KOICA Ceramic Water Filter Workshop Project at Ourossogui, Senegal: Final Report of Potters Without Borders
October 17, 2008
Project History
Korea International Cooperation Agency -KOICA invited Potters Without Borders (PWB) of Canada to assist with the initial factory set up of a ceramic water filter workshop at Ourossogui, Matam Region Senegal. The Matam region is located 12 hours overland in the eastern area of the country near the Mauritanian border. This is a very dry desert area of the country. The project set up was scheduled for mid-September to mid October 2008. Representing KOICA were project engineers Je-Min Lee and Joohyun Kim.
The project was begun in collaboration with personnel from the Matam region Water department. In beginning the workshop set up KOICA has committed to an initial trial production of 1000 filter units, (filter element, plastic receptacle and spigot) It is hoped that an arrangement can be made for the transfer of the filter workshop to government or non-governmental hands.
Before arriving in Senegal, Burt Cohen of PWB collaborated with Mr. Je-Min Lee and Mr. Joohyun Kim on the material requirements for the factory set up. Materials were located and contracts set for the manufacture of brick for the ceramic kiln. It was agreed that Potters Without Borders would arrange for the manufacture and transport of a portable press and aluminum filter molds for forming filters.
Upon arrival in Dakar a few days were spent in purchasing materials necessary for the project, which could not be easily found in the Matam region. These materials included plastic bags, scales, propane bottles and burners for the new kiln. Arrangements were made for the transportation of brick from the brickyards near Dakar to the project site. The new English pug mill, which had been shipped to Dakar, could not be located before our departure for Matam region.
The project site is a Matam Region Water Department warehouse which KOICA has been renovating for a number of months. The site is located in the town of Ourossogui not far from the town centre. Drying shelves, a large soaking tank for flow testing filters, as well as an outside metal roofed drying area have been built by the team engineers.
In the preparation for the project start-up, Mr. Kim and Mr. Lee made arrangements with the Matam Region Water Department for contracting government workers to be trained for ceramic water filters production. They also developed relationships with a nearby mechanics shop capable of producing materials needed for the workshop.
Two potters, Ramata Lam and Oan Ji,from the village of Ogo were hired to provide assistance with the forming of the clay filters as well as to identify suitable clays for filters.
Set up period
The first task of the factory set up was the assembly and testing of the ceramic water filter press that had been brought from Managua and Canada.
Three generations of potters form filters at Ourossugui. Ramata Lam (r)- Oan Ji with baby Rugen Ji (l)
Once the filter press was re assembled the staff were trained in the forming of ceramic water filters. Because the pug mill was not available it was necessary to mix materials by hand.
In order to evaluate the clay, simple field tests for plasticity were made. A trip was made to the nearby village of Ogo where our project potters live in order to dig clay. Clay was brought back to the workshop and a local man was contracted to sack and transport clay to the workshop by donkey cart.
A batch series test of filters consisting of six separate proportions of clay/sawdust was formed in order to establish the correct mixture for producing filters.
Mix#1-85.6% clay
Clay 5992 gr85.6%
Sawdust 1008 gr14.4%
Total7000 gr100%
Water 2500 gr.
Mix #2-82.7% Clay
(Represents 20% more sawdust in comparison with mix #1)
Clay 5790 gr82.7%
Sawdust 1210 gr17.3%
Total7000 gr100%
Water 2500 gr.
Mix#3-80 % Clay
Clay 5600 gr80%
Sawdust 1400 gr20%
Total7000 gr100%
Water 2500 gr.
Mix#4-77.5% Clay
Clay 5425 gr77.5%
Sawdust 1575 gr22.5%
Total7000 gr100%
Water 2500 gr.
Mix#5-75% Clay
Clay 5250 gr77.5%
Sawdust 1750 gr22.5%
Total7000 gr100%
Water 2500 gr
Mix#6-72.5% Clay
Clay 5075 gr77.5%
Sawdust 1925 gr22.5%
Total7000 gr100%
Water 2500 gr
A result of the flow testing of the filters (shown below) made during the set up period .
The filters formed showed excellent dry strength. The Mix #1 and Mix #2 filters had excellent physical appearance. An evaluation of the two firings of the kiln showed that many of the filters still had some carbon remaining in the wall of the fired filters. Only two of the Mix #1 filters were in the 1-2.5 li/hr target flow rate but because they still had carbon remaining they too were not acceptable.
Mix# %Clay#of filters testedAverage flow rate li/hr
Mix#185.615 3.7
Mix #2 82.7 8 3.6
Mix#380 7 4
Mix#477.5 7 3.85
Mix #575 2 4+
Mix#672.5 1 4+
Total 40
During the set up period a Mani type kiln was built with brick manufactured near Dakar. The kiln type was chosen because there wasn’t certainty about availability of fuel at Ourossogui. Because of this uncertainty six full propane tanks were purchased in Dakar and transported to the workshop. In addition because the largest size propane tanks available were 10kg tanks too small to maintain adequate pressure, Mr. Kim and Lee also purchased waste oil as a back up fuel and during the kiln set up we constructed a storage and delivery system for burning the waste oil.
Waste oil storage for new kiln
In order to dry newly constructed kiln, local bush wood was purchased from local wood vendors. This wood was easily available locally from vendors and proved of a high calorific value. The first firing achieved temperature-using wood. Although wood is not considered to be a sustainable fuel, it is being used as an interim measure until a better source of fuel is located. In discussion with the government doctor in charge of the regional health department recommended a source of nearby rice husk and numbers of bags were obtained. It is hoped that a forced air delivery system can be set up for using rice husk in the kiln.
Three-donkey cart load of bush wood
New mani-kiln at Ourossogui
The kiln was warmed first before the kiln was put into operation to remove water from construction. During the set up period the kiln was fired twice to a temperature of 900 C.
First Firing Schedule
Hour / Time / Target temperature (Celsius) / Actual top reading / bottom reading30minutes / 7:15am / 32.5
1hr / 7:45am / 75 / 46
1hr 30 minutes / 8:15am / 112.5
2hr / 8:45am / 150 / 151 / 70.5
2hr 30 minutes / 9:15am / 187.5 / 187
3 hr / 9:45am / 225 / 228
3hr 30 minutes / 10:15am / 262.5 / 277 / 160
4hr / 10:45am / 300 / 341 / 252
4 hr 30 minutes / 11:15am / 350 / 362 / 295
5hr / 11:45am / 400 / 392 / 315
5 hr 30 minutes / 12:15am / 450 / 473 / 350
6 hr / 12:45pm / 500 / 518 / 372
6 hr 30 minutes / 1:15pm / 550 / 550 / 406
7 hr / 1:45pm / 600 / 607 / 423
7 hr 30 minutes / 2:15pm / 650 / 663 / 473
8 hr / 2:45pm / 700 / 701 / 504
8hr 30 minutes / 3:15pm / 750 / 753 / 538
9 hr / 3:45pm / 800 / 793 / 583
9 hr30 minutes / 4:15pm / 850 / 867 / 615
10 hr / 4:45pm / 900 / 901 / Soaking
10hr 30 minutes / 5:15pm / 900 / Soaking
11hr / 5:45pm / 900 / Kiln off
Pyrometric cones distributed through the kiln indicated an even temperature in the kiln. But the digital readings showed a substantial difference of temperature from the kiln top and bottom in excess of 200 C. Before returning home I reduced the size of the opening into the chimney flue by 30%.
Final Recommendations on project implementation
Two additional proportions of 15 filters each must be formed, dried, fired and flow tested, in order to make filters in the 1 to 2.5 liter per hour range. Another 15 filters of Mix #1 should be included in this firing.
Mix# A- (10 % less sawdust then Mix #1)
Clay 6092.8 gr87%
Sawdust 907.2 gr 13%
Total7000 gr100%
Water 2500 gr.
Mix# B- (20 % less sawdust then mix #1)
Clay 6192.8 gr88.5%
Sawdust 807.2 gr 11.5%
Total7000 gr100%
Water 2500 gr.
Forming filters- When pressing new filters, the workshop should now leave two coins thickness between the molds in order to make the wall of the filters slightly thicker.
Drying of filters- a large number of filters cracked while drying in the workshop at Ourossogui. In part this was caused by the need to allow drying to proceed more rapidly in order to conduct testing of filters. In part it was only a problem caused by the fact that the proportions of clay/sawdust were not correct.
After forming, new filters are left in the plastic bag in which they are made. The plastic bag is separated from the sides of newly made filter, but not removed. Because of the great heat in the workshop, bags are left around the newly formed filters for two days.
After two days the bag is pulled down to allow the filter’s top to become stiff or leather hard.
The filter is next removed from the plastic bag turned upside down, the bottom plate is removed, the edges are smoothed, a piece of newspaper is put on the bottom and the new filter is put right side up on a wood board. The filter is again loosely covered with plastic to prevent fast drying.
Filters must be carefully and slowly dried. They must be kept out of direct sunlight and if they are drying unevenly they must be rotated. Drying filters must be examined first thing in the morning and before finishing work at the end of the day. Once filters are completely dried, they are no longer a problem as long as they are prevented from getting wet.
Kiln Operation
Before leaving Ourossogui, I reduced the size of the flue (chimney) exit in the kiln floor in order to make temperature in the kiln more even.
The kiln’s firing schedule should be changed to help eliminate carbon in the wall of the filters. I would suggest that the soaking period at the end of the firing should be increased from 45 minutes to 1hr and 15 minutes. The digital pyrometer readings should be carefully recorded to indicate the difference in temperature. Draw trials should be removed from the lower area of the kiln at the end of the firing to examine it for signs of carbon. The next day when the kiln is cool enough to open, if there are any filters that are cracked, they should be broken and examined for signs of carbon in the wall.
Increase the space between the filters when they are stacked in the kiln by using broken pieces of filters as well as clay balls, in order to allow heat to penetrate the filter more completely. Note that filters that have carbon on them may be re fired in the next firing.
Once filters of a correct flow rate are being produced I recommend that the project begin to look for another sponsor capable of continuing production of filters beyond the 1000 filters, which have been agreed to. It is unlikely that the government of Senegal’s regional water authority will be able to maintain the workshop even after the new facility has begun production.
Pyrometric cones may be ordered through Axner and Company in Miami Florida on-line at:
Supplies for digital pyrometer can be ordered from Ward and company at: