In a sense, a new era in Dissolved Gas Analysis (DGA) has arrived. And not too soon; as so much of this understanding has been more of an art than a science in the past. With the ongoing and much –reported age-wave retirement, much intimate knowledge of transformers is leaving the utilities. The embedding and systemization of their knowledge into interpretation software programs has been something we have now added to our Dissolved-Gas-Analysis (DGA) transformer monitoring devices.
Automatically-generated and actionable, Dissolved-Gas-Analysis (DGA) diagnostics for transformers.
Paradigm shift: takes the industry from a detective-corrective mode to a strategic-preventative mode.
These days, the electric utility world is stepping up its investment in an array of technical developments in systems that have the ability to accurately capture, track and analyze outage information. The impetus is partly explained by the growing movement by commissions to regulate quality through Utility performance.
While useful, this emphasis belongs in the tactical arena of detective-corrective – a kind of closing of the barn door after the proverbial cows have been stolen.
Leading companies have focused their commitment to quality and energy reliability through the development of devices that significantly prevent transformer failures. It is appreciated now, that the average age of transformers in America is about 42 years and are approaching the end of their designed life expectancy. It is not overly dramatic to predict that all of these transformers will eventually fail, and sooner rather than later.
With on-line dissolved-gas monitoring and now, with automatically- generated analysis and interpretation of the monitoring data – there is a clearer way to quickly understand the health of transformers and to take action to prevent failures. One could describe this progress as “strategic- preventative” as opposed to “detective-corrective”.
Fortunately, in this case, doing the right thing and saving lots of money - happen together.
Power transformer failures increasingly result in tank-rupture, fires, extensive damage to other equipment and disastrous consequences including blackouts. The economics for on-line DGA monitoring deliver a compelling business case for utilities world-wide. With automated interpretation of the data, we continue to learn about the contributing factors that affect transformer health and how to detect and act on the information. The combination of automated interpretation tools and the monitoring that generates the underlying data should revolutionize the way utilities manage these critical transformer assets and deliver impressive ROI.
It has taken 2 years to monitor and gather data on the equivalent of 220,000 transformers.
Transformer users would have to carry out manual DGA on 220,000 transformers twice a year to equal the volume of transformer gas data that we gathered and analyzed in the past two years.
Becoming authorities, by virtue of the database we have grown and diagnostic tools employed, has enabled the drawing of correlations and trends between the DGA data and the transformer-events.
Standards and Guidelines; Making sense of them.
There are a number of Standards and Guidelines for diagnostic tools supporting DGA interpretation and these include the North American Standards setting body, the Institute of Electrical and Electronics Engineers (IEEE) who have issued Std C57.104-1991: IEEE Guide for the Interpretation of Gases Generated in Oil-Immersed Transformers” (approved June 27, 1991). IEEE PC57.104 Draft 11d: Draft Guide for the Interpretation of Gases in Oil-Immersed Transformers” (most recent draft is dated April 21, 2004; currently being balloted for approval).
Then there is the International Electro-technical Commission (IEC) – Geneva, Switzerland (IEC 60599-1999: The Interpretation of Gases in Transformer and Other Oil-filled Electrical Equipment in Service), and CIGRE of France.
The table at left represents the common knowledge about fault conditions and the dissolved gases that indicate those conditions. The table above shows dissolved Hydrogen as being common to most faults. The problem with Hydrogen, and our caveat here is, that Hydrogen is not really a diagnostic gas.
H2 is a very light molecule with extremely low solubility in the oil. There is little doubt that H2 due to any type of discharge rises in bubbles due to buoyancy & oil flow, collects in small gas pockets, and ultimately escapes through gaskets, welds & oil preservation system. H2 has little likelihood of showing up (especially at a drain valve) particularly with directed-flow as used by most large power transformers. H2 may have time to dissolve during a day-in/day-out thermal problem but the type of oil preservation system makes a huge difference as to how much is accumulated.
On-line DGA more & more shows such differences. In addition, H2 can be caused by a variety of conditions inside a transformer – rust, moisture in a galvanized pipe fitting (i.e. connected to a drain valve), and sunlight. Almost everything causes H2 in mineral oil. Hydrogen is not a diagnostic gas.
On-Line DGA data populating diagnostic tools delivers useful interpretation power and trending; the insight is in rate of change (ROC), not in snapshots.
An inherent difficulty for IEEE (& IEC) DGA interpretation guides is that they cover the entire range of transformer sizes, design types, oil quantity, oil preservation systems, operating conditions, maintenance history & age. Just about everything makes a difference.
Therefore, it is the Rate-of-change (ROC) that becomes more important than absolute values.
Results from periodic oil sampling from transformer drain valves are highly variable for many reasons & can be downright misleading. On-line monitoring of all the individual gases is discovering this every day.
The opportunity therefore is on-line diagnostic tools that deliver rates of change not only of individual gases, but of the ratios that provide enhanced (and actionable) information about the health of the transformers. Therefore, the next evolution is: automation of diagnostics with inputs from on-line DGA – a new capability.
