Analysis and retrofitting of fixed steel platforms under low seismic loads

1.2 Literature review 1.2.1 Seismic hazard analysis The seismic hazard analysis refers to the estimation of some measure of the strong earthquake ground motion expected to occur at a

Seismic waves generated at the epicenter can be classified into two groups: high frequency waves, which have high intensity but damp out rapidly during propagation; low frequency waves, which have large displacement properties and damp out relatively slowly Because the high frequency waves damp out quickly during propagation, the seismic waves at long distances are often rich in low frequency waves Although the peak ground acceleration (PGA) of the low frequency waves may be very low, the induced motions may have disproportionally high displacement and possibly high velocity characteristics In addition, such low frequency waves may

be amplified several folds through the soft soil layers; if the natural period of the soil

at the site is close to the predominant period of the ground motion at bedrock (this kind of amplification due to resonance is called site amplification effects) Furthermore, such amplification may be further enlarged, if the natural period of the

Chapter 1 Introduction

structures supported on such soil sites is close to the predominant period of the site Therefore, due to the large displacement properties that low frequency waves possess and the amplification by the soil, the fixed steel platforms may be subjected to large displacements that may cause some concern

The platforms which are not designed for earthquake may still resist certain level

of earthquake loading This is due to reserve strength The possible sources of reserve strength are (1) actual strength of the material used in construction is higher than the strength used in the design; (2) effects of structural elements that are not included in the prediction of lateral load capacity; (3) effects of minimum requirements on member sections in order to meet the stability and serviceability limits; and (4) redistribution of internal forces in the inelastic range (Rahgozar and Human 1998) Thus there is a need to evaluate the reserve strength in a fixed steel platform designed for wave load to see whether there is sufficient capacity to meet the demand due to low seismic loads and to investigate possible seismic retrofitting techniques if needed

1.2 Literature review

1.2.1 Seismic hazard analysis

The seismic hazard analysis refers to the estimation of some measure of the strong earthquake ground motion expected to occur at a selected site This is necessary for the purpose of earthquake resistant design of a new structure or for estimating the safety of an existing structure of importance, like dams, nuclear power plants, long-span bridges, high-rise buildings, offshore structures, etc at that site By taking into account the entire available database on seismicity, tectonics, geology and attenuation characteristics of the seismic waves in an area of interest, the seismic hazard analysis provide an estimate of the site-specific design ground motion (Dravinski et al., 1980;

Chapter 1 Introduction

Westermo et al., 1980) One important application of hazard analysis is the preparation of seismic zoning maps for generalized applications (Lee and Trifunac, 1987; Trifunac, 1989a, 1990a; Anderson and Trifunac, 1977, 1978a, 1978b) By estimating the amplitudes of a parameter describing the ground motion at a closely spaced grid of sites covering the complete area of a big city or an entire state, zoning maps are developed by contouring the sub-areas with equal hazard Such maps find useful applications in the earthquake-resistant design of common types of structures, for which it is not possible to carry out the detailed site-specific studies The zoning maps are also useful for land-use planning, assessing the needs for remedial measures, and estimation of possible economical losses during future earthquakes (Trifunac, 1989b; Trifunac and Todorovska, 1998)

Seismic hazard analysis involves the quantitative estimation of ground-shaking hazards at a particular site Seismic hazards may be obtained deterministically (deterministic seismic hazard analysis, DSHA), when a particular earthquake scenario

is assumed, or probabilistically (probabilistic seismic hazard analysis, PSHA), in which uncertainties in earthquake size, location, and time of occurrence are explicitly considered

1.2.1.1 Deterministic seismic hazard analysis

The deterministic seismic hazard analysis aims at finding the maximum possible ground motion at a site by taking into account the seismotectonic setup of the area around the site and the available data on past earthquakes in the area (Krinitzsky, 1995; Romeo and Prestininzi, 2000) For this purpose, the magnitude of the largest possible earthquake is estimated for each of the seismic sources around the site of interest The commonly used forms of seismic sources are the line, area, dipping plane, and the volume sources The point source is also used sometimes when the epicenters

Chapter 1 Introduction

are concentrated in a very small area far away from the site of interest The maximum magnitude in each of the sources is assumed to occur at the closest possible distance from the site The DSHA involves the development of a particular seismic scenario on which the ground motion hazard evaluation is based The scenario consists of the assumed occurrence of an earthquake of a specified size occurring at specified location Earthquake sources may be identified from the records of historical seismicity If sufficient data are available, the maximum intensity can be determined and used to estimate the location of the earthquake epicenter and magnitude of the event A typical DSHA has four-step process (Reiter, 1990) as described below:

1 Identification and characterization of all earthquake sources that can produce significant ground motion at the site

2 Selection of a source-to-site distance parameter for each source zone In DSHA, the shortest distance between the source zone and site of interest is selected

3 Selection of the controlling earthquake is generally expressed in terms of some ground motion parameter, at the site The selection is made by comparing the levels of shaking produced by earthquakes assumed to occur at the distance identified in step 2 The controlling earthquake is described in terms of its size and distance from the site

4 The hazard at the site is defined, usually according to the ground motions produced at the site by the controlling earthquake Peak acceleration, peak velocity and response spectrum ordinates are used to characterize the seismic hazard

Although the deterministic seismic hazard analysis seems to be a very simple procedure and provides a direct framework for evaluation of worst-case ground motions, it cannot present information on the likelihood of occurrence of the

