Power Transformer Health Concepts

All matters affecting the health of a power transformer are discussed in this section.

Dissolved Gas Analysis

Use this free analysis template to analyse your dissolved gas data to establish general condition.

LEDT - Low Energy Degradation Triangle

The LEDT is a contemporary method to assess the condition of the transformer based on low energy incipient faults. It provides early detetcion of incipient faults which start off from low energy insulation degradation.

Case Studies

Review cases studies of power transformer failures.

Thursday, June 8, 2017

Case Study: Investigation into Combustible Gases in Selector of Tap Changer

The following report highlights the problem of the production of combustible gases in the selector of the tap changer of gen/motor transformer 1 at A Pumped Storage Scheme.  This may be caused either by abnormal arcing or a leak between the diverter and selector compartments.


From the oil results in 2002 it was noticed that there was an increase in the level of combustible gases in the selector tank of gen/motor transformer 1. This gas level was monitored and a gradual increasing trend is noticed. However the gases fluctuate around a level that is above limits. Last year (September 2002) it was proposed that an internal inspection be carried out in the selector to identify a possible cause for the high gassing. Due to production reasons and an outage not being available it was decided to rather monitor the transformer with regular oil samples (weekly to three weekly) until the outage in April 2003 where an internal inspection of the selector and diverter compartment is carried out.


The first possible cause highlighted has to do with the operating limits of the unit linked to the range of tap changing. The tap changer is operated at least twice a day across the change over tap 9 which would result in a small arc that could be the source of the combustible gases. Comparing the levels of gases on all the other units, Unit 1 is at an elevated level suggesting an area of higher levels of arcing. If this is the case we need to as soon as possible identify and sort the problem out before it develops into an electrical fault resulting in the failure of the transformer.

The second possible cause is that there might be a possible leak between the diverter tank and selector. The diverter because of its operation will produce high levels of acetylene and combustible gases. If there is a leak between these compartments there will be a migration of the gases into the selector. 


Inspection of the U2 tap changer at the Power Station was carried out due to suspected faults as a result of high combustible gases in the selector tank. The following is an explanation by the OEM of the cause. It must however be stressed that the level of the U1 gases are far above that experienced on U2.


The dissolved gas analysis of oil samples taken from the selectors of the generator transformers indicated hydrogen and acetylene values higher than normal. This initiated an internal inspection of the selector switch on Unit 2 in February 1999.

This inspection revealed some evidence of sparking on the changeover contacts of the selector. The operations of these transformers are somewhat different from the other generator transformers in the  network. The pumped storage operation requires generating and pumping modes using the same transformer. This then gives rise to frequent tap changer operations and also a different tapping range. Unlike other generating stations the tapping range includes the changeover position on a daily basis.
During the changeover operation the polarity of the tap winding is reversed. To allow this to happen the tap winding is disconnected electrically. When the reconnection takes place even when it is not under load, the capacitive coupling to the other windings on the same phase induces some voltage in this winding. The contact to which it then connects is at a different voltage and then a minor arc occurs when this contact makes. This only happens to the changeover contact and it happens in both directions – tapping up and down. It is better explained when referring to the rating plate diagram. 

Figure 1:

The changeover occurs on position 9.  9A and 9B are the transitional positions but have the same output voltage as position 9. This is where the polarity of the tapping winding is reversed. Consider the operation from tap position 8 to 10. The taps occur from 8 to 9A. Then to 10 via 9 and 9B without stopping on 9 or 9B.  At the point of 9 the tap winding is completely disconnected and then connected in the reverse polarity on position number 9B. When going back from position 10 back to 8 the actual arc takes place between position 9 and 9A.


