平均应变对0Cr18Ni10Ti管道钢随机应变-疲劳寿命关系的影响规律

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

试验研究了平均应变对0Cr18Ni10Ti管道钢随机应变-寿命关系的影响规律.从节约试样和试验费用角度,采用改进的极大似然疲劳试验法,应变比分别为-1、-0.52、-0.22、0.029、0.18和0.48的条件下,完成了104个试样的应变控制疲劳试验.在材料具有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情况下,首先基于Coffin-Manson方程,应用广义极大似然法有效地测定出各应变比下材料的随机应变-寿命关系.通过比较各应变比下疲劳寿命均值、对数疲劳寿命均方差和概率疲劳寿命,揭示出0Cr18Ni10Ti管道钢高可靠性时,应变比大于0有大约1.3～1.6的正面效应,小于0为负面效应.现有平均应变效应研究仅考虑疲劳寿命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 on the mean strain effects on the random cyclic strain-life relations of the new nuclear material, 0Cr18Ni10Ti pipe steel. In order to save costs of specimens and tests, an improved maximum likelihood fatigue test method is applied to manage the present strain-controlled fatigue tests. Six straining ratios, respectively, -1, -0.52, 0.22, 0.029, 0.18, and 0.48, are applied to study the effects. Total 104 specimens are fatigued. Since the material exhibits an entirely relaxation effect of mean stress under the six ratios and, in addition, there is no effectively method for the description of the mean straining effects under this case, previous Zhao's random strain-life relations are therefore applied for effective characterization of the scattering test data under the six ratios on a basis of Coffin-Manson equation. Then the effects of the ratios are analyzed respectively on the average fatigue lives, the standard deviations of the logarithms of fatigue lives, and the fatigue lives under different survival probabilities and confidences. The results reveal that the ratios greater than zero exhibit a positive effect of about 1.3 to 1.6 times under the survival probability of 0.999 and the confidence of 95%. A negative effect is exhibited for the case of the ratios less than zero. In addition, the assessment of the effects from the sense of average fatigue lives might result in a wrong conclusion for the practice of higher reliabilities. The effects can be appropriately assessed from a probabilistic sense to take into account the average lives, the scattering regularity of test data, and the size of sampling.