Why Field Lessons Travel Further Than Training Materials

The lessons that stick in engineering practice are rarely the ones from training materials. They are the ones from field experience — the situations where the training met reality and the gap between the two became visible in a timeline, a procurement call, or an inspection finding that changed the scope of the whole project.

This article documents the field lessons from one real transformer maintenance project — a 220 kV unit, nineteen years in service, scheduled for a standard OLTC service that the Day 1 inspection revealed was anything but standard. The lessons are organised by the phase of the project where they arose: readiness, inspection, response, and documentation.

Readiness Lessons

Lesson R1: The condition data you have is only as good as the method used to collect it.

The four quarters of DGA sampling history for this transformer showed within-limits gas concentrations. That history was one of the primary inputs into the decision to plan a standard service scope. On Day 1, the first oil sample from the correct sampling point — the drain at the bottom of the OLTC diverter compartment rather than the conservator connection used for quarterly sampling — showed dielectric breakdown voltage below the specification floor and carbon contamination significantly above the quarterly samples had indicated.

The quarterly samples had been taken from the wrong point. The maintenance programme had been calibrated to data that underrepresented the contamination at the component that actually mattered.

Field lesson: Before every project, confirm that the condition data you are relying on was collected correctly. Oil sampling point. Temperature at time of sampling. Sample handling and submission timeline. These factors affect what the data shows. If the method is wrong, the data is misleading — and a maintenance plan built on misleading data is a plan that will be wrong in ways that become visible on Day 1.

Lesson R2: The spares for the expected scope are not the spares for the actual scope.

The spare kit for this project was built for the expected scope — standard service, anticipated contact wear within the refurbishment threshold, oil top-up rather than full change, motor drive adjustment rather than linkage replacement.

The actual scope required contact set replacement, full oil drain and refill, and motor drive linkage replacement. The Tier 1 spares covered the contact set replacement. They did not cover the full oil quantity or the motor drive linkage components.

The motor drive linkage identification and expedited supply was handled efficiently through EMR Global's spares support — the correct components for the specific motor drive configuration were confirmed within an hour and arranged for expedited delivery within the extended outage window. But the delivery wait extended the project beyond what pre-positioned contingency spares would have allowed.

Field lesson: The spare kit that matches the expected scope is the Tier 1 order. The Tier 2 order — contingency spares for the worst reasonable case findings — should be placed at the same time, six weeks before the outage. The Tier 2 components that are not needed get returned. The ones that are needed are on site rather than in transit.

Lesson R3: Vendor availability is not the same as vendor readiness.

The specialist vendor for the OSR relay calibration was available when called. That availability was fortunate. It was not planned. The vendor had not been pre-scheduled for the outage window, and the scheduling confirmation and travel arrangement that happened on Day 2 consumed time that a pre-scheduled appointment would not have required.

Field lesson: Availability means the vendor can come. Readiness means they are already scheduled to come. The difference is the coordination overhead of an unplanned scheduling call during an active outage — which is always more complicated and more time-consuming than the same coordination done six weeks before the window opens.

Inspection Lessons

Lesson I1: Measurement tells you things that visual assessment doesn't.

The diverter switch contacts in this transformer were assessed visually at the previous inspection and flagged as minor discolouration without immediate action. Calibrated contact thickness measurement at this inspection found three of four positions beyond the wear replacement threshold.

Visual assessment and calibrated measurement are not the same diagnostic. Visual assessment tells you whether something looks acceptable. Calibrated measurement tells you whether it meets the specification. For components where the specification limit is a dimensional threshold, visual assessment is not a substitute for measurement.

Field lesson: Every component with a defined specification limit should be measured, not assessed visually. The measurement takes minutes longer. The information it provides is categorically different. A maintenance record built on measurements rather than visual assessments supports better decisions at the next inspection — and the inspection after that.

Lesson I2: The motor drive is the upstream cause of many OLTC condition findings.

The elevated contact wear and oil contamination in this transformer were both explained by the motor drive timing drift caused by linkage play. The contacts and the oil were the visible findings. The motor drive was the root cause.

A maintenance programme that assesses the contacts and the oil but not the motor drive timing is a programme that addresses downstream consequences without identifying upstream causes. The same condition findings will present at the next maintenance interval — because the root cause was not addressed.

Field lesson: Motor drive timing verification should be a standard step in every OLTC service scope. Not because timing drift is always the cause of contact and oil findings, but because when it is, identifying it changes the maintenance response from symptom treatment to root cause correction.

Lesson I3: Protection device calibration is not optional.

The OSR relay was found miscalibrated at a threshold above the correct value for the specific OLTC design. Recalibration took less than an hour. The consequence of not recalibrating — a protection device that does not function as designed — is a system operating with a structural protection gap of indeterminate duration.

Field lesson: OSR relay calibration verification is mandatory on every OLTC service. Not because it is likely to be wrong. Because when it is wrong, the consequence of returning the transformer to service without correcting it is a protection scheme that fails at the moment it is most needed.

Response Lessons

Lesson Re1: The operations department needs the contingency timeline before the outage, not during it.

When the expanded scope was confirmed on Day 1 afternoon and the revised five-day timeline was communicated to operations, the response was genuine surprise and a scramble to adjust the load management plan, the control centre communication, and the assessment of adjacent scheduled work impacts.

All of that adjustment was correct and was handled professionally. It was also avoidable. If the pre-outage operations briefing had included the contingency scenario — explicitly stating that if the Day 1 inspection confirmed the findings the condition data suggested, the outage would extend to five days — the operations response would have been pre-planned rather than improvised.

Field lesson: Brief operations on both the base and contingency timelines before the outage opens. The briefing takes an extra thirty minutes during the planning phase. The improvised load management adjustment that it prevents takes half a day during an active outage.

Documentation Lessons

Lesson D1: The maintenance record is a reference document, not a compliance record.

The maintenance record from the previous service described what had been done. It did not record what had been found. The as-found contact thickness measurements that justified the previous refurbishment decision were not in the record. The oil test results from the previous service were not in the record.

This meant the team could not assess the rate of condition progression between the previous service and this one. The baseline comparison that should have been available from the previous record was not there.

Field lesson: Every measurement taken during a maintenance project — as-found and as-left — should be in the maintenance record before the team leaves site. Not summarised. Not approximated. The actual number, against the specification limit, with the instrument used and the date noted. The team who opens this transformer in two years will thank the team that recorded this data correctly today.