Filter Effluent Creep and a Sinking Finished-Water pH
Your 6 MGD conventional plant treats a surface water with raw turbidity of 14 NTU and total alkalinity of 30 mg/L as CaCO3. Alum is fed at 32 mg/L and there is no other coagulant, polymer, or pH adjustment. Over three days, settled turbidity has climbed from 2.5 to 4.0 NTU, filtered effluent from 0.08 to 0.32 NTU against an internal limit of 0.10, and finished pH has fallen from 7.2 to 6.1. Filter runs have shortened from 40 hours to 22 hours. A jar test this morning at the existing alum dose gave settled turbidity of 3.8 NTU at pH 6.1 and 1.1 NTU when the same dose was applied with pH raised to 6.9. The state requires finished water pH of at least 7.0 for corrosion control, and lead/copper monitoring is due next quarter. The plant has caustic soda and hydrated lime on site.
1. What is the most likely cause of the turbidity and filter-run deterioration? Support the diagnosis with the numbers given.
Read the worked answer
The arithmetic points at coagulant-driven alkalinity consumption, not at a bad raw water day.
- Alum consumes roughly 0.45 to 0.5 mg/L of alkalinity (as CaCO3) per mg/L of alum. At 32 mg/L that is about 15 mg/L consumed, leaving only ~15 mg/L of the original 30 mg/L.
- With that little bicarbonate left, finished pH drops to 6.1. The jar test is the decisive evidence: same dose, same water, but pH 6.9 gives 1.1 NTU settled versus 3.8 NTU at pH 6.1. The coagulant dose is not the controlling variable; pH is.
- Poor floc at depressed pH produces pin floc that passes the filters (0.08 to 0.32 NTU) and blinds the media early (40 h to 22 h runs).
Do not respond by simply raising the alum dose. That would consume more alkalinity and drive pH lower still, making the problem worse.
2. What would you change today, and how would you verify the change before committing to a full filter run?
Read the worked answer
Feed an alkali ahead of the alum and prove the dose in a jar test before it reaches the filters.
- Estimate the alkali demand: you need to recover the ~15 mg/L of alkalinity the alum consumes plus enough excess to land finished pH between 7.0 and 7.4. Caustic at roughly 15 to 20 mg/L as CaCO3 equivalent is the starting point; hydrated lime is the backup if caustic supply is tight.
- Confirm in jars at the existing 32 mg/L alum, then re-run the series at lower alum doses (24, 28 mg/L) once pH is correct, since the optimum dose often drops once pH is in range.
- Feed point matters: add alkali before the coagulant flash mix so hydrolysis sees the right chemistry, and check for scale or lime carryover in the clearwell.
- Verify with settled turbidity (target 2 NTU or less), filter effluent turbidity trend, filter run length, and finished pH/alkalinity every two hours until stable.
3. Raising pH has a disinfection cost. What else must be re-examined and how would you handle the trade-off?
Read the worked answer
Higher pH shifts free chlorine from hypochlorous acid toward the far less effective hypochlorite ion, so the CT credit you earn per mg/L of residual drops as pH rises. At pH 7.5 roughly half the free chlorine is HOCl; by pH 8.0 it is closer to a quarter.
- Recompute CT at the new pH and the coldest water temperature, not at today's temperature. If the CT margin was thin at pH 7.2, it may now be noncompliant.
- Balance rather than maximize: hold pH at the low end of the corrosion-control target (7.0 to 7.2) rather than 7.4 to 7.8, and pick up any lost CT with contact time or residual rather than pH.
- Watch DBP formation: more chlorine and higher pH both push THM/HAA formation. Track TTHM/HAA5 at the next quarterly set.
- Re-verify the corrosion target itself: the pH that solves the coagulation problem should also satisfy the lead/copper optimum, which is why the pH target must be documented before you turn the caustic pump up.