Thermal Imaging

Definition: Thermal imaging is a non-contact inspection technique that uses an infrared camera to convert the infrared radiation a surface emits into a visible heat picture called a thermogram. In maintenance, it makes temperature differences visible before they become failures: a hot termination in a panel, a bearing running warmer than its twin, a motor housing that cannot shed heat.

What Is Thermal Imaging?

Objects above absolute zero emit infrared radiation, and the intensity of that radiation rises with the temperature of the emitting surface. A thermal camera is a calibrated radiometer: it samples this radiation across a field of view and converts it into a map of surface temperatures. The result is a heat picture of equipment under real operating load, captured without contact and without stopping production. That combination, live-load data with zero contact, is why thermal imaging became a standard condition monitoring technique for electrical and mechanical equipment.

How Thermal Imaging Works

Industrial cameras operate in the long-wave infrared band, roughly 8 to 14 micrometers, where room-temperature objects radiate strongly and common lens materials transmit well. Inside the camera, a detector absorbs this radiation and responds in proportion to it.

Two detector families dominate. Microbolometers are uncooled thermal detectors whose electrical resistance changes with absorbed radiation; they are compact, affordable, and standard in industrial cameras. Cooled photon detectors require cryogenic cooling but deliver higher sensitivity and frame rates, which matters more for scientific and high-speed applications than for routine plant routes.

The camera renders each pixel as a temperature value, and radiometric files store the full temperature array so images can be reanalyzed after the walk. Operators use spot meters for single points, isotherms to isolate a temperature band, and span adjustments to reveal small differences. One practical limit deserves attention: the measurement spot grows with distance, so a component imaged from 20 meters averages a much larger area than one imaged from 2 meters. Keep the target large in the frame, or the reading blends the component with its surroundings.

Thermal Imaging for Predictive Maintenance

A fault that generates friction, electrical resistance, or load imbalance shows up as heat before it shows up in other measurements. That is the core value: one survey pass through an electrical room or a production line captures dozens of assets and flags the ones deviating from their neighbors. In a structured predictive maintenance program, thermal surveys sit alongside vibration and oil analysis, each catching faults the others miss. Electrical resistance heating, for example, is invisible to vibration routes but obvious in a panel image, while deep internal gear damage shows in vibration long before it warms a housing.

Survey discipline matters as much as camera capability. Routes should run at consistent load, consistent distance, and documented emissivity, and each asset should carry a baseline thermogram from healthy operation. Trending an asset against its own history catches slow degradation that a single snapshot hides. Surveys also anchor the wider asset condition monitoring picture: they establish which assets run hot and where fixed sensors earn their placement.

Electrical Inspection: Hot Spots, Connections, and Breakers

Electrical equipment is where thermal imaging earns its keep first, because resistance heating concentrates at small points. Common electrical findings include:

  • Loose or corroded connections: resistance at a termination generates intense, tightly localized heat at a lug or screw.
  • Phase imbalance: one phase conductor or its terminations run noticeably hotter than the other two under the same load.
  • Overloaded conductors and breakers: uniform heating along a cable or across a breaker body rather than at a single point.
  • Open or failing fuses: a fuse whose end caps differ in temperature from adjacent fuses in the same holder.
  • Transformer and MCC components: bushings, bus joints, contactor lugs, and splice points develop distinct signatures as contact degrades.

Two cautions apply. First, the hottest surface point marks the path heat takes to escape, not necessarily the fault origin; a hot panel cover may point to a breaker behind it. Second, load rules the reading: a connection problem at 20% load may barely register, so record load at capture and treat low-load surveys as incomplete.

Mechanical Inspection: Bearings, Motors, and Couplings

Mechanical faults generate heat through friction, and conduction carries that heat to the housing where the camera sees it. Typical mechanical signatures include:

  • Failing bearings: a localized warm ring at the bearing housing, often with one end of a motor running warmer than the other.
  • Coupling misalignment: heat concentrated at the coupling between two shafts that otherwise read similar.
  • Motor cooling problems: blocked fins or a failed fan show as hot spots or a warmer side on the housing rather than a uniform rise.
  • Low gearbox oil: a sharp horizontal temperature line at the oil level inside the sump, with the splash zone above running hot.
  • Hydraulic and process equipment: blocked coolers, failed steam traps, and internal leakage produce characteristic patterns of their own.

Conduction also smears mechanical heat: a small hot bearing spreads its temperature across a large casting, so the surface rise understates the fault. For that reason, thermal imaging on rotating equipment works best as a screening layer, and abnormal findings get confirmed with vibration analysis, which resolves component-level faults the housing surface has blurred.

Interpreting Temperature Anomalies and Severity

Interpretation starts with a comparison, not an absolute number. The most reliable reference is a similar component under the same load: one phase against the other two, one pump against its twin, one end of a motor against the other. The temperature difference between them, the delta-T, indicates how far the component has drifted from normal.

Severity then combines three inputs. Temperature rise comes first: many maintenance programs treat a rise of a few degrees above a comparable component as a watch item, a double-digit rise as a scheduled correction, and any rise approaching material limits as urgent. Absolute limits matter second: winding insulation classes cap allowable temperature, and Class F insulation, common in industrial motors, is rated to 155 °C, so a rising winding temperature escalates as it nears its class rating. Load is the third input: a 10 °C rise at light load is more alarming than the same rise near full load, because the problem will worsen as load increases.

