Filter Breakthrough With Two Filters Out of Service

Water quality first: take Filter 4 out of service. Its individual effluent is above the 0.3 NTU benchmark, it is still climbing, and combined effluent will follow within the hour. Because the effluent butterfly valve does not seat tightly, close the influent or isolation valve as well and verify no flow is passing - a "closed" valve that leaks leaves you with an uncontrolled filter.

Then do the capacity math before you commit:

• Now: 7.2 MGD / 4 filters = 1.8 MGD each = 1,250 gpm; 1,250 / 320 ft2 = 3.9 gpm/ft2, essentially at the 4 gpm/ft2 design. You have no margin today.
• With Filter 4 out: 7.2 MGD / 3 = 2.4 MGD each = 1,667 gpm; 1,667 / 320 = 5.2 gpm/ft2, above design. At 4 gpm/ft2 the three remaining filters can carry only 3 x 1,280 gpm = 3,840 gpm = 5.5 MGD.

So removing Filter 4 is not a lever you can pull without cutting production to about 5.5 MGD or less. Check clearwell and distribution storage to see whether 5.5 MGD gets you through the day, alert the superintendent that this is now a capacity decision, and be prepared to ask larger customers to curtail. The wrong move is to leave a failing filter in service to protect throughput.

Low headloss with early breakthrough means the bed is not doing the work - particles are penetrating and passing through instead of being captured in the upper media, so the filter never builds head. That points upstream or to bed condition, not to the filter being "full."

Check, in roughly this order:

• Settled water quality and floc character - turbidity, floc size and strength, coagulant dose, pH, rapid-mix energy, flocculator speed. Weak, small floc shears and penetrates.
• Filter profile - sample at 6-inch depth intervals through the bed. Contaminants near the surface means coagulation problems; deep penetration means media or rate problems.
• Backwash - watch one. Media in the trough or a boiling surface means media loss; inadequate backwash rate or duration leaves dirt that later sloughs; over-expansion can stratify or disturb gravel.
• Media condition - depth versus original, mudballs, cracks, surface boils, gravel displacement, air binding (which shows as erratic or high headloss, not low).
• Hydraulic rate and surges - a filter run near or above design rate, or flow steps from clearwell level control, will punch floc through.
• Ripening - if turbidity was high right after backwash and then improved, it is ripening, not breakthrough. Six hours of steady climb is not ripening.

It reframes the whole event: Filter 4 is a symptom, not the disease. If settled water has tripled, the remaining three filters will break through next, and re-running Filter 4 will only put a clean bed under the same bad water.

Actions:

• Stop chasing filter fixes and go to the coagulation/sedimentation train: verify coagulant pump calibration and solution strength, check for a plugged feed line or a lost prime, confirm rapid mix and flocculator operation, check the sludge blanket, weirs, and whether a hydraulic surge accompanied the raw water change.
• Confirm with a plant jar test using current settled water and the current chemical, not a bench test made up with clean water.
• Hold Filter 4 out of service (or return it only when settled water recovers and its profile is clean), since putting it back early simply reloads a bed.
• Reduce plant flow to buy sedimentation time and protect the three filters that are still good; watch combined filter effluent against 0.3 NTU, and be ready to notify the Illinois EPA if that limit is not met.
• Document the settled water turbidity trend - it is the number that explains the event to the state and to your own staff in review.