After it was discovered that there is an increasing trend of combustible gases especially acetylene in the selector tank, special DGA sampling at weekly intervals were established. This was then increased to every three weeks when the level stabilised. While this was monitored the tapping profile was also recorded for 17/10/2002 to 4/11/2002. This information is discussed below. The next long outage on this unit was only planned in April 2003 and it was agreed that the tap changer be closely monitored until the outage or if there were further increases in the gas level to warrant an emergency outage. The oil results of the selector may be found in appendix 1 and the tapping profile data may be found in appendix 2. Figure 2 below clearly portrays the tapping profile of gen/motor transformer 1. As was explained in section 4 above the taps move across taps 8-10 crossing the transition tap 9 on most occasions thus producing a small amount of arcing which results in the production of combustible gases. 

Figure 2:


Figure 3 illustrates the number of taps that are changed per day for the period 17/10/02 to 4/11/02. The average number of tap changes for one day is estimated at 8 taps per day. This is considered to be frequent for a generator/motor transformer but not unusual considering its application.

Figure 3:

Figure 4 below gives an indication of the number of taps per mode. From this profile it is evident that most taps are during generating and SCO modes.

Figure 4:

TEMPERATURE & HYDRAN

Weekly oil and winding temperature readings taken reveal that gen/motor transformer 1 is also relatively and consistently warmer that its other three counterparts. This increase has been noticed from just six months now. The Hydran reading has also increased a bit to 162 ppm were it was previously sitting at mid 130’s. 

OUTAGE – APRIL 2003

During the outage in April it is proposed that an internal inspection be carried out to identify the cause of the high gassing. The monitoring of the gases over the past three months gives no indication of a decreasing trend. The risk of a possible failure would thus be considered high. If the internal inspection is not carried out in the April outage Peaking will have to live with the knowledge of a possible failure.

The oil samples taken for the past three months was taken to keep track of further increases so as pick up any problems as quick as possible. The hydran on the transformer is only monitoring the main tank which is isolated from the selector by a shut off valve. This further emphasises the need for this inspection. 

CONCLUSION

From the report above it is clear that there is an arcing problem in the selector of gen/motor transformer 1. We however need to establish the criticality and source of this arcing. The possibility of a leak between the selector and diverter must also be investigated so that plans can be put in place for the replacement of the diverter barrel if such leak exists.
It must also be noted that if the recommendations made below are followed it must be coordinated with the conservator bag project which is scheduled in the April 2003 outage due to interfaces on the handling of oil. 

RECOMMENDATIONS

The recommendations below can be implemented in the April 2003 outage.

1.  Inspect tap changer (selector & diverter) tank for source of combustible gases.
2.  Test diverter and selector to identify if there is a leak between selector and diverter tanks.
3.  If tap changer is scheduled for 18 month service, carry out the service.

Wednesday, June 7, 2017

Rogers Ratio Method

The Rogers Ratio method is another ratio method very similar to the Dornenburg method however only having the three ratios R1:CH4/H2R2:C2H2/C2H4, R5:C2H4/C2H6. IEEE standard C57.104 has the following flowchart to explain the process.


Figure 1




























Figure 2 below also taken from IEEE standard C57.104 provides a tabular list of the different conditions of the transformer based on the three ratios.




Figure 2























Use the following link to the "Analysis" section to get the Rogers Ratio diagnosis of the oil samples. Enter the oil sample under "Sample 5" to get the diagnosis.



Tuesday, June 6, 2017

An Effective Asset Management Strategy for Power Transformer Health

A major challenge to transformer engineers of power utilities and major industries is the management of an ageing transformer fleet. Predicting transformer failure is as unpredictable as that of predicting share prices in the market. This makes it difficult to effectively plan mitigation strategies to reduce the impact of such failures. The other challenge is replacing assets too soon without getting the maximum benefit out of its useful life.
The following paper defines an asset management strategy for Power Utilities on how to manage power transformers. This becomes extremely important, especially for the effective management of an ageing fleet of transformers. Failure of critical transformers can leave a utility crippled with resultant production costs up to ten times the capital cost of equipment. The aim of this paper is to clearly map out a process to establish a risk profile of the transformer fleet and a strategy on how to mitigate such risk by incorporating effective condition monitoring techniques, design and investment strategies for the utility.