1.2.1.2 Probabilistic seismic hazard analysis

During the past 50 years, the use of probabilistic concept has made the uncertainties in the location, size and rate of recurrence of earthquakes and ground motion characteristics to be explicitly considered in the evaluation of seismic hazards Cornell (1964) studied the probability distribution of a dependent variable and derived its relationship with other independent variables whose probability distributions are known or can be assumed Esteva (1966) and Rosenblueth (1968) studied the earthquake ground motions, their dependence on magnitude and distance, and the relationship between the frequency of occurrence of earthquakes and the frequency of occurrence of ground motions at a site The PSHA can be described as a procedure involving four steps (Reiter, 1990):

Chapter 1 Introduction

1 Identification and characterization of earthquake sources It is like the first step of DSHA, except that the probability distribution of potential rupture locations within the source must be characterized

2 Characterization of the seismicity or temporal distribution of earthquake recurrence In this step, a recurrence relationship that represents the average rate at which an earthquake of some size will be exceeded is used to characterize the seismicity of each source zone

3 The ground motion produced at the site by earthquakes of any possible size occurring at any possible point in each source zone must be determined with the use of predictive relationship

4 The uncertainties in earthquake size, earthquake location and ground motion parameter prediction are combined to obtain the probability that the ground motion parameter will be exceeded during a particular time period

In the PSHA, the “logic-tree” formulation is often used to incorporate the effects

of both aleatory and epistemic uncertainties (Coppersmith and Youngs, 1986; SSHAC, 1997; Savy et al., 2002) The logic-tree methodology considers a large number of different probabilistic models and model parameters, and computes the hazard for all the combinations of parameter values defined by the end branches of the logic-tree Each of the input parameters is assigned an appropriate weight to define a discrete

probability density function for the frequency of exceeding a value of a strong motion

parameter One can then obtain the various statistical estimates of the frequency of exceeding

1.2.1.3 Disaggregation

It has become common to display the relative contributions to the hazard by

different random components, specifically, the magnitude, M, the source-to-site

Chapter 1 Introduction

distance, R andε , a measure of the deviation of the ground motion from the predicted value The above results, which are obtained separately for each fault and successively combined for all the faults in the region, are called the disaggregation of the PSHA The disaggregation has two major objectives: get the contributions to a fixed hazard level in terms of fundamental quantities and provide seismological parameters describing the earthquake that contribute most to a fixed hazard value The first application of disaggregation was introduced by McGuire and Shedlock (1981) and the relevance of disaggregation was pointed out by the National Research Council (NRC) (1988) and recognized lately by U.S Regulatory Commission (NRC) and the U.S Department of Energy (DOE) McGuire (1995) published a disaggregation method that finds an earthquake representative of the disaggregated uniform hazard-response spectrum The calculation involves:

1 execution of the PSHA analysis in terms of magnitude, distance and ε for each ground motion prediction equation considered

2 analysis of the contributions of each seismic source to check which source dominates the hazard

3 selection of the representative magnitude-distance combination

4 modification of ε until the computed spectral ordinates reproduce the values of the target uniform hazard spectrum

The disaggregation describes the conditional probability distribution of M, R and

ε of an event for which spectral acceleration exceeds a specified level at the site In probability theory, the different representations of the disaggregated hazard are called,

respectively, the marginal probability mass function (PMF) of M, and the joint PMF

of M-R and M-R-ε , based on condition that the spectral acceleration is greater than a

specified level at the site

Chapter 1 Introduction

The disaggregation of the hazard in terms of the joint M-R- ε distribution was

introduced and recommended by Stepp et al (1993), Chapman (1995), McGuire (1995), Abrahamson (1996), Kramer (1996), Abrahamson and Silva (1997) Recently, Spudich (1997), Harmsen and Frankel (2001) used four dimensions: latitude,

longitude, M, andε for the joint probability distribution This disaggregation scheme permits the display of the hazard on a typical map of the faults surrounding the site, allowing an immediate identification of the locations on the faults dominating the hazard Practically speaking, this formulation, along with the knowledge of the most likely magnitude, may be very helpful in establishing the specific earthquakes that present the greatest hazard to the site

1.2.2 Capacity Assessment

It is important to answer the question how to determine the capacity of the offshore platform by means of advanced analyses Usually, in a design situation the approach is to obtain a characteristic capacity higher than the characteristic environmental loads with a return period of typically 100 years multiplied by some partial safety factor for loads and resistance Using conventional design procedures for ultimate limit state design (ULS), the characteristic capacity is normally taken as first yield or first component buckling decided on the basis of linear analysis Obviously this method is conservative Figure 1.1 (Skallerud and Amdahl, 2002) illustrates a typical offshore steel platform’s deformation curve with environmental loads The structural response is typically such that after some plasticity has developed, a critical compression member becomes unstable Although the load at member failure could be said to be the capacity of the structure, the structure often regains resistance because the reduction in load-carrying capacity is compensated by load redistribution to still intact members When more members fail by buckling or

Chapter 1 Introduction

yield, the structure at last reaches a limit load With the increasing interest in taking advantage of the system strength beyond first component failure to resist extreme loading events, it is important to understand how the various parameters (number of legs, bracing configuration, numer of bays, and so on) are critical to strength and how variations in these parameters may affect strength In order to correctly assess the ultimate strength after the first component failure and post-ultimate behavior, static pushover analysis (Kallaby and Millman, 1975) is used to identify the reserve and residual strength of the platform Usually, the reserve strength is referred to pre-ultimate performance measures and the residual strength is referred to post-ultimate performance measures However, there is no consensus regarding which performance measures should be used in deciding the above parameters Indeed, many performance measures are used in the industry

Figure 1.1 Global load versus global displacement for an offshore platform

1.2.2.1 Reserve strength

Reserve strength exists at the component level to allow for uncertainties in both the resistance of the component and the loading to which it is subjected Based on

