基于BP神经网络的开采沉陷预计参数求取

发布时间：2018-10-21 19:53

【摘要】：煤炭资源的开采导致地表移动变形,可能引发地表塌陷、裂缝等灾害。严重影响了矿区的安全生产以及周围居民的正常生活,对区域经济的发展造成巨大的威胁。所以,开采沉陷预计模型对地表塌陷区的范围和形变量的确定有重要指导意义。本文以安阳矿区的实测资料为依托,以收集到的全国多个矿区的实测资料为基础数据,对实测数据进行分析并得出预计结果,同时对安阳矿区的概率积分法参数进行预计,文章的主要研究内容有:(1)基于对矿区沉陷的预计,本文选取概率积分法。首先介绍了概率积分法的预计参数与地质采矿条件的关系,为论文的进行奠定基础。由于矿区沉陷预计使用最为广泛的就是概率积分法,所以概率积分法参数的预计精度就决定了地表沉陷预计的精度。(2)为降低实验数据对网络的误差,使用Origin8.0对原始数据进行平滑处理。随后,分析了实验数据的不同选取方式对预计结果的影响。(3)介绍了BP神经网络模型,并运用其对参数进行预计,同时对预计结果的精度进行分析;再运用遗传算法对BP神经网络进行优化,优化后的网络对参数进行预计并分析其精度。对比两种预测结果的精度,分析结果差异产生误差的原因。(4)根据安阳矿区的实测数据,运用遗传算法优化BP神经网络的方法,对安阳矿区概率积分法的参数进行预测,然后与安阳矿区实际参数对比分析,得出结论。
[Abstract]:The mining of coal resources leads to surface movement and deformation, which may lead to surface collapse, cracks and other disasters. It seriously affects the safe production of mining area and the normal life of the surrounding residents, and poses a great threat to the development of regional economy. Therefore, the prediction model of mining subsidence is of great significance to determine the range and shape variables of surface subsidence. Based on the measured data of Anyang mining area and the measured data collected from many mining areas in China, this paper analyzes the measured data and obtains the predicted results. At the same time, the parameters of probability integration method in Anyang mining area are predicted. The main contents of this paper are as follows: (1) based on the prediction of mining subsidence, the probability integration method is selected in this paper. The relationship between the predicted parameters of the probabilistic integration method and the geological and mining conditions is introduced, which lays a foundation for the paper. As the probability integration method is the most widely used method for predicting subsidence in mining area, the precision of prediction parameters of probability integration method determines the accuracy of surface subsidence prediction. (2) in order to reduce the error of experimental data to the network, Use Origin8.0 to smooth the raw data. Then, the influence of different selection ways of experimental data on the predicted results is analyzed. (3) the BP neural network model is introduced, and the parameters are predicted, and the accuracy of the predicted results is analyzed. Then genetic algorithm is used to optimize the BP neural network. The optimized network predicts the parameters and analyzes its accuracy. Comparing the accuracy of the two kinds of prediction results and analyzing the causes of the error caused by the difference of the results. (4) according to the measured data of Anyang mining area, using genetic algorithm to optimize the BP neural network, the parameters of the probability integration method in Anyang mining area are forecasted. Then compared with the actual parameters of Anyang mining area, draw a conclusion.
【学位授予单位】：西安科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TD327;TP183

【参考文献】