Managing Editors’ Note:
This article is Part 1 of a 2-part series by Bhaba Das and Anthony Urlich from Dynamic Ratings, Australia. When I first received this article from Bhaba, who is a valued member of our Technical Advisory Board, I planned to publish it in full. I believe it represents one of the best practical approaches to Mitigate Hidden Risks of Rapid Cyclic Loading with Condition Monitoring, so we are splitting it into Part 1 and Part 2, with the latter coming in May magazine.

Read part 2 here.

Introduction

The global energy transition is transforming the way power systems operate. Traditional load patterns, relatively stable, with predictable seasonal peaks, are being replaced by highly variable, rapid cycling driven by renewable generation, electric vehicle (EV) charging, and grid connected battery energy storage systems (BESS). Transformers, once accustomed to slow seasonal variations, are now exposed to multiple daily load cycles [1], with sharp morning and evening peaks.

While thermal ageing of insulation under high load has long been understood [2], a less visible but equally important phenomenon is emerging: the interaction between cyclic thermal stress, moisture migration, and dielectric strength of insulating oil [3], [4]. This interplay has direct consequences for breakdown voltage (BDV), reliability, and ultimately, transformer life expectancy.

These risks are often invisible in standard offline tests or average load/temperature data. Transformers are built with thermal margins and robust materials to handle overloads and variations in demand. Asset managers often focus primarily on average load, oil temperature rise, and hot-spot temperatures. If these values stay within specification, the transformer is often considered “safe.”

Yet, beneath the surface, hidden stresses are at work. Rapid cyclic loading affects not only temperature but also moisture migration dynamics within the oil-paper insulation system. The lag between fast load swings and the slower moisture diffusion process means that localized regions of high moisture concentration can form. These transient conditions may temporarily weaken dielectric strength, lowering breakdown voltage. This is what makes rapid cyclic loading a hidden risk. There have been recent publications on importance of considering both thermal and moisture dynamics in transformer operation from asset management, operational reliability, and strategic decision-making point of view [5]-[6].

Online Condition monitoring (OCM) is a vital tool to uncover this hidden risk. By providing real-time insights, OCM allows end users to detect hidden risks early, make informed operational adjustments, and prevent costly failures. This article explores the hidden risks of rapid cyclic loading, explains how condition monitoring mitigates these risks, and provides practical guidance for end users aiming to enhance transformer reliability.

Traditional Oil Testing

A power transformer’s insulation system is a composite of solid and liquid insulation materials working together to withstand electrical, thermal, mechanical, and chemical stresses. Oil quality testing has long been the primary window into transformer insulation health. Even though the critical aging occurs in the solid insulation, oil is easier to sample and analyse, hence the most tested. Two of the most common transformer oil tests are:

1. Dielectric Breakdown Voltage (BDV): The dielectric breakdown voltage of transformer oil is a direct indicator of its insulating capability. In testing, a controlled voltage is applied across electrodes immersed in the oil until an electrical discharge (breakdown) occurs.
2. Moisture Content: Moisture is one of the most critical aging accelerators in the transformer insulation system. While oil can dissolve some water, the majority is held within the solid insulation (paper/pressboard), which is far more vulnerable to degradation. The gold-standard laboratory method for measuring absolute water content in ppm is Karl Fischer Titration (KFT).

Breakdown Voltage (BDV)

There are two most used test standards for determining the BDV of transformer oil:

• IEC 60156 [7]: This test method is accepted by IEC 60296 (new oil standard), IEC 60422 (maintenance guide) and other various alternative fluid standards.
• ASTM D1816 [8] & ASTM D877 (older standard): Mostly used in North America. IEEE c57.106 accepts ASTM D1816 (and D877) for BDV evaluation of in-service oil.

IEC 60422 sets out maintenance thresholds for BDV values (e.g., >60 kV for ≥170kV transformers for new oil, corrective action below ~50 kV). Similar suggested limits are also available in IEEE c57.106. However, IEEE and IEC both (and CIGRE) emphasize that BDV is a screening test only:

• Non-specific: BDV reduction may result from moisture, particulate contamination, oxidation products-the test does not distinguish between causes.
• Sample handling sensitivity: Even minor contamination during sampling or transport can cause artificially low BDV values.
• Method scatter: Differences between IEC and ASTM test methods lead to measurable variation; single-sample values must be treated cautiously [9].
• Not predictive: A high BDV value today does not guarantee dielectric safety under future cyclic thermal stresses.

