Asst. Prof. Yuguang Fu
School of Civil and Environmental Engineering, Nanyang Technological University
Title: Noise-robust structural health monitoring leveraging advanced signal processing and artificial intelligence
Abstract:
Measurement noise is a common and critical source of uncertainty in Structural Health Monitoring, arising from the very first stage of data acquisition. It can significantly affect the accuracy and reliability of SHM applications, including displacement estimation, event detection, and damage localization. This talk presents the development of noise-robust SHM systems from three perspectives. First, reference-free displacement estimation is achieved through the design of a finite impulse response (FIR) filter using Tikhonov regularization, addressing the low-frequency noise amplification issue associated with double integration of acceleration signals. The proposed approach has been further implemented on edge devices for applications such as interstory drift estimation during earthquakes and rapid bridge assessment under train-crossing events. Second, sudden damage detection and localization are realized by integrating signal decomposition with blind source separation techniques. The method effectively identifies subtle signal discontinuities caused by structural damage while suppressing measurement noise. Laboratory experiments demonstrate its capability for effective real-time online damage detection. Third, scalable impact detection and localization are enabled through wavelet-based denoising and Bayesian information fusion. The proposed framework has demonstrated strong robustness under noisy measurement conditions and has been applied to deep-space habitat scenarios for meteoroid impact detection and protection.
Biography:
Dr. FU Yuguang is currently an Assistant Professor in the School of Civil and Environmental Engineering, Nanyang Technological University (NTU), Singapore. He received his B.Sc. and M.Sc. in civil engineering from Tongji University, China in 2012 and 2014, respectively. He earned his Ph.D. in civil engineering from the University of Illinois at Urbana-Champaign (UIUC) in 2019. Prior to joining NTU in 2021, he was a research scientist at Embedor Technologies to commercialize the sensor technology developed in the Ph.D. study, funded by NSF-SBIR. He then became a postdoctoral research associate in the Resilient Extra-Terrestrial Habitats Institute at Purdue University, funded by NASA. His research objective is to advance state-of-the-art smart IoT sensing and digital twin modeling techniques for structural health monitoring and digital construction management to make our built environment sustainable and resilient.
Dr. Tingting Sun
Institute for Risk and Reliability
Title: Physically Driven Dimension-Reduced Probability Density Evolution Equation for Efficient Stochastic Response and Dynamic Reliability Analysis of High-Dimensional Nonlinear Systems
Abstract:
Engineering structures are typically high-dimensional nonlinear stochastic systems involving complex sources of uncertainty. Accurate prediction of their probabilistic dynamic response and time-dependent reliability forms a fundamental theoretical basis for ensuring structural safety and functionality. However, this task remains highly challenging. The difficulty lies not only in the high dimensionality of the governing dynamics, but also in the strong coupling between nonlinearity and randomness, which often invalidates or severely limits classical random vibration approaches.
This talk presents the recently developed physically driven dimension-reduced probability density evolution equation (DR-PDEE) framework for efficient stochastic response and dynamic reliability analysis of high-dimensional nonlinear systems. The DR-PDEE governs the evolution of the instantaneous probability density function (PDF) of general path-continuous stochastic processes. For a specified physical quantity of interest, a corresponding low-dimensional DR-PDEE can be established, thereby effectively overcoming the curse of dimensionality. Moreover, the framework eliminates the Markovian assumptions commonly required in classical random vibration methods. It therefore provides a promising approach for determining the instantaneous PDFs and first-passage reliability of high-dimensional nonlinear dynamical systems.
The theoretical foundation and numerical implementation of the DR-PDEE will be introduced in detail. It will be shown that the framework is applicable to a broad class of stochastic dynamical systems, including essentially non-Markovian systems, systems with random parameters, systems under anticipating excitations, and systems driven by general non-white and non-stationary stochastic processes. For first-passage reliability analysis, particular attention will be given to the non-exchangeability between imposing absorbing boundaries and formulating the DR-PDEE. This issue reveals a critical theoretical constraint in probability-density-evolution-equation-based reliability analysis methods and provides important guidance for data-driven modeling methods.
Several numerical and engineering examples, including large-scale structural systems, will be presented to demonstrate the accuracy, efficiency, and applicability of the DR-PDEE framework. The talk will conclude with a discussion of future developments and potential extensions of the method for complex stochastic dynamical systems.
Biography:
Tingting Sun is a postdoctoral researcher at the College of Civil Engineering, Tongji University. She is currently a visiting researcher at the Institute for Reliability and Risk, Leibniz University Hannover. She received her Ph.D. in Engineering from Tongji University, China. Her research interests include stochastic dynamics and engineering reliability analysis, with a particular focus on advancing the theory and applications of the probability density evolution method (PDEM) and dimension-reduced probability density evolution equation (DR-PDEE) theory. She has authored or co-authored eight journal papers, including publications in Journal of Computational Physics and Computer Methods in Applied Mechanics and Engineering.
