Recognition of airfoil wake structure based on deep learning

Abstract

Inspired by the perception ability of flows by organs of aquatic organisms, an identification method of airfoil wake structure based on deep learning is proposed. The wake structure of a NACA0012 airfoil at different Reynolds numbers and angles of attack collected by vertical soap film experimental device is used as the training object. A convolutional neural network is established to extract the features of wake structure of the airfoil. Bivariate identification, including the Reynolds number and angle of attack, is realized with high accuracy. Firstly, the variation law of airfoil wake structure is analyzed, and the loss function and the accuracy of verification set of three convolution neural networks with different structures are compared under the condition of small samples. The network with higher recognition ability is selected as the training model. Then, the data set is divided into training set, verification set and testing set with the ratio 3:1:1. The test set that does not participate in the training is only used for the evaluation, and the test results are expressed by a confusion matrix. The results show that the overall recognition accuracy of the model is above 92%.