Moisture Content

The KFT method remains the reference for measuring absolute water content in ppm:

• IEC 60814 [10]: This test method is accepted by IEC 60296 (new oil standard), IEC 60422 (maintenance guide) and other various alternative fluid standards.
• ASTM D1533 [11]: IEEE c57.106 accepts ASTM D1533 for moisture content evaluation of in-service oil.

Although oil samples are easy to obtain and measure (e.g., using KFT), they are only an indirect indicator of the moisture state of the paper insulation. Since 90–98% of the total water in a transformer is stored in the paper(cellulose), not the oil, relying on oil data alone has significant limitations:

• Moisture Partitioning
  • Moisture is distributed between oil and paper according to temperature-dependent equilibrium curves (CIGRE TB 349 [12]).
  • At high temperatures, oil can hold more water (ppm), so oil may appear “wet” while paper has released moisture.
  • At low temperatures, oil ppm may be very low even though paper remains dangerously wet.
  • Non-Uniform Moisture Distribution in Paper
  • Paper near the winding hot spots often contains more moisture than cooler outer layers.
  • Oil samples from the bulk tank reflect only the “average equilibrium”, not the local critical moisture content in winding hot spots where dielectric risk is highest.
• Sampling and Interpretation Limitations
  • Oil ppm alone cannot be reliably back calculated to paper moisture without a temperature profile, equilibrium models, and paper/oil ratios.
  • Different fluids (mineral, esters) have different water solubilities, making cross-comparison difficult.
• Temperature and Load Cycling Effects
  • Rapid thermal swings (e.g., from renewable or BESS-driven cycling) cause water to migrate back and forth between paper and oil.
  • A single oil sample represents only the condition at that moment and misses these dynamics.
  • Paper can remain chronically wet even when oil appears “dry”during hot operation.
  • Karl Fischer ppm values do not directly indicate relative saturation (RS).
  • Even with “low ppm” values, R may exceed safe limits at local hot spots, increasing risk of bubble formation and dielectric failure.

What do we want to measure – The Moisture Cloud

One of the emerging concepts in transformer condition assessment is the moisture cloud approach. Traditional practice has relied heavily on point measurements of water content in oil (ppm), often obtained through periodic oil sampling. However, these values alone provide limited insight into the dynamic behaviour of water in the cellulose–oil system, especially under cyclic load and temperature variations.

The moisture cloud approach provides a more holistic way of interpreting transformer moisture dynamics by plotting the relationship between oil temperature and moisture-in-oil content over time. When data is trended in this two-dimensional space, characteristic hysteresis patterns called moisture clouds become visible. These clouds represent the continuous exchange of water between insulating paper and insulating oil during heating and cooling cycles. Figure 1 explains the simple formation of the moisture cloud.

Figure 1: Simulated response to load step change: Oil temp, Moisture & %RS

The oil temperature changes as a response to a step change in the load current. The moisture follows the “heating path” and the “cooling path”.

• Heating Path: As the transformer heats under load, water desorbs from the cellulose insulation into the oil, increasing moisture concentration at higher oil temperatures.
• Cooling Path: During cooling, the reverse process occurs—water is reabsorbed by the paper. However, the time constant for the cooling path is significantly higher than the heating path. This cooling path does not retrace the heating line exactly, leading to hysteresis as shown in Figure 2.

Figure 2: Simulated Oil temp Vs Moisture (Hysteresis)

The shape, size, and density of the cloud carry diagnostic value.

• A narrow cloud may indicate relatively dry insulation with low mobility of water.
• A wide or displaced cloud suggests higher bulk moisture in paper, with stronger desorption/absorption effects.
• Asymmetry in the cloud may point to slow diffusion processes or uneven moisture distribution.

Compared with single-point moisture measurements, the moisture cloud approach is particularly powerful for diagnosing transformers subjected to rapid cyclic loading (such as in wind, solar, or EV/BESS integration). It not only quantifies average moisture levels but also visualizes the dynamic interaction between oil and paper, offering a more realistic picture of insulation health and risk.

While moisture-in-oil concentration (ppm) is useful, it does not directly reflect the risk of condensation or reduced dielectric strength. This is because the same ppm value corresponds to very different moisture risk levels at different oil temperatures. Relative saturation (%RS) accounts for this by normalizing the dissolved water content against the oil’s temperature-dependent saturation limit.

CIGRE 741 recommends using relative saturation (%RS) as the preferred parameter for continuous monitoring. Figure 3 shows the %RS vs. Oil Temperature Moisture Cloud. It bridges laboratory-based understanding of oil-paper equilibrium with practical online monitoring, making it especially important for transformers exposed to renewable and storage-driven cyclic loading.

