was assumedequal to 500 ton. The โjacketโ platform is represented bya simplified 2D model with mean mechanical properties.The joints were selected according to their global structuralcapacity contribution. Joints 1, 2, 3, 4, and 6 were selected forfinding the first time interval for maintenance, whereas joints5, 7, 8, and 9were selected for finding the second time interval(once the joints 1, 2, 3, 4, and 6 were totally reconstructed).
5.1. Evaluation of Structural Capacity over Different TimeIntervals. The lateral resistance of the structure was obtainedby means of nonlinear static analyses using twenty differentsimulated profiles of lateral loads. These profiles are relatedto the acting forces corresponding to the simulated wavesthat produced the maximum base shear on the structure.The appearance and growth of cracks, which give place toa reduction of the structural capacity, are considered asdamage condition. Two crack points were considered for eachstructural element connected to the joints.
The average crack size simulation process of the selectedpoints subject to random load was performed using theMonte Carlo technique. Operational and storm waves wereconsidered. The arrival times between storms were assumedwith exponential distribution.The storm wave heights followa Gumbel distribution. The statistical parameters used forsimulation of the crack size were taken from Silva and
Mathematical Problems in Engineering 5
Selected solution for the first time intervalSelected solution for the second time interval
Repairing action
Time (years)
Interval ofvalues forrepairing
action
๐conf
(a)
Selected solution for the first time intervalSelected solution for the second time interval
Repairing action
Time (years)
d0
DI
(b)
Selected solution for the first time intervalSelected solution for the second time interval
Time (years)
E[CT
C]
(c)
Figure 4: Projections of selected solutions corresponding to the first and second time intervals for repairing actions.
Heredia [33]. The crack size was obtained using the modifieddifferential equation by Paris and Ergodan [34] and Sobczykand Spencer [35]. In the differential equation, the randomload was replaced by an equivalent cyclic load whose ampli-tude and frequency were expressed in terms of the meanproperties of the random process. After obtaining the timehistory of the crack size, the capacity of the joint was reducedby a linear correction factor as follows [36, 37]:
๐๐= ๐๐(1 โ (
๐ดcrack๐ด joint
)) , (3)
where ๐๐is the capacity of the cracked joint; ๐
๐is the capacity
of the intact joint which can be calculated by means ofthe API recommendations [38]; ๐ด joint and ๐ดcrack are theareas of the transverse section and of the crack, respectively,
corresponding to all the elements connected to the joint. Asan example, percentages of simulated cracked areas at the endof different time intervals, corresponding to joint 2, are shownin Figure 6.
Themedian of structural capacity,๐ถ, represented in termsof the global displacement (in meters, m) of the platform,and the standard deviation of its natural logarithm, ๐ln๐ถ|ฮ๐กare shown in Table 1 for different time intervals (ฮ๐ก =
0, 1, 2, . . . , 15 years). It is assumed that both the structuralcapacity and the structural demand follow a lognormaldistribution in any time [17โ21].
5.2. Evaluation of the Structural Demand over Different TimeIntervals. The structural demand (given a maximum waveheight) at the end of a time interval was obtained by means
6 Mathematical Problems in Engineering
โ45.110m
โ31.394m
โ17.678m
โ6.906m
+3.658m 13 14
12
9
6
32
1
45
78
1011
Figure 5: 2D model of the โjacketโ platform.
020406080
100
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15Time (years)
Ajo
int/A
crac
kร
100
Figure 6: Percentages of simulated cracked areas. Joint 2.
Table 1: Median of structural capacity, ๐ถ, and standard deviationof the natural logarithm of the structural capacity at different timeintervals.
ฮ๐ก (years) ๐ถ (m) ๐ ln๐ถ|ฮ๐ก
0 0.2520 0.01401 0.2516 0.05742 0.2492 0.09013 0.2462 0.10544 0.2429 0.11305 0.2378 0.15006 0.2322 0.18137 0.2297 0.19028 0.2272 0.19189 0.2206 0.197110 0.2190 0.197811 0.2179 0.207912 0.2151 0.212213 0.2116 0.213714 0.2106 0.230515 0.2092 0.2334
of dynamic โstep by stepโ analysis, using simulated timehistories of waves that correspond to different maximumwave heights. The crack growth simulations and the capacityreductions at selected joints considered to evaluate the capac-ity ๐ถ over time (Section 5.1) were also used in this sectionas damage condition. Figures 7(a) and 7(b) show the median
structural demand as a function of the platform globaldisplacement, and the corresponding standard deviation ofthe natural logarithm of the demand, for different intervalsฮ๐ก, respectively. The expression that fits the median valuesof the structural demand, ๐ท, given a maximum wave height,โmax, at the end of different time intervals ฮ๐ก, is given by๐ท(ฮ๐ก) = (๐+๐ โ ฮ๐ก) โ โ
๐
max (a similar expression was proposedby Tolentino et al. [24]). According to the above, the medianstructural demand, given amaximumwave height, is equal to๐ท(ฮ๐ก) = (3.75๐ธ โ 04 + 5.0๐ธ โ 08 โ ฮ๐ก) โ โ
2.0
max. Similarly, thecorresponding standard deviation of the natural logarithm ofthe demand is given by ๐ln๐ท|โmax ,ฮ๐ก
= (1.65๐ธโ 03 + 1.5๐ธโ 05 โ
ฮ๐ก) โ โ1.5
max.
