Much of the talk today goes round and round on whether or not granular media filters are capable of effectively preventing the passage of Giardia Lamblia and Cryptosporidium into potable water distribution systems. They can and do. What seems to get lost in the talk is that many of the same principles that apply to granular media filters also apply to other types of filters, including many membrane filter types.

When service flow rates are increased beyond a certain point, contaminant breakthrough will occur regardless of the filter type. Let a filter reach its solids holding capacity and breakthrough will occur without regard for granules or membranes. Plugging a filter just means its neighboring filter(s) take on an additional service load, or total production is reduced. Choose the wrong granular media or membrane material/pore opening and either effluent quality will be compromised or intolerable headlosses will result. Inefficient cleaning of a granular media bed or reusable membrane reduces the capacity of the next filter run. And the list goes on.

The 32' x 62' filter shown was originally built with a dual-parallel-lateral ceramic block underdrain design. A stainless steel panel with punched slots was installed over each block to delete the need for a gravel base beneath the filter sand and filter coal. Open areas of the punched slots were engineered based on measurements of backwash mal-distribution.

The science of water treatment/purification has not yet defined the minimum number of viable units of Giardia Lamblia orCryptosporidium that constitute a health risk. Those of us over 45, who grew up in small population centres, well remember the number of times the "stomach flu" made the rounds in our schools. It seems the symptoms of that mysterious "stomach flu" are the same as those of a mild infection from Cryptosporidium. Could it be? And yet, those of normal health recovered in a matter of days. That is still the case today in mild infections from Cryptosporidium.

While the science is not definitive, this is known. Viable units of Giardia Lamblia are the size often removed by a granular media filter operated at so-so efficiency. Typically, granular media filters should be expected to remove units greater than 10 microns. Reducing the number of particles in the 2-10 micron range to single digits or low double digits requires optimization, and that usually means rehabilitation. In rehabilitation, optimization means everything in the treatment train will be brought up to today's standards. The key word is everything.

As the size of particles gets smaller, their number increases. So, reducing the number of these particles greatly reduces the risk of passing sufficient numbers of viable Cryptosporidium units to result in ill health in humans.

To accomplish removal of such fine particulate often requires treatment train rehabilitation, rather than just filter rehabilitation. Little is to be gained from rehabilitating a filter if the contaminants to be removed are not rendered filterable by some chemical pretreatment program. The entire process cannot be sustained if parts of the train cannot keep up with those rehabilitated.

So, the first step in any rehabilitation program is a comprehensive evaluation of the treatment train and its attendant hardware. AWI's Filter Audit^TM accounts for everything down to the seemingly insignificant, as does a financial audit.

To start at the real beginning requires a thorough analysis and interpretation of the raw water's chemistry. Developing filtration for contaminants often goes beyond chemical coagulation, flocculation, settling, and clarification. Determination of the particle size distribution in the water to be filtered is a must. Compact portable membrane filters and portable particle counters quickly establish the efficiency of the chemical treatment program.

Assessing the filterability of the contaminants permits design of a granular media bed and filtration procedure. It is now a proven fact that highly angular sand produces water of lower NTUs and has a higher solids holding capacity than round sand, all other things being equal. Consciously choosing sand shape, effective size, and uniformity coefficient is one of those seemingly small things, but one which helps in keeping the Crypto out of consumers' taps.

On-site pilot testing can develop the data needed to choose a mono-medium or dual media bed, specifications of the media chosen, and media depths. Old media bed designs, using granular material larger than could be fluidized during backwash, compromised effective bed size because of retained particulate. This condition produced invisible channeling, which is little more than nature's way of increasing the rate of filtration without you knowing it. A more effective bed is one that can be fluidized to the correct degree during backwashing, thereby providing a cleaned filter bed close to 100% of its original capacity.

Service flow rates are site-specific, plain and simple. Original plant designs used different criteria than criteria pertinent today. Technology has provided instruments capable of measuring much smaller particulate that the regulator says must be removed. The outstanding plant design of a decade ago may not be capable of producing the quality needed today. Rehabilitation, therefore, must take that "good old plant" and make it better than new. And its being done every day at a cost well below new plant price or conversion to membrane filtration. The beauty of rehabilitation of existing processes and hardware is continuing flat-line operating costs as compared to the cost of regular membrane replacement.

Complete rehabilitation of the removal process is comprised of three equal parts: chemical pretreatment (i.e. making contaminants filterable), filtration (i.e. media selection, bed specifications, and operating procedures), and sustainability.

Sustainability begs an answer to these questions. How can the media bed be cleaned to very nearly 100% of its rehabilitated design at the conclusion of each filtration cycle? Is air scour a fit? How is better cleaning accomplished with less water? Answer these questions correctly and the concern goes away over shrinking effective bed size, mudballs, channeling, dead spots, and media migration/gravel mounding.

If the filter underdrain in use performs well elsewhere, but poorly in the plant in need of rehabilitation, then it too must undergo work that will remove its in situ deficiencies, regardless of the manufacturer's original claims. Assuming pumps provide adequate volumes of backwash water, the underdrain must be capable of distributing water within a fraction of zero mal-distribution. It's simple! Mal-distribution results in some areas of the bed getting more water than others, resulting in different localized cleaning efficiencies. So what? So watch the effluent quality deteriorate over time.

If the filter in need of rehabilitation still has the old style multiple layers of support gravel, watch seemingly insignificant levels of mal-distribution produce gravel migration, gravel mounding, and loss of media into the underdrain through a supposedly impenetrable gravel bed.

It is an interesting phenomenon that was grappled with for years, but is now understood. Media and support bed migration is a never-ending work in progress, and in some instances the movement can be measured after each backwash cycle.

The late J.B. Hambley, P.Eng., founder of AWI, agonized over these and other thorny treatment issues, but found the answers to gravel mounding/migration and mal-distribution. He developed a series of orifice formulas that prevented mal-distribution in underdrain headers and laterals. Designing a highly efficient air scour delivery system is a science of its own. Hambley made air scour an integral part of his Flexscour^TM Stainless Steel Filter Underdrain, but left the choice to use it or not with the client. A further rehabilitation advancement is a stainless steel device, the AWI Phoenix Panel System that makes many block and false-bottom filters gravel-free while also mitigating mal-distribution of the existing underdrain.

For more information, contact: R.G. Mondoux, P.Eng., 2209 Lakeshore Road, Burlington, Ontario L7R 1A8, Tel/Fax: (905) 632-3080.

Reprinted from September 1998 issue.