Latest advances in UV-disinfection
Hydrodynamic simulation and Relation to practical experiences

Presented at 
Aquatech '96
Amsterdam,
September 1996
by

M.M. Baas, Eur.Ing.
Berson UV-techniek B.V., Nuenen, The Netherlands

Contents

Summary
1. Introduction
2. Calculation of UV intensity
3. UV dose
4. Correlation between theoretical calculations and practical experiments
References

Summary

The results presented give important evidence that it is possible to predict UV disinfection performance based on flow simulation calculations delivered by Cyclone Fluid Dynamics B.V. and UV-C intensity calculations provided by Berson UV-techniek. Historical methods assumed plug flow, no radial dispersion, no dead spaces and homogeneous UV-intensity throughout the entire volume of the reactor. In real circumstances the parameters vary considerably with all associated problems of correct assessment. The program developed by Cyclone Fluid Dynamics and Berson UV-techniek analyses the flow and UV-intensities for every point in space in detail. The flow domain is divided in a very large number of cells and the Navier-Stokes equations are solved in each individual cell. The resulting streamlines are solved in each individual cell. The resulting streamlines are fed into the Berson UV-techniek UV-C intensity program. It shows very good correlation to the actual situation of the user with respect to pressure drop and bacterial kill rates. Correlation of the computer simulation model with a big-assay of the UV-system will reduce the number of units subject to certification via successful extrapolation of the big-assay results with a calculation model applied to a number of units. An European approval via CEN-TC-164 will be investigated.

1. Introduction

In this presentation the possibility of certification based on theoretical intensity calculations and flow simulations will be discussed and related to practical experience. Although over 50,000 UV-disinfection systems are in operation all over the world, a proper world standard has not been determined. In various countries legislation and/or certification make an effort to limit exaggerated claims on UV lamps and bad workmanship on irradiation chambers to safeguard the quality of potable water. For many years water has been disinfected to render it suitable for drinking by the use of either chlorine or chlorine dioxide (Ref. 1). The first application of an Ultra Violet (UV) Low Pressure mercury vapour discharge lamp was in Marseille/ France in 1901. However, UV disinfection became widely applied in Europe for potable water from 1955 onwards. In that year UV disinfection equipment was installed in Switzerland, Austria and Norway. Following the discovery of the formation of halogenated hydrocarbons (THM's) during chlorination, UV disinfection became popular in most European countries (Ref. 2).

Pioneering work has been undertaken by KIWA in The Netherlands on the development of innovative UV disinfection systems for potable water disinfection. The results of a 5 year test, beginning April 1981, prove the successful use of UV-disinfection upon predicted UV-dosage in the irradiation chamber for Low Pressure mercury vapour discharge lamp 9 mJ/cm2 and for Medium Pressure mercury vapour discharge lamp 6 mJ/cm2 at same kill rates of 99,9%.

UV-disinfection in the Netherlands
(source: KIWA)

From approximately 200 measurements, bacteria from the cold group were detected in the raw water 116 times, from which 65 could be confirmed as E-coli. After UV disinfection, bacteria from the cold group were detected 10 times, from which 5 could be confirmed as E-coli during this 5 year period (Ref. 3). In a few cases no complete kill was achieved because of a too low UV-dosage caused by low transmissivity / dirty quartz sleeves. However, in the following years, with more accurate monitoring and effective mechanical wiping of quartz sleeves, no further breakthrough of E-coli occurred. New UV disinfection systems based on traditional Low Pressure mercury vapour discharge lamps (LP-Hg-lamps) or Medium Pressure mercury vapour discharge lamps (MP-Hg lamps) have been developed. The MP-Hg lamp technology results in compact and simplified installations. Hydraulics in a UV-disinfection system are important: 1. The flow through the UV-system should be plug flow for the optimalisation of the UV-system, thus each element of fluid passing through the reactor resides for the same period of time. 2. The flow motion should be turbulent radially from the direction of flow, allowing the water particles to be exposed to the same average intensity in a non-uniform intensity field. 3. The entire volume of the UV-system must be used, avoiding dead spaces and/or short residence time.

Powerful software has come to help us ....

The accurate hydrodynamic calculation method established by Cyclone Fluid Dynamics determines the residence time and path of a particle in an irradiation chamber. The Berson-3D program can calculate UV intensity as function of position in the irradiation chamber. Now, the exact time/intensity can be established in every point of the chamber in mWsec/cm2.

2. Calculation of UV intensity

The UV-intensity is defined as the quantity of UV-light falling on a unit area of surface. Having measured/determined the intensity and UV-C output of the UV mercury discharge lamp we can start calculating the intensity as a function of position in the irradiation chamber.

