基于基因表达式编程的矿山边坡变形预测模型研究

发布时间：2018-07-17 04:14

【摘要】：矿山边坡变形是矿山普遍会出现的现象,边坡变形一旦超出了限度,就会造成危害,甚至会形成滑坡等灾难性后果。掌握边坡变形的趋势,提前采取相应措施可以有效降低灾害的发生概率和减小灾害发生带来的损失。目前现有的预测方法都存在着一定的局限性,因此采用新的方法有效地预测矿山边坡变形有重要的意义。本论文首先介绍了矿山边坡变形预测国内外研究现状和基因表达式编程的国内外研究现状;其次阐述了矿山边坡变形监测原理及矿山边坡几种预测模型;然后利用小波阈值去噪的方法,开发了一个小波阈值去噪程序,对含有噪声的矿山边坡变形监测数据进行预处理;另外,根据基因表达式编程原理及算法流程,在Visual Studio编程环境下,利用C#编程语言设计了基于基因表达式编程矿山边坡变形预测程序;并利用德兴市某矿山边坡监测数据作为训练数据预测了德兴市某矿山边坡的变形值,同时分别利用BP神经网络预测模型和GM(1,1)灰色预测模型预测了德兴市某矿山边坡的变形值。结果表明,基于GEP预测模型得到预测值与实际值的误差为1-5mm,基于BP神经网络预测模型得到的预测值与实际值的误差为4-10mm,基于GM(1,1)预测模型得到的预测值与实际值误差为8-13mm。证明了基于基因表达式编程的矿山边坡预测模型效果更好,更适合应用于预测德兴市某矿山边坡变形,从而找到了一种预测矿山边坡变形的有效方法。
[Abstract]:Slope deformation is a common phenomenon in mines. Once the slope deformation exceeds the limit, it will cause harm and even form catastrophic consequences such as landslides. To grasp the trend of slope deformation and take corresponding measures in advance can effectively reduce the probability of disasters and reduce the losses caused by disasters. There are some limitations in the existing prediction methods, so it is very important to use new methods to predict the slope deformation effectively. This paper firstly introduces the research status of mine slope deformation prediction at home and abroad and the research status of gene expression programming at home and abroad, secondly, expounds the monitoring principle of mine slope deformation and several prediction models of mine slope. Then, a wavelet threshold de-noising program is developed to pre-process the mine slope deformation monitoring data with noise, in addition, according to the gene expression programming principle and algorithm flow, a wavelet threshold de-noising program is developed. In Visual Studio programming environment, the program of mine slope deformation prediction based on gene expression is designed by using C # programming language, and the deformation value of a mine slope in Dexing city is predicted by using the monitoring data of a mine slope in Dexing as training data. At the same time, BP neural network prediction model and GM (1t1) grey prediction model are used to predict the deformation value of a mine slope in Dexing city. The results show that the error between the predicted value and the actual value based on the GEP model is 1-5 mm, the error between the predicted value and the actual value based on the BP neural network model is 4-10 mm, and the error between the predicted value and the actual value based on GM (1K1) prediction model is 8-13 mm.. It is proved that the prediction model of mine slope based on genetic expression programming is more effective and more suitable for predicting the deformation of a mine slope in Dexing city. Thus an effective method for predicting the slope deformation of a mine is found.
【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】：TD80

【相似文献】