INMR Issue 105, Q3 2014

www.inmr.com

INMR VOLUME 22 NUMBER 3 • QUARTER THREE - 2014

Issue 105 - Quarter 3 - 2014

SPECIAL DOUBLE ISSUE Contains

2015

WORLDWIDE DIRECTORY OF HV/HP LABORATORIES

A. J. (Tony) Carreira Receives 2014 Claude de Tourreil Memorial Award More than 8½ years have passed since that cold morning in February when I received a message from Claude de Tourreil that he was dying. His words were to the point – typical of his style in everything he did – “diagnosis of acute AML type leukemia … prognosis bad …. only 30% chance of remission.” I could hardly digest the words. Within only weeks, I was rushing to a hospital in Lausanne to say my final farewell to a dear friend and mentor – the person who had inspired me to take up the challenge of creating INMR and keep it at a level he would approve.

I am delighted to announce that A.J. (Tony) Carreira has been designated the 2014 recipient of the Claude de Tourreil Memorial Award for Lifetime Achievement in the Field of Electrical Insulators.

The reason it is a particular pleasure for me is that I have known Tony since 1990 – even before the launch of this journal and indeed even before I first met Claude. During a telephone interview with him that year, Tony was the first person to talk to me about composite insulators – at that time still widely viewed as an unproven technology with an uncertain future.

Yet, in spite of this, I can still recall him passionately recounting all the reasons why these insulators were going to change not only the face of the industry but also how power lines are designed and maintained.

I found it easy to be infected by his enthusiasm and accept his belief that the world of outdoor insulation was about to change. Later that same month, I was introduced to Claude who was working as a research scientist at the IREQ laboratories of Hydro Québec. He confirmed to me that insulation technology was indeed 2

In one of his last editorial contributions to INMR in 2005 – perhaps sensing the end was near – Claude wrote: “If I have ever done anything to offend someone, I express my apologies. All I was trying was to do my work as best I could.”

Well, far from offending, Claude made a lasting contribution to the discipline of outdoor electrical insulation and helped lay a foundation on which the industry is still building. That justifies preserving his legacy – not only for those who knew and respected him but also for the growing numbers of young power engineers who may never have heard his name yet still benefit from all that he accomplished.

undergoing a dramatic shift and that someone should be reporting on it.

Well … the rest as they say is history.

Tony seems to have been destined to one day be honored with an award in Claude’s memory because he too has been a steady contributor to the development of today’s insulator industry.

After receiving his Bachelor of Applied Science in Electrical Engineering from the University of Waterloo, he began a career at one of Canada’s largest power suppliers, Ontario Hydro (now Hydro One), where he held engineering as well as management positions

over 13 years. This extensive work on the user side of the business gave him an excellent perspective on the design, testing and application requirements of insulators – information that would serve him well starting in 1990, when he was hired as President of Toronto-based K-Line Insulators.

Since that time, he has helped transform that firm from a local supplier of mostly lower voltage insulators to an international player covering most overhead line needs. Somehow, in the process, he also made the time to volunteer for a range of external work as part of diverse committees within CIGRE, IEC, CSA, CIGELE and as Senior Member of IEEE. Tony has authored many technical papers on insulators, which he presented at conferences across the globe, including ICOLIM, IEEE and the INMR WORLD CONGRESSES. On behalf of INMR and its readers throughout the world, congratulations Tony!

Marvin Zimmerman Publisher

Tribute T

The past months have witnessed the passing and the retirement of two longtime contributors to the world’s insulator and power industries. Both will indeed be missed.

his past May, Peter Pfaff succumbed to a battle with cancer that had afflicted his family for generations. A charming and vivacious man with a passion for baseball, he dedicated much of his life to the firm he founded, Glasforms – a manufacturer of pultruded FRP core rods and other advanced composites.

Over the years I knew him, I was always struck by his constant good humour and seemingly endless optimism, even when I last saw him a year ago in Vancouver.

There, he told me of his illness and that this could well be his last INMR conference. Yet he was resolute in the face of the terrible battle he faced.

I met Peter in 1997 at the INMR WORLD CONGRESS in Miami. Just like he apparently never missed a game of his beloved San Francisco Giants he too came to virtually every successive event after that, always being among the first to praise the organization or location. After the INMR Congress in Barcelona in 1999, we became close and Peter joined me for a week touring the Spanish Costa del Sol as well as playing golf.



Profound condolences from INMR as well as his many friends throughout the industry go out to Peter’s wife, Haeran, his family and to his colleagues at Glasforms.

