Study On Dynamic Prediction Of Foundation Pit Monitoring Time Series Analysis Methods Based On Wavelet Technology

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.