[SME] SEMINAR KHOA HỌC THÁNG 4

Structural Health Monitoring (SHM) and Condition Monitoring (CM) have become increasingly important across various engineering fields, including civil, mechanical, aerospace, and offshore structures. A key aspect of SHM and CM technologies is signal and data processing, which plays a crucial role in their successful implementation. Long-term monitoring of large and complex structures – such as bridges, buildings, ships, aircraft, wind turbines, and pipelines – presents the challenge of processing and interpreting vast amounts of collected data while accounting for the influence of environmental and operational variations (EOVs). When data used for damage or fault detection are collected under these changing conditions, false-positive and/or false-negative detection scenarios may occur. To address this challenge, most recently developed solutions have leveraged artificial intelligence (AI) and machine learning (ML) techniques, such as learning, classification, and adaptation. However, these solutions remain inefficient in handling the influence of EOVs. A promising alternative is to explore well-established time series analysis methods from the fields of econometrics and statistics. Despite their potential, this approach has yet to gain significant interest within the engineering community. Over the past decade, Prof. Dao has focused on developing statistics-based methods for SHM and CM. In this talk, three methods will be introduced: (1) a cointegration-based method for removing EOVs and detecting structural damage, (2) an Augmented Dickey-Fuller (ADF) test-based method for analysing data stationarity, and (3) a Wilcoxon rank-sum test-based method for monitoring operational states and enabling automated fault detection in wind turbines.