| 摘要: |
| 含裂纹管道评估模型预测精度的分布情况对研究管道失效概率和改进模型都至关重要。基于89组实验数据,结合极大似然估计法和Kolmogorov-Smirnov检验对7种含外表面轴向裂纹管道安全评估模型的预测精度进行统计推断,并根据其分布规律对不同模型的预测性能进行对比分析。结果表明:除Battelle和CorLAS模型的预测精度分别服从Gumbel分布和正态分布外,其他规范的预测精度均服从对数正态分布;7种模型的优先级顺序和推荐情况为R6-3、R6-1(或API)、R6-2和SINTAP规范;BS7910和GB规范应适用于要求比较高的项目;CorLAS和Battelle模型可用于项目的预估计;采用失效评定图(FAD)评估模型,两种FAD曲线无明显差异,受断裂比的影响较大,因此可通过改进断裂比的计算精度提高模型的预测水平。 |
| 关键词: 含外表面轴向裂纹管道 预测精度 统计推断 极大似然估计 Kolmogorov-Smirnov检验 |
| DOI：10.3969/j.issn.1673-5005.2021.02.017 |
| 分类号::TU 81 |
| 文献标识码:A |
| 基金项目: |
|
| Statistical inference on prediction accuracy of evaluation models for a pipeline with axial surface cracks |
|
GUO Lingyun1,2, ZHOU Jing1,2
|
|
(1.State Key Laboratory of Coast and Offshore Engineering, Dalian University of Technology, Dalian 116024, China;2.Institute of Earthquake Engineering, Faculty of Infrastructure Engineering, Dalian University of Technology, Dalian 116024, China)
|
| Abstract: |
| The distribution of the prediction accuracy for evaluation model plays an important role in studying the failure probability and improving the burst capacity model for the crack-containing pipeline. Based on 89 sets of test data, the prediction accuracy of seven evaluation models was statistically inferred by the maximum likelihood estimation and the Kolmogorov-Smirnov test. According to the distribution law, the prediction performance of different models was compared and analyzed. The results show that, the prediction accuracy of Battelle and CorLAS obeys the Gumbel distribution and the normal distribution respectively, while those of others all obey the lognormal distribution. The priority order and application of the seven models are R6-3, R6-1 (or API), R6-2 and SINTAP specifications. BS7910 and GB should be applied to the evaluation projects with high demand, and CorLAS and Battelle are suitable for the pre-estimation of the projects. The models based on failure assessment diagram (FAD), where the two FAD curves are not significantly different, are influenced by the fracture ratio. Therefore, the prediction performance of the model can be enhanced by improving its computational accuracy. |
| Key words: pipeline with axial surface cracks prediction accuracy statistical inference maximum likelihood estimation Kolmogorov-Smirnov test |
