【作者】 肖海斌；

【导师】 冯夏庭；

【作者基本信息】 中国科学院研究生院（武汉岩土力学研究所） ， 岩土工程， 2006， 博士

【摘要】 本文以糯扎渡水电站高心墙土石坝为研究对象,针对心墙掺砾料最佳掺砾比例的选择、坝料力学参数的辨识及最优坝料分区结构和坝体稳定性分析等关键问题,采用试验和综合智能分析方法开展了系统的研究工作:1、本文针对不同击实功、不同配比的掺砾料分别进行了压实、固结、渗透、剪切等试验研究,结果发现,混合土料掺砾后可大大改善其压缩性能和强度,并能保证其作为防渗材料的低渗透性;开展了心墙混合料、坝壳Ⅰ区及Ⅱ区堆石料的应力-应变特性试验研究工作;开展了堆石料和掺砾料的流变试验研究工作。这些试验研究为结构优化和坝体结构长期稳定性分析提供了基本资料。2、基于试验数据,引入了改进的遗传算法对坝料邓肯-张非线性弹性本构模型参数和广义开尔文流变本构模型参数进行智能识别,高精度地获得了模型参数。3、将安全性和经济性相结合作为优化目标,运用人工智能方法,建立了进化神经网络优化模型;在坝体分区方案控制因素的变化范围内,采用均匀设计的方法设计分区方案,建立学习与测试样本,以描述方案控制因素与目标函数值之间的神经网络映射关系,采用改进的遗传算法确定了坝体的最优分区方案。4、采用以上所得到的最优材料参数和模型及最优材料分区方案,开展了坝体的稳定性、材料参数的敏感性和坝体的长期稳定性分析,结果表明,本文所确定的设计方案在所考虑的不同分析工况条件下是稳定的。更多还原

【Abstract】 Aiming at the key problems in the construction of super-high rockfill dam with core wall of Nuozhadu Hydraulic Power Station, such as selection of the best proportion of gravel, the intelligent identification of mechanical parameters of composition materials of dam body, the stability analysis on optimizing zone structure of composition materials of dam body and etc, the systemic study work was conducted in this paper in which the experiments and the synthesis intelligent methods were adopted.1. In this paper, a series of experiments including compaction, consolidation, penetration and shear tests were performed on gravelly soil with different compaction effort and proportion. The results indicate that the gravelly soil behaves the more predominant mechanical property than mixing earth material does. In order to obtain the stress-strain relationship of the mixing material of core wall and rockfill materials of dam shell in zonesⅠandⅡ, triaxial tests were performed to offer basic data for the analysis of optimizing structure and stability of dam body. The rheological deformation of rockfill material and gravelly soil is usually very obvious, which may have great effect on the stability of dam body. Therefore, rheological tests were performed to study the rheological properties of rockfill material and gravelly soil, based on which the analysis of long-term stability of dam body structure can be carried out.2. In this paper, we introduced genetic algorithm and modified it. Parameter identification can be performed for composition materials of dam body used Duncan-Chang nonlinear elastic constitutive model and generalized Kelvin rheological constitutive model based on the experimental data above.3. In order to solve the zone optimization for composition materials of dam body, this paper regarded the best boning point of stability and economy as optimizing object and set up a model of evolutionary neutral networks used artificial intelligent更多还原