A jack-up rig has to be designed for extreme storm conditions in its elevated mode during operations. Guidelines of ISO 19905-1 [1] and SNAME TR-5-5A [2] for site specific assessment of jack-up rigs explain in detail such analysis and assessment requirements. It is well known that for higher water depths and extreme environment, structural dynamics and spudcan-soil interaction plays a very significant role. The extreme storm response can be determined either by a two-stage deterministic storm analysis procedure using a quasi-static analysis that includes an inertial load set or by a more detailed fully integrated (random wave) dynamic analysis procedure that uses a stochastic analysis. More commonly, however, jack-up rigs are assessed using a two-stage deterministic wave model along with steady wind loads, since this is much simpler. In two-stage deterministic analysis, the first step is the determination of the inertial load set and structural analysis for all the environmental and gravity loads. To include the effects of the spudcan and soil foundation, an initial rotational stiffness or fixity is assumed that depends on the soil type and the preload. Assessment of the foundation is performed thereafter using the yield interaction approach. This is normally an iterative approach to arrive at the right fixity that satisfies the assessment.
The two-stage approach, although simpler could be conservative leading to adverse conclusions for the suitability of a jack-up at a site. As indicated, the other approach is the one-stage approach involving random time domain analysis which is normally not used and reported in the literature probably due to its complexity and difficulty.
The present paper describes random wave time domain analysis of a specific jack-up using a 3D model in 400 ft water depth using USFOS software [3]. This software has the spudcan-soil interaction integrated, to simulate the foundation behavior. In this one-stage analysis, the assessment for the foundation is performed through an iterative approach inside the software using yield and bounding surfaces. Extreme values of some of key responses are compared with traditional deterministic analysis. Benefits and limitations of random wave time domain analysis are explained and quantified. These benefits are sometimes so significant that one-stage analysis may lead to favorable conclusions where the conservative two-stage analysis approach had failed to show the adequacy of the rig. These observations and the overall methodology of analysis used here could be beneficial to any rig’s applicability at a specific site.
TSTNN: Two-Stage Transformer Based Neural Network for Speech Enhancement in the Time Domain
