2.7 Filter Backwash, Surface Wash, Air Scour & Filter Surveillance

Filter Backwash, Surface Wash, Air Scour & Filter Surveillance

Filtration is the last physical barrier between Cryptosporidium and the customer. The ABC Water Treatment outline names "backwash data," "filter performance," "perform filter surveillance tests," and "measure and analyze filter media to determine compliance with design specifications" as distinct tasks. This section covers the wash cycle and the surveillance program that keeps it honest.

---

1. Why Filters Are Washed

A filter run ends for one of three reasons, and the reason matters diagnostically:

A filter that always terminates on time but never on headloss is being over-washed, wasting water and shortening media life. A filter that terminates on turbidity breakthrough is a coagulation problem, not a filter problem.

---

2. Backwash Hydraulics

The purpose of backwash is fluidization: expanding the bed so grains move freely and abrade each other, shearing off attached floc.

Rate and expansion

• Typical backwash rate: 15 to 20 gpm/sq ft for dual-media beds in warm water, adjusted upward in cold water because higher viscosity increases drag and lifts the bed more at a given rate.
• Target bed expansion: 20 to 50 percent of the unexpanded media depth.
• Rise rate conversion: rise rate (in/min) = backwash rate (gpm/sq ft) x 1.604.

Worked example

A filter bed measures 20 ft by 22 ft and is washed at 7,000 gpm.

1. Area = 20 x 22 = 440 sq ft
2. Backwash rate = 7,000 / 440 = 15.9 gpm/sq ft
3. Rise rate = 15.9 x 1.604 = 25.5 in/min
4. If the wash runs 8 minutes: volume = 7,000 gpm x 8 min = 56,000 gallons
5. If the filter produced 3.2 million gallons during the run, backwash water is 56,000 / 3,200,000 = 1.75 percent of production - a healthy figure; above about 4 percent warrants investigation.

Cold-water correction

Water at 4 degrees C is roughly 1.5 times as viscous as water at 20 degrees C. The same backwash rate therefore produces materially more expansion in winter. Plants that never reduce the winter wash rate throw media over the wash troughs - which shows up months later as measured media depth loss.

---

3. Auxiliary Scour

Water alone does not clean media well because fluidized grains stratify and move in parallel, with little inter-grain abrasion. Auxiliary scour supplies that abrasion.

• Surface wash - fixed or rotating arms with nozzles above the media, operated at roughly 2 gpm/sq ft for 1 to 2 minutes before and during the early part of the backwash. Targets the mudball-forming top few inches.
• Air scour - air injected through the underdrain at roughly 3 to 5 scfm/sq ft, usually with no or low simultaneous water flow in a "collapse-pulsing" regime. Air scour is more effective than surface wash for deep beds and for anthracite.

Sequencing rules

1. Never air scour with the bed fully fluidized at high water rate - the air and water together can carry media over the troughs and can drive fines into the gravel.
2. A typical combined sequence is: drain to a few inches above the media, air scour alone 2 to 4 minutes, air plus low-rate water 2 to 3 minutes, air off, then full-rate water wash to fluidize and carry solids out, then a slow rewash or filter-to-waste.
3. Rate-of-rise should be ramped, not stepped. A sudden full-rate start can lift the gravel support and cause gravel upset - a permanent defect that shows as sand in the clearwell and localized boils during wash.

---

4. Ripening and Filter-to-Waste

Immediately after a wash, clean media has poor particle-capture efficiency and the bed still holds backwash remnant water. The result is the ripening spike - a turbidity peak that can reach 0.3 to 1.0 NTU and that carries a disproportionate share of the filter run's total Cryptosporidium risk.

Filter-to-waste (rewash) diverts effluent to the washwater recovery basin or drain until turbidity settles. Practical operating targets:

• Run to waste for 5 to 30 minutes, or until individual filter effluent turbidity is below 0.10 NTU.
• If the plant has no filter-to-waste piping, use a slow-start / delayed-start ramp - bringing the filter back at 25 to 50 percent of normal rate for the first 10 to 15 minutes - which substantially blunts the spike.
• Never return a filter to service at full rate immediately after backwash if the plant is at peak demand; that is the moment the ripening spike goes straight into the clearwell.

Filter aid polymer (a nonionic or anionic polymer at 0.01 to 0.05 mg/L) applied to the filter influent strengthens floc attachment, shortens ripening, and delays breakthrough - at the cost of faster headloss buildup.

---

5. Turbidity Compliance and Reporting

Two different turbidity limits apply, and candidates confuse them:

The IFE follow-up triggers under the Interim Enhanced Surface Water Treatment Rule:

• Above 1.0 NTU in two consecutive 15-minute readings - report to the primacy agency with the cause.
• Above 0.5 NTU in two consecutive readings at the end of the first four hours of operation after backwash - report with the cause.
• Above 1.0 NTU in two consecutive readings in three consecutive months - conduct a filter self-assessment.
• Above 2.0 NTU in two consecutive readings in two consecutive months - arrange a comprehensive performance evaluation (CPE) by the state or a third party.

---

6. The Filter Surveillance Program

Surveillance answers a question turbidity alone cannot: is the media still what the design assumed? A full annual inspection includes:

1. Media depth measurement. Probe or excavate at several points and compare to design. Losses above roughly 10 percent of design depth, or anthracite loss above about 2 inches per year, indicate wash-rate or trough problems.
2. Core sampling and sieve analysis. Determines effective size (d10) and uniformity coefficient. Anthracite typically d10 0.9 to 1.1 mm with a uniformity coefficient below 1.5; sand typically d10 0.45 to 0.55 mm.
3. Floc retention analysis. Media samples from several depths are washed and the recovered solids measured. High retention in the lower bed means solids are penetrating too deeply.
4. Mudball measurement. Media is sieved and mudballs volumetrically measured. Under about 0.1 percent by volume is good; above about 1 percent is a problem; above 5 percent is severe.
5. Expansion test. A calibrated probe or a series of small collection cups on a rod measures actual expansion at operating wash rate. Verify 20 to 50 percent, and verify it is uniform across the bed - non-uniform expansion means underdrain plugging or gravel displacement.
6. Visual inspection of the drained bed. Look for cracks (indicating a compacted, shrinking bed), craters, boils during wash, sidewall shrinkage (which creates a short-circuit along the wall), and uneven media surface.
7. Underdrain and gravel inspection whenever media is removed - the only time gravel upset can be confirmed and corrected.