ADB610Steel Fatigue Crack Propagation Statistical Distribution Research

Even the experimental process being controlled strictly, the fatigue crack propagation has undetermined performance in the statistical sense. Although the random difference of fatigue crack propagation can’t be described by determined method, it follows certain statistical regularity.So it is necessary to use probability and statistics to study the undetermined performance. The newly developed low-carbon bainite steel ADB610is currently widely used in pressure steel pipe and volute of hydropower station.Therefore,it is significant to investigate the undetermined performance of fatigue crack propagation for ADB610steel.In this thesis, based on the tensile test of ADB610steel12C(T) samples were cut from ADB610steel plate and machined to the desired size.Then test of fatigue crack propagation were carried out by using continuous cycle reduction K method. The data of crack length(a) and crack propagation life(N) were obtained.The fatigue crack propagation rate(da/dN) of each sample was calculated by using fitting of seven data method and secant method respectively.And based on Paris formula(da/dN=C(△K)m),the relation formula between da/dN and stress intensity factor range(△K) of each sample were established.Using least square method.the lgC and m of each sample under these two methods was obtained. Normal distribution, logarithmic normal distribution, exponential distribution, and two-parameter Weibull distribution were used to fit12data of lgC and m using least square method and K-S test.Distribution study shows that the best distributions of lgC and m calculated by using seven data method and secant method respectively are both normal distribution. Then statistical research on standard deviation and mean value of lgC and m under these two methods shows that the standard deviation and mean value of lgC and m have insignificant difference. That is to say that even data method and secant method both can be used to calculated the fatigue crack propagation rate when using continuous cycle reduction K method test.The standard deviation and mean value confidence intervals of lgC and m using even data method were studied on the base of normal distribution.It was calculated using seven data method for crack propagation life(N) of each sample under five same crack lengths(a) during the early stage, middle stage and last stage of fatigue crack propagation respectively. Statistical research on N shows that the best distribution is two-parameter Weibull distribution.Based on the Paris expression of each sample obtained by using seven data method,five△K were randomly evaluated during the early stage,middle stage and last stage of fatigue crack propagation respectively.Then the da/dNof each sample were calculated under same AK. Statistical research on lg(da/dN)shows that the best distribution is normal distribution during the early stage and middle stage of crack propagation, and during the last stage that the best distribution is two-parameter Weibull distribution. In this thesis all calculations were programmed using MATLAB.