Shiyin Wei

Prof. Shiyin Wei

School of Civil Engineering, Harbin Institute of Technology, Harbin 150090, China
orcid

Research Focus & Metrics

Academic Career Length

5Years (2022-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
26 April 2023 in Applied Sciences
Verified
Journal Article
Open access
PDF file
FDFan Deng
Fan Deng
Xiaoming TaoXiaoming Tao
Xiaoming Tao
PWPengxiang Wei
Pengxiang Wei
Shiyin WeiShiyin Wei
Shiyin Wei
Data-driven methods have shown promising results in structural health monitoring (SHM) applications. However, most of these approaches rely on the ideal dataset assumption and do not account for missing data, which can significantly impact their real-world performance. Missing data is a frequently encountered issue in time series data, which hinders standardized data mining and downstream tasks such as damage identification and condition assessment. While imputation approaches based on spatiotemporal relations among monitoring data have been proposed to handle this issue, they do not provide additional helpful information for downstream tasks. This paper proposes a robust deep learning-based method that unifies missing data imputation and damage identification tasks into a single framework. The proposed approach is based on a long short-term memory (LSTM) structured autoencoder (AE) framework, and missing data is simulated using the dropout mechanism by randomly dropping the input channels. Reconstruction errors serve as the loss function and damage indicator. The proposed method is validated using the quasi-static response (cable tension) of a cable-stayed bridge released in the 1st IPC-SHM, and results show that missing data imputation and damage identification can be effectively integrated into the proposed unified framework.

11 November 2023 in Engineering Structures
Verified
Journal Article
Closed access
FDFan Deng
Fan Deng
Shiyin WeiShiyin Wei
Shiyin Wei
Yang XuYang Xu
Yang Xu
HLHui Li
Hui Li

22 May 2025 in Engineering Structures
Verified
Journal Article
Closed access
QZQunhong Zhong
Qunhong Zhong
Shiyin WeiShiyin Wei
Shiyin Wei
YHYong Huang
Yong Huang
KJKailiang Jia
Kailiang Jia
HLHui Li
Hui Li

2 June 2022 in Structural Health Monitoring
Verified
Journal Article
Closed access
XLXinyi Lei
Xinyi Lei
ZCZhicheng Chen
Zhicheng Chen
HLHui Li
Hui Li
Shiyin WeiShiyin Wei
Shiyin Wei
Damage detection or structural condition assessment is an important objective of structural health monitoring (SHM). The damages or adverse changes in structural conditions can usually be manifested as pattern changes in damage-sensitive features (DSFs) extracted from SHM data; this enables us to shift damage detection to DSF change detection. Online monitoring can accumulate huge amounts of data, finding the changes from the massive DSF data through manual inspection is impractical; thus, automatic detection tools are required. If possible, relevant significance test is also desired to make a rational judgment on the existence of a change. In this sense, the change-point detection technique is an attractive choice, which is increasingly proved to be a powerful change detection tool in various SHM applications. However, existing change-point detection methods in SHM are mainly used for scalar or vector data, thus incapable of detecting changes in features represented by complex data, for example, the probability density functions (PDFs). Detecting abrupt changes in the distributions (represented by PDFs) of the DSF data is of crucial concern in structural condition assessment. However, relevant automatic diagnostic tools have not been well developed in the SHM community. To this end, a novel change-point detection method is developed in the functional data analytic framework for this task. The proposed approach has advantages in detecting changes for massive data and directly handling general PDFs. Considering that the major challenge in PDF-valued data analysis comes from the nonlinear constraints of PDFs, the PDFs are embedded into the Bayes space to develop the detection methodology by using the linear structure of the Bayes space. Comprehensive simulation studies are conducted to validate the effectiveness of the proposed method as well as demonstrate its superiority over the competing method. Finally, a case study involving cable condition assessment of a long-span bridge demonstrates its practical utility in SHM.

8 April 2023 in Mechanical Systems and Signal Processing
Verified
Journal Article
Closed access
XLXinyi Lei
Xinyi Lei
ZCZhicheng Chen
Zhicheng Chen
HLHui Li
Hui Li
Shiyin WeiShiyin Wei
Shiyin Wei

26 July 2025 in Structural Health Monitoring
Verified
Journal Article
Closed access
YZYunkai Zhang
Yunkai Zhang
Shiyin WeiShiyin Wei
Shiyin Wei
Yong HuangYong Huang
Yong Huang
YSYawu Su
Yawu Su
SLShanshan Lu
Shanshan Lu
HLHui Li
Hui Li
Existing computer vision-based structural damage identification models demonstrate notable accuracy in categorizing and localizing damage. However, these models present several critical limitations that hinder their practical application in structural health monitoring. Primarily, their ability to recognize damage types remains constrained, preventing comprehensive analysis of the highly varied and complex conditions encountered in real-world structures. Second, these models lack linguistic capabilities, rendering them unable to articulate structural damage characteristics through natural language descriptions. With the continuous advancement of artificial intelligence, multimodal large language models (MLLMs) have emerged as a transformative solution, enabling the unified encoding and alignment of textual and visual data. These models can generate detailed Structural Damage Image descriptions while demonstrating robust generalization across diverse scenarios and tasks. This study introduces General Language Model for Structural Damage Identification (SDIGLM), an innovative MLLM for Structural Damage Identification, developed based on the open-source VisualGLM-6B architecture. To address the challenge of adapting MLLMs to the intricate and varied operating conditions in civil engineering, this work integrates a U-Net-based semantic segmentation module to generate defect segmentation maps as visual chain of thought (CoT). Additionally, a multiround dialogue fine-tuning dataset is constructed to enhance logical reasoning, complemented by a language CoT formed through prompt engineering. By leveraging this multimodal CoT, SDIGLM surpasses general-purpose MLLMs in structural damage identification, achieving an accuracy of 95.24% across various infrastructure types. Moreover, the model effectively describes damage characteristics such as hole size, crack direction, and corrosion severity.