Constant Rate of Momentum Change Ejector: simulation, experiments and flow visualisation

Constant Rate of Momentum Change Ejector: simulation, experiments and flow visualisation

An ejector is a momentum-transfer device that requires no external mechanical input or moving parts. However, ejectors have low performance due to irreversibilities such as viscous losses and shocks in the primary stream and diffuser. It has previously been argued that by maintaining a constant rate of momentum change along the ejector duct, shock losses could be eliminated or at least minimised, and so the Constant Rate of Momentum Change (CRMC) ejector was introduced. The CRMC configuration appears to have significant potential, but the CRMC design prescription relies on: (1) an arbitrary choice for the constant rate of momentum change along the length of the duct; and (2) complete mixing between primary and secondary streams at the entrance to the duct. This thesis investigates the themes of shock losses and mixing within a CRMC ejector using physical experiments and computational simulation.

The CRMC ejector duct and the primary nozzle were manufactured using 3D printing technology and then an experimental test bench using air as the working
fluid was assembled and successfully tested. The primary nozzle had a throat diameter of 3.2mm and an exit diameter of 13.6 mm; the CRMC duct had a throat diameter of 25.48 mm. Extensive experimental tests were carried out for primary pressure between 200 kPa and 270 kPa, and secondary pressure between 0.6 kPa and 5 kPa. The results demonstrate the primary nozzle exit position within the entrainment region has a limited effect on the ejector performance in terms of the entrainment ratio and critical back pressures. A gas dynamic model was used to compare the performance of the present CRMC ejector with different ejector profiles (both conventional and CRMC) working with different fluids. The CRMC ejector showed a slightly better performance in terms of entrainment ratio and compression ratio. When CFD simulations of the present CRMC ejector were compared with a conventional ejector at a similar operating condition, the total pressure of the CRMC ejector remained 15% larger than the conventional ejector but this higher performance was due to different primary flow shock structures, not due to improvements in the compression process within the diffuser. Differences in the primary flow structure are thought to be caused by the different contraction angle of the secondary flow area. Higher entrainment ratio and compression ratio were simulated for the CRMC ejector relative to the conventional ejector but were not as high as expected from the CRMC design.

To investigate the mixing of the flow within the CRMC ejector, a laser-based visualization technique was developed. A transparent CRMC ejector test section was designed, fabricated, and operated in the ejector system using air as the working fluid. The laser-based flow visualisation used a laser light beam of diameter of 1mm to illuminate the seeded secondary flow and thus, the unmixed primary flow was defined. The wall static pressure of the seeded flow agrees well with that of the unseeded flow which indicates that the seeding has a very small effect on the flow. Analysis of the images by digital image processing tools enabled identification of the jet core flow length which was found to lie between 65mm and 95mm from the nozzle exit at the selected operating conditions.

The primary and secondary flows entering the CRMC duct are certainly not fully mixed as assumed in the CRMC design prescription. Furthermore, enhancement of the distribution of the wall static pressure and centreline total pressure is not directly attributable to the CRMC prescription. The modest performance improvements associated with the present CRMC design relative to the performance of a conventional duct should be balanced against the added complexity associated with manufacturing a CRMC duct when considering the CRMC design for future applications.