Page 210 -

HIGH-RATE GRANULAR MEDIA FILTRATION 8.29

Each filter's effluent turbidity should be monitored and recorded continuously (max-
imum 15-min intervals) using an on-line turbidimeter to detect variances from normal op-
eration immediately. Observations of the effects of fluctuations in source water quality,
rate changes, equipment malfunctions, chemical feed variations, filter washing, and other
such occurrences contribute to the operator's understanding of the plant's performance
and increase the ability to deal with such situations.
Turbidity measurement is sometimes used to automatically initiate a filter wash cycle
or to actuate an alarm whenever the filter effluent reaches a preset maximum turbidity
level. Most water treatment plants establish a filtered water turbidity goal (say, 0.1 ntu)
well below current regulatory standards.
Turbidity of the waste washwater and filter-to-waste water can also be monitored to
assess the performance of filter washing and ripening, respectively. A different type of
turbidimeter, capable of measuring higher turbidity levels, can monitor waste washwater.
Sample piping should be designed so that it does not collect air bubbles that can dis-
tort readings (Letterman, 1994). Air bubble traps, available from most turbidimeter sup-
pliers, can be used for this purpose. On-line turbidimeters should be located where they
are readily accessible because they require periodic cleaning and calibration to ensure ac-
curate readings.
Additional information on turbidity measurement and methods can be found in the
USEPA Guidance Manual for Compliance with the Interim Enhanced Surface Water
Treatment Rule Turbidity Provisions (1999).

Particle Counting. Particle counting is rapidly gaining acceptance to monitor filter per-
formance (Lewis, Hargesheimer, and Ventsch, 1992). Particle counters are instruments
that can quantify and size particles in water by light-scattering techniques and can be ei-
ther discrete sample or on-line type. Particles ranging in size from approximately 1 to 500
/xm can be quantified by particle counters. Particle counters can provide a direct mea-
surement of the number of particles in a particular size range and serve to verify the log
removal of particles achieved.
Particle counters are particularly useful in determining the log removal of particles in
the Giardia and Cryptosporidium size ranges. Giardia cysts generally fall in the 5- to
15-/xm range, and Cryptosporidium oocysts fall in the 4- to 7-/zm range; however, the
correlation between particle counts and Giardia and Cryptosporidium concentration has
not been fully established. Filtration theory indicates that 1- to 3-/xm particles are the most
difficult to remove, but Moran et al. (1993) found earliest breakthrough of 3- to 7-/zm
particles in test filters. The 3- to 7-/xm size range includes cysts and oocysts currently of
primary concern.
Current standards require a 2- to 2.5-1og removal of Giardia cysts and a 2-log removal
of Cryptosporidium oocysts by a filtration process, depending on the type of filtration.
Future regulations may increase the Cryptosporidium oocyst log removal requirement. For
very clean source waters, particle counters may not demonstrate proper log removals be-
cause of the low required filtered water particle counts. Even in these situations, fluctuations
in filtered water particle counts can still give an excellent indication of filter performance.
Engineers and plant operators should be aware that at present no standard method of
analysis has been established, so particle counts can vary from instrument to instrument.
Sensors associated with particle counters may have different operating principles such as
particle size detection limit, particle concentration range, sample flow rate, or pressure re-
quirements. Bends and flow-altering devices should be avoided upstream of particle coun-
ters because they can alter performance. Chemical and physical cleaning of particle coun-
ters is required and should be considered in their installation. The operator should be aware
that stable particle counts are difficult to obtain when samples are switched periodically
between filter effluent lines.