Ask why an operating limit is set where it is, why a bypass line exists, why one exchanger fouls faster than its twin, or what actually happened during the 2009 trip — and on most units the answer lives in someone’s memory, not in the document management system. That was survivable when careers were long and turnover was slow. It is not survivable through the retirement wave, contractor-heavy staffing, and ownership changes that operating sites are now experiencing simultaneously.
How the loss actually presents
Knowledge loss rarely announces itself. It presents as symptoms: repeat failures of a problem the unit solved twenty years ago; operating envelopes narrowed “to be safe” because nobody can reconstruct the basis for the original limit; turnaround scopes inflated because the history that would justify exclusions cannot be produced; troubleshooting that starts from zero on a fault the previous crew had a worn path through. Each symptom has a cost; none of them gets booked against the real cause.
The incident connection is direct
Investigation reports — including the CSB’s recent finalized set — keep finding units that had lost the knowledge of their own hazards: damage mechanisms identified at design and forgotten in operation, safeguards whose purpose nobody could state, near-misses that predicted the event but were never indexed anywhere a planner would look. A unit that cannot remember its hazards re-learns them the expensive way.
Retention is a program, not an exit interview
The workable response is targeted, not encyclopedic. Capture the basis behind limits and alarms, the history behind recurring problems, and the reasoning behind non-obvious design features — for the highest-consequence systems first. Tie capture to events that already happen: turnarounds, incident reviews, engineering studies, and the last two years before a known retirement. Store it where the next engineer will actually look: the equipment data sheet, the operating procedure basis, the unit hazard summary.
- Interview toward documents: structured sessions with senior operators and engineers, converted into records with owners
- Attach “why” fields to limits, alarms, and safeguards — a limit without a basis is a limit that will be casually changed
- Index near-misses and unit history so they surface in planning, not in hindsight
- Make knowledge capture a defined deliverable of every study, turnaround, and investigation the site already pays for
Institutional knowledge is an asset with no line item — until it leaves, when it becomes a liability with several.