Why Is Transformer Moisture High After Dry Air Generator Circulation?

In field substation maintenance, reducing moisture within the transformer insulation system is critical to maintaining dielectric strength and preventing premature aging. A common technical bottleneck encountered by field service engineers is the phenomenon where the moisture levels remain unacceptably high even after executing a continuous dry air circulation process.

Field operators often report that the dry air generator for transformer has run for 48 to 72 hours, with the online dew point meter indicating a stable output of -50°C to -60°C, yet the post-treatment insulation testing reveals that the moisture content within the solid insulation has not decreased to target levels. This technical analysis isolates the thermodynamic, physical, and chemical reasons behind this issue and details field-proven diagnostic and corrective methodologies.

The Core Culprit: The “Hide-and-Seek” of Moisture in Oil-Paper Insulation System

To diagnose why moisture remains high, a field engineer must look at the mass distribution of water within the transformer tank. A transformer insulation system is a composite of liquid insulation (mineral oil) and solid cellulose insulation (pressboard, kraft paper, and laminated wood).

The distribution of water between these two media is highly unequal:

• The 95% vs. 5% Rule: Under normal equilibrium conditions, more than 95% to 99% of the total water mass inside a transformer is trapped within the solid cellulose matrix. Less than 1% to 5% is dissolved in the insulating oil.
• Cellulose Hydrophilia: Cellulose molecules contain a high density of hydroxyl groups (-OH), which form strong hydrogen bonds with water molecules. This chemical affinity makes it highly difficult to extract water from the paper compared to extracting it from the oil or the gas phase.

When a field engineer deploys a dry air machine, the ultra-dry air passes through the gas headspace or flows through the oil phase, creating a steep moisture concentration gradient. The dry air rapidly absorbs moisture from the gas phase and the surface layer of the oil. However, this process only addresses the surface-level moisture. The water molecules embedded deep within the thick pressboards and winding insulation layers must first migrate to the surface before they can be removed by the circulating air.

Mass Transfer Limitations and Diffusion Time Lag

The primary physical mechanism controlling moisture removal during dry air circulation is molecular diffusion, which is governed by Fick’s Second Law:

∂C/∂t=D∇2 C

Where C is the moisture concentration, t is time, and D is the diffusion coefficient of water in cellulose.

In field environments, the diffusion coefficient D of water within solid pressboard is extremely small, typically ranging from 10⁻¹¹ to 10⁻⁸ cm²/s, depending heavily on temperature. Because of this low diffusion rate, a significant time lag exists between the drying of the insulation surface and the migration of water from the core of the insulation to that surface.

During a standard dry air circulation cycle, the dry air generator for transformer lowers the dew point of the internal atmosphere, creating a temporary, local equilibrium at the immediate boundary layer of the paper. This creates a “false dryness” or a deceptive equilibrium. The moisture sensors on the exhaust line of the dry air machine will register a low moisture content because the air is moving faster than the rate at which the deep cellulose layers can release their water content. Once the circulation stops and the transformer is sealed, the internal system moves back toward a true thermodynamic equilibrium. Water from the deep insulation layers diffuses outward into the drier surface layers and the oil, causing the measured moisture values to “rebound” or increase significantly after 24 to 48 hours.

Common Field Blind Spots Causing “Fake Dryness”

Beyond the fundamental physics of diffusion, three practical engineering limitations in the field prevent dry air circulation from achieving deep dehydration.

Blind Spot 1: Absence of High-Vacuum Sublimation Forces

Dry air circulation works at atmospheric pressure or with a small positive pressure (about 0.03–0.05 MPa) to avoid the intrusion of moist ambient air. The only factor that drives water removal is the differential in the partial pressure of water vapor between the cellulose surface and dry air.

Unlike high-vacuum desiccation processes (where absolute pressures are dropped below 133 Pa), atmospheric dry air circulation cannot lower the boiling point of water. It lacks the mechanical pulling force required to break the capillary forces holding bound water inside the microscopic pores of aged cellulose.

Blind Spot 2: Inadequate or Non-Uniform Temperature Gradient

The diffusion coefficient D is an exponential function of temperature, modeled by the Arrhenius equation:

D=D0·e ^{− Ea/R·T} 

Where Ea is the activation energy for diffusion, R is the universal gas constant, and T is the absolute temperature in Kelvin.

If the field engineer runs the dry air machine at ambient temperatures ( like 20°C to 30°C ), the diffusion rate just stays pretty low, so you won’t really get deep drying in a time period that’s economically viable. To release bound water, the core winding temperature must be raised uniformly to between 60°C and 70°C. If there is a poor temperature gradient—such as the top of the tank being hot while the bottom remains cool—moisture will simply migrate from the warmer zones and condense in the colder, uncirculated zones of the transformer.

Blind Spot 3: Micro-Leakage and Internal Pneumatic Dead Zones

Large power transformers have complex internal geometries, consisting of core cooling ducts, phase-to-phase barriers, and radiating structures. During a circulation cycle, air follows the path of least resistance.

If the internal baffling or the circulation hose layout is incorrect, pneumatic “dead zones” are created where the dry air remains stagnant. The moisture in these dead zones does not get evacuated. Furthermore, if any flange gaskets, valves, or bush mountings exhibit micro-leakage, the constant introduction of ambient moisture will neutralize the drying capacity of the incoming air from the dry air generator for transformer.

