Troubleshooting

Troubleshooting is the method by which an observed symptom is traced to its cause in a system where the processes are coupled. In a recirculating system almost every parameter influences several others, so a symptom rarely indicates its own cause and the ordering of the investigation matters more than the individual checks.

• The processes are coupled, so treating a symptom as its own cause frequently makes the underlying problem worse.
• The system has a live biomass, so the cost of a slow or wrong diagnosis is measured in fish rather than in downtime.
• Several distinct causes produce very similar symptoms, and distinguishing them requires method rather than experience alone.
• Changes made in combination cannot be evaluated, so undisciplined intervention destroys the information needed to find the cause.
• The same problem recurs indefinitely if the correction addressed the symptom rather than the cause.

• Provide sampling and measurement points where diagnosis will need them, particularly across each treatment process so that its performance can be assessed independently.
• Provide flow and pressure measurement around the loop, since head loss across a process localises a fault immediately.
• Sub-meter major electrical loads, because power against flow is a diagnostic as well as an energy record.
• Enable isolation of individual tanks and processes, so a hypothesis can be tested without disturbing the whole plant.
• Log continuously and retain the history, since diagnosis depends on comparison with a period when the system was well.
• Provide the standing capability to stabilise: supplementary oxygen, the means to increase exchange, and the means to reduce loading.

• Keep records good enough that the question 'when did this start' has an answer.
• Cross-check any reading before acting on it.
• Change one variable at a time, and write down what was changed and when.
• Record the whole investigation, including the hypotheses that were wrong, because the next occurrence begins there.
• Escalate rather than persisting alone when the biomass is at risk and the cause is not converging.
• Distinguish the correction that stabilised the system from the correction that fixed the cause, and complete the second one.

• Calibrate instruments on schedule, since the entire method rests on trustworthy measurement.
• Keep reagents within date and stored correctly, and hold a second measurement method for the critical parameters.
• Maintain the trend record and its backup, because a gap in the history is a gap in every future diagnosis.
• Keep the incident record indexed and readable, so past findings are retrievable.
• Verify that isolation valves and sampling points still work before an incident requires them.

• Diagnosing before stabilising, with the biomass deteriorating throughout.
• Acting on a single unverified reading.
• Changing several things at once.
• Correcting a symptom without identifying what produced it.
• Assuming the cause is disease before excluding water quality and handling.
• Failing to record the investigation, so the same problem is diagnosed from scratch next time.
• Continuing alone when the situation warrants escalation.
• Stopping once the symptom clears, leaving the cause in place.

Coupled systems produce delayed and displaced symptoms. An event in the biofilter appears as ammonia days later; a change in feeding appears as oxygen demand within hours and as solids loading over a longer period. Because of that lag, the moment a symptom appears is rarely the moment the cause acted, and the investigation should look back across a window rather than at the present state alone.

Structured post-incident review is where the durable improvement comes from. Reconstructing the sequence, identifying what was detectable earlier and was not detected, and asking why the existing monitoring did not catch it, turns a single event into a change in the monitoring or the routine. Without that step each incident teaches only the individual who happened to be present.

Diagnosis on a live system follows a fixed order: stabilise the biomass, verify the measurement, establish when the symptom began, list what changed, then trace outward through the coupling. Change one thing at a time and record everything, including the wrong hypotheses. Finish by separating the action that stabilised the system from the correction that removed the cause, and complete the second.