ABS TECHNICAL PAPERS 2003 ASSESSMENT OF CORROSION RISKS TO AGING SHIPS USING AN EXPERIENCE DATABASE

Transcription

1 Proceedings of OMAE nd International Conference on Offshore Mechanics and Arctic Engineering 8-13 JUNE 2003, CANCUN, MEXICO OMAE ASSESSMENT OF CORROSION RISKS TO AGING SHIPS USING AN EXPERIENCE DATABASE Ge Wang 1, John Spencer, Haihong Sun American Bureau of Shipping Northchase Drive, Houston, TX, USA, : gwang@eagle.org ABSTRACT Damages to ships due to corrosion are very likely, and the likelihood increases with the aging of ships. Risk and reliability approaches are more and more frequently applied in design and maintenance planning. These advanced approaches require reliable data reflecting the structural condition of ships in service. Such data is scarce. This paper presents a database of corrosion wastage. It is based on over 110,00 thickness measurements recently collected from 140 trading tankers. This database is larger than most other corrosion databases in the public domain. Corrosion wastage exhibits a high level of variability. In addition to thickness measurements of individual structural members, this database also has information on hull girder s geometrical properties and strength of ships in service. Corrosion wastage has an influence on the hull girder strength. Statistical interpretations of the database are used to represent corrosion wastage in oil tankers. The severity of corrosion is ranked by three levels: slight, moderate and severe levels corresponding respectively to 50, 75 and 95% cumulative probability on the database. The risks of corrosion wastage to aging ships structural integrity are assessed using the observations of the corrosion wastage database. The investigated risks are loss of local member s strength, loss of global hull girder strength, and shortened inspection intervals. The experience database can be used in many aspects, such as design requirements for corrosion additions and wastage allowance for plate renewal, establishment of limits to hull girder strength of FPSOs, time variant reliability approach and risk based inspection schemes. INTRODUCTION Figure 1 shows the underdeck area of a 22-year-old tanker (ABS 2001). The deck plates and deck longitudinals suffered various degrees of corrosion. In some locations, the web plate of some deck longitudinals was totally wasted away. This caused loss of support of deck plates from deck longitudinals. The unsupported span of the deck plate increased, with a corresponding decrease in buckling strength. In heavy seas, buckling repeatedly occurred under the action of the cyclic wave loads. Plastic deformation accumulated and eventually cracks appeared. Statistics reveal that corrosion is the number one cause for marine casualties in old ships (Harada et al. 2001). Damages to ships due to corrosion are very likely, and the likelihood increases with the aging of ships. The consequences of corrosion wastage can be local or minor, but also can be very serious in some circumstances. Severe corrosion has resulted in deck cracks across almost the entire ship width (ABS 2001), and has even resulted in the loss of ships (JMT 1997). Structures deteriorate over time due to corrosion. This causes variability in structural properties and capability. Traditional engineering and analysis use simplified deterministic approaches to account for this time-variant random process; in most cases some nominal values are predefined for corrosion additions (e.g., Wang et al. 2002). A more rational and direct approach is to model the uncertainties probabilistically. There is a clear trend that engineering analysis and design standards are moving toward reliability-based formats. Assessment of Corrosion Risks to Aging Ships Using an Experience Database 149

2 Figure 1. Heavily corroded under-deck of a 22 year old oil tanker (ABS 2001) Originally, the structural reliability approach was introduced for establishing safety factors. Probabilistic presentations of global and local loads have been developed, and structural failure modes and limit states have been extensively studied. As a result, the reliability approach has been refined and applied to some engineering problems (Guedes Soares et al. 1989, Mansour 1997, Wang et al. 1996, Melchers 1999). Recently, there is an increased interest in developing and demonstrating the time variant reliability (TVR) approach to explicitly address the uncertainties due to structural deterioration (e.g., Guedes Soares et al. 1996, Wirsching et al. 1997, Sun & Bai 2001, Ivanov et al. 2003, Qin and Cui 2002, Paik et al. 2003). The TVR approach is more suitable to the assessment of the strength of ships in service and new constructions, and can also be used in maintenance or inspection planning, and