Damage Diagnostics of Miter Gates Using Domain Adaptation and
Normalizing Flow-Based Likelihood-Free Inference
Yichao Zeng
1
, Zhao Zhao
2
, Guofeng Qian
3
, Michael D. Todd
4
, and Zhen Hu
5
1,2,5
Department of Industrial and Manufacturing Systems Engineering,
University of Michigan-Dearborn, 4901 Evergreen Rd, Dearborn, MI 48128, US
zyichao@umich.edu, zhaozhao@umich.edu, zhennhu@umich.edu
3,4
Department of Structural Engineering, University of California San Diego,
9500 Gilman Drive Mail Code 0085 La Jolla, CA 92093
g4qian@ucsd.edu,mdtodd@ucsd.edu
ABSTRACT
Miter gates are vital civil infrastructure components in inland
waterway transportation networks. To provide risk-informed
insights for decisions related to repair and maintenance, sen-
sors have been installed on some miter gates for monitor-
ing. Despite the monitoring system’s ability in collecting a
large volume of monitoring data, accurately diagnosing dam-
age state in such large structures remains challenging due to
the lack of labeled monitoring data, since these structures are
designed with high reliability and for a long operation life.
This paper addresses this challenge by proposing a damage
diagnostics approach for miter gates based on domain adap-
tation. The proposed approach consists of two main modules.
In the first module, Cycle-Consistent generative adversarial
network (CycleGAN) is employed to map monitoring data of
a miter gate of interest and other similar yet different miter
gates into the same analysis domain. Subsequently, a nor-
malizing flow-based likelihood-free inference model is con-
structed within this common domain using data from source
miter gates whose damage states are labeled from historical
inspections. The trained normalizing flow model is then used
to predict the damage state of the target miter gate based on
the translated monitoring data. A case study is presented to
demonstrate the effectiveness of the proposed method. The
results indicate that the proposed method in general can ac-
curately estimate the damage state of the target miter gate in
the presence of uncertainty.
1. INTRODUCTION
Navigational locks in inland waterway are infrastructure that
allows ship and barge traffic to move through different water
Yichao Zeng et al. This is an open-access article distributed under the terms
of the Creative Commons Attribution 3.0 United States License, which per-
mits unrestricted use, distribution, and reproduction in any medium, provided
the original author and source are credited.
elevations. One critical component in navigational locks is a
miter gate, which seals the chamber during a locking opera-
tion. Many miter gates have been in service for over their 50-
year design life, raising safety and reliability concerns (Foltz,
2017). Therefore, routine structural evaluations are crucial
for early fault detection and timely intervention. Tradition-
ally, the inspection requires expensive and labor-intensive de-
watering process for trained inspectors to get access to dif-
ferent components. Additionally, these inspection results are
often inconsistent due to inevitable human bias from inspec-
tors (Eick et al., 2018a; Wang, Huang, & Du, 2010; Vega, Hu,
Fillmore, Smith, & Todd, 2021). Recently, structural health
monitoring (SHM) have increasingly gained attention as tools
for reducing human-inspection effort in assessing structural
integrity (Estes, Frangopol, & Foltz, 2004; Nemani, Thelen,
Hu, & Daining, 2023; Eick et al., 2018b). It is crucial for life-
cycle management of structures but needs careful design and
implementation to maximize its benefits (Vistasp M. Karb-
hari, 2009).
The current SHM methods for damage diagnostics can be
broadly categorized into three groups, namely (1) data-driven
approaches; (2) physics-based approaches; and (3) hybrid
approaches that combines physics-based method with data-
driven approaches. For instance, one of the most common
used physics-based methods is Bayesian inference method
which leverages computational simulations and Bayesian
techniques to solve inverse problems (Thelen et al., 2022).
Recently, several Bayesian-based SHM approaches have been
developed for detecting damage in inland waterway infras-
tructure like miter gates (Ramancha, Vega, Conte, Todd, &
Hu, 2022; Levine, Golecki, Gomez, Eick, & Spencer, 2023;
Qian, Zeng, Hu, & Todd, 2024; Qian, Wu, Hu, & Todd, 0).
While Bayesian inference is a powerful tool for damage diag-
nostics in SHM, its effectiveness can be compromised by the
quality or availability of monitoring data for a specific miter
1
