基于改进粒子群算法的边坡工程参数辨识研究

发布时间：2018-03-15 04:10

本文选题：改进粒子群算法　切入点：边坡稳定性　出处：《大连理工大学》2015年硕士论文　论文类型：学位论文

【摘要】：边坡的变形和受力状态分析的难题之一便是如何恰当的估计边坡的力学参数和初始应力场,毫无疑问,实验室测试和现场试验是解决这一问题的有效方法,但以上两种方法各有其局限性,比如边坡的非均匀特性,基于实验室内小试样的测试或局部的有限现场试验得到的边坡力学参数存在较大的随意性,并且实验结果代表性不强,数据离散,使得与实际的边坡力学参数有较大偏差,进而导致在一定程度上按照这些参数计算的理论分析结果和现场实测结果有较大的误差。反分析方法为合理确定边坡力学参数提供了一条有效的途径。伴随着计算机技术的发展,正分析的理论和计算方法逐渐成熟,观测仪器的精度也逐步提高,根据现场观测数据进行边坡力学模型参数反演具有良好的应用前景,本文主要研究成果如下：(1)分析了基本粒子群算法的代数和解析特性。(2)探讨了一种基于自主学习的改善粒子群算法,通过赋予粒子一定的自主性来改善种群的全局广度搜索与局部深度搜索能力,分析了该算法的计算效率,并通过实际测试函数验证了该算法比基本粒子群算法具有更好的寻优能力和更快的收敛速度。将该算法应用于边坡工程力学反演参数运算中,得到了比较满意的反演参数。(3)在主动学习的改进粒子群算法的基础上,将轮形结构和非线性函数调整参数权值结合起来,提出了一种全新的改进粒子群优化算法,并讨论了其收敛性。通过测试函数对其进行了函数优化对比分析,并将改进的粒子群算法应用于构建边坡工程力学参数反演问题中,结果表明该方法在边坡反演参数中是行之有效的。
[Abstract]:One of the difficult problems in slope deformation and stress analysis is how to properly estimate the mechanical parameters and initial stress field of the slope. There is no doubt that laboratory tests and field tests are effective methods to solve this problem. However, each of the two methods has its own limitations, such as the non-uniform characteristics of the slope, the mechanical parameters of the slope obtained from the test of small samples in the laboratory or the local limited field test have greater randomness, and the experimental results are not representative. The data are discrete, which leads to a big deviation from the actual mechanical parameters of the slope. Therefore, the theoretical analysis results calculated according to these parameters to a certain extent and the field measured results have a large error. The inverse analysis method provides an effective way to reasonably determine the mechanical parameters of the slope. The development of computer technology, The theory and calculation method of forward analysis is maturing gradually, and the precision of observation instrument is improved gradually. The inversion of parameters of slope mechanics model based on field observation data has a good application prospect. The main research results of this paper are as follows: (1) the algebraic and analytical properties of the basic particle swarm optimization (PSO) are analyzed. (2) an improved PSO algorithm based on autonomous learning is discussed. By giving the particle some autonomy to improve the global breadth search and local depth search ability of the population, the computational efficiency of the algorithm is analyzed. The experimental results show that the proposed algorithm has better optimization ability and faster convergence speed than the basic particle swarm optimization algorithm. The proposed algorithm is applied to the inversion of slope engineering mechanics parameters. Based on the improved particle swarm optimization algorithm, a new improved particle swarm optimization algorithm is proposed, which combines the wheel structure with the parameter weights adjusted by nonlinear function. The convergence is discussed. The function optimization is compared and analyzed by the test function, and the improved particle swarm optimization algorithm is applied to the inversion problem of the mechanical parameters of slope engineering. The results show that the method is effective in the inversion of slope parameters.
【学位授予单位】：大连理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TU43;TP18

【相似文献】