8. CONCLUSIONS
8. CONCLUSIONS
Based on the experience of, converting a full scale conventional filter into a dual media filter, the performance of such a filter, and the other investigations carried out as a part of this study, following major conclusions can be drawn:
Engineering Aspects
(1) Indigenous bituminous coal can be used effectively as a filter media in dual dedia filters with performance on field, comparable to that of anthracite.
(2) In case of new water treatment plants, adoption of bituminous coal-sand dual media filter is going to be significantly more economical as compared to any conventional plant both in respect of capital as well as operation costs.
(3) Conversion of existing rapid sand filters to coal-sand dual media filters can increase the capacity of the existing plant by three times or more and such conversion can be made in less than 30% of the cost of building new installations.
(4) For longer filter runs and consistent low filtrate turbidities,better pretreatment of raw water is essential in case of dual media filtration.
(5) More efficient backwashing is needed for dual nedia filter than the rapid sand filter. Use of high backwash rate (735 1pm/sq.m.) or increased air scouring would be adequate in this respect but provision of surface wash system is preferable.
(6)Dual media filter systems would generally require better operational control. Thus, although operational staff requirement for dual media filter is less per mld capacity, the staff would have to be of a higher technical calibre.
Scientific Aspects
(1) A dual media filter acts essentially as a two stage filter, two in one filter,with top ooal layer removing major portion of the coarse turbidity and lower sand layer doing the work of polishing the filtrate.
(2) Floc characteristics like also of the floc particles and their zeta potentials have pronounced effect on the turbidity removal pattern and distribution of deposits in a dual media bed.
(3) Formation of surface mat in both coal and sand layers and subsequent straining action is essential for efficient filter operation. Ive's relationships like λo α V^-1 or λo α dm^-1 do not seem to be applicable to these surface set layers.
(4) Contribution of lower filter layers to total removal,depends on the pretreatment conditions. The coal layer shows dependence of λo α V^-1 but no definite relationship is observed between λo α dm for the different layers of the same filter bed. Thus the bed of much larger sized coal is able to remove suspended matter as efficiently as the much finer sand does. Since coal has higher specific surface, the higher removals in coal layer could probably be attributed to this.
(5) On the basis of results obtained, Ives' relationship for λo could be modified as follows to be more realistic 
The parameters C and φ for determination of filter coefficient during the filter run (λ) are also seem to have similar relationship,though no exact relation could be derived from the experimental results. It follows then that filter coefficient at any time during the filter run could well be predicted in a better way with the use of similar relationship as proposed for λo.
(6) Overall removal pattern can be explained fully on the basis of combined effect of removal mechanisms like straining, gravity settling, interception and electrokinetic phenomena.
