The Work That Keeps The Lights On
There is a layer of operational work that everyone in the energy sector knows exists but nobody on social media ever writes about. It involves the unglamorous, repetitive, physically demanding tasks that prevent catastrophic failures. I spent about eight years on site at a combined cycle plant in Ohio watching the same problems recur every third winter. The paperwork called it preventive maintenance. We called it shoveling coal. The term describes any work that doesn't produce revenue directly but absolutely determines whether the system survives a stress event. It includes inspections, calibration, filter changes, valve exercises, and the mundane data entry that creates an audit trail three years later when someone asks why a trip occurred. Most organizations try to minimize this labor because it shows up as cost on a quarterly report. That is a mistake. I have seen plants cut inspection cycles by half to free up manhours. Within fourteen months, two unexpected shutdowns cost more than three years of conservative maintenance budgets combined. The math is not complicated. It just requires someone to actually do the math instead of guessing.
Why This Category Of Work Gets Ignored
The primary reason is visibility bias. When shoveling coal happens correctly, nothing happens. The alarm stays silent. The grid stays stable. The report comes back blank. Leadership interprets blank as unnecessary. I watched a maintenance supervisor get pushed to reallocate his crew to a commissioning project because the old equipment was running fine. It ran fine for eleven more days. Then a pressure transducer failed during a ramp event, took out a control loop, and tripped the unit. The fix required a full teardown that would have taken two weeks instead of the forty-five minutes a scheduled inspection would have caught. The second reason is measurement difficulty. You cannot easily attach an ROI to a catastrophe that did not occur. Standard financial models reward visible output. They penalize hidden effort. This structural incentive means the people doing the actual invisible work rarely get credit for it, and they get blamed when something goes wrong despite their best efforts.
How To Identify Your Own Shoveling Coal Tasks
Start by mapping every recurring task on your team's calendar that does not generate a deliverable anyone outside operations cares about. Filter for tasks that prevent failures rather than create features. These are your candidates. Rank them by what happens if you skip them for ninety days. The ones that cause cascading problems belong in the protected category. The ones that are just habit belong in the elimination category. Here is where most teams make a mistake. They protect the wrong things. They keep tedious monthly reports that nobody reads while letting quarterly safety inspections slide because the last inspector said everything looked fine. The workaround I used at that Ohio plant was brutal but simple. I started requiring every inspector to document exactly one anomaly per visit, even if it was trivial. This forced attention to detail and created a paper trail that made neglect impossible to hide. It also meant I could point to specific examples when budget people asked why we needed the time.
Get the Full Details
The Counter-Intuitive Part
More shoveling coal actually reduces total labor. This sounds backwards until you trace the failure chain. A missed gasket inspection leads to a seal failure. The seal failure causes a pressure drop. The pressure drop triggers an unplanned shutdown. The shutdown requires diagnostic time, parts expediting, overtime, and possible replacement of secondary components that were only stressed, not broken. The total hours expended across all those downstream events usually exceed twenty times the hours that a proper inspection would have consumed. I have run these calculations repeatedly. The ratio rarely favors cutting the visible labor. Another nuance that beginners miss is that the quality of invisible labor matters more than the quantity. A sloppy inspection that checks boxes without actually checking anything is worse than no inspection at all, because it creates false confidence. At one site I worked, the vibration analysis contractor was submitting reports that showed all green statuses. The equipment kept bearing failures every six months. I pulled the raw data myself and found they were sampling at the wrong frequencies and missing the early-stage resonance patterns that precede failure. We switched to an in-house team after three months. The learning curve was steep. The payback came within a single outage cycle.
Edge Case: When The Protocol Itself Is The Problem
I encountered a situation where following the published maintenance schedule actually increased risk. The procedure called for exercising isolation valves every thirty days. The valves were older globe types with packed stems. Thirty-day exercise was creating wear that led to stem leaks. The leaks were minor until a cold snap made the packing brittle, at which point we had a visible emission during a high-load period. The fix was switching to a ninety-day cycle with a torque check and lubrication step in between. The procedure manual did not allow for that variation without engineering sign-off, which took weeks. So I documented the failure modes, calculated the risk reduction, and got informal concurrence from the reliability engineer. We changed the interval, logged the deviation, and monitored for six months. No leaks. Same compliance posture. The lesson is that blind adherence to schedule can be as dangerous as skipping the schedule entirely. You do not need to fight cultural battles to preserve essential invisible labor. You need mechanisms that make the work self-evident. Here is what has worked in my experience: First, attach metrics that reflect avoidance. Track mean time between forced outages. Track inspection-to-failure ratios. Track the number of issues caught during scheduled windows versus issues discovered during unscheduled downtime. These numbers tell a story that headcount arguments cannot.
Second, rotate the work. When the same three people always do the unglamorous tasks, they burn out or resent it. Rotation creates shared ownership and surfaces problems faster because more people understand what normal looks like. Third, make the deliverables visible to people who control budgets. A checklist buried in a binder does not protect you. A one-page summary showing thirty anomalies prevented this quarter does. Format matters more than content in these situations. Leadership responds to brevity and color, not to detailed technical narratives.
When Shoveling Coal Is Not The Answer
This approach has limits. If the underlying equipment is fundamentally obsolete and parts are no longer available, no amount of inspection will keep it running reliably. At a facility I visited in West Virginia, the control system was running on hardware that had been end-of-life for twelve years. The vendors had stopped supporting it. The team was spending roughly sixteen hours per week on workarounds and manual compensations. The right call was replacement, not more diligent shoveling. I fought to make that case for nine months before it got approved. The downtime during replacement was four days. The annualized cost of the workaround labor and the risk exposure was far higher. Sometimes the invisible labor is a signal that the system needs to change, not that you need to work harder. Another scenario where this breaks down is when the labor itself is poorly trained. I once inherited a team that performed thorough inspections but recorded nothing useful. The tags were inconsistent, the photos were cropped badly, and the notes used vernacular that only two people understood. The work was technically competent but operationally worthless because the knowledge could not transfer. We spent six weeks standardizing the documentation format before the inspection program became effective. Training investment here was not optional. It was the difference between data and noise.
What To Do Next Week
Pick one recurring task your team considers invisible. Document exactly what happens if it is skipped for sixty days. Find one person in a position of influence who cares about continuity and show them the documentation. Ask for a ninety-day trial where the task is protected from reallocation. Report the results with numbers, not opinions. Repeat until the pattern becomes structural rather than personal. The work does not advertise itself. It rarely gets praised. But the systems that endure are the ones that stop treating essential labor as optional. The alternative is always the same. Something breaks at the worst possible time, and someone is suddenly very grateful for the people who used to do the thing nobody wanted to do.
