Foam Filtration Reflection Report
Primary Authors: Rachel Philipson, Catherine Hanna, Melissa Shinbein, Kevin Wong
Primary Editor: Rachel Philipson
AguaClara Reflection Report
Cornell University
School of Civil & Environmental Engineering
Ithaca, NY 14853-3501
Date Submitted: 24/09/2010
Date Revised: 28/19/2010
Abstract
The purpose of the foam filtration team is to create a new and effective way of filtering water for a variety of systems both on large and small scales. Over the summer, experiments were run in hopes of developing a foam filtration unit to reduce the effluent turbidity from an AguaClara plant from approximately 5 NTU to below 1 NTU. Since it was determined that foam filtration is not feasible for an AguaClara plant, we will be designing a point of use filtration unit for apartment complexes or schools. We will test the performance of foam filtration for a variety of influent turbidities, pore sizes and determine the head loss through the foam. Additionally, we will submit our filtration unit design to the EPA P3 competition.
Introduction
Polyurethane foam was first used in air filters before research revealed its applicability in the water treatment field. It is a viable filter medium alternative to rapid sand filtration for the removal of colloidal particles. Because the foam is highly porous and permeable, the head loss across the foam can be significantly less than that of conventional filtration media. However, the use of polyurethane foam has yet to achieve the same removal efficiencies per unit depth compared to conventional filtration. In our study, we will model polyurethane foam as a depth filter and determine the filtration parameters for optimal removal of turbidity. From our previous studies, we have found parameters for the foam that are ideal for filtering water, such as pore size, depth, and flow velocity. Additional tests are needed to determine the head loss through the foam, the effect of layering different pore sizes of foam, and the range of influent turbidities that are treatable. We propose that it is possible to design a foam filtration system that can reliably treat turbid water on a large scale (schools or apartment buildings) and on a small scale (households). Our research goal is to gather the experimental data needed to support the system design.
Previous work:
During summer 2010, the foam filtration team worked on designing a foam filtration unit that could reliably treat the typical AguaClara effluent water with a turbidity of about 5 NTU to a turbidity of less than 1 NTU. We ran multiple experiments with different velocities, foam depths, influent turbidities, and alloy coating. The important insights we gained from the project were as follows:
(1) The experiments revealed that filtration performance (pC*) decreased with increasing velocity (Figure 1).
Figure 1 - Maximum average pC* versus water velocity
(2) We determined that a flow velocity of 3 mm/s in the filter was able to achieve the US turbidity standards of .3 NTU? for the longest period of time.
(3) Some of the experiments developed a head loss sufficient to compress the foam and terminate the experimental trial. We need a way to predict when the foam will compress based on measurements of head loss.
(4) With increasing filter depth, particles are more likely to become trapped in the porous media (Figure 2). It is interesting to note that we observe an inversely proportional relationship between effluent turbidity and filter depth, which is consistent with conventional sand filtration theory which suggests that filtering capacity is a function of filter depth.
Figure 2: Average best effluent turbidity for filter depths of 12.7, 25.4 and 38.1 cm (approach velocity of 6 mm/s and influent conditions of 5 NTU and 1.5 mg/L alum).
(5) Filter performance decreased with increasing pore size, as smaller particles become more difficult to remove.
(6) Effluent turbidity for foam that was previously exposed to alum resulted in similar performance to that with an alum feed. Even better results were achieved using an alum feed and foam that had not yet been exposed to alum
Experimental Design
This semester, we will run a variety of experiments to test the filtration properties of polyurethane foam such as the head loss through the foam, varying influent turbidities and the effect of layering pore sizes of foam. We will be using the same experimental apparatus used in summer 2010 (Figure 3). Our first task will be to determine the head loss through the foam with clean water running through the filtration unit. Additionally, we need to determine the head loss through the foam at collapse, or failure. When enough particles build up on top of the foam layer, the pore size in the foam approaches zero and the foam sheets collapse under the pressure exerted by the flow rate in the filter. We will also take pictures of the foam column throughout this experiment so we can have visual documentation of the collapse. The one parameter we will be measuring in this experiment is the pressure across the foam layer, which will be measured using a pressure sensor. Unfortunately, because of noise in the pressure sensor measurement, we will not be able to obtain an exact reading for the head loss in the "unclogged" foam, but rather an order of magnitude approximation for the head loss through the foam. In future experiments, we will also be monitoring the influent and effluent turbidity in the system, however we do not yet have turbidimeters, and in order to start getting experimental results, we've chosen to run this experiment without monitoring the turbidity. The influent turbidity will be approximated using a measured mass of clay added to a set volume of water, and this synthetic raw water will be dosed with alum to coagulate the particles and facilitate the clogging of the filter. We will use a 1.5 mg/L alum dose, which was the dose used in previous semester's experiments.
