边坡稳定的有限元可靠度分析方法及参数敏感性研究

发布时间：2018-03-01 15:05

本文关键词： 边坡稳定可靠度响应面法最优化算法　出处：《贵州大学》2015年硕士论文　论文类型：学位论文

【摘要】：边坡岩土体的性质参数具有不确定性和相关性,传统的定值分析方法已经不能全面的评价边坡的稳定性,探讨更为有效的评价边坡稳定性的方法已成为当前重要的研究方向。为了更全面的评价边坡的稳定性,在传统确定性分析的基础上,研究人员提出了对边坡的稳定性进行可靠度评价的思想。鉴于此,论文结合土性参数对边坡稳定的敏感性,对边坡可靠度分析方法——响应面法做了研究。首先,结合边坡土体破坏机理,建立了适合边坡工程的二次多项式响应面表达式,探讨了二次多项式响应面法的MATLAB编程实现过程。选取土质边坡采用此方法进行可靠度分析,以蒙特卡洛法的分析结果作为对比,两者误差小于5%。说明此方法应用于边坡稳定可靠度分析是可行的。其次,分析了土性参数对边坡稳定性影响的敏感性,分析结果对于简化边坡的可靠度分析具有重要意义。分析表明:土的粘聚力c是影响边坡可靠度的最敏感的因素,容重γ是最不敏感的因素,而内摩擦角φ的敏感性介于两者之间;土的粘聚力和内摩擦角之间的负相关性对边坡的可靠度有利,在边坡可靠度分析时可以适当考虑其有利影响。最后,针对二次多项式响应面法的不足,探讨了将BP神经网络响应面法应用于边坡的可靠度分析。运用BP神经网络拟合边坡稳定状态响应面,结合最优化算法分析计算边坡的可靠指标和失效概率,通过均质土和非均质土的算例分析表明,该方法应用于边坡的可靠度分析中是可行的,并且简单高效。
[Abstract]:The property parameters of slope rock and soil are uncertain and relevant, so the traditional method of fixed value analysis can not evaluate the stability of slope comprehensively. It has become an important research direction to explore more effective methods to evaluate slope stability. In order to evaluate slope stability more comprehensively, based on the traditional deterministic analysis, The researchers put forward the idea of reliability evaluation of slope stability. In view of this, considering the sensitivity of soil parameters to slope stability, the reliability analysis method-response surface method (RSM) is studied in this paper. According to the failure mechanism of slope soil, the quadratic polynomial response surface expression suitable for slope engineering is established, and the MATLAB programming process of quadratic polynomial response surface method is discussed. Compared with the results of Monte Carlo method, the errors of the two methods are less than 5. It is proved that the method is feasible for the reliability analysis of slope stability. Secondly, the sensitivity of the influence of soil parameters on slope stability is analyzed. The analysis results are of great significance to simplify the reliability analysis of slope. The analysis shows that the cohesion of soil c is the most sensitive factor affecting the reliability of slope, and the bulk density 纬 is the least sensitive factor. The sensitivity of internal friction angle 蠁 is between them, and the negative correlation between cohesive force and internal friction angle of soil is favorable to the reliability of slope. In view of the shortcomings of quadratic polynomial response surface method, the reliability analysis of slope using BP neural network response surface method is discussed. The BP neural network is used to fit the slope stable state response surface. The reliability index and failure probability of slope are analyzed and calculated by the optimization algorithm. The example of homogeneous soil and heterogeneous soil shows that this method is feasible and simple and efficient in the reliability analysis of slope.
【学位授予单位】：贵州大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TU43

【参考文献】