桥梁健康监测系统数据处理与分析技术研究

发布时间：2018-03-02 21:22

本文选题：桥梁健康监测　切入点：预处理　出处：《重庆大学》2015年硕士论文　论文类型：学位论文

【摘要】：几十年来,随着中国的经济飞速发展,桥梁的建造数量越来越大,但桥梁的坍塌事件也是时有发生,因此对桥梁的安全极为重视,桥梁健康监测系统也应运而生,并已成为国内外学者的研究热点。桥梁健康监测系统通过安装的传感器获取桥梁的各种状态,并据此进行安全评估,因此传感器数据的正确性是监测系统的基础和关键。为此需要对传感器数据的可信度进行评价,检测并剔除异常数据,以及对剔除数据进行插补。虽然桥梁健康监测技术经过多年的发展取得了较大的进展,监测数据的采集、存储技术也日臻完善。但现有的数据可信度评估的研究还不多,而目前针对桥梁健康监测数据异常检测方法也存在着一些问题,数据插补技术还需要做针对性改进。同时监测数据中还有很多隐含待挖掘的信息,时间序列分析和数据挖掘技术已经成功应用在这上面,但还存在一些问题需要进一步研究。本文研究了桥梁健康监测数据的预处理和分析,主要针对数据可信度评估、数据插补技术、监测数据的时间序列分析。对于数据的可信度评估,提出了基于灰色关联度分析的传感器数据可信度评估方法,并用灰色关联度分析定位不可信数据的时间段,为桥梁结构预警提供依据。针对监测数据的异常值,采用基于统计特征的方法对异常值进行剔除。为了对剔除异常值的数据进行插补,提出了基于灰色关联度分析的预测模型与时间序列ARMA模型预测的结合方案。长期的桥梁监测数据的趋势反映了桥梁的健康状况的变化趋势,本文以多种时间角度的对监测数据的整体发展趋势进行了观察分析,并利用ARMA模型对其建立了最佳的拟合模型,利用此模型对桥梁进行预警。并利用此预测模型结合基于关联度分析的预测模型对异常值做了预测分析实验,表明基于灰色关联度分析的预测模型与时间序列ARMA模型结合的预测方法的预测效果比单独ARMA模型预测效果好。本文的主要创新点是把灰色关联度理论应用于数据的预测上,并把这种预测方法与时间序列模型预测结合起来,消除了时间序列模型预测的误差。最后通过实验验证了论文所提出的数据处理和分析方法,结果符合预期要求,并且已成功应用于多座桥梁监测系统中。
[Abstract]:In recent decades, with the rapid development of Chinese economy, the number of bridges has been increasing, but the collapse of bridges has occurred from time to time. Therefore, the safety of bridges has been attached great importance, and the bridge health monitoring system has emerged as the times require. The bridge health monitoring system obtains the various states of the bridge through the sensor installed, and carries on the safety evaluation accordingly. Therefore, the correctness of sensor data is the basis and key of the monitoring system. Therefore, it is necessary to evaluate the credibility of sensor data, detect and eliminate abnormal data. Although the bridge health monitoring technology has made great progress after years of development, the monitoring data collection and storage technology are also improving, but the existing data credibility evaluation research is still few. At present, there are some problems in the anomaly detection method of bridge health monitoring data, and the data interpolation technology needs to be improved. At the same time, there are many hidden information to be mined in the monitoring data. Time series analysis and data mining techniques have been successfully applied in this field, but there are still some problems that need to be further studied. In this paper, the preprocessing and analysis of bridge health monitoring data is studied, which is mainly aimed at the reliability evaluation of bridge health monitoring data. Data interpolation technology, time series analysis of monitoring data. For the reliability evaluation of data, a method of reliability evaluation of sensor data based on grey correlation analysis is proposed, and the grey correlation analysis is used to locate the time period of untrusted data. The method based on statistical features is used to eliminate the outliers. In order to interpolate the outliers, the outliers of the monitoring data are interpolated. This paper puts forward a scheme of combining prediction model based on grey correlation analysis with ARMA model of time series. The trend of long-term bridge monitoring data reflects the changing trend of bridge health status. In this paper, the overall development trend of monitoring data is observed and analyzed from various time angles, and the best fitting model is established by using ARMA model. The model is used to predict the bridge, and the prediction model based on the correlation analysis is used to predict and analyze the outliers. The results show that the prediction method based on grey correlation degree analysis and time series ARMA model has better prediction effect than that of single ARMA model. The main innovation of this paper is to apply the grey correlation degree theory to the prediction of data. Combining this method with time series model prediction, the error of time series model prediction is eliminated. Finally, the data processing and analysis method proposed in this paper is verified by experiments, and the results meet the expected requirements. And has been successfully applied to many bridge monitoring systems.
【学位授予单位】：重庆大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U446

【相似文献】