滤波技术在沈阳地铁变形监测数据处理中的应用研究

发布时间：2018-04-15 21:38

本文选题：变形监测 + 自动化　；参考：《沈阳建筑大学》2014年硕士论文

【摘要】：本文以沈阳中街人防工程为背景,对人防工程与地铁重叠段进行变形监测。因地铁处于运营阶段,采用自动化监测系统——静力水准监测系统进行实时监测。但是该系统输出的数据,与人工二等水准测量结果相比,数据跳动较大,存在17%——26%的误差,需对监测数据进行去噪处理。本文在VS2010下编写滤波软件,首先采用经典卡尔曼滤波、自适应卡尔曼滤波和小波滤波三种方法分别对监测数据进行去噪处理,通过实验结果发现经典卡尔曼滤波能够有效剔除掉的误差为38%,自适应卡尔曼滤波能够有效剔除掉的误差为55%,小波滤波能够有效剔除掉的误差为52%,三种方法单独使用时都不能满足剔除60%以上误差的要求。针对这一问题,本文提出使用滤波模型组合技术,根据单独使用一种滤波方法时发现的特点,即：1、经典卡尔曼滤波不适合长期使用；2、自适应卡尔曼滤波在低频部分有明显优势；3、小波滤波在高频部分有降噪效果明显。进行滤波模型组合。根据滤波先后顺序确定了四种滤波模型组合BCa1、BCa2、CalB和Ca2B,并由理论知识推测BCal降噪效果最好。通过工程实例的验证,滤波模型组合方法在降噪效果方面优于单独使用一种滤波技术,滤波模型组合方法能够有效剔除掉65%——75%的误差。其中滤波模型组合BCal降噪效果最好,能够有效剔除掉75%的误差,大大提高了自动化监测数据的准确性和可靠性,达到了预期有效剔除60%的误差的目的,得出处理中街人防工程地铁自动化监测数据的最优方法是滤波模型组合BCal。
[Abstract]:Based on the civil air defense engineering of Shenyang Central Street, deformation monitoring of overlapped section between civil air defense engineering and subway is carried out.Because the subway is in operation stage, the automatic monitoring system-static leveling monitoring system is used for real-time monitoring.However, the output data of the system is larger than the result of manual secondary leveling, and the error of 17% to 26% exists, so it is necessary to Denoise the monitoring data.In this paper, the filtering software is written under VS2010. Firstly, three methods of classical Kalman filter, adaptive Kalman filter and wavelet filter are used to Denoise the monitoring data.The experimental results show that the classical Kalman filter can effectively eliminate the error of 38 parts, the adaptive Kalman filter can effectively eliminate the error of 55 steps, the wavelet filter can effectively eliminate the error of 52 parts.The time can not meet the requirement of eliminating more than 60% error.In order to solve this problem, this paper proposes to use filter model combination technique, according to the characteristics found when a single filtering method is used.That is, the classical Kalman filter is not suitable for long-term use, adaptive Kalman filter has obvious advantages in the low frequency part and wavelet filter has obvious noise reduction effect in the high frequency part.Filter model combination is carried out.According to the sequence of filtering, four filtering models, BCA _ 1C _ (2) Ca _ (2) CalB and Ca _ (2) B _ (2) B, are determined, and it is inferred from the theoretical knowledge that BCal has the best effect on noise reduction.Through the verification of engineering examples, the filter model combination method is superior to one kind of filtering technique in noise reduction effect, and the filter model combination method can effectively eliminate the error of 65% to 75%.The filter model combined with BCal has the best denoising effect and can eliminate 75% of the error effectively, which greatly improves the accuracy and reliability of the automatic monitoring data, and achieves the goal of eliminating 60% of the expected error effectively.It is concluded that the optimal method to deal with the subway automatic monitoring data of civil air defense engineering in the middle street is the combination of filter model and BCal.
【学位授予单位】：沈阳建筑大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U456.3;U231.3

【相似文献】