Interpretation Techniques

As mentioned in the Gas Production Mechanisms section above, the fundamental principle used to infer the condition of the transformer relies on the correlation between gas generation rates and proportions and the energy that produces them.

Since Dissolved Gas Analysis (DGA) started being used as a diagnosis technique, a number of patterns and associations have been found, documented and used as diagnosis tools and guidelines.

While there are a variety of interpretation techniques and guidelines, in general they rely on one or more of the following factors:

The gas profile of a particular sample, typically the latest one.
The relationship (ratios) between concentrations of different gases in a given sample.
The rate of change of these concentrations between samples taken at different times.

In almost all cases, the methods attempt to find and establish statistically significant correlations between the observed gas behaviour and particular modes of operation or failure of the transformers in that population. A sense of what is the “normal” or expected gas behaviour of a certain population, is achieved by analysing the patterns of transformers without any known active failure modes (18) (19). On the other hand, by selecting and categorising known defects and observing their historical trends it is possible to establish some general guidelines about gassing behaviours that correspond to particular failure modes (9) (20).

Transformers are complex pieces of equipment and their construction varies from unit to unit. Factors such as design and manufacturing practices can have an impact in the behaviour of any particular unit.

Whilst the correlations found for a certain population of transformers are useful to gain generic insights and derive broad interpretation guidelines, an in-depth knowledge of the particular characteristics of the transformer population being analysed will result in more accurate and confident interpretations and condition assessments.

The reader might have heard before that the interpretation of gases dissolved in transformer oil is both an “art” and a “science”. As the person using DGA to monitor the condition of power transformers develops expertise and familiarity with a particular fleet or model of transformers; his or her ability to further analyse the fleet’s data, correlate with specific failure modes and gain additional insights becomes more granular and specific.

Furthermore, DGA is only one of the tools in the toolbox necessary to produce a meaningful assessment of a transformer’s condition. In order to arrive at an actionable conclusion, the diagnosis has to be looked at from multiple angles and using all available data including: electrical tests (21), oil quality tests (22), external and internal inspections, paper condition (23), components condition (i.e. OLTC, bushings, etc.), known design or manufacturing issues, loading, ambient conditions, undue electrical stresses (i.e. transients and through-faults), overloading, etc.

Analysing all this information in a comprehensive manner is the “art” component of dissolved gas interpretation that engineers in this industry often refer to. Given the variability in the available data as well as the transformers’ design and construction, the publicly available guidelines tend to document more universal and broader behaviours. Whilst the overall physics and statistical behaviours, in other words the “science”, of gas production is well documented and understood, the more precise and reliable assessments are produced by years of experience in this field and careful consideration of all condition indication factors, the “art”.

Although not strictly part of the DGA, the oil analysis reports often include results for other oil quality tests such as moisture, interfacial tension, acidity, dielectric breakdown, etc. In this case we’ll focus on the interpretation of the DGA component. The references section of this chapter points to additional references (22) in the interpretation of this additional oil analysis information for the interested reader.

xDGA was created with the express purpose to facilitate the application of the most relevant interpretation guidelines used in this industry.

In this first version of xDGA, the following techniques have been implemented: