1. INTRODUCTION
Most of
the water treatment plants in India
are being overloaded in an attempt to cope up with the ever increasing water
demand. The existing rapid sand filters form a bottleneck for such overloading
with the result that part of the water has to be bypassed without filtration,
or it finds its way through cracked filter beds, thereby escaping treatment.
Poor pre-treatment and non-availability of skilled personnel for operating the
plant make the situation worse, resulting in large and unpredictable changes in
treated water quality. A fact-finding survey by NEERI revealed deteriorated
conditions of water treatment plants in India during 1972 and the
conditions have remained practically the same till now. It is felt that
conversion of existing rapid sand filters to high rate dual media filters may
form the most practical and economical interim measure that could be adopted to
improve the conditions.
No doubt, the solution to the problem
lies in improvement in pre-treatment conditions and augmentation of plant
capacities. But it must be admitted that skilled personnel and equipment
necessary for proper control of coagulation and availability of financial
resources for installation and maintenance of improved pre-treatment
facilities, or construction of supplementary units are beyond the reach of
most of the water works in India.
Does it mean that the present state of affairs should
continue till these obstacles are overcome?
Fortunately,
dual media filters have a capacity of functioning even under conditions of
poor pre-treatment and increased flow. Experience in the West as well as in India confirm
this. Another advantage in adoption of these filters is that existing filters
can be converted to dual media filters with minor changes at a very low cost.
The design of such filters is really a difficult task as no established norms
can.be adopted due to limitations imposed by existing conditions. The present
paper tries to evolve a rational approach for design of such converted filters.
Design of dual media filters for new
installations has not been considered here as it would be based on entirely
different aspects.
2. DESIGN CONSTRAINTS
Design of any filter is
based on suspension characteristics, filter structure, operating parameters
and expected filtrate quality. Considering these, the constraints imposed by
prevailing conditions in the water treatment plants can be described as
follows:
(i) Suspension Characteristics:
Due to uncontrolled alum dose, ineffective
flocculation and short circuiting in settling tanks, the water applied to the
filters has large concentration of turbidity, with significant percentage of
large settleable flocs alongwith partially stabilized but unfloccu- lated
colloids. The observations at Kanpur Water Works indicated that the turbidity
range is from 15-30 FTU, with about 40 per cent of large floc turbidity which
invariably gets strained in the top filter layers.
This can
be taken as a representative characteristic of applied water, as in most of the
cases the situation is more or less identical. Moreover, along with rise in
turbidity levels, there is a change in nature of turbidity also. Thus in rainy
season the turbidity levels may shoot upto 100 FTU and during late winter and
summer "difficult to flocculate" algal turbidity may pose problems.
(ii) Filter Structure:
Most of the filtration plants are provided
with rate controllers and have submerged inlet flow conditions, thus limiting
the standing water depth over the filter: medium. The structure of the filter
does not provide any allowance for increasing the depth of filter medium,
which is restricted to 0.6 to 0.75 m. The limiting head loss available for
filtration seldom exceeds 2.0 m, thereby preventing longer filter runs
(iii) Operating Parameters:
The filters are operated
by semiskilled personnel and in some plants, even by coolies; as such, the
control of filter operation is very poor. The head loss gauges, rate
controllers and discharge measuring devices are out of order in most of the
cases due to negligence. The backwashing is done on a routine time cycle,
decided on the convenience of shift duties. As a result the filter run is
terminated only on the basis of time interval without considering limiting
head loss, turbidity of filtrate, etc. Any new design of the filter must take
into consideration this lacuna in operation, which may not be altered unless
skilled personnel are employed. Therefore the filter should be designed to give
good filtrate quality and acceptable filtration rate during the entire filter
run which should be a pre-determined time interval.
(iv) Expected filtrate quality:
The current design
practice in developed countries aims at filtrate turbidity to be consistently
below 0.2 FTU. However, the filtrate turbidity at such a low level, though
desirable, is neither economically feasible under present conditions, nor
essential from the point of view of acceptable drinking water standards which
permit turbidity up to 5 FTU. Considering the techno-economical feasibility,
the filtrate turbidity of 1 FTU may be taken for design.
3. SELECTION OF HIGH RATE FILTER
Of all the
high rate filters available, dual, multimedia and mixed media filters can be
adopted easily for converting old filters. The multimedia filters, though more
suitable, may not be practically feasible due to difficulty in the availability
and cost of procuring and placing different filter media. Mixed media filters
advocated by Culp and Culp suffer from the same drawbacks. Dual media filters,
though not truly coarse to fine graded, still provide an effective and
feasible alternative.
