深基坑变形监测及变形预测研究

发布时间：2018-07-16 18:56

【摘要】：本文主要较为深入的研究了深基坑工程施工过程中的变形监测及预测。以黑龙江省鸡西市万达广场深基坑工程为背景,全面总结了深基坑监测技术,验证了灰色GM(1,1)模型、神经网络模型和灰色神经网络组合模型在深基坑变形预测中的可靠性与实用性。主要研究内容及结论如下:1.介绍了深基坑变形机理及其常规变形监测技术,对变形监测系统的布设、监测数据处理等进行了探讨,介绍了用于深基坑变形预测的灰色GM(1,1)模型、BP神经网络模型、Elman神经网络模型的基本理论,分析了灰色模型和神经网络模型的串联型组合在深基坑变形预测中的应用。2.灰色GM(1,1)模型能够对变形监测数据进行合格的拟合和预测,运用灰色GM(1,1)模型进行预测时,不宜将预测的时间段设计的太长。3.训练后的BP和Elman神经网络在深基坑的变形预测中可以达到较高的拟合精度,完全符合实际工程应用的要求。通过对样本数据的预测结果进行研究分析得出,BP神经网络模型的预测精度最高。Elman神经网络的预测精度与BP神经网络接近,与应用于其他方面的预测相比,其并没有表现出预测精度方面的优越性,BP神经网络应该更适合应用于基坑变形预测方面。4.结合具体实例研究发现,灰色神经网络组合模型的预测精度要高于单一的GM(1,1)模型,这两个不同模型的相互组合可以使组合模型对其各自所具有的优点进行发挥,既可以利用神经网络的高度非线性,又可以利用累加数据的规律性及灰色模型弱化数据的随机性。5.通过对本文的几种模型预测结果进行对比分析可以得出,GM(1,1)模型、神经网络模型、灰色神经网络组合模型都能够预测出较为准确的结果,能够有效的指导基坑工程的施工。
[Abstract]:In this paper, the deformation monitoring and prediction in the construction process of deep foundation pit are studied deeply. Based on the deep foundation pit engineering of Wanda Square in Jixi City, Heilongjiang Province, the monitoring technology of deep foundation pit is summarized, and the grey GM (1Q1) model is verified. Reliability and practicability of neural network model and grey neural network combined model in deep foundation pit deformation prediction. The main contents and conclusions are as follows: 1. This paper introduces the deformation mechanism of deep foundation pit and its conventional deformation monitoring technology, and discusses the layout of deformation monitoring system, monitoring data processing and so on. This paper introduces the basic theory of Elman neural network model based on grey GM (1 + 1) model and BP neural network model for deep foundation pit deformation prediction. The application of series combination of grey model and neural network model in deep foundation pit deformation prediction is analyzed. The grey GM (1K1) model can fit and predict the deformation monitoring data. When the grey GM (1K1) model is used to predict the deformation monitoring data, it is not appropriate to design the predicted time period for too long. 3. The trained BP and Elman neural networks can achieve high fitting accuracy in the prediction of deep foundation pit deformation, which fully meet the requirements of practical engineering application. By studying and analyzing the prediction results of sample data, it is concluded that the prediction accuracy of BP neural network model is the highest. Elman neural network is close to BP neural network, and compared with the prediction applied in other aspects. It does not show the superiority of prediction accuracy. BP neural network should be more suitable for foundation pit deformation prediction. It is found that the prediction accuracy of the combined grey neural network model is higher than that of the single GM (1K1) model, and the combination of these two different models can make the combined model give full play to their respective advantages. It can be used not only to make use of the high nonlinearity of the neural network, but also to use the regularity of the accumulated data and the weakening of the randomness of the data by the grey model. By comparing and analyzing the prediction results of several models in this paper, we can conclude that GM (1t1) model, neural network model and grey neural network combination model can all predict more accurate results and can effectively guide the construction of foundation pit engineering.
【学位授予单位】：长安大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TU196.1

【参考文献】