There are two known intensity calculation methods. One used by the EPA in the USA (UV-DIS 3.1 from Hydroqual Incorporated) and one by Berson UV (Berson-3D Intensity calculation). (Ref. 7 and 8)

Calculation UV-intensity in irradiation chamber: E.P.A. vs. Berson-3D

Both methods use similar parameters in order to reach an intensity value. They use the following parameters:

1. Number of segments, N. per lamp
2. Distance, R. from the lamp
3. UV-power emitted, P
4. Transmission factor in 10 mm of the water, T10
5. Multiple reception points, lp.

Both the EPA and Berson-3D methods divide the lamp into a number of segments "N" which are considered to be near point sources. The methods take into account the contribution of each lamp to the intensity (Watts/m2) at a point "n" in the irradiation irradiation chamber. The EPA method calculates the intensity at each of a number of points "n" in a plane "Z" at the base of the lamp. The average intensity for this plane "Z" is the sum of all reception point intensities " Ip" divided by the number of points "n" (max. 4000) in the plane Z. To calculate the average intensity for the irradiation chamber, a chamber factor (Fz) is arbitrarily introduced. The chamber factor Fz is a correction factor for the position of plane Z. The average intensity of a chamber is equal to the average plane "Z" intensity multiplied with Fz. The Berson UV-techniek method calculates the intensity at each of the reception points "n" in the irradiation chamber. For each point the contribution of all lamp segments UN" to the intensity, in relation to the distance "R" and water transmissivity" T10", is calculated.

The average intensity of a chamber is equal to the sum of the reception point intensities " Ip" divided by the number of points "n" (max. 700,000). A colour graph depicting the UV intensity in mW/cm2 as function of position, is produced by the Berson-3D calculation. UV-intensity is indicated by a specific shade / colour.

3. UV dose

Microbial inactivation depends on the UV-C dose which is described as UV intensity multiplied by exposure time.

As seen above, it is possible to calculate the intensity in mW/cm2 based on the assumed UV-C power emission of the lamp. However it is still necessary to determine the exposure time of a particular particle, during its passage through the irradiation chamber. Important progress has been made by Cyclone Fluid Dynamics in co-operation with Berson UV-techniek to describe the path and velocity of a particle through the chamber using hydrodynamic calculations supported by powerful software. Two examples were compared. 1. Traditional irradiation chamber with a perpendicular inlet and a perpendicular outlet with a parallel flow along the lamps. A wide range can be observed between minimum dose (mJ/cm2) and maximum dose (mJ/cm2), especially at lower transmissions. In short, the system is vulnerable to short circuiting and also applies an overdose in the vicinity of the lamp at the expense of energy cost.

2 InLine chamber irradiation having a straight inlet-outlet and with a perpendicular flow to the lamps, has been compared. Vastly improved hydrodynamic behaviour means this design has high energy efficiency and reduces the risk of short circuiting. The flow through both designs was numerically analysed using the finite volume method. The finite volume method is based on solving the conservation equations, a suitable casted form of the Navier-Stokes equations (the equations which describe the motion of a fluid together with the continuity equation), for each small volume or cell. The most attractive feature of the finite volume method is that the resulting solution would imply that the integral conservation of quantities such as mass and momentum is exactly satisfied over any group of volumes and, of course, over the whole calculation domain regardless of the number of cells. Other strengths of the finite volume method are the easy implementation of additional (physical) models, for example turbulence models, combustion models and multi phase models, and the economical use of computer resources. Hence the majority of today's general purpose fluid dynamics codes use the finite volume method. For this project an implicit unstructured code was used which employs the Simple algorithm for the velocity pressure coupling. The classic textbooks of Patankar and Roache provide an introduction to the world of computational fluid dynamics. (Ref. 9/10).

Meshes of the traditional UV-system and of the new InLine UV-system
(please click the icons in order to get lager images)

Every flow domain is divided into more than 100,000 cells. For every cell a calculation is made of:

• velocity components u, v, w
• turbulence properties (including viscosity)
• pressure p (linked to velocity)

The velocity can be shown as a function of the x, y and z-coordinates in the irradiation chamber.

Velocties in the traditional UV-system and in the new InLine UV-system
(please click the icons in order to get lager images)

In the past it was only possible to calculate average UV dose throughout irradiation chambers, disregarding flow inhomogeneity, radiation inefficiency and short circuiting. The possibility of short circuiting is not taken into consideration in historical calculation methods since the velocity, and thus exposure time, is taken as an average. The importance of knowing the velocity distribution throughout the chamber in flow simulations become important in predicting whether there are circumstances for short circuiting, i.e. short exposure time plus low intensity situations.

The following figures visualise the exposure path of a particle passing through a UV irradiation chamber:

Particle tracks in the traditional UV-system and in the new InLine UV-system
(please click the icons in order to get lager images)

The comparison has been made between a traditional UV irradiation chamber and the Berson UV-techniek InLine system.