5.3. Damage Index. In this study the damage index is definedas follows:
DI =๐ถ0โ ๐ถ (ฮ๐ก)
๐ถ0โ ๐ท (ฮ๐ก)
, (4)
where ๐ถ0is the maximum global displacement of the
structure without damage, ๐ถ(ฮ๐ก) represents the maximumglobal displacements that consider structural deteriorationat instant ฮ๐ก, and ๐ท(ฮ๐ก) is the ultimate global displacementdemanded by the structure at instant ฮ๐ก.
5.4. Confidence Factor and Expected Cumulative Total Cost.The confidence factor values, ๐conf (0, ฮ๐ก), were calculatedusing (A.1). The parameters ๐ผ and ๐ฝ were fitted according to(A.4), based on the median values of the structural capacity,๐ถ, shown in Table 1. The epistemic uncertainties associatedwith the structural demand, ๐
๐๐ท, and with the structural
capacity, ๐๐๐ถ
, were assumed equal to 0.15. The parameters ๐and ๐ fitted to thewave hazard curvewere equal to ๐ = 5.0๐ธ03and ๐ = 5, which correspond to a maximum wave heightequal to โmax = 23m and to a return interval of 1485 yearsfor the failure condition [39].
The expected cumulative total cost was evaluated using(B.1). An annual discount rate ๐พ = 6% was assumed. Theinspection cost by jointwas 3,518USD [40] and the repair costwas 20,000USD for each critical point [41].The cost of failureincludes costs due to equipment damage, pollution, deferredproduction, and indirect losses [25].
5.5. Simulation of the Three Objectives. 1024 profiles ofdamage index, confidence factor, and expected cumulativetotal cost, considering cumulative damage at the end of atime interval [0, ฮ๐ก), were simulated, following the procedurepresented in Figure 2. The simulations were based on themedian values and the standard deviations of structuralcapacity and structural demand. Lognormal distributionfunctions were assumed.
5.6. Multiobjective Optimization before Reconstruction (FirstTime Interval for Maintenance). The simulations were codi-fied in order to perform the optimization by means of NSGAII. A population size of 50 was considered. A set of 12 optimalsolutions in 24 generations were found. Different views of
Mathematical Problems in Engineering 7
0.00
0.05
0.10
0.15
0.20
0.25
0 5 10 15 20 25
Without damageฮt = 3 yearsฮt = 6 years
ฮt = 9 yearsฮt = 12 yearsฮt = 15 years
hmax (m)
Med
ian
of d
eman
d,D
(m)
(a)
0.00
0.05
0.10
0.15
0.20
0.25
0 5 10 15 20 25
Without damageฮt = 3 yearsฮt = 6 years
ฮt = 9 yearsฮt = 12 yearsฮt = 15 years
๐lnD|h
max,ฮ
t
hmax (m)
(b)
Figure 7: (a) Median of structural demand and (b) standard deviation of the natural logarithm of the demand, for different time intervals.
Table 2: Comparison between solutions A, B, C, D, E, and F.
Solution Time interval(years)
Damage index(DI)
Confidence factor(๐conf)
Expected cumulativetotal cost E[CTC]
(USD)A 6 0.2345 1.0965 1105415B 5 0.2387 1.1375 930566C 5 0.2421 1.1452 850778D 4 0.2762 1.2246 786761E 4 0.2879 1.2381 818191F 3 0.2939 1.2378 665032
the Pareto front that include damage index, confidence factor,and the expected total cost over a time interval are shownin Figures 8(a), 8(b), and 8(c). Figure 8(d) shows the sameresults in 3D view. It can be observed that the damage indexvaries between 0.23 and 0.29, the confidence factor presentsvalues between 1.10 and 1.24, and the expected cumulativetotal cost takes values between 6.62๐ธ + 05 and 1.10๐ธ + 06
USD. In order to compare different time intervals to performmaintenance actions, six optimal solutions (A, B, C, D, E,and F) were selected (see Figures 8(a)โ8(d)). A comparisonamong the solutions selected is shown in Table 2.
With the aim of establishing time intervals of structuralmaintenance, solutions associated with values of ๐conf โค
1.15 were selected for repairing actions. Then, in accordanceto Table 2, maintenance actions on joints 1, 2, 3, 4, and 6should be performed between 5 and 6 years (correspondingto solutions A, B, and C).
5.7. Multiobjective Optimization after Reconstruction (SecondTime Interval for Maintenance). In this section it is assumedthat the maintenance (repairing) actions on joints 1, 2, 3,4, and 6 are performed in accordance with solution A, and
alternatively, with solution C. It means that the first repairis done after 6 years of installing the platform (solution A),or alternatively, after 5 years if solution C is selected (seeTable 2). Also, it is considered that after maintenance, thejoints recover their full capacity. Taking this into account, thejoints that should be repaired in the next maintenance planare joints 5, 7, 8, and 9 (which were selected in accordancewith their contribution to global structural capacity).
