Thermal Monitoring: How VIE Tells the Difference Between a Hotspot and a Cooling Problem
Not all thermal anomalies in a power transformer point toward the same failure mode. VIE's multi-height sensor placement distinguishes between two conditions that produce similar alarm language but require different responses: insulation stress in the upper winding region, and cooling system obstruction. Confusing the two leads to the wrong intervention at the wrong time.
Why Surface Temperature Alone Is Not Diagnostic
A transformer's surface temperature at any given point is a function of four variables: the heat generated internally by load-related losses, the transformer's thermal resistance at that location, the ambient temperature, and any local anomaly that breaks the expected thermal gradient.
A single temperature reading — from a spot measurement or a periodic infrared scan — cannot separate these contributions. A unit running at high load in hot weather will have a higher surface temperature than the same unit at low load in cool conditions. That temperature difference tells you nothing about whether the transformer is healthy or not.
VIE's thermal model addresses this by computing the surface temperature the transformer should produce at each sensor height given its current load and ambient conditions. The metric VIE produces is not absolute temperature. It is the residual: the difference between what the model predicts and what the sensor measures. A transformer running hot because of high load in hot weather produces a residual near zero — the model accounts for both. A transformer producing excess heat relative to its load and ambient conditions produces a positive residual that the model cannot explain through normal operating parameters. That unexplained excess is the diagnostic signal.
Excess Heat Flux at the Top of the Tank
When excess heat flux appears at the uppermost sensor positions, the most likely causes are localized to the upper winding region: insulation degradation, conductor overheating, or oil quality problems that reduce the fluid's ability to carry heat away from the winding surface.
The upper portion of the transformer winding is typically the hottest point in the unit under normal loading. It is where the oil exits the winding channel before being returned to the cooling radiators. If insulation in the upper winding region is degrading — through aging, moisture ingress, or sustained overtemperature — the thermal losses from that section increase, producing excess heat flux that concentrates at the upper sensors.
Rising excess heat flux at the top of the tank calls for increased monitoring frequency and correlation with Dissolved Gas Analysis (DGA) trends. If DGA shows rising dissolved combustible gas alongside the thermal anomaly, the combination is a priority signal. If DGA is clean and only the thermal metric is elevated, the recommended response is to increase VIE's sampling frequency and watch for convergence with other metrics before escalating.
Excess Heat Flux at Lower Sensor Heights
When excess heat flux appears at lower sensor positions, or when the thermal gradient across the tank does not match the expected top-weighted profile, the cause is more likely to be a cooling system problem than an internal winding condition.
A transformer's normal thermal gradient runs from coolest at the base to hottest near the top, following the natural convection of heated oil rising through the winding channels. When that gradient is disturbed — flattened, inverted, or showing anomalous readings at specific heights — the disruption typically reflects something interfering with oil circulation or heat transfer.
The most common causes are: blocked or fouled radiator fins reducing heat dissipation, oil sludging that increases viscosity and slows circulation, cooling pump or fan failure in forced-cooled units, or oil contamination that reduces thermal conductivity. Each of these produces a characteristic pattern in VIE's thermal metrics that differs from the upper-winding hotspot pattern.
When excess heat flux is concentrated at lower sensor heights, the recommended response is to inspect the cooling system before ordering electrical tests. In many cases, restoring adequate cooling — cleaning fins, addressing oil quality, repairing a cooling pump — eliminates the thermal anomaly without any winding fault being present. Acting on a cooling problem as though it were a winding fault wastes diagnostic resources and delays the correct intervention.
When Both Patterns Appear Together
When excess heat flux appears at both upper and lower sensor positions simultaneously, or when the thermal metric is rising across the entire sensor array, the combination points toward oil quality degradation affecting the whole unit.
Degraded oil is both a thermal conductor and a thermal generator. As oil quality deteriorates, its ability to carry heat away from the windings decreases, raising temperatures across all sensor heights. Simultaneously, degrading oil may produce additional heat through dielectric loss and chemical reactions. When VIE's oil health metrics (V2P and S2P) are also rising at the same time as a whole-tank thermal anomaly, treat the combination as a priority: order an oil quality lab test immediately. Addressing oil quality directly may prevent the thermal condition from escalating to insulation damage.
The Thermal-Vibration Model
VIE's thermal analysis integrates surface temperature measurements with vibration data. The same sensor that measures surface temperature also measures the vibration signature from which winding mechanical health is assessed. This integration is what makes VIE's thermal diagnostic more than a surface measurement: thermal anomalies are evaluated alongside the mechanical state of the winding structure and the condition of the oil that transmits both heat and vibration.
VIE's thermal-vibration model was submitted to the CIGRE 2025 International Symposium. The model's approach to separating load-driven temperature from fault-driven excess is the technical foundation for the residual heat flux metric and its interpretation framework described in this article.