On Statistic Alignment Performance for Enhancing damage localisation across a Population of Heterogeneous shear-frame Structures

The development of machine learning algorithms for Structural Health Monitoring (SHM) is rapidly advancing. However, their application for real-world structures finds a high number of complications. One is the need for comprehensive data for training the proper algorithms. The theory of Population-based Health Monitoring (PBSHM) overcomes these challenges by sharing information between different structures. In this framework, it is necessary to understand to what extent knowledge can be shared, especially for heterogeneous datasets. This study implements a simple domain-adaptation technique based on Statistic Alignment (SA) on a population of heterogeneous structures to investigate how the performance changes because of the variations within the population. The study focusses on the numerical simulation of a population of bi-dimensional shear-frames under multiple sources of heterogeneity and damage conditions. The knowledge transfer within the population is investigated by performing damage localisation on multiple pairs of source and target domains to highlight how variations in the structures' topology, materials and geometry affect the transfer-learning and monitoring performance.