CFD modelling storage for disinfection
Figure 2: Radiation intensity field from UV lamp in a water tank (G = fluence) CFD model: A Boghi

Water disinfection is a key issue in the water industry. There are three main techniques used for this purpose: chlorination, ozonation and ultra-violet disinfection. In order to increase the effectiveness of these techniques, the disinfection unit must be efficiently designed. At Cranfield University, we believe computational fluid dynamics (CFD) is a valid tool for this purpose.

The treatment of drinking water is of vital importance for water distribution systems. The lack of adequate disinfection can result in microbial contamination, and is a serious violation of national and EU regulations. In the UK, Escherichia Coli (E. Coli) organisms are chosen as the indicator for bacterial pollution, while in the USA, the surface water treatment rules use Giardia Lamblia (G. Lamblia).

The main problem in distribution storage management is the conflict between two requirements: quality and quantity. Quality requires the water to spend the least possible time in the tank in order to avoid the proliferation of bacteria; quantity demands the storage of a large amount of water to satisfy the requirements of the distribution network.

In order to eliminate the bacterial contamination, oxidation must be promoted in the tank. This can be achieved two ways: the introduction of chemicals, eg ozone, chlorine, chlorine dioxide and chloramines, or the use of ultra-violet (UV) lamps. These two strategies present different advantages and disadvantages.

CFD model of disinfection tank
Figure 1: Disinfection bio product iso-surface and water velocity vector at the inlet of a chlorine contact tank (Pr = bio-product mass fraction) CFD model: A Boghi

 

Chemical-based disinfection is easy to handle and control; the amounts of chemicals are easy to measure and the system has a low installation cost. However, chemical-based disinfection has some health risks due to the formation of disinfection by-products (DBPs), which arise from excess residual disinfectant in contact tanks (see Fig.1).

CFD modelling storage for disinfection
Figure 2: Radiation intensity field from UV lamp in a water tank
(G = fluence)
CFD model: A Boghi

The use of UV light (see Fig.2) eliminates the risks of harmful by-products. Nevertheless, it is more difficult to assess the disinfection level of this technique, and the operation costs are normally higher than those associated to the chlorine contact tanks (CCT).

The performance of the disinfection tanks depends on their design, which is still largely empirical and relies on field data. The fundamental assumption, often made in the design of disinfection tanks, is that the flow is ‘plug-like’. Unfortunately, this condition can never be achieved due to viscous (wall) effects and turbulence, which inevitably induce a degree of dispersion of fluid elements, and promote tridimensional flow patterns and vortices (see Fig.3).

velocity streamlines in a baffled ozone contact tank
Figure 3: Velocity streamlines in a baffled ozone contact tank
(p = relative pressure)
CFD model: A Boghi

A tank that is designed neglecting the complexity of the local fluid dynamics is likely to have hydraulic losses and not comply with the regulation on water disinfection. Furthermore, the experimental determination of local flow velocities is very difficult and time consuming. The flow-through curve (FTC) can be experimentally determined and used to provide gross information on the hydraulic characteristics of chlorine contact tanks. In addition, the bioassay tests are conventionally employed to evaluate the performance of UV devices. Nevertheless, these experiments are disruptive to the treatment process, and costly.

CFD – a better solution?

A valid alternative to assessing the performance of contact tanks is to simulate the fluid flow using computational fluid dynamic (CFD) models. These have the advantage of being less expensive and less time-consuming than tracer studies.

CFD studies have been extremely useful in showing the details of the three-dimensional fluid flow, and the presence and location of the recirculation regions that trap the micro-organisms and prevent efficient disinfection. The recirculation zones promote short-circuiting and the disinfectant scarcely interacts with the pathogens, which are trapped in these regions. An efficient design should avoid the formation of these areas and improve mixing and disinfection.

The most common solution to eliminate dead zones and short-circuiting, and direct the current to a plug-type flow inside the chlorine contact tanks, is the use of baffles (flow-directing panels), which force the flow into long narrow channels. In a similar way, the position of the bulbs and the UV light intensity must be chosen to deal with the flow patterns.

Challenges

Several approaches have been used in the past to study disinfection with CFD, ranging from depth-average to full three-dimensional models. As with every model, CFD is affected by uncertainties. The first is the treatment of turbulence, which affects almost all engineering applications involving fluid processing. However it should be pointed out that in the case of disinfection tanks, this problem is simplified by the fact that the typical turbulence intensity is low, and therefore more accurate methods can be used for simulation.

The two-phase nature of the flow (air-water) can potentially lead to complication. However, in standard operating conditions, the ratio between the kinetic and the potential energy of the water is so small that there is essentially no sloshing at the water-air interface, and the influence of the air can usually be neglected.

From a modelling point of view, the features that present more difficulties are bacterial inactivation and UV fluence distribution. Indeed the problem of bacterial behaviour modelling is common in many biological sciences, and many proposals put forward have tried to address this issue. Modern water disinfection tanks suggest that the Chick-Watson equation is generally sufficient to describe the inactivation of the micro-organisms in chlorine contact tanks, while the Incomplete Gamma Hom model was found to best describe the bacteria survival curve in ozone contact tanks. As far as the UV fluence rate prediction is concerned, there are many models available. In this case the choice of model is dictated by the availability of computational resources.

Put all this together and it is clear that water disinfection represents an extremely complex problem. CFD modelling, we believe, may be the surest way to deliver the desired outcome in the most efficient manner. By using CFD, industry can ensure that the drinking water reaching the public is contaminant-free, and delivered faster and cheaper than ever before.

www.cranfield.ac.uk