The Experimental Verification of Extreme Value Response of a Moored Tanker

Abstract

Abstract For the design of a mooring system, extreme values of the motion response of Floating Production Storage and Offloading vessel moored in random seas must be investigated. The tank test results of a model of such a tanker anchored in 800m of water are analysed. The extreme valueestimation for the high and low frequency response was done by methods detailed in Naess I, and the API RP2P2. The main conclusions are that the Gaussian method underestimates the prediction. The API RP2P underestimates the heave and the roll motion of the tanker, but shows good comparison for surge, sway and yaw. The SRSS (Square rootof sum of squares) gives results which are equivalent to the API estimation. Both the Least Square method and the Modified SRSS overestimate in some cases. The underestimation of the loading in the heave direction by the API code has been taken into account in the design. The statistical parameters are compared with similar published data to show the dominance of the 2nd order effect on motion response and the tensions in the mooring lines. Further analysis was done by dividing the time history inoverlapping intervals as proposed by Kinoshita3. The influence on the estimation of the extreme values are discussed. The significance aspect of the work is the prediction of extreme values from either the time series simulation or the experimental data of limited length to help in load estimation of mooring lines. Introduction In the future, the oil and gas industry is heading towards exploration and production in deeper waters, the ability to understand and predict the dynamic behaviour of moored floating structures is therefore of great importance. The significance of the low frequency response in relation to thewave induced responses and the total responses is an important matter. In the analysis of the motion response of a floating structure moored in random seas, it is important to investigate the extreme values of the motion. This motion response has a quadratic component, which is normally called the slow-drift motion. The quadratic response is non-Gaussian in nature, although the total response is generally treated as Gaussian (Ref. 3). It is known that the dynamic behaviour of moored floating structures to a large extent is characterised by non-linear phenomena. This makes experimental work particularlyimportant. For complete studies of the key phenomena of the motion behaviour of the system and the resulting extreme responses, the combination of experimental work with parallel theoretical/numerical work is often an optimal solution. The estimation of extreme values by mathematical techniques have been proposed by various authors in the past. Naess4 gave an explicit expression for probability distribution of 1st and 2nd order response. But the Formula was too complicated for practical use. Kato5 presented an approximate procedure for calculation of the combined response distribution which is much easier to use.