Reserve strength is more commonly defined as the ability of a structure to sustain loads in excess of the design value The Reserve Strength Ratio (RSR) (Tius and Banon, 1988) may be defined as:

RSR=

LoadDesign

ResistancePlatform

Chapter 1 Introduction

terms of the unfactored global environmental load to cause collapse of structure

1.2.2.2 Residual strength

The concept of residual strength is particularly important in assessing the capacity

of a structure which has been damaged LLoyd and Clawson (1984) defined residual strength in terms of a Residual Resistance Factor (RIF) given by:

RIF=

structureundamaged

ofstrength ultimate

structuredamaged

ofstrength ultimate

ofstrength ultimate

structureintact

ofstrength residual

at absorbedenergy

statedamaged

at absorbedenergy

(1.4)

In equation (1.2), when comparing different structural configurations a clear definition of the damaged state should be used (eg Structural load corresponding to twice the displacement at ultimate load) Figures 1.2 and 1.3 (Skallerud and Amdahl, 2002)illustrate the definition of reserve and residual strength for a structure subjected

to increasing applied loads The design load is the load at which the most highly utilized component reaches its maximum allowable stress or at which the summation

of loads multiplied by the relevant load factors reaches the strength multiplied by the resistance factor Therefore the design load can be taken as the load at first component failure (point C in Figure 1.2) divided by the safety factor The ultimate capacity of the structure is defined as the maximum load that can be sustained by the platform (point A in figure 1.2) The RSR ratio is taken as the ultimate capacity of the structure

Chapter 1 Introduction

Ultimate capacity of intact structureFirst component failure

Structure design load

Residual capacity of damaged structure (loading from an intact state)

Reserve capacity of a damaged structure (loading from a damaged state)

Residual capacity of damaged structure (loading from an intact state)

Ultimate capacity of intact StructureLoading from an intact state

Loading from a damaged state

divided by the design load When the load is removed, the damaged platform will return to a new ‘deformed’ position (point F in Figure 1.3) rather than to its original undeformed configuration (point E in Figure 1.3) Under a new loading condition, the damaged platform will deform in a more flexible manner The residual strength is taken as the ultimate capacity of the jacket, when subjected to loading from a damaged state (point B’ in Figure 1.3)

Figure 1.2 Definition of reserve and residual strength – loading from

an intact state (Skallerud, 2002)

Figure 1.3 Definition of reserve and residual strength – loading from

a damage state (Skallerud, 2002)

Chapter 1 Introduction

1.2.3 Seismic analysis on offshore steel platform

Severe seismic events occur rarely and thus, a dual approach has become widely accepted for offshore design The structure is designed for two levels of earthquake load (see Table 1.1): (1) Strength level earthquake, which is 100-200 year event, and stresses should not exceed the yield stress; (2) Ductility level earthquake, which has a 500-5,000 year return period, and the calculated stresses may exceed yield stress but the structure must not collapse These requirements are summarized in Table 1.2 The strength requirement ensures that the structure can sustain a reasonable probable earthquake without major damage It is an elastic requirement: the structure withstands the earthquake by absorbing its energy as strain energy in the members The ductility requirement ensures that the structure can survive an extreme earthquake without collapse, even though local damage may occur It is an inelastic requirement, and the structure withstands the earthquake by absorbing energy through elastic deformation and dissipating energy through plastic deformation

Table 1.1 Comparison of earthquake and wave loads (North Sea)

Occurrences (per year) Return period (years)

Strength level earthquake 0.005-0.01 100-200

Ductility level earthquake 0.002-0.0002 500-5,000

Table 1.2 Comparison of strength and ductility level analyses for offshore design

Strength level earthquake Ductility level earthquake Typical return period 100-200 years 500-5,000years

Performance criterion Strength (stresses less than yield) Survival (structure should not collapse) Typical peak horizontal ground

Structural behavior Principally elastic Considerable inelastic Acceptable damage Minor (non-structural elements only) Considerable (but overall safety should not be impaired)

Chapter 1 Introduction

The seismic analysis methods fall into three main groups of ascending sophistication: 1 Simple methods: static coefficient method This can often be carried out by hand calculation It provides an order of magnitude estimate of the response and is usually conservative; 2 Linear methods: response spectrum analysis and time domain analysis (normal modes) These are the main methods for dynamic seismic analysis The response spectrum method is the most common owing to its simplicity and the ability to easily verify results by manual checks Time history analysis will normally be used only when the response spectrum analysis cannot give the required results For example, relative deflections of points responding primarily in different modes would be best estimated using time history analysis; 3 Non-linear methods: static pushover and time domain analyses (non-linear) These analyses are usually complex A non-linear time history analysis may be used when it has not been possible to demonstrate structural adequacy using linear methods, or for particularly important structures Table 1.3 gives a summary and comparison of the methods of

analysis

Table 1.3 Comparison methods of seismic analysis

Design data Analysis method Models ductility

Non-linear

methods

Non-linear time domain Time history Dynamic Yes

Chapter 1 Introduction

The static coefficient method is conservative and is a very useful method for carrying out an initial analysis where seismic loading is not expected to dominate the design

The response spectrum method is the most common method for seismic analysis

of multi-degree of freedom structures It is recommended for detailed analysis of response to strength level earthquakes and may be used for response to ductility level earthquakes where the seismic loading does not dominate the design The response spectrum method only gives the peak response in each mode The peak responses in all the modes are unlikely to occur simultaneously and empirical modal combination rules are used to calculate the likely total response Similarly, the modal responses in the three axis directions are also unlikely to occur simultaneously and the response due to all directions together is obtained using similar rules called spatial combination rules (Brebbia and Walker, 1979) If the structure is loaded in more than one coordinate direction, the spatial responses must be combined The usual approach is to calculate the individual response to each earthquake direction acting separately and then to obtain the combined effect of the independent direction responses using a suitable empirical rule

