Abstract

HV Bushings, albeit transformer or CT-mounted, account for one of the most significant single causes of failure in HV substations. Worse still, their failure mechanisms tend to develop to a critical level at a mid-life point for the surrounding assets and such mechanisms generally result in a sudden and catastrophic failure of an explosive nature. As such, HV bushings deservedly merit a close observation from before their mid life point. Technologies now exist in affordable, reliable, and flexible configurations to permit asset owners to detect impending failures in a timely manner. With a major proportion of the world’s monitored bushings being in Australia, a new benchmark now exists in the region for the ‘best practice’ stewardship of this asset class.

1 Introduction … Quantifying the Problem

Whilst merited, one of the dangers in focussing primarily on bushings in this paper is that they seldom ‘act alone’ in their functionality and failure consequences in a substation. Nor, in some cases, might their failure modes be unrelated to outside influences or to the asset to which they are mounted. Accordingly, a wider perspective will be introduced in Sections 3 and 5 but the reader must at all times be cognisant of the powerful driver toward improving total substation reliability that bushings represent.
The literature agrees in general that problems with bushings are a major source of transformer failures. One study [1] of some 106 aged (23-39 years, 110-500 kV) transformers noted bushing defects in 70% of the transformers surveyed (or just over 14% of the sampled bushing population). Indeed, the same survey drew the conclusion that transformer life is limited by the deterioration of accessories, of which were listed just three: bushings, LTC, and the cooling system. Over 46% of transformer defects were found to be attributable to these items [1,24]. Another study [11] from the 63 members of the Edison Electric Institute essentially concurs, it listing bushings amongst the three most common transformer failure modes reported.
Australasian reliability statistics [23] on 2096 transformers over 1970-1995 found similarly, concluding that bushings were second only to tap changers as the component initially involved in failure and were amongst the top three contributors to costly transformer failures.
Cigre’s WG2.18 collected and tabulated typical failure modes of large power transformers ( from at least a 5000 unit population database ) and concluded [2] that most failures occur due to abnormal change in equipment condition over life. Shortened life due to accelerated deterioration of bushings and OLTC were highlighted. Park [14] and Krieg [15] concur, Krieg citing HV CT failure rates of one major failure for every 1000 unit years, and observing that bushings and HV CTs are in effect the key to HV substation reliability.
More alarming still is reported findings from European statistics [3] indicating that bushing failures not only initiate 30% of transformer failures [3,24,27,30] but that 80% of bushing failures occur between 12-20 years, or mid life of the transformer population. Another source [30] agrees that early failure scenario dominates, citing 80% of bushing failures occur in the 10-12 year area and only 30% after 20-25 years, but reasons that an aged failure mode still predominates. Sokolov [24] cites recently reported information published by Turley [28,29] to endorse the latter observation, commenting that the occurrence of PD and paper overheating failures in numerous bushings in the 2-15 year period “…confirms urgency for appropriate studies”.
Certainly, the concerns have not gone unheeded with Cigre establishing a WG ‘Bushings Reliability’ in 2004 with the aim ‘to improve the bushing reliability or at least prevent the decrease of bushing performance (trend due to economic pressure)’ [24].
What is also universally clear in the literature [2,3,4,5,6,7,8,9,10,17,19] is that a bushing or HV CT failure is not untypically followed by a catastrophic event such as tank rupture, violent explosion of the bushing (propelling large broken shards of porcelain some hundreds of metres at velocities enough to imbed the material in concrete walls), and fire (Figure 1). Clearly, the risk and likelihood of collateral and personnel damage is a major concern in such an eventuality.
Following a catastrophic failure, it would be an understatement to observe that the damage bill may well be extensive. This aspect will be reviewed further in Section 5.

Explosive failure: Failed 275 kV HV CT clearly showing degree of porcelain fracture, oil spillage, and burned core element papers Fire: Tank Rupture: Figure 1: Examples of classic outcomes from bushing failure scenarios.

