Study On Corrosion Fatigue Crack Growth Model Of Marine Steel Under Dry And Wet Cyclic Conditions

The trend towards the use of steel structure in ocean offshore makes it necessary to study corrosion fatigue crack growth rate (CFCGR) which has an influence on steel structure durability. The marine steel not only receives the chemical corrosion of seawater, but also be impacted by ocean currents. And many studies have showed that waves splash zone and tidal zone are the most serious corrosion areas to marine steel. In the current corrosion iatigue studies, most of researches mainly focused on the study of pure corrosion ignoring the iatigue impact of the waves, and others study the electrochemical properties of materials under wet and dry environment without considering the change of the wet/dry time ratio. Offshore platforms at different heights has different wet and dry environments. Therefore, the study of CFCGRunder wet and dry cyclic environment is necessary.In this article, the CFCGR under alternating dry and wet environments were studied in three areas:1) The influence of wet and dry time ratio on CFCGR under the fixed loading frequency condition; 2) The influence of frequency on CFCGR under the fixed wet and dry time ratio condition; 3) The research of CFCGR model under complex environment. This paper presents an investigation on the CFCGR of D36 offshore platformsteel with three-point bending fatigue test. Test conditions involve different alternating wet and dry environments, frequency, stress ratio, cathodic protection potential. This paper use BP neural network algorithm based on genetic algorithm to fit the coefficients of Paris equation, and then get the crack growth rate model.The experimental results reveal that CFCGR are closely related to k and loading frequency f. The loading frequency will affect the corrosion rate of alternating wet and dry, and the lower the frequency the greater the impact. When subjected to alternating wet and dry time ratio, CFCGR shows a trend of first increase and then decrease with the increasing of k in the bw the stress intensity factor amplitude (△K) phase.To low frequency and low △K phase, When subjected to alternating wet and dry time ratio, CFCGR shows a trend of first increase and then decrease with the increasing of k. And there is a critical ratio value which is vary with frequency to make the CFCGR maximum. Considering the complexity of environmental factors (loading frequency, stress ratio, alternating wet and dry longer than, applied potential, etc.), the paper establishs a BP neural network algorithm based on genetic algorithm (GA-BP) to the coefficients of Paris CFCGR model. And compared to the traditional network model, this model has high accuracy, good applicability, scalability, etc..Based on GA-BP model, the paper presented that the dry and wet time ratio and crack growth rate were in line with a cubic polynomial relationship at low stress intensity factor amplitude for assumes special circumstances. Extreme value of cubic polynomial corresponded to the critical dry and wet time ratiounderthis kind of dry and wet cyclic condition.