Quality and hydraulics answer different questions
If TMP rises at the same flux, investigate hydraulic resistance and operating conditions. If particles increase in the filtrate, investigate their pathway. A module can be fouled while retaining particles, and a small leak can exist without an obvious decrease in hydraulic performance. TOC or conductivity can also change with the source water without UF having lost integrity.
Retain both series when diagnosing a problem: temperature-corrected permeability and filtrate quality. Add valve status, flow, temperature and timestamps for backwash or cleaning. A visual correlation between two curves helps formulate a hypothesis, but does not establish a break or identify the component at fault.
| Observation | Hypothesis to investigate | Useful check |
|---|---|---|
| Low permeability; stable particulate quality | Resistance, temperature or measurement | Normalisation, sensors and response to authorised washing |
| Particles in filtrate; similar hydraulics | Leak, unfiltered path or downstream contamination | Train-level samples and direct integrity test |
| Signal only during a transient | Air, sequence or sampling | Valve status and analytical comparison |
| Integrity test fails | Membrane or test-circuit components | Locate leaks and repeat under the specified procedure |
Confirm where quality deteriorates
Before working on modules, compare the online reading with a representative sample and check the measuring equipment: cleanliness, calibration, air and flow through the cell. Establish whether the sample comes from an individual train's filtrate, a common header or a tank. Contamination downstream of UF needs a different investigation from passage through the membrane.
Locate the problem using available sampling points and comparisons between trains operating under similar conditions. Review recent events: maintenance, module replacement, chemical dosing changes, shutdowns or pressure incidents. While determining the cause, manage out-of-specification filtrate through the diversion or storage provided by the project, preventing automatic delivery to the protected process.
What integrity testing establishes and what can interfere
EPA distinguishes indirect quality monitoring from direct tests; turbidity sensitivity can be insufficient for certain integrity breaches. DuPont describes possible leaks through seals, valves or fibres and a pressure-decay test for its modules. Apparently acceptable turbidity does not replace a direct test when diagnosis or validation requires integrity verification.
Apply the procedure for the installed module, including its pressure, duration and acceptance criterion. A failed result can include leaks in connections or the test circuit; it does not automatically imply broken fibres. EPA also warns of false positives when membranes are not fully wetted. Inspection and component isolation need to preserve the conditions required for a valid test.
Close the diagnosis and demonstrate return to service
When a leak is confirmed, document its location and the authorised repair or replacement. If integrity passes but poor quality persists, investigate instrumentation, sequence, common pipework and storage. If particulate quality remains stable while permeability decreases, focus the investigation on resistance and operating recovery, without inferring barrier loss from that reading alone.
After intervention, verify the applicable integrity test and product quality, and compare hydraulics under equivalent conditions. Retain before-and-after results, test conditions and the action performed. For purchasing and integration, request train-level sampling access, isolation capability, alarm records and a return-to-service protocol. These provisions allow an incident to be investigated and resolved with evidence, alongside any cleaning that is required.