On August 27, 2016, Southwest Airlines flight 3472 from New Orleans to Orlando had to perform an emergency landing when a fan blade separated from the engine hub and destroyed the cowling and punctured the fuselage. Initial findings from the metallurgical examination conducted in the National Transpiration Safety Board Materials Laboratory found that the fracture surface of the missing blade showed curving crack arrest lines consistent with fatigue crack growth. Fatigue is often cause by resonate vibrations. Modal analysis is a method that can model the natural frequencies and bending modes of turbomachinery blades. When performing modal analysis with finite element solvers like Mechanical ANSYS, images are often generated to help an engineer identify mode shapes created by nodal displacements. Manually identifying mode shapes from these generated images is an expensive task. This research proposes an automated process to identify mode shapes from gray-scale images of turbomachinery blades within a jet-engine. This work introduces mode shape identification using principal component analysis (PCA), similar to approaches in facial and other recognition tasks in computer vision. This technique calculates the projected values of potentially linearly correlated values onto P-linearly orthogonal axes, where P is the number of principal axes that define a subset space. Classification was performed using support vector machines (SVM). Using the PCA and SVM algorithm, approximately 5300 training images, representative of 16 different modes, were used to create a classifier. A test set was created with approximately 2000 unknown mode images. The classifier achieved on average 98.4% accuracy on the test set when using the bilinear Eigenface method. The accuracy was 98.6% when using the triangle interpolate Eigenface method. In addition, The results suggest that using digital images to perform mode shape identification can be achieved with better accuracy and computation performance compared to previous work. Potential generalization of this method could be applied to other engineering design and analysis applications.



College and Department

Ira A. Fulton College of Engineering and Technology; Mechanical Engineering

Date Submitted


Document Type





Modal Analysis, Computer Vision, PCA, Machine Learning