Wave Rotor Application for Micro Gas Turbines

: (Alternative Format Thesis)

Abstract

There is an increasing demand for compact, portable and endurable power sources in both the civilian and military markets. Gas turbines are a proven technology excelling at high energy and power density, reliability and flexibility. However, downscaling turbomachinery entails severe penalties in component and system efficiency as well as service life due to increased rotational speeds, greater tip clearance and larger relative roughness effects.
In this thesis, an alternative to turbomachinery is proposed using on a non-axial, throughflow wave rotor turbine that combines shaft power extraction as well as fluid compression and expansion through shock and rarefaction waves within a single rotating device. The thesis is written in the alternative thesis format and is based on five academic articles that address the performance characteristics of such a pressure-exchange device.
The first paper introduces the concept of a micro-wave rotor gas turbine based on a throughflow wave rotor with non-axial channel shape. The study uses a time-marching quasi-one-dimensional wave action model that is able to accurately capture the wave pattern within a wave rotor channel. The model is coupled with a steady-flow combustor model and identifies leakage and combustor pressure drop as major performance affecting parameters.
The second publication addresses an experimental study centred on an innovative four-port, three-cycle micro-wave rotor. The study seeks to characterise the wave rotor and further elaborates on the challenges in exploiting wave rotor technology. In addition, reasonably high compression and expansion efficiencies are achieved. In the third paper, the results from the second paper are then used in conjunction with literature data to validate and calibrate the quasi-one-dimensional model.
The fourth paper introduces a shape-optimisation study that seeks to enhance the shaft power output of the baseline wave rotor turbine. To do so, a reduced order quasi-two-dimensional transient CFD model is coupled to a hybrid algorithm. The optimisation technique trains a Kriging surrogate model with data from a genetic algorithm and alternates the search area based on global exploration and local exploitation, thus ensuring improved coverage of the design space compared to standard evolutionary algorithms. The optimisation routine results in a wave rotor candidate design that predicts a power output improvement of 80%. The numerical results indicate that the increase in shaft power stem from increased incidence losses leading to a 3% drop in pressure ratio.
Finally, the fifth and final paper verifies the numerical results through an experimental campaign comparing both a baseline and an optimised rotor. The experimental results support the validity of the reduced-order CFD model to accurately predict performance trends. Implications of the camberline change on performance are discussed as well.

Date of Award	18 Nov 2020
Original language	English
Awarding Institution	University of Bath
Supervisor	Colin Copeland (Supervisor), Giovanni Vorraro (Supervisor) & Sam Akehurst (Supervisor)