Field-Proven Solutions: Beyond Standard Dry Air Circulation

When standard dry air circulation fails to meet the moisture specifications outlined in standards such as IEEE Std C57.106 or IEC 60422, field engineers must upgrade the processing methodology based on the severity of the moisture ingress.

Drying Method	Core Mechanism	Field Advantages	Limitations & Best Use Case
Dry Air Circulation	Vapor pressure differential at atmospheric pressure.	Safe for old windings; low mechanical stress on the tank; active online monitoring.	Limited to surface and gas-phase moisture; ineffective for deep pressboard insulation (>3% moisture).
Hot Oil Circulation + High Vacuum	Thermal activation paired with absolute pressure reduction (<133 Pa).	Lowers the boiling point of water; extracts bound water from deep cellulose layers.	Requires full oil management; high risk of tank deformation if the vacuum rating is exceeded.
Low-Frequency Heating (LFH) + Vacuum	Direct copper heating via low-frequency current under vacuum.	Uniform heating directly from the inside of the insulation outward.	Requires specialized high-power LFH equipment; typically reserved for critical site repairs.

Engineering Checklist for Field Execution

To maximize the efficiency of a dry air machine and prevent false equilibrium readings, field engineers should follow this verification protocol:

• Verify Input Dew Point: Ensure the dry air generator for transformer consistently delivers air with a dew point below -55°C at the rated volumetric flow rate (m3/h).
• External Tank Insulation: Wrap the transformer tank in thermal blankets to minimize heat loss and eliminate internal cold spots that cause moisture trapping.
• Implement the 48-Hour Stabilization Rule: Never perform definitive moisture tests (such as Karl Fischer titration of oil or Frequency Domain Spectroscopy) immediately after turning off the dry air apparatus. Allow the transformer to rest in an unforced thermal state for a minimum of 48 hours to permit moisture equilibrium to stabilize across the oil-paper boundaries.

Field Frequently Asked Questions

Q1: What level of moisture content in transformer insulation paper is considered severe moisture ingress? How is this classified according to international standards?

A1: According to IEEE Std C57.91 and IEC 60422 standards, the moisture content (by weight percent) of transformer solid insulation paper is generally classified into the following four moisture ingress levels:

• Dry: <1.5%. This is the ideal state for newly commissioned transformers or units after deep dehydration processing.
• Moderately Wet: 1.5%−3.0%. This is a common condition found during industrial operation and requires close monitoring.
• Wet (Severe): 3.0%−4.5%. This state significantly accelerates the aging rate of the cellulose, causing a rapid decline in the mechanical strength (Degree of Polymerization – DP) of the paper, and markedly reduces the Partial Discharge Inception Voltage (PDIV).
• Very Wet (Extremely Dangerous): >4.5%. In this state, during high-temperature operation of the transformer or sudden short-circuiting, the moisture inside the insulation paper turns easily into vapor and forms gas bubbles (the Bubbling Effect), resulting in catastrophic dielectric breakdown. If the field measurements are greater than 3.0%, then the root problem cannot be solved by the normal dry air circulation, but only by the implementation of a deep vacuum dehydration process.

Q2: During the dry air circulation process, why is the dryness still considered substandard after shutdown, even if the exhaust air dew point has remained stable at -50°C for 12 consecutive hours?

A2: This is a classic case of a “gas dynamic flushing illusion.” When a large volume of dry air circulates within the transformer tank, the passing dry air rapidly sweeps away the trace amounts of moisture released from the surface of the pressboard.

Because the diffusion rate of moisture from the deep layers of the cellulose fibers to the surface (limited by the diffusion coefficient) is far slower than the rate at which the air circulation removes it, the moisture in the exhausted air becomes highly diluted. Consequently, the online dew point meter displays an artificially low value.

This metric only indicates that the “gas headspace” and the “surface layer of the pressboard” inside the transformer tank are dry; it does not prove that the core of the solid insulation is dehydrated. The equipment must be shut down and isolated for 24 to 48 hours to allow deep-layer moisture to fully diffuse outward and reach a new thermodynamic equilibrium with the internal air. Only then does the measured equilibrium dew point hold true as an engineering reference value.

Q3: How can field engineers accurately determine whether a transformer’s oil-paper insulation system has achieved true moisture equilibrium?

A3: In field operations and maintenance, determining true equilibrium relies primarily on the following two methodologies:

• The Temperature-Pressure Equilibrium Curve Method: Following the shutdown of the dry air generator and closing of the transformer tank, the top oil temperature or the winding average temperature should be continuously measured in addition to the gas dew point within the transformer tank or the moisture content of the oil in parts per million. Record these data points continuously for at least 48 hours. Plot the measured “stable temperature” and “stable moisture/dew point” onto standard Oommen equilibrium curves or Paoletti curves. If the calculated moisture content (%) of the insulation paper derived over 24 consecutive hours remains constant and shows no further drift, the system has achieved true physical equilibrium.

• Frequency Domain Spectroscopy (FDS): Use a transformer dielectric response analyzer to perform a sweep of the dissipation factor (tanδ) across low frequencies (typically from 10−4Hz to 103Hz). FDS technology directly measures the response of the solid insulation. The curve characteristics in the low-frequency band directly reflect the true moisture content and conductivity inside the paper fibers, completely unaffected by short-term moisture fluctuations in the oil. It is currently the most precise technical method available for determining real insulation dryness in the field.