Coagulant Underfeed After a Storm in a Conventional Surface-Water Plant

The proposal treats a symptom. Polymer is a filter aid: it strengthens floc at the filter and helps with shear, but it cannot make floc that never formed or never settled. When settled turbidity rises from 1.1 to 4.2 NTU at the same dose, the coagulant is underdosed relative to the new load. Storm water brings more particles and, importantly, more NOM (UV254 rose); NOM consumes coagulant before particles get destabilized. The result is small, weak floc that penetrates the filters, plugs pore spaces, and shortens runs. Excess polymer can also blind media and add TOC.

First actions, in order:

1. Run jar tests on the current raw water across a range of alum doses and pH; do not guess.
2. Verify the alum feed is actually delivering 30 mg/L (drawdown test, rotameter, pump stroke, strainer).
3. Check rapid mix and flocculator operation; a failed flash mixer changes everything.
4. Profile each filter: turbidity and headloss immediately after backwash and through the run. One filter at 0.42 NTU while the combined is 0.26 points to a filter-specific problem (media disturbance, backwash upset, channeling) on top of the coagulation problem.
5. Only after a jar test, raise the coagulant and keep polymer as a modest secondary tool.

Alum consumes roughly 0.45 to 0.5 mg/L of alkalinity as CaCO3 per mg/L of product. At 30 mg/L, consumption is about 14 mg/L, leaving roughly 31 to 32 mg/L of alkalinity and a pH near 7.5. Coagulation is not alkalinity-limited now, which is why the dose can be raised.

If the jar test optimum is 50 mg/L, consumption rises to about 23 mg/L, leaving roughly 23 mg/L alkalinity and dropping pH into the mid-6s. That range often works well for alum, but it leaves little buffer and requires the finished-water pH to be recovered for corrosion control.

If the optimum is 60 to 70 mg/L, alkalinity consumption reaches 27 to 32 mg/L and pH depression becomes the limiting factor, not alkalinity per se. Options:

• Feed caustic soda or soda ash after coagulation to restore pH; confirm the corrosion-control target can still be met.
• Switch to a polyaluminum chloride coagulant, which typically removes the same turbidity with less alkalinity demand.
• Use a cationic polymer as a primary coagulant or aid to lower the metal-salt demand. Whatever is chosen, confirm the dose in jar tests measuring settled turbidity, filtered turbidity, pH, and alkalinity, and record the change.

Buy time with filtration and flow control rather than chemistry alone:

• Put the worst filter (0.42 NTU) on filter-to-waste and return it to service only when its individual turbidity is stable and low.
• Use filter-to-waste on any filter for the ripening period after backwash; ripening spikes are the usual cause of individual filter exceedances.
• Slow filter starts and avoid rate changes during a run; hydraulic surges shear floc through the media.
• Reduce plant throughput if clearwell inventory allows: at 70 percent of 1.1 MG there is about 770,000 gallons in storage, which provides hours of buffer at 12 MGD. Lower loading rate means lower filtered turbidity.
• Track each filter individually (turbidity and headloss) every shift, not just the combined value, so the trend shows whether the coagulant change is working before the combined value moves.
• Confirm disinfection contact time and residual are maintained if flow is shifted between clearwells, and record everything for the compliance file. If combined effluent approaches 0.30 NTU, treat it as an emergency: cut flow, take the marginal filter out of service, and notify the primacy agency as your plan requires.