自动化沉降监测数据在线处理

发布时间：2019-06-27 11:30

【摘要】：随着传感器技术发展,自动化监测技术越来越得到广泛的运用。自动化沉降监测技术是目前建筑物沉降监测的主要发展方向,其中的关键问题就是数据的在线处理。所以本文旨在研究自动化沉降监测数据在线处理方法,以提高获得的沉降量的精度,为以后的工程应用提供一定的依据。首先在自动化沉降监测数据异常处理方面,为实现自动化沉降监测数据的异常在线检测建立滑动窗口,并提出基于预测模型的沉降异常检测方法,该方法降低了异常检测的误报率,从而提高了检测异常的准确性。然后通过模拟数据对该方法的异常处理的灵敏度进行了测试,并确定了合适的滑动窗口大小等参数,并对检测出的异常进行修复,保证了数据的连续性并使其满足监测数据精度。然后在自动化沉降监测数据处理中,需要将自动化沉降监测数据信号转换为对应的工程实体的沉降量,由于滑动窗口内的自动化沉降监测数据都是不同的,使用单一阈值与拟合模型并不能很好适应窗口内数据,为提高获得的沉降量的精度,提出了基于小波及拟合的数据处理方法。该方法依据窗口内数据的均方根误差与信噪比动态选取阈值与拟合模型。通过实验表明,该方法获得的沉降量不仅可以满足人工监测要求而且可以提高获得的沉降量的精度,与人工监测得到的沉降量的差值降低了大约0.1-0.4mm。最后为了实现自动化沉降监测数据的在线处理,提出了基于滑动窗口的在线处理方案,该方案通过建立缓冲区,将异常处理和基于小波及拟合的数据处理方法嵌入其中,从而对数据实现在线处理。通过实验可以得到该方案完成了自动化沉降监测数据的在线处理,得到沉降曲线与沉降速率。各监测点人工监测成果与在线处理成果差异值最大为0.9mm,最小为0.3 mm,获得的沉降量与人工监测数据基本一致。
[Abstract]:With the development of sensor technology, automatic monitoring technology is more and more widely used. Automatic settlement monitoring technology is the main development direction of building settlement monitoring at present, and the key problem is the on-line processing of data. Therefore, the purpose of this paper is to study the online processing method of automatic settlement monitoring data in order to improve the accuracy of settlement and provide a certain basis for future engineering application. Firstly, in the aspect of abnormal processing of automatic settlement monitoring data, a sliding window is established to realize the online detection of anomalies in automatic settlement monitoring data, and a settlement anomaly detection method based on prediction model is proposed, which reduces the false alarm rate of anomaly detection and improves the accuracy of anomaly detection. Then the sensitivity of the anomaly handling method is tested by the simulated data, and the appropriate sliding window size and other parameters are determined, and the detected anomalies are repaired to ensure the continuity of the data and make it meet the accuracy of the monitoring data. Then, in the automatic settlement monitoring data processing, it is necessary to convert the automatic settlement monitoring data signal into the settlement of the corresponding engineering entity. Because the automatic settlement monitoring data in the sliding window are different, the single threshold and fitting model can not adapt to the data in the window very well. In order to improve the accuracy of the settlement, a data processing method based on wavelet and fitting is proposed. In this method, the threshold and fitting model are selected dynamically according to the root mean square error and signal to noise ratio of the data in the window. The experimental results show that the settlement obtained by this method can not only meet the requirements of manual monitoring, but also improve the accuracy of the settlement, and the difference between the settlement obtained by this method and the settlement obtained by manual monitoring is reduced by about 0.1 脳 0.4 mm. Finally, in order to realize the online processing of automatic settlement monitoring data, an online processing scheme based on sliding window is proposed. By establishing buffer, the anomaly processing and the data processing method based on wavelet and fitting are embedded in the scheme, so as to realize the online processing of the data. Through experiments, the online processing of automatic settlement monitoring data can be completed, and the settlement curve and settlement rate can be obtained. The maximum difference value between the manual monitoring results and the on-line treatment results is 0.9mm, and the settlement obtained by the minimum 0.3mm mm, is basically consistent with the manual monitoring data.
【学位授予单位】：西南交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TU196.2

【参考文献】