Exploiting DNS in 3D Design

研究目的目前，非定常数值模拟仍然依赖于时间平均量。这意味着，即使在今天，非稳定流动仍然被从根本上稳定的角度来看待，没有现有的框架来切换到真正不稳定的角度。建立一个真实的非定常框架来描述流动是本项目的第一个目标，在此基础上，开发的非定常框架将应用于径向叶轮，以预测非定常损失，特别是来自叶尖和壁面之间的间隙(叶尖间隙)的非定常损失。这填补了两个知识空白：首先，目前已知减小径向叶轮的叶尖间隙可以减少损失，但对于固定叶尖间隙，例如如果最小叶尖间隙受到制造约束的限制，如何减少损失仍是未知的。其次，对径向叶轮性能的预测能力落后于轴流压气机。实现这一目标将在减少这种差异方面向前迈进一步。最后，从DNS数据和非定常框架出发，开发出一个降阶的径向叶轮损失模型，该模型足够快，可以用于3D设计阶段。研究方法为了实现第一个目标，将在叶轮的叶栅版本(这是几何形状的简化版本)上进行DNS，以测试不同的非定常框架。一旦选择了框架，将通过在完全运行条件下执行具有尖端间隙的全径向叶轮几何形状的DNS来实现第二个目标。利用非定常框架，非定常流动行为可以与损失联系在一起，从而可以预测径向叶轮的性能。实现第三个目标的方法可能涉及识别主导损失的径向叶轮内的流动特征，而其他的将从损失预测模型中剥离。新的工程/物理科学用于表征非定常和湍流的复杂的非定常框架开发具有固定叶顶间隙的径向叶轮的损失减少方法更高精度的径向叶轮性能预测

Research AimsCurrently, unsteady numerical studies with DNS still rely on time averaged quantities. This means that, even today, unsteady flows are still viewed from a fundamentally steady point of view with no existing framework to switch to a truly unsteady perspective. Establishing a truly unsteady framework with which to characterise flows is the first aim of this project.Following on from this, the unsteady framework developed will be applied to a radial impeller to predict the unsteady losses, especially from the gap between the blade tip and the wall (tip gap). This serves to fill two gaps in knowledge: firstly, it is currently known that reducing the tip gap in radial impellers reduces the loss, but it is still unknown how the loss can be reduced for a fixed tip gap e.g. if it minimum tip gap is limited by manufacturing constraints. Secondly, the ability to predict radial impeller performance lags behind that for axial compressors. Achieving this aim will provide a step forward in reducing this discrepancy.Finally, from the DNS data and unsteady framework, develop a reduced order model for radial impeller loss that is fast enough to be used in the 3D design stage.Research ApproachTo achieve the first aim, DNS will be performed on a cascade version of the radial impeller (this is a simpler version of the geometry) in order to test different unsteady frameworks.Once a framework has been selected, the second aim will be achieved by performing DNS of the full radial impeller geometry with tip gaps under full operating conditions. Using the unsteady framework, unsteady flow behaviour can be linked to loss which will allow prediction of radial impeller performance.Methodology for achieving the third aim will likely involve the identification of the flow features within a radial impeller which dominates loss, with others being stripped from the loss prediction model.Novel Engineering/Physical SciencesTruly unsteady framework for characterising unsteady and turbulent flowsDevelopment of loss reduction methods in radial impellers with a fixed tip gapHigher accuracy radial impeller performance prediction