UV reactor flow visualization and mixing quantification using three-dimensional laser-induced fluorescence
Highlights
► Flow in a lab-scale UV reactor was visualized for the first time using 3DLIF. ► Mixing in the UV reactor was highly unsteady not only spatially but also temporally. ► Flow past a UV lamp was characterized by the presence of recirculation zone and vortex shedding. ► Trajectory of tracer strongly depended on upstream point of entry. ► Sufficient downstream length would be required to obtain homogenous water sample at the outlet.
Introduction
UV disinfection has been gaining popularity in drinking water treatment over the past decade due to the discovery of the efficient inactivation of Cryptosporidium parvum oocysts and Giardia lamblia cysts at relatively low doses (Clancy et al., 1998, Linden et al., 2002) with much less concern on the formation of disinfection by-products as compared to chemical disinfectants (Bellar et al., 1974, Glaze et al., 1993). However, a spatial and temporal assessment of the UV dose delivered and the reactor performance has been severely limited for utilities practicing UV disinfection. The commonly used validation method, biodosimetry, treats the UV reactor as a “black-box” and hence cannot account for the dependence of the dose delivery on the complex hydrodynamics and the spatial variation in UV intensity. Development of proper reactor design would be ideally pursued through an understanding of the fluid behavior that determines how microorganisms accumulate UV dose as they spend varying amounts of time in regions of fluctuating light intensity (Lawryshyn and Cairns, 2003).
The unsteadiness of flow in UV reactors arises mainly due to the placement of cylindrical UV lamps perpendicular to flow that leads to separation forming unsteady large-scale vortices and consequently significant fluid mixing (Williamson, 1996, Zdravkovich, 1997). These so called von Karman vortices (repeating pattern of alternating vortices produced downstream of a bluff body) compound the complexity and unsteadiness in the flow region behind the lamp, as they involve the interactions of three shear layers, i.e., a boundary layer around the sleeve, a separating free shear layer, and a highly turbulent wake (Williamson, 1996). The presence of multiple lamps in a staggered configuration further complicates the flow, rendering the prediction of unsteady hydrodynamics significantly difficult. Furthermore, inlet configuration (Sozzi and Taghipour, 2006), upstream pipe bends (Zhao et al., 2009), and the presence of modifications such as baffles (Blatchley et al., 1998, Wols et al., 2010a), rings (Janex et al., 1998), or “wave-like” walls (Chiu et al., 1999a) have been found to significantly affect the hydrodynamics and reactor performance.
Due to the complexity of the flow, computational fluid dynamics (CFD) has been increasingly used to model the hydrodynamics in UV reactors based on the time-averaged Reynolds Averaged Navier Stokes (RANS) approaches (Sozzi and Taghipour, 2006, Alpert et al., 2010, Wols et al., 2010b). However, the velocity distributions in the RANS simulations have been found to differ from the experimental results typically obtained using particle image velocimetry (PIV). Alpert et al. (2010) and Wols et al. (2010b) determined that RANS simulations under-predicted the flow complexity in dynamic wake regions and dead zones due to the poor capture of large vortices and turbulent motions. These studies concluded that resolving the unsteady turbulent motions is essential to provide an accurate representation of the microorganism trajectories and more significantly the UV dose received by each microbe.
Dose distributions consist of both spatial and temporal components. While many past studies have focused on the former, the authors are unaware of any experimental or computational studies that considered the temporal component in the reactor. In this study, a three-dimensional laser-induced fluorescence (3DLIF) (Tian and Roberts, 2003) was applied for the first time to examine the hydrodynamics in a lab-scale model UV reactor both spatially and temporally. In the 3DLIF system, a planar monochromatic laser sheet is created and scanned across the width of the reactor to obtain three-dimensional images. The laser causes a tracer dye to fluoresce which is captured by a high-speed CCD camera (Guiraud et al., 1991). Using this technique, instantaneous, 2D and 3D mixing characteristics in a model UV reactor were visualized and quantitatively examined.