 

2 Drivers to Monitor Bushings

2.1 Aging Asset Base

Industry observers [1] consider that the global task of the electric power industry for the next 20 or so years will be to manage the serviceability of a huge transformer and bushing population that has already been in service for 25-40 years. Noting already the propensity for early bushing failure before end of life of the transformer asset, whilst recognising both the high population base of installed HV bushings and their propensity to fail rapidly once the mechanism begins (Section 3), time is not on our side to begin a pro-active on-line condition monitoring programme.

2.2 Insurance Companies

Whilst an HV bushing may be worth in the range of $10-20,000 to purchase, one quickly finds it is not a simple matter to consider the low-cost option of automatically replacing them at the mid life of a transformer. Practicalities such as compatibility with existing transformer designs or internal and external connection geometries, seldom make this an easy or practical option.
On the other hand, a bushing-initiated failure can produce losses to insurers [12] of between USD $1 and $3 million for Physical Damage and for Business Interruption in the same order of magnitude, although an exceptional business Interruption loss may escalate to tens of millions of US Dollars. Typical of the latter would be a failure of a large GSU with little spare capacity to service the output of the generator to which it attaches: commitments of the generator company to supply the lost power may see them having to not only repair the physical losses but also to buy in the committed power from the spot market at something like $250/MWh, easily costing up to $1 million per day for a reasonable plant size. Similarly frightening numbers would confront a large process industry such as an aluminium smelter where units of over 100MVA are not uncommon and losses in production and order commitments levy major additional consequences. A natural weighting of the largest cost-of-consequence failures occurs [12] in the transformers of the generation, transmission, and distribution industries.
It is no wonder, then, that the Insurance Industry takes a keen interest in the risks of transformer failure [11,26] and often have a their own full-time transformer specialists [13] involved in the assessment of their risk and allocation of associated premiums to the asset owner.
With some process industries in this region paying insurance premiums of around $40 million with $5 million excess to self-insure the likes of a transformer loss it is clear where the real drivers to monitoring originate. The value of even the largest pieces of HV substation assets is really only the tip of an economic ‘iceberg” when it comes to the true costs of a bushing failure.
A spokesperson for the insurer HSB Power Generation ( a unit of AIG Global Energy NY ), commenting in late 2001 [11] on the aging population generation and transmission equipment, commented typically of the Industry’s views: “We have seen utilities take two paths—those that run to failure and then replace, or those that demonstrate reasonable risk while investing in maintenance and reliability programmes. It is this latter group we wish to retain as clients”.

2.3 Related Industrial Issues (Including Health & Safety)

Given the violence of bushing explosions, somewhat determined in proportion to system voltage, coupled with the huge coverage area that the jagged porcelain fragments may fly, Health and Safety issues and Union advocacy of them is now a major driver toward bushing monitoring [5, 15, 18, 19, 20, 21].

2.4 Liability-Statutory Requirements

Asset owners find themselves exposed to a legal and statutory risk if an asset fails. Another driver is that legal precedents have adjudged a human life to be worth sufficient to merit a $10 million claim. Environmental concerns after such an event now weigh heavy also, ranging from modest soil cleanups to the potential loss of one’s ISO614000 certification.

2.5 Regulatory Penalties

Losses may well contribute to significant performance issues that the regulators will rule upon.

2.6 Exposure to Compensatory Legal Action

Major losses affecting large parts of a city or region may no longer be regarded as an event of the most unlikely scenario. For what appear unrelated reasons, the incidences of such events are rising rapidly and will continue to do so over the next ten to 15 years as the average HV asset population reaches the end of predicted service life.
Given that cities effectively cease to function once power is lost, a major driver is now emerging from the absolute shambles that such an event, whatever the cause, will almost certainly precipitate. Auckland and Sydney have experienced such issues, as have New York State, Central London, and Italy in recent times.
Legal contractual protection is unlikely to prevent major lawsuits being lodged. Asset owners will now inevitably be facing direct questions from the new quarter of the “average” consumer who might have remained silent until now but, having now woken to the possibility of their power suppliers not having taken all possible steps to assure themselves of the condition of assets using currently available “best practice” technologies, might well have a very strong case.
Clearly, the best defence for an asset owner henceforth under a climate of such potential interrogation is monitoring to prevent consequences of unforeseen failures. As we shall see shortly, such technologies do exist now and are economically justifiable.

2.7 Economic Factors

Economic consequences of losses are still the major driver toward monitoring and will be reviewed further in Section 5. CFO’s of asset owners are now integrally critical to implementing monitoring programmes as they are driven by the same drivers as the network engineers: maintaining adequate reliability and security of supply with minimum cost., and offsetting future asset replacement and consequential loss costs by prudent net present value expenditure.

2.8 Political Fallout and Loss of Reputation

Whenever a loss of supply occurs, and more so if there is a spectacular event associated with it, now attracts ferocious political and media interest. This process quickly counts the human and commercial costs of the loss and allows little sympathy for a weak technical response from the asset owner. One need only perform web searches under the obvious headings “transformer failure” and the like to see the massive amount of material that may be easily read in regard to historic events. To its credit, the Industry attempts to pre-empt loss of reputation by posting interim or completed failure reports on the web directly, often with an excusable positive spin. CEO’s of such companies involved are generally politically and performance monitored and incidents of a major nature, no matter how small their origin, may signal a premature end to their career. Accordingly, much of the driver toward monitoring in recent times is now coming from the CEO.

2.9 Other

Other drivers include avoidance of costs from unplanned outages, collateral damage ( estimated by EPRI [5] to be an order of magnitude higher than an HVCT replacement cost ), or premature demise of the asset. Reduction in current maintenance costs is also cited.

3 Bushing Failure Mechanisms

With the literature suggesting good levels of reliability and longevity from Resin Impregnated Paper (RIP) condenser-graded bushings [29], our discussion will be focussed in the main upon paper-oil-condenser bushings. These are cleverly designed [16, 17] to establish a near-uniform dielectric stress across the radial distance between core and ground, achieving this typically by way of wound layers of oil-impregnated Kraft paper trimmed longitudinally by strips of conductive material to form a graded capacitance divider string (figure 2).
It is primarily via various mechanisms causing that controlled voltage gradient to become significantly uneven that the bushing ultimately fails.
Notwithstanding the preceding comment above, there exists enough evidence [30] to suggest that resin-bonded paper bushings have shown that fault-development processes and diagnostic parameters are similar to those in oil-impregnated bushing types.
Whilst some specific designs have known flaws and failure modes as they age [16,17, 28, 29,30], defects in a bushing [2, 3, 4, 8, 16,17] may originate from one or more of four main parts of the construction: core, core surface, oil, and the porcelain inner surface. Examining each area in more detail and referring to Table 1 ( after [4,30] ), we now consider the mechanisms in turn.

Figure 2: Bushing Design. Practical bushings have layers of conductive foil wound in to the dielectric in such a way as to form a uniform capacitive voltage division from conductor to ground potential. This results in an essentially uniform voltage gradient throughout the dielectric material.

3.1 Core Defects

Defects may be of an overall nature such as ingress of moisture or air, or high losses of impregnated liquid, all of which might result from leaking or deteriorated gaskets. Alternatively, they may be localised such as residual moisture, poor impregnation of oil into paper ( resulting in X wax build up on paper and partial discharge at foil ends), migration of conductive graphite ink used in some bushings instead of foils, shorted layers, overstressing, or dielectric heating. Wrinkles and delamination in papers may also cause defects.

3.2 Core Surface Defects

Defects may arise from surface moisture, contamination from oil aging products, and deposits of metal or carbon particles form the transformer. These may give rise to PD on the core surface, reduced surface resistance, and increased dielectric losses. Failure modes resulting from core problems include ionisation, gassing, thermal runaway, puncture, flashover, and explosion.

3.3 Aging Defects of the Bushing Oil

As the oil deteriorates from the effects of temperature ( localised core heating even causing carbon in the oil ), moisture, electrical field, bushing solid materials, the oil loses its dielectric strength and also results in colloid-type contamination ( containing copper, aluminium, zinc etc ) and semi conductive sediment. As the oil/paper ratio is so small in the bushing, even small quantities of oil loss or moisture ingress from atmosphere or core may significantly degrade the insulation qualities inside the bushing.

3.4 Effects on Internal Porcelain Surface

This surface may receive deposits of carbon, and semi conductive sediment from the oil breakdown and even from particles of iron from pump bearing wear. Impurities in the porcelain may give rise to PD and catastrophic failures from this mechanism have certainly been noted in epoxy bushing construction. Failure modes from oil and porcelain surface problems include PD, surface discharge, and gassing. Sokolov [30] observes that flashover along the internal surface of the lower porcelain constitutes ‘the typical unexplained failure mode”.

3.5 Conductor Defects

Connections at the top, and bottom of the bushing have been known to overheat and this may be exacerbated by use of dissimilar metals. Overheating of the draw rod may result, as may gassing and sparking.

3.6 External Defects

Cracks, contamination, and surface discharge may occur and result in a flashover. Animal contact flashover and failures are not unknown. Improper storage of spare bushings has been known to promote failures. Problems with seals may quickly cause major defects from loss of oil or moisture ingress.

3.7 Result of Local Defects in a Bushing Core

Core faults as discussed above result [4] in the development of two types of physical fault:
• electric-destructive ionization at the place of overstressing;
• thermal-dielectric overheating and thermal instability.
Each results in a defective area with excessive conductance appearing between two or more core layers. Ultimately, this develops into an increasing conductivity and localised tan delta, resulting in a burning through between papers and the occurrence of a short circuit between two of several layers ( Figure 3 ).
As core layers are shorted, C1 is effectively increased (+10% over nameplate being regarded as very serious [17]), reducing the reactance of C1. The increased C1 current that results from the increased dielectric losses may precipitate further localised heating and finally a state of thermal runaway: an explosion is practically inevitable from this condition. This situation affects in turn the overall tan delta of the core and generates a change in the partial conductance between the central tube and the test tap.

Figure 3 A bushing failure at a 400 kV Generator Step-up unit in the UK.
Note the classic failure pattern, in this case occurring in an inner capacitive layer of the bushing. It begins to heat causing the paper to dry and then burn outward.
As the paper burns, hydrocarbons are created that cause gases to be generated in the bushing. Eventually these gases build up pressure and an explosion and fire results.

3.8 Detectable Signatures of Failure Mechanisms

Signatures [4, 6, 30] of bushing core faults are primarily:
• change in the C1 dissipation factor of the core tan delta (on and off-line);
• change in the leakage current due to C1 changes.
Condition assessment of critical aging of the oil and of the formation of semi-conductive residue on the inner surface of the porcelain is also evident in a decreasing C1 Tan Delta, possibly exacerbated by increasing temperature [4].
Indeed, an increased variation of tan delta with bushing temperature (increased temperature coefficient ) and applied voltage is a further acknowledged symptom of bushing deterioration [4,6,18].
In general terms these signatures constitute the basis of the on-line monitoring technologies employed for bushings.

4 MONITORING OPTIONS (after [3, 4, 5, 6, 9, 16, 18])

4.1 Literature Review

A review of the literature in the regard is summarised in Table 2 below.

