【作者】 赵燕容；

【导师】 袁宝远；

【作者基本信息】 河海大学 ， 地质工程， 2006， 硕士

【摘要】 当前,深大基坑的应用越来越广,对于大型建筑物来说,基坑的施工影响到整个建筑物的施工安全和将来的运营。准确地进行动态预测不仅可为基坑工程的安全施工提供可靠保证,而且还能为新技术的推广和应用提供经验。通常采用仪器对基坑工程进行安全监测,并及时处理和分析监测数据。目前,常用的监测数据的处理方法有回归分析法、灰色理论分析法及神经网络方法等。近年来,时间序列分析方法因其可以处理动态数据、在短期内预测精度较高等优点,在工程实际中得到了广泛的应用。 但是,在基坑监测数据采集过程中,不可避免的会受到施工以及其它不确定因素(如天气、环境等)的影响,从而采集到的监测数据会含有一定程度的噪声,如果此时直接将采集得到的监测数据用时间序列分析方法进行分析,那么噪声的存在必将影响分析结果的精度,因此,对监测数据采用时序分析前最好能够将隐藏在信号中的噪声去除掉以达到更好的分析结果。鉴于此,论文提出了基于小波技术改进时间序列分析的方法。 小波技术具有多分辨率性,可以将一个信号分解为不同频段的信号,而信号与噪声恰好处在不同的频段,信号处在低频段,在时域与频域上是局部的;噪声处在高频段,在时域空间是全局分布,在整个监测时段内处处存在。因此,利用这一分布特征,只要选取一定的高频系数阈值对高频系数处理就可以有效地的除噪而保留信号。 论文结合润扬长江公路大桥南汊悬索桥南锚碇采用排桩冻结法新技术的工程实际,对深基坑监测数据提出了基于小波技术的时间序列分析优化方法,旨在将时间序列分析的多步预报功能与小波对信号精加工的功能相结合,对监测数据进行提纯处理。论文采用小波改进时间序列分析方法建立了时间序列动态预测模型,分析、预测基坑性态,及时修改设计方案,为基坑的安全施工提供了技术保障,为排桩冻结法新技术应用于深大基坑工程提供了支撑。更多还原

【Abstract】 At present, deep foundation pit is applied more and more fargoing. As for the large-scale building, deep foundation pit’s construction affect the whole building’s safety and exertion. The precise dynamic forecast can not only provide reliable guarantee for the deep foundation pit’s safety but also provide value experience for the technical spread and application. Generally, deep foundation pit is monitored by using instrument. In time deal with and analyse monitoring data. Nowadays, there are many methods of dealing with monitoring data, for example, regression analysis, grey theory analysis and neural network. Recently, time series analysis method is comprehensive put on. Because it can deal with dynamic data and have good precise of prediction in short-term.However, in the process of collecting data, it is inevitable that it is affected by construction and others factors(such as weather and environment). So the collected data include a certain extent noises. If analyse the monitoring data by time series directly, the noise will influence the precise of the analysis result. Therefore, before analyzing the monitoring data by time series, it is the best that wipe off the noises of the signal. So, the paper put forward the time series analysis method based on wavelet reducing-noises.Wavelet technology possess the multi-differentiated character. It can decompose the signal into different frequency signal. Signal and noises lie in different frequency apropos. Signal lie in low frequency, it is partial on time zone and frequency zone;Noises lie in high frequency, it is general on time zone. So, making use of its character, it can wipe off the noises and save the signal by using certain high frequency coefficient scope to deal with the high frequency coefficient.The paper study base on the engineering fact of the Run Yang Yangtse Bridge’s south anchor. The south anchor apply new technology of frozen row piles methods. The paper put forward new method of data processing based on time series analysis of wavelet technology. It can combine the multi-step forecasting function of time series analysis and the refined data processing function of wavelet. The paper carry out time series analysis based on wavelet technology and establish the dynamic forecast model to analyse and forecast deep foundation pit state. So it can amend the design in good time and afford reliable guarantee for construction of deep foundation pit and much support for the deep foundation pit of using the frozen row piles methods.更多还原