改进基因表达式编程在深基坑变形预测中的应用研究

发布时间：2018-11-20 15:14

【摘要】：随着我国城镇化建设的深入推进,涌现出大量高层与超高层建筑物,这些建筑物对地基基础的要求也颇为严格。其中,地基基础中的深基坑就是他们的典型代表。建筑物或构筑物在使用过程中会产生一定的变形与沉降,这对建筑物使用而言非常不利。因此,对深基坑进行科学的评估与形变预测对建筑物安全使用具有重要现实意义。本文采用基因表达编程算法作为研究方法,利用其超强的发现能力和独特的算法优势,并对该算法进行进一步的改进,使其更加符合实际应用的要求,主要研究工作包括:首先,阐述了深基坑变形监测和基因表达式编程的国内外研究现状及背景意义;其次,对传统基因表达式编程算法的基本原理和实际应用中存在的缺陷进行分析;再次,利用云模型理论的基本思想并根据种群适应度的大小利用X正态云发生器选取遗传控制参数来改进传统的基因表达式模型;最后,以深基坑两个监测点的前20期观测数据作为训练样本,利用改进前后的预测模型对深基坑后5期形变数据进行预测并对其数据精度及空间分布影响进行分析。通过实例分析与比较可得知改进后的模型较传统的模型在预测精度上提高了一倍多,证明改进后的基因表达式模型在收敛的速度和预测精度方面都有所提高,从而体现该改进模型在深基坑变形预测领域的研究价值。
[Abstract]:With the further development of urbanization in China, a large number of high-rise and super-tall buildings have emerged, and the requirements of these buildings for foundation are quite strict. Among them, deep foundation pit in foundation is their typical representative. Buildings or structures in the process of use will produce certain deformation and settlement, which is very disadvantageous to the use of buildings. Therefore, the scientific evaluation and deformation prediction of deep foundation pit are of great practical significance to the safe use of buildings. In this paper, the gene expression programming algorithm is used as the research method, and its super discovery ability and unique algorithm advantages are utilized, and the algorithm is further improved to make it more suitable for practical application. The main research work includes: firstly, the research status and background significance of deep foundation pit deformation monitoring and gene expression programming at home and abroad are expounded. Secondly, the basic principle of the traditional gene expression programming algorithm and the defects in its practical application are analyzed. Thirdly, using the basic idea of cloud model theory and the size of population fitness, the genetic control parameters are selected by X normal cloud generator to improve the traditional gene expression model. Finally, using the first 20 observation data of two monitoring points of deep foundation pit as the training sample, the prediction model before and after the improvement is used to predict the deformation data of the later five periods of deep foundation pit, and the accuracy of the data and the influence of spatial distribution are analyzed. Through the analysis and comparison of examples, it can be found that the improved model is more than twice as accurate as the traditional one. It is proved that the improved gene expression model has improved both the convergence speed and the prediction accuracy. The research value of the improved model in the field of deep foundation pit deformation prediction is demonstrated.
【学位授予单位】：江西理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TU433

【参考文献】