development of new designs. The success of these state-of-the-art technologies depends to a large extent on reliable estimates of corrosion wastage of various structural members. There are very few databases of corrosion wastage available in the literature. The Tanker Structure Co-operative Forum guidance (TSCF 1992) is based on thickness Table 1. Main details of the corrosion wastage database and comparisons with other database of oil tankers introduced in the public domain The present database TSCF (1992) Harada et al. (2001) Paik et al. (2003) Ship type Single hull oil tankers Single hull tankers Single hull tankers Single hull tankers Data sources SafeHull Condition Assessment Owner, class Gauging records Gauging reports Vessels >100 Gauging reports 157 Not known 346 Not known Thick. measurements 110,082 Not known > 250,000 33,820 Info. Hull strength Yes, 599 sections No No No Ship size 168 ~ 401 meters > 150, 000 DWT 100 ~ 400 meters Not known Service years 12 ~ 26, 32 years ~ 25 years ~ 23 years 12 ~ 26 years Class ABS, LR, NK, DnV, KR ABS, DnV, LR, NK NK KR, ABS Ship built Mostly 1970 s, some 1980 s 1960s ~ 1980s Not known Not known Ship measured Not known Not known Not known 150 Assessment of Corrosion Risks to Aging Ships Using an Experience Database

3 measurements of 52 oil tankers. Yamamoto and Ikegama (1998) introduced a database of 50 bulk carriers. There was a probabilistic corrosion rate estimation model developed from and calibrated with the measurements of 44 bulk carriers (Paik et al. 1998), and more than 100 oil tankers (Paik et al. 2003). These databases are, however, relatively small in size, and some are not representative of commercial ships of today. Harada et al. (2001) collected a database from 197 oil tankers. This database has been circulated with a working group of the International Association of Classification Societies (IACS), and has not been released to the public. There is a need to develop a sizable database that reflects, as close as possible, the structural conditions of ships in service. This paper presents a database of corrosion wastage of oil tankers. It is aimed to provide a more realistic picture of corrosion wastage of oil tankers. This newly developed database has been analyzed, and general trends of corrosion wastage, which change over the service life, have been studied. Discussion is given to some safety issues of tankers from the standpoints of both local strength of individual structural members and global hull girder strength. It is expected that the database will enhance and update the knowledge about corrosion wastage in oil tankers, and also provide more realistic estimates of corrosion for structural members that can form a reliable basis for a quantitative assessment of structural integrity of ships in service. A NEW CORROSION WASTAGE DATABASE A new corrosion wastage database was built recently at ABS. It is an integral part of the efforts to develop reliability based design standards. Database particulars The database has more than 110,000 corrosion wastage measurements of various structural members, which are collected from 157 gauging reports of 140 tankers. Most of the ships are still in service. Some have been or will be converted to FPSOs. The ships are classed with five major classification societies. The ship length ranges from 170 m to 400 m. They were 12 to 33 years old when thickness measurements were taken. Table 1 summarizes some of the main details of this database. Table 1 also includes corrosion databases on oil tankers that have been introduced in the literature. Obviously, there are only a limited number of databases on corrosion wastage. The present database is one of the largest of its kind, second only to Harada et al. (2001). It provides up-to-date information about corrosion in oil tankers. The database also includes information about the hull girder strength, as this is calculated and used for assessing the ship s structural adequacy for its intended service. This database is the only one that has Frequency 20% 15% 10% 5% 0% 12 Distribution of Vessel Age at the Time of Gauging (158 Records) information about hull girder sectional properties for ships in service (see Table 1). Figures 2 and 3 are ship age and length profiles of the sampled ships. These ships are representative of modern single hull oil tankers. Data sources The data comes from the ABS SafeHull Condition Assessment Program (CAP). CAP is a service separate from and a supplement to classification (Horn et al. 1994). The CAP offers an evaluation of ship structure Vessel Age at the Time of Gauging 32 Figure 2. Profile of ship age at the time of thickness measurement (157 gauging reports, 140 oil tankers) Frequency 40% 30% 20% 10% 0% Distribution of Ship Length (140 Vessels) Ship Length (m) Figure 3. Profile of ship length (157 gauging reports, 140 oil tankers) 420 Assessment of Corrosion Risks to Aging Ships Using an Experience Database 151e