Power transformers are one of the key components of a power circuit, especially for power utilities and heavy industries. The high capital costs involved in the purchase of a transformer means that for the best investment the replacement should be differed as late as possible or to the optimum point.

Do you ask the following questions?
  • How is my transformer fleet doing?
  • Do I need to plan for refurbishment?
  • When will a transformer fail?
  • What is the risk of failure?
  • Which are my high risk transformers and how does one mitigate the risk?
  • How does one determine the optimum replacement point of a transformer? 
Managing your transformer fleet properly and knowing the condition of your transformers will help immensely in getting to the answers of the above questions. It is of utmost importance to have a structured approach to solving these problems. Time and human resources are limited and every transformer cannot get the same attention. It is up to the transformer engineer to formulate a system to effectively identify problem transformers for focussed attention.


 When one starts out with a large transformer fleet it can be an overwhelming task to identify which transformers need special attention or replacement. A transformer fleet at a power utility serve many applications and have varying levels of importance. The following factors play and important role in focussing the attention. These are rating, voltage levels, importance, redundancy, spares, and similar type in fleet. The first task of a transformer engineer is to record as much information on the transformer fleet as possible. Start with recording all details on the name plate of the transformer. This information must be stored in a database and forms the basis for analysis.

For , starting from a transformer fleet of 70 transformers, Figure 1 below gives an immediate picture of where most of the transformers are. The key transformers are the higher rated Generator Step-Up (GSU) and Station transformers. 

Figure 1

















Age Profile
Transformers no matter how well they run have a limited design life. Establish an age profile of the transformer fleet. Record the year of manufacture and take this as the reference age. Calculate the present age of the transformers and plot on a graph to give a quick picture of the age profile for the current transformer fleet.
Figure 2 















Profile the Transformer Importance
It is very important to establish what percentage of the transformer fleet are critical transformers and the impact of these transformers. Establish how many of the same type of transformers are available and number of spares. Of importance is the MVA rating, low and high voltage levels and percentage impedance. Focus must be given to the high impact transformers. For Power Utilities, these are usually GSU and station transformers and to some extent unit transformers. 


The basic approach to condition based assessment is to use the oil as the primary indicator of the internal condition of the transformer. Condition based assessment allows the transformer engineer to identify what is happening inside the transformer without the risk of going into the transformer. Insulation life is a function of thermal ageing of the system. It is reasonable to expect that the transformer insulation systems will last the expected 30-40 years at full load if the system is maintained in a good condition. Failure or end of life occurs as the result of a variety of mechanisms including poor oil preservation systems. Thus, it is necessary to examine the causes of deterioration of the paper and oil properties. This is based on dissolved gas analysis (DGA) and electrical testing.

Young identified the following as being some of the monitoring offered by the industry; winding and oil temperatures, internal partial discharge, oil moisture content, on-line DGA, tap changer monitoring, oil and air flow, external hotspots by infra red photography, pump and fan motor bearing wear [1].

Ranking the transformer according to health criteria would assist in refining the focus transformers. Transformer ageing is caused primarily from three effects. These are hydrolysis, pyrolysis and oxidation [2, 3]. It follows from here that the levels of moisture, temperature and oxygen must be controlled to within limits.

Studies on Furan compounds have provided some correlation to the degree of polymerisation and the mechanical strength of the paper insulation [4-6]. Due to paper insulation having a direct correlation to the life of a transformer, Furan results could give a rough idea of the condition of the paper without having to get a paper sample. This crude method can be used as a high level guide on the paper insulation ageing of the transformer. This method is however limited by the replacement or processing of the oil.

