基于HyperFLOW平台的客机标模CHN-T1气动性能预测及可信度研究

近年来,针对实际飞行器外形的CFD气动性能预测及可信度研究逐步得到重视,国内也召开了第一届航空CFD可信度研讨会(AeCW-1)。本文首先基于自主研发的CFD软件平台HyperFLOW对NACA0012翼型低速绕流进行了网格收敛性研究,验证了软件对简单湍流问题的模拟能力且具备良好的网格收敛性。其次,针对AeCW-1提供的客机标模CHN-T1,选用其中的两个算例:(1)定升力系数的网格收敛性研究;(2)考虑模型支撑和模型静气动弹性变形的抖振特性研究,研究了计算结果的网格收敛性及模型支撑、静气动弹性变形和湍流模型等对气动特性预测精度的影响。结果表明:观测精度阶和网格收敛性指数显示数值结果具有良好的网格收敛性和可信度;是否考虑模型支撑对力矩的预测精度影响较大,引入尾撑和弹性变形后,数值结果与实验结果吻合较好;对于CHN-T1标模,采用QCR关系式对原始SA模型进行修正对标模力矩特性有一定影响。

基于小波尺度函数的WSK-SV算法及其气动性能预测

提出了一种将小波的尺度函数与SV(support vector)算法相结合的WSK-SV(wavelet scalingkernel-support vector)新算法,并将Daubechies小波以及Shannon小波的尺度函数分别构成尺度核函数,而且分别作为SV算法中一个可容许的支持向量核函数使用.该算法充分利用了Daubechies小波函数的紧支集与正交等特点以及小波的MRA(multi-resolution analysis,多分辨分析),并注意了尺度核函数能够满足Mercer条件.该算法除了具有通常SVM(support vector machine)所具有的优点外,还具有很好的收敛性以及泛化能力,能够有效地提高学习与预测效率.典型算例选取了不同的小波尺度函数,数值计算表明:在一维、二维和三维问题中,这些小波的尺度函数均可以用于WSK-SV算法,进而显示了这个新算法的可行性与通用性.

基于深度度量学习的导弹气动系数预测

传统多输出深度神经网络在导弹气动性能系数预测任务中,通常采用均方误差(Mean square error,MSE)和平均绝对误差(Mean absolute error,MAE)来训练网络,但在小样本及无物理方程约束的情况下,MSE与MAE对导弹性能系数之间的约束和不同导弹样本之间的区分就会降低。针对该问题,提出一种基于深度度量学习的K最近邻大边距损失函数(K-nearest neighbor large margin,KNNLM),它通过边距约束将大差异输出样本推开,拉近相近输出样本,以此来解决样本及样本间的约束区分问题。以导弹气动外形及工况参数作为输入,4种气动系数作为输出,在反向传播神经网络(Backpropagation neural network,BPNN)和多任务学习神经网络(Multi-task learning neural network,MTLNN)中分别采用MSE、MAE、KNNLM进行实验对比,实验结果表明:KNNLM在BPNN和MTLNN中的精度相比于MSE和MAE最大能够提升14.44%和16.35%,最少提升3.72%。KNNLM能够在少样本及无物理知识约束的情况下,能更好地对导弹样本进行约束区分,使深度神经网络模型的预测精度更高,且鲁棒性更强。

飞机NACA进气口性能的数据驱动预测模型研究

冲压空气进气口是民机环控系统性能实现的关键。针对飞机不同结构参数的NACA进气口,提供三种数据驱动的预测模型,对其性能进行预测。分析多元线性回归模型、二阶多项式模型、人工神经网络三种预测模型的原理,建立相应的数学模型,并在MATLAB实现三种预测模型的代码编译。以某一型号客机为例,进行ANSYS CFX流场仿真,获取不同结构参数对应性能指标的数据库,基于数据库对比不同预测模型下的误差,并分析出适用于飞机NACA进气口不同结构参数性能预测的模型。对较优的BP神经网络模型进行改进,得到更加适合的BP神经网络改进模型。

Prediction of Wing Aerodynamic Performance in Rain Using Neural Net

A new method for prediction of wing aerodynamic performance in rain condition was presented.Three-and four-layer artificial neural networks based on improved algorithm for error Back Propagation(BP)network were respectively built.Detailed approaches to determine the optical parameters for network model were introduced and the specific steps for applying BP network model to predict wing aerodynamic performance in rain were given.On this basis,the established optimal three-and four-layer BP network model was used for this prediction.Results indicate that both of the network models are appropriate for predicting wing aerodynamic performance in rain.The sum of square error level produced by two models is less than 0.2%,and the prediction accuracy by four-layer network model is higher than that of three-layer network.

Numerical and Experimental Study on Aerodynamic Performance of Small Axial Flow Fan with Splitter Blades

To improve the aerodynamic performance of small axial flow fan, in this paper the design of a small axial flow fan with splitter blades is studied. The RNG k-e turbulence model and SIMPLE algorithm were applied to the steady simulation calculation of the flow field, and its result was used as the initial field of the large eddy simulation to calculate the unsteady pressure field. The FW-H noise model was adopted to predict aerodynamic noise in the six monitoring points. Fast Fourier transform algorithm was applied to process the pressure signal. Experiment of noise testing was done to further investigate the aerodynamic noise of fans. And then the results obtained from the numerical simulation and experiment were described and analyzed. The results show that the static characteristics of small axial fan with splitter blades are similar with the prototype fan, and the static characteristics are improved within a certain range of flux. The power spectral density at the six monitoring points of small axial flow fan with splitter blades have decreased to some extent. The experimental results show sound pressure level of new fan has reduced in most frequency bands by comparing with prototype fan. The research results will provide a proof for parameter optimization and noise prediction of small axial flow fans with high performance.

Modification of k-ω Turbulence Model for Predicting Airfoil Aerodynamic Performance

Predicting wind turbine S825 airfoil＇s aerodynamic performance is crucial to improving its energy efficiency and reducing its environmental impact. In this paper, a numerical simulation on the wind turbine S825 airfoil is con- ducted with k-to turbulence model at different attack angles. By comparing with experimental data, a new method of modifying k-to model is proposed. A modifying function is proposed to limit the production term in ω equation based on fluid rotation and deformation. This method improves turbulent viscosity and decreases separating re- gion when the airfoil works at large separating conditions. The predictive accuracy could be improved by using the modified k-to turbulence model.