4 recognizing the effects of corrosion with respect to yielding, buckling and fatigue. Based on extensive surveys, the CAP database provides a wealth of information regarding the structural condition of ships in service. The database reflects the condition of single hull oil tankers in service. For ships in CAP, plate thickness measurements of heavily wasted structural members are recorded and are not excluded from the thickness measurement reports. They are recorded as is, and repair work, if necessary, is recommended after the ship s condition is assessed. Traditional gauging reports for ships in service as required by classification societies, which almost all the available databases are based upon, may not include thickness measurements below the wastage limits. Thickness measurements obtained as part of CAP evaluations may give a more realistic picture of the actual corrosion wastage trends. On the other hand, the vessels assessed in CAP may be in relatively good condition. The ship owner probably believes that his ship can be used for service for a few more years. Substandard ships, though a small percentage of the fleet, may not be found in CAP. In this sense, data gained from CAP may not include the worst cases. Nevertheless, thickness measurement data from ships in CAP are very good records of the condition of ships in service. Corrosion wastage Wastage due to corrosion is calculated as the difference between the as-built thickness and the measured residual thickness. Thickness measurements are relevant to general corrosion, where the plates are assumed to be uniformly wasted. Pitting and grooving are generally not fully reflected in gauging reports. Replaced plates As usual with databases based on gauging reports, the data may include plates that have been replaced. However, such plates occupy only a small percentage of the total. They do not have a prominent influence that would skew the statistical characteristics of the database. During the 3rd or 4th special survey, oil tankers in the range of 150,000 to 300,000 deadweight tons may have to replace up to 380 tons of steel (TSCF 1992). The hull of a 137,000 deadweight ton tanker weighs about 22,000 tons. The steel renewal in the 3rd or 4th special survey accounts for, at the maximum, about 1.7% of the total steel weight. If it is a VLCC, the maximum percentage of replaced steel can be less than 1.0% of the hull s steel weight. The replaced steel plates, if there are some, possibly occupy a very small percentage of the entire population. Nevertheless, thickness measurements corresponding to probably replaced plates have been removed from the database. Plates with very small wastage, say less than 0.01 mm, are screened out; they are probably plates renewed after the ship s delivery. SOME OBSERVED TRENDS The wastage measurements are categorized according to location (structural member) and usage space. The locations are deck, side, bottom and longitudinal bulkheads. Both plates, and web and flanges of longitudinals are investigated. In line with classification rules for new construction designs, two usage spaces are considered, i.e., cargo tanks and ballast tanks. The database provides a lot of information about the trends of corrosion wastage in oil tankers. Table 2 and Figures 4 and 5 are snapshots of the database. Table 2 summarizes the mean values, standard deviations and maximum values of corrosion wastage measurements of various structural members for 20 years of service. Figure 4 shows the wastage measurements for deck plates in cargo tanks in millimeters for ships of 12 to 32 years old. One diamond mark represents one measurement. Figure 5 shows the loss of hull girder section modulus at the deck over the past year. One diamond mark represents one section of a ship. Usually, a ship has about three girth belts (transverse sections) gauged in one thickness measurement survey. Corrosion wastage exhibits a high level of variability The maximum corrosion wastage is much higher than the average. For example, for 20 years old ships (Table 2 and Fig. 4), the maximum observed wastage in deck plate in cargo tanks is 8.70 mm, while the average wastage is 1.1 mm. Corrosion wastage measurements spread over wide ranges. Some structural members exhibit standard deviations higher than the averages, e.g., deck plates, bottom shell plates, and bottom longitudinal flanges in cargo tanks (Table 2 and Fig.4). The maximum corrosion wastage seems to be higher in cargo tanks than in ballast tanks (Table 2). The average corrosion wastage does not seem to depend on the usage spaces (cargo or ballast tank). See Table Assessment of Corrosion Risks to Aging Ships Using an Experience Database

