publications
2024
2023
2022
2021
- Homomorphic Encryption as a secure PHM outsourcing solution for small and medium manufacturing enterpriseHa Eun David Kang, Duhyeong Kim, Sangwoon Kim, David Donghyun Kim, Jung Hee Cheon, and Brian W AnthonyJournal of Manufacturing Systems 2021
Small and medium manufacturing enterprises (SMEs) often lack skills and resources required to perform in-house PHM analytics. While cloud-based services provide SMEs the option to outsource PHM analytics in the cloud, a critical limiting factor to such arrangement is the data owner’s unwillingness to share data due to data privacy concerns. In this paper, we showcase how homomorphic encryption, a cryptographic technique that allows direct computation on encrypted data, can enable a secure PHM outsourcing with high precision for SMEs. We first outline a two-party collaborative framework for a secure outsourcing of PHM analytics for SMEs. Next, we introduce a frequency-based peak detection algorithm (H-FFT-C) that generates a machine health diagnosis and prescription report, while keeping the machine data private. We demonstrate the secure PHM outsourcing scenario on a lab-scale fiber extrusion device. Our demonstration is comprised of key functionalities found in many PHM applications. Finally, the extensibility and limitation of the approach used in this study is summarized.
- Factory 4.0 Toolkit for Smart Manufacturing TrainingJoseph Dennis Cuiffi, Haifeng Wang, Josephine Heim, Brian W Anthony, Sangwoon Kim, and David Donghyun KimIn ASEE Virtual Annual Conference Content Access 2021
The rapid pace of technology development in the field of smart manufacturing has left educational systems scrambling to keep pace and adapt learning outcomes, resulting in inadequate preparedness and readiness of workforce at all levels. Often, smart manufacturing training materials are either broad and conceptual or a specific technical deep dive with little context. We have developed an educational toolkit that leverages an inexpensive, bench scale extrusion platform to provide lab activities and feature-rich data to explore fundamental concepts of smart manufacturing in a production context for an audience of both undergraduate engineering students and current manufacturing workforce members. Through investigation of the mock production platform and associated data, concepts and applications of modern datadriven tools are explored in the topic areas of data collection and the industrial internet of things, data analytics and predictive modeling for production data, simulation and digital twinning, and process and manufacturing systems optimization. The activities culminate in the exploration of advanced feedback control algorithms and optimization of operating conditions, balancing throughput, quality, and power consumption, using digital twins. The combination of overview conceptual materials along with in-depth activities on an actual process allows us to tailor the scope of the specific training to the intended audience. Select modules of the Factory 4.0 toolkit were delivered in an undergraduate course and in a training workshop for manufacturing personnel. Pre- and post-attitude surveys, along with participant comments, were used to assess the training approach and content. We found that the proper technical scope is critical for a given audience and that all types of manufacturing personnel, from technicians and engineers to operations and management, benefit from foundational smart manufacturing concepts and examples. We also found that for technical materials, student audiences required more of the fundamental instrumentation and statistical analysis topics, while current technical practitioners desired specific deep dives into data analytics, digital twinning, and process optimization after introductory overviews. Both educational experiences exposed a need for preparedness in programming and statistical analysis software tools to take advantage of these smart manufacturing concepts.
2020
- Model-free tracking control of an optical fiber drawing process using deep reinforcement learningSangwoon KimMIT M.S. Thesis 2020
A deep reinforcement learning (DRL) approach for tracking control of an optical fiber drawing process is developed and evaluated. The DRL-based control is capable of regulating the fiber diameter to track either steady or varying reference trajectories in the presence of stochasticity and non-linear delayed dynamics of the system. With about 3.5 hours of real-time training, it outperformed other control models such as open-loop control, proportional-integral (PI) control, and quadratic dynamic matrix control (QDMC) in terms of diameter error. It does not require analytical or numerical model of the system dynamics unlike model-based approaches such as linear-quadratic regulator (LQR) or model predictive control (MPC). It can also track reference trajectories that it has never experienced in the training process.