The evaluation of the structural capacity and the struc-tural demand (given a maximum wave height) at the endof a time interval [0, ฮ๐ก), after the first maintenance action,was performed in a similar way as in Sections 5.1 and 5.2;however, it was now considered that joints 5, 7, 8, and 9presented an initial cumulative damage (ICD).The ICD takesinto account the deterioration due to fatigue that the jointssuffered during the time interval between the installation ofthe structure and its first repair. Figure 9 shows an exampleof the percentages of cracked areas of joint 5 at the end ofdifferent time intervals, considering an ICD corresponding to5 years (which is associated with solution C). It is noticed thatin Figure 9 there is an initial cumulative damage (representedby ๐ด joint/๐ดcrack ร 100) at ฮ๐ก = 0 years (on the contrary,
8 Mathematical Problems in Engineering
Con
fiden
ce fa
ctor
1.25
1.20
1.15
1.10
1.05
1.00
Damage index0.22 0.24 0.26 0.28 0.30
Dominated solutions23rd generationPareto front
A
BC
D E F
(a)Ex
pect
ed cu
mul
ativ
e tot
al co
st U
SDร 1000 1800
1600
1400
1200
1000
800
600
1.00 1.05 1.10 1.15 1.20 1.25Confidence factor
Dominated solutions23rd generationPareto front
A
BC D
E
F
(b)
0.22 0.24 0.26 0.28 0.30Damage index
Expe
cted
cum
ulat
ive t
otal
cost
USD
ร 1000 1800
1600
1400
1200
1000
800
600
Dominated solutions23rd generationPareto front
A
BC
DE
F
(c)
Expe
cted
cum
ulat
ive t
otal
cost
USD
ร 1000
1800
1600
1400
1200
1000
800
600
Confidence factor
1.001.05
1.101.15
1.201.25 Dam
age index
0.220.24
0.260.28
0.30
Dominated solutionsPareto front
A
BC D E
F
(d)
Figure 8: Pareto front corresponding to the three objective functions and optimal solutions A, B, C, D, E, and F.
Figure 6 does not indicate any initial cumulative damage injoint 2).
Tables 3 and 4 show the median of the structural capac-ity, ๐ถ, represented in terms of global displacement of theplatform, corresponding to solutions A and C, respectively,for different time intervals (ฮ๐ก) after the first maintenanceactions. The standard deviations of the natural logarithmof the structural capacity are also shown in Tables 3 and4.
The median of the structural demand (given a maximumwave height, โmax) was calculated as in Section 5.2. The
expressions obtained are ๐ท(ฮ๐ก) = (3.55๐ธ โ 04 + 4.1๐ธ โ 08 โ
ฮ๐ก) โ โ2
max and ๐ท(ฮ๐ก) = (3.4๐ธ โ 04 + 4.0๐ธ โ 08 โ ฮ๐ก) โ โ2
max,for solutions A and C, respectively. The standard deviationsof the natural logarithm of the demand, for a maximumwaveheight, at the end of different time intervals (ฮ๐ก) were fittedas ๐ln๐ท|โmax ,ฮ๐ก
= (1.5๐ธ โ 03 + 1.5๐ธ โ 05 โ ฮ๐ก) โ โ1.55
max and๐ln๐ท|โmax ,ฮ๐ก
= (1.4๐ธ โ 03 + 1.4๐ธ โ 05 โ ฮ๐ก) โ โ1.55
max for solutionsA and C, respectively.
A total of 1024 simulations corresponding to solution A,and alternatively, to solution C were performed followingthe procedure presented in Figure 2. The simulations were
Mathematical Problems in Engineering 9
0
20
40
60
80
100
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15Time (years)
Ajo
int/A
crac
kร
100
Figure 9: Percentages of the simulated cracked area on joint 5, withan ICD corresponding to solution C.
Table 3: Median of the structural capacity, ๐ถ, and standard devia-tion of the natural logarithm, at the end of different time intervals,corresponding to solution A.
ฮ๐ก (years) ๐ถ (m) ๐ ln๐ถ|ฮ๐ก
0 0.2530 0.01291 0.2515 0.03362 0.2499 0.08853 0.2474 0.11414 0.2448 0.14125 0.2420 0.15926 0.2373 0.18037 0.2365 0.18708 0.2325 0.19669 0.2304 0.202610 0.2296 0.203311 0.2286 0.219312 0.2249 0.217613 0.2246 0.224714 0.2227 0.236215 0.2213 0.2578
codified for each case in order to realize the optimizationprocedure using NSGA II. A population size of 50 wasconsidered. A set of 13 optimal solutions with 23 generationsfor solution A was found, and 12 optimal solutions with25 generations for solution C were obtained. Figures 10and 11 show the set of solutions (Pareto fronts) for casesA and C, respectively. Next, some optimal solutions areselected in order to handle different options for struc-tural maintenance. A comparison of two sets of solutionsselected is shown in Tables 5 and 6 (see Figures 10 and11).