5 Table 2 Corrosion wastage of various structural members at 20 years old (unit: mm) Structure Tank Mean value Deviation Maximum 50 percentile 75 percentile 95 percentile Dk pl Cargo Ballast Dk long web Cargo Ballast Dk long fl Cargo Ballast Side shell Cargo Ballast Side long web Cargo Ballast Side long fl Cargo Ballast Btm shell Cargo Ballast Btm long web Cargo Ballast Btm long fl Cargo Ballast Long bhd pl Btw cargo Others Bhd long web Cargo Ballast Bhd long fl Cargo Ballast Abbreviations: btw between, bhd bulkhead, dk deck, fl flange, long longitudinal, pl plate One factor that may have influenced the data is whether or not the space has been coated. Ballast tanks generally have a corrosion protection system, whereas cargo tanks may not. The presence or absence of a coating is not noted in the database. With the aging of ships, more steel is wasted. The average corrosion wastage exhibits an increasing trend with the passage of time (Fig. 4). With the aging of ships, the spread of wastage measurements becomes more prominent. The standard deviations tend to increase with the passage of time. Figure 4 shows that corrosion wastage does not always increase with the ship s age. This observation is not new, and has been demonstrated in previous studies. Most oil tankers are scraped at about years old and older (Harada et al. 2001). This database does not include scraped ships, nor do any other databases. Therefore, the worst conditions of ships much older than 23 years are not covered in the database. There are fluctuations in the average values and standard deviations of corrosion wastage (Fig. 4). The measurements come from a fleet of ships, and do not represent a trend of a single plate in a specific ship. The variability may be attributed to measurements not being taken from a single ship, or at the same location. The different maintenance of ships may also contribute. Assessment of Corrosion Risks to Aging Ships Using an Experience Database 153e

6 Corrosion wastage has an influence on the hull girder strength The information about the hull girder sectional properties is extracted from the calculation results of the ABS SafeHull Condition Assessment program. Ships in CAP are evaluated for their local and global strength. Figure 5 shows the reduction of section modulus to the deck as a function of the vessel age. The mean values, 75 and 95 percentile curves are also shown. The maximum SM reduction is close to 16% of the as-built condition, which is for ships about 20 years old. This may be the minimum strength that the present design standards expect of a tanker. The majority of ship sections, say at 95% probability for a given age, have a maximum reduction of about 10%. This is in line with the IACS UR S7 requirement that ships in service be at least 90% of the section modulus required for new construction. The average SM reduction increases with ship s age. The lines of 75 and 95% percentile also increase with ship s age. The drop at 24 years old is because most tankers are scraped at 22 to 23 years, and the corrosion wastage database does not include scraped ships. As expected, as ships become older, the hull girder section modulus reduces further. SLIGHT, MODERATE AND SEVERE LEVELS OF CORROSION WASTAGE Because of the shown high variability, it appears that the mean values and standard deviations are not sufficient for presenting corrosion wastage. Statistical interpretations of a large volume of records, such as the present database, give more information, and should be used to provide a more realistic picture of corrosion wastage in commercial ships. Despite continuous efforts on corrosion protection, the mechanisms of corrosion in tankers are still not fully understood. The inherent complexity casts questions about the attempts to develop physical models for predicting corrosion wastage, because the physical models (e.g., Melchers 2001, Gardiner and Melchers 2001) are usually limited to some well-defined conditions, while it is recognized that there are a vast variety of possible situations and causal factors. There is a need to develop a more reliable, yet easy to use, scheme to quantitatively describe corrosion wastage in commercial ships. Cumulative probability One way to present this highly variable problem is to assign cumulative probability values, and derive corrosion wastage from the database accordingly. The values of corrosion wastage as thus determined would measure the extent of structural deterioration in a probabilistic manner. Corrosion Wastage (mm) Deck Plates in Cargo Tanks M easured Average 95% 75% 50% Age (Year) Figure 4. Corrosion wastage of deck plate in cargo tanks (4665 thickness readings, 157 gauging reports, 140 oil tankers) Loss of SM (% as-built) 20% 15% 10% 5% 0% Loss of Section Modulus to Deck section average 95% 75% Age (years) Figure 5. Loss of hull girder section modulus to deck over time (599 sections) Table 2 includes values of corrosion wastage corresponding to 50, 75 and 95% cumulative probability at 20 years. 154 Assessment of Corrosion Risks to Aging Ships Using an Experience Database

