Analysis of thermoelastic stresses in filament wound composite pressure vessels using evolutionary deep learning and many-objective optimisation

Thermoelastic stresses induced in filament wound composite pressure vessels by temperature changes significantly influence the resulting load of the pressure vessel. This study presents the analytical approaches based on classical lamination theory and netting theory, which can be used for computing these thermoelastic stresses in the basic parts of the pressure vessel, such as the cylinder part, the end dome, and the junction area between them. Two material configurations were considered – glass-epoxy and carbon-epoxy. Concerning the knowledge of the thermoelastic stresses in analysed essential parts of the vessel, the Hoffman failure index analysis was performed. Using this failure index analysis, the critical places of the pressure vessel can be easily detected. Using data-driven evolutionary algorithms, the invariance of the pressure vessel geometry concerning the thermoelastic stresses was verified. Knowing the critical places of the pressure vessel may improve the designer’s decision during the development and design process.