PSI - Issue 14

Ilyin A.V. et al. / Procedia Structural Integrity 14 (2019) 964–977 Author name / Structural Integrity Procedia 00 (2018) 000 – 000

972

9

following formulae summarize the experimental data gained in air medium.   (1 0.75 ) th 0 r K K K ef       for r  0,

(11a)

K



  (1 0.5 ) th 0 r K   (11b) For loading in air the values of coefficients in (10) recommended by Sadkin et al (2016) are С = 1.0  10 -13 , q = 3.1 when SIF is measured in N/mm 3/2 , crack growth rate in mm/cycle. These figures practically do not depend on strength of steel. According to the known standards, the effect of corrosion medium can be taken into account by double to quadruple magnification of the crack growth rate in relation to loading in air. Simultaneously, there are recommendations to simulate in tests an actual loading frequency of a structure f that is in most cases impossible. The key meaning has an answer to the question: if any the most damaging test frequency can be found to get conservative data for all structures? Our trials have shown that an answer is attainable when the corrosion medium effect is reduced to anodic dissolution of plastically deformed metal in a crack tip. In this case two processes compete: solution of metal ahead a crack tip resulting in faster crack growth, and a crack tip blunting resulting in the decrease of strain level within process zone. Tests have been carried out at f differing in two orders, from 1 to 0.01 Hz. Figure 5 shows the test results for sinusoidal round-the-clock cycling at 0.01 Hz frequency obtained by Sadkin et al (2016) in the low-cycle area. 1 0.77  r K   ef  for r < 0.

Figure 5 – Effect of loading frequency under uninterrupted cyclic load (1,  – in air; 2, + – 0.1 Hz in a corrosion medium; 3,  – 0.01 Hz in a corrosion medium) and with long breaks under load in a corrosion medium (4,  ). In this case not more than 2.5 times increase of da / dN was observed in relation to loading in air. However when long breaks in cycling under the maximum load are simulated, after each break a crack stops for some cycles, and if a time ratio “pause/cycling” is equal to 1:2, the test result is practically the same as in air. Hence, to predict the service life of structures long time operated in corrosion media it should be checked whether such a structure is actively deformed during the whole time of operation.