摘要
参数反演确定堆石体的强度变形参数已成为堆石体参数取值的有效途径之一。在目前的堆石体参数反演方法中,未考虑不同材料分区堆石体力学特性的差异和各自分区的应力变形特点,对整个坝体或某个断面采用一个目标函数进行参数反演。本文考虑堆石坝不同材料分区的受力变形特点,建立不同材料分区的反演目标函数,并基于非支配排序的多目标遗传算法(NSGA-Ⅱ),兼顾各分区间材料特性对变形的相互影响,提出了基于NSGA-Ⅱ和RBF神经网络的堆石坝多目标参数反演方法。以水布垭堆石坝为例,采用该方法对坝体主堆石和次堆石区的瞬变-流变参数进行了反演分析,并与单目标参数反演方法进行了对比分析。计算结果表明,所有测点的变形量和变形趋势上与实测值基本一致,反演效果明显优于单目标参数反演结果,表明该方法对于多分区堆石坝的材料参数反演更为合理有效。
Parameter inversion has become an effective method for determining strength and deformation parameters of rockfill and other material parameters for practical projects. All the existing inversion methods use a single objective function in analysis of a dam section or even the entire dam body, neglecting the differences in mechanical properties and characteristics of stress and deformation between dam zones of different materials. This paper develops a multi-objective inversion method for the transient and rheological parameters of rockfill dams on the basis of a non-dominated sorting multiobjective genetic algorithm-Ⅱ（NSGA-Ⅱ） and a radial basis function（RBF） neural network. In this method, we consider variations in stress and deformation across different material zones and adopt a different objective function for each zone, so that interaction between material characteristics and deformation in each zone can be taken into account through calculation using NSGA-Ⅱ. The method has been applied to the Shuibuya concrete face rockfill dam（CFRD） for joint inversion analysis on the transient and rheological parameters of its two dam zones of different rockfill materials, namely major and minor material zones. Comparison with the single-objective inversion method shows that the NSGA-Ⅱ calculations at all the monitored points agree well with the measured deformations and their variation trends, thus achieving a significant improvement.
作者
温少雄
周伟
李少林
马刚
李鹏鹏
段炼
WEN Shaoxiong ZHOU Wei LI Shaolin MA Gang LI Pengpeng DUAN Lian(State Key Laboratory of Water Resource and Hydropower Engineering Science, Wuhan University, Wuhan 430072 , Changjiang Institute of Survey, Planning, Design and Research, Wuhan 430010)
出处
《水力发电学报》
EI
CSCD
北大核心
2017年第1期86-95,共10页
Journal of Hydroelectric Engineering
基金
国家自然科学基金(51579193
5132290
51509190)
中国博士后科学基金面上资助(2016T907272)
关键词
堆石坝
参数反演
多目标优化
非支配排序
遗传算法
rockfill dam
parameter inversion
multi-objective optimization
non-dominated sorting
genetic algorithm
