Application of computational fluid dynamics to water and wastewater treatment plants (1996)

The capital infrastructure for water and wastewater treatment will have to be increased considerably in the immediate future. It is important that new and existing plants are designed to operate as efficiently as possible. The increased activity in this sector will put pressure on experienced designers. One way to promote efficient design and to de-bottleneck existing equipment is to use modern computer techniques.

All water and wastewater treatment equipment rely on through flow of water. Some processes require this flow to be well mixed, whereas others require plug-flow. Examples of the former are activated sludge plants, chemical dosing zones and anaerobic digesters while sand filters (in both filtration and back-flush modes), clarifiers, adsorption columns (ozone, activated carbon and ion exchange) and dissolved air flotation cells are examples of the latter. Some processes such as nutrient removal activated sludge plants require the combination of both plug flow and completely mixed reaction zones.

Laboratory-scale experiments are usually operated under ideal flow conditions; unfortunately this is often not carried through to full-scale plants, due to the greater difficulties involved. For example, observation of clarifiers has often indicated regions of excessive up-flow velocity. The flow rate through the clarifier must be limited to a value that maintains the concentration of suspended solids, which are carried over into the clarified product water, at an acceptable level. If the flow distribution throughout the clarifier could be improved to achieve more uniform upflow through the judicious addition of baffles, or through modifications to the inlets or outlets, then the clarifier would either produce the better product quality at the same flow, or an increased flow at the original quality. Either case would result in reduced operating costs or delayed capital expenditure.

Techniques identified in a previous Water Research Commission project, Project No. 363, entitled The Development of Small-scale Potable Water Treatment Equipment enabled the quantification of overall fluid distribution in equipment. These techniques are purely diagnostic, and do not predict what remedial measures can be taken to overcome the causes of flow maldistribution. Computational Fluid Dynamic (CFD) modelling has a mechanistic basis, so that it has the potential to both quantify the degree of maldistribution, and to make estimates concerning the improvements that could be achieved by postulated remedial measures. CFD is therefore a technique which can yield solutions for problems, rather than simply identifying and quantifying them.

Questions have frequently arisen, during the course of other Water Research Commission projects, which seemed to require the use of CFD to obtain satisfactory answers. For instance, in Project No. 238, Research on the Design Criteria for Cross-flow Microfiltration, trials were carried out on increasing the concentration of suspended solids in an anaerobic digester at the Durban Corporation Northern Waste Water Treatment Works. Concern was expressed by the Corporation as to the effects that the increased sludge viscosity will have on the mixing characteristics of the sludge. In view of the very large size of the digester, experimental investigation of this issue would be very difficult and costly. CFD would provide a means to estimate the effect, and help evaluate suggested methods of improving the mixing. Residence time distribution tests on the digester were also undertaken. Initial indications were that that the actual sludge residence time is significantly less than the design value, the implication being that there were significant dead volumes in the digester. The use of CFD might suggest ways to enable the full utilisation of the digester volume. Another field in which CFD modelling might find profitable application is that of mixing and flocculation, a topic that was also explored in WRC project No. 238. Flocculation and floc break-up are critically dependent on shear rate and turbulent intensity, and these generally vary tremendously from point to point in typical mixing chambers found at water treatment works. It is frequently very difficult to determine representative values to characterise such equipment. With its detailed representation of the fundamental flow mechanisms, CFD has the potential to provide major advances in the understanding of large scale flocculation processes.

It is noteworthy that designers and operators of water treatment plant in Europe and America, such as Thames Water and Lyonnaise des Eaux, have begun to use CFD techniques, and to employ CFD specialists. CFD expertise also exists in South Africa, notably at the Potchefstroom University for Christian Higher Education, however the application of CFD techniques in the South African water industry has been limited. It was one of the goals of this project to promote the use of CFD in the local industry by demonstrating this potential to solve practical problems, as well as by starting to train students in the hope of creating a core of CFD practitioners for the future.