Real situations · field reasoning

Decisions.

A monitoring system rarely tells you what is wrong. It tells you what it observed.

These are real situations where the first explanation turned out not to be the useful one.

FIELD DECISION 01 · FUJAIRAH

My pH Sensor Woke Up on Monday and Chose Violence

A station reported pH of approximately 9.8 while the other parameters remained stable. That reading was an observation—not a diagnosis.

01Replacement sonde checked at the station.

02The anomaly returned after the technician began the approximately 200 km journey back.

03The removed sonde behaved normally on the bench.

Before the answer: what would you check next?

The investigation widened to the full measurement path. The cable passed continuity but failed leakage testing. The cable and connectors were serviced or replaced as relevant, and the issue resolved.

An anomalous reading is an observation, not a diagnosis.Troubleshoot the path of the measurement, not only the instrument producing it.

FIELD DECISION 02 · COMMUNICATIONS

The Station Was Offline. Except It Wasn't.

The station disappeared from one office/server view, then returned with historical data. The client monitoring room could still see it.

01Measurement and local storage had not necessarily stopped.

02One receiving endpoint had lost visibility while another still received data.

03The useful question changed from “Is it offline?” to “Offline where?”

Reveal the communication fault

The installed modem was 3G and mismatched the required specification. Replacing it restored reporting. Geographic distance was not treated as the cause.

Offline where?Separate measurement, local storage, communications and the receiving endpoint or display.

FIELD DECISION 03 · WINTER OPERATIONS

What Changed When Winter Arrived?

Three stations at a major Middle East aluminium facility developed recurring winter battery problems. Batteries were recharged or replaced, and two additional solar panels did not resolve the recurrence.

01The required output was one record every 15 minutes.

02Healthy-station comparison exposed different operating-program behaviour.

03The affected systems remained active much more than the requirement demanded.

Reveal what changed

Duty-cycle optimisation—sleep → wake → stabilise → measure → sleep—resolved the recurrence. Excessive internal recording also contributed to logger-memory behaviour.

Reporting interval ≠ operating interval.Before adding hardware, check whether software is creating the load. Winter weather alone was not declared the cause.

FIELD DECISION 04 · BU TINAH ISLAND

Laugh Late, Laugh Great

Planning used a forecast of around 0.1 m waves. On arrival, the antenna was broken; fisherman or fishing-vessel interaction was suspected but never confirmed. Conditions deteriorated after the team reached the station.

01The field team reduced speed, adjusted and waited as conditions required.

02The logger canister, battery canister and broken antenna were removed.

03The sonde remained deployed and self-logging every 15 minutes.

04The manager was updated once mobile coverage returned at the marina.

Field safetyA favourable forecast is never permission to continue unsafe offshore operations. If sea state, wind, visibility, vessel or crew condition becomes unsafe, stop, withdraw, wait or change the plan. Field safety comes before station recovery.

Use forecasts for planning. Use observations for decisions.Loss of telemetry ≠ loss of measurement.

The method that emerged

Let’s try to fail successfully.

If the system were healthy and I wanted to create exactly this symptom, what would I have to break?

This is not Decision #5. It is the habit that grew out of the four real situations above.