Intelligent Crack Damage Identification for Compressor Blade Based on Transfer Graph Neural Network

Advanced sensing techniques and intelligent algorithms have greatly contributed to the development of structural health monitoring & damage identifications. As the most common-used electrical equipment in modern industry, compressors play an irreplaceable role in petrochemical and other process industries, usually accompanied by complex operation conditions such as variable speed. Its weak and vulnerable components, blades, will inevitably suffer varying degrees of damage in service. This study introduces the graph structure-based decision-making into the compressor's blade crack damage identification under variable operating conditions and thereby proposes the transfer graph neural network enhanced damage identification (TGNNeDI) method. Firstly, damage signal-based affinity graphs are constructed to express the latent damage information beyond Euclidean space. Then the developed multi-order graph convolutional network and domain discriminator layers are used to extract the domain-invariant damage features, which will be fed into the linear layer for class prediction. Finally, real-measured experiments under the compressor experimental platform are utilized to cross-verify the established TGNNeDI method under variable rotating speed conditions.