粒子群算法改进及内变量本构模型参数反演

发布时间：2018-06-27 06:35

本文选题：岩土工程反分析 + 本构模型　；参考：《煤田地质与勘探》2017年02期

【摘要】：为了研究深埋煤矿巷道通常存在长时间、大变形问题,拓展岩土工程反分析的手段,改善岩土工程反分析的效率和精度,首先基于自然选择、自适应变惯性权重、异步变化学习因子的策略改进了粒子群算法并完成了程序实现,通过Sphere和Rastrigrin两函数测试了改进算法的优越性;其次以Matlab软件为平台,联合大型有限元软件ABAQUS,编制了岩土反分析程序Geo PSOInverse.m;最后应用所编程序反演了以不可恢复应变为变量的、不显含时间的泥岩蠕变模型参数。结果证实:改进的粒子群算法在岩土工程参数反演计算中体现出了可靠的反演能力和很快的收敛速度,可进行复杂采矿工程的实践应用。
[Abstract]:In order to study the problem of long time and large deformation, expand the means of back analysis of geotechnical engineering and improve the efficiency and precision of back analysis of geotechnical engineering in deep buried coal mine roadway, firstly, based on natural selection, adaptive variable inertia weight is adopted. The strategy of asynchronous change learning factor improves the particle swarm optimization algorithm and completes the implementation of the program. The superiority of the improved algorithm is tested through Sphere and Rastrigrin functions. Secondly, the Matlab software is used as the platform. A geotechnical back-analysis program Geo PSOInverse.m is developed with Abaqus, and the parameters of creep model of mudstone with unrecoverable strain as variable are retrieved. The results show that the improved particle swarm optimization algorithm has a reliable inversion ability and a fast convergence rate in the inversion calculation of geotechnical engineering parameters. It can be used in the practice of complex mining engineering.
【作者单位】：中煤科工集团西安研究院有限公司;
【基金】：国家自然科学基金项目(51404295)~~
【分类号】：TD26;TP18

【相似文献】