新拌RCC单轴压缩试验的PFC模拟及细观参数反演

发布时间：2018-02-28 23:24

本文关键词： 新拌碾压混凝土单轴压缩试验 PFC模拟细观参数反演自适应差分进化算法　出处：《水力发电学报》2017年09期 　论文类型：期刊论文

【摘要】：为从细观角度研究碾压混凝土(RCC)坝料的压实机理,需确定被压RCC的细观参数,本文通过颗粒流程序(Particle Flow Code,PFC)模拟新拌RCC单轴压缩试验来反演细观参数。首先,为真实模拟实际RCC的配比和骨料级配,提出了基于筛分统计的各级颗粒数目确定方法,并采用三维扫描技术进行不规则骨料的精细建模;接着,提出了新拌RCC单轴压缩的PFC模拟方法,并建立了代替复杂PFC正分析过程的BP神经网络;最后,以单轴压缩物理试验的应力应变曲线作为标定目标,提出了基于自适应差分进化算法的参数反演方法。实例表明,反演的参数经PFC模拟所得压缩曲线与物理试验曲线吻合较好。本文所建PFC模型及反演的细观参数可为进一步研究RCC的振动压实过程与层间结合的细观机理提供基础。
[Abstract]:In order to study the compaction mechanism of RCC dam materials from a mesoscopic point of view, it is necessary to determine the meso-parameters of compacted RCC. In this paper, the mesoscopic parameters are inversed by particle Flow code (Particle Flow code) to simulate the uniaxial compression test of freshly mixed RCC. In order to simulate the actual RCC proportioning and aggregate gradation, a method based on sieving statistics is proposed to determine the number of particles at different levels, and the fine modeling of irregular aggregate is carried out by using three-dimensional scanning technique. The PFC simulation method of fresh RCC uniaxial compression is put forward, and the BP neural network instead of complex PFC forward analysis process is established. Finally, the stress-strain curve of uniaxial compression physical test is taken as the calibration target. A parameter inversion method based on adaptive differential evolution algorithm is proposed. The parameters obtained by PFC simulation are in good agreement with the physical test curves. The PFC model established in this paper and the mesoscopic parameters can provide a basis for further study on the meso-mechanism of the vibratory compaction process of RCC and the interlaminar bonding.
【作者单位】：天津大学水利工程仿真与安全国家重点实验室;交通运输部天津水运工程科学研究院;
【基金】：国家自然科学基金(51321065;51679164)
【分类号】：TV431

【相似文献】