4. DESIGN OF DUAL MEDIA FILTER
Filter Media:
Due to
non,.availability of anthracite in India, various other materials like high
grade bituminous coal (Paramsh:vam1 et al 1973, Ranade et al'" 1975),
crushed coconut shell (Kardile8 1972), berry seeds (Bhole and Nashikkar1 1974)
and kernels of stone fruits like apricots (Ranade and Agrawal 1974) have been
tried and were found to be suitable as coarse filter media. Considering the
cost, availability and filtration properties, high grade bituminous coal seems
to be a better choice. However when local conditions favour, use of indigenous
filter media may be made as the coarse media. Recent attempts at using plastic
granules as filter media may be explored further in. this respect. The lower
layer would be of sand for all such filters(ii)
Size Gradation:
High percentage of large
flocs in the influent to filter suggests the use of much coarser medium to
prevent surface mat formation, whereas improper flocculation of some colloids
necessiatates greater depth of coarse medium followed by fine medium, with a
sharp interface. However, it is reported that use of fine sand produces
intermixing. The use of high density fine medium like garnet immediately below
the coarse medium may be tried, in order to achieve optimum design. Such combination,
though it seems to be strange, may give better performance. How
ever, from availability and cost points of view, use
of garnet sand may not be feasible.
Opinions differ regarding the desirable
amount of intermixing at the junction of coarse and fine media. Culp and Culp
recommend significant intermixing to achieve ideal "coarse to fine"
gradation and claim better filtrate quality and longer filter runs. However
there are sufficient reasons to adopt a sharp interface for the filters
designed to suit Indian conditions. The turbidity of applied water is of two
distinct types and requires two-stage treatment consisting of fine media
filterfollowing a coarse media filter. There is less control over the grading
of filter media and if the gradation is selected to have significant
intermixing there may remain large portion of fine material which may cause
dense packing of intermixing layer, thus defeating the purpose of coarse to
fine gradation. This suggests that some degree of in termixing is inevitable
under present conditions but the design should aim at minimising it.
As regards selection of
size grading of filter media, the norm of equal hydraulic settling may be
followed as a guide line. But to ensure minimum intermixing, the 10% (wt basis)
coarsest grains of top medium and 10% (wt basis) and finest grains of fine medium (Effective size)
should be considered as representative sizes, as they would determine the
degree of intermixing. According to Brossman and Malina if the size ratio is in
the range of 2 : 1 for coarse and fine media (in case of coal and sand),
separation of layers would occur and the same norm can be adhered to in the
design. If the densities of coarse and fine media are assumed to be 1.2 and
2.65 respectively as in case of coal and sand, the specifications may be
stipulated as follows:
Coarse media:
Size range 0.85 to 1.6 mm
E.S. = 1.00 mm U.C. = 1.3 to 1.5
Fine media:
Size range 0.55 to 0.9 mm
E.S. = 0.60 mm U.C. = 1.3 to 1.5
Depth of Media:
Experiments conducted
with various combinations of relative depths of individual media, under high
rates of filtration have conclusively proved that coarse media with a depth of
40 to 50 cm can effectively remove large floc turbidity without excessive head
loss build-up and can flocculate unflocculated particles. Turbidity reaching
the lower sand layers is always less than 5 FTU, even under varying influent
turbidity concentrations. The greater depth also prevents clogging of beds when
algal suspensions are encountered. Considering the permissible standards for
filtrate turbidity, sand layer
of 15 cm thickness is more than enough to ensure
filtrate turbidity of less than 1 FTU during the entire filter run.
Use of greater depth of
coarse medium also helps in preventing surface mat formation at the interface.
It may be argued that a sharp interface would favour such surface straining,
but it should be pointed out that the turbidity reaching interface being low,
the surface mat formation would take a much longer time as compared to that in
a conventional rapid sand filter. The backwash interval for Dual media filters
would be much less than for conventional filters, depending on turbidity
storage capacity of coarse layer. An increase of head loss due to surface mat
formation would not limit the filter run.
normal rates of filtration, becomes a
controlling factor at higher rates as the head loss increase with the square of
velocity. Fortunately a change in the number of openings on the existing
laterals can solve the problem. However, this must be done judiciously after
studying the hydraulics of the system, otherwise backwashing may become
non-uniform and ineffective after such modifications.