% of flowvolume vs. residence time

It is clear that in the commonly used UV irradiation chamber there are some slow particles and some very fast particles. The InLine systems velocity distribution is very much more homogeneous. The bar graph shows clearly a small number of low exposure particles in the traditional design.

% of flowvolume vs. dosis

In graphical presentation, the homogeneity in exposure time distribution is vastly improved using the novel design of an InLine system. Thus, knowing the path of the particle in the UV irradiation chamber and knowing the intensity in the chamber as a function of position "X", combining both calculations, the UV dose as function of the path is known.

4. Correlation between theoretical calculations and practical experiments

In order to verify the theoretical calculations of the Cyclone setup and the calculated intensity of the Berson UV lamps a number of practical experiments were undertaken in the market, in order to prove the feasibility of the proposed system of certification / approval of UV-disinfection systems.The results with the leak-circuit prove of the danger of inhomogeneous irradiation. The co-operation with Cyclone Fluid Dynamics will avoid this.

1. Osaka Japan - effluent disinfection Very important experiments in order to assess the effect of "short circuit". Approximately 2% of total flow was fed through the unit without UV-exposure. This resulted in a maximum microbial reduction at 90% despite doubling of UV-dose at 24 mJ/cm2. Avoiding the 2% short circuit, the microbial reduction was 99.2% at UV-dose at 24 mJ/cm2.
2. Kyoto Japan - effluent disinfection Small scale testing proving systems performance. The calculation proved the experimental results. Microbial reduction 99.97% at UV-dose of 30 mJ/cm2.
3. Concarneau France - sea water disinfection for fisheries Looking for reducing faecal streptococcus requiring a higher UV-dose, system required a different UV-dose. The calculated 100 mJ/cm2 proved to realise the guaranteed microbial reduction.
4. Blaricum Netherlands - effluent. Important testing of A.H. Havelaar/Th. J. Nieuwstad describing UV-disinfection performance. Via a number of UV-units in series the UV-dose was increased from 20 to 80 mJ/cm2. At lower dose practical results correspond with theory.
5. Bodkins WWTP - effluent disinfection This closed MP-Hg-lamp channel started up in 1988. The system proves already for 10 years its stabile performance. The practical results are in line with the theoretical expectation.
6. Hillsborough WWTP - effluent disinfection The customer required 100% microbial reduction. Paying specific emphasis on hydrodynamical design, it has been achieved with U V-dose of 102 mJ/cm2.
7. Culemborg Netherlands - drinking water. The UV-system used after novel Berson InLine design. First correlative test using the Cyclone method 99.999% kill rate achieved.

References

1. J. C. Kruithof: "Chlorination byproducts: Production and Control" A.W.W.A. RFRF/KiwaManualatJ.C.Kruithof/Denver1986
2. J.J. Rook: "Formation of Haloforms during chlorination of natural water", Water treatment exam 23. page 234-245/ 1974
3. J. C. Kruithof, R. Chris van der Leer, W.R.M. Heinen, Nuhn, Houtepen, Fijn: "Ultraviolet disinfection of carbon filter drinking water", In ozon + UV in the treatment of water and other liquids etc. Ref. 12)
4. P. Gelzhauser: "UV plants for water disinfection requirements test and construction characteristics" in Wasser Berlin 1989
5. Dr. G.R. Egberts: "Recent developments in the technology of Low Pressure mercury UV lamps" in Wasser Berlin 1989
6. Philips training course UV lamps /booklet nr. 5
7. United States Environmental Protection Agency: Second draft User Manual for UV DIS version 3.1, UV-disinfection process design manual, Hydroqual Inc., New Jersey. April 1992
8. P. Van Staveren: "Berson Case Berson UV systems; hydraulics and its UV-dose"
9. Patankar, S. V.: "Numerical Heat Transfer and Fluid Flow", Hemisphere, London, 1983
10. Roache, P. J. : "Computational Fluid Dynamics", Hermosa Publ., New Mexico, 1985
11. W. Elenbeas: "High Pressure Mercury Vapour Lamps and their Applications" Philips Technical Library, 1965
12. Th. J. Nieuwstad, A. H. Havelaar: "De desinfectie van gezuiverd afvalwater met ultraviolet licht op technische schaal", Werkdocument 92.139X
13. P. Kreft et. al.: "Hydraulic studies and cleaning evaluation of Ultraviolet disinfection units" Journal WPCF Volume 58 nbr. 12
14. D. Schoenen: "The influence of inhomogeneous irradiation in UV disinfection experimental findings and theoretical considerations" J. Water SRT- Aqua Vol. 45, No. 3, pp. 120- 129, 1996

Please adress all enquiries about this site to our webmaster.
Last updated: 25 januari 1999
© Copyright 1996-1999, Cyclone Fluid Dynamics B.V.