Section snippets
3DLIF system
Detailed information regarding the 3DLIF system used in this study is given in Kim et al. (2010). Briefly, a laser beam generated by an argon ion laser (Innova 90, Coherent®, Palo Alto, CA) at a wavelength of 514 nm and an intensity of 1.5 W was used to excite a fluorescent dye tracer, Rhodamine 6G (Sigma–Aldrich, St. Louis, MO). The laser beam was first directed toward a mirror which oscillated vertically at high frequency to produce a 2D laser sheet that passed through the center width of a
Flow visualization using 3DLIF
Fig. 2 shows how three-dimensional flows in a UV reactor can be visualized using 3DLIF at a high resolution (corresponding to millions of sampling points), which is not possible with traditional dye tracer test techniques. These images were obtained from a 3DLIF experiment performed with dye injected at the point in the center of the y-z plane. The region from the sleeve to the outlet is dissected in the stream-wise (x-z plane), span-wise (x-y plane), and cross-stream (y-z plane) directions.
Conclusions
This study applied the 3DLIF technique for the first time to visualize and quantitatively analyze the flow across a UV lamp in a model reactor used for drinking water treatment. In addition to three-dimensional mixing, the technique successfully visualized the two-dimensional, transient mixing behaviors within the reactor, which has not been possible with traditional tracer test techniques. It is also noteworthy that the 3DLIF technique is non-intrusive, i.e., there is no disturbance in the
Acknowledgments
This research was partially funded by Water Research Foundation (Project No. 4134) and Korea Water Resources Corporation (Kwater).
References (25)
- et al.
Modeling the UV/hydrogen peroxide advanced oxidation process using computational fluid dynamics
Water Research
(2010) - et al.
Process modeling of ultraviolet disinfection
Water Science and Technology
(1998) - et al.
Turbulence properties in the cylinder wake at high Reynolds numbers
Journal of Fluids and Structures
(2006) - et al.
Laser measurements of local velocity and concentration in a turbulent jet-stirred tubular reactor
Chemical Engineering Science
(1991) - et al.
Numerical simulation of UV disinfection reactors: evaluation of alternative turbulence models
Applied Mathematical Modelling
(2007) - et al.
Influence of inlet positions on the flow behavior inside a photoreactor using radiotracers and colored tracer investigations
Applied Radiation and Isotopes
(2007) - et al.
Flow visualization around a circular cylinder near to a plane wall
Journal of Fluids and Structures
(2002) - et al.
Experimental investigation of a confined flow downstream of a circular cylinder centred between two parallel walls
Journal of Fluids and Structures
(2008) - et al.
Flow past a circular cylinder between parallel walls at low Reynolds numbers
Ocean Engineering
(2010) - et al.
The weaknesses of a k-epsilon model compared to a large-eddy simulation for the prediction of UV dose distributions and disinfection
Chemical Engineering Journal
(2010)
Evaluation of experimental techniques to validate numerical computations of the hydraulics inside a UV bench-scale reactor
Chemical Engineering Science
Occurrence of organohalides in chlorinated drinking waters
Journal American Water Works Association
Cited by (15)
Experimental study of flow visualisation using fluorescent dye
2022, Flow Measurement and InstrumentationEffect of dissolved oxygen on efficiency of TOC reduction by UV at 185 nm in an ultrapure water production system
2019, Water ResearchCitation Excerpt :These results differ significantly from the general phenomenon that the efficiency of UV irradiation is proportional to the UV dosage. This result can be explained by the concept of the mixing effect in the UV reactor (Gandhi et al., 2011). Because water can strongly adsorb UV-185, approximately 90% of UV-185 is adsorbed within 5-mm thickness of water from the sleeve of the lamp (Han et al., 2004).
LIF study of temporal and spatial fluid mixing in an annular downcomer
2019, Annals of Nuclear EnergyExperimental investigation of fluid mixing inside a rod bundle using laser induced fluorescence
2019, Progress in Nuclear EnergyFlow visualization and mixing quantification in a rod bundle using laser induced fluorescence
2016, Nuclear Engineering and Design