T

his past month, I received a message from Bernhard Staub that, at 77, he was finally closing his engineering consultancy business and retiring from an industry he had devoted his life to. Bernhard had a long and distinguished career in the insulator industry in Switzerland and across other countries as well. As an expert in porcelain insulators, first with

Laufen and later with Ceram, he carried out valuable work for CIGRE, recording the extremely low failure rates of porcelain long rods across Germany, Austria and Switzerland – where they are especially popular.

Good luck and long life to you as a pensioner, Bernhard, and in your newfound occupation promoting the readings, theater plays and exhibitions of your relative, Lina Boegli, who traveled the world from 1892 to 1902 and later wrote about her experiences.

Marvin L. Zimmerman

Ready to Serve You with a Complete Series of Instrument Transformers to Meet All Your Needs • Cast epoxy resin 0.5-35 kV • Oil/paper insulation with porcelain 35-500 kV

• SF

insulated with composite housings up to 252 kV 6

We are proud to be ranked among the Top Three Instrument Transformer Manufacturers in China We supply customers in Europe, Asia, Africa & South America

Dalian North Instrument Transformer Group (NORIT) Fengrong Industrial Park, Pulandian, Dalian, P.R.China Tel: +86 138 40897579 DLBF@DLBF.COM or NORIT.DLBF@GMAIL.COM

Hebei Xinhua High Voltage Electrical Equipment CO., LTD Rigid composite insulators represent an advanced insulator technology employing high temperature injection molding to ensure consistently high strength, fully sealed designs. This revolutionary technology offers excellent hydrophobic properties and other electrical performance characteristics while also being resistant to damage due to sand, wind, climbing by line workers or bird pecking. These rigid insulators also eliminate all risk of brittle fracture during service.

Gold medal for product design

Silver medal for manufacturing technology

The 42nd international exhibition of inventions in Geneva, Switzerland took place at the Palexpo in April 2014. Over 4 days of comparison of the many entries by the review committee, Hebei Xinhua High Voltage Electrical Equipment was awarded the Gold Medal for ‘Adhesive Binding Disc Composite Suspension Insulators’. Our manufacturing technology for this design won the Silver Medal.

Contact us to learn more about how this unique technology can better serve your insulator needs across a range of different applications.

Mailing Address: P.O. Box 19-918, Renqiu City, Hebei Province, China 062550 Tel: +86-317-2217266 Fax: +86-317-2218518 http://www.hbxhgy.com Email us at: xhe_exp@hbxhgy.com

Advanced rigid composite insulator Advanced production technology Advanced rigid composite insulating material

Contents Issue 105 Quarter 3 − 2014 Volume 22 − Number 3

46

Advertisers in This Issue

30

74 2 Claude de Tourreil Memorial Award for 2014 4 Tribute 12 Editorial Will Technology Help

‘De-Commoditize’ the Insulator?

14 Commentary by Pigini Dielectric Strength and Air Density

16 From the World of Testing Is Enough Being Done to Assure No Major Power Disruptions?

Insulators 30 Silicone Insulators: Lessons from 30 Years Experience in China

46 Latest Principles for Selection of DC Insulation

58 Resolving External Insulation Problems at HVDC Converter Stations

Testing 68 Innovations in Type & Commissioning Testing of High Voltage Cables

18 Reporting from CIGRE Reviewing CIGRE’s Recent

Maintenance 74 Monitoring System

20 Transient Thoughts ‘Fine Tuning’ An Aesthetic

82 Preventing Wildlife Outages

2014 Session

Replacement Line

22 Scene from China Evaluating Hazards of Lightning on Transmission Lines

Detects Critical Insulator Contamination on Overhead Lines at Substations

Cable Accessories 86 New Design of Dry Type Cable Terminations up to 170 kV

24 Woodworth on Arresters Fittings Do Your Station Class Arresters 90 Manufacturer Expands Have Adequate Fault Current Withstand Capability?

Production of Insulator Fittings

2015 WORLDWIDE 26 Focus On Cable Accessories DIRECTORY OF HV/HP New Outdoor Termination LABORATORIES Designs for Polymeric HV Cables 94 Quick Reference Charts 100 Alphabetical Listing of Laboratories & Capabilities 8

