Conference Agenda

Dissolved Gas Analysis (DGA) and offline electrical testing are fundamental tools for diagnosing faults in power transformers. Although these two techniques differ in their approach, they complement each other by providing a comprehensive view of the equipment's operational condition, which is crucial for ensuring its reliability and extending its service life.

DGA does not require the transformer to be disconnected and is highly sensitive to early-stage thermal and electrical faults. It allows detection and quantification of gases generated in the insulating oil due to overheating, electrical arcing, or partial discharges, among other phenomena. These gases, such as hydrogen (H₂), methane (CH₄), ethane (C₂H₆), acetylene (C₂H₂), and other gases are analyzed to identify the type and severity of the fault. On the other hand, offline electrical tests provide a detailed evaluation of the transformer’s electromechanical and dielectric condition. These tests include winding resistance , transformer turns ratio (TTR) , tan delta (tan δ) , and dielectric response tests, among others. However, unlike DGA, offline electrical testing requires the transformer to be de-energized and isolated from other devices, which can present a limitation in situations where service interruption is complex or undesirable.

It is common for an unusual increase in certain gases detected by DGA to indicate an internal problem that is then confirmed with greater accuracy through offline electrical testing. For instance, if the DGA reveals an increase in gases related to partial discharges, such as hydrogen, tests that assess insulation conditions could confirm this deterioration. Similarly, an abnormal value in winding resistance measurements may correlate with hot spots detected through DGA, suggesting loose connections or thermal dissipation issues.

However, there are cases where an abnormal condition diagnosed by DGA is not confirmed by an electrical test, or vice versa. This discrepancy can occur because each technique has its own limitations and sensitivity to different types of faults. For example, a low-magnitude dielectric fault may not produce enough gases to be detected by DGA, but it could be identified through a dielectric response test.

This paper presents various case studies illustrating the correlation (or lack thereof) between DGA results and electrical tests. The combination of both techniques allows for a more comprehensive and accurate diagnosis, helping to improve decision-making in predictive maintenance. Furthermore, this correlation aids in interpreting potential discrepancies between methods, optimizing inspection intervals, and managing assets more efficiently. In conclusion, the integration of these tools enhances fault detection accuracy, resulting in more reliable transformer management.

The dynamic thermal modelling of transformers is subject to many constraints which have always been a challenge to match the design phase with the real operation under different conditions such as variable loading, short- and long-term emergency loading. With the increasing awareness of real-time modelling and the growing trend nowadays to use accurate and detailed models that give a clearer picture of the thermal behavior of the transformer in service, the improvement of the available models has become essential.

One of the most important key factors that affecting the performance, lifetime and health condition of transformers is the thermal performance. Temperature affects the insulation system by increasing ageing factors such as moisture distribution, gassing, bubble formation, degree of polymerization, etc. The insulation system mutually interacts with the cooling process and therefore the health condition of the insulation system is essential to maintain its main functions.

Natural ester liquids that are characterized by higher biodegradability and lower CO2 footprint compared to other liquids beside relatively lower market price, leading to a significant growth in their demand. Therefore, the operational reliability for asset management and lifetime expectancy of a transformer filled with natural ester is essential to achieve a higher sustainability value. On the other hand, compared to mineral and synthetic ester liquids, natural esters present more challenges in cold climate operation due to their relatively high pour point. Monitoring requires a reliable dynamic transformer thermal model (DTTM) to accurately determine the response to the real-time loading profile under environmental conditions for different insulating liquids allowing energy maximization without compromising sustainability.

The DTTM has been receiving increased attention because it has a direct impact on the lifetime of the transformer and is dependent on the transformer load profile and environmental conditions which cannot be easily estimated when evaluating transformers onsite. The implementation of sophisticated and rapid computational models for monitoring dynamic thermal behavior provides opportunities for real-time control and optimization of transformer performance, maintenance, and loadability while ensuring insulation system integrity and avoiding misleading operational decisions.

This study presents investigations of the dynamic thermal behavior of a 66 kV KFWF transformer filled with natural ester insulating liquid, designed for offshore wind applications operating in cold climates. The transformer is equipped with fiber optic sensors to monitor the temperature distribution of the winding and liquid during the heating and cooling processes. The case study transformer is investigated through a heat run test by injecting a load while the transformer is exposed to sub-zero temperatures for 100 hours, with the entire transformer placed in a controllable temperature cold chamber. The study includes an extended DTTM that considers the variation of liquid properties with temperature to predict the dynamic temperature behavior of the entire transformer during the test.