Why Experienced Technicians Detect Failures Before Instruments Do
Photo: Unsplash — Industrial maintenance technician at work
There's a particular kind of walk that a seasoned crane maintenance technician does during a shift start-up check. It's unhurried. It pauses in places that aren't on any checklist. The hand rests on a motor housing for a moment longer than necessary. The ear tilts toward the hoist drum. The nose registers something faint before the brain has even formed a question. By the time that technician pulls out a clipboard, they've already run a diagnostic that no installed sensor was designed to perform.
This piece is about that walk. It's about what accumulates over ten, fifteen, twenty years of working with the same equipment in the same environment — and why that accumulation represents a form of knowledge that is genuinely difficult to replicate through instrumentation alone. It's not mysticism. It's pattern recognition, embodied experience, and the deep familiarity that comes from being present when things went wrong.
"I knew that crane was going to give us trouble before the week was out. The long travel drive had a different sound when it decelerated — not louder, just different. Twelve days later, the brake liner was metal-on-metal. The vibration analyser caught it on day nine. I'd flagged it on day one."
— Overhead Crane Maintenance Supervisor, Steel Plant, Central India (experience shared with author)The Instrument's View of the World Is Narrow
Modern condition monitoring equipment is, by any objective measure, an extraordinary achievement. A triaxial accelerometer mounted on an EOT crane gearbox pedestal captures vibration data in three planes simultaneously, at sampling rates that far exceed human perception. A thermal imaging camera reveals temperature gradients invisible to the naked eye. An online partial discharge monitor in a HT panel detects dielectric stress that would be entirely imperceptible through any physical sense a person possesses.
And yet each of these instruments has a fixed field of view. The accelerometer on the gearbox tells you nothing about the condition of the runway rail joints three bays over, where the crane wheel flange is developing an unusual wear pattern. The thermal camera shows you surface temperature — not the subtle change in the crane cabin's electrical panel that a technician would notice from the slightly acrid trace of overheating insulation on a hot afternoon. The partial discharge monitor watches one specific set of HT cables; it says nothing about the earth fault that's developing in the pendant control cable because it's been dragging on an abrasive surface for months.
Instruments, in other words, measure what they are pointed at. An experienced technician walks through the whole picture.
The Coverage Problem
Consider what a fully instrumented overhead crane in a steel plant actually has sensors covering. A well-specified system might include vibration monitoring on hoist gearbox and drive end bearings, end-carriage gearbox bearings, and long-travel motors. It might include rope tension load cells, motor current monitoring for overload detection, and thermal monitoring at panel MCC level. This is genuinely comprehensive by the standards of most facilities in operation today.
But step back and count the mechanical interfaces on that same crane that are not instrumented. The rail clips and fishplates. The conductor bar section joints. The bridge structure itself — the box girder, the end trucks, the cross-girder connections. The pendant station mounting bracket. The hook latch mechanism. The crane cabin floor lining, which over years of thermal cycling and vibration can develop micro-cracks that compromise the platform's mechanical integrity. Every bolted joint on the crane. An experienced technician's eyes and hands cover all of these in a thorough inspection. A sensor network, even a generous one, does not.
How the Body Reads a Machine
The experienced maintenance professional's body is, in a very real sense, a multi-modal sensing platform. It's imprecise by the standards of digital instrumentation — it cannot give you a bearing defect frequency in Hertz or a thermal gradient in degrees Celsius per centimetre. But it integrates information across modalities in ways that even the most sophisticated data fusion algorithms struggle to replicate. Here's how each sense contributes to early fault detection in a steel plant crane and electrical maintenance context.
Vision
Grease colour, contamination, weeping seals. Rust patterns that indicate abnormal moisture ingress. Wear dust accumulation in unexpected locations. Slight misalignment in couplings visible as uneven wear marks.
→ Greyish metallic grease on a gearbox vent: bearing material in the lubricant
Hearing
Subtle changes in running tone. Intermittent clicks or pops absent at previous inspection. Changes in brake engagement sound — sharper, softer, hesitant. Wind noise through a crack.
→ Soft rhythmic knock at full load on hoist lift: wire rope strand fatigue
Touch
Temperature differentials across a motor housing. Surface vibration felt through the palm that differs in character from known baseline. Roughness in a cable outer sheath where insulation is hardening.
→ Asymmetric heat across opposite ends of a long-travel motor: partial winding issue
Smell
Overheated insulation has a distinct acrid note. Ozone-like smell near switchgear indicates arc tracking. Burnt rubber smell near cable trays. Contaminated gear oil has an identifiable sour quality.
→ Faint phenolic smell in HT panel: tracking current on busbar insulation
Intuition (Pattern Integration)
The integration of multiple subtle signals that individually mean little but together form a recognisable pattern. The mind comparing the present state against hundreds of remembered reference states — including states that preceded failures. This is the hardest to articulate and the most difficult to transfer.
→ "Something isn't right" — often the precursor to finding something genuinely wrong that no single sense fully identified
That last entry — intuition as pattern integration — deserves more than a grid cell. Research in naturalistic decision-making, particularly the work of Gary Klein on Recognition-Primed Decision (RPD) theory, provides a compelling cognitive science framework for what experienced maintenance professionals do when they sense impending failure. Klein's research with military officers and firefighters — and subsequently with maintenance engineers — found that experts don't reason from first principles the way novices do. They pattern-match. They rapidly compare the current situation against a mental library of prior situations and, when the match is close enough, they act on the resulting recognition. The recognition feels like intuition. It is actually the product of structured experience.
Experience Is Not Just Time Served
There's a temptation to think of expertise as simply the accumulation of years. Spend enough time around cranes and you'll eventually know what you're doing. This is partially true and mostly misleading. Time spent in an environment generates experience — but not all time generates equally useful experience. The quality of what's accumulated depends heavily on whether the person was paying active attention, whether they received feedback on their interpretations, and whether they had the opportunity to see — and understand — the consequences of early signs they either caught or missed.
Years 1–3: Building Vocabulary
The early maintenance technician learns to name things — components, failure modes, standard procedures. They follow checklists. They observe experienced colleagues. They learn what normal looks, sounds, and feels like in routine conditions. This is the foundation, and it matters.
Years 4–8: Encountering Failure
This is where the mental library starts to fill with the entries that actually matter. The first time you see what a wire rope looks like just before it fails. The first gearbox you open that shows catastrophic gear tooth pitting. The first earth fault that caused a production stoppage and whose early signs you missed. Each incident — especially the ones you didn't prevent — writes a lasting entry.
Years 9–15: Pattern Completion
The library is large enough that incomplete patterns trigger recognition. You see one sign, and your mind retrieves the associated cluster of signs you've learned to look for. You no longer need all the evidence — you know where to look for the rest of it. This is when technicians start making early calls that look like intuition to their colleagues.
Years 15+: Integration and Judgement
The most experienced technicians don't just detect — they triage. They can tell the difference between the bearing that needs monitoring and the bearing that needs replacing today. They know which warning signs are urgent and which can wait for the next planned window. This calibration is extremely difficult to formalise in a procedure or a checklist.
The Japanese concept of tacit knowledge — knowledge that cannot be fully articulated, that is embodied rather than codified — is the most useful intellectual framework here. The philosopher Michael Polanyi, who coined the phrase "we can know more than we can tell," was describing exactly the kind of expertise we're discussing. The experienced overhead crane technician knows things about their cranes that they cannot fully verbalise — and this knowledge is a legitimate, valuable, safety-critical asset.
The Steel Plant as Teacher
A steel plant is one of the more demanding environments for developing maintenance expertise, precisely because it offers so many failure mechanisms operating simultaneously and at scale. Heat, dust, electromagnetic interference, vibration, corrosive atmospheres, extreme load cycling, and the economics of continuous production — all of these shape how equipment degrades and how people learn to read that degradation.
The overhead crane fleet in a typical integrated steel plant spans everything from light-duty auxiliary cranes in workshop areas to the ladle cranes and charging cranes in the steelmaking bays — machines operating in ambient temperatures that can exceed 60°C near furnace mouths, handling loads that may approach or reach rated capacity multiple times per shift. The failure modes relevant to these machines are not identical to those of a crane operating in a neutral-temperature warehouse environment. The experienced steel plant crane technician's mental map is calibrated to these specific conditions.
Specific Patterns the Plant Teaches
In high-heat steelmaking bays, grease selection and condition reading is a discipline of its own. The wrong grease — or good grease that has undergone thermal breakdown — changes consistency, colour, and smell in ways that an experienced lubrication hand recognises immediately. The standard vibration monitoring system reports no anomaly. The experienced technician opens the grease point and knows the bearing is running without adequate film lubrication.
Runway rail wear patterns tell stories over months. Rail head surface cracking, side wear profiles on the gauge face, and the wear dust distribution around crane wheel flanges all carry information about wheel-rail alignment, crane skewing behaviour, and the condition of the crane bridge structure itself. A technician who has walked the same runway every week for a decade reads this information without consciously thinking about it.
Electrical systems in steel plants accumulate pollution loading over time — conductive dust, metallic particulates, moisture cycling — that creates leakage current pathways in control panels and motor terminal boxes that would be invisible in a clean-room environment. The experienced electrical maintenance person in this context develops a heightened sensitivity to signs of insulation stress: discolouration, tracking marks, the specific smell profile of different insulating materials under different thermal and electrical stresses.
π What Instruments Catch First
- High-frequency bearing defect frequencies (>1kHz)
- Precise temperature values above threshold
- Electrical overcurrent / overload events
- Load cell deviations from rated capacity
- Limit switch proximity outputs
- Motor winding resistance (formal testing)
π️ What Technicians Catch First
- Changes in running quality before measurable vibration shift
- Grease / lubricant condition and contamination
- Structural micro-changes: loose fasteners, hairline cracks
- Conductor wear and insulation surface condition
- Abnormal smell signatures in panels and cable trays
- Operational behavioural changes noticed by crane operators
When the Knowledge Walks Out the Door
Across heavy manufacturing sectors globally, there is a demographic transition underway that has significant safety implications. A large proportion of the most experienced maintenance workforce in steel plants, power stations, and heavy engineering facilities is approaching retirement age. In many organisations, a meaningful share of the tacit maintenance knowledge accumulated over decades exists only in the heads of people who will leave the workforce within the next five to ten years.
This is not an abstract concern. When an experienced overhead crane technician retires, what leaves with them is not just their formal qualifications and their knowledge of written procedures. What leaves is their mental model of every crane they have worked with — the specific way each machine sounds at different load levels, the recurring quirks that never appear on a defect notification but are part of the working knowledge of the maintenance team, the historical context of every repair and modification that has shaped the current condition of the equipment.
Replacing this with a more junior technician and a better sensor package is not an equivalent substitution. The sensor package will catch what it was designed to catch. The junior technician will follow the procedures they were trained to follow. But the gap — the early warning space between "normal operation" and "an instrument threshold crossed" — will be thinner, and some of the things that the experienced person would have noticed and investigated will pass unremarked.
"Every experienced maintenance person on this crane fleet has a list of things in their head that aren't written down anywhere. Small things. Soft things. Things they notice and keep an eye on. When that person goes, those things disappear. And some of them, eventually, become incidents."
— Plant Maintenance Manager, Integrated Steel Facility (shared with author)Preserving the Irreplaceable
Not all tacit knowledge can be made explicit. This is the nature of the thing — if it could be fully articulated, it would have been written into a procedure already. But more of it can be captured and transferred than most organisations currently attempt. The methods that work require deliberate investment and a cultural commitment to treating experiential knowledge as a valued organisational asset.
Practical Approaches to Knowledge Retention in Maintenance Teams
- Structured shadowing programs — Pair junior technicians with experienced colleagues on specific tasks with explicit discussion of what to observe and why. Don't just send them along; debrief afterward.
- Failure-mode storytelling — Create regular forums where experienced technicians narrate past fault cases: what they noticed, what the investigation found, what they would look for in future. Video these sessions where possible.
- Annotated inspection routes — Supplement standard checklists with site-specific experiential annotations: "On this crane, pay particular attention to the long-travel brake lining — this machine has a history of rapid wear on the operator side."
- Defect photography libraries — Build a visual reference collection of defects found on the specific equipment in the specific plant. Generic textbook images are less useful than photographs of the actual wear patterns on the machines your people maintain.
- Pre-retirement knowledge extraction interviews — Structured conversations with departing senior technicians, conducted by a facilitator, documenting site-specific knowledge before it leaves the organisation.
- Operator feedback loops — Formalise the channel between crane operators and maintenance teams. Operators notice changes in machine behaviour — subtle changes in travel feel, control response, or unusual sounds — well before instrument thresholds are crossed. Their observations are often the earliest warning available.