The linear time domain analysis method involves calculating the time history response of each mode as a single degree of freedom system (by transforming to a normal mode coordinate system) The total response is then obtained by summing the contribution from individual modes The response in each mode is known at all points

in time and conventional linear superposition rules apply for combination of the modal responses

In non-linear methods, static pushover analysis is an alternative and possibly simpler method of assessing the ultimate capacity without recourse to non-linear

Chapter 1 Introduction

dynamic analysis It is carried out by imposing static lateral loads on the structure which are successively increased to induce yield or brace buckling until collapse occurs or some performance criterion is met Krawinkler and Seneviratna (1997) found that the pushover analysis could provide a good estimate of global and local inelastic deformation demand The technique would detect the weakness in the structure such as the formation failure mechanisms, excessive deformation demands, strength and stiffness irregularity, overloads on structural elements and connections, and global instability of the structural system Compared with full nonlinear dynamic analysis, there is a great saving in time when performing the pushover analysis A pushover analysis gives a valuable indication of the general post-yield behavior However, it cannot consider cyclic effects, nor is it a dynamic analysis

When the seismic risk and consequence of failure are high, the use of a more detailed non-linear time domain analysis should be considered During severe shaking serval types of non-linear behavior may occur: plastic hinge formation; brace buckling; P−Δ effects; member strength decay under cyclic load; soil pore water pressure generation under cyclic load and soil strength degradation under cyclic loading A realistic nonlinear dynamic time domain response analysis of a fixed offshore structure was carried out and compared with three simplified methods for determining ultimate earthquake resistance by Gates et al.(1977): Ductility Modified Response Spectrum (Newmark and Hall, 1973); Ductility Evaluation under Equivalent Static Loads (Kallaby, 1975)and Reserve Energy Technique (Blume, 1962) Through comparison with three simple methods, it was found that the non-linear time domain analysis gives far more information about the behavior of individual members as well as the overall structure Degradation of members under numerous cycles of inelastic excursion is not accounted for in any of the simplified

Chapter 1 Introduction

methods Concentration of damage due to force redistribution is also overlooked in all the simplified methods Furthermore, the nonlinear time domain analysis provides valuable insights into different failure modes; equipment support design motion, and levels of potential damage

From the above review, it can be seen that the pushover analysis is a simple and quick method to obtain the ultimate strength capacity of an offshore structure under the earthquake load However, if wave load and earthquake load are considered simultaneously then one has to resort to nonlinear time domain analysis

1.2.4 Retrofit of steel structure using fiber reinforced polymers (FRP)

If the seismic adequacy of steel platform is insufficient, seismic retrofitting may

be required Seismic retrofitting refers to wise modification of the structural properties

of an existing structure, in order to enhance its strength and ductility for future earthquake (Aoyama and Yamamoto 1984)

For steel structures, current methods of retrofitting typically utilize steel plates that are bolted or welded to the structure However, constructability and durability drawbacks are associated with this method Steel plates require heavy lifting equipment and can add considerably more dead load to the structure, which reduces their effectiveness of strengthening The added steel plates are also susceptible to corrosion, which leads to an increase in future maintenance costs In many cases, welding is not a desired solution due to fatigue problems associated with weld defects (Kulak and Grondin 2002) On the other hand, mechanical details such as bolted connections, which have better fatigue life, are time consuming and costly

The need for adopting durable materials and cost-effective retrofit techniques is evident One of the possible solutions is to use high performance, nonmetallic

Chapter 1 Introduction

materials such as fiber reinforced polymers (FRP) The superior mechanical and physical properties of FRP materials make them quite promising for repair and strengthening of steel structures Bonding of FRP materials to metallic structures was first used in mechanical engineering applications Carbon fiber reinforced polymer (CFRP) laminates have been successfully used to repair damaged aluminum and steel aircraft structures (Armstrong et al, 1995) Bonding of composite laminates was also shown to have many advantages for marine structures (Allan et al, 1988; Hashim, 1999) For civil engineering structures, previous work conducted on the strengthening

of metallic structures using CFRP has been focused in three main areas: strengthening

of iron or unweldable steel girders, rehabilitation of corroded steel girders, and repair

of fatigue damaged riveted connections Due to the light weight of FRP composite materials, it is expected that they could be installed in less time than by strengthening with the equivalent number of steel plates The second factor that favors composites, especially CFRP, is its higher tensile strength in comparison to the yield strength of steel, provided that adequate means of bonding are introduced

Gillespie et al (1996), Bassetti et al (1998), Buyukozturk and Gunes, (2003), Tavakkolizadeh and Saadatmanesh, (2003) and Mosallam et al (1998) conducted fatigue experiments on girders, plates, beams and frame connections using CFRP to resist the cyclic load The results show no evidence of CFRP plate debonding The crack growth rate was drastically decreased and the fatigue life was increased by a factor up to twenty The CFRP repair provided the higher ductility with an increase of more than 1.25 times the ductility of the fully welded control specimens

Toutanji, and Dempsey (2001) performed a theoretical study on the effects of winding three different types of FRP sheets, including glass, aramid, and carbon, around damaged steel pipelines (see Figure 1.4) The circumferential pretensions

Chapter 1 Introduction

applied during the winding are converted into radial stress A mathematical model was developed to predict the stress levels induced due to the effects of soil loads, traffic loads, and the internal pressure The results of the analysis showed that carbon fiber sheets provide better performance than glass or aramid in improving the internal pressure capacity of pipes at ultimate stress

