Filter Run Collapse After Coagulant Feed Change
A 12 MGD conventional plant with four dual-media filters has run 30-day filter runs for months. On Monday the coagulant is switched from liquid alum to a blended polyaluminum chloride product at the same 18 mg/L dose. By Thursday, two filters have dropped to 11-hour runs. Turbidity on those filters climbs from 0.08 NTU to 0.22 NTU within 4 hours of a backwash, then creeps to 0.35 NTU before ripening never fully settles. Headloss rises faster than run time predicts. Finished water turbidity averages 0.14 NTU and the chlorine residual is stable. The jar tests from Monday showed good floc at 18 mg/L, but the operator notes the floc looked 'pinpoint and dense' compared with the old alum floc. Filter aid polymer feed is unchanged at 0.05 mg/L.
1. What is the most likely cause of the shortened runs, and what specific tests would confirm it before you change anything?
Read the worked answer
The pattern points to a coagulant chemistry mismatch, not a mechanical filter failure. Evidence: runs collapsed only after the coagulant switch; floc is described as pinpoint and dense (small, strong flocs that penetrate media and blind the surface); turbidity breaks through early and never ripens; headloss climbs faster than expected. PACl blends have different alkalinity demand and a different optimum dose/pH than alum, so an equal-mass dose is not an equal-performance dose.
Confirm before acting:
- Run jar tests on current raw water at a dose sweep (12, 15, 18, 21, 24 mg/L) and measure settled and filtered turbidity, not just floc appearance.
- Check raw and settled pH and alkalinity against the pre-switch baseline; PACl may be depressing pH or over/under-reacting.
- Compare filtered turbidity vs. time curves (filter ripening profile) for a switched filter and a not-yet-switched filter.
- Verify the actual feed rate with a drawdown or gravimetric check; 'same dose' assumes the product density and strength match, which it may not.
- Inspect media on one filter for surface matting or mudballs to rule out physical causes.
2. The jar tests now show 15 mg/L gives the best settled turbidity. Describe how you would implement the change and what you would watch for over the next two weeks.
Read the worked answer
Implement deliberately and in a controlled way:
- Confirm the 15 mg/L optimum across at least two raw water conditions (morning and afternoon), since PACl performance is more pH-sensitive than alum.
- Recalculate the actual chemical feed pump setting using the product's density and active strength; don't just scale the old number.
- Adjust one filter at a time or one half of the plant if piping allows, so you keep a comparison.
- Verify alkalinity and pH after the change; add alkalinity if pH drops below the target for corrosion control.
- Re-optimize filter aid polymer separately, because the new floc size changes the polymer demand.
Watch over two weeks:
- Filter run time and ripening turbidity profile per filter.
- Filter effluent particle counts if available; turbidity alone can miss small-particle passage.
- Backwash efficiency: check for media loss, bed expansion, and mudball formation.
- Finished water turbidity trend and any change in chlorine demand.
- Document the dose change, jar test results, and dates so the change is traceable.
3. A board member asks why you can't just raise the filter aid polymer to fix the short runs. How do you respond, and what is the operational risk of that shortcut?
Read the worked answer
Raising polymer treats the symptom, not the cause. The early turbidity breakthrough and fast headloss indicate the floc is too small and penetrating the bed; more polymer may improve capture temporarily but can blind the filter faster, shorten runs further, and create a sticky surface mat that is hard to backwash. Overdosing polymer also risks finished-water polymer carryover, which can raise turbidity, affect disinfection, and create a compliance problem.
The honest answer to the board: the root cause is coagulant chemistry and dose, and the fix belongs at coagulation, not at the filter. The polymer is a polishing tool, not a substitute for proper coagulation. I would explain that we are testing the correct dose, will change it in a controlled way, and will report run times and turbidity weekly. The risk of the shortcut is shorter runs, media fouling, possible finished-water quality exceedance, and a harder problem to unwind later.