On can establish a rating system based on the DGA, condition of the oil and Furan levels. Dissolved gas analysis has been extensively utilised for incipient fault detection by the following methods; key gas analysis [7], the Dornenburg [8], Duval [9, 10],Rogers [11] gas ratio method and those highlighted in the ANSI/IEEE standard [12]. Use limits specific for similar transformers as each transformer would exhibit different gassing profiles according to design and materials used. The limits can be based on hydrocarbon gas limits and furan levels.
The total number of short-circuit and voltage surge events also contribute to the health of the transformer. Although a transformer might have been designed and tested to withstand mechanical stress resulting from external faults, the latter must nevertheless be considered as an ageing factor. The clamping force of many transformers is reduced over time due to the shrinking of insulating material. Historical data of incidents are very important and every time an incident occurs these must be recorded and the assessment of the transformer revisited. 

Power utility transformer fleets have remained relatively reliable over the last decade resulting in few alarms and focus by the engineer, however it has become evident that most utilities are ending up with an ageing transformer fleet. Compounded by this is the increase in demand for electricity which results is assets being operated more at and beyond its capabilities. This has a dual effect as assets age faster and causes the probability of failure to increase.
With the network being further strained the probability of incidents are increasing causing the ageing transformer fleet to be exposed to high number of through fault conditions. Through faults are infamous for exposing ageing and weak insulation which results in instantaneous failure of the transformer. These failures are difficult to pick up and remain a high risk for the transformer. It is thus quintessential to make sure that the transformer electrical protection is reviewed and settings appropriately set to eradicate fault conditions as soon as possible.  


Lapworth and Mcgrail have highlighted the following transformer asset management strategies; replace on age, replace on failure and replace on condition [13]. The replace on age offers a low risk option but this option is capital intensive and does not allow for full utilisation of the asset. Replace on failure on the other hand makes full use of the asset but failures are at in-opportune times with risk of damage to neighbouring plant, forced outages with related penalties, and compromise of safety of personnel. The third strategy is to replace the transformer when its reliability to satisfy system requirements cannot be met anymore. This also enables for a planned replacement which allows adequate time for quality reviews and manufacture and better economic management.


Figure 3 highlights a typical life cycle of a power transformer with high seven steps.

Figure 3  
















Step 1: Specify and Manufacture
This step is where the detailed specification for the transformer is specified together with the manufacture and transport of the transformer to the site. This is very important step as it defines the operating conditions of the transformer and it is where one has control of the internal structure of the transformer. A bad design can mean a short lifespan of the transformer. It is highly recommended that an intense design review is conducted as a hold point. Transport is also a high risk process and proper quality plans and checks must be in place to manage transport over sea and road.
Step 2: Install & Commission
This step introduces many risks especially on the interface points of the transformer. These must be clearly defined and comprehensive detail provided. If the contractor is responsible for the installation of the transformer, all scope of work must be clearly defined. The project program must include all activities and proper records must be provided for every step in the process.
Step 3: Operate

This step represents the flat part of the bath tub curve. It is where the transformer operates without any major problems. Normal condition monitoring must be carried out during this phase. This includes six monthly dissolved gas sampling and analysis to identify incipient faults. This can also be supplemented by on-line DGA monitoring. The oil condition must also be monitored by routine sampling for dielectric strength, moisture and acidity. Paper condition can be monitored crudely by taking yearly Furan sampling which can be trended to establish breakdown rates. The furan levels are affected by oil processing and oil replacement which makes this method is ineffective if carried out. Routine thermal scanning of the transformer also helps in establishing localised hotspots and problems with the cooling system. This can be done three monthly.

Step 4: Inspect & Maintain
This step covers the normal routine inspection and maintenance of the transformer. Routine inspections must be carried monthly especially by operating personnel. This involves visual checks on the general condition, recording of oil and winding maximum temperatures (reset max dial), tap changer readings and general components like cooling and protection.
Maintenance also plays an important role in ensuring the good operating condition of the transformer. This includes maintenance of the tap changer, cooling fans and pumps, tan delta testing of bushings.  
Step 5: Repair
Small repairs are necessary to maintain the transformer life. This includes repairs to oil leaks, tap changer, fan and pump motors and rust with repainting of the tank.
Step 6: De-commission

Decommissioning of the transformer takes place when it is replaced or when it has faulted and cannot be reused. This includes for the proper dismantling and storage in a temporary storage area. Arrangements must be made for the removal of the oil. This especially becomes a challenge when the Utility undergoes a comprehensive replacement strategy which results in hundreds of thousands of litres of oil.  