Figure 1.4 Schematic of the repairing (a) and section of the wound pipe (b)

Vatovec et al (2002) tested rectangular steel tubes retrofitted with different configurations of 50 mm x 1.2 mm CFRP strips, attached to the tension and compression flanges using simple beam tests To avoid local buckling of the tubes’ upper flange, the middle half of the specimens’ length was filled with concrete Test results showed that the ultimate moment capacity was increased from 6 percent for the tube reinforced with one strip attached to the compression flange, to 26 percent for the specimen that had two strips attached to the tension flange and one strip attached

to the compression flange The governing failure mode of all specimens was delamination of the CFRP strips on the compression flange, followed by the delamination of the strips on the tension flange The CFRP strips on the compression

to strengthen VHS tubes under axial tension loading

Shaat and Fam (2006) conducted axial compression tests on short and long square hollow structural steel columns (HSS) retrofitted using carbon fibre reinforced polymers Twenty-seven short-column and five long-column HSS specimens were tested (see Figure 1.5) The effect of CFRP sheet orientation in longitudinal and transverse directions was studied for short columns For long columns, CFRP sheets were oriented in the longitudinal direction only The results showed that a maximum strength gain of 18% was achieved for short columns with two transverse CFRP layers For long columns, the maximum strength gain of 23% was achieved with three longitudinal CFRP layers applied on four sides In all CFRP-strengthened long columns, lateral deflections were reduced after retrofitting

Chapter 1 Introduction

Figure 1.5 FRP retrofit configurations: (a) short columns (b) long columns

A nonlinear model (Shaat and Fam, 2007) based on the concepts of equilibrium

and strain compatibility was developed to predict the behavior of concertrically loaded cold-formed square hollow structural section (HSS) slender columns retrofitted using externally bonded CFRP sheets The model predicted the load versus axial and lateral displacements, and accounted for plasticity of steel, the built-in through-thickness residual stresses, geometric nonlinearity, initial out-of-straightness imperfection, and the contribution of CFRP sheets The model was verified using experimental results and showed good agreement: the percentage increase in strength ranged between 11 and 39% Axial stiffness is also increased by up to 46%, regardless

of the value of out-of straightness; the effectiveness of CFRP retrofitting increases as the values of out-of-straightness of the column increases

In order to investigate the application of CFRP composites to tubular structures under a marine environment, Michael et al, (2007) conducted four point bending test employing a 2.2m span length with seven 2.4m long steel circular tubes One tube was used as a control specimen, whereas the remaining six were wrapped with CFRP

Chapter 1 Introduction

materials Two wrapped specimens were prepared under standard curing conditions, whereas the other four were wrapped underwater The test results show that the ultimate bending strength, flexural stiffness and rotation capacity of the wrapped beams, relative to the reference beam were increased for both performance parameters (underwater and air) However the composite members wrapped and cured underwater were not able to attain the flexural capacity of those cured in air For the tubes wrapped and cured in air, the ultimate strength increased by 16% and 27%, and the flexural stiffness by 7% and 18% When wrapping and curing were performed under seawater, the ultimate strength increased in the range of 8–21% and the flexural stiffness increased in the range of 3 to 10% No serious debonding problems were found in any of the specimens, suggesting that the fibers bonded adequately regardless of the application and curing conditions involved Therefore, using CFRP for retrofitting offshore structural members is a viable procedure, but to date no work has been reported on what extent the buckling strength of tubular members with circular cross section would be enhanced when retrofitted by CFRP

1.3 Objectives and Scope

The main objective of the research work reported herein is to evaluate the seismic vulnerability of offshore steel platforms in Northern Borneo and Northeastern Kalimantan, designed primarily for wave loads, when they are subjected to far field effects of earthquakes and to propose retrofitting schemes for platforms which are vulnerable More specifically, the study sought to find:

1 the seismic hazard curve in Northern Borneo

2 the performance of fixed steel platforms at this site and at another nearby site at Northeastern Kalimantan, designed for wave loads, under simultaneous action of

Chapter 1 Introduction

wave load and earthquake load

3 the performance of platform retrofitted with high strength grout and CFRP under simultaneous action of wave load and earthquake loads

In order to achieve the above objectives, a seismic hazard analysis based on the PSHA in Northern Borneo was first conducted The worst earthquake scenario (accelerograms at bedrock) in Northern Borneo and the amplified accelerograms at the surface of the selected site are obtained The dynamic analyses of 40m and 80m deep platforms with different bracing configurations at Northern Borneo and Northeastern Kalimantan under the wave load and earthquake load were then carried out Based on the results obtained, two retrofitting techniques were applied on the platforms: viz retrofitting using high strength grout and CFRP In order to investigate the CFRP effects on offshore platform, long circular steel columns (89mm X 4mm) were retrofitted with CFRP sheets in various configurations and tested to study the effects of CFRP retrofit on overall buckling behavior Using the experimental results,

a nonlinear model based on the concepts of equilibrium and strain compatibility was developed and a new approach of modeling steel-CFRP beam element in USFOS is adopted in evaluating the performance of the retrofitted platform

1.4 Organization of the thesis

This thesis has seven chapters with Chapter 1 devoted to background information and literature review In Chapter 2, the background material on seismology is presented and seismicity of Northern Borneo is described Seismic hazards analysis for the platform site was carried out and seismic hazard curves were obtained Chapter

3 describes basic theory about idealized structural unit method adopted in USFOS The design of fixed steel platforms of different configuration in different water depths

