长江科学院院报 ›› 2019, Vol. 36 ›› Issue (8): 159-164.DOI: 10.11988/ckyyb.20171386

• 水工结构与材料 • 上一篇    下一篇

基于Kriging模型的重力坝非概率可靠性分析

占良红,魏博文,钟紫蒙,徐镇凯   

  1. 南昌大学 建筑工程学院,南昌 330031
  • 收稿日期:2017-12-15 修回日期:2018-01-15 出版日期:2019-08-01 发布日期:2019-08-01
  • 作者简介:占良红(1993-),男,江西上饶人,硕士研究生,主要从事水工结构安全监测与数值优化方面的研究。E-mail:zlh201412@163.com
  • 基金资助:
    国家自然科学基金资助项目(51779115,51569014);
    南昌大学研究生创新专项资金项目(CX2018052)

Non-probabilistic Reliability Analysis of Gravity Dam Based on Kriging Model

ZHAN Liang-hong,WEI Bo-wen,ZHONG Zi-meng,XU Zhen-kai   

  1. School of Civil Engineering and Architecture, Nanchang University, Nanchang 330031, China
  • Received:2017-12-15 Revised:2018-01-15 Published:2019-08-01 Online:2019-08-01

摘要: 为解决概率可靠性方法分析受限于重力坝不确定参数的统计信息较难获取,及其功能函数高度非线性且较难用显式解析表达式给予刻画等问题,构建了基于不确定参数界限的重力坝单元非概率可靠性(Non-probabilistic Reliability,N-PR)指标计算模型,在此基础上,融合Kriging模型较强预测和拟合优势,发展了一种基于Kriging模型的重力坝N-PR指标计算方法。经实例考证,所给模型与方法可有效规避概率可靠性分析中要求不确定参数随机变化且计算结果对参数高度敏感的局限,较好适应了重力坝功能函数高度非线性甚至难以显式表达的特点,可较深入分析重力坝可靠性。

关键词: 重力坝, 非概率可靠性指标, Kriging模型, 区间变量, 抗滑稳定

Abstract: In the reliability analysis for gravity dam, statistical information is challenging to be obtained by probability reliability analysis due to the limitations posed by uncertain parameters, and the functions are highly nonlinear hence are difficult to be described with analytical expression. In view of this, a non-probabilistic reliability(N-PR) index calculation model of gravity dam in consideration of uncertain parameter ranges is constructed. On this basis,the calculation method of N-PR index for gravity dam in association with Kriging model’s strong prediction and fitting advantages is built.A practical engineering example demonstrates that the present method effectively overcomes the limits that the uncertain parameters shall vary randomly in probability reliability analysis and the calculation results are highly sensitive to the parameters. The method also adapts to the highly nonlinear function of gravity dam.

Key words: gravity dam, non-probabilistic reliability, Kriging model, interval variables, anti-slip stability

中图分类号: