Fatigue assessment of offshore tubular structures

Abstract

Fatigue assessment of offshore structures has always been an important area of study due to a large number of cyclic loading acting on the structure. Traditionally, offshore structures are simulated in time-dependent dynamic analysis to calculate the stress response at critical locations of the structure. Sorting of relevant stress cycles is done using a suitable cycle counting algorithm (usually rainflow counting) and then an appropriate S-N curve and Miner’s cumulative rule are used to estimate the total fatigue damage at these critical locations. Offshore structures are usually designed for a minimum fatigue life of 20 years, which takes a lot of computational time in FEM. If loads are converted to an equivalent frequency domain and the structure is analyzed in frequency dependent dynamic analysis, computational time can be reduced. This will help engineers perform faster optimizations on new structural developments.
For frequency analysis, using Dirlik’s semi-empirical method, a good estimation of fatigue damage are achieved according to literature. Application of Dirlik’s method on various structures has been validated by comparing the damage values with equivalent rainflow counting method. Unlike rainflow counting algorithm, Dirlik’s method calculates probability density function of the stress ranges with empirical formulae and uses S-N curve and Miner’s rule to estimate total fatigue damage. Dirlik’s method is easy to understand, easy to program and estimate fatigue damage with good accuracy. This might also help us in calculating fatigue damage in the stinger, due to bow deflection.
To fully understand the method of dynamic frequency analysis with Dirlik’s method, the objective of the research is to study its feasibility on simple models by conducting simulations in FEM and to establish the fatigue calculation procedure. Two models are selected in this research, a cantilever beam and a simple frame model (for tubular joints). Simulation on both the models are conducted in frequency analysis and fatigue estimation is done using Dirlik’s method. To verify the estimated damage values, equivalent time-dependent analyses are also conducted with rainflow counting method.
Comparable fatigue damage results are achieved for cantilever beam but, for frame model, results obtained could not be compared due to issue found by conducting detailed analysis. Although, the overall objective of the research is not fully complete, results and observations found in this research will be helpful for conducting additional simulations with different strategies to verify the frame model. With this research, a firm basis is developed which could motivate the company to further investigate the method.