Numerical investigations of the performance and effectiveness of thermoacoustic couples.

Abstract

Thermoacoustics is a field of study which includes devices purpose-built to exploit the phenomenal interaction between heat and sound. Thermoacoustics has been demonstrated as an effective technology which can potentially serve a variety of purposes such as cryogenics, cost-effective domestic refrigeration or electricity generation, without adverse environmental impact or commercial drawbacks such as expensive construction or maintenance costs or high part counts. The mechanisms by which thermoacoustic devices operate at low amplitudes have been identified and effective design tools and methods are available, but the precise heat and mass transfer which occurs deep inside the core of thermoacoustic devices at high amplitudes cannot at present be precisely determined experimentally, and to date have been estimated using only relatively simple or one-dimensional computational domains. It is expected that thermoacoustic devices will need to operate at relatively high pressure amplitudes for commercial and practical applications, to achieve power densities similar to competing technologies. Clearly, advancement of these models and the methods used to investigate them will enable a better understanding of the precise heat and mass transfer that occurs within such devices. Previous numerical studies have modelled a ‘thermoacoustic couple’ which consists of a single or several plates (often modelled with zero thickness) and channels within an oscillatory pressure field. In this thesis several improvements to the ‘thermoacoustic couple’ modelspace are introduced and modelled, and compared with published results. Using the commercial CFD software Fluent, a two-dimensional, segregated and second-order implicit numerical model was developed which solves equations for continuity of mass, momentum and energy. These equations were computed using second-order and double-precision discretisation of time, flow variables and energy. A computational domain is presented which is capable of modelling plates of zero or non-zero thickness, is ‘self-resonant’ and able to capture the entrance and exit effects at the stack plate edges. Studies are presented in which the acoustic pressure amplitude, the thickness of the plate (‘blockage ratio’) and the shape of the plate are varied to determine their influence upon the rate of effective heat transfer, flow structure and overall efficiency. The modelling of thermoacoustic couples with finite thickness presented in this thesis demonstrates that the finite thickness produces new results which show significant disturbances to the flow field and changes to the expected rate and distribution of heat flux along the stack plate. Results indicate that the thickness of the plate, t[subscript]s, strongly controls the generation of vortices outside the stack region and perturbs the flow structure and heat flux distribution at the extremities of the plate. Increases in t[subscript]s are also shown to improve the integral of the total heat transfer rate but at the expense of increased entropy generation. Another contribution of this thesis is the study of the effect that leading and trailing edge shapes of stack plates have on the performance of a thermoacoustic couple. In practice, typical parallel or rectangular section stack plates do not have perfectly square edges. The existing literature considers only rectangular or zero-thickness (1-D) plates. Hence a study was performed to evaluate the potential for gains in performance from the use of non-rectangular cross sections, such as rounded, aerofoil or bulbous shaped edges. Consideration of various types of stack plate edges show that performance improvements can be made from certain treatments to the stack plate tips or if possible, stack plate profiles. This thesis also considers the influence of thermophysical properties and phenomena associated with practical thermoacoustic devices to investigate the applicability of the numerical model to experimental outcomes. Comparisons made between results obtained using the numerical model, linear numerical formulations and experimental results suggest that the numerical model allows comparative study of various thermoacoustic systems for design purposes but is not yet of sufficient scope to fully characterise a realistic system and predict absolute levels of performance. However, the presented method of modelling thermoacoustic couples yields increased insight and detail of flow regimes and heat transportation over previous studies.