Thermal buckling has become an important issue for HPHT (high pressure high temperature) subsea pipelines design. The uncertainty of design loads could result in a large cost impact in terms of the measures required for controlling pipeline buckles and expansion, such as seabed intervention, buckle trigger device and so on. The uncertainty is represented with condition load effect factor, γc, in the format of Load and Resistance Factor Design (LRFD) in the design guideline DNV RP-F110. The applied loads are multiplied by condition load effect factor, γC, to have design loads. Without addressing the uncertainties involved in the design load properly, it could result over-conservative or under-conservative design. The standard calibration methods of the condition load effect factor, γc, are introduced together with a design example in this paper. The uncertainties of soil resistances are discussed specifically and explain how the large variation and non-symmetric upper and lower bound associated with soil resistance affect the condition load effect factor, γc, calibration. In addition, the statistics of pipeline penetration on the clay and sandy seabed is given based upon the real pipeline as-laid survey data. The pipeline penetration on the sandy seabed has highly non-symmetric probability distribution compared with that in clay seabed. In order to investigate the uncertainties of resulting bending moment response of pipeline under thermal buckling, a statistical analysis approach stemming from the Markov Chain Monte Carlo (MCMC) Bayesian techniques has developed.
The uncertainty of design loads significantly affects the analysis results of subsea pipeline under thermal buckling. It could have a large cost impact in terms of the measures required for controlling pipeline buckles and expansion towards spools etc. These include seabedintervention, buckle trigger device and so on