(v) Operation of the Filter :
variable conditions of turbidity and
filter components, it is recommended that the filter should be designed to operate
with declining rate range of 300 Ipm/sqm to 200 lpm/sqm with 12 hours of filter
run. The design is on a much conservative side and would guarantee good
filtrate quality. There may be objection to adoption of 12 hours' filter run as
the present practice is to have 24 hours filter run. But it should be
remembered that the filtration capacity of the converted filter is more than
double that of the original filter and a little increase in operation cost is
fully justified as the percentage of backwash water requirement is constant.
Moreover this can be adopted without any special training to operators. The
backwash procedure would remain unchanged for the converted filters(vi) Performance
of the Filter:
Ample data
are available on the performance of dual media filters, both on pilot scale
and field scale, it has been established that these filters can operate at high
rates of filtration with consistent high removal efficiency of turbidity and
bacteria. Use of such filters in existing water treatment plants would improve
filtrate quality and augment the capacity at a relatively low cost.
It must be pointed out here that conversion
of filter is only an interim measure and improvement in the pre-treatment
conditions cannot be overlooked.
As the funds become available pretreatment can be
improved by:
1. Proper coagulant dose and efficient
mixing arrangement.
2. Use of polyelectrolytes to streng
then the floc.
3. Effective flocculation of all the colloids.
4. Improvement in settling tanks and augmentation
of their capa
cities by use of tube settlers.
5-. Employment of skilled personnel
and special training facilities.
Once these changes come into effect
the same dual media filters would give excellent filtrate turbidities with
longer filter runs of 24 to 36 hours, as the
inlet water would then have turbidity
of less than 5
F. T.U. and would be of
filterable nature.
(vii) Cost of Conversion :
The cost
of the conversion includes the cost of modifications in influent and
underdrainage system and cost of placing new media. From the experience gained
at Kanpur and Ichhalkaranji Water Works it can be said that this cost is around
Rs. 6,000 per mld of plant capacity or Rs. 3,000 per mId of increased capacity
If a conversion of a fair sized filtration plant is done the
cost can be further reduced5.
CONCLUSION
Dual media filters offer
the most appropriate and economical solution -to the present problem of
overloaded and inefficient water treatment plants in India. The novelty of these filters
1s that they can be incorporated in the existing system without much change in
plant structure or method of operation. The technical know-how for such
conversion is fully developed in India.
and the process of conversion of filters can start on
a large scale, awaiting only the decision of the concerned authorities to
solve such -urgent problems.
REFERENCES
1.
"Survey of Water Treatment Plants", Technical Digest No.3, CPHERI, Nagpur, (January 1971).
2. Culp G.
L. and Culp R. L., "New Concepts in Water
Purification", Van
Nostrand Reinhold Company, New York,
(1974).
3. Conley, W. R., "High Rate
Filtration",
J.A.W.W.A.
64, 3, p. 203,
(1972).
4. Ranade
S. V., "Engineering and Theoretical Investigations on Dual Media Filters
using Indian Bituminous Coals", Ph.D. Thesis, J.I.T., K;anpur,
(July 1976).
5. Ranade
S. V., Agrawal G. D. and Misra Y. D., "Conversion of Rapid Sand Filter
into Dual Media Filter", J. Inst. Public Health Engrs., J., 2, p. 12,
(1975).
6. "Manual on Water
Supply and Treatment", Central Public Health and Environmental Engineering
Organization, Government of India,
(1977).
7. Paramsivan et al., "Bituminous
Coal A substitute for Anthracite in two Layer Filtration of Water", Indian
J. Env. Health, 15, p. 178, (1973).
8. Kardile, J. N. (1972) "Crushed
Coconut Shell as a New Filter Media for Dual and Multilayer Filters", J.I.W.W.A.
1, 1, p. 28, (1972).
9. Bhole A. G. and Nashikkar J. T.,
"Berry
Seed Shell as Filter Media", J. Inst. o! Engrs 54, PH 2, p.
45, (1974).
10. Ranade S. V. and Agrawal G. D.,
"Use of Vegetable Wastes as a Filter Media" presented at the
Conference on Engg. Materials and Equipment, The Association of Engineers,
Cal lcutta (1974).
11. Brossman, D. R. and Malina J. F.
Jr., "Intermixing of Dual Media Filters and Effects on Performanl;e",
Technical Report EHE 72-4 CRWR 86, Env.
Engg. Lab., University
of Texas, Austin, Texas,
U.S.A., (1972).
13. Gadgil J. M., "Effect of
Media Depths on Performance of Dual Media Filters", M.E. Thesis, Walchand College of Engg., Sangli, (1979).
14. Ranade S. V., Agrawal
G. D. and Misra Y. D., "FulJ Scale Trials on Converted Dual Media
Filter", J. I.W.W.A., (October 1976).
