基于改进HHT的桥梁结构震后损伤识别

发布时间：2018-03-13 03:26

本文选题：桥梁结构健康监测　切入点：损伤识别　出处：《西南交通大学》2017年硕士论文　论文类型：学位论文

【摘要】：桥梁结构健康监测系统是既有桥梁结构安全运营的重要保障,损伤识别作为其中至关重要的一环,其准确与否与桥梁结构健康监测系统的效率直接相关,因此研究高精度的损伤指标具有极高的理论意义和实际应用价值。本文综述了结构损伤识别的研究现状和发展趋势,并简要分析了目前广泛应用的结构损伤指标之不足,提出了基于改进HHT的桥梁结构震后损伤识别指标,并以某斜拉桥振动台试验模型为例,对文中所提方法进行验证,所做的主要工作及结论如下:(1)针对传统线性滤波器在处理时变与非线性响应信号时的不足,采用近几年被引入到结构响应信号处理的形态学滤波器作为降噪手段,并以仿真信号为例确定了圆形结构元素作为最优结构元素,降噪结果表明形态学滤波器能在降噪的同时最大程度的保留结构响应信号。(2)介绍了三种模态混叠的改进方法——EEMD、CEEMD、AEEMD,并分析了以上三种方法的优点和局限,在此基础上结合形态学滤波器与排列熵提出了一种改进的HHT方法——PAEEMD,并以斜拉桥振动台试验模型实测动力响应对以上方法的模态混叠改进效果进行验证。结果发现EMD模态混叠问题最为严重,但其分解速度最快;EEMD(CEEMD)改进效果有限,分解速度仅次于EMD;AEEMD大幅改善了模态混叠问题,但分解速度太慢,不适合于实时桥梁结构健康监测;借助形态学滤波器的降噪前处理,PAEEMD在改善模态混叠方面略优于AEEMD,同时大幅提高了 AEEMD的分解速度,较适合于桥梁结构健康监测中损伤指标的构建。(3)将改进的HHT方法——PAEEMD应用于桥梁结构震后损伤识别,提出了 IMF特征能量比损伤指标,分别对五种工况下桥梁结构损伤状况进行评估,并与传统的频率损伤指标进行对比。结果表明,文中提出的损伤指标能有效的识别结构微弱损伤和准确的定位损伤,是一种有效的结构损伤指标。
[Abstract]:The health monitoring system of bridge structure is an important guarantee for the safety operation of existing bridge structure. Damage identification is one of the most important parts of the system, and its accuracy is directly related to the efficiency of the health monitoring system of bridge structure. Therefore, it is of great theoretical significance and practical value to study the damage index with high accuracy. This paper summarizes the research status and development trend of structural damage identification, and briefly analyzes the deficiency of structural damage index which is widely used at present. The damage identification index of bridge structure based on improved HHT is proposed, and the method proposed in this paper is verified by taking the shaking table test model of a cable-stayed bridge as an example. The main work and conclusions are as follows: (1) aiming at the shortcomings of traditional linear filters in dealing with time-varying and nonlinear response signals, the morphological filter introduced into structural response signal processing in recent years is used as a means of noise reduction. Taking the simulation signal as an example, the circular structure element is determined as the optimal structure element. The denoising results show that the morphological filter can retain the structural response signal to the maximum extent while reducing the noise. In this paper, three improved modal aliasing methods, EEMDN CEEMDN AEEMDM, are introduced, and the advantages and limitations of the above three methods are analyzed. On the basis of this, an improved HHT method, PAEEMD, is proposed by combining morphological filter and permutation entropy. The improved effect of modal mixing is verified by the measured dynamic response of the shaking table test model of cable-stayed bridge. The results show that the improved method is effective. The problem of EMD mode aliasing is the most serious. However, the speed of decomposition is limited, and the decomposition speed is second only to that of EMD-AEEMD, but the decomposition speed is too slow to be suitable for real-time bridge structure health monitoring. With the help of morphological filter, the preprocessing of noise reduction is slightly better than AEEMD in improving mode aliasing, and the decomposition speed of AEEMD is greatly improved. The improved HHT method is applied to the damage identification of bridge structure after earthquake, and the damage index of IMF characteristic energy ratio is put forward, which is more suitable for the construction of damage index in bridge structure health monitoring. The damage status of bridge structure under five working conditions is evaluated and compared with the traditional frequency damage index. The results show that the damage index proposed in this paper can effectively identify the weak damage and accurate location damage of the bridge structure. It is an effective structural damage index.
【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】：U446

【相似文献】