Coagulation Failure at a 30 MLD Surface-Water Plant During a Turbidity Spike
A 30 MLD surface-water treatment plant draws from a river subject to monsoon spates. During a storm, raw water turbidity rises from 20 NTU to 800 NTU, pH is 7.8, alkalinity is 90 mg/L as CaCO3, and water temperature is 26°C. The plant has two flash mixers, three flocculators, four clariflocculators, six rapid sand filters and chlorination. The alum dose is fixed at 25 mg/L and there is no coagulant aid. Within six hours, filtered-water turbidity rises to 3–5 NTU, filter runs fall from 30 h to 8 h, residual aluminium reaches 0.25 mg/L and chlorine demand increases sharply. Operators have increased backwashing frequency, but the filters now reach breakthrough sooner. The plant must maintain filtered-water turbidity below 1 NTU and a chlorine residual of 0.2 mg/L at the consumer end. The laboratory has a six-place jar tester but no streaming-current detector.
1. Diagnose the likely causes of filter breakthrough and short filter runs.
Read the worked answer
Short runs and early breakthrough point to poor coagulation and flocculation rather than a filter defect alone. At 800 NTU the solids load on the filters is about 40 times the normal load; if the alum dose stays at 25 mg/L, the coagulant is likely underdosed. The floc may be small or sheared by excessive mixing, so it passes the clarifier and blinds the filter surface.
The rise in residual aluminium to 0.25 mg/L supports incomplete reaction or soluble aluminium carry-over. The filters may also have media loss, mudballs or inadequate backwash, but the simultaneous drop in performance of all six filters after the storm points to upstream carry-over. Chlorine demand rises because the organic load and ammonia associated with the turbidity spike consume chlorine.
The first step is to stop relying on a fixed dose and run jar tests. Backwashing alone will not solve a coagulation problem; it may even worsen media loss.
2. Outline a jar-test programme and explain how you would select the coagulant dose and pH.
Read the worked answer
Collect a fresh 20 L raw-water sample and measure turbidity, pH, alkalinity, temperature and UV254. Use six jars:
- rapid mix at 200 rpm for 1 min;
- flocculate at 30–40 rpm for 15 min;
- settle for 30 min.
Vary alum from 20 to 60 mg/L in steps. In a second series, adjust pH to 6.5, 6.8, 7.0 and 7.3 with lime or acid. Add a polymer at 0.05–0.2 mg/L in a third series if the floc is weak. Measure settled turbidity, filter a subsample through a 0.45 µm membrane or a small sand column, and record headloss and residual aluminium.
Select the lowest dose that gives settled turbidity below 5 NTU and filtered turbidity below 0.5 NTU, with residual aluminium below 0.2 mg/L. Verify that the chosen pH is within the plant's alkalinity budget; 90 mg/L as CaCO3 is adequate for alum at 25–40 mg/L but may need lime at higher doses.
3. What operational measures would you take if raw turbidity reaches 1,500 NTU?
Read the worked answer
Protect the filters by reducing plant flow to 60–70% and using pre-sedimentation or a roughing filter if available. Increase the coagulant dose according to the jar test, add a coagulant aid, and adjust pH to the optimum. Run filters to waste until turbidity falls below 1 NTU, then return them to service; do not send the first filtrate to the clear well.
Increase backwash frequency but avoid over-washing that causes media loss. For disinfection, maintain Ct = 15–20 mg/L·min at pH 7.8 and 26°C. If chlorine demand is high, use chlorine dioxide or ozone pre-oxidation if available, and monitor trihalomethane precursors. Notify the regulator if turbidity cannot be maintained; a boil-water advisory may be needed. Record raw and filtered turbidity continuously and revise the coagulation control strategy after the event.