01
Data Analysis & Data-Driven Design
- Data-driven optimization and analysis
- Digital twin-based design
- Fault detection and diagnosis
- Prognostics and health management
The Smart System Analysis and Design Laboratory analyzes and models data collected from the geometry, materials, performance, and operation of mechanical systems — advancing fault diagnosis, prognostics, and design optimization through modern statistical and machine-learning methods.
Thank you for visiting our laboratory. Recent advances in IoT and sensor technologies have made vast quantities of operational data available across mechanical systems — yet finding engineers who combine statistics, mechanical engineering, and design optimization remains difficult for industry.
Our group bridges that gap. We work across the full pipeline: data preprocessing, big-data analysis, statistical modeling, and design optimization under uncertainty. Active collaborations include projects with the Korea Agency for Defense Development, the National Research Foundation of Korea, and ongoing research forums with Hyundai Heavy Industries, LG Electronics, Hanwha Aerospace, and Samsung Heavy Industries.
We are actively recruiting graduate students and postdoctoral researchers in mechanical engineering, industrial engineering, and statistics. If you are interested, please send your CV and a brief research statement to yoonoh@pusan.ac.kr.
Our work spans the analysis of uncertainty, the design of reliable systems under it, and the use of data to drive both. Each area informs the others.
Our research has produced applied tools across manufacturing, defense, shipbuilding, and consumer electronics. A selection of recent projects:
Motor fault classification (normal/demagnetization/inter-turn/bearing) and severity estimation for predictive maintenance.
End-to-end MLOps pipeline for acoustic fault classification of air conditioning systems with continuous learning.
Hybrid statistical and AI anomaly detection for 18 turbo air compressors deployed in shipyard operations.
Lifecycle PHM (design-development-decision making) framework for piping systems with international collaboration.
Pareto optimization across six objectives to design endurance test conditions for wheeled and tracked military vehicles.
GNN, PointNet, DeepONet models trained on transient CFD simulation data for cycle-aware temperature field prediction.
Our alumni shape engineering teams across Korea’s leading manufacturers, research institutes, and shipbuilders.
Conference participation, invited talks, and lab announcements.
Lab participation in the 16th WCSMO conference, May 18–23, 2025.
Lab presentations at the PHM-AP conference, December 8–11, 2025.
Open positions for those interested in data-driven analysis, RBDO, and design optimization.
Five-year project on lifecycle data utilization for integrated health prediction.
We welcome inquiries from prospective students, industrial partners, and visiting researchers. For graduate admissions, please include your CV, research interests, and academic transcripts.
Contact the Director