Feasibility of a simplified model for a latent thermal energy storage unit for integration with multiphysics systems

• Autor(a) principal: Thiago Torres Martins Rocha
• Data de Publicação: 2023
• Tipo de documento: Tese
• Idioma: eng
• Título da fonte: Repositório Institucional da UFMG
• Texto Completo: http://hdl.handle.net/1843/63459
• topic: Latent thermal energy storage; Phase-change material; Computational fluid dynamics; Porosity constant; Effective thermal conductivity; Engenharia mecânica; Calor - Armazenamento; Energia; Porosidade; Dinâmica dos fluidos computacional; Condutividade térmica; Dinâmica dos fluidos

The goal of this work is to assess the suitability of a simplified model for the melting of a phase-change material, verifying its accuracy and simulation time. Initially, a mathematical model is formulated and solved with Ansys® Fluent, being validated with two experimental datasets, using lauric acid. The effects of altering the porosity constant, material properties, and their interaction were assessed. It was evidenced that, by holding the porosity constant of validation in another geometry, absolute average deviations as high as 38.0% could be achieved, whereas this error could be reduced to 5.7% with a proper adjustment. The numerical results of this computational fluid dynamics model were also used to derive an effective thermal conductivity, for later use. Subsequently, a simplified model based on the pure conduction was developed and implemented in Python language, including the natural convection of the liquid phase-change material trough the effective thermal conductivity. A horizontal shell-and-tube unit was chosen, since it is one of the most employed configuration in heat exchangers using phase-change materials. The results of this simplified model revealed maximum and absolute average deviations of 8.17% and 4.32%, compared to the computational fluid dynamics model, and 7.43 % and 1.67%, compared to experimental data, respectively, with a 3500 fold reduction on the simulation time. Therefore, the proposed model can be considered feasible for integration with multiphysics systems.