ABB Components & Insulation Inside Cover CSL Silicones - SiCoat Outside Back Cover Chengdu Line Fittings/Power China 56-57 Dalian Composite Insulator DCI 4 & 89 Dalian HiVolt Power Systems 3 Dalian Insulator Group 10-11 Dekuma Rubber & Plastic 92-93 Desma Elastomertechnik 41 Dextra Power 3 DNV GL KEMA 109 EGU HV Laboratory 114 Glasforms PolyOne 9 Haefely Hipotronics 110 Hebei Xinhua HV Electrical Equipment 6-7 HIGH VOLT 95 High Voltage 107 Himalayal 113 Hubbell Power Systems Inside Back Cover Hübers Verfahrenstechnik 37 Integrated Engineering Software 119 Jinan Meide Casting 37 KERI 121 Kinectrics 123 Manitoba Hydro International 127 Motic Electric 21 Nanjing Electric (Group) 65 National Engineering Lab. for UHV Tech. 97 NORIT Instrument Transformers 5 Phenix Technologies 107 Reinhausen Power Composites 13 SGD La Granja 1 Sediver 17 Shanghai Jiuzhi SAMGOR 99 SINTEF Energy Research 131 STRI 133 Shaanxi Taporel Electrical Insulation 19 Sichuan YiBin Global Group SYGG 28-29 Taizhou Huadong Insulation 35 TE Connectivity 27 TransGard Systems 81 Trench Test Systems 9 Tridelta Surge Arresters 25 Uvirco Technologies 5 WS Test Industries 107 Wellwin Precision Moulds 21 Wenzhou Yikun Electric 51 Yizumi Rubber Machinery Front Cover Zhejiang Fuerte 21 Zhengzhou Jingwei Electric 53 Zhengzhou Xianghe Group Electric 60-61 Zibo Taiguang Electrical Equipment 23

INMR Issue 105 www.inmr.com ISSN 2290-5472, E-mail: info@inmr.com Editor & Advertising Sales: Marvin L. Zimmerman mzimmerman@inmr.com 1-514-939-9540 中国地区联系方式：余娟女士 电话: 135 1001 6825 / juan.inmrchina@gmail.com

Magazine Design: Cusmano Design and Communication Inc. 1-514-509-0888 corrado@cusmanodesign.com Contents of this publication are protected by international copyrights and treaties.

Reproduction of the publication, in whole or in part, without express written permission of the publishers is prohibited. While every effort is made to verify the data and information contained in this publication, the publishers accept no liability, direct or implied, for the accuracy of all information presented.

Everything a Well-Built Network Needs ….

Composite Long Rod Insulators, Post Insulators & Hollow Core Insulators

As China’s best, largest and most diversified insulator manufacturer, we can offer a complete range of line, solid core post as well as hollow porcelain and composite insulators to meet the needs of any overhead line or substation application up to 1000 kV AC or ±800 kV DC. We also supply a full range of hardware and special fittings for insulators and overhead lines.

Porcelain Line Disc Insulators

Dalian Insulator Group Co., Ltd

No. 88 Liaohe East Road, DD Port, Dalian Economic & Technological Development Area, Liaoning 116600, China Tel: 86-411-84303112/ 84305786/ 84342270 Fax: 86-411-84305689 E-mail: info@insulators.cn·ISO 9001 Certified Plant National High-Tech Enterprises

And All From One Trusted Source Porcelain Hollow & Solid Core Insulators

With a tradition of insulator know-how going back to 1915, our newly-built factories are among the most modern anywhere, with efficient production flows and state-of-the-art manufacturing and testing equipment.

All this is supported by an experienced engineering and production staff dedicated to maintaining quality all along the line.

That means every order is made to the highest standards and also completed ready to ship according to your leadtime requirements.

Choose Dalian Insulator Group and benefit from dealing with one trusted source that can reliably meet all your insulator needs.

Insulator Hardware & Line Fittings

www.insulators.cn

EDITORIA Will Technology Help ‘De-Commoditize’ the Insulator?

When the first high strength ceramic insulators were developed during the 1960s, the capability to produce these much-improved products was limited to a handful of suppliers.

Two decades later, when composite line insulators were first launched, the know-how behind their revolutionary design and manufacturing was shared among only a few innovators. Then, in the late 1980s, there was only one supplier capable of offering composite hollow core insulators for applications up to 500 kV. In all these cases, there were obvious benefits that allowed technology leaders to capture large shares of their respective market segments based on better performing products.

Alas, Pandora’s Box of insulator technologies has since been pried open. The once closely guarded secrets behind superior porcelain, glass and composite insulators have become available to almost any firm willing to invest in the necessary machinery and know-how.

That means competing insulator manufacturers these days are battling over global market share but without the same clear and demonstrable product advantages as in the past.

Not surprisingly, this market dynamic has made price and delivery key drivers in this business and created great pressure to lower cost and shorten production. At the same time, because there are no longer obvious technology leaders, users have grown to regard insulators as commodities and the common denominator to many is that they meet all applicable standards.

But is it correct to think of insulators as interchangeable commodities? Experience has demonstrated that comparable insulators from different manufacturers can be equivalent – but not always. While some competing products offer similar performance and life expectancy under the same service conditions, not every insulator will measure up. Subtle differences in material, design and manufacturing can impact how different insulators perform and age. In other words, different products that pass all the tests set out in the standards may still end up having much different service histories.