7 Figure 4 also includes wastage of deck plates in cargo tanks for 50, 75 and 95% cumulative probability values. The lines of 50, 75 and 95% percentile demonstrate an increasing trend over time. They fluctuate also because of the sampling, etc. Table 3. Slight, moderate and severe corrosion levels based on the cumulative probability of corrosion wastage in the database Levels Slight Moderate Severe Cumulative probability 50% 75% 95% For deck plates in cargo tanks after 20 years of service, a 1.10 mm corrosion wastage corresponds to a 75% cumulative probability. This means that the cumulative probability of wastage measurements less than 1.10 mm is 75%, or, the wastage measurements below 1.10 mm occupies 75% of all deck plate measurements taken at 20 years. Slight, moderate and severe levels of corrosion It seems reasonable to categorize the corrosion wastage based on the cumulative probability as follows: Slight corresponds to a 50% percentile. Moderate corresponds to a 75% percentile. Severe corresponds to a 95% percentile. The corrosion wastage approximately doubles when the cumulative probability is changed from 50% to 75%, and roughly triples at 95%. Most of the structural members have about 0.5 mm wastage for a 50% probability, approximately 1.0 mm for a 75% probability, and roughly 1.5 mm for a 95% probability. Exceptions are deck plates and bottom shell plates, which have much higher corrosion wastage than other structural members. This ranking, summarized in Table 3, provides a convenient and practical vehicle for presenting a highly variable problem CORROSION RISK TO AGING SHIPS Corrosion causes change in the thickness of structures. With the aging of a ship, more and more steel is wasted away, increasing the risks to the ship s safety. The majority of marine casualties involving ships older than about 22 years is found to be due to corrosion wastage. Two sample ships will be used for following discussions on some aspects of corrosion risks to aging ships. Their details are listed in Table 4. Table 4. Particulars of a single hull and a double hull tanker Ship SHT DHT Ship type Single hull Double hull Construction Conversion to FPSO New build Length (m) Breadth (m) Depth (m) Ship built Section modulus 103.2% required 103.6% required Deck plate (mm) Material HT36 HT32 Long. Sp. (mm) Table 5. Buckling strength of deck plates for different levels of corrosion wastage Ship Corrosion Thick (mm) Buckling/yield SHT DHT As-built Slight Moderate Severe As-built Slight Moderate Severe Table 6. Hull girder section strength for different levels of corrosion wastage Ship SHT DHT As-built 100.0% 100.0% Slight 97.0% 96.7% Moderate 94.5% 94.0% Severe 88.5% 87.3% The differences in corrosion wastage between single hull tankers and double hull tankers are not considered, though such differences are recognized. Assessment of Corrosion Risks to Aging Ships Using an Experience Database 155e

8 Corrosion causes loss of strength of individual structural members. Some recent oil tanker incidents took place when ships were loaded in a sagging condition. Deck plates were under compression, and buckling and ultimate strength were reduced due to wastage, which led to catastrophic failure (ABS 2001). Table 5 shows the loss of buckling strength of deck plates assuming that they are 20 years old and have different levels of corrosion wastage. The plates are compressed at the shorter edges from longitudinal bending of the hull girder. The slight, moderate and severe corrosion levels, corresponding to the 50, 75 and 95% percentiles, are based on Table 2 (for ships 20 years old). They may be regarded as the results of different maintenance practices, though other factors such as coating condition may also play a role. In the case of severe corrosion, the buckling strength of deck plate is reduced by about 7% for the single hull tanker (SHT in Table 4), and by 14% for the double hull tanker (DHT). Combined with the reduced hull girder strength, the deck plates may buckle under heavy seas. Corrosion causes loss of hull girder strength. Hull girder section modulus is a well-accepted parameter measuring the longitudinal bending strength of ships. This is perhaps the single most important design parameter describing hull girder strength. Hull girder section modulus to the deck often determines the bending