ABSTRACT The circumferential welds on steel tubular berthing monopiles, namely pile splices, are potential fatigue hot spots. This paper presents the latest progress in linear elastic fracture mechanics (LEFM)-based fatigue safety assessment for welded splices of steel berthing monopiles. Particularly, the fatigue assessment against vessel impacts is studied at the current stage. Practical methods for determining stress intensity factor (SIF) and stress concentration factor (SCF) at the circumferential welds given bending moments are reviewed and summarized. This paper also proposes a new modification factor for Paris law, as well as a Beta distribution for characterizing the hot-spot stress range caused by vessel impacts, which is often doubly bound by operational water levels. The findings of this paper provides practical information for performing LEFM-based fatigue safety assessment of splice welds of steel berthing monopiles. INTRODUCTION Steel marine piles often work against cyclic and harsh operational and environmental loadings. Due to welding flaws or defects (undercuts, cracks, incomplete penetration, and gas pores), circumferential welds on steel piles, also known as splice welds, transversal welds, or girth butt welds, have been recognized as potential fatigue hot spots, given the frequently repeated loadings. Insufficient fatigue strength of welded pile splices may cause severe structural failures (Dailey et al., 1987a, 1987b; Weidler et al., 1987). This paper studies the fatigue reliability of circumferential welds of steel berthing monopoles. Fatigue loadings for a berthing monopile include berthing impacts, wave effects, current oscillations, etc. In many cases, for example well-sheltered harbours, the berthing impact dominates as a major fatigue loading for splice welds. The berthing impact normally leads to substantial bending moments on a monopile. This paper focuses on the cases dominated by berthing impacts. Typically, hot-spot fatigue strength can be assessed with either S-N curves or linear elastic fracture mechanics (LEFM).

ABSTRACT: Distributions of wave phases were studied. Three different sources of wave records were used for comparison. These are: mechanically generated directional waves with a prescribed JONSWAP target spectrum; wave records measured in the northern part of" Taiwan; and surface fluctuations acquired during wind wave flume experiments. Three models were used to study the possible distributions of wave phases. They are: a uniform distribution according to the usual assumptions: a model due to Tayfun & Lo (1989); and finally a beta distribution. The latter two models were found to provide better fits than the usually assumed uniform distribution. INTRODUCTION In the study of ocean waves, the phases of surface waves are often assumed to be uniformly distributed over the interval [0, 2re]. A natural consequence of this assumption is then that wave components are all independent of each other. Accordingly, surthce fluctuations can be taken to have a normal distribution. When the surface fluctuations are normally distributed, together with the assumption of narrow-bandedness, it can then be shown that the wave heights will have a Rayleigh distribution (Longuet-Higgins, 1952). This model was found by many researchers to be a good approximation for a Variety of wave conditions, ranging from North Atlantic storm (Chakrabarti & Cooley, 1977) to Indian monsoon waves (Dattatri et al., 1979). However, wave data were not always/bund to comply to the Rayleigh distribution (Forristall, 1978). In tact, there are many researchers who have reported that a one- or two-parameter Weibull distribution would be more appropriate (see, tbr example, Krogstad, 1985; Mase, 1989; Sundar et al., 1993). The assumption of normal distribution of the water surlhce is not always true. According to the Stokes finite wave theory, nonlinear waves will have higher crests and flatter troughs as compared to linear, sinusoidal waves.

ABSTRACT: Stochastic models can be used efficiently in order to model the uncertainties in the activity cost and schedules of the offshore platform construction projects. However, they have some disadvantages. For example, the Program Evaluation and Review Technique (PERT) which is a Level II model, will introduce additional bias problems in the estimation of the activity duration. Level III stochastic models (e.g. Monte Carlo simulation), which are developed to avoid the bias problems of the PERT require lots of computational efforts. In this study, the available stochastic models used in the determination of the optimum time and cost trade-off of the offshore construction projects are compared with respect to their efficiencies. Monte Carlo simulations of the network are performed and a network planning model is suggested. These models are applied to Gulf of Mexico offshore platform (Bullwinkle) project. INTRODUCTION Since the search of energy moves into deeper water, offshore operations are extending the limits of platform construction technology. Hence, the degree of the uncertainties in the project schedule and cost estimation increases. The limitations on wave, wind and soil data, the reliability of long term information and design methodologies are the main uncertainties in the planning of offshore operations. In order to achieve the project Objective on time and budget, the offshore operations are optimized by using stochastic network planning techniques, which can model the uncertainties. In this paper, the sensitivity of the techniques to the probability distributions which are used to model the construction activities are investigated. In order to solve the bias problems associated with the Level II stochastic models, the project system can be simulated. However, simulation is limited by the constraints of economy and computer capability. Therefore, a network planning model is suggested by modifying the Level II models.