Storm Load, Fixed Alum Dose, and a Request to Push Production
A 6.0 MGD conventional surface water plant treating a lake supply has run at 5.2 MGD since a storm. Raw turbidity rose from 11 to 46 NTU, raw alkalinity fell to 42 mg/L, and the plant kept the same alum dose of 26 mg/L. Settled turbidity is 3.4 NTU against a goal of 2.0 or less, finished pH is 7.0, filter runs have dropped from 58 hours to 20 hours, and combined filter effluent turbidity is bouncing between 0.22 and 0.38 NTU. Free chlorine leaving the clearwell is 1.1 mg/L. The filters have not been profiled in two years and no jar tests have been run since the storm began. Your superintendent wants to raise production to 6.5 MGD to refill the elevated tank before the weekend.
1. Diagnose the most probable cause of the turbidity and short filter runs, and list the specific data you would examine to confirm it before changing any chemical feed.
Read the worked answer
Most probable cause: the plant is underdosing coagulant for the new raw water. Storm runoff roughly quadrupled the particle and organic load (11 to 46 NTU) while the alum feed stayed at 26 mg/L, so the floc is small and weak. Weak floc does two things at once: it blinds the filters (runs drop from 58 h to 20 h) and it passes through them (0.22 to 0.38 NTU). Low raw alkalinity makes it worse: 26 mg/L of alum consumes roughly 12 mg/L of alkalinity as CaCO3, which drags pH toward or past the edge of the coagulation optimum.
Confirm before you touch the feed:
- Jar tests on today's raw water at 26, 32, 38 and 44 mg/L alum, with and without caustic or lime, reading settled turbidity at 20 to 30 minutes and noting floc size and settle rate.
- Coagulant feed verification: draw-down or calibration-cylinder test of the metering pump, stroke and speed against the dose you think you are feeding, and day-tank inventory against the meter totalizer. A 20 percent feed error looks exactly like this.
- Rapid-mix energy and flash-mix detention. A partially plugged static mixer produces poor floc even at a correct dose.
- Individual filter effluent turbidimeters and headloss trends: are runs ending on turbidity breakthrough or on headloss? Breakthrough at 20 hours points at coagulation; headloss points at solids loading and rate.
- Settled turbidity and particle counts as the process-control numbers, not just finished water turbidity.
If jar tests show a clear optimum and the pump is calibrated, the problem is downstream in mixing or filtration, not chemistry.
2. Do you raise production to 6.5 MGD? Give the decision and the specific actions you would take before and during the rate change.
Read the worked answer
No. Do not raise to 6.5 MGD while filtered turbidity is intermittently over 0.3 NTU. Higher rate means higher filter loading, shorter runs, and a much greater chance of a reportable filtered-water exceedance. Adding flow to a plant in turbidity trouble tests its worst failure mode.
What to do instead:
- Hold rate, or reduce to about 4.5 MGD. The headloss math is simple: cutting flow roughly 15 percent typically extends run time by more than that, and it protects the 0.3 NTU limit while you fix chemistry.
- Tell the superintendent the trade plainly, in writing: the tank can be refilled at a lower sustained rate overnight, or from another source, but not by pushing a plant that is already passing floc.
- Run the jar tests and optimize dose and pH first. Consider a filter-aid polymer only after coagulation is right; polymer on top of poor coagulation just moves the problem into the media.
- Re-check disinfection: with 1.1 mg/L free chlorine at pH 7.0, verify the CT calculation at the flow and clearwell level you will actually run. A rate change changes CT, and the clearwell may not be full.
- Log the decision, the supporting numbers, and the notification. If you are overruled, the record shows what you knew and when.
3. Combined filter effluent is reaching 0.38 NTU. At what point does this become a compliance and reporting problem, and what would you do about the individual filters?
Read the worked answer
The combined filter effluent limit is 0.3 NTU in at least 95 percent of monthly measurements, and never above 1.0 NTU. At 0.22 to 0.38 NTU you are already intermittently over the line, so this is a compliance issue now, not later.
What to watch and do:
- Individual filters: 0.3 NTU or more in two consecutive 15-minute readings requires follow-up (self-assessment and repairs, escalating if it repeats over months); 1.0 NTU or more in two consecutive 15-minute readings is an immediate notification event. Read each filter's turbidimeter, not the combined number.
- Record every reading with time and filter ID. You cannot reconstruct compliance after the fact, and reconstructed data looks like what it is.
- Profile the filters. They have not been profiled in two years. Run a turbidity and headloss profile across the bed at the end of a run to find channeling, media loss, mudballs, or a leaking or mispositioned valve.
- Take the worst filter out of service for media inspection if optimizing the dose does not bring its effluent back under 0.3 NTU within a run or two. Running a known-bad filter to keep rate is the choice that ends in a violation and a notice.
- Keep the jar-test sheets and filter logs together; they are your evidence that you detected the problem and acted.