Stacked Rapid Sand Filtration

Introduction

A vertically stacked filtration system meets many of the AguaClara Project constraints. Both normal filtration and back wash operations are gravity driven and require no electricity. In additon, it is an open system. The required construction materials(PVC pipes, sand, concrete, and rebar) are relatively cheap and abundant in Honduras. Most importantly, sand filtration is proven to consistently lower the effluent NTU standard to below 1 NTU.

A stacked rapid sand filtration unit is preferable to a conventional rapid sand filtration unit as it requires less water to back wash since the filters are stacked, and use the same water for all filter layers to back wash. In addition, a stacked filtration system requires less planned area to achieve the same level of filtration.

Figure 1: Vertically Stacked Filtration Unit Design

For more information, please see the explanation of stacked rapid sand filtration theory and unit design.

Keep up to date with [Reflection Reports] and Future Challenges for the Stacked Filtration Team!

Current Research

Bench Scale Model Design
To test the theory behind the stacked rapid sand filter, a bench scale model was designed and built. Using this model, a series of experiments will be conducted to test the filtering capacity of a stacked rapid sand filter.

[Back Wash]
The most important challenge to prove is the ability to clean the stacked rapid sand filter by the back wash method. In order to back wash the filter, we must prove that the individual filter layers can be fluidized in sequence, and thus result in a fully fluidized filter. Once fully fluidized, the filter can then be back washed, and require significantly less water than a comparable conventional single layer filter.

[Performance Comparison]
In order to learn if the flow distribution in the stacked rapid sand filter system is working reasonably well, it is necessary to conduct an experimental comparison to a conventional single layer filter of the same depth as a single layer of our stacked filter (20 cm).

Head Loss Measurement
The head loss across the filter over time is the governing factor that determines the effective filtration time. The build up of head loss ultimately determines how often the filtration unit will require back washing.

Future Research

Filtration Unit Design
It is necessary to design a filtration unit that can be built at our existing water treatment plants, and be incorporated into future plant designs. This includes a hydraulic analysis to calculate the critical elevations for how the filter fits in with the sedimentation tank effluent weir and the distribution tank. This analysis must be done using the approach for the design tool with correctly named variables, and with the standard inputs of flow rate and then automated design of the filter system. This automated design must include full hydraulic design of all of the piping required.

[Plane Assumption Test]
In our initial design, we assumed that as long as the spacing between the filtration tubes in each layer is small in comparison to the depth of a layer of sand, then the flow of water out of the tubes will converge and form a plane of filtration. For our purposes, we assumed that a ratio of 1 to 10 of tube spacing versus filter depth was small enough. Laboratory experiments should be conducted to find the depth which will still allow the tubes to be modeled as a plane. The number of inlet and outlet tubes per plane can also be varied to see if our modeling of the layer of tubes as planes of filtration is accurate. We can decrease the space between the tubes to find the necessary ratio between the sand layer depth and the necessary spacing between the tubes that will allow us to model the layer of tubes as a plane.

Figure 2: Plane Area Concept

[Performance Study]
A number of parameters will need to be studied in order to find the optimal design conditions for a filtration unit. The variables to be studied experimentally include diameter of sand grains, depth of each filter layer, horizontal spacing between pipes, and filtration velocity. A robust study of all these variables will require a long period of time, and it is thus recommended that we proceed with the design and installation of a filtration unit with conservative design parameters. Once full-scale performance can be analyzed, we can proceed with an exhaustive performance study to further optimize the stacked rapid sand filtration unit design.