平均应变对0Cr18Ni10Ti钢随机循环应力-应变关系的影响规律

Effects of Mean Strain on the Random Cyclic Stress-Strain Relations of 0Cr18Ni10Ti Pipe Steel

试验研究了平均应变对0Cr18Ni10Ti管道钢随机循环应力-应变关系的影响规律。从节约试样和试验费用角度,提出了改进的极大似然疲劳试验法,以应变比分别为–1、–0.52、–0.22、0.029、0.18和0.48完成了104个试样的应变控制疲劳试验。揭示出材料具有Masing行为和完全平均应力松弛特征,现有平均应变理论无法表征其影响。因此,首先基于Ramberg-Osgood方程,应用广义极大似然法有效地测定出各应变比的材料随机循环应力-应变关系。通过比较各应变比的循环应力幅均值、循环应力幅均方差和概率循环应力幅,揭示出高可靠性管理时,平均应变具有降低循环应力幅作用,应变比接近0时最显著,偏离逐渐减弱。说明了研究平均应变效应,仅考虑均值,可能给出错误评价;必须综合考虑均值、均方差和样本量3因素才能给出合理评价。

Experimental study is performed for the effects of the mean strain on the random cyclic stress-strain relations of the new nuclear material, 0Cr18Ni10Ti pipe steel. From saving the size of specimens, an improved maximum likelihood fatigue test method is proposed to operate the present strain-controlled fatigue tests. Six straining ratios, –1, –0.52, –0.22, 0.029, 0.18, and 0.48, respectively, are applied to study the effects. Fatigue test has been carried out on totally 104 specimens. The test results reveal that the material exhibits a Masing behaviour and the saturation hysteresis loops under the six ratios hold an entirely relaxa- tion effect of mean stress. There is no effectively method for the description of the mean straining effects un- der this case. Previous Zhao’s random stress-strain relations are therefore applied to characterizing effectively the scattering test data under the six ratios on a basis of Ramberg-Osgood equation. Then the effects of the ratios are analyzed respectively on the average stress amplitudes, the standard deviations of the stress ampli- tudes, and the stress amplitudes under different survival probabilities and confidences. The results reveal that the ratios act a relatively decreasing effect to the stress amplitudes under higher survival probabilities and confidences. The strongest effect appears at the ratio of 0.029, and a weaker effect acts as the distance in- crease of the ratio from the zero. In addition, it is indicated that the effects from the sense of average fatigue lives might result in a wrong conclusion. The effects can be appropriately assessed from a probabilistic sense to take into account the scattering regularity of test data and the size of sampling.