strength of the entire hull girder. Table 6 shows the loss of hull girder section modulus to deck as a result of different levels of corrosion wastage. When every structural member is severely corroded, the single hull tanker (SHT) has a 11.5% reduction in hull girder strength, and the double hull tanker (DHT) has a 12.3% reduction. It is assumed that every member at the same location (e.g., every strake at deck) has the same level of corrosion. This assumption may not be realistic, but is used here for convenience and demonstration purposes. Figure 5 is a realistic picture of hull girder strength of corroded ships. Severe corrosion requires more frequent inspection or maintenance. Figure 6 is the estimated time-dependent annual reliability index of a stiffened panel. Details of the structural dimensions are in Table 7. This panel is at the bottom of a cargo hold of a single hull tanker 232 meters in length. The three corrosion levels specified in Tables 3 and 2 are assumed. The corresponding corrosion rates obtained from Table 2 are assumed to remain constant beyond 20 years old. Discussions on corrosion rates are detailed in Wang et al. (2003). This bottom panel is acted upon by in-plane compression due to longitudinal bending and lateral loads due to water pressure. The ultimate strength of the panel is calculated and compared with the external loads. Annual Reliability Index Slight Moderate Severe Age (Years) Figure 6. Annual reliability index of a stiffened panel at a tanker s bottom for different corrosion levels Annual Reliability Index Slight 2.2 Moderate Severe Age (Years) Figure 7. Annual reliability index of a stiffened panel at a tanker s deck for different corrosion levels Table 7. Dimension of analyzed stiffened panels (mm) Plate Web Flange b t h w t w b f t f Fig Fig It is assumed that plates are replaced at special surveys when failing the requirements of classification societies. The spikes in Fig. 6 reflect the effects of plate renewal. Details of this time-variant reliability assessment can be found in Sun & Bai (2001) and Sun & Guedes Soares (2003). 156 Assessment of Corrosion Risks to Aging Ships Using an Experience Database

9 The renewal criteria in ABS Steel Vessel Rules were used. Plate components that are wasted by 20% were assumed to be renewed. If corrosion remains slight, inspections at five-year intervals will be sufficient, and no plate renewals are needed for more than 30 years. When experiencing moderate level of corrosion, inspections at five-year intervals seem sufficient for maintaining the reliability index at reasonable level, though plate renewals are expected after 30 years in service. When experiencing severe level of corrosion, inspections at five-year intervals can not prevent the reliability index from becoming too low. The curve of the reliability index declines quickly. Within 5 years, the reliability index decreases from 3.28 to 3.12, and plate renewals are necessary at every special survey. In order to maintain enough margin when severe corrosion is anticipated, inspections should be conducted at intervals shorter than 5 years. Similar conclusions can be drawn from the analyses on a deck panel (Fig. 7) in the cargo hold and on the hull girder (Fig. 8) of the same tanker. Details of structural dimensions are also listed in Table 7. APPLICATIONS OF THE DATABASE The database can be used in some other applications, in addition to those described in the previous section. A sizable database is the key to the development of corrosion wastage allowance in design standards. Classification Societies have set safety standards requiring that structural scantlings of ships be designed with a certain allowance for corrosion wastage. This allowance is often referred to as corrosion addition (TSCF 1992). Ships in service are periodically surveyed and inspected. While deemed necessary according to defined criteria, i.e., the wastage allowances (TSCF 1992), wasted plates are recommended to be replaced. To a large extent, the relevant requirements for corrosion addition and wastage allowance were empirically derived from experience. One of the key issues is that there is very limited data, and a quantitative assessment is nearly impossible. The corrosion wastage database in this paper has extensive data, which makes it possible to quantitatively evaluate corrosion in oil tankers. A more refined approach for developing standards regarding corrosion wastage should be based on thickness measurement data, and use probabilistic interpretations of the data. The approach includes: constructing a database of corrosion wastage measurements, properly assigning the level of confidence for these records, and obtaining the corresponding values from the experience database. For structural