In a similar way to the previous section, the time intervalsfor the second maintenance actions are performed whenthe structural confidence values are lower than 1.15. Tables5 and 6 show that maintenance actions could be done attime intervals between 5 and 8 years (solutions a
1, b1, c1,
and d1) if the first maintenance actions were done after 6
years of installing the platform, or alternatively, at intervalsbetween 7 and 10 years (solutions a
2, b2, c2, and d
2) if the
Table 4: Median of the structural capacity, ๐ถ, and standard devia-tion of the natural logarithm, at the end of different time intervals,corresponding to solution C.
ฮ๐ก (years) ๐ถ (m) ๐ ln๐ถ|ฮ๐ก
0 0.2534 0.01461 0.2518 0.03832 0.2503 0.08633 0.2495 0.11964 0.2476 0.12505 0.2468 0.13986 0.2448 0.14337 0.2417 0.15098 0.2392 0.16029 0.2375 0.166810 0.2364 0.170211 0.2342 0.184112 0.2315 0.191513 0.2300 0.192014 0.2285 0.202615 0.2273 0.2332
Dominated solutionsPareto front
0.300.28
0.260.24
0.22 Damage index
1.001.05
1.101.15
1.201.25
Confidence factor
2600
2400
2200
2000
1800
1600Expe
cted
cum
ulat
ive t
otal
cost
USD
ร 1000
a1
b1c1
d1
e1f1
Figure 10: Pareto front with 13 optimal solutions consideringsolution A for the first time interval for maintenance and a set ofsolutions selected (a
1, b1, c1, d1, e1, and f
1) for the second interval
for maintenance.
first maintenance actions were performed after 5 years ofinstalling the platform.
Figures 12(a), 12(b), and 12(c) showprojections of solutionA (see Table 2) together with projections of solutions a
1, b1,
c1, and d
1corresponding to the three objectives (damage
index, confidence factor, and expected cumulative total cost).Similarly, Figures 13(a), 13(b), and 13(c) show projections of
10 Mathematical Problems in Engineering
Table 5: Comparison between solutions a1, b1, c1, d1, e1, and f1. ICD corresponding to 6 years.
Solution Time interval(years)
Damage index(DI)
Confidence factor(๐conf)
Expected cumulativetotal cost E[CTC]
(USD)a1 8 0.2277 1.0859 2319250b1 7 0.2388 1.0935 2190760c1 6 0.2435 1.1119 2095360d1 5 0.2448 1.1136 2014650e1 3 0.2489 1.1630 1794310f1 3 0.2535 1.1751 1756520
Table 6: Comparison between solutions a2, b2, c2, d2, e2, f2, and g2. ICD corresponding to 5 years.
Solution Time interval(years)
Damage index(DI)
Confidence factor(๐conf)
Expected cumulativetotal cost E[CTC]
(USD)a2 10 0.2284 1.0827 2133420b2 9 0.2309 1.0886 2040560c2 8 0.2371 1.1200 1870170d2 7 0.2375 1.1255 1813980e2 5 0.2680 1.2029 1665650f2 4 0.2734 1.2284 1523060g2 4 0.2980 1.2458 1554500
Dominated solutionsPareto front
0.300.28
0.260.24
0.22 Damage index
1.001.05
1.101.15
1.201.25
Confidence factor
2200
2000
1800
1600
1400Expe
cted
cum
ulat
ive t
otal
cost
USD
ร 1000
a2 b2
c2d2
e2
f2g2
Figure 11: Pareto front with 12 optimal solutions consideringsolution C for the first time interval for maintenance and a set ofsolutions selected (a
2, b2, c2, d2, e2, f2, and g
2) for the second interval
for maintenance.
solution C together with projections of solutions a2, b2, c2,
and d2.
In summary, in accordance with the criterion selected forperforming maintenance actions (values of ๐conf < 1.15),Figures 12(a) and 12(b) show that the first optimal timeinterval for reconstructing joints 1, 2, 3, 4, and 6 can beafter 6 years of installing the offshore โjacketโ platform (seeTable 2, solution A). If this solution is selected, then thesecond optimal time interval for repairing actions on joints5, 7, 8, and 9 should be after 5, 6, 7, or 8 years (solutions a
1, b1,
c1, or d1) of reconstructing the structure. On the other hand,
if the maintenance actions for the first time are done after 5years of installing themarine structure (solutionC inTable 2),then the repairing actions on joints 5, 7, 8, and 9 should beperformed after 7, 8, 9, or 10 years after reconstructing thestructure (solutions a
2, b2, c2, or d
2), as shown in Figures
13(a) and 13(b).
6. Conclusions
An approach based on multiobjective optimization in orderto find adequate time intervals for maintenance, by meansof genetic algorithms, is applied. A set of nondominatedsolutions is found based on three objectives: (1) confidencefactor, which is represented by closed-form mathematicalexpressions that take into account the structural deteriorationover time, (2) damage index, expressed in terms of theratio between the structural capacity and structural demand,and (3) expected cumulative total cost, which considersinspection, repair, and failure costs.
