TY - JOUR
T1 - Probabilistic prediction of the spudcan extraction from sand-over-clay based on monitored data
AU - Wang, Y.
AU - Hu, Pan
AU - Li, J.
PY - 2023/8/1
Y1 - 2023/8/1
N2 - Upon completion of the operation on a site, spudcan foundations are pulled out of the seabed before relocating the jack-up platform to another candidate site. It is critical to predict the peak extraction resistance before commencing the jack-up extraction. Though sand-over-clay seabed is commonly found in active marine oil and gas exploration zones, the models for estimating the extraction resistance are very rare. This paper proposes a Bayesian prediction model based on 51 groups of centrifuge tests on foundations extracting from sand overlying clay soils. Model factors are introduced to define the uncertainties in the prediction. Combined with the monitored data during foundation extraction, possible breakout range of resistance and depth is predicted and updated in real time. The most likely breakout resistance and depth in the area are also estimated. The results show that (1) the predicted range of breakout resistance and depth may occur are concentrated around the measured values with the update of the monitored data; (2) the relative error between the most likely breakout depth and the measured value is kept within 10%; and (3) regardless the shape of the foundation, the Bayesian prediction model can effectively predict the extraction behaviour of the foundation.
AB - Upon completion of the operation on a site, spudcan foundations are pulled out of the seabed before relocating the jack-up platform to another candidate site. It is critical to predict the peak extraction resistance before commencing the jack-up extraction. Though sand-over-clay seabed is commonly found in active marine oil and gas exploration zones, the models for estimating the extraction resistance are very rare. This paper proposes a Bayesian prediction model based on 51 groups of centrifuge tests on foundations extracting from sand overlying clay soils. Model factors are introduced to define the uncertainties in the prediction. Combined with the monitored data during foundation extraction, possible breakout range of resistance and depth is predicted and updated in real time. The most likely breakout resistance and depth in the area are also estimated. The results show that (1) the predicted range of breakout resistance and depth may occur are concentrated around the measured values with the update of the monitored data; (2) the relative error between the most likely breakout depth and the measured value is kept within 10%; and (3) regardless the shape of the foundation, the Bayesian prediction model can effectively predict the extraction behaviour of the foundation.
UR - https://hdl.handle.net/1959.7/uws:73730
U2 - 10.1016/j.oceaneng.2023.114787
DO - 10.1016/j.oceaneng.2023.114787
M3 - Article
VL - 281
JO - Ocean Engineering
JF - Ocean Engineering
M1 - 114787
ER -