基于基因表达式编程的大坝变形预测模型研究

发布时间：2018-03-19 01:14

本文选题：大坝　切入点：变形预测　出处：《江西理工大学》2014年硕士论文　论文类型：学位论文

【摘要】：为保障大坝的安全运营状态，需要根据已有的观测数据预测大坝未来的变形量。而目前对于大坝变形的预测方法较多，传统的预测方法存在预测过程复杂、速度慢、预测精度低等问题。基因表达式编程是在遗传算法和遗传编程的基础上发展起来的一种新型演化算法，它被广泛应用在灾害预警等预测领域。因此，非常有必要开展基于基因表达式编程的大坝变形预测研究。本文首先阐述了大坝变形预测和基因表达式编程的国内外研究现状，介绍了大坝变形监测理论和大坝预测模型；其次，根据基因表达式编程原理及算法流程，确定模型构建包括函数集与终止符集的确定、种群初始化、染色体解码、适应度评价、遗传操作等过程，利用Visual Studio平台下的C#编程语言，，完成了模型的构建工作；然后，利用C#语言设计Fibonacci加权滑动窗口法程序，对大坝监测数据进行预处理；最后，用MATLAB软件建立灰色GM(1,1)模型和BP神经网络模型，分别利用基于基因表达式编程预测模型、灰色GM(1,1)预测模型和BP神经网络模型对某大坝进行了变形预测，并分析了三种预测模型的预测结果。通过计算可知，三种模型在水平位移的平均相对误差分别为1.43%、3.85%和3.08%；在垂直位移的平均相对误差分别为1.87%、5.54%和4.83%。实验结果表明，基因表达式编程的大坝变形预测模型精度相对较高，优于灰色GM(1,1)模型和BP神经网络模型的预测精度。由此可知基于基因表达式编程的大坝变形预测模型为大坝变形预测提供了一种新的方法。
[Abstract]:In order to ensure the safe operation of the dam, it is necessary to predict the future deformation of the dam according to the existing observation data. Genetic expression programming is a new evolutionary algorithm developed on the basis of genetic algorithm and genetic programming. It is widely used in prediction fields such as disaster warning. It is necessary to study dam deformation prediction based on genetic expression programming. In this paper, the research status of dam deformation prediction and gene expression programming is introduced, and the dam deformation monitoring theory and dam prediction model are introduced. The construction of deterministic model includes the determination of function set and terminator set, population initialization, chromosome decoding, fitness evaluation, genetic operation and so on. Using C # programming language based on Visual Studio, the model is constructed. The program of Fibonacci weighted sliding window method is designed by C # language to preprocess the dam monitoring data. Finally, the grey GM1 / 1) model and BP neural network model are established by MATLAB software, and the prediction model based on genetic expression programming is used, respectively. The deformation prediction of a dam is carried out by using the grey GMX1) prediction model and the BP neural network model, and the results of three kinds of prediction models are analyzed. The average relative errors of the three models in horizontal displacement are 1.433.85% and 3.08%, respectively, and the average relative errors in vertical displacement are 1.875.54% and 4.833%, respectively. The precision of dam deformation prediction model based on gene expression programming is relatively high. The prediction accuracy of the model is better than that of the grey GMX1) model and the BP neural network model, which shows that the dam deformation prediction model based on genetic expression programming provides a new method for dam deformation prediction.
【学位授予单位】：江西理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TV698.11

【参考文献】