Home >> Publication
ELSEVIER
2023년 234권 1호 p.1 ~ p.12
A versatile strategy for hybridizing small experimental and large simulation data: A case for ceramic tape-casting process
A versatile strategy for hybridizing small experimental and large simulation data: A case for ceramic tape-casting process
Jeong-Hun Kim, Hyunseok Ko
Dong-Hun Yeo. Zeehoon Park, Upendra Kumar, Kwan-Hee Yoo, Aziz Nasridinov, Sung Beom Cho
- Abstract -
In manufacturing industry, finding optimal design parameters for targeted properties has traditionally been 
guided by trial and error. However, limited data availability to few hundreds sets of experimental data in typical 
materials processes, the machine-learning capabilities and other data-driven modeling (DDM) techniques are too 
far from it to be practical. In this study, we show how a versatile design strategy, tightly coupled with physics- 
based modeling (PBM) data, can be applied to small set of experimental data to improve the optimization of 
process parameters. Our strategy uses PBM to achieve augmented data that includes essential physics: in other 
words, the PBM data allows the inverse design model to ‘learn’ physics, indirectly. We demonstrated the ac- 
curacy of both forward-prediction and inverse-optimization have been dramatically improved with the help of 
PBM data, which are relatively cheap and abundant. Furthermore, we found that the inverse model with 
augmented data can accurately optimize process parameters, even for ones those were not considered in the 
simulation. Such versatile strategy can be helpful for processes/experiments for the cases where the number of 
collectable data is limited, which is most of the case in industries.
- Key Words -
Machine learning Simulation Data deficiency Inverse design Tape-casting
Machine learning Simulation Data deficiency Inverse design Tape-casting
- 논문 구분 -
SCI(E)
- 사사 정보 -
We acknowledge the support from Ministry of Trade, Industry & Energy (20004367) and National Research Foundation (RS-2023- 00209910). This research was also supported by the MSIT(Ministry of Science and ICT), Korea, under the Grand Information Technology Research Center support program(IITP-2023-2020-0-01462) supervised by the IITP(Institute for Information & communications Technology Planning & Evaluation). The computations were carried out using re- sources from Korea Supercomputing Center (KSC-2022-CRE-0348).
- Downloads -
1-s2.0-S0264127523007724-main.pdf