Lord Consulting
Power Industry Asset Management Consultants
Lord Consulting
Power Industry Asset Management Consultants
Home » Technical Papers » Dissolved Gas Analysis
Analysis of Intelligent Electronic Device and instrument generated dissolved gas data …prior to and after the repair of a critical transformer
On Monday September 30th, 2002, a 650 MVA GSU was taken out of service urgently at one of Mirant’s major generating plants. Data used to make this decision consisted primarily of dissolved gas measurements generated by an on-line dissolved hydrogen and water monitor, and of on-site DGA measurements made using a portable multi-gas analyser.
Mirant operates several generating plants in North America, Asia and the Caribbean. In late 2001, routine DGA (Dissolved Gas Analysis) performed on a Mirant Mid-Atlantic 650 MVA GSU built in 1969, lead the electrical maintenance crew to believe an incipient fault was developing in this transformer. Table 1 below provides an example of the DGA values obtained at that time.
Routine DGA tests during the 6-month period following the above mentioned analysis showed no significant changes in fault gas concentrations. However, the decision was made to increase the DGA testing frequency and to install an on-line dissolved hydrogen and water monitor on the suspect transformer. The selected IED (Intelligent Electronic Device), a Calisto monitor from Morgan Schaffer Systems in Montreal, Canada, was installed. Figure 1 describes the main features of the instrument which continuously measures the concentration of dissolved hydrogen in oil in ppm. Dissolved water content is also continuously measured and can be reported in ppm, %RS (Relative Saturation) at 25 °C, or %RS at a specific transformer oil temperature if this value is available as an input to the monitor (4-20 mA or J-type thermocouple input).
As opposed to other sensors of this type, the Calisto monitor features its own oil circulation and conditioning system, i.e. the circulating oil is either cooled or heated to a set point prior to making dissolved hydrogen and water measurements.
Dissolved gases are continuously extracted using permeable PTFE fibres and selective measurement of dissolved hydrogen in the gas sample is achieved using a specially designed thermal conductivity detector (TCD). Continuous dissolved water measurement is performed using an IC (integrated circuit) sensor located directly in the oil flow.
As the IED continuously circulates oil to get access to representative dielectric fluid from the transformer, connection to independent supply and return valves on the transformer is required. Standard copper tubing and off-the-shelf fittings are used to make the oil line connections. The manufacturer also offers cut-to-length stainless steel flexible hose and fitting kits if added security or protection is required (e.g. use in corrosive environments).
The installation procedure involves (1) mounting of the IED on the side of the transformer control cabinet or on a mounting plate, (2) connection of the AC supply, communication and alarm lines, (3) connection of the oil lines and flushing of the oil circuit and (4) start-up of the instrument. No field calibration or signal adjustments are required once the instrument has been started.
The IED was installed in May 2002 and decision was made to degas the transformer in order to reset the dissolved gas reference to zero. Laboratory and on-site DGA measurements were performed regularly in the period of June to mid- September 2002. During that period, the IED showed a rate of dissolved hydrogen generation that varied according to Table 2 and Figure 2 below:
From this data, one can observe that the hydrogen generating rate significantly increased on September 13th, thus indicating a sudden change in the gas generating conditions. Figure 3 shows the dissolved water variation for the same period. As the amount of water in oil is dependent, amongst other parameters, on the load on the transformer, Figure 3 provides qualitative information on the load variation for the period. Based on this information, it is interesting to observe that the hydrogen generating rate does not significantly vary with load. Moreover, the sudden change in hydrogen generating rate on September 13th was not induced by an increase in load thus indicating a true deterioration of the transformer condition. On-site DGA tests performed during the period are shown in Table 3. Though large increases in fault gas concentrations are observed, percent changes between tests for both absolute concentrations and key ratios remained more or less constant.
Figure 4 shows the period of September 27th to September 30th where very large amounts of hydrogen were produced under fairly constant load conditions. This graph demonstrates the speed with which a transformer condition can degrade and with which the hydrogen concentration can increase. Table 4 shows the maximum concentration and hydrogen generating rate reached during periods A, B and C. It should be noted that the ability of the sensor to pick up such increase in hydrogen concentration is very dependent on the oil circulation within the transformer and on the ability of the sensor to have access to this representative oil. Results from an on-site DGA performed on September 30th, are provided in Table 5.
The authors analysed the data contained in Table 3 and Table 5 using the well-known Duval Triangle method. Based on this method, a diagnostic of thermal fault of medium range temperature T2 (300 to 700 °C) was established from July 8 to September 17 (it may be noted that a similar fault T2 was also present prior to November 11, 2001, as indicated from Table 1). The September 30th data then led to a high temperature T3 (>700 °C) thermal fault diagnostic. From these diagnostics, one can suspect that a source of high electrical resistance was initially present in the transformer (e.g. a bad connection). This high resistance created a resistive hot spot of increasing resistance, thus the observed increase in hydrogen, methane, ethylene and ethane.
At some point between September 17 and 30 (possibly on September 27th), additional energy, probably due to circulating currents, increased significantly the temperature of the metal and oil at the fault location. This led to accelerated production of hydrogen, methane, ethylene and a small relative amount of acetylene, which are indicative of overheating of the oil above 700 °C.
Once the transformer was taken out of service, it was found that a high voltage lead was burned and ready to fail. This finding was consistent with the DGA data, and a technical paper dedicated to providing information on the fault cause and subsequent transformer repair will be published in the coming months.
The above experience demonstrates the value of on-line monitoring of dissolved hydrogen for both life extension and transformer protection purposes. Such value lies in the ability of the IED to provide real time and long term data that is meaningful and truly representative of the transformer condition.
This ability of the IED to provide such information for any type of transformer (new, old, containing contaminants, etc.) and in any type of environment is also of paramount importance.
In order to maximise this value, the following recommendations can be made:
The Mirant GSU was repaired and re-energised on October 21st, 2002. Since then, dissolved hydrogen concentration has remained constant at approximately 30 ppm.
Acknowledgments: The authors would like to thank Mr. Michel Duval from the Institut d’Hydro Québec (IREQ) and Dr. James J. Dukarm from Delta-X Research for their help in analysing the DGA data of the present case.