Good examples of models that deliver actionable, automatically-generated interpretations and insights are the Duval triangle and Rogers’ ratios. Rogers’s ratios are recognized in the IEEE Standards & Guidelines and are equivalent to the “Basic Gas ratios” in the IEC.
Duval’s Triangle Model is recognized in the IEC Guidelines and this as been automated also. The CO2 vs. CO ratio is a third useful metric as an indicator of thermal decomposition of cellulose insulation with high potential for detecting emerging problems (IEC 60599-1999).
The relationships of gas concentrations can be displayed in 3-dimensional graphical representations on-line, together with the gas concentration data – to provide a diagnostic of root causes, and where necessary a call-to-action. (See Figure 1 and side bar of Rogers Ratio, to the left.)
Populating diagnostic models with on-line data enables new insights
Duval Triangle (IEC 60599 Appendix, developed by IREQ/Hydro Quebec) combined field service evidence with laboratory experiments published in 1989 followed by enhancements in 2002. This method uses the individual ppm of 3 gases – CH4, C2H4 & C2H2 – relative to the total ppm of those three gases, to locate a point within the triangle. Sections within the triangle designate: thermal fault < 300 C; thermal fault 300-700 C; thermal fault > 700 C; low-energy discharge; high energy discharge; and partial discharge.
In addition to the automation of the Rogers Ratios model (equivalent to Basic Gas ratios per IEC), we have also automated the Duval model. It is a more recent development that according to experts, delivers a useful and reliable interpretation of transformer health. Figure 2 shows a comparison of the Duval triangle and Rogers Ratio methods.
The take-away from this chart is that as one’s experience grows in the category, the greatest value lies not in the absolute values, but in the trending. Trending is the way to catch a problem.
How a multi-million dollar transformer was saved with on-line monitoring and automated diagnostics of the data – a real life example
Looking at the gases individually, each are in the “normal” PPM range. However, both Duval and Rogers ratios say it’s hot (hot metal). Later, (months), individual gases demonstrated that condition.
However on-line, automatically generated diagnostics delivered an earlier, actionable insight.
The utility was able to manage load, monitor diagnostic ratios, and nurse the unit until repair. The concentration of dots on the Duval triangle indicates close to a T3, but still in T2 range due to load management and monitoring.
It was the combination of DGA monitoring data and the diagnostic tools that provided the insight to a problem that would not have been obvious otherwise.
The cause was a defective brazing, an excellent example of the value of the automation of high-level diagnostic capability – to manage transformer performance and reduce the risk of unplanned failure.
Summary
Reliable, affordable, smart technologies exist today that can manage critical utility assets. Given the criticality of the transformers fleets, it makes sense to begin thinking seriously about wide programs of transformer DGA monitoring with the use of on-line automatically generated diagnostic functionality. It makes good sense and it makes for good governance.
Automatically-generated and actionable, Dissolved-Gas-Analysis (DGA) diagnostics for transformers.
Paradigm shift: takes the industry from a detective-corrective mode to a strategic-preventative mode.
These days, the electric utility world is stepping up its investment in an array of technical developments in systems that have the ability to accurately capture, track and analyze outage information. The impetus is partly explained by the growing movement by commissions to regulate quality through Utility performance.
While useful, this emphasis belongs in the tactical arena of detective-corrective – a kind of closing of the barn door after the proverbial cows have been stolen.
Leading companies have focused their commitment to quality and energy reliability through the development of devices that significantly prevent transformer failures. It is appreciated now, that the average age of transformers in America is about 42 years and are approaching the end of their designed life expectancy. It is not overly dramatic to predict that all of these transformers will eventually fail, and sooner rather than later.
With on-line dissolved-gas monitoring and now, with automatically- generated analysis and interpretation of the monitoring data – there is a clearer way to quickly understand the health of transformers and to take action to prevent failures. One could describe this progress as “strategic- preventative” as opposed to “detective-corrective”.
Fortunately, in this case, doing the right thing and saving lots of money - happen together.
Power transformer failures increasingly result in tank-rupture, fires, extensive damage to other equipment and disastrous consequences including blackouts. The economics for on-line DGA monitoring deliver a compelling business case for utilities world-wide. With automated interpretation of the data, we continue to learn about the contributing factors that affect transformer health and how to detect and act on the information. The combination of automated interpretation tools and the monitoring that generates the underlying data should revolutionize the way utilities manage these critical transformer assets and deliver impressive ROI.
It has taken 2 years to monitor and gather data on the equivalent of 220,000 transformers.
Transformer users would have to carry out manual DGA on 220,000 transformers twice a year to equal the volume of transformer gas data that we gathered and analyzed in the past two years.
Becoming authorities, by virtue of the database we have grown and diagnostic tools employed, has enabled the drawing of correlations and trends between the DGA data and the transformer-events.
Standards and Guidelines; Making sense of them.
There are a number of Standards and Guidelines for diagnostic tools supporting DGA interpretation and these include the North American Standards setting body, the Institute of Electrical and Electronics Engineers (IEEE) who have issued Std C57.104-1991: IEEE Guide for the Interpretation of Gases Generated in Oil-Immersed Transformers” (approved June 27, 1991). IEEE PC57.104 Draft 11d: Draft Guide for the Interpretation of Gases in Oil-Immersed Transformers” (most recent draft is dated April 21, 2004; currently being balloted for approval).
