No abstract

No abstract

Data-driven machine learning models, compared to numerical models, shown promising improvements in detecting damage in Structural Health Monitoring (SHM) applications. In data-driven approaches, sensors' data are used to train a model either in a centralized server or locally inside each sensor unit node (decentralized model similar to edge computing). The centralize learning model suffers from issues including wireless transmission costs and data sensitive data vulnerability. The decentralized model also poses different challenges such as feature correlations and relationships loss in decentralized learning. To handle the shortcomings of both models, we proposes a new Federated Learning model augmented with tensor data fusion to detect damage in SHM. Our approach enables the central machine learning model to gain experience from diverse datasets located at different sensor locations. It also trains a shared centralized machine learning model using datasets stored and distributed across multiple sensor nodes. Our experimental results on real structural datasets demonstrate promising damage detection accuracy without the need to transmit the actual data to the centralized learning model. It also shows that the data correlations and relationship from all participating sensors are preserved.

No abstract

No abstract