Chapter 1 Introduction

is presented Static pushover analysis is then used to evaluate the ultimate capacity of the fixed steel platforms Chapter 4 introduces the procedure to obtain accelerograms

at bedrock for the worst earthquake scenario in Northern Borneo obtained in Chapter

2 The amplified accelerograms at the surface of the selected site are then obtained and used in a 3D seismic analysis of offshore platforms at Northern Borneo and Northeastern Kalimantan Chapter 5 describes the experimental study on steel brace retrofitted with CFRP An analytical model is developed to predict the behavior of axially loaded slender braces strengthened with CFRP sheets Chapter 6 describes the modeling of grouted members and introduces a new approach for modeling steel-CFRP beam element in USFOS The effectiveness of retrofitting the tubular members

of the platforms by grout and CFRP is discussed The conclusion of the study and recommendations for the future work are presented in Chapter 7

Chapter 2 Seismic Hazards Evaluation on Northern Borneo

Chapter 2 Seismic Hazards Evaluation on Northern

Borneo

2.1 Introduction

This Chapter evaluates the potential seismic hazard at a site on the northern Borneo due to the seismic activity in the surrounding region Major potential seismic faults or sources in the region which could generate significant bedrock motions at the site of interest were identified Using the past-earthquake data, the recurrence rate of various magnitude earthquakes at each identified source was developed to characterize the seismicity of each source Using the attenuation models for ground motion acceleration available in the literature, the peak ground (horizontal) acceleration at the chosen site for different return periods due to all possible sources are estimated using the probabilistic seismic hazard analysis (PSHA)

2.2 Background to seismology and seismicity of Northern Borneo

2.2.1 Background of seismology

The earth is a spherical body measuring 6400km in radius It consists of three layers of different nature: the core, the mantle, and the crust (see Figure 2.1) The core comprises the central part of the earth and is spherical with a radius of 3500km; the mantle envelopes the core and is 2900km in thickness; the crust is the outermost layer

of the earth with thickness varying from 5 to 40km

The tectonic plates are made up of the crust and the upper part of the mantle layer underneath Together the crust and the upper mantle are called the lithosphere and they may extend up to 150km deep The theory of plate tectonics was developed on

Chapter 2 Seismic Hazards Evaluation on Northern Borneo

the hypothesis of sea-floor spreading during the past few decades It mainly explains the movements on the earth’s surface and the locations of earthquakes and volcanoes

As shown in Figure 2.2, the rigid lithosphere is divided into six major sized plates (African, American, Antarctic, Australian-Indian, Eurasian and Pacific Plates) and is drifting above the asthenosphere with an imperceptible slow speed

continental-Figure 2.1 Structure of the earth

Plate Boundaries

Three distinct types of plate boundaries shown in Figure 2.3 have been identified with regard to the movement associated with each plate, namely spreading ridge boundaries, subduction zone boundaries and transform fault boundaries The characteristics of the plate boundaries do influence the nature of the earthquakes that occur along them

Spreading ridge boundary: In certain areas the plates move apart from each other at

boundaries known as spreading ridges or spreading rifts Molten rock from the underlying mantle rises to the surface where it cools and becomes part of the spreading plates Figure 2.4 illustrates the formation of Mid-Atlantic Ridge and Atlantic Ocean as the result of American and African Plates moving away from each other

Chapter 2 Seismic Hazards Evaluation on Northern Borneo

Chapter 2 Seismic Hazards Evaluation on Northern Borneo

Figure 2.3 Interrelationships among spreading ridge, subduction zone and

transform fault plate boundaries (Kramer, 1996)

Figure 2.4 Spreading ridge boundary - Magma rises to surface and cools in gap

within and becomes part of the spreading plates

Subduction zone boundary: Opposite to spreading ridge boundary, consumption of

plate material occurs at subduction zone boundary where the relative movement of two plates is converging and toward each other At the point of contact, one plate

American Plate Mid-Atlantic Ridge African Plate

Magma Atlantic Ocean American Plate Spreading Ridge Boundary African Plate

Magma

Chapter 2 Seismic Hazards Evaluation on Northern Borneo

plunges, or subducts beneath the other, as shown in Figure 2.5 Subduction zone boundaries are usually found near the edge of continents because the oceanic crust is generally cold and dense that it sinks under its own weight beneath the lighter continental crust

Figure 2.5 Cascadia subduction zone off the coasts of Washington and Oregon

(After Noson et al., 1988)

Transform fault boundary: Transform faults occur where plates move past each

other without creating new crust or consuming old crust They are often found offsetting spreading ridges as shown in Figure 2.3 While Figure 2.6 illustrates that along the boundary only segment of the fracture zone between the spreading ridges is referred as the transform fault Whereas the inactive portions are considered not producing earthquakes

Figure 2.6 (a) Oblique and (b) plan views of transform fault and adjacent inactive

fracture zones (Kramer, 1996)

Chapter 2 Seismic Hazards Evaluation on Northern Borneo

Faults

In some regions, the plate breaks into smaller plates which are trapped between the larger plates Philippine Plate, for instance, is encircled by major plates like Eurasian, Pacific and Australian Plates Locally, the movement between the two portions of the crust will occur on new or pre-existing offsets in the geologic structure

of the crust known as faults Fault movement is generally categorized into dip slip components (normal and reverse faulting) and strike-slip components (left lateral and right lateral faulting) according to the direction of movement occurring on the fault plane

Dip slip movement: It is the movement that occurs primarily in the vertical direction

as shown in Figure 2.7 In a normal fault, the upper of the two adjacent blocks of rock slips relatively downward In contrast, a reverse fault is a rupture resulted from vertical motion of two adjacent blocks In a reverse fault, the upper of the two adjacent blocks moves relatively upward

