Principal component analysis of hygrothermal ageing in a castor oil-based polyurethane foam using engineered degradation descriptors

Abstract

Accelerated ageing studies of polymeric foams often report individual mechanical properties separately, which can mask the overall degradation pattern needed for durability screening in sustainable construction and structural repair-related applications. This study proposes nine physically based descriptors that describe geometry, mechanical integrity, directional degradation, structural anisotropy, and energy absorption. These descriptors were obtained from open-access compression and 3D digital image correlation data for a castor oil-based polyurethane foam aged at three temperatures under elevated relative humidity. The experimental context is described through material synthesis, conditioning in a climatic chamber, and a two-camera compression-test setup. Representative optical images provide qualitative structural context for cellular heterogeneity and edge densification, whereas directional dependence is established by the direction-resolved mechanical measurements. Principal component analysis (PCA) was applied to determine the main degradation modes. Two principal components explain most of the total variance. The first is interpreted as a general degradation-severity index, with integrity, retention, geometry, and anisotropy descriptors loading in a consistent way; regression shows that temperature is the main statistically significant driver, in line with the published Arrhenius kinetics. The second component is linked mainly to the maximum energy-absorption efficiency and showed no statistically significant association with the thermal inputs in the present dataset, suggesting batch-level variability rather than a second thermal degradation route. The study therefore offers a physically grounded, two-parameter framework for multivariate ageing monitoring of bio-based foams and supports their future durability screening for sustainable construction, structural repair, and low-carbon infrastructure applications.

The authors' abstract, as published at the source. Discover Applied Sciences, 2026 · DOI ↗