基于综合测量的岩土坡体稳定裕度的研究
|
摘要
针对目前现有的地质滑坡的形成条件、诱发因素错综复杂,使用传统的测量手段存在实时性差,准确度低的情况,提出一种基于BP神经网络的滑坡预警模型。通过多个节点传感器综合测量得到滑坡发生的输入参数,对多组输入参数进行降维处理,并使用遗传算法调整输入层、隐含层以及输出层的权值和阈值,提高神经网络的训练精度,使得输出层的预测值更接近理想期望值。通过对多组神经网络的样本训练之后,对训练后的神经网络进行泛化,并将其投入到实际的应用场景,对地质滑坡进行预测分析。同时,搭建滑坡预警的信息采集显示系统,通过下位机多个节点采集传感器信号,传至主节点加以显示并在上位机监测软件实时显示。通过实验验证采用神经网络的智能学习算法,得到的预测结果与实际情况基本一致,验证了神经网络对于滑坡预测的切实可行性。
In view of the existing geological landslide formation conditions, the predisposing factors are intricate, the use of traditional means of measurement there is real-time poor, low accuracy. A landslide warning model based on BP neural network is proposed. Use the genetic algorithm to adjust the input layer, hidden layer and output layer weights and thresholds, and the training precision of the neural network is improved. So that the prediction value of the output layer is closer to the ideal expected value. Through the training of several groups of neural networks, the neural network after training is generalized and put into practical application scenarios to predict and analyze the geological landslide. At the same time, set up a landslide warning information collection and display system, through the next machine multiple nodes to collect sensor signals, to the main node to be displayed and real-time display in the host computer monitoring software. The experimental results show that the neural network is the same as the actual situation, and the feasibility of neural network for landslide prediction is verified;
In view of the existing geological landslide formation conditions, the predisposing factors are intricate, the use of traditional means of measurement there is real-time poor, low accuracy. A landslide warning model based on BP neural network is proposed. Use the genetic algorithm to adjust the input layer, hidden layer and output layer weights and thresholds, and the training precision of the neural network is improved. So that the prediction value of the output layer is closer to the ideal expected value. Through the training of several groups of neural networks, the neural network after training is generalized and put into practical application scenarios to predict and analyze the geological landslide. At the same time, set up a landslide warning information collection and display system, through the next machine multiple nodes to collect sensor signals, to the main node to be displayed and real-time display in the host computer monitoring software. The experimental results show that the neural network is the same as the actual situation, and the feasibility of neural network for landslide prediction is verified;
引文
[1] 郑颖人,赵尚毅,邓楚建等. 有限元极限分析法发展及其在岩土工程中的应用[J]. 中国工程科学,2006,26(8): 39-61.
[2] 贺可强,雷建和. 边坡稳定性的神经网络预测研究[J]. 地质与勘探. 2001,37(6):72-75.
[3] 葛哲学,孙志强. 神经网络理论与MATLABR2007实现[M]. 北京:电子工业出版社, 2008.
[4] 农吉夫,金龙. 基于MATLAB的主成分RBF神经网络降水预报模型[J]. 热带气象学报,2008,24(6):713-717.
[5] 韩力群,人工神经网络理论设计及应用[M].北京:化学工业出版社,2007.
[6] 程玉萍,袁志强,周博等. 遗传算法优化BP神经网络在滑坡灾害预测中的应用研究[J]. 水文地质工程地质,2012,39(1): 114-119.
[7] 黄声享,尹晖,蒋征. 变形监测数据处理[M]. 武汉: 武汉大学出版社,2003:118-123.
[2] 贺可强,雷建和. 边坡稳定性的神经网络预测研究[J]. 地质与勘探. 2001,37(6):72-75.
[3] 葛哲学,孙志强. 神经网络理论与MATLABR2007实现[M]. 北京:电子工业出版社, 2008.
[4] 农吉夫,金龙. 基于MATLAB的主成分RBF神经网络降水预报模型[J]. 热带气象学报,2008,24(6):713-717.
[5] 韩力群,人工神经网络理论设计及应用[M].北京:化学工业出版社,2007.
[6] 程玉萍,袁志强,周博等. 遗传算法优化BP神经网络在滑坡灾害预测中的应用研究[J]. 水文地质工程地质,2012,39(1): 114-119.
[7] 黄声享,尹晖,蒋征. 变形监测数据处理[M]. 武汉: 武汉大学出版社,2003:118-123.