On the parametric assessment of fatigue disparities疲劳差异的参数化评估
![]()
Kufoin EN, Susmel L, 2024. On the parametric assessment of fatigue disparities. Probabilistic Engineering Mechanics, 77: 103651.DOI: 10.1016/j.probengmech.2024.103651
整合不同来源的疲劳数据集是提高疲劳评估和设计可靠度的有效策略，同时还能降低成本并缩短时间。统计参数分析方法可应用于疲劳数据集，以确定其在统计学上差异显著 (不相似) 与否 (相似)。本文系统地采用统计参数检验假设来评估显著性。为验证此方法，本文以不同缺口试件生成的疲劳数据集为案例研究，孔径范围为 0-3 mm，同时结合文献数据。尤其是使用总应力来确保仅通过统计分析识别疲劳数据集的差异。该方法对缺口几何形状差异小至 1 mm 的情形表现良好，并且能识别铸铁的缺口不敏感性。因此，该方法可根据统计参数而非其他物理参数来区分疲劳数据集。关键词: 统计学, 显著性, 检验统计量, 共线性, 疲劳Efficiently merging fatigue datasets from diverse sources has proven to be a strategic approach for enhancing the reliability of fatigue assessment and design within industry, while concurrently streamlining costs and time. Statistical parametric analysis is an approach that can be applied to fatigue datasets to determine whether the datasets can be deemed statistically significant (different) or statistically insignificant (similar). This paper systematically employed statistical parametric test-statistic hypotheses to assess significance. To validate this approach the paper used as a case study, fatigue data sets generated from varied notched specimens with hole diameters ranging from 0 mm to 3 mm, in addition to data from the literature. In particular, gross stresses were utilized to ensure that the only means to identify differences in the fatigue datasets was through statistical analysis. This approach was observed to work well for geometries with differences in notch geometry as small as 1 mm and was able to identify notch insensitivity in cast iron. Thus, this method can be used to differentiate fatigue datasets based on statistical parameters rather than other physical parameters.
Keywords: Statistical; Significance; Test-statistic; Collinear; Fatigue.图 1: 检验两条 S-N 曲线统计显著性的流程图Fig. 1. Flow chart to illustrate how to test the statistical significance of two S-N curves
图 2: 样本几何形状: (a) 平面图; (b) 带给定直径的圆形缺口Fig. 2. Specimen geometry in mm: (a) Plane; (b) Notched with a circular notch of diameter ∅
Fig. 3. Some samples of failed specimens after fatigue tests
图 4: 95% 置信水平与 5% 显著性水平的散点带: (a) 钢及其含 1 mm 直径缺口的样品; (b) 钢及其含 2 mm 直径缺口的样品; (c) 钢及其含 3 mm 直径缺口的样品; (d) 含 1, 2 mm 直径缺口的钢样品; (e) 含 1, 3 mm 直径缺口的钢样品; (f) 含 2, 3 mm 直径缺口的钢样品Fig. 4. Scatter bands at 95% level of confidence and 5% level of significance: (a) S and S_1 mm; (b) S and S_2 mm; (c) S and S_3 mm; (d) S_1 mm and S_2 mm curves; (e) S_1 mm and S_3 mm curves; (f) S_2 mm and S_3 mm curves
图 5: 95% 置信水平与 5% 显著性水平的散点带: (a) 黄铜及其含 1 mm 直径缺口的样品; (b) 黄铜及其含 2 mm 直径缺口的样品; (c) 黄铜及其含 3 mm 直径缺口的样品; (d) 含 1, 2 mm 直径缺口的黄铜样品; (e) 含 1, 3 mm 直径缺口的黄铜样品; (f) 含 2, 3 mm 直径缺口的黄铜样品Fig. 5. Scatter bands at 95% level of confidence and 5% level of significance: (a) Br and Br_1 mm; (b) Br and B_2 mm; (c) Br and Br_3 mm; (d) Br_1 mm and Br_2 mm; (e) Br_1 mm and Br_3 mm; (f) Br_2 mm and Br_3 mm curves
图 6: 95% 置信水平与 5% 显著性水平的散点带: (a) 铸铁及其含 1 mm 直径缺口的样品; (b) 铸铁及其含 2 mm 直径缺口的样品; (c) 铸铁及其含 3 mm 直径缺口的样品; (d) 含 1, 2 mm 直径缺口的铸铁样品; (e) 含 1, 3 mm 直径缺口的铸铁样品; (f) 含 2, 3 mm 直径缺口的铸铁样品Fig. 6. Scatter bands at 95% level of confidence and 5% level of significance for: a) CI and CI_1 mm, b) CI and CI_2 mm, c) CI and CI_3 mm, d) CI_1 mm and CI_2 mm, e) CI_1 mm and CI_3 mm and f) CI-2mm and CI_3 mm
图 7: Louks 与 Susmel 给出的锻造 (悬臂和弯曲) 重要数据集在 95% 置信水平与 5% 显著性水平下的散点图Fig. 7. Scatter band at 95% level of confidence and 5% level of significance for significant data set of as-forged (cantilever) and as-forged (bending) in Louks & Susmel (2015)
作者信息 | Authors
英国谢菲尔德大学 (University of Sheffield) 土木与结构工程系
Luca Susmel, 通讯作者 (Corresp.)英国谢菲尔德大学 (University of Sheffield) 土木与结构工程系Email: l.susmel@sheffield.ac.uk
律梦泽 M.Z. Lyu | 编辑 (Ed)
P.D. Spanos | 审校 (Rev)
陈建兵 J.B. Chen | 审校 (Rev)
彭勇波 Y.B. Peng | 审校 (Rev)
