Generalization enhancement of artificial neural network for turbulence closure by feature selection

Feature selection targets for selecting relevant and useful features,and is a vital challenge in turbulence modeling by machine learning methods.In this paper,a new posterior feature selection method based on validation dataset is proposed,which is an efficient and universal method for complex systems including turbulence.Different from the priori feature importance ranking of the filter method and the exhaustive search for feature subset of the wrapper method,the proposed method ranks the features according to the model performance on the validation dataset,and generates the feature subsets in the order of feature importance.Using the features from the proposed method,a black-box model is built by artificial neural network(ANN)to reproduce the behavior of Spalart-Allmaras(S-A)turbulence model for high Reynolds number(Re)airfoil flows in aeronautical engineering.The results show that compared with the model without feature selection,the generalization ability of the model after feature selection is significantly improved.To some extent,it is also demonstrated that although the feature importance can be reflected by the model parameters during the training process,artificial feature selection is still very necessary.

钝度对钝锥/楔熵层稳定性的影响研究

本文利用稳定性理论和数值模拟方法研究头部钝度对高超声速钝锥与钝楔的熵层稳定性的影响.结果表明钝楔的熵层不稳定性相比钝锥更强.随着钝度增加,熵层厚度逐渐变厚,熵层模态的不稳定性频率逐渐减小,但其不稳定区间不断向下游拓宽,表明熵层中的扰动将有更多机会影响边界层的转捩过程.数值模拟实现了熵层模态激发边界层第一模态的过程.

High-Order and High Accurate CFD Methods and Their Applications for Complex Grid Problems

The purpose of this article is to summarize our recent progress in high-order and high accurate CFD methods for flow problems with complex grids as well as to discuss the engineering prospects in using these methods.Despite the rapid development of high-order algorithms in CFD,the applications of high-order and high accurate methods on complex configurations are still limited.One of the main reasons which hinder the widely applications of thesemethods is the complexity of grids.Many aspects which can be neglected for low-order schemes must be treated carefully for high-order ones when the configurations are complex.In order to implement highorder finite difference schemes on complex multi-block grids,the geometric conservation lawand block-interface conditions are discussed.A conservativemetricmethod is applied to calculate the grid derivatives,and a characteristic-based interface condition is employed to fulfil high-order multi-block computing.The fifth-order WCNS-E-5 proposed by Deng[9,10]is applied to simulate flows with complex grids,including a double-delta wing,a transonic airplane configuration,and a hypersonic X-38 configuration.The results in this paper and the references show pleasant prospects in engineering-oriented applications of high-order schemes.

Progress in flight tests of hypersonic boundary layer transition

Boundary layer transition(BLT)can cause a sharp rise in heat flux and skin friction,which can seriously affect the flight performance and safety of hypersonic flight vehicles.Therefore,the mechanism,prediction and control of transition have become important issues that must be dealt with for the development of advanced flight vehicles,and it is also a research hotspot of particular interest to major aerospace countries.Compared to other transition research approaches,model flight tests can better present the transition problems under real flight conditions,thus have been carried out extensively over the past 30 years.The United States,Germany,France,Australia,and other countries have carried out transition research based on flight tests,such as the Pegasus wing-glove crossflow transition and the Hypersonic Boundary Layer Transition(HyBOLT)transition control flight test of the United States,the joint research project of the Hypersonic International Flight Research and Experimentation-1(HIFiRE-1)circular cone and the HIFiRE-5 elliptic cone transition flight tests between the United States and Australia,the flight test of compression surface transition of the scramjet forebody(LEA)in France and so on.Although these flight tests suffered various setbacks,they still obtained valuable transition data.Recently,the United States is carrying out the concave-surface transition flight tests of Hypersonic Boundary Layer Transition(BOLT)and BOLT-II.Since its first model flight test mission for verification purpose launched successfully in 2015,several hypersonic BLT flight tests have been conducted by China Aerodynamics Research and Development Center(CARDC).The flight tests have measured valid transition data under flight conditions,obtained the transition front and its dynamical variation on blunt cones at various angles of attack and a lifting body Hypersonic Transition Research Vehicle(HyTRV).The crossflow traveling waves in high-altitude flight were measured for the first time,and our understanding of hypersonic BLT has been greatly improved.

Design and transition characteristics of a standard model for hypersonic boundary layer transition research

To understand fundamental problems in hypersonic laminar-turbulent boundary layer transition for three-dimensional complex vehicles,a new standard model with typical lifting-body features has been proposed,named as hypersonic transition research vehicle(HyTRV).The configuration of HyTRV is fully analytical,and details of the design process are discussed in this study.The transition characteristics for HyTRV are investigated using three combined methods,i.e.,theoretical analyses,numerical simulations,and wind tunnel experiments.Results show that the fully analytic parameterization design of HyTRV can satisfy the model simplification requirements from both numerical simulations and wind tunnel experiments.Meanwhile,the flow field of HyTRV reveals typical transition mechanisms in six relatively separated regions,including the streamwise vortex instability,crossflow instability,secondary instability,and attachment-line instability.Therefore,the proposed HyTRV model is valuable for fundamental researches in hypersonic boundary layer transition.

From primary instabilities to secondary instabilities in Görtler vortex flows

We have studied the transformation process from primary instabilities to secondary instabilities with direct numerical simulations and stability theories(Spatial Biglobal and plane-marching parabolized stability equations)in detail.First Mack mode and second Mack mode are shown to be able to evolve into the sinuous mode and the varicose mode of secondary instability,respectively.Although the characteristics of second Mack mode eventually lose in the downstream due to the synchronization with the continuous spectrum,second Mack mode is found to be able to trigger a sequence of mode resonations which in turn give birth to highly unstable secondary instabilities.In contrast,first Mack mode does not involve in any mode synchronization.Nevertheless,it can still“jump”to a sinuous mode of secondary instability with a much larger growth rate than that of the first Mack mode.Therefore,secondary instabilities of Görtler vortices are highly receptive to the primary instabilities in the upstream,so that one should consider the primary instability in the upstream and the secondary instability in the downstream as a whole in order to get an accurate prediction of the boundary layer transition.

Application of surrogate models to stability analysis and transition prediction in hypersonic flows

To increase the efficiency and robustness of stability-based transition prediction in flow simulations, simplified methods are introduced to substitute direct stability analyses for rapid disturbance growth prediction. For low-speed boundary layers, these methods are mainly established based on self-similar assumptions, which are not applicable to non-similar boundary layers in hypersonic flows. The objective of this article is to investigate the application of surrogate models to stability analysis of non-similar flows over blunt cones, focused on parameterization of boundary-layer (BL) profiles. Firstly, correlations between BL edge and profile parameters are analyzed, along with self-similar flow parameters and discrete points on BL profiles, which present four groups of BL characteristic parameters. Secondly, using these parameters as inputs, surrogate models are built for disturbance growth prediction over an MF-1 blunt cone. Results show that, surrogate models using four BL edge parameters and a BL shape factor {Ue, Te, ρe, ηe, H12} for stability analysis can achieve comparable accuracy with those using 16 discrete BL profile parameters, which are more precise than those using merely self-similar parameters or BL edge parameters. Thirdly, the established surrogate models are validated by stability analysis and transition prediction over the MF-1 blunt cone in flight experiments at the instants of t = 17 s ~ 22 s. Compared with direct linear stability analyses, the mean relative error of predicted disturbance growth rates by surrogate models is 8.0% and the maximum relative error of N factor envelopes is 6.6%, which indicates feasible applications of surrogate models to stability analysis and transition prediction of non-similar boundary layers in hypersonic flows.