Figure 3: Filter Foam Experimental Setup
Results and Discussion
In the past two weeks, we have worked on the preliminary designs for a point of use foam filtration unit:
Influent water flow rate control: For the filter to operate, we want to design a flow control device that will maintain a constant flow rate as the main flow changes. It is not practical to treat the influent water on demand because flow rates are often variable and it would leave the system idle for long times. It is proposed that we use a gravity system with a stock tank that utilizes a float valve, which closes the opening of the tank when too much water has been added. This system is used to maintain a constant amount of head in the tank, allowing for the control of the effluent flow rather than the influent flow of water.
The foam filtration unit itself will consist of using a 5-gallon cylindrical non-tapered bucket to hold foam pieces of diameter approximately equivalent to 11 inches (ppi count and depth to be determined). A tube connected to an entrance port at the top of the bucket will transfer constant flow into the filtration unit from the storage tank. Multiple holes at the bottom of the bucket will allow flow to exit the container. The number and size of holes at the bottom are determined so that the flow into the container is equivalent to the flow out of the container. In other words, the holes will be designed to allow for constant flow through the filter. Another container will be placed below the filtration unit to catch the exiting filtered water. This container sends captured water to the chlorination dose controller.
The exit chamber of the filtration unit will be comprised of a chlorination tank, a chlorine doser, and a storage tank large enough to store 24 hours worth of water. The chlorine doser will have the same design as those employed by current AguaClara plants and will drip a steady concentration of approximately 0.1125 mg/L of chlorine into the water.
Future Work
Because we have not yet built a water distribution system and have no turbidimeters, the experiments have been off to a slow start. Because of these delays, we are mainly focusing on designing the filter unit for the EPA P3 competition. For the EPA P3 grant, team members have been designing the foam filtration unit on a larger scale to hopefully provide for larger groups of people in places such as schools and apartment buildings in areas without a municipal water treatment facility.
Once the apparatus is completed, the first experiment will be to measure the head loss through the foam both while functioning and at failure. This head loss value can still be determined for water without a specific influent turbidity. When the turbidimeters arrive, experimentation on aspects such as varying turbidities and pore size will commence.
Team Reflections
Because of we are starting our work from scratch in terms of analyzing foam filtration at the point-of-use level versus our previous analysis at the AguaClara filtration plant level; we have not yet begun experiments. Instead, our group has been focusing on setting up our experimental apparatus to model a point-of-use filtration unit as best as possible and on the design for our full-scale point-of-use foam filtration unit.
Given the decision to examine foam filtration as a point-of-use system, our experimental apparatus has required considerable modification from the past year. Also, previous supplies accessible to our group were being used by another class. Thus, in the past few days, we have focused on revising our materials list to match our new apparatus, and on ordering more turbidimeters and acquiring all necessary materials for experiments. The turbidimeters are expected to arrive within the next two weeks, and our experimental apparatus is nearly assembled.
Even though we have not yet begun to obtain physical data, the foam filtration team has accomplished a lot in terms of large-scale design. Our group has learned about expectations for the design of the desired final apparatus through our preparations for the EPA P-3 competition. By letting each member design a part of the filtration unit and present their proposal to the rest of the group, everyone is updated on information about the design as a whole and are all able to discuss and brainstorm different ways of improving the unit.