Pressure gain combustion is a form of combustion that uses pressure waves to transfer energy and generate a rise in total pressure during the combustion process. Pressure gain combustion shows potential to increase the cycle efficiency of conventional gas turbine engines if used in place of the steady combustor. However, one of the challenges of integrating pressure gain combustion into a gas turbine engine is that a turbine driven by pulsing flow experiences a decrease in efficiency. The interaction of pressure pulses with a turbine was investigated to gain physical insights and to provide guidelines for designing turbines to be driven by pulsing flow. An experimental rig was built to compare steady flow with pulsing flow. Compressed air was used in place of combustion gases; pressure pulses were created by rotating a ball valve with a motor. The data showed that a turbine driven by full annular pressure pulses has a decrease in turbine efficiency and pressure ratio. The average decrease in turbine efficiency was 0.12 for 10 Hz, 0.08 for 20 Hz, and 0.04 for 40 Hz. The turbine pressure ratio, defined as the turbine exit total pressure divided by the turbine inlet total pressure, ranged from 0.55 to 0.76. The average decrease in turbine pressure ratio was 0.082 for 10 Hz, 0.053 for 20 Hz, and 0.064 for 40 Hz. The turbine temperature ratio and specific turbine work were constant. Pressure pulse amplitude, not frequency, was shown to be the main cause for the decrease in turbine efficiency. Computational fluid dynamics simulations were created and were validated with the experimental results. Simulations run at the same conditions as the experiments showed a decrease in turbine efficiency of 0.24 for 10 Hz, 0.12 for 20 Hz, and 0.05 for 40 Hz. In agreement with the experimental results, the simulations also showed that pressure pulse amplitude is the driving factor for decreased turbine efficiency and not the pulsing frequency. For a pulsing amplitude of 86.5 kPa, the efficiency difference between a 10 Hz and a 40 Hz simulation was only 0.005. A quadratic correlation between turbine efficiency and corrected pulse amplitude was presented with an R-squared value of 0.99. Incidence variation was shown to cause the change in turbine efficiency and a correlation between corrected incidence and corrected amplitude was established. The turbine geometry was then optimized for pulsing flow conditions. Based on the optimization results and observations, design recommendations were made for designing turbines for pulsing flow. The first design recommendation was to weight the design of the turbine toward the peak of the pressure pulse. The second design recommendation was to consider the range of inlet angles and reduce the camber near the leading edge of the blade. The third design recommendation was to reduce the blade turning to reduce the wake caused by pulsing flow. A new turbine design was created and tested following these design recommendations. The time-accurate validation simulation for a 10 Hz pressure pulse showed that the new turbine decreased the entropy generation by 35% and increased the efficiency by 0.04 (5.4%).

Trois techniques de mesure différentes ont été utilisées pour étudier la distribution des fluides dans un lit fixe : la tomographie à fils, le collecteur de liquide et un ensemble de thermistances. La tomographie à fils, dont la première application dans le cadre des lits fixes est décrite ici, a permis, tout comme le collecteur de liquide, d’obtenir des résultats intéressants concernant l’influence de paramètres tels que la distribution initiale, le type de chargement et les débits de fluides sur la distribution du liquide. L’étude a notamment montré l’importance de bien définir la maldistribution de liquide en termes de la grandeur mesurée et a apporté un regard critique vis-à-vis des consensus généraux concernant l’effet sur la distribution de liquide de paramètres tels que le débit de gaz (dont les expériences ont montré qu’elle dépend du distributeur utilisé) et le type de chargement (l’hypothèse selon laquelle le chargement dense disperse mieux le liquide dans la direction radiale par rapport au chargement lâche n’a pas été vérifié). Une étude réalisée en régime à haute interaction a permis aussi d’observer la relation étroite qui existe entre la distribution initiale et le régime d’écoulement / Three different measuring techniques were used to study the fluid distribution inside a trickle-bed reactor: the wire mesh tomography, the liquid collector and a set of thermistors. The liquid collector and specially the wire mesh tomography, whose first application in trickle bed reactors is described here, yielded interesting results concerning the influence of variables such as the initial liquid distribution, the loading method and the fluid flow rates on liquid maldistribution. Among the main observations, the study illustrates the importance of well defining liquid maldistribution in terms of the measured quantity and prompts to some caution when referring to some “normally accepted facts” like the advantages in terms of liquid distribution obtained when increasing the gas flow rate (which depends, according to this study, on the quality of initial liquid distribution) or when using a dense loading of the catalyst (the hypothesis according to which, compared with a sock loading, dense loading favors radial dispersion was not verified by the study). Also, a study performed under high interaction conditions showed the intimate relationship between the inlet distribution and the flow regime observed inside the reactor