PhD Student, Islamic Azad University, Iran
Shayan Ghazimoghadam is a Ph.D. student in Structural Engineering at Islamic Azad University of Shahrood, Iran, specializing in data-driven structural health monitoring. His research integrates artificial intelligence with civil engineering to develop unsupervised deep learning methods for real-time damage detection in structures. Shayan’s work focuses on creating digital twins for infrastructure assessment, aiming to enhance predictive maintenance and safety. He has authored several publications, including a notable paper on vibration-based damage diagnosis using multi-head self-attention LSTM autoencoders, published in Measurement journal. Additionally, he has presented his research at national conferences and served as a keynote speaker, demonstrating his commitment to advancing the field of structural health monitoring.
Profile
Google Scholarβ
Education
Shayan Ghazimoghadam completed his Bachelor of Science in Civil Engineering at Islamic Azad University of Gorgan, Iran, in 2018, where he excelled in design courses, achieving a GPA of 19/20 in Steel Structures and a perfect 20/20 in Concrete Structures. He then pursued a Master of Science in Structural Engineering at Lamei Gorgani Institute of Higher Education, Gorgan, graduating in 2022 with a GPA of 19.07/20. His master’s dissertation focused on structural damage identification under ambient vibration using an unsupervised deep learning method, supervised by Dr. Seyed Ali Asghar Hosseinzadeh. Currently, Shayan is a Ph.D. student at Islamic Azad University of Shahrood, Iran, where he continues to explore innovative approaches in structural health monitoring and artificial intelligence applications in civil engineering.β
ExperienceΒ
Between 2022 and 2023, Shayan Ghazimoghadam served as a Research Assistant at Golestan University, Gorgan, Iran, under the guidance of Dr. Seyed Ali Asghar Hosseinzadeh. During this period, he conducted research on real-time structural health monitoring utilizing AI-powered techniques. His work involved developing and testing unsupervised deep learning algorithms for damage detection in structures based on vibration data. Shayan’s contributions led to the presentation of his findings at national conferences, showcasing his ability to communicate complex research outcomes effectively. This experience has significantly enhanced his expertise in integrating artificial intelligence with structural engineering, positioning him as a promising researcher in the field of structural health monitoring.β
Awards and HonorsΒ
Shayan Ghazimoghadam has been recognized for his academic excellence and research contributions. He ranked first among M.Sc. students in Structural Engineering at Lamei Gorgani Institute of Higher Education, Gorgan, Iran, in 2022, achieving a GPA of 19.07/20. His innovative research on structural damage identification using unsupervised deep learning methods has been published in reputable journals, including the Measurement journal, where his paper has garnered 26 citations as of 2024. Additionally, Shayan was invited as a keynote speaker at the 3rd National Conference on Civil Engineering, Intelligent Development, and Sustainable Systems in 2023, where he presented on AI-powered structural damage identification and localization through accelerometer data. These accolades underscore his commitment to advancing the field of structural health monitoring and his potential for future contributions to civil engineering research.β
Research Focus
Shayan Ghazimoghadam’s research focuses on the integration of artificial intelligence with structural health monitoring (SHM) to develop innovative solutions for infrastructure maintenance. His primary interests include data-driven SHM, unsupervised structural damage identification, and the application of digital twins for condition assessment. Shayan aims to enhance the accuracy and efficiency of damage detection in structures by employing unsupervised deep learning techniques, particularly multi-head self-attention LSTM autoencoders. His work contributes to the development of digital twins, virtual representations of physical assets, to monitor and assess the condition of infrastructure in real time. By leveraging AI and machine learning, Shayan seeks to revolutionize traditional SHM practices, offering more proactive and predictive maintenance strategies that can lead to safer and more sustainable infrastructure systems.
Publication Top Notesβ
π 1. A Novel Unsupervised Deep Learning Approach for Vibration-Based Damage Diagnosis Using a Multi-Head Self-Attention LSTM Autoencoder
Authors: S. Ghazimoghadam, S.A.A. Hosseinzadeh
Journal: Measurement, Volume 229, Article 114410
Year: 2024
Citations (as of 2025): 26
DOI: Measurement 229, 114410 (sample link, please verify)
π Summary:
This publication introduces a novel unsupervised deep learning method for real-time structural damage detection using only ambient vibration data. The approach combines Long Short-Term Memory (LSTM) autoencoders with multi-head self-attention mechanisms, enabling the system to effectively learn temporal features and focus on critical data patterns without the need for labeled damage data.
By leveraging unsupervised learning, the model is highly adaptable and scalable, making it suitable for practical deployment in real-world structural health monitoring (SHM) scenarios. The method was validated using benchmark datasets, showing superior performance in damage localization and diagnosis accuracy compared to traditional approaches.
π 2. Transformer-Based Time-Series GAN for Data Augmentation in Bridge Monitoring Digital Twins
Authors: V. Mousavi, M. Rashidi, S. Ghazimoghadam, M. Mohammadi, B. Samali
Journal: Automation in Construction, Volume 175, Article 106208
Year: 2025 (Under Review)
DOI: Automation in Construction 175, 106208 (check for final link when published)
π Summary:
This paper presents a Transformer-based Generative Adversarial Network (GAN) for augmenting time-series sensor data in bridge monitoring systems. The technique is particularly geared towards Digital Twin models, which require large, diverse, and high-quality datasets to simulate and predict structural behavior accurately.
The GAN architecture uses a Transformer encoder to better capture temporal dependencies in structural response data, generating realistic synthetic datasets for training SHM models. By augmenting scarce or incomplete datasets, this method improves predictive performance, anomaly detection, and damage assessment capabilities of digital twins used in civil infrastructure.