Rate of change completes the picture. A component that has sat at the same elevated temperature across two annual surveys is a different risk than one that added 8 °C since last quarter. This is where the interpretation discipline known as infrared analysis formalizes the process, pairing measured deltas with load data and component ratings to produce a severity classification and a recommended action window.

Emissivity: The Setting That Makes or Breaks a Reading

Emissivity is the ratio of radiation a surface emits compared with a perfect emitter at the same temperature, on a scale from 0 to 1. Non-metals, painted surfaces, oxidized metal, rubber, and water emit efficiently, typically 0.85 or higher. Polished metals are the problem: polished copper or aluminum can sit near 0.05, emitting almost nothing and reflecting the surroundings instead.

That reflection is the classic error. A camera pointed at a shiny busbar does not read the busbar; it reads the reflected temperature of the surroundings, including the thermographer. A human body at 37 °C reflected in a panel surface can manufacture a false hot spot out of nothing.

Practical fixes are straightforward. Set emissivity to match the surface. On reflective targets, apply a high-emissivity reference patch (electrical tape or paint, both near 0.95) and measure the patch. Avoid extreme viewing angles on reflective surfaces, and correct the reflected apparent temperature input when precision matters. Airflow adds its own error by convectively cooling the surface, so note wind and fan cooling when comparing assets.

Thermal Imaging vs Infrared Analysis vs Thermal Monitoring

The three terms overlap, and they are not interchangeable. Thermal imaging is the camera survey that captures a heat picture. Infrared analysis is the interpretation discipline that turns the picture into a severity call and a corrective action. Thermal monitoring is the continuous side of the practice: fixed temperature sensors watch instrumented points and alarm between surveys. A survey finds what nobody thought to instrument; monitoring watches what someone did.

Aspect Thermal Imaging Infrared Analysis Thermal Monitoring
What it is A camera survey that captures surface heat as an image The interpretation discipline applied to thermal images and temperature data Continuous or periodic measurement at fixed points with installed sensors
Equipment Handheld or mounted infrared camera The same cameras plus analysis software and thermography training RTDs, thermocouples, or fixed infrared spot sensors, wired or wireless
Data collected A radiometric image of each asset at survey time Anomaly classification, severity rating, recommended corrective action A numeric temperature series over time with alarm thresholds
Best at Finding unknown anomalies across many assets in one route Turning an image into a defensible repair decision and priority Catching drift and change on known-critical points between surveys
Limitation A snapshot; validity depends on load and emissivity at capture Adds interpretation cost and depends on thermographer skill Covers only instrumented points and misses faults elsewhere

The Bottom Line

Thermal imaging gives maintenance teams a fast, non-contact way to see the heat that electrical resistance and mechanical friction produce on the way to failure. The technique rewards discipline: correct emissivity, surveys at real load, comparisons against similar components, and a trend record from previous inspections. Run that way, an infrared route finds the developing faults that vibration never touches, documents them in a form a repair planner can act on, and turns a walk through the electrical room into a prioritized work list.

Bring Thermal Inspections Into Your Condition Monitoring Program

Tractian combines thermal, vibration, and operational signals in one asset health platform, so a hot connection or a warming bearing reaches the right person before the failure does.

Explore Tractian Condition Monitoring

Frequently Asked Questions

How accurate are thermal imaging cameras?

Most industrial handheld cameras specify accuracy of about ±2 °C or 2% of the reading, whichever is larger. The larger error sources are usually outside the camera: emissivity settings, reflections from shiny surfaces, distance-to-spot ratio, and airflow across the target. A correctly configured camera on a high-emissivity surface at close range gives readings reliable enough for condition monitoring decisions.

Can thermal imaging see through walls or metal?

No. An infrared camera reads radiation from the first surface it reaches, so it shows the temperature of a panel cover, not the breaker behind it. Heat generated inside a component conducts to the surface, which is why internal hot connections still show, but the image belongs to the surface. Waiting for thermal steady state and comparing external patterns against similar components helps interpret what is underneath.

What causes false readings in a thermal image?

The most common causes are wrong emissivity settings, reflections from low-emissivity surfaces, surveys taken at low load, airflow cooling the target, and measuring from too far away so the spot size includes the background. Water films, oil, and dust also change effective emissivity. Each is controllable with survey discipline: documented camera settings, adequate load, and consistent distance.

How often should thermal imaging inspections be scheduled?

Cadence follows asset criticality and failure history. Many maintenance programs survey critical electrical distribution monthly or quarterly, general equipment annually, and any asset after repair to reset its baseline. Interval decisions should also respond to what the data shows: if a watch-list component rose sharply between surveys, shorten the cycle for that asset rather than for the whole route.

How is thermal imaging different from installing temperature sensors?

A thermal camera covers an entire asset in one image and finds faults nobody thought to instrument, while fixed temperature sensors watch specific points continuously and alarm on change. They work at different stages: sensors catch drift on known-critical points between surveys, and camera surveys sweep for anomalies across everything else. Most plants use both, with sensors on a handful of critical locations and survey routes covering the rest.

Related terms