成组法疲劳试件个数选取方法研究

目的 在成组法疲劳试验过程中，确定最少疲劳试件个数判据，获取有效可靠的疲劳试验结果。方法基于疲劳可靠性的基本原理，在疲劳寿命对数正态分布的假设下，根据t分布理论，结合船舶及海洋工程一般构件具有95%置信度和可靠度p=97.72%下的P-S-N曲线的基本要求，推算出相应置信度和可靠度下5%误差限度内的疲劳试件个数与允许的最大变异系数数值的对应关系，以此作为成组法疲劳试验最少试件个数的判据，开展典型节点的疲劳试验。结果 得到了具有95%置信度的百位估计值Np，并通过线性相关系数r，判断对数疲劳应力Y=lg S与X=lg Np之间线性相关的程度，采用最小二乘法对对数疲劳寿命lg Np和对数疲劳应力lg S进行线性拟合，得到了具有95%置信度的P-S-N曲线，与IIW规范中相应的曲线一致。结论 文中给出的成组法疲劳试验个数确定方法是合理、可靠的，为船舶与海洋工程结构物典型焊接接头成组法疲劳试验个数的确定提供了依据。

Selection Method of the Number of Fatigue Specimens by Group Method

The work aims to determine the minimum number of fatigue specimens in the process of fatigue test by group method and obtain effective and reliable fatigue test results.Based on the basic principle of fatigue reliability, under the assumption of lognormal distribution of fatigue life, according to the t distribution theory and combined with the basic requirements of P-S-N curve with 95% confidence and 97.72% reliability for general components of ships and offshore engineering, the corresponding relationship between the number of fatigue specimens within 5% error limit of corresponding confidence and reliability and the allowable maximum variation coefficient value was calculated, which was used as the criterion for the minimum number of specimens in the fatigue test by group method, and then the fatigue test of typical nodes was carried out. The estimated value Np with 95% confidence was obtained, and the degree of linear correlation between logarithmic fatigue stress Y=lgS and X=lgNp was judged by linear correlation coefficient r. The logarithmic fatigue life lgNp and logarithmic fatigue stress lgS were linearly fitted by least square method, and a P-S-N curve with 95% confidence was obtained, which was consistent with the corresponding curve in IIW code. The method for determining the number of fatigue tests by group method given is reasonable and reliable, which provides a basis for determining the number of fatigue tests by group method for typical welded joints of ships and offshore engineering structures.