4.2 Selecting Monitoring Method According to Strategic Priority

Given the overwhelming published evidence of bushing failures ultimately being extremely rapid and erratic, particularly in the final stages of deterioration, as well as the indicators of deterioration being most easily confirmed in an energised condition, strategic bushing and CT assets require a continuously on-line monitoring method in order to provide a timely detection of such situations. Krieg [20,22], Sokolov [25,30], Lau [9], Bradley [33], and Cigre [4] all support the latter observation.
A variety of on-line bushing monitoring technologies exist [9, 30, 33 et al], the more serious contenders comprising of unbalanced neutral current, leakage current, and tan delta methods, each of which are capable of covering a large part of the probable defects. Of these methods there is debate about the relative merits, but tests [33] conducted to compare them have concluded that a comparative tan delta system would appear to be as comprehensive as may be found for any single method applied to the task, more sensitive at providing an early warning, stable in field conditions, and capable of providing unique identification of the affected bushing from those under monitoring observation. This method is certainly one of the most widely implemented and successful in the Australasian region.
By the same argument, the condition of assets judged to be of a second tier strategic priority is most effectively monitored via an episodically interrogated relative on-line tan delta system. To be effective, sample intervals must be closely spaced, a weekly interval being sensible. Off-line methods are clearly only suitable for assessing the lowest priority assets.

4.3 Brief Technical Review of the Relative Tan Delta Monitoring Method

Work on the development of a practical relative tan delta monitoring system may be traced back to the early 1990’s research by CSIR in South Africa [6]. Partnering in 1996 with AVO International USA to refine and commercialise the designs, the resulting product has been a commercial success in a variety of refinements and continues to be further enhanced by the current technology owners, On-Line Monitoring Inc ( “OMI” ) USA. EPRI USA established a CT Evaluation project in 1997 at 4 research labs, reporting [5] the benefits of monitoring at rated voltage and operating temperature vs. off-line methods. Bonneville Power HV Lab provided independent verification of the comparative tan delta technique soon after. Over the past year EPRI have continued this research in partnership with OMI, whereby US utilities are confirming the benefits of an episodically downloaded on-line relative tan delta system to conventional off-line methods.

Figure 4: Schering Bridge circuit used as the basis of Tan Delta testing of Bushings and HV CT’s

As with conventional off-line tan delta test sets, the monitor employs a conventional Schering bridge concept (Figure 4), with the important distinction that in the on-line system the standard reference capacitor is derived instead from a second bushing in the substation. The signals Ur and Ux are sampled and processed under software control to calculate the tan delta and capacitance relative to that unit. The principle of cross-referencing is extended in present systems to a minimum of two other bushings, a total of three bushings thus being needed for a functional installation.

In practice the system employs the voltage output from the bushing, accessed via the bushing tap, this varying in direct proportion to the C1 condition. The software corrects for phase angle differences. To achieve the coupling a custom-matched “Bushing Tap Coupler” is threaded into the bushing (Figure 5). This produces a normalised measurement voltage which is fed to the monitoring equipment containing a capacitor network whose primary reactance is much less than the bushing C1 value. The output of this network is a voltage proportional to the current flowing in the bushing insulation.
As each comparison is made on line and in a closed loop configuration, common mode effects such as ambient temperature [30], operating voltages, load conditions are removed from direct influence, permitting instead a trended change in relative condition. Having achieved this, one is able to assess more reliably the degradation in bushing condition from a number of mechanisms that are exhibited via the behaviour of the individual bushing tan delta with temperature [30].
Alarm status is drawn from both changes in baseline data for each bushing and from the degree of relative change from a given bushing to another, the latter criterion justifiably receiving a higher weighting factor.
A proprietary four step statistical analysis produces reliable alarms upon the determination of adverse relative condition status, as opposed to the likes of actual measured values.

Figure 5: A Bushing Tap Coupler Installed.

5 Economic and Commercial Justification Process for On-line Monitoring of Bushings

Traditionally guiding the process for any company was the quest to balance the achievement of the highest possible system reliability at the lowest reasonable cost to do so ( this in turn balancing capex and opex). Aside from this traditional balance, what asset owners are now more so being faced with is the maximisation of asset life whilst balancing the cost and risk of doing so [31].
Impacting on activities are the various drivers identified in Section 2 earlier. HV assets in generation, transmission, distribution, and process industries will inevitably have varied weightings applied to each factor, as well as criteria specific to that industry segment such as Regulator requirements for the energy sector.
In his excellent paper on the subject in 2002, Krieg [20] identifies a new framework by which to approach the place of on-line monitoring in the asset management issue, one which is also endorsed by Kingsmill and Phillips’ work in 2003 [21], both arising from successful such implementations:
• Asset management linked to corporate objectives
• Economically justified decisions. Criteria such as net present value of future benefits, discount rate, performance index, payback, and internal rate of return assessed against such outcomes as: deferred capital expenditure, reduced maintenance and routine inspection costs, longer asset life, ability to nurse an asset until a timely and planned replacement may be implemented, and costs of premature asset damage and unplanned outage from catastrophic failure.
• Balanced CAPEX/OPEX
• Information for strategic decisions
• Risk framework for decisions
• Data mining
• Use of expert systems to extract maximum diagnostic and signature information, returning optimum benefit from the monitoring investment.
Both agree that technical and non-technical factors (including safety) are readily integrated into the economic models.
In regard to more recent developments in the Australian context, health and safety has become one of the major drivers toward on-line bushing monitoring in itself, while other economic models have seen ready justification of the on-line concepts as part of mid-life transformer refurbishment programs across all sectors of the Industry.
Also only very recently emerging as clear drivers in their own right are those of “non-negotiable” corporate risk decisions in regard to the likes of basement zone substations in CBD areas, or the main power supply hardware supplying large process industries where in each case the relatively minor cost of adding bushing and ancillary monitoring is viewed as ‘acceptably low’ in relation to the alternative cost of brand/corporate damage or consequential production loss, respectively.

6 On-Line Relative Tan Delta Bushing Monitoring case studies

6.1 Current implementation Levels

Aside from those systems used in early trials and EPRI projects before 1997 ( of which at least 9 systems may be counted in South Africa [6]), it is currently believed that over 1650 bushings and CT’s on 94 sites internationally are now being monitored on-line via the current generation comparative tan delta technology. With nearly 20% of the monitored bushings being in Australia, it is pertinent to note also that Australia has some of the largest such installations, ElectraNet in Adelaide [22] installing their first system in 1997 following a catastrophic failure of a 275 kV CT at Torrens Island and ultimately installed 4 systems totalling 195 bushings and CT’s. One Australian power transmission company recently installed a 66 point system, whilst another embraced the concept for a mid-life HV transformer refurbishment. Projects are well advanced in the generation sector also focussed on mid-life refurbishments, whilst the distribution sector is rapidly making in-roads on prioritised sites.

6.2 Selected Case Studies

Fricker [6] outlines two case studies of successful intervention, one being an 88 kV floor bushing and another a 400 kV CT.
ElectraNet reported [22] the early detection of 3 abnormal samples.PSE&G [18] discuss a timely intervention due to loss of oil in a 500 kV CT, permitting a controlled repair. Figure 5 shows a tracking screen from the relative on-line tan monitoring system. The red line indicates clearly the rapid change in bushing insulation condition over a very short time (one day). This monitoring system’s annual depreciated cost per point was USD $622 and the avoidance of this one failure alone is estimated to have avoided an expenditure of USD $31,870, or about 30% of the initial monitoring system cost
.Figure 6: 500kV CT bushing failure caught in time by relative tan monitor.
(Note short timeframe of fault development.)

7 FURTHER INNOVATIONS WITH REGARD TO ON-LINE RELATIVE TAN DELTA SYSTEMS

With the technology of bushing tan delta well established, there has been a great deal of further research recently to add further innovations of transformer and bushing condition measurement via both the common access point of the bushing capacitive tap and the optional expansion of the associated measurement platform itself. This effort has now made possible the on-line transformer/bushing partial discharge detection, whilst research is proceeding apace currently to add the technical capabilities of passive transformer frequency response analysis [32].
Expanding upon the above concept by way of employing the relative tan delta hardware as a point of interface for other, related, monitoring of associated substation equipment, one maker now even permits the reliable monitoring of surge arrestors (via the relative leakage current monitoring concept ), as well as the ability to interface to any DC output signal from the likes of on-line hydrogen, moisture, transformer/ambient temperature, humidity, and switchgear, monitoring and trending each parameter via the one SCADA ‘point of contact’ [27].

