Finite element analysis and Krylov subspace-based model order reduction method can be used to construct massive frequency responses for a compressor blade according to various cracks, and the big data are converted to MS similarity map showing frequen...
Finite element analysis and Krylov subspace-based model order reduction method can be used to construct massive frequency responses for a compressor blade according to various cracks, and the big data are converted to MS similarity map showing frequency response similarity depending on the existing cracks. Throught deep learning trained from the images, the location and sized of blade cracks can be predicted.