Kuo-I Chang

Mr. Kuo-I Chang

orcid

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

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

Representative Publications

View All Publications
19 April 2024 in Sensors
Verified
Journal Article
Open access
PDF file
GSGeorgia Stavropoulou
Georgia Stavropoulou
Konstantinos TsitseklisKonstantinos Tsitseklis
Konstantinos Tsitseklis
LMLydia Mavraidi
Lydia Mavraidi
KCKuo-I Chang
Kuo-I Chang
AZAnastasios Zafeiropoulos
Anastasios Zafeiropoulos
Symeon PapavassiliouSymeon Papavassiliou
Symeon Papavassiliou
In the highly competitive field of material manufacturing, stakeholders strive for the increased quality of the end products, reduced cost of operation, and the timely completion of their business processes. Digital twin (DT) technologies are considered major enablers that can be deployed to assist the development and effective provision of manufacturing processes. Additionally, knowledge graphs (KG) have emerged as efficient tools in the industrial domain and are able to efficiently represent data from various disciplines in a structured manner while also supporting advanced analytics. This paper proposes a solution that integrates a KG and DTs. Through this synergy, we aimed to develop highly autonomous and flexible DTs that utilize the semantic knowledge stored in the KG to better support advanced functionalities. The developed KG stores information about materials and their properties and details about the processes in which they are involved, following a flexible schema that is not domain specific. The DT comprises smaller Virtual Objects (VOs), each one acting as an abstraction of a single step of the Industrial Business Process (IBP), providing the necessary functionalities that simulate the corresponding real-world process. By executing appropriate queries to the KG, the DT can orchestrate the operation of the VOs and their physical counterparts and configure their parameters accordingly, in this way increasing its self-awareness. In this article, the architecture of such a solution is presented and its application in a real laser glass bending process is showcased.

19 July 2025 in The International Journal of Advanced Manufacturing Technology
Verified
Journal Article
Open access
PDF file
KCKuo-I Chang
Kuo-I Chang
ADAabhash Dhakal
Aabhash Dhakal
TKTorsten Kraft
Torsten Kraft
A critical challenge in manufacturing process optimization is that simulation-based surrogate predictive models often fail when confronted with real-world measurement uncertainties. In this study, we present a robust surrogate modeling approach applicable to simulation-based manufacturing process optimization, while accounting for real-world measurement uncertainties. Our methodological framework comprises the following: (1) generation of high-fidelity simulation data from finite element modeling, (2) training of diverse machine learning algorithms as surrogate models, (3) implementation of multi-objective hyperparameter optimization (MOHPO) to simultaneously optimize prediction accuracy and robustness against perturbations, and (4) quantification of real-world robustness through Monte Carlo sampling under simulated measurement uncertainties. Unlike previous predictive methods that focus solely on tuning prediction accuracy or incorporate robustness through model-specific techniques, our methodology systematically balances both objectives in a model-agnostic manner, requiring only simulation data for training. Using glass forming as a case study, we quantitatively evaluate six machine learning algorithms under temperature measurement uncertainties of ±3°C. In our experiments, multi-layer perceptrons achieve the best overall performance with mean squared error of nodal deviation < 0.2 while maintaining high robustness 0.6. Our approach generates a diverse set of Pareto-optimal solutions that allows post-training-and-optimization selection of the ideal model based on specific manufacturing requirements, eliminating the need to predefine the exact balance between accuracy and robustness before model development. This work represents a significant advancement in bridging the gap between idealized simulations and practical industrial applications by systematically accounting for measurement uncertainties in a model-agnostic manner.

16 April 2025 in Research Square
Verified
Preprint
Open access
PDF file
KCKuo-I Chang
Kuo-I Chang
ADAabhash Dhakal
Aabhash Dhakal
TKTorsten Kraft
Torsten Kraft
A critical challenge in manufacturing process optimization is that simulation-based surrogate predictive models often fail when confronted with real-world measurement uncertainties. In this study, we present a robust surrogate modeling approach applicable to simulation-based manufacturing process optimization, while accounting for real-world measurement uncertainties. Unlike previous predictive methods that focus solely on tuning prediction accuracy or incorporate robustness through model-specific techniques, our methodology simultaneously optimizes for both accuracy and robustness, requiring only simulation data for training. Using glass forming as a case study, we quantitatively evaluate six machine learning algorithms under temperature measurement uncertainties of \((\pm)\)3°C. In our experiments, Multi-layer perceptrons achieve the best overall performance with mean squared error of nodal deviation \((<)\) 0.2 while maintaining high robustness (0.6). Our approach generates a diverse set of Pareto-optimal solutions that allows post-training-and-optimization selection of the ideal model based on specific manufacturing requirements, eliminating the need to predefine the exact balance between accuracy and robustness before model development. This work represents a significant advancement in bridging the gap between idealized simulations and practical industrial applications by systematically accounting for measurement uncertainties in a model-agnostic manner.