Study of composite polymer degradation for high pressure hydrogen vessel by machine learning approach

• Published in: Energy storage (Hoboken, N.J. : 2019) Vol. 6; no. 4
• Main Authors: Kadri, K., Kallel, A., Guerard, G., Abdallah, A. Ben, Ballut, S., Fitoussi, J., Shirinbayan, M.
• Format: Journal Article
• Language: English
• Published: Chichester, UK John Wiley & Sons, Ltd 01.06.2024; Wiley Subscription Services, Inc
• Subjects: Algorithms; behavior law; classification; Cluster analysis; Clustering; Composite materials; Damage assessment; damage law; Degradation; Dendrogram; External pressure; feature classification; Fiber volume fraction; Finite element method; High pressure; Hydrogen; hydrogen pressure vessel; Machine learning; RVE; Static pressure; Tanks; time series
• ISSN: 2578-4862; 2578-4862
• DOI: 10.1002/est2.645

The aim of this article is to study the degradation of a composite material under static pressure. The high pressure condition is similar to the one encountered inside hydrogen tanks. Damage modeling was used to evaluate the behavior of hydrogen tanks to high pressure. A practical approach, coupling a finite element method (FEM) simulation and machine learning (ML) algorithm, is suggested. The representative volume element (RVE) was used in association with a choice of a behavior law and a damage law as an input data. Algorithms for ML classification such as K‐nearest neighbors (k‐NN) and a special k‐NN with a dynamic time warping metric were used. The hierarchical clustering through dendrograms visualizations allowed to exhibit the impact of composite parameters in relation to fiber, matrix properties and fiber volume fraction on the strain degradation under external static pressure. Continuing this, the optimum RVE which shows a low degradation value will be exhibited.