design purposes, corrosion additions may be based on a moderate corrosion level, at about the 75% percentile. For renewal criteria, corrosion wastage allowance may be based on a severe corrosion level, at Annual Reliability Index Slight Moderate Severe Age (Years) Figure 8. Annual reliability index of the hull girder strength of an oil tanker for different corrosion levels approximately the 95% percentile. This study is ongoing and will be reported in a future paper. The experience gained in trading tanker designs provides useful information for establishing limits to strength of FPSOs. Because of the limited experience of designing and operating FPSOs, experience gained from trading tankers is often considered. FPSOs are generally designed based on site-specific environments. It is necessary to introduce limits to keep design parameters from going too low. These limits reflect successful experience, not to inadvertently create a re-ordering of the dominant structural failure modes, and to avoid the introduction of new controlling limit states (ABS 2000). It has been recognized that limits to the minimum allowable hull girder strength should be established for FPSOs to take into account the inevitable corrosion risks. Oil tankers have exhibited possible strength reduction of about 10 to 16%, see Figure 5. The same level of strength reduction may also need to be taken into account at the design stage for FPSOs. The database can be incorporated into a time variant reliability approach. One of main advantages in structural reliability analysis is the recognition of the inherent uncertain nature of various random variables. There is a need to estimate the reliability of a structure over its lifetime to take account of inspection and repairs. Time variant reliability explicitly addresses the effects of corrosion wastage on the structural integrity of ships. This is a more refined reliability approach. One of Assessment of Corrosion Risks to Aging Ships Using an Experience Database 157e

10 the keys to the successful application of the time variant reliability approach is the prediction of corrosion wastage of structures over time. In addition to Figs. 6 and 7, Figure 8 illustrates an application of the time variant reliability approach to the hull girder strength of a single hull tanker 232 meters in length. Estimation of corrosion rates is detailed in a separate paper (Wang et al. 2003). Plate renewals are assumed to be conducted at special surveys when the wastage exceeds the limits specified by classification societies. The ultimate strength of the hull girder is calculated using a program based on the Smith s method (Sun and Bai 2001). Hull girder failure is defined as the total bending moment exceeding the maximum hull girder bending capacity, both of which are expressed in probabilistic terms. The database can be incorporated into a risk based inspection planning scheme. One of the major objectives of inspections is to detect defects of any kind, and remedy the situation before the defect develops into an unwanted event, for example, loss of containment or failure of structures. Inspections can possibly be conducted in a smarter way if the likely situations can be predicted in advance, and the associated risks can be properly assessed. Corrosion wastage is the number one causes for marine casualties in old ships. Predictions of corrosion wastage over a ship s life are very important. Risk is often defined as the product of failure consequence and probability of failure. According to the failure consequence and failure type, the lower limit of safety level of a component or structural system can be defined in order to keep the component or structural system free from failure. The likelihood of failure can be determined by statistical studies, analytical solutions, or both. The database can provide the foundation to evaluate the risk due to corrosion damage and help to determine inspection planning. CONCLUSIONS This paper presented a database of corrosion wastage that contains more than 110,000 wastage measurements collected from 140 oil tankers. This database also has information about the hull girder strength of corroded ships. The following conclusions are reached: Corrosion wastage exhibits high variability. Corrosion wastage exhibits an increasing trend with the passage of time. Corrosion wastage has an influence on the hull girder strength. Based on the cumulative probability of measurements in the database, corrosion wastage may be ranked in three levels, slight, moderate and severe. This ranking scheme provides a convenient vehicle to represent a highly variable problem. The risks of corrosion wastage to aging ships structural integrity are discussed. The investigated risks are loss of local member