The Dissolved Gas Analyis
Determination of thermal and electrical condition
Transformers play an important role for the production and distribution of electric energy. Therefore, the operators should have a vivid interest in recognising rapidly the faultless operating condition of the transformer and detecting faults fast and exactly. Today the Dissolved Gas Analysis “DGA” is one of the most frequently applied diagnostic methods for the recognition and assessment of faults in oil-filled electric equipment. This is due to the characteristics of insulating oils dissolving split gas along with atmospheric air which occur with natural oxidation or thermal and electric faults from oil or cellulose material. The extraction of all the gas from oil and their quantitative determination leads to a transformer assessment in regard of its thermal and electric condition. DGA grants the opportunity of recognising creeping or critical defaults in time and identifying them with typical key gases or concomitant gases [1,2]. The chronological development of the gas concentration (long-term gas reactions) leads to a trend analysis permitting to undertake appropriate measurements.
The basic requirement for an instructive analysis is technical correct and representative oil sampling as well as its conformity with standards. Generally, three different tanks are used for sampling : Gas-tight syringe, sample cylinder (metal or glass) and bottle (metal or glass). Normally, one prefers the bottle and the syringe. Bottles especially aluminum bottles have the advantage of being unbreakable and usage without any problem on-site. Apart from a conventional transport this is possible due to the unbreakability of the sample tank. Furthermore, due to the large volume of samples even oil samples with a low gas content, e.g. acceptance tests of new transformers, can exactly be examined . Glass syringe (with a low sample content) presents itself for electric devices with only a low oil content, e.g. for instrument transformers. The Dissolved Gas Analysis is realised in two steps: First the dissolved gases are extracted from oil, then they are analysed by means of gas chromatography.
Extraction of Gases from Oil
The first step for realising the DGA is to degas the oil sample. IEC 60567 describes the following extracting methods in detail: multi-periodic vacuum extraction by means of the Toepler-Pump, vacuum extraction by partial degassing, stripping method und headspace method.
Gas Chromatographical Determination
In a second step the analysis of the extracted gas combination is carried out. Apart from the atmospheric gases nitrogen (N2) and oxygen (O2) also the split gases carbon dioxide (CO2), carbon monoxide (CO), hydrogen (H2), methane (CH4), acetylene (C2H2), ethylene (C2H4) and ethane (C2H6) are determined qualitatively and quantitatively. Propylene (C3H6) and propane (C3H8) are useful indications for a reasonable interpretation of DGA in case of an accumulation of different sources of error. After the gas has been collected it is injected into the gas chromatograph. By means of an inert carrier gas (mobile phase) key gases and concomitant gases are transported through a separating column (stationary phase). Different physical and chemical characteristics of each component evoke various reciprocal effects with the surface within the column and thus cause different operating delays of each compound. The gas components arriving at the end of the column are registered by detectors and recorded as a chromatogram. The most frequently used detector is the flame ionisation detector (FID), which can be used for all flammable gases. It burns gases in a hydrogen flame and registers the change of conductivity provoked by ionisation in the flame. Before this carbon dioxide and carbon monoxide are catalytically reduced to methane and thus can also be determined by this method. Therefore hydrogen, used as combustion gas in FID cannot be determined this way. Oxygen, nitrogen and hydrogen are determined by means of a thermal-conductivity detector (TCD). The signals (peaks) for each gas are allocated by comparing them with reference gases, by assessing them quantitatively with surface integration below the peaks and comparing them with a calibration gas of known concentration.
Even during the normal operation characteristic dissolved gases are produced by the thermal degradation of the paper-oil dielectric. Consequently, these gas characteristics may not be interpreted as a fault, but must be considered as a typical “gas formation” due to normal ageing. If the result of DGA deviates from the common scheme of “natural ageing”, it is necessary to check whether there is a creeping or acute fault. In practice one has to distinguish between thermal and electric faults. Thermal overheating at a hot spot leads to the pyrolysis of hydrocarbons. At the same time for different temperature levels characteristic key gases are produced concomitted by less characteristic gases. Under the 300°C level most of all propylene is produced concomitted by ethylene. In a hot spot in the temperature field of 300°C to 700°C mostly ethylene is produced concomitted by propylene and minor portions of methane and hydrogen. Above 700°C mainly ethylene is produced concomitted by methane and minor portions of hydrogen and propylene, above 1000°C also acetylene. Electric discharges with high energy (flashover and sparking) cause hydrogen and acetylene splitting off from methane and ethylene. Partial discharges of low energy mainly lead to the production of hydrogen and methane as well as minor portions of ethane. If electric faults occur along with solid insulation, carbon monoxide and carbon dioxide are produced as decomposition products in an even wider extent as occurring with the normal natural cellulose decomposition. The gas concentration measurable in oil not only depends on the type, energy and duration of the fault, but also on the oil volume absorbing the gas. Consequently, the assessment of the analysis result focuses more on the allocation of gas than their absolute concentration. In the past one tried to describe this with quotients [1,2]. The recent approach was described in IEC 60599  and is based on the quotients C2H2/C2H4, CH4/H2 and C2H4/C2H6 with a very general interpretation scheme . Apart from this the ratio CO2/CO is also examined because it can be a sign whether solid insulation is damaged or not. In general, this standard and the referring interpretation scheme may only be considered as a guide, which cannot substitute years of experience of the realisation and assessment of the Dissolved Gas Analysis. The final interpretation of DGA and each decision for resulting actions should only be realised with a technical competent assessment.
- Müller R., Soldner K. and Schliesing H., Elektrizitätswirtschaft 76 (11), pp. 345-349 (1977)
- Dörnenburg E. and Hutzel O., ETZ-A 98 (3), pp. 211-215 (1977)
- IEC 60567, Mineral oil-filled electrical equipment - Sampling of gases and analysis of free and dissolved gases - Guidance
- IEC 61181, Mineral oil-filled electrical equipment - Application of dissolved gas analysis (DGA) to factory tests on electrical equipment
- IEC 60599, Mineral oil-impregnated electrical equipment in service - Guide to the interpretation of dissolved and free gases analysis