基于基因表达式编程方法的大坝动态变形智能预测模型研究

项目名称： 基于基因表达式编程方法的大坝动态变形智能预测模型研究

项目编号： No.41261093

项目类型： 地区科学基金项目

立项/批准年度： 2013

项目学科： 天文学、地球科学

项目作者： 刘小生

作者单位： 江西理工大学

项目金额： 48万元

中文摘要： 为了保证大坝的正常运营，必须对大坝进行监测，分析监测数据质量及变形规律，反演探讨影响因素，研究有效预测模型，但目前的预测模型有缺陷。而基因表达式编程（GEP）是新型自适应演化算法，有极强的函数发现能力和很高的数据挖掘效率。根据大坝的特点，对大坝变形数据进行预处理；利用GEP的功能，挖掘出变形影响因素；针对单测点模型缺点，提出变形强度，利用GEP聚类算法进行变形强度的空间聚类，从而体现多测点的整体空间形变关系；基于大坝位移有限元数值计算值和原始观测值所建立的目标函数,研究大坝及岩基物理力学参数优化反演问题，确定大坝变形监控指标；进行大坝变形自动建模数值常量处理；设计大坝动态变形预测智能模型适应度函数和选择策略；建立大坝动态变形智能预测模型；设计与开发基于GEP的大坝动态变形智能预测原型系统，从而完善大坝变形预测的理论与应用体系，为大坝、边坡、建筑物、桥梁等变形体的变形预测提供决策依据。

中文关键词： 大坝变形；基因表达式编程；智能预测；模型；数据处理

英文摘要： In order to ensure the normal operation of the dam, it is necessary to monitor the dam, analyze the quality of the monitoring data and the deformation regulation, explore the influence factors by inversion and study the effective Predicting Model, but the current forecasting models have defects. However, the Gene Expression Programming is a new self-adaptive evolution algorithm that has strong ability of identifying functions and high efficiency of data mining. It can preprocess the data of the dam deformation according to the characteristics of the dam and dig out the deformation factors with the help of GEP's functions; In view of the defects of single surveying point model , it puts forward deformation strength; And make spatial clustering of deformation strength by using GEP clustering algorithm, so as to express the integral space deformation relationship of surveying points; With the objective function established base on the calculated value of the dam deformation finite element numerical and the original observation data to research the inversion problem of physical and mechanical parameters optimization for dam and batholith, then ascertain the dam deformation monitoring index; make processing of the automatic modeling numerical constants of dam deformation; and design the intelligent forecasting model

英文关键词： Dam deformation；gene expression programming；intelligent prediction；model；data processing