8 Conclusions

HV bushings and CT’s are prone to catastrophic and often unpredictable failures earlier in life on average than assets to which they interface. Examples of such failures are widely documented in the literature and are not at all uncommon. Consequences are generally severe, highly disruptive, and result in damage significantly beyond the value of the failed device itself.
As such, HV CT’s and bushings pose a major concern to the reliability of strategic assets. Risk assessment processes all are unanimous in concluding that the only reliable means of timely intervention is via
on-line monitoring.
The present technological status of relative tan delta approach to monitoring offers a reliable platform from which to gain this assurance.
A significant uptake of the technology in the region, coupled with a recently-expanded range of associated upgrade and expansion options to monitor related parameters and assets, has firmly established the concept as a ‘best practice’ approach.

Bibliography

1 “Transformer Life Management”, V. Sokolov, II Workshop on Power Transformers-Deregulation and Transformers Technical, Economic, and Strategical Issues, Salvador, Brazil, 29-31 Aug 2001.

2 “Transformer Risk Assessment Considerations” by Sokolov, Bassetto, Mak, and Hanson, EuroTechCon 2002.

3 “Bushing Failure Rates/Mechanisms etc” by J Stead. Weidmann 2002 LV Conference Presentation.

4 “Life Management Techniques for Power Transformers” Prepared by CIGRE WG A2.18, 20 Jan 2003.

5 “Comparison of a new Technique for Power Factor Measurement at Rated Voltage with the Standard Off-Line Test on Bushings and HVCTs”, EPRI Project Opportunity-Substation Operation and Maintenance, Feb 2003

6 “On-Line HV Insulation Condition Monitoring…Substation On-line System”, by RK Fricker, Power Vision Programme, CSIR, South Africa. Presented at the 1997 AVO International Technical Conference, Dallas, Texas, Sept. 14-17, 1997.

7 Managing High Voltage Current Transformers and Bushings Using On-line Insulation Monitoring Techniques”, by Terry Krieg, ElectraNet SA, and Jeff Benach, AVO International, TechCon Asia Pacific, 2002.

8 “On-line Monitoring of High-Voltage Bushings”, by Sokolov and Vanin, Proc. Of the 1995 International Conference of Doble Clients Sect 3-4.

9 “500 kV Bushings Failures and Bushing Sampling Program” by Mike Lau, BC Hydro, 2001

10 “Huntly Power Station Transformer Incident-Why it was Catastrophic”, by Harvey O’Sullivan, Tri-Sheras Ltd, EEA Conference, Auckland NZ, June 1999.

11 “Operations-Extreme Loading”, by Mark Janick, “Electrical World T&D” magazine, Nov/Dec 2001.

12 “Failure of large Oil Cooed Transformers”, IMIA 16-66 (96) E. Working Group of the 29th conference of the International Machinery Insurer’s Association, Sept. 1996

13 “Risk Assessment of Power Systems Assets…The Insurance Perspective”, by Paul Boman, Hartford Steam Boiler, USA, AVO New Zealand / LORD Consulting Third International Technical Conference, Methven NZ, Oct 15-17, 2002.

14 “Condition Monitoring of High Voltage Electrical Equipment ( with an Emphasis on Transformers)”, Ron Park, Park Consultants Ltd, Presented at the Third AVO New Zealand/LORD Consulting International Technical Conference Methven NZ, Oct 15-17, 2002.

