桥梁主梁气动导纳和影响因素精细化分析的大涡模拟研究

Aerodynamic admittances are the paramount parameters to determine buffeting forces, and further to evaluate buffeting response of long-span bridges. Currently, wind tunnel test is the only way to identify them. However, many problems still exist in wind tunnel test concerning turbulence and spectrum, such as significantly small length scale, obvious deficiency at low frequency while large value at high frequency, rapid decay of turbulent intensity, ignoring turbulent cross spectrum, assuming equivalent aerodynamic admittance, etc. It should be pointed out that those problems are solvable by using suitable computational fluid dynamics methods. This may allow to improve accuracy of aerodynamic admittance, and help to develop an alternative approach to effectively obtain aerodynamic admittance.. The present research will focus on the large eddy simulation and the parallel computation. The cost-effective numerical method on large eddy simulation will be first studied, as to develop Fortran-based code spatially discretized on finite volume cells, followed by parallelization. Considering several typical forms of cross section of bridge girders, each simulation will be carried out with suitable oncoming turbulent wind flow at inlet satisfying desired wind spectrum, as well as boundary-condition consistency. The aerodynamic admittances can then be identified based on the recorded fluctuating wind and buffeting forces time histories. Wind tunnel test may help to validate and improve the CFD method. The effects on aerodynamic admittance concerning turbulent intensity, length scale and correlation of turbulent wind, length scale and coherence of buffeting forces, as well as characteristic turbulent flow will be further studied. The assessment on ignoring cross spectrum between along wind and vertical wind, as well as the method of equivalent aerodynamic admittance and cross-spectrum identification will also be carried out.

气动导纳是确定抖振力，进而计算大跨度桥梁抖振响应的关键参数，需要基于风洞试验获得。但风洞试验方法存在诸多问题，如格栅紊流积分尺度明显偏小、低频成分显著不足、高频成分过高、紊流度衰减快、忽略脉动风互谱、等效导纳处理等。这些问题可在合理的CFD模拟中克服，因而有望提高气动导纳识别精度，并开辟识别气动导纳的另一有效途径。.项目将以大涡模拟方法和并行计算为核心。研究大涡模拟的有效算法并编制基于Fortran平台的有限体积法程序，通过并行化处理，开发桥梁主梁非定常绕流模拟软件包。针对几种典型断面桥梁，通过入口满足风谱要求的随机脉动风场施加并满足自保持特性，由计算的抖振力和脉动风时程识别气动导纳。开展细致的风洞试验有助于验证并改进数值方法。研究来流湍流度、脉动风和抖振力的积分尺度和相关性、特征紊流等因素对气动导纳的影响，评价忽略脉动风互谱、等效导纳处理和互谱方法的适用性，基于流场可视化揭示微观机理。

风洞试验识别气动导纳存在如格栅紊流积分尺度明显偏小、低频成分显著不足、高频成分过高、紊流度衰减快、忽略脉动风互谱等诸多问题。但CFD模拟能部分解决这些问题，但有效的方法是基于大涡模拟识别气动导纳，这有助于获得气动导纳的精细化结果。.大涡模拟的关键问题是入口来流风场的模拟和CFD求解的数值实现。以圆柱和扁平钢箱梁为例开展了大涡模拟的网格、时间步和算法、边界条件影响、以及在并行计算机上的实现研究，研究了其气动力特征。重点研究了满足LES要求的入口紊流风场生成方法。提出采用DSRFG(Discretizing and Synthesizing Random Flow Generation)方法合成了满足目标湍流度、积分尺度、脉动风速谱及空间相关性等参数的各向异性脉动风速时程。通过自开发的UDF(User Define Functions)程序，在FLUENT中将其赋给大涡模拟的入口边界，获得了入口脉动特性在计算域内沿流向的变化规律。研究了不同网格尺寸和时间步长对湍流脉动特性的影响。.以薄平板和扁平钢箱梁为例，开展了气动导纳识别的大涡模拟研究，研究了网格尺度、时间步大小、湍流度等参数对气动导纳识别结果的影响，基于流场可视化揭示了湍流流动的微观机理。