Mathematical Problems in Engineering 11
Con
fiden
ce fa
ctor
1.45
1.40
1.35
1.30
1.25
1.20
1.15
1.10
1.05
1.00
Time (years)0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Solution ASolution a1Solution b1
Solution c1Solution d1
(a)D
amag
e ind
ex
0.25
0.20
0.15
0.10
0.05
0.00
Time (years)0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Solution ASolution a1Solution b1
Solution c1Solution d1
(b)
0 1 2 3 4 5 6 7 8 9 10 11 12 13 140
200400600800
10001200140016001800200022002400
Expe
cted
cum
ulat
ive t
otal
cost
USD
ร 1000
Time (years)
Solution ASolution a1Solution b1
Solution c1Solution d1
(c)
Figure 12: Different projections of solution A and of solutions a1, b1, c1, and d
1.
The evolutionary Nondominated Sorting Genetic Algo-rithmNSGA IIwas selected as a tool to solve the optimizationproblem. Different sets of nondominated solutions werefound in order to recommend different time intervals formaintenance actions of a fixed steel โjacketโ platform sub-jected to structural deterioration over time.The implementedGA is very efficient since it only requires a few seconds to finda set of nondominated solutions based on different objectives;it also requires less than 30 generations in order to obtain thePareto front.
The results show that, for the offshore โjacketโplatform analyzed, it is recommended to make the
maintenance actions on the first set of selected joints,between 5 and 6 years after installing the platform. In thecase of performing the first maintenance action after 5 years,the second time interval for repairing the second set of jointsis recommended to be done between 7 and 10 years. On theother hand, if the first maintenance actions are performedafter 6 years of installing the platform, the second timeinterval for maintenance actions should be between 5 and 8years.
The multiobjective optimization approach presented,together with the closed-formmathematical expression used,constitutes an efficient tool for makingmaintenance plans for
12 Mathematical Problems in Engineering
Con
fiden
ce fa
ctor
1.45
1.40
1.35
1.30
1.25
1.20
1.15
1.10
1.05
1.00
Time (years)0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Solution CSolution a2Solution b2
Solution c2Solution d2
(a)D
amag
e ind
ex
0.25
0.20
0.15
0.10
0.05
0.00
Time (years)0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Solution CSolution a2Solution b2
Solution c2Solution d2
(b)
2200
2000
1800
1600
1400
1200
1000
800
600
400
200
0
Expe
cted
cum
ulat
ive t
otal
cost
USD
ร 1000
Time (years)0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Solution CSolution a2Solution b2
Solution c2Solution d2
(c)
Figure 13: Different projections of solution C and of solutions a2, b2, c2, and d
2.
offshore structures; moreover, it can be adapted to considerdifferent kinds of structures (buildings, transmission towers,bridges, etc.). The criterion can be extended to take intoaccount different types of maintenance actions in order tokeep the structural system within prescribed performancelevels.
A limitation of the approach presented here is thatit is adequate only for making decisions about structuralmaintenance actions for systems subjected to ordinaryexcitations; it does not apply to the identification ofrepair actions that should be taken after the occurrenceof extraordinarylinebreak environmental events, forexample, after the action of hurricanes with a return
period much longer than that considered here (seeSection 5.4).
Appendices
A. Confidence Factor That Takes into Accountthe Structural Deterioration
The confidence factor over the time interval [0, ฮ๐ก) can beexpressed as follows [21, 24]:
๐conf (0, ฮ๐ก) =๐ โ ๐ถ
๐พ โ ๐ท]0[ฮฉ (0, ฮ๐ก)
ฮ๐ก]
โ๐/๐
, (A.1)
Mathematical Problems in Engineering 13
where
๐ = exp [โ ๐
2๐(๐2
ln๐ถ|ฮ๐ก + ๐2
๐๐ถ|ฮ๐ก)] , (A.2)
๐พ = exp [ ๐2๐
(๐2
ln๐ท|โmax ,ฮ๐ก+ ๐2
๐๐ท|ฮ๐ก)] , (A.3)
๐ถ = ๐ผ โ ๐ฝฮ๐ก, (A.4)
๐ท]0 = (๐ + ๐ฮ๐ก) (
]0
๐)
โ๐/๐
, (A.5)
ฮฉ (0, ฮ๐ก) =๐๐ผ
๐ฝ (๐ โ ๐)(
๐๐ฝ
โ๐ผ๐ + ๐ฝ๐)
โ๐/๐
ร [โ๐น (๐; ๐; ๐; ๐ฅ) + ๐น (๐; ๐; ๐; ๐ฅ (ฮ๐ก))
โ โ โ (1 + (๐๐ฝฮ๐ก
๐ฝ๐)
โ๐/๐
)(1 +๐ฝฮ๐ก
๐ผ)
ร (๐ผ + ๐ฝฮ๐ก
๐ + ๐ฮ๐ก)
โ๐/๐
(๐ผ
๐)
๐/๐
] ,
(A.6)
where ๐ and ๐พ are the capacity reduction factor and thedemand amplification factor at instant ฮ๐ก, respectively;๐2
ln๐ท|โmax ,ฮ๐กand ๐2ln๐ถ|ฮ๐ก are the variances of the natural log-
arithm of the demand, ๐ท, given a maximum wave height,โmax, and of the structural capacity,๐ถ, respectively;๐
2
๐๐ถ|ฮ๐กand
๐2
ln๐๐ท|ฮ๐ก represent the epistemic uncertainties associated withthe structural demand and the structural capacity respec-tively; ๐ถ is the median of the structural capacity at the limitstate of interest, which vary linearly in the interval [0, ฮ๐ก);๐ท]0
is the median of the structural demand caused by environ-mental loads related to the tolerable annual exceedance rate]0, at time ฮ๐ก; ฮฉ(0, ฮ๐ก) represents the correction function
of the expected number of failures which considers thevariation in the interval [0, ฮ๐ก) of the structural demand(given a maximum wave height) and the structural capacitysimultaneously.