Figure 2.7 Dip slip movement: (a) normal faulting and (b) reverse faulting (After

Noson et al., 1988)

Strike-slip movement: It referred as the horizontal motion of one block relative to

another along a fault plane as shown in Figure 2.8 A strike-slip fault occurs when the

Chapter 2 Seismic Hazards Evaluation on Northern Borneo

movement is parallel to the strike Strike-slip faults are usually nearly vertical in fault plane and can produce large movements The faults are further divided into right lateral and left lateral strike-slip faults depending on the relative direction of the block motion

Figure 2.8 Strike-slip fault (left lateral strike-slip faulting) (Noson et al., 1988)

is released If the rock is strong and brittle, the rupture of the rock will release the stored energy explosively, partly in the form of heat and partly in the form of the stress waves that are felt on ground as earthquake There are two major regions of earthquake activity One is the Circum-Pacific Belt which encircles the Pacific Ocean, and includes New Zealand, New Guinea, Philippines, Japan, the Aleutian Islands, Alaska, and the west coasts of North and South America It has been estimated that about 75% of the world’s earthquakes occur in this Circum-Pacific seismic belt And 22% of earthquakes occur in the other region called Alpide Belt or Mediterranean-Asiatic Belt It slices through Europe and Asia, passing under the Atlas, Alpine,

Chapter 2 Seismic Hazards Evaluation on Northern Borneo

Iranian, Himalayan, and Burmese Mountains, the island chains of Indonesia and New Guinea, the Solomon Islands, and the New Hebrides Earthquakes in these two regions are interplate earthquakes pertaining to subduction zone boundaries and transform fault boundaries Then the remaining 3% of the events are intraplate earthquakes that occur within the plate area

To describe the location of an earthquake, it is important to use accepted descriptive terminology Earthquake result from rupture of the rock along a fault, and even though the rupture may involve thousands of square kilometers of fault plane surface, it must begin somewhere The point at which rupture begins and the first seismic waves originate is called the focus (or hypocenter) of earthquake (Figure 2.9) Although fault rupture can extend to the ground surface, the focus is located at some focal depth (or hypocentral depth) beneath the ground surface The point on the ground surface directly above the focus is called the epicenter The distance on the ground surface between an observer or site and the epicenter is known as the epicentral distance, whereas the distance between the observer or site and the focus is called the focal distance (or hypocentral distance)

Figure 2.9 Notation for description of earthquake location (Kramer, 1996)

Chapter 2 Seismic Hazards Evaluation on Northern Borneo

When an earthquake happens, different types of seismic waves are produced: body

waves and surface waves; Body waves: travel through the interior of the earth and are

of P-waves (primary waves) and S-waves (secondary waves) P-waves are the fastest seismic waves and the first to arrive a site P-waves oscillate the ground back and forth along the direction of wave travel; involve successive compression and rarefaction of the material through which they pass While S-waves involve particle motion from side to side, perpendicular to the direction of wave propagation Surface waves result from the interaction between body waves and the surface of the earth Because of the nature of the interactions required to produce them, surface waves are more prominent at distances farther from the source of the earthquake At distances greater than about twice the thickness of the earth’s crust, surface waves, rather than body waves will produce peak ground motions

Seismic Magnitude

Richter magnitude (M L ): In 1935, Charles Richter first introduced the idea of

earthquake magnitude, which now known as local magnitude as the logarithm (base 10) of the maximum trace amplitude (in μm) recorded on a Wood-Anderson seismometer located 100km from the epicenter of the earthquake This measure is found to be adequate for shallow, small local earthquake

Surface wave magnitude (M S ): Surface wave magnitude (Gutenberg and Richter,

1936) is commonly used to describe shallow (less than about 70km), distant (farther than about 1000km) and moderate to large earthquake The magnitude scale is based

on the amplitude of Rayleigh waves with a period of about 20 seconds It is obtained from

MS = log A + 1.66 log Δ + 2.0 (2.1)

Chapter 2 Seismic Hazards Evaluation on Northern Borneo

where A is the maximum ground displacement in μm and Δ is epicentral distance

of the seismometer in degrees (360° corresponding to the circumference of the earth, which is equal to 2π×6400km)

Body wave magnitude (m b ): Body wave magnitude (Gutenberg, 1945) is introduced

for deep-focus earthquake because surface waves are often too small to permit reliable evaluation of MS The magnitude scale is based on the amplitude of the first few cycles of p-waves which are not strongly influenced by the focal depth (Bolts, 1989)

It can be expressed as

mb = log A – log T + 0.01 Δ + 5.9 (2.2) where A is the p-wave amplitude in μm and T is the period of the p-wave, which

is usually about 1 second

Japan magnitude (M JMA ): In Japan, the Japanese Meteorological Agency uses

long-period waves to determine a local magnitude scale for Japanese earthquakes MJMA is defined by the intensity of the JMA scale observed or calculated at an epicentral distance of 100km

Figure 2.10 Saturation of various magnitude scales (Idriss, 1985)

Chapter 2 Seismic Hazards Evaluation on Northern Borneo

As shown in Figure 2.10, the abovementioned magnitude scales (ML, MS, mb,

MJMA) saturate for higher magnitudes because their respective amplitudes tend to reach limiting values, thus these magnitude scales may not reflect accurately the size

of very large earthquakes The moment magnitude (Hanks and Kanamori, 1979), which is not obtained from ground motion characteristics and consequently does not saturate, is able to describe the size of any earthquake