Step 7: Dispose

This phase of the life cycle covers the proper removal and disposal of the transformer. Before the transformer can be scrapped a post mortem must be carried out to gain valuable experience on the mechanism of failure and weak areas. This is particularly important if there are similar transformers within the fleet. The teardown can be done according to “IEEE Guide for Failure Investigation, Documentation, and Analysis for Power Transformers and Shunt Reactors”. This guideline provides a procedure to perform a failure analysis and is primarily focused on power transformers used on electric utility systems where it encourages the establishment of routine and uniform data collection procedures during the failure analysis process [14].

Phases A & B
Phase A of the life cycle is composed of steps 3, 4 and 5 which consist of the longer term process of the transformer lifecycle. It is noted that over the past years due to the reliability of transformers being relatively good most of the focus had been on this phase of the lifecycle. This phase is the normal operating of the asset, maintenance with timely inspections and minor repairs to improve reliability and availability of the asset.
Phase B however comes into play when there is a need to purchase and replace a transformer either due to failure or planned replacement. This phase is made up of steps 1, 2, 6 and 7. This phase can take up to two years from compiling the specification, issuing and evaluating the tender enquiry, design, manufacture, factory acceptance, decommissioning of the old transformer, installation and commissioning of the new transformer and finally disposal of the old transformer.
Of recent years due to most of the assets reaching end of life and premature failures transformer engineers have been focusing more on procurement, design, commissioning, decommissioning and disposal of assets. Due to these activities not being everyday activities the expertise for such work is limited which result in some delays and longer down times on the plant. It is thus of utmost most importance to have a structured approach and to document and share experiences within the organisation on typical learning’s to improve the overall quality of the asset management process.
An important part of the asset management process it is to have a clear picture of the longer term plan for each transformer. This map provides all time frames covering present life, remaining life, life extension possibilities, system health life and major maintenance activities. These are defined in more detail below. This plan can then be used for the life cycle costing and financing of the projects.
Station Life
Start the process by first establishing the remaining life of the Utility, Power Station or specific unit. This sets up the scene for the future decision making. Plot this in years starting from the present year. In figure 4 an example is made for Gariep Power Station with 2007 as the base year.
Design Life
Record the design life of the transformer. This is usually taken according to history or trends in the transformer fleet, industry experience or experience of similar type of transformers in the world. For the power station transformers this was taken as 30 years. The present age of the transformer is then subtracted from this age to estimate a crude remaining life of the transformer. In the case of GSU transformers 1 and 2 this was -4 years which means that the transformer should have already been replaced. This method however does not take into account the transformer condition and only provides a guide.  The next step is to estimate the System Analysis Life

Figure 4















System Analysis Life

After carrying out condition monitoring on the transformer the transformer health is established and the remaining life is refined to establish a more accurate estimate. From the above example it is estimated that GSU transformers 1 and 2 have a remaining life of 2 and 3 years respectively after 2007. This step brings the influence of the actual condition of the transformer into consideration and provides a more accurate estimate.

Major Maintenance and Testing

It is advantageous to include all major maintenance activities on the plan so that it can be budgeted and planned for in the long term. In this example consideration is made for oil filtering, oil replacement and tan delta testing of the bushings. 


Life Extension Possibilities
After assessing the profile and replacement strategy it is prudent to explore for life extension possibilities. There are no rules to the criteria for selecting a life extension strategy but one can assess this by evaluating the remaining life and present condition around 5-10 years before the end of station life. It is important however to maintain a healthy condition of the transformer from inception. Key life extension methods would entail timely maintenance especially of the on load tap changer and bushings, reduction of oxygen and moisture levels in the transformer, removal of sludge and reduction of acidity levels, replacement of oil depending on condition, reduction of hotspot temperatures by reducing frequent overloading and through fault conditions.

