Filter Breakthrough After a Storm at a 6-MGD Conventional Plant
Riverton WTP treats 6.0 MGD of reservoir water with alum (32 mg/L), flocculation, sedimentation and four dual-media filters of 350 sq ft each. Design loading is 3 gpm/sq ft. A two-day storm pushed raw turbidity from 4 to 26 NTU and color from 10 to 45 units, and raw pH fell from 7.2 to 6.6. Settled water turbidity is 1.6 NTU and climbing. The alum feed has not been changed since before the storm. Filter effluent turbidity was 0.08 NTU last week; it is now 0.24 NTU on two filters and 0.38 NTU on filter 4, the oldest unit. Combined filter effluent is 0.19 NTU. Run times have dropped from 62 hours to 26 hours, and headloss climbs fastest on filter 4. Free chlorine residual and CT are on target. The chief operator is off-site and the board wants to know whether the plant can stay in service.
1. Before you touch the alum feed, what do you check, and what does each check tell you?
Read the worked answer
Separate the plant-wide problem from the single-unit problem.
A filter that fails while settled water is still 1.6 NTU points at the filters; a filter that fails while settled water is climbing points at coagulation. With 26 NTU and 45 color units of raw water the coagulant demand has almost certainly gone up, so the chemistry is the prime suspect - but filter 4 is behaving differently and must be explained separately.
Checks, in order:
- Jar test at the current 32 mg/L and one or two higher doses, using today's raw water. This is the only way to know whether the dose is short rather than guessing.
- Filter loading rate: 6.0 MGD = 4,167 gpm; 4 x 350 = 1,400 sq ft, so 4,167/1,400 = 2.98 gpm/sq ft - already at design. If anyone takes a filter out of service the rate goes to 3.97 gpm/sq ft, above design, so do not backwash one filter at a time carelessly.
- Filter 4 specifically: media condition, bed depth loss, media boil or mudballs, underdrain air binding, backwash rate and duration, and whether Run 4's polymer feed is actually reaching the filter.
- Settled water turbidity and pH every two hours; settled turbidity above about 2 NTU will overwhelm any filter.
- Headloss and turbidity versus time since backwash - a terminal headloss curve that flattens means surface blinding; a curve that rises fast with early turbidity breakthrough means the media is passing floc.
That combination tells you whether you need a chemistry change or a filter repair before you spend the adjustment on the wrong thing.
2. Combined filter effluent is 0.19 NTU and filter 4 is 0.38 NTU. Can the plant stay online? Give the decision rule you would use.
Read the worked answer
The numbers alone do not force you offline, but the trend says filter 4 is failing and you are one step from doing so.
- The combined filter effluent limit under the SWTR is 0.3 NTU in at least 95% of monthly measurements, never above 1.0 NTU. At 0.19 NTU you are compliant today, but a single filter at 0.38 NTU can drag the combined value over 0.3 as it degrades.
- The individual filter effluent limit is 1.0 NTU, so 0.38 NTU is not a violation. It is a warning.
Decision rule I would apply:
- Run the jar test and adjust the coagulant; give it one filter run to show a response.
- Reduce plant rate to about 5 MGD (2.48 gpm/sq ft) to buy settling time - this is the cheapest control you have and it does not compromise compliance.
- Put filter 4 on filter-to-waste if you have it, and increase its polymer slightly.
- If individual filter effluent on filter 4 reaches 1.0 NTU, take it out of service and confirm the remaining three filters stay at or below design rate at your reduced flow.
- Document every 15-minute reading. Federal filter-assessment triggers and your state's thresholds attach to sustained individual filter readings and to monthly combined values, so the record you keep now decides whether you owe an assessment later.
Tell the board: stay in service, at reduced rate, with filter 4 on a short leash.
3. The jar test shows 40 mg/L alum gives 0.9 NTU settled water at pH 6.4; 32 mg/L gives 1.7 NTU at pH 6.8. Raw alkalinity is 42 mg/L as CaCO3. What are the tradeoffs of raising the dose now?
Read the worked answer
Better settled water versus a corrosion-control problem you will own for months.
- Settled water at 0.9 NTU instead of 1.7 NTU roughly halves the solids load on the filters and would likely restore run times. That is a real operational gain.
- Alkalinity and pH: alum consumes roughly 0.45–0.5 mg/L of alkalinity per mg/L of alum. Going from 32 to 40 mg/L consumes about 3.6–4.0 mg/L more alkalinity out of only 42 mg/L, and the jar test already shows pH dropping from 6.8 to 6.4. At pH 6.4 you are at the edge of the optimal alum coagulation range and you begin to risk poor floc and post-precipitation of aluminum in the distribution system.
- Distribution impacts: lower pH water leaving the plant affects lead and copper corrosion control and the lime or caustic feed must be able to compensate. Check that the post-treatment pH trim can hold your target before you commit.
- Practical decision: raise the dose to about 36–40 mg/L combined with a compensating caustic or lime feed to hold settled pH near 6.8–7.0, rather than raising alum alone. Confirm with a jar test that includes the pH adjustment - a jar test without the caustic is testing a condition the plant will never see.