Given that not all insulators will deliver the same performance, why is it that many users have come to regard them as commodities? This is a challenging question. But it's one that should be preoccupying an industry where there were once an impressive array of different design and production patents?

At the IEEE PES T&D Show in Chicago last April, the keynote speaker was Daniel Burrus – a self-described ‘technology futurist’. According to Burrus, new technologies can change the rules of competition in every field and again allow manufacturers to develop innovative products that are better differentiated from the competition. But what should be the driver for new technologies in the field of electrical insulators? Is it only a question of further reducing their cost?

The old industry axiom goes: Insulators represent 5 to 8 percent of investment in new power infrastructure but account for the bulk of subsequent maintenance expenses. They are also responsible for most of the costly disruptions due to flashover or other failure.

It would therefore seem that the insulator industry should be focused more on improved performance rather than on cost reduction. For example, if technology could further reduce the cost of insulators by 10 percent, the impact on total investment in new power networks would be marginal.

Let’s get back to Burrus. During his speech at IEEE/PES, he predicted that sensors are the way of the future and will revolutionize what users expect from traditional products. For example, a Swiss drug-maker has reportedly just struck a deal with Google to develop ‘smart’ contact lenses that help diabetics track blood glucose levels through a device that analyses eye fluids and sends the data wirelessly to a mobile device. This technology is potentially life changing for diabetics, who must stab their fingers several times a day to check blood sugar.

Are there not opportunities to add sensors to insulators as well? Devices that would alert power companies when they are in need of replacement or other maintenance, e.g. when it has become necessary to wash them to avoid risk of pollution flashover (e.g. see article on page 74). That could offer huge potential savings to the power industry given that the interval between successive inspection cycles for insulators tends to be shorter than for items such as towers or conductors.

Another important message from Burrus was, “if you don't do it, someone else will”. That means that if the latest technological advances will indeed again allow for clearly superior insulators, the industry may undergo a ‘shake-up’. As in the past, the most innovative suppliers will once attain market leadership based on superior products that deliver more value added to users, e.g. because they can monitor themselves.

At last, smart insulators for smart grids.

Marvin L. Zimmerman mzimmerman@inmr.com 12

While a number of atmospheric parameters can affect the dielectric strength of external insulation, relative air density (δ) and absolute humidity are considered the most significant. Here, I will focus on the former, which plays a key role in developing transmission systems at high altitude, as for example foreseen in China up to 5000 m.

Dielectric Strength & Air Density

Studying the role of air density started over a century ago but accelerated with development of EHV and the need to optimize line and substation design under switching overvoltages. High altitude tests on large clearances were carried out in Russia (1967 Bazeylan & 1968 Volkova and al: tests up to 3370 m), in the U.S. (1967 Phillips and al.: tests up to 3500 m), in Japan (Harada and al. 1970: tests up to 1850 m) as well as in Italy, South Africa and Mexico (Pigini and al. 1989: comparative tests up to 3000 m). The latest such research comes from the need to optimize design of UHV projects at high altitudes in China, with systematic testing in Wuhan (35 m), Beijing (50 m), Chengdu (500 m), Yinchuan (1000) m, Lanzhong (1500 m), Kunming (2100 m), Xining (2260) m, Qinghai (3000) m and Tibet (4300m). Huge climate chambers, such as the one at CEPRI, have been built to simulate altitudes up to 6000 m.

At present, there are different approaches in the standards on how to account for changing air density with altitude. IEC 60060-1, for example, conceived for correcting laboratory tests, uses: U=Uo*K, where U and Uo are the dielectric strengths at high altitude and at standard atmospheric conditions respectively and where K is the air density correction factor given by K=δm with δ being relative air density at high altitude. IEC 60071-2, conceived for insulation coordination, makes direct reference to site altitude (H), being δ under simplified assumptions related to H by δ=e(H/8150). The main problem is determining the parameter m, which depends on type of voltage stress, insulation configuration, type of insulator and environmental conditions (e.g. dry, wet, contaminated).

Fig. 1 shows an example of the range in ‘m’ values found by various researchers for positive Switching Impulse for different configurations with and without insulators. Results are plotted as a function of gap clearance. In this same chart, the continuous curves represent the correction approach adopted in the old IEC 60 relating m to clearance. The new approach under IEC standards 60060 and 60071, where I contributed, attempted to better rationalize available information (then limited to 3500 m) relating the factor m to stress parameters instead of clearance. However, the approaches in the two are sometimes contradictory, even if starting from the same basic data, and they are difficult to apply as well. Moreover, they do not take into account new information from tests up to 5000 m. An urgent need therefore exists to update and harmonize such correction approaches while taking into account latest results, as recommended by IEC and supported by CIGRE, where two new working groups look at the influence of altitude on clean insulators (WG D1.50) and