KIPS Transactions on Computer and Communication Systems (정보처리학회논문지:컴퓨터 및 통신 시스템)

Korea Information Processing Society (한국정보처리학회)
권호 표지
DOI QR코드

DOI QR Code

Study of Polymor Properties Prediction Using Nonlinear SEM Based on Gaussian Process Regression

가우시안 프로세서 회귀 기반의 비선형 구조방정식을 활용한 고분자 물성거동 예측 연구

  • 문경렬 (한국소재융합연구원 생산기술연구단) ;
  • 박건욱 (한국소재융합연구원 혁신소재연구단)
  • Received : 2023.10.17
  • Accepted : 2023.11.30
  • Published : 2024.01.31
Download PDF
⟨ Previous Next ⟩

Abstract

In the development and mass production of polymers, there are many uncontrollable variables. Even small changes in chemical composition, structure, and processing conditions can lead to large variations in properties. Therefore, Traditional linear modeling techniques that assume a general environment often produce significant errors when applied to field data. In this study, we propose a new modeling method (GPR-SEM) that combines Structural Equation Modeling (SEM) and Gaussian Process Regression (GPR) to study the Friction-Coefficient and Flexural-Strength properties of Polyacetal resin, an engineering plastic, in order to meet the recent trend of using plastics in industrial drive components. And we also consider the possibility of using it for materials modeling with nonlinearity.

고분자 분야의 개발 및 양산과정에는 제어가 안되는 많은 변수가 있으며, 화학적 조성, 구조, 가공 조건 등 작은 변화에도 물성편차가 크게 발생하기에 보편적인 환경을 가정한 기존의 선형적 모델링 기법으로는 현장 데이터 적용시 많은 오차가 발생한다. 이에 본 연구에서는 최근 산업용 구동부품의 플라스틱 채용경향에 맞추어 엔지니어링 플라스틱인 Polyacetal 수지의 내마모성 및 내굴곡성 강화 연구에 다변량 분석기법인 구조방정식과 가우시안 프로세스 회귀를 결합한 모델링 방식(GPR-SEM)을 제안하고, 비선형성을 가지는 물질 모델링에 활용 가능성을 고찰하고자 한다.

Keywords

References

  1. N. Lee, Y. Shin, and D. Shin, "Prediction of mechanical properties and behavior of polymer matrix composites based on machine learning," Journal of the Korean Institute of Gas, Vol.25, No.2, pp.64-71, 2021. 
  2. M. Kim, S. Kang, and K. Min, "Searching for next-generation battery cathode materials using machine learning," Korean Society of Mechanical Engineers, 2021 Conference, pp.1166-1168, 2021. 
  3. S. M. Kim, "Optical characterization and prediction with neural network modeling of various composition of perovskite usinga hyper regression method," Polymer Science and Technology, Vol.33, No.6, pp.548, 2022. 
  4. J.-H. Sim, "Characteristics analysis of highly elastic materials according to the graphite content and a simulation study of physical properties prediction using a nonlinear material model," Textile Coloraation and Finishing, Vol.34, No.4, pp.250, 2022. 
  5. K.-M. Lee, K.-Y. Kim, U. Oh, S.-K. Yoo, and B.-S. Song, "Prediction of multi-physical analysis using machine learning," Journal of IKEEE(Institute of Korean Electrical and Electronics Engineers), Vol.20, No.1, pp.99-100, 2016. 
  6. D. Lee I.-S. Hwang, "Analysis on the dynamic characteristics of a rubber mount considering temperature and material uncertainties," Korea Society of Computational Structural Engineering Journal, Vol.24, No.4, pp.387-388, 2011. 
  7. C. E. Rasmussen and C. K. I. Williams, "Gaussian Processes for Machine Learning," pp.19-46, pp.79, pp.114-122, pp.123-146, 2006. 
  8. C. M. Bishop, "Pattern recognition and machine learning," pp.291-339, 2006. 
  9. J.-P. Yu, "The Criticisms and considerations of structural equation modeling," Journal of Product Research, Vol.34, No.4, pp.83-93, 2016.  https://doi.org/10.36345/KACST.2016.34.4.009
  10. C.-S. Yoo, "A performance evaluation of fit indices for structural equation model," pp.2-7, 2019. 
  11. T. Eckert, F. C. Klein, P. Frieler, O. Thunich, and V. Abetz, "A new experimental design method for targeted property optimization in polymer development," Polymers, 13-03147-v2, pp.6-10, 2021. 
  12. S. Pruksawan, G. Lambard, S. Samitsu, K. Sodeyama, and M. Naito, "Prediction and optimization of epoxy adhesive strength from a small dataset through active learning," Science and Technology of Advanced Materials, Vol.20, No.1, pp.1010-1011, 2019. 
  13. M. Hossin and M. N. Sulaiman, "A review on evaluation metrics for data classification evaluation," International Journal of Data Mining & Knowledge Management Process (IJDKP), Vol.5, No.2, pp.5, 2015.