Moment magnitude (M w ): It is based on the seismic moment, which is a direct

measure of the factors that produce rupture along the fault It is given by

Mw = 10.7

5.1

log

−

o M

where M is the seismic moment in dyne-cm o

From Figure 2.10, it is known that in the range Mw<5, Mw = ML = MS = mb =

MJMA, in the range Mw<7.5, Mw = MS = MJMA, in the range Mw>7.5, Mw > MS >

MJMA> ML > mb Bolt (1989) suggested that ML or mb be used for shallow earthquakes of magnitudes 3 to 7, MS for magnitudes 5 to 7.5, and Mw for magnitudes beyond 7.5

Ground motion

Ground motion during an earthquake is measured by a strong motion accelerograph which records the acceleration of the ground at a particular location Three orthogonal components of the motion, two horizontal and one vertical, are recorded by the instrument The instruments may be located on free field or mounted

on structures The ground motion is influenced by factors such as

- Earthquake magnitude

- Distance from the focus to measurement location

- Condition at the measurement site

Chapter 2 Seismic Hazards Evaluation on Northern Borneo

- Geological characteristics of the rock along the wave transmission path from source to site

- Source mechanism of the earthquake

The influence of some of these factors is better understood than others and, in fact, the detailed influence of some factors such as the source mechanism and the transmission path of the geology may never be understood except in a general way

2.2.2 Geology of target site

Figure 2.11 Map of Southeast Asia with major tectonic plates and plate boundaries

(chosen sites are indicated by symbol ‘X’)

The site of interest is located at the northern Borneo shown in Figure 2.11, spanning from 111° E to 117° E in longitude and between 4° N and 7° N in latitude, it has always been considered as non-seismically active zone since it is located away from the boundary of tectonic plates and there is no seismic fault running through the

1

2

Chapter 2 Seismic Hazards Evaluation on Northern Borneo

island However it is surrounded by the adjacent Philippine Islands and Sulawesi Island, where the seismic faults are active

Figure 2.12 Tectonic faults running through Philippines Archipelago

At southern Philippines, the Negros Trench runs north to south west of the country from northern Luzon down to Sulu Sea; Cotabato Trench runs off western and southern Mindanao in the Sulawesi Sea; while Sulu Trench is considered active The development of these trenches was attributed to the collision of the Philippine archipelagic landmass with the Eurasian plate in the west and the oceanic Philippine Sea Plate Consequent to movements in these trenches, a major fracture running 1300

km along the length of the archipelago has developed as a response to the oblique

Chapter 2 Seismic Hazards Evaluation on Northern Borneo

compressional forces of the subduction at the trenches The fracture is an active, left lateral strike-slip fault called the Philippine Fault

Figure 2.13 Tectonic faults running through North Sulawesi Island

Sulawesi Island is located at the junction between the Pacific-Philippine, Australian Plates, and the Sunda Block, i.e., the southeastern edge of the Eurasian Plate The North Sulawesi Trench lies slightly above the north arm of the Sulawesi Island The western boundary fault of North Sulawesi Trench system is the Palu-Koro fault which runs from the town of Palu to the Gulf of Bone Earthquake focal studies have indicated left-lateral strike slip motion along the Palu-Koro fault

In order to consider all potential seismic sources capable of generating significant ground motion at the site, the nearby region to the site has been divided into four main zones, where Zone 1 is further divided into four parts to differentiate the effects of different faults running through the region, as shown in Figure 2.14

Zone 1: Subduction zone of Philippines Sea Plate below Eurasian Plate that combines Negros Trench, Sulu Trench, Cotabato Trench and a small part of Philippine Fault

Chapter 2 Seismic Hazards Evaluation on Northern Borneo

Zone 1A [latitude: 4° N-11° N; longitude 117° E-122° E]

-includes the Sulu Trench (about 450km in length) and half-length of the Negros Trench (about 680km in total length) which is nearer to the site The closest point of Sulu Trench is less than 300km from the east edge of the rectangular region of the site while the furthest end of Negros Trench is about 700km away from the site

Zone 1B [latitude: 4° N-11° N; longitude 122° E-125° E]

-includes the Cotabato Trench (about 600km in length), and the nearest distance to the site is about 700km

Zone 1C [latitude: 11° N-13° N; longitude 122° E-125° E]

-includes the small segment of Philippine Fault which is roughly 380km long and

is pretty far away from the site more than 900km in distance

Zone 1D [latitude: 11° N-13° N; longitude 117° E-122° E]

-includes another half-length of Negros Trench The closest point is about 700km from the site

Zone 2 [latitude: 2° N-4° N; longitude 117° E-125° E]: North Sulawesi Trench lies along the subduction of Celebes Sea The shortest distance to the site is about 350km

Zone 3 [latitude: 4° S-2° N; longitude 117° E-122° E]: The closest end of Palu-Koro Fault (of 380km long) is less than 400km away from the site It is a strike slip fault and energy is stored by shear interlock while the two sides of the fault move in opposite direction to each other

Zone 4[latitude: 0° N-13° N; longitude 105° E-117° E]: The northern part of Borneo Island connected to South China Sea is free from any tectonic featured faults lying beneath it This zone is taken into consideration to capture any intraslab or intraplate earthquakes

Chapter 2 Seismic Hazards Evaluation on Northern Borneo

Figure 2.14 Seismic fault zone classifications

2.2.3 Recurrence relationship

The list of earthquake activities in the identified zones is mainly obtained from the following sources: Earthquake listings held online by National Earthquake Information Center (NEIC), U.S Geological Survey (USGS, Appendix A) of the United States (http://www.usgs.gov/); Indonesia earthquake listing prepared by ASEAN Earthquake Information Center; Philippine historical earthquake listing provided online by Philippine Institute of Volcanology & Seismology (http://www.phivolcs.dost.gov.ph/); Earthquake catalog 1900-1989 by Pacheco and