The Case for Combined Intelligence — Not Replacement
The conversation about AI and advanced instrumentation in industrial maintenance sometimes has a sub-text of replacement — as though the goal is to eventually make the experienced maintenance professional unnecessary. This framing misses the point in ways that have practical safety consequences.
The most defensible position, supported by both operational experience and the research literature on human-machine teaming, is that instrumentation and human expertise are most powerful when they are genuinely integrated — when each informs the other rather than competing. Instruments should extend the reach of experienced people into domains where human senses cannot operate (real-time bearing load cycles, for example, or continuous partial discharge monitoring in live HT equipment). Experienced people should interpret instrument outputs in context, catching the false alarms that would otherwise create alert fatigue, and investigating the physical reality behind anomaly flags that an instrument cannot contextualise.
This is a partnership that requires deliberate design. It means ensuring that vibration analysis reports reach the people who actually know the cranes, not just the maintenance planner's inbox. It means building inspection routines that take advantage of both instrument outputs and physical observation, not treating them as alternatives. It means respecting the experienced technician's instinct enough to investigate when they flag something — even when no instrument has yet confirmed their concern.
In the author's own experience with overhead crane fleets in steel plants, some of the most valuable near-miss preventions have come from exactly this kind of collaboration: an instrument flagging a trend that was ambiguous enough to be deprioritised, combined with an experienced technician's ground-level observation that gave the trend enough context to warrant immediate investigation. Neither signal alone would have prompted action. Together, they did.
The Walk Still Matters
There is nothing sentimental about arguing for the continued centrality of experienced human expertise in maintenance safety. It's not a resistance to technology or a nostalgia for older ways of working. It's a clear-eyed assessment of what the technology can and cannot do, and a recognition that the gap — the space between what instruments measure and what experienced people sense — is where many of the most important early warnings live.
The walk that the experienced technician takes at shift start has a scientific basis in pattern recognition and tacit knowledge theory. It serves a safety function that no current sensor network fully replicates. And it is at risk — from retirement demographics, from organisational assumptions that instrumentation is a substitute for expertise, and from the simple tendency to stop valuing what has become routine.
Investing in the development and retention of experienced maintenance personnel isn't a cost centre. It's one of the most direct investments a steel plant can make in its own safety record — alongside, not instead of, the instrumentation systems that extend and quantify what those experienced people can do.
The bearing that the technician noticed three days before the vibration analyser did? That's not luck. That's knowledge. And that knowledge, in a plant where the consequences of getting it wrong can be catastrophic, is worth fighting to preserve.
Sources & References
- Klein, G.A. (1998). Sources of Power: How People Make Decisions. MIT Press, Cambridge MA. [Recognition-Primed Decision theory]
- Polanyi, M. (1966). The Tacit Dimension. Doubleday, New York. [Tacit knowledge — "we can know more than we can tell"]
- Nonaka, I. & Takeuchi, H. (1995). The Knowledge-Creating Company. Oxford University Press. [Tacit to explicit knowledge conversion]
- Bureau of Indian Standards. IS 807:2006 — Design, Erection and Testing (Structural) of Cranes and Hoists. BIS, New Delhi.
- Bureau of Indian Standards. IS 3177:1999 — Code of Practice for Electric Overhead Travelling Cranes and Gantry Cranes. BIS, New Delhi.
- ISO 10816 / ISO 20816. Mechanical Vibration — Evaluation of Machine Vibration by Measurements on Non-Rotating Parts. International Organisation for Standardization.
- Health and Safety Executive (UK). (2009). Tacit Knowledge and Experience in Maintenance of Safety-Critical Plant. HSE Research Report RR737. hse.gov.uk
- Ericsson, K.A., Krampe, R. & Tesch-RΓΆmer, C. (1993). "The Role of Deliberate Practice in the Acquisition of Expert Performance." Psychological Review, 100(3), 363–406.
- World Steel Association. (2023). Workforce Development in the Steel Sector: Skills for the Future. worldsteel.org
- Reason, J. & Hobbs, A. (2003). Managing Maintenance Error: A Practical Guide. Ashgate Publishing.
- Nyce, J.M. & Timpka, T. (1993). "Work, Knowledge and Action: The Role of Expertise in Maintenance." Human Factors and Ergonomics in Manufacturing, 3(4), 297–310.
- Deloitte Global / The Manufacturing Institute. (2021). Closing the Skills Gap in Advanced Manufacturing. deloitte.com
No comments:
Post a Comment