To determine the effects of stress ratio and relative humidity of laboratory air on the fatigue crack growth rates of grade 260 rail steels of the draft European Rail Standard. To measure the scatter in fatigue crack growth rates measured in six laboratories under conditions within the European Rail Standard draft, and to propose the test method and acceptance limits for inclusion in the European Rail Standard for grade 260 rail.
The Part 1 tests showed that stress ratio had a significant effect on the slope and intercept of the stable crack growth rate region of the log - log plot of the da/dN vs delta K relationship. The project has concluded that there was no significant difference in the crack growth rate behaviour between the specimens from different rail manufacturers and that there was no significant difference between the results from the six laboratories. It was seen that there was an interaction between humidity and laboratory in one laboratory where, due to the use of a manual crack measurement technique, the tests were interrupted overnight. It is recommended that interruption of the test, although permissible in BS 6835 "Determination of the rate of fatigue crack growth in metallic materials", should be avoided and the test should be continuous.
A five point moving average method for the slope of the log - log plot of da/dN against delta K is recommended for determining the stable crack growth rate behaviour. The slope and intercept of the stage 2 behaviour were directly related in these tests and the mean value of the slope was close to 4. A proposal for the test conditions and method of analysing the results has been submitted for the consideration of CEN TC256. Examples have been found where software used to calculate da/dN and delta K have used incorrect methods to calculate these values, and it is recommended that all software is validated before use.
The setting of acceptance limits for qualifying tests on rails requires that the scatter in results due to test technique is sufficiently small so that the test can identify acceptable rails and reject unacceptable ones. A previous exercise to determine acceptance limits worked on the mean of results + 3 standard deviations as being the maximum crack growth rate allowed. Even with this limit there were differences in results between laboratories which resulted in acceptable rails failing the proposed specification when tested in some laboratories. The cause of these differences needs to be identified so that the test procedure and the acceptance limits can be set in the standard.
The rails from four manufacturers were machined to produce ten single edge notched bend specimens, as described in the draft standard, from each rail. Part 1 of the project determined the effects of different factors, namely rail source, laboratory, stress ratio at values of 0.2 and 0.5, and relative humidity on the fatigue crack growth rate by carrying out twenty tests in two laboratories using an experimental design. Part 2 tests were carried out under a stress ratio of 0.5, with test temperature and relative humidity recorded but not controlled beyond the ambient laboratory condition, with cyclic test frequency in the range 10 - 120 Hz added as a factor in the experimental design. Test frame and crack monitoring details were also recorded, but not explicitly studied.
Funding SchemeCSC - Cost-sharing contracts
4465-591 Sao Mamede De Infesta
NE1 7RU Newcastle Upon Tyne