Many waterworks utilize sand filters that rely on microorganisms for water treatment. When new, these biofilters require several months for the growth of microorganisms before they become functional. This study is the most thorough documentation of this start-up process to date. Results show that the onset of removal of individual substances proceeds in a well-defined order and that the different filter depths have different functions, suggesting which activities may shorten the start-up period.

This paper concerns the extension and tuning of a genetic algorithm used for the automated design of optimal drinking water distribution networks. Different settings and extensions are tested for their effect on the speed and reproducibility with which the algorithm can produce good results. The fastest combinations are reported. Speed and reproducibility are key conditions for drinking water utilities to include the use of optimization algorithms in the regular design process of mains.