Refurbishment/Replacement
This decision must be planned at least two years before the estimated end of life of the transformer. The decision for replacement or refurbishment becomes a business decision and takes into account the present condition of the tank and core, return on investment, quality of rewind companies within the country, transport costs and potential for up rating. If transformers have aged prematurely with inherent design defects it is possible to consider a redesign of the windings. This option can be more cost effective than purchasing a new transformer.


Condition monitoring techniques form an important part of the asset management strategy.  The trend is moving towards on-line methods that are non intrusive. The advantage of this is that information on the transformer condition is continuous and allows for early detection of incipient faults so that planned mitigation and recovery strategies can be put in place.

Such methods consist of on-line DGA which consist of equipment measuring up to eight dissolved gasses. These trends can be obtained either as 4-20mA outputs or via communications to be linked directly onto the station SCADA system for easy access.

On-line measurement of oil and winding temperature consisting of 4-20mA analogue output are now becoming the standard. These are also transmitted to the station SCADA system. Transformer manufacturers are now starting to adopt fibre optic sensors to get direct hotspot temperatures of the windings. These also have a 4-20mA output for easy connection to data acquisition systems.  

Online moisture measurement is also fast becoming part of the condition monitoring tool. These are usually being added as supplementary to the on-line DGA equipment. These devices help to access the flow of moisture from the oil to paper to oil and establish cyclical load dependence.

Centralised data acquisition systems are the next phase to condition monitoring. It is becoming more important for manufacturers of condition monitoring equipment to focus not only on the technology of sensing key data but how this data can be effectively transmitted and analysed. International standards like IEC 61850 are assisting in the convergence of data processing and should be clearly highlighted when specifying such equipment. It is important for all this information to feed into one place where some intelligence can be built in to form automatic analysis of data. This provides the transformer engineer with more meaningful focussed information and saves time by alerting to abnormal conditions.


Over the last seven years the Power Station has been involved in numerous transformer projects. This consisted of 3 GSU transformer rewinds and five GSU transformer purchases. The replacement program has offered numerous opportunities for learning due to the tight project schedules and availability of plant. 

The most important part of the replacement program is to have a clear picture of what transformers need replacement. It is important to get all stakeholders involved at the start of the project. These include plant operators, maintenance and production engineers and related plant engineers. The specification and assessment of tenders is the start of a very important process in the future asset management of transformers. It is critical that a proper design review is undertaken by experienced personnel. The customer must also insist on a Factory Acceptance Test (FAT) of the transformer and ensure that the international acceptance standard IEC76 is clearly complied with.

Transformer interfaces are the highest risk areas. Careful attention needs to be paid to this aspect of the design. These interfaces consist of low and high voltage bushings connections to the plant and system busbars, cooling water interface, electrical and control interfaces to the plant, fire protection systems. Interface drawings play a key role to a successful transformer installation. It is thus important to have as built drawings when issuing a tender for a new transformer. First step is to measure and survey the site with the supplier to make sure that all interface points are clearly noted and planned for.

The logistics of a transformer replacement program must be taken into account in the overall project plan. Key aspects like oil disposal if not properly planned can eat into valuable outage time resulting in delays of the overall project schedule.    

Over the recent years a power Utility was heavily exposed to premature failures of transformers and it was evident that a reassessment was necessary in respect of the spares philosophy of critical Generator Step-up (GSU) transformers. It is recommended that an assessment is carried out on the transformer fleet to identify transformers that can be inter changed with minor modifications and holding spare transformers between such stations or transformers. 
Figure 5
The landscape is constantly changing and focus on the transformer fleet must be continuous. Figure 5 provides a picture of how the Asset profile for the GSU transformer changes in a span of ten years from 2002 to 2012.

Graph A provides the age profile at the year 2002. There were 18 GSU transformers including 2 spare transformers. There were 2 GSU transformers in the 30-40 age groups which fall under the high risk category and 14 in the 20-30 categories which are medium to high risk.

From the current transformer replacement program which started in 2003 the profile has changed to that represented in graph B. There were 2 rewinds and 3 transformer purchases with installations up to 2007. This mitigated the risk of the 2 premature aged Palmiet GSU transformers and 4 transformers from Gariep, Vanderkloof and Drakensberg Power Stations.    
Graph C provides a snapshot of the profile of the  GSU transformers in year 2012. Most of the high risk GSU transformers would have been replaced by new or rewound transformers. The 7 GSU transformers in the 30-40 age groups are the Acacia and Port Rex Gas Fired GSU transformers. Although these transformers are in this age category they are still in a satisfactory health condition and are not highly loaded with a low Station load Factor. There is also a spare transformer to reduce the impact of a transformer failure.

The following paper attempts to provide a picture of the nature of a typical transformer life cycle and the components that go into forming a transformer asset management strategy. There are numerous ways of achieving this and the author attempts to provide a structured approach for asset management decision making. Transformers are a huge investment for a Utility so it is important to balance the decision to defer the replacement/refurbishment to as late as possible and that of preventing a catastrophic failure.
The most important aspect of asset management is to know as much as you can about your assets. Then formulate techniques on achieving a focussed approach for identifying abnormal conditions and potential transformers failures.


  1. Young, W., Transformer Life Management - Condition Monitoring. 1998, The Institution of Electrical Engineers: Savoy Place, London.
  2. Emsley, A.M. and G.C. Stevens, Review of Chemical indicators of degration of cellulosic electric paper insulation in oil-filled transformers. IEE Proc.-Sci. Meas. Technology, 1994. 141(5): p. 324-334.
  3. Wang, M., A.J. Vandermaar, and K.D. Srivastava, Review of Condition assessment of Power Transformers in Service. IEEE Electrical Insulation Magazine, 2002. 18(6): p. 12-25.
  4. Oomen, T.V. and L.N. Arnold. Cellulose Insulation Materials Evaluated by Degree of Polymerization Measurements. in IEEE Proc. 15th Electrical/Electronics Insulation Conference. 1981. Chicago, IL, USA.
  5. Shroff, D.H. and A.W. Stannett, Review of paper ageing in power transformers. IEE Proc. C, 1985. 132(6): p. 312-319.
  6. Allan, D., C. Jones, and B. Sharp. Studies of the Condition of Insulation in Aged Power Transformers. 1. Insulation Condition and Remnant Life Assessments for In-service Units,. in IEEE Proc. 3rd International Conference Properties and Appl. Dielectric Materials. 1991.
  7. IEC Publication 599, Interpretation of the analysis of gases in transformers and other oil-filled electrical equipment in service. IEC Publication, 1978.
  8. Dornenburg, E. and W. Stittmater, Monitoring oil cooling transformers by gas analysis, in Brown Boveri Rev. 1974. p. 238-274.
  9. Duval, M., A Review of Faults Detectable by Gas-in-oil Analysis in Transformers. IEEE Electrical Insulation Magazine, 2002. 18(3): p. 8-17.
  10. Duval, M. and J. Dukarm, Improving the reliability of transformer gas-in-oil diagnosis. IEEE Electrical Insulation Magazine, 2005. 21(4): p. 21-27.
  11. Rogers, R.R., IEEE and IEC codes to interpret incipient faults in transformers using gas in oil analysis. IEEE Trans., Electrical Insulation, 1978. 13(5): p. 349-354.
  12. ANSI/IEEE std C57.104-1991, IEEE guide for the interpretation of gases generated in oil-immersed transformers,. IEEE Power Engineering Society, 1992.
  13. Lapworth, J. and T. Mcgrail, Transformer failure modes and planned replacement.
  14. C57.125, I., IEEE Guide for Failure Investigation, Documentation, and Analysis for Power Transformers and Shunt Reactors. 2005, The Institute of Electric and Electronic Engineers.