Numerical study on converging jets
Using turbulent jet flows is an important means to dilute pollutant released to ambient water (e.g. in ocean outfalls). Conventional engineering design rules are based on semi-empirical formulae, which involve time-averaged flow quantities. Jet nozzles are commonly aligned parallel to each other ass...
Saved in:
Main Author: | |
---|---|
Other Authors: | |
Format: | Theses and Dissertations |
Language: | English |
Published: |
2015
|
Subjects: | |
Online Access: | https://hdl.handle.net/10356/62947 |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Institution: | Nanyang Technological University |
Language: | English |
Summary: | Using turbulent jet flows is an important means to dilute pollutant released to ambient water (e.g. in ocean outfalls). Conventional engineering design rules are based on semi-empirical formulae, which involve time-averaged flow quantities. Jet nozzles are commonly aligned parallel to each other assuming minimal momentum impact in the merging process of the component jets. A potential alternative is to have jets emitting diagonally to each other— mutually-impinging jets—where momentum exchange plays a key role in shaping the turbulent core of the compound jet, thus affecting the mixing capability at high Reynolds number (Re). This study’s target is to formulate practical equations to characterise the turbulent jet flows from three impinging jets, thereby elucidating the advantage of such flow configuration as opposed to a single outlet, or an array of parallel aligned jets. To achieve this target, the author carried out incremental research from a single free jet to a system of two interacting jets and finally three co-planar jets. The open source computer program Gerris has been applied to study high-Re viscous flows. The solver functionality is extended with a simple LES model based on Smagorinsky’s formula. Results of single jet simulations are compared to Papanicolaou (1984) ’s experimental data, where Re ranges from 5000 to 10000. The simulated velocity profiles follow a typical Gaussian curve, with peak velocity being underestimated about 10% to 20%. The input viscosity parameter (to counter the effect of grid coarseness) is found to be an order of magnitude larger than the physical viscosity.
A mathematical model for two impinging jets has been suggested. The double shear layer flow field formed by these two jets is assumed to radiate from the zone of impact. The velocity profiles of this double shear layer is characterised by the mean velocity, U, and the layer thickness, h. This model is only valid for small values of the angle 2α between the two jets. The simulated flow field was compared to experimental results by Jia Xin (NTU/MRC study group). For 3-jet cases, simulations with a typical Re of 10^5 has been performed with various jet configurations: 2α = 15°, 30°, 45° and an discharge-equivalent single jet. The cross-sectional shape of the turbulent core in the compound flow shows tendency to become circular from a distance of about 25d (nozzle diameter). There is no apparent functional relationship between the shape characteristics and α. Finally, the mixing capability of a 3-jet configuration has been examined through simulation. Larger α and larger in-between nozzle spacing, s, had both shown more favourable dilution. Increasing s leads to an improvement in entrained flow into the jet core; a relationship between the entrainment rate and s had been proposed. The correlation between entrainment rate ∆Q/∆x and s² has been found to be reasonable, with a correlation coefficient of R² ≈ 0.85. |
---|