Figure 3: Simulated %RS vs Moisture (Hysteresis)

When plotted as %RS vs oil temperature, the transformer’s dynamic moisture behaviour forms a distinct cloud pattern like the ppm temperature case, but with sharper diagnostic meaning:

• Heating Path: As oil warms, its moisture-holding capacity increases. This causes %RS to decrease even if the absolute ppm of moisture rises, since the oil can now hold more water without approaching saturation.
• Cooling Path: As oil cools, saturation capacity drops, so %RS rises more steeply. If the cellulose releases water into oil during this phase, the %RS can spike, sometimes crossing critical thresholds (>50%), indicating significant risk reduction in breakdown voltage.

Figure 4: Simulated example of Moisture Clouds (%RS vs Oil Temperature)

The shape, size, and density of the %RS cloud (Figure 4) carry significant diagnostic value:

• A tight, low-lying cloud indicates a dry system withstable insulation moisture.
• A broad cloud extending into higher %RS reflects significant paper moisture content and strong exchange during load cycling.
• Clouds touching or exceeding 100% RS suggest supersaturation and possible water droplet formation, which is highly detrimental to dielectric strength.
• Asymmetry between heating and cooling curves highlights hysteresis in the paper–oil system and can reveal sluggish diffusion or uneven drying or even show us the oil sensors are installed incorrectly.

Compared with ppm-based plots, %RS clouds are more directly related to dielectric risk [14] and moisture saturation events. Figure 5 shows the averaged BDV vs %RS plot. This makes %RS moisture cloud a valuable tool for identifying when rapid cooling (e.g., in solar or wind transformers during generation dips, or BESS (or EV) charge/discharge cycles) could push the insulation system into dangerous operating regions.

Figure 5: Average %RS vs BDV curve – New Mineral oil, Aged Mineral Oil

References
[1] I. Alvarez Fernandez, et.al, “Transformer thermal ratings with evolving load profiles,” 27th International Conference on Electricity Distribution (CIRED 2023), pp. 2431-2435, Italy, 2023.
[2] Tapan Kumar Saha and Prithwiraj Purkait, “Transformer Insulation Materials and Ageing,” in Transformer Ageing: Monitoring and Estimation Techniques, Chapter 1, pp.1-33, Wiley-IEEE Press, 2017.
[3] O. Roizman, “Moisture equilibrium in transformer insulation systems: Mirage or reality? Part 1, Transformer Magazine, Vol 6, Issue 2, 2019.
[4] O. Roizman, “Moisture equilibrium in transformer insulation systems: Mirage or reality? Part 2, Transformer Magazine, Vol 6, Issue 3, 2019.
[5] A. Al-Abadi, A. Gamil and A. Sbravati, “Determination of Moisture Content during Dynamic Loading of Liquid-Filled Distribution Transformers,” 2024 IEEE Electrical Insulation Conference (EIC), Minneapolis, USA, pp. 134 138, 2024.
[6] A. Al-Abadi, J. Bobrowski and A. Gamil, “Enhanced Transformers Thermal and Moisture Distribution Modelling for On-Line Assessment of Insulations Condition,” 2025 IEEE International Conference on Dielectric Liquids (ICDL), Lodz, Poland, pp. 1-4, 2025.
[7] IEC 60156:2025, Insulating liquids – Determination of the breakdown
voltage at power frequency – Test method.
[8] ASTM D1816-12(2019), Standard Test Method for Dielectric Breakdown Voltage of Insulating Liquids Using VDE Electrodes.
[9] Vitaly Gurin and Marius Grisaru, “The statistical scatter of breakdown
voltages of transformer oil – Part I”, Transformer Magazine, Vol 11, Issue 4, 2024.
[10] IEC 60814:1997, Insulating liquids – Oil-impregnated paper and pressboard – Determination of water by automatic coulometric Karl Fischer titration.
[11] ASTM D1533-20, Standard Test Method for Water in Insulating Liquids by Coulometric Karl Fischer Titration.
[12] CIGRE Technical Brochure 349, Moisture equilibrium and moisture
migration within transformer insulation systems, WG A2.30, 2008.
[13] CIGRE Technical Brochure 741, Moisture measurement and assessment in transformer insulation – Evaluation of chemical methods and moisture capacitive sensors, WG D1.52, 2018.
[14] Vaisala White paper, “The Effect of Moisture on the Breakdown Voltage of Transformer Oil”, 2013.