Thermo-Elastic Analysis of an Axisymmetric Layered Cylinder under Constant Thermal Loading with Artificial Intelligence-Based Validation

Abstract

This study presents a numerical investigation of the thermo-elastic behavior of a two-layer disk composed of partially stabilized zirconia (PSZ) and aluminum subjected to uniform (constant) thermal loading. The analysis focuses on the evaluation of radial and circumferential stress distributions as well as radial displacement responses under different prescribed temperature levels. The governing thermo-elastic equations are formulated under plane stress assumptions and solved numerically by discretizing the disk geometry into finite radial segments.The material layers are assumed to be homogeneous, isotropic, and perfectly bonded, while the elastic properties are considered temperature-independent within the investigated temperature range of 12.5 °C to 100 °C. The numerical formulation ensures continuity of radial displacement and radial stress across the material interface. The results demonstrate that increasing temperature levels significantly influence the magnitude of thermo-elastic stresses and radial displacements. In particular, pronounced stress gradients are observed in the vicinity of the material interface, highlighting the effect of thermal expansion mismatch between PSZ and aluminum. The circumferential stress component is found to be more sensitive to temperature variations compared to the radial stress, while radial displacement increases monotonically with temperature. Beyond numerical analysis, the generated stress and displacement data are employed as a structured dataset for artificial intelligence–based validation. A supervised learning framework is developed to predict thermo-elastic responses based on temperature and radial position inputs. The AI model demonstrates strong agreement with numerical results, confirming its capability to accurately reproduce stress and displacement trends under constant thermal loading conditions. The combined numerical–AI approach provides a reliable and efficient tool for analyzing thermo-elastic behavior in layered disk structures.