Full metadata record
| DC Field | Value | Language |
|---|---|---|
| dc.contributor | Department of Civil and Environmental Engineering | en_US |
| dc.contributor.advisor | Zhu, Songye (CEE) | en_US |
| dc.creator | Zhu, Zimo | - |
| dc.identifier.uri | https://theses.lib.polyu.edu.hk/handle/200/12984 | - |
| dc.language | English | en_US |
| dc.publisher | Hong Kong Polytechnic University | en_US |
| dc.rights | All rights reserved | en_US |
| dc.title | Dynamic response reconstruction with multi-type multi-rate sensor data fusion and unknown input | en_US |
| dcterms.abstract | Structural health monitoring (SHM) systems play a pivotal role in accurately capturing structural dynamic loads and responses and evaluating structural conditions, a fundamental prerequisite in assessing the integrity and performance of civil structures. However, for large-scale civil structures characterized by many degrees of freedom (DOFs), achieving full observations of responses at all DOFs is often impractical and unnecessary. Practical constraints in accessibility and constructability further complicate sensor installations and measurements, resulting in partially observed systems in which only the responses at the selected DOFs are measured. Addressing the challenge of reconstructing unmeasured structural responses in such partially observed systems is essential to achieving SHM objectives. Moreover, monitoring external excitation forces, particularly distributed loads, poses another formidable challenge. Consequently, the development of robust and systematic dynamic response reconstruction algorithms for partially observed structural systems under unknown inputs becomes imperative. The integration of new sensing technologies introduces additional complexities, requiring further refinement of existing response reconstruction algorithms to accommodate multi-modal sensor data with varying sampling frequencies. To solve these problems, this thesis presents a comprehensive investigation into structural dynamic response reconstruction, covering algorithm developments and experimental validations. Special attention is given to the algorithm application to wind turbine (WT) structures in the later part of this thesis. | en_US |
| dcterms.abstract | This work begins with a study of dynamic response reconstruction algorithms when the external input is known. First, a Kalman filtering (KF)-based adaptive mode selection algorithm was introduced, aimed at minimizing reconstruction errors in multiple DOF (MDOF) systems. A modal signal-to-noise ratio (MSNR) was defined as the ratio of the estimated modal response variance to the corresponding estimation error variance. Only modes with MSNR values higher than an analytically derived threshold were selected. Results from the numerical and experimental studies indicate that utilizing all vibration modes or a complete numerical model with all DOFs can lead to reduced response estimation accuracy. | en_US |
| dcterms.abstract | Furthermore, the KF-based response reconstruction algorithm was revised to incorporate computer vision (CV) technology in SHM. The multi-rate Kalman filtering (MRKF) technique with Rauch–Tung–Striebel (RTS) smoothing was proposed to fuse data from multiple sensor types sampled at varying rates. This innovative approach enables the integration of consumer-grade cameras with low sampling rates for structural displacement tracking, offering feasible alternatives for current SHM applications. The MRKF algorithm was further improved to allow the direct fusion of asynchronous sensor data. Through the presented asynchronous Kalman filtering (ASKF) algorithm, sensor data sampled randomly can be fused directly during response reconstruction, and intermittently missing sensor data could also be properly recovered. | en_US |
| dcterms.abstract | When external excitation remains unknown, response reconstruction becomes even more challenging. Existing algorithms typically require a full-rank input feedthrough matrix in observation equations, necessitating acceleration measurements at unknown input locations, which cannot be satisfied in many practical scenarios. To overcome this limitation, the recently emerging unified linear input and state estimator (ULISE) algorithm was applied to joint load and response estimation (JLRE) problems. This ULISE algorithm was validated through numerical simulations and experimental studies on shear frames, showcasing its effectiveness. It was then applied experimentally to a scaled WT model. This model was designed on the basis of the 5 MW reference WT presented by the National Renewable Energy Laboratory (NREL). The influence of blade rotation on the WT tower was evaluated through various WT working scenarios. The interface horizontal force, bending moment, and unmeasured structural responses were estimated using the presented method. | en_US |
| dcterms.abstract | To address asynchronous sensor data fusion challenges in scenarios with unknown external inputs, the ASKF algorithm was extended to the ULISE framework, resulting in the asynchronous ULISE (AS-ULISE) algorithm. RTS smoothing technique under unknown inputs has been applied in this AS-ULISE algorithm. The AS-ULISE algorithm enables the estimation of the unknown input and unmeasured state vectors from sensor data sampled at different rates. Additionally, to facilitate the application of the digital image correlation (DIC) technique in WT SHM, this AS-ULISE algorithm was designed to accommodate unsteady sampling frequencies common in DIC cameras. The AS-ULISE algorithm was validated on the 1:50 scaled WT model. This algorithm facilitated the integration of DIC cameras into the WT SHM sensing system, and the abnormal sensor data were smoothed using the RTS technique. | en_US |
| dcterms.abstract | By integrating theoretical derivations, numerical simulations, and rigorous experimental investigations, this research aims to advance the existing response reconstruction (also known as the virtual sensing) technique in the field of SHM by fusing multi-type and multi-rate sensor data. It introduces novel solutions for response reconstruction and sensor data fusion in scenarios involving both known and unknown inputs. Considering that partially observed system, asynchronous sensor data, and unknown input information are common issues in the field of SHM, the advancements presented in this thesis are of practical significance, particularly for large-scale structures located in complex environments (e.g., megawatt WTs). The developed algorithms will also pave the way toward the digital twin paradigm, aiming to replicate the responses and conditions of a physical model accurately on a virtual model. | en_US |
| dcterms.extent | xxx, 317 pages : color illustrations | en_US |
| dcterms.isPartOf | PolyU Electronic Theses | en_US |
| dcterms.issued | 2024 | en_US |
| dcterms.educationalLevel | Ph.D. | en_US |
| dcterms.educationalLevel | All Doctorate | en_US |
| dcterms.LCSH | Structural analysis (Engineering) -- Mathematics | en_US |
| dcterms.LCSH | Structural health monitoring | en_US |
| dcterms.LCSH | Hong Kong Polytechnic University -- Dissertations | en_US |
| dcterms.accessRights | open access | en_US |
Copyright Undertaking
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item:
https://theses.lib.polyu.edu.hk/handle/200/12984