Hypergeometric functions implicit in (A.6) can be solvedusing a simpler method by means of the following hypergeo-metric series [42, 43]:
๐น (๐, ๐; ๐; ๐ฅ)
= 1 +๐๐
1!๐๐ฅ +
๐ (๐ + 1) ๐ (๐ + 1)
2!๐ (๐ + 1)๐ฅ2+ โ โ โ
+๐ (๐+1) โ โ โ (๐ + ๐ โ 1) ๐ (๐+1) โ โ โ (๐+๐โ1)
๐ (๐ + 1) โ โ โ (๐ + ๐ โ 1) ๐!๐๐,
(A.7)
where
๐ = 1 โ๐
๐, ๐ = โ
๐
๐, ๐ = 2 โ
๐
๐,
๐ฅ =๐๐ผ
๐๐ผ โ ๐๐ฝ, ๐ฅ (ฮ๐ก) =
๐ (๐ฝฮ๐ก + ๐ผ)
๐๐ผ โ ๐๐ฝ,
(A.8)
where๐, ๐, and ๐ฅ represent real numbers, and ๐must be aninteger.The value ๐will depend on the problem; for example,for the case analyzed in this study, the hypergeometric seriesis developed up to ๐ = 5, in order to obtain a goodapproximation.
B. Expected Cumulative Total Cost
The expected cumulative total cost function is defined as thesummation of the total expected cost of inspection, repair,and failure at the end of a time interval [0, ฮ๐ก) as follows [25]:
๐ถTotal (0, ฮ๐ก) = ๐ถ๐|ฮ๐ก โ ๐โ๐พ(ฮ๐ก)โ๐
๐น(0,ฮ๐ก)
+
๐
โ
๐=1
๐ถ๐๐|ฮ๐ก
โ ๐ (๐ท๐ (ฮ๐ก) โฅ ๐) โ ๐
โ๐พ(ฮ๐ก)โ๐๐น(0,ฮ๐ก)
๐ถ๐|ฮ๐ก
๏ฟฝ๏ฟฝ
โ
๐=1
๐๐น(๐ก๐โ ๐ก๐โ1) ๐โ๐พ(ฮ๐ก)โ๐
๐น(0,๐ก๐โ๐ก๐โ1),
(B.1)
where ๐ถ๐|ฮ๐ก
, ๐ถ๐๐|ฮ๐ก
, and ๐ถ๐|ฮ๐ก
are the expected cost of inspec-tion, repair, and failure, respectively; ๐พ is the discount rate;๐๐นis the expected number of failures at the end of a time
interval considering the variation of the structural capacityand structural demand over time [0, ฮ๐ก) [24]; ๐ is the numberof elements to be repaired; ๐(๐ท
๐(ฮ๐ก) โฅ ๐) is the conditional
probability that element ๐ with certain level of cumulativedamage ๐ท, at time ฮ๐ก, is higher or equal to an establisheddamage level ๐; ๏ฟฝ๏ฟฝ is the number of time intervals of interest;๐๐น(๐ก๐โ ๐ก๐โ1) is the expected number of failures for the time
intervals of interest, and ๐ก1, ๐ก2, . . . , ๐ก
๐= 1, 2, . . . , ๏ฟฝ๏ฟฝ years.
Equation (B.1) implies that the structure survives up to theend of the time interval of interest.
Acknowledgments
This research project had the support of DGAPA-UNAM(PAPIIT IN107011-3). The first author thanks CONACYT forthe economic support to develop his PhD research.
References
[1] J. S. Kong and D. M. Frangopol, โLife-cycle reliability-basedmaintenance cost optimization of deteriorating structures withemphasis on bridges,โ Journal of Structural Engineering, vol. 129,no. 6, pp. 818โ828, 2003.
[2] A. Miyamoto, K. Kawamura, and H. Nakamura, โBridge man-agement system and maintenance optimization for existingbridges,โ Computer-Aided Civil and Infrastructure Engineering,vol. 15, no. 1, pp. 45โ55, 2000.
[3] M. Liu and D. M. Frangopol, โOptimal bridge maintenanceplanning based on probabilistic performance prediction,โ Engi-neering Structures, vol. 26, no. 7, pp. 991โ1002, 2004.
[4] H. Furuta, T. Kameda, and M. Erami, โA practical bridgemanagement system using new multi-objective genetic algo-rithm,โ in Bridge Maintenance, Safety, Management, Life-CyclePerformance and Cost, P. J. S. Cruz, D. M. Frangopol, and L. C.
14 Mathematical Problems in Engineering
Neves, Eds., pp. 131โ132, Taylor & Francis Group, London, UK,2006.
