基于影响线的中小桥梁荷载识别技术研究

发布时间：2018-10-13 10:18

【摘要】：近年来,国内多地灾难性的桥梁倒塌事故频频发生,引发了人们对桥梁结构安全问题的广泛关注。桥梁结构健康监测也以成为当今桥梁工程研究的热点,移动荷载识别技术能进一步完善健康监测系统。然而,现有的多数荷载识别方法存在不同的局限性,需要实际工程的进一步论证。分别针对单车和多车移动荷载的高精度识别,本文提出了基于影响线并考虑荷载横向分布的单车移动荷载识别方法以及基于影响线和BP神经网络的多车移动荷载识别方法,并用数值模拟和实验室车桥试验对上述方法的可行性和适用性进行了验证。本文主要研究内容及结论如下：(1)为了提高移动荷载的识别精度,提出并建立了基于影响线并考虑荷载横向分布的移动荷载识别方法,并结合数值模拟算例,对未考虑荷载横向分布的方法和考虑了荷载横向分布的方法进行了比较,结果表明未考虑荷载识别的方法不适合解决空间问题,而考虑了荷载横向分布的方法,对于车辆行驶于桥面上任意位置的空间问题,其识别精度都很高,且抗噪性能优异。(2)通过实验室车桥试验,进一步研究了考虑荷载横向分布的影响线方法在实际工程应用中的可行性。车速识别的结果显示,车速识别的误差基本能控制在±5%以内,且车速识别的精度对车重识别影响较大。车重识别的结果表明,利用考虑荷载横向分布的影响线方法进行荷载识别,车重相对误差可控制在±10%以内,且93%的样本能控制在±5%以内。(3)由于无法建立理想的数学模型对多车移动荷载进行识别,本文根据应变影响线中包含车重信息,再结合BP神经网络方法,建立了基于影响线和BP神经网络的移动荷载识别方法。然后针对该方法在实验室进行了双车移动荷载试验研究,结果表明该方法能精确识别车辆所处车道位置信息,在识别车辆荷载时,所有样本的相对误差能控制在±10%以内,且97%的样本能控制在±5%以内。
[Abstract]:In recent years, many catastrophic bridge collapse accidents occurred frequently in China, which caused widespread concern about bridge structure safety. Health monitoring of bridge structures has also become a hot topic in bridge engineering nowadays. Moving load identification technology can further improve the health monitoring system. However, most existing load identification methods have different limitations and need further demonstration of practical engineering. Aiming at the high precision identification of moving load of single vehicle and multi-vehicle, this paper presents a method of identification of moving load of single vehicle based on influence line and considering the transverse distribution of load, and a method of identification of moving load of multi-vehicle based on influence line and BP neural network. The feasibility and applicability of the above method are verified by numerical simulation and vehicle bridge test in laboratory. The main contents and conclusions of this paper are as follows: (1) in order to improve the accuracy of moving load identification, a moving load identification method based on the influence line and considering the transverse distribution of load is proposed and established. The method without considering the transverse distribution of load is compared with the method of considering the transverse distribution of load. The results show that the method without load identification is not suitable to solve the spatial problem, but the method of transverse distribution of load is taken into account. For the spatial problem of vehicle traveling on the bridge deck at any position, its recognition accuracy is very high, and the anti-noise performance is excellent. (2) through the test of the vehicle bridge in the laboratory, The feasibility of the influence line method considering the transverse distribution of load in practical engineering is further studied. The result of speed recognition shows that the error of speed recognition can be controlled within 卤5%, and the accuracy of speed recognition has great influence on the recognition of vehicle weight. The result of vehicle weight identification shows that the relative error of vehicle weight can be controlled within 卤10% by using the influence line method considering the transverse distribution of load. And 93% of the samples can be controlled within 卤5%. (3) because it is impossible to establish an ideal mathematical model to identify the moving loads of multiple vehicles, according to the information of vehicle weight contained in the strain influence line, and then combined with the BP neural network method, A moving load identification method based on influence line and BP neural network is established. The experimental results show that the method can accurately identify the lane position information of the vehicle, and the relative error of all samples can be controlled within 卤10% when the vehicle load is identified. And 97% of the samples could be controlled within 卤5%.
【学位授予单位】：东南大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U446

【参考文献】