基于改进的BP神经网络高层建筑物沉降规律分析
|
摘要
随着我国经济建设的不断发展,多层建筑物已经被高层和超高层建筑所替代。高层建筑物对单元地面所产生的压力骤然增加,建筑物自身所存在的荷载相应增加。本文主要利用数字水准仪对高层建筑H楼进行沉降观测,设置15个周期,主体施工阶段每2层观测一期数据,封顶之后观测了5期数据。取3个点作为实验分析数据,得出了沉降变化曲线。利用BP神经网络、改进BP神经网络,对沉降数据进行预测,取期间的沉降数据和期间累计沉降数据作为训练样本,根据两个沉降数据预测值的大小,选择合适的训练样本,提高预测精度。
With the need of our country' s economic construction,multi-storey buildings have been replaced by high-rise and super high-rise buildings. The pressure in the floor of a high-rise building increases suddenly,and the load of the building itself increases.In this paper,the digital level is used to observe the settlement of H building in high-rise building. The whole observation time is 15 cycles. During the main construction stage,the data are observed every 2 layers,and 5 times after the roof is observed. Take three points as the experimental data,the settlement curve,using the improved BP neural network and BP neural network,prediction of settlement data from settlement data during and cumulative settlement data as the training sample,according to the forecast value of two settlement data,select the appropriate training samples,improve the prediction accuracy.
With the need of our country' s economic construction,multi-storey buildings have been replaced by high-rise and super high-rise buildings. The pressure in the floor of a high-rise building increases suddenly,and the load of the building itself increases.In this paper,the digital level is used to observe the settlement of H building in high-rise building. The whole observation time is 15 cycles. During the main construction stage,the data are observed every 2 layers,and 5 times after the roof is observed. Take three points as the experimental data,the settlement curve,using the improved BP neural network and BP neural network,prediction of settlement data from settlement data during and cumulative settlement data as the training sample,according to the forecast value of two settlement data,select the appropriate training samples,improve the prediction accuracy.
引文
[1]黄声享,尹晖.变形监测数据处理[M].武汉:武汉大学出版社,2003.
[2]吴子安.工程建筑物变形观测数据处理[M].北京:测绘出版社,1999.
[3]王楠,张郁.超高层建筑物沉降观测方法[J].测绘地理信息,2017,42(3):109-112.
[4]孙玉强,马鑫程,袁兴明.基于SPSS曲线拟合在矿山岩移监测预测分析中应用[J].矿山测量,2017,45(3):53-56.
[5]张健雄,蒋金豹,张建霞.高层建筑沉降监测与灰色预测[J].测绘科学,2007,32(4):56-59.
[6]王银平.高层建筑沉降变形监测分析[J].江苏建筑,2014,162(3):37-39.
[7]钱思众.高层建筑物地基沉降监测与分析[J].西安科技大学学报,2014,34(3):285-289.
[8]马鑫程,孙玉强,袁兴明.小波神经网络变形监测分析中的应用[J].矿山测量,2016,44(6):44-47.
[9]王富麟.改进BP算法的灰色神经网络模型在高铁沉降预测中的应用研究[D].南昌:江西理工大学,2013.
[10]魏健.小波人工神经网络在建筑沉降预测中的应用研究[D].北京:北京交通大学,2013.
[11]姚文举.改进的GM(1,1)模型在沉降数据处理中的应用[J].北京测绘,2015(8):95-97.
[12]袁兴明,曲国庆,任婷.基于AR模型的矿山GPS监测数据分析预报[J].山东理工大学学报,2011,25(4):79-81.
[2]吴子安.工程建筑物变形观测数据处理[M].北京:测绘出版社,1999.
[3]王楠,张郁.超高层建筑物沉降观测方法[J].测绘地理信息,2017,42(3):109-112.
[4]孙玉强,马鑫程,袁兴明.基于SPSS曲线拟合在矿山岩移监测预测分析中应用[J].矿山测量,2017,45(3):53-56.
[5]张健雄,蒋金豹,张建霞.高层建筑沉降监测与灰色预测[J].测绘科学,2007,32(4):56-59.
[6]王银平.高层建筑沉降变形监测分析[J].江苏建筑,2014,162(3):37-39.
[7]钱思众.高层建筑物地基沉降监测与分析[J].西安科技大学学报,2014,34(3):285-289.
[8]马鑫程,孙玉强,袁兴明.小波神经网络变形监测分析中的应用[J].矿山测量,2016,44(6):44-47.
[9]王富麟.改进BP算法的灰色神经网络模型在高铁沉降预测中的应用研究[D].南昌:江西理工大学,2013.
[10]魏健.小波人工神经网络在建筑沉降预测中的应用研究[D].北京:北京交通大学,2013.
[11]姚文举.改进的GM(1,1)模型在沉降数据处理中的应用[J].北京测绘,2015(8):95-97.
[12]袁兴明,曲国庆,任婷.基于AR模型的矿山GPS监测数据分析预报[J].山东理工大学学报,2011,25(4):79-81.