[5] L. A. C. Neves, D. M. Frangopol, and P. J. S. Cruz, โProba-bilistic lifetime-oriented multiobjective optimization of bridgemaintenance: single maintenance type,โ Journal of StructuralEngineering, vol. 132, no. 6, pp. 991โ1005, 2006.
[6] N. M. Okasha and D. M. Frangopol, โLifetime-oriented multi-objective optimization of structural maintenance consideringsystem reliability, redundancy and life-cycle cost using GA,โStructural Safety, vol. 31, no. 6, pp. 460โ474, 2009.
[7] K. Deb, A. Pratap, S. Agarwal, and T. Meyarivan, โA fastand elitist multiobjective genetic algorithm: NSGA-II,โ IEEETransactions on Evolutionary Computation, vol. 6, no. 2, pp. 182โ197, 2002.
[8] R. Skjong, โReliability based optimization of inspection strate-gies,โ in Proceedings of the 4th International Conference onStructural Safety and Reliability (ICOSSAR โ85), vol. 3, pp. 614โ618, Kobe, Japan, 1985.
[9] H. O. Madsen, J. D. Sรธrensen, and R. Olesen, โOptimal inspec-tion planning for fatigue damage of offshore structures,โ in Pro-ceedings of the 5th International Conference on Structural Safetyand Reliability (ICOSSAR โ89), pp. 2099โ2106, San Francisco,Calif, USA, August 1989.
[10] J. D. S. Sรธrensen, M. H. Faber, R. Rackwitz, and P. Thoft-Christensen, โModeling in optimal inspection and repair,โ inProceedings of the 10th International Conference on OffshoreMechanics and Arctic Engineering (OMAE โ91), pp. 281โ288,Stavenger, Noruega, 1991.
[11] P.Thoft-Christensen and J. D. Sรธrensen, โOptimal strategies forinspection and repair of structural systems,โ Civil EngineeringSystems, vol. 4, no. 2, pp. 94โ100, 1987.
[12] T. Moan, O. T. Vardal, N.-C. Hellevig, and K. Skjoldli, โIn-service observations of cracks in North Sea jackets. A studyon initial crack depth and POD values,โ in Proceedings of the16th International Conference on Offshore Mechanics and ArcticEngineering (OMAE โ97), pp. 189โ197, Yokohama, Japan, April1997.
[13] D. Straub, J. D. Sรธrensen, J. Goyet, and M. H. Faber, โBenefitsof risk based inspection planning for offshore structures,โ inProceedings of the 25th International Conference on OffshoreMechanics and Arctic Engineering (OMAE โ06), pp. 1โ10, Ham-burg, Germany, June 2006.
[14] M. H. Faber, โReliability based inspection planning of fatiguedamaged offshore platforms,โ in Proceedings of the 1st Interna-tional Symposium on Analysis of Structural Risk and Reliabilityfor Offshore Facilities (IMP โ01), Cancun, Mexico, 2001.
[15] C. E. O. Estrada and D. de Leon Escobedo, โDevelopment of acost-benefitmodel for inspection of offshore jacket structures inMexico,โ in Proceedings of the 22nd International Conference onOffshore Mechanics and Arctic Engineering; Offshore TechnologyOcean Space Utilization, pp. 87โ98, Cancun, Mexico, June 2003.
[16] J. D. Sรธrensen andG. Ersdal, โRisk based inspection planning ofageing structures,โ in Proceedings of the 27th International Con-ference on Offshore Mechanics and Arctic Engineering (OMAEโ08), pp. 399โ408, Estoril, Portugal, June 2008.
[17] C. A. Cornell, F. Jalayer, R. O. Hamburger, and D. A. Foutch,โProbabilistic basis for 2000 SAC federal emergency man-agement agency steel moment frame guidelines,โ Journal ofStructural Engineering, vol. 128, no. 4, pp. 526โ533, 2002.
[18] C. A. Cornell, โCalculating building seismic performance relia-bility: a basis for multi-level design norms,โ in Proceedings of the
11th World Conference on Earthquake Engineering, 1996, paperno. 2122.
[19] Federal Emergency Management Agency (FEMA), โRecom-mended seismic evaluation and upgrade criteria for existingwelded steel moment-frame buildings,โ Tech. Rep. FEMA 351,SAC Joint Venture, Washington, DC, USA, 2000.
[20] S.-Y. Yun, R. O. Hamburger, C. A. Cornell, and D. A. Foutch,โSeismic performance evaluation for steel moment frames,โJournal of Structural Engineering, vol. 128, no. 4, pp. 534โ545,2002.
[21] M. A. Torres and S. E. Ruiz, โStructural reliability evaluationconsidering capacity degradation over time,โ Engineering Struc-tures, vol. 29, no. 9, pp. 2183โ2192, 2007.
[22] D. Celarec, D. Vamvatsikos, and M. Dolsek, โSimplified esti-mation of seismic risk for reinforced concrete buildings withconsideration of corrosion over time,โ Bulletin of EarthquakeEngineering, vol. 9, no. 4, pp. 1137โ1155, 2011.
[23] D. Vamvatsikos and M. Dolsek, โEquivalent constant rates forperformance-based seismic assessment of ageing structures,โStructural Safety, vol. 33, no. 1, pp. 8โ18, 2011.
