Design and performance of a ThermoAcoustic Electric Generator powered by waste-heat based on linear and nonlinear modelling

This article presents the design, numerical modeling, and performance evaluation of a multistage ThermoAcoustic Electric Generator (TAEG), aimed at converting thermal energy from internal combustion engine exhaust gases into electricity. The proposed TAEG adopts a double-stage looped resonator configuration, using helium as the working fluid, with an internal static pressure limited to 20 bar for safety reasons. Gas-to-gas hot heat exchangers were specifically designed to recover waste heat at approximately 530 K. Due to practical constraints, commercial audio speakers were employed as acoustic-to-electric transducers, despite their lower impedance compared to ideal linear alternators. Initial linear thermoacoustic simulations conducted using DeltaEC software optimized geometric and operational parameters, predicting an electrical power output around 300 W (150 W per stage) with a resistive load of 10 Omega. However, recognizing the inherent limitations of linear modeling, particularly the omission of nonlinear thermo-fluid dynamics, a computational fluid dynamics (CFD) analysis was conducted using OpenFOAM. The CFD model integrated novel nonlinear porous media formulations tailored for oscillatory flow conditions within the thermoacoustic core. Comparisons between a purely linear DeltaEC and OpenFOAM results revealed excellent qualitative agreement but quantitative differences, primarily due to minor losses caused by abrupt geometric discontinuities, and conical segments. A second DeltaEC model including all minor losses based on the steadystate approximation reveals that dissipation were overestimated compared to the CFD model. Therefore, a third DeltaEC model was built by calibrating minor losses based on CFD data. The findings emphasize the critical role of accurately modeling nonlinear effects to reliably predict TAEG performance and the limitation of the local pressure drop coefficients based on steady-state analysis.