Investigations of the origin of turbulent superstructures in a flat-plate boundary layer with pressure gradient

本文利用流向长积分域对无和有流向压力梯度的平板边界层湍流进行了数值研究。为了避免湍流超结构发展的可能偏差，应遵循Schlatter等人（2009）的成功方法，通过积分域内的层流-湍流过渡产生湍流。这具有独特的优点，即不符合完整Navier-Stokes方程的人工扰动量保持在最小值。模拟旨在达到尽可能大的雷诺数超过Re_Theta = 5000，其中Theta是动量厚度。在逆向压力梯度的影响下，湍流产生的最大值与雷诺数的增加类似地远离壁面。因此，应利用压力梯度"加速"流向的流动发展，以便更容易观察到大型上部结构的形成。通过选择合适的压力梯度，可以将靠近壁面和远离壁面区域的不同机制相互分离，分别和联合研究。应检查和测试用于探测和跟踪湍流上层结构的适当方法。因此，检测到的结构，然后可以比较与线性稳定性理论的本征函数和线性化的Navier-Stokes方程的Re_Theta = const的结果，这是获得的平均局部流或过滤流场。这种比较的目的是调查是否湍流超结构可以解释，建模和预测的本地不稳定性。根据该假设，湍流上层结构可能会在上游产生，然后通过流动对流到下游，应使用特征跟踪的数据进行评估。目前的直接数值模拟补充了优先计划1881“湍流上层建筑”内的实验研究。他们产生额外的数据和理解的假设检查，尽管他们的局限性较小的雷诺数比大多数实验。这使人们能够进一步深入了解优先方案实施过程中可能出现的现有假设和新假设的流程和理论研究。一旦方案结构最后确定，就有可能与从事边界层工作的所有伙伴开展合作，并打算这样做。在编写提案期间已经交换了材料。工作方案中说明了已经确定的合作。

This contribution is devoted to the numerical study of turbulence in flat-plate boundary layers without and with streamwise pressure gradients using long integration domains in streamwise flow direction. To avoid a possible bias of the development of turbulent superstructures turbulence shall be generated via laminar-turbulent transition within the integration domain, following the successful approach of Schlatter et al. (2009). This has the distinctive advantage that the amount of artificial disturbances which do not comply with the full Navier-Stokes equations is kept to a minimum. The simulations are intended to reach as large as possible Reynolds numbers beyond Re_Theta = 5000, where Theta is the momentum thickness. Under the influence of an adverse streamwise pressure gradient the maximum of turbulence production moves away from the wall similar to the increase of the Reynolds number. Therefore, the pressure gradient shall be used to "speed-up" the flow development in streamwise direction such that formation of large-scale superstructures can be observed easier. Different mechanisms in close-to-the-wall and far-from-the-wall areas can be separated from each other by a suitable choice of pressure gradient, studied separately and conjointly. Appropriate means for the detection and tracking of turbulent superstructures shall be inspected and tested. Thus, detected structures can then be compared with eigenfunctions of linear stability theory and results of linearized Navier-Stokes equations for Re_Theta=const which are obtained for the Reynolds-averaged local flow or filtered flow fields. This comparison is intended to investigate whether turbulent superstructures can be explained, modeled and predicted by local instability. The hypothesis, according to which turbulent superstructures might be generated further upstream and then convected downstream by the flow, shall be evaluated using data from feature tracking. The present direct numerical simulations complement experimental investigations within the priority programme 1881 "Turbulent Superstructures". They yield additional data and understanding for examination of hypotheses, despite their limitations to smaller Reynolds numbers than in most experiments. This allows additional insight into the flow and theoretical studies of existing as well as new hypotheses that may appear in the course of the priority programme. Cooperation with all partners who work on boundary layers is possible and intended as soon as the structure of the programme is finalized. An exchange of material during preparation of the proposals has already occurred. Already identified cooperations are described in the work programme.