Research on Fault Diagnosis of Industrial Air Conditioning Based on Hierarchical Fusion Self-Attention
YU Zepei
ZENG Xingjie
WANG Tao
Abstract：The failure of industrial air conditioners can cause serious impact on users,and in a practical application environ-ment,the training of artificial intelligence diagnostic models is challenging due to the occasional failure and various types of fail-ures.Based on federal learning,this paper proposes a fault diagnosis model based on hierarchical fusion self-attention,which can expand the training data without sharing the original data and reduce the model non-convergence caused by data heterogeneity.Ex-periments demonstrate that the novel fusion algorithm can handle heterogeneous data better than traditional FedAvg in most cases,with an average F1 score improvement of about 5％.
Keywords：federated learningindustrial IoTfault diagnosis
Publication Date：2025-08-20
Online Publishing Date：2025-12-12(First online date of this platform, not the publication date of the document)
Pages：6( 2162-2166,2180 )