The same COD can describe very different wastewater
A cream loss and a whey discharge contribute different proportions of fats, particles, colloids and dissolved compounds. EPA and IDB Invest guidance describe this variability and the importance of preventing product losses at source.
Total COD alone therefore cannot establish a DAF effluent target. Measure total and filtered COD using a stated procedure, alongside suspended solids and fats. Filtered COD is an operational approximation of the soluble fraction: the filter matters, and the result is neither a lactose measurement nor proof that all remaining COD is biodegradable.
Define the flotation duty
Microbubbles help aggregates of fats and solids rise for collection. Some proteins and emulsified fats may require coagulation and flocculation before they become readily separable. Test conditioning on actual wastewater: a switch from milk to cheese production or a change in detergent can change the response.
The aim is often to relieve biological treatment while producing manageable sludge. Clear-looking water can still carry substantial dissolved COD. Judging performance by appearance alone hides that load and can leave downstream treatment undersized.
Include cleaning and product-loss events
Sampling should cover production, start-ups, product changeovers and clean-in-place cycles. Record flow, pH, temperature and conductivity with each sample so a difficult event can be traced to its origin. First rinses may justify separate management after their volume and composition have been established.
Equalisation can buffer variations, but its volume cannot be selected from daily average flow alone. Map the discharge schedule against the planned DAF feed rate. Capturing recoverable product before it reaches the drain also reduces material that would otherwise need separation or biological degradation.
Use a load balance to set expectations
Illustrative example: 100 m³/day at 2,000 mg/L total COD carries 200 kg COD/day. If filtered COD is 1,500 mg/L, that fraction contributes 150 kg/day. Even hypothetical complete separation of the other 50 kg/day would leave those 150 kg/day. This is an accounting example, not a predicted DAF removal rate. This assumes the filtered fraction remains unchanged; it is not a minimum effluent COD, because some fine filter-passing colloids may respond to coagulation and flotation.
A trial should assess both clarified water and the collected sludge, rather than only the liquid above a jar-test floc. Measure residual COD, solids and fat, and examine float stability and dewatering behaviour. These observations allow the actual load benefit to be weighed against chemical use and sludge production.
Check the complete treatment train
Before selecting equipment, establish the load the biological stage can accept and the stream it is meant to protect. Where soluble lactose dominates, enlarging flotation does not replace a biological capacity check. Where fat and particulate material dominate, a properly tested upstream separation step can be valuable.
Include sludge storage, liquid returns and seasonal product changes. Representative trials provide a better basis than transferring results from a dairy with different products and cleaning practices.
Common questions
Does DAF remove lactose?
Truly dissolved lactose is not directly removed by flotation. It needs a process suited to the dissolved fraction, often biological treatment when characterisation confirms treatability.
Does lower COD mean the discharge problem is solved?
Not by itself. Check the complete effluent quality against its intended destination and verify the capacity and stability of subsequent treatment stages.