Then there is the International Electro-technical Commission (IEC) – Geneva, Switzerland (IEC 60599-1999: The Interpretation of Gases in Transformer and Other Oil-filled Electrical Equipment in Service), and CIGRE of France.
The table at left represents the common knowledge about fault conditions and the dissolved gases that indicate those conditions. The table above shows dissolved Hydrogen as being common to most faults. The problem with Hydrogen, and our caveat here is, that Hydrogen is not really a diagnostic gas.
H2 is a very light molecule with extremely low solubility in the oil. There is little doubt that H2 due to any type of discharge rises in bubbles due to buoyancy & oil flow, collects in small gas pockets, and ultimately escapes through gaskets, welds & oil preservation system. H2 has little likelihood of showing up (especially at a drain valve) particularly with directed-flow as used by most large power transformers. H2 may have time to dissolve during a day-in/day-out thermal problem but the type of oil preservation system makes a huge difference as to how much is accumulated.
On-line DGA more & more shows such differences. In addition, H2 can be caused by a variety of conditions inside a transformer – rust, moisture in a galvanized pipe fitting (i.e. connected to a drain valve), and sunlight. Almost everything causes H2 in mineral oil. Hydrogen is not a diagnostic gas.
On-Line DGA data populating diagnostic tools delivers useful interpretation power and trending; the insight is in rate of change (ROC), not in snapshots.
An inherent difficulty for IEEE (& IEC) DGA interpretation guides is that they cover the entire range of transformer sizes, design types, oil quantity, oil preservation systems, operating conditions, maintenance history & age. Just about everything makes a difference.
Therefore, it is the Rate-of-change (ROC) that becomes more important than absolute values.
Results from periodic oil sampling from transformer drain valves are highly variable for many reasons & can be downright misleading. On-line monitoring of all the individual gases is discovering this every day.
The opportunity therefore is on-line diagnostic tools that deliver rates of change not only of individual gases, but of the ratios that provide enhanced (and actionable) information about the health of the transformers. Therefore, the next evolution is: automation of diagnostics with inputs from on-line DGA – a new capability.
Good examples of models that deliver actionable, automatically-generated interpretations and insights are the Duval triangle and Rogers’ ratios. Rogers’s ratios are recognized in the IEEE Standards & Guidelines and are equivalent to the “Basic Gas ratios” in the IEC.
Duval’s Triangle Model is recognized in the IEC Guidelines and this as been automated also. The CO2 vs. CO ratio is a third useful metric as an indicator of thermal decomposition of cellulose insulation with high potential for detecting emerging problems (IEC 60599-1999).
The relationships of gas concentrations can be displayed in 3-dimensional graphical representations on-line, together with the gas concentration data – to provide a diagnostic of root causes, and where necessary a call-to-action. (See Figure 1 and side bar of Rogers Ratio, to the left.)
Populating diagnostic models with on-line data enables new insights
Duval Triangle (IEC 60599 Appendix, developed by IREQ/Hydro Quebec) combined field service evidence with laboratory experiments published in 1989 followed by enhancements in 2002. This method uses the individual ppm of 3 gases – CH4, C2H4 & C2H2 – relative to the total ppm of those three gases, to locate a point within the triangle. Sections within the triangle designate: thermal fault < 300 C; thermal fault 300-700 C; thermal fault > 700 C; low-energy discharge; high energy discharge; and partial discharge.
In addition to the automation of the Rogers Ratios model (equivalent to Basic Gas ratios per IEC), we have also automated the Duval model. It is a more recent development that according to experts, delivers a useful and reliable interpretation of transformer health. Figure 2 shows a comparison of the Duval triangle and Rogers Ratio methods.
The take-away from this chart is that as one’s experience grows in the category, the greatest value lies not in the absolute values, but in the trending. Trending is the way to catch a problem.
How a multi-million dollar transformer was saved with on-line monitoring and automated diagnostics of the data – a real life example
Looking at the gases individually, each are in the “normal” PPM range. However, both Duval and Rogers ratios say it’s hot (hot metal). Later, (months), individual gases demonstrated that condition.
However on-line, automatically generated diagnostics delivered an earlier, actionable insight.
The utility was able to manage load, monitor diagnostic ratios, and nurse the unit until repair. The concentration of dots on the Duval triangle indicates close to a T3, but still in T2 range due to load management and monitoring.
It was the combination of DGA monitoring data and the diagnostic tools that provided the insight to a problem that would not have been obvious otherwise.
The cause was a defective brazing, an excellent example of the value of the automation of high-level diagnostic capability – to manage transformer performance and reduce the risk of unplanned failure.
Summary
Reliable, affordable, smart technologies exist today that can manage critical utility assets. Given the criticality of the transformers fleets, it makes sense to begin thinking seriously about wide programs of transformer DGA monitoring with the use of on-line automatically generated diagnostic functionality. It makes good sense and it makes for good governance.
