von Gunten, UrsTournefier, Alan2021-04-202021-04-202020-10-05https://infoscience.epfl.ch/handle/20.500.14299/177440Saline groundwater by the shore of Lake Victoria in Eastern Uganda is unsuitable for human consumption, making the contaminated lake water the only possible source for drinking water. In such context, five water kiosks were constructed in 2015 and 2018 to provide safe water for children and households within the school catchment area of some communities in Busia and Namayingo districts. The small-scale treatment units are based on gravity-driven membrane (GDM) ultrafiltration technology and designed to not depend on any other energy supply than their own solar pumping system. Up to 10’000 L can be treated per kiosk on a daily basis with only little maintenance requirements. The study evaluated the water quality of the five GDM kiosks, their technical and financial performance, as well as the adequacy of two automatic chlorination devices specifically designed for the systems. A field trip of three months conducted in Autumn 2019 focused on the implementation of the T-Chlorinator and AkvoTur and data collection. The microbial quality was assessed by quantifying the faecal indicator Escherichia coli (E. coli) and the filtration performance determined by measuring the removal of probiotic enterococci added in the raw water, using HyServe Compact Dry plates. The stability of the chlorine dosing was addressed with daily and weekly free residual concentration (FRC) measurements. In addition, 35 households participated in an experiment to assess the residual disinfection capacity in 20L jerrycans after 24 hours of storage. The technical and financial performance of the kiosks was estimated based on pressure, power and flowrate measurements and from book records kept by the operators. Results showed that a yearly deep maintenance of the filtration module is required to ensure the integrity of the membranes. Nonetheless, the GDM kiosks could treat the lake water with a bacteria log-removal > 4.5. Permeate did not contain E. coli in 81.2% of the samples. Moreover, only 2.4% of the analysis were attributed to a high risk of faecal related diseases. Nevertheless, the treated water was subjected to recontamination in the safe storage tank or at the tap, mainly due to a poor maintenance of the systems. A monthly shock chlorination was evaluated as insufficient to guarantee the water safety. Instead, a continuous process showed satisfying results in term of water quality. Particularly, the T-Chlorinator was judged as an adequate technology and was able to dose FRC within 1.5 - 2.5 mg/L in 87.5% of the tests. The operator could properly handle the device, which would represent less than 15% of the potential kiosk revenue. However, the residual capacity after 24 hours of storage was negligible (< 0.2 mg/L). Chlorine decay of 1.9 mg/L was observed. Further investigations on long term sustainability and the implementation of the device in other kiosks is recommended. The GDM filtration (> 65 mbar) could cover the daily water demand of 1.7-2.9 m3 in less than 6 hours. Membrane fluxes stabilised between 4-12 L/m2/h which is consistent with other studies. The assessment of a monthly flushing procedure demonstrated partial flux recovery up to 117%. Under optimal conditions, it was found that the solar pumping system could abstract and transport 3.7 - 7.1 m3 depending on the sites. Nevertheless, a constant operation was compromised due regular shutdowns. Furthermore, a new type of solar pump should be implemented in two kiosks due to a high total dynamic head. Despite some operation, maintenance and management issues, the GDM kiosk were able to operate until now. To ensure the long term sustainability of the water kiosks, further investigations should look into the commitment of new organisations to supervise the business management and the water quality monitoring.drinking water treatmentmembrane filtrationautomatic chlorinationdrinking water qualitylow income countrywater kioskrecontaminationstudent work::master thesis