15 “Innovations in Hardware and Software Architecture Applied to On-line Monitoring of Strategic HV Substation Assets”, by Terry Krieg, Power and Water, Darwin. Presented at the Third AVO New Zealand/LORD Consulting International Technical Conference Methven NZ, Oct 15-17, 2002.

16 “High Voltage Bushings”, Young & Caruso ( Lapp Insulator Company USA ), and Gill ( Gill Engineering ), presented at the AVO International Technical Conference, Dallas Texas, 1996.

17 “A Bushing Presentation”, by John Leech, KC&KM Consulting, AVO International Technical Conference, Dallas Texas, Sept. 14-17 1997.

18 “PSE&G Insulates Itself from Failure”, by George Binnas, PSE&G. ‘Transmission and Distribution World’, Sept. 1, 2001.19 “Transformer Bushing and HV CT Monitoring”, by Trevor Lord and Jeff Benach, ‘Australasian Power Transmission and Distribution’ magazine, Aug/Sept 2003.

20 “Successful Economic Justification of On-line Monitoring Systems for Power System Assets”, Terry Krieg, Power and Water Darwin, Presented at the Third AVO New Zealand/LORD Consulting International Technical Conference, Methven NZ, Oct 15-17, 2002.

21 “Use of Online Condition Monitoring in the New South Wales Electricity Transmission Network”, A. Kingsmill and P. Phillips, TransGrid, TechCon Asia Pacific, 8-9 May 2003.

22 “Innovations in Hardware and Software Architecture Applied to On-Line Monitoring of Strategic HV Substation Assets”, T. Krieg, presented at Third AVO New Zealand/LORD Consulting International Technical Conference, Methven NZ, Oct 15-17 2003.

23 “Australia/New Zealand Transformer Reliability Survey 1996 Report”. Western Power Transmission Projects Branch.

24 “Changing World Perspectives-A Report from CIGRE”, V. Sokolov, Fourth AVO New Zealand International Technical Conference, Methven, New Zealand, April 2005.

25 “Transformers condition-based Ranking on the basis of design review and comprehensive oil analysis”, V. Sokolov, Fourth AVO New Zealand International Technical Conference, Methven, New Zealand, April 2005.

26 “Analysis of Transformer Failure…the Insurance Company Perspective”, W. Bartley, Hartford Steam Boiler Inspection and Insurance Company, USA, Fourth AVO New Zealand International Technical Conference, Methven, New Zealand, April 2005.

27 “Strategic Advantages of Relative Tan Delta Monitoring of HV Bushings and CT’s”, J. Benach and R. Anand, On-line Monitoring Inc, USA, Fourth AVO New Zealand International Technical Conference, Methven, New Zealand, April 2005.

28 “Recent Failure Experience with HSP Bushings”, M.V. Turley, Doble Engineering Company, USA, presented at Doble 2004 Conference.

29 “Discussion of the Marshall F. Turley Paper: Recent Failure Experiences with HSP Bushings”, R. Krump, HSP Hochspannunsgerae Porz GMBH, presented at Doble 2004 Conference.

30 “Evaluation and Identification of Typical Defects and Failure-Modes of 110-750 kV Bushings”, V. Sokolov, B. Vanin, ZTZ-Service Co, Ukraine.

31 “On-line Monitoring as a Strategic Tool to Enhance Supply System Reliability”, G. Hodge and T. Lord, LORD Consulting, New Zealand, 2003. Presented at D2003 Adelaide, Australia, Nov 2003; EEA NZ, Christchurch, NZ, June 2004; TechCon Sydney May 2004.

32 ‘Development of an On-Line Frequency Response Analysis System Based Upon Random Transients”, EPRI Contract EP-P12664/C6307, EPRI Substation and Diagnostic Conference X111, March 6-9 2005, New Orleans, LA.

33 “Detection of Imminent Failure of Oil Filled Current Transformers-Final Report”, D.A. Bradley, Bonneville Power Administration Engineering and Technical Services, Vancouver, WA, USA, 2000.