[24] D. Tolentino, S. E. Ruiz, and M. A. Torres, โSimplified closed-form expressions for the mean failure rate of structures con-sidering structural deterioration,โ Structure and InfrastructureEngineering, vol. 8, no. 5, pp. 483โ496, 2012.
[25] D. Tolentino,Optimizacion multiobjetivo para el mantenimientode estructuras considerando la influencia del dano acumulado[Ph.D. thesis], Graduate Program in Engineering, UniversidadNacional Autonoma de Mexico, 2013 (Spanish).
[26] D. E. Goldberg, Genetic Algorithms in Search, Optimizationand Machine Learning, AddisonWesley Longman Inc., Boston,Mass, USA, 1989.
[27] M. Fragiadakis, N. D. Lagaros, andM. Papadrakakis, โPerform-ance-based multiobjective optimum design of steel structuresconsidering life-cycle cost,โ Structural and MultidisciplinaryOptimization, vol. 32, no. 1, pp. 1โ11, 2006.
[28] O. Lavan and G. F. Dargush, โMulti-objective evolutionaryseismic design with passive energy dissipation systems,โ Journalof Earthquake Engineering, vol. 13, no. 6, pp. 758โ790, 2009.
[29] K. Deb,Multi-Objective Optimization Using Evolutionary Algo-rithms, John Wiley & Sons, Chichester, UK, 2001.
[30] D. E. Goldberg and K. Deb, โA comparative analysis of selectionschemes used in genetic algorithms,โ in Foundations of GeneticAlgorithms (Bloomington, IN, 1990), G. J. E. Rawlins, Ed., pp.69โ93, Morgan Kaufmann, San Mateo, Calif, USA, 1991.
[31] H. M. Muhlenbein and D. Schlierkamp-Voosen, โPredictivemodels for the breeder genetic algorithm,โ Evolutionary Com-putation, vol. 1, no. 1, pp. 25โ49, 1993.
[32] The MathWorks, Global Optimization Toolbox 3 Userโs Guide,Version 3.1, 2010.
[33] F. L. Silva and E. Heredia, โEffect of uncertainties on thereliability of fatigue damaged systems,โ inProceedings of the 23thInternational Conference onOffshoreMechanics and Arctic Engi-neering (OMAE โ04), vol. 2, pp. 427โ434, Vancouver, Canada,2004.
[34] P. Paris and F. A. Erdogan, โCritical analysis of crack propa-gation laws,โ Journal of Basic Engineering, vol. 85, pp. 258โ534,1963.
[35] K. Sobczyk and F. B. Spencer, Random Fatigue: From DataTheory, Academic Press, Boston, Mass, USA, 1992.
[36] A. Stacey, J. V. Sharp, and N. W. Nichols, โStatic strengthassessment of cracked tubular joints,โ in Proceedings of the
Mathematical Problems in Engineering 15
15th International Conference on Offshore Mechanics and ArcticEngineering (OMAE โ96), vol. 3, pp. 211โ224, June 1996.
[37] F. M. Burdekin, โThe static strength of cracked joints in tubularmembers,โ Offshore Technology Report 2001/080, Health andSafety Executive, London, UK, 2002.
[38] American Petroleum Institute (API), Recommended Practicefor Planning Designing and Construction of Fixed OffshorePlatforms Load and Resistance Factor Design, API RP 2a-LRFD,Washington, DC, USA, 1993.
[39] PEMEX, โDesign and evaluation of fixed marine platforms inCampeche Bay,โ NRF-2003-Pemex-2000, Mexico, 2000.
[40] A. Raine, โThe development of Alternating Current Field Mea-surement (ACFM) technology as a technique for the detectionof surface breaking defects in conducting material and its usein commercial and industrial applications,โ in Proceedings of the15th World Conference on Non-Destructive Testing, Rome, Italy,2000.
[41] The Marine Technology Directorate Limited, Review of Repairsto Offshore Structures and Pipelines, vol. 94โ102, 1994.
[42] L. R. Ford,Differential Equations,McGraw-Hill BookCompany,New York, NY, USA, 1955.
[43] E. D. Rainville, Intermediate Course in Differential Equations,John Wiley & Sons, New York, NY, USA, 1961.
Submit your manuscripts athttp://www.hindawi.com
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
MathematicsJournal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Mathematical Problems in Engineering
Hindawi Publishing Corporationhttp://www.hindawi.com
Differential EquationsInternational Journal of
Volume 2014
Applied MathematicsJournal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Probability and StatisticsHindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Journal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Mathematical PhysicsAdvances in
Complex AnalysisJournal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
OptimizationJournal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
CombinatoricsHindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
International Journal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Operations ResearchAdvances in
Journal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Function Spaces
Abstract and Applied AnalysisHindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
International Journal of Mathematics and Mathematical Sciences
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
The Scientific World JournalHindawi Publishing Corporation http://www.hindawi.com Volume 2014
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Algebra
Discrete Dynamics in Nature and Society
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Decision SciencesAdvances in
Discrete MathematicsJournal of
Hindawi Publishing Corporationhttp://www.hindawi.com
Volume 2014 Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Stochastic AnalysisInternational Journal of