s strength, loss of global hull girder strength, and shortened inspection intervals. The experience database can be used to develop (1) design requirements for corrosion additions and wastage allowance for oil tankers, (2) design limits to the hull girder strength of FPSOs, (3) a time variant reliability approach, and (4) risk based inspection schemes. ACKNOWLEDGMENTS The authors appreciate very much the contributions of Yongjun Chen, Tarek Elsayed and Sara Irwin in building up the database. The authors wish to thank many colleagues for their valuable comments and reviews, especially those from J. Card, D. Diettrich, L. Ivanov, J. Baxter, Y. Shin, P. Rynn and K. Tamura. The authors are indebted to Jo Feuerbacher for editing the manuscript. REFERENCES ABS, 2000, Guide for building and classing floating production installations, American Bureau of Shipping. ABS, 2001, Final report of Investigation into the damage sustained by the M.V. Castor on 30 December 2000, American Bureau of Shipping. Gardiner C.P., Melchers R.E., 2001, Bulk carrier corrosion modeling, International Offshore and Polar Engineering Conference, IV: , Stavanger, Norway. Guedes Soares, C. and Ivanov, L. D., 1989, Time Dependent Reliability of the Primary Ship Structure, Reliability Engineering and System Safety, 26, Guedes Soares, C. and Garbatov, Y., 1996, Reliability of Maintained Ship Hulls Subjected to Corrosion, Journal of Ship Research, 40(3), Harada S., Yamamoto No., Magaino A., Sone H., 2001, Corrosion analysis and determination of corrosion margin, Part 1&2, IACS discussion paper. Horn G.E., Johnson M.D., 1994, Using the ABS SafeHull Approach to Optimize repairs ABS SafeHull Condition Assessment Services, Ship Repair & Marine Maintenance 94, New Orleans, USA. Ivanov, L. Spencer, J., Wang, G., 2003, Probabilistic evaluation of hull structure renewals for aging ships, The Eighth International Marine Design Conference (IMDC), 5-8 May 2003, Athens, Greece. JMT, 1997, Report on the investigation of causes of the casualty of Nakhodka, Japan Ministry of Transport, The committee for the investigation on causes of the casualty of Nakhodka. Mansour A.E., 1997, Assessment of reliability of ship structures, Ship Structure Committee report SSC Assessment of Corrosion Risks to Aging Ships Using an Experience Database

11 Melchers, R E., 1997, Structural reliability analysis and prediction, John Wiley & Sons, New York. Melchers R.E., 2001, Probabilistic models of corrosion for reliability assessment an maintenance planning, 20th International Conference on Offshore Mechanics and Arctic Engineering, Rio de Janeiro, Brazil, June 3-8, Paik J.K., Kim S.K., Lee S.K., 1998, Probabilistic corrosion rate estimation model for longitudinal strength members of bulk carriers, Journal of Ocean Engineering, 10, Paik J.K., Wang G., Thayamballi A.K., Lee JM., 2003, Time-variant risk assessment of aging ships accounting for general / pit corrosion, fatigue cracking and local dent damage, SNAME annual meeting, San Francisco, CA. Qin S., Cui W., 2003, Effect of corrosion models on the time-dependent reliability of steel plated elements, Marine Structures, 16, Sun H.H., Bai Y., 2001, Time variant reliability of FPSO hulls, SNAME annual meeting, Orlando, FL. Sun H.H., Guedes Soares C. 2003, A corrosion model and reliability-based inspection for ship-type FPSO hulls, Journal of Ship Research, submitted. TSCF (Tanker Structure Co-operative Forum), 1992, Condition evaluation and maintenance of tanker structures, Witherby & Co. Ltd, London. TSCF (Tanker Structure Co-operative Forum), 1997, Guidance manual for tanker structures, Witherby & Co. Ltd, London. Wang G., Tang S., Shin Y., 2002, Direct calculation approach and design criteria for wave slamming of an FPSO bow, International Journal of Offshore and Polar Engineering. 12 (4). Wang G., Chen Y.J., Zhang H., Peng H., 2002, Longitudinal strength of ships with accidental damages, Marine Structures, 15, Wang G., Spencer J., Elsayed T., 2003, Estimation of corrosion rates of oil tankers, 22th International Conference on Offshore Mechanics and Arctic Engineering, Cancun, Mexico, 8-13 June Wang, X., Jiao, G., Moan, T., 1996, Analysis of Oil Production Ships Considering Load Combination, Ultimate Strength and Structural Reliability, Trans. SNAME, 104: Wirsching, P.H., et al., 1997, Reliability with Respect to Ultimate Strength of a Corroding Ship Hull, Marine Structures, 10: Yamamoto N., Ikegami K., 1998, A study on the degradation of coating and corrosion of ship s hull based on the probabilistic approach, Journal of Offshore Mechanics and Arctic Engineering, 120, Assessment of Corrosion Risks to Aging Ships Using an Experience Database 159e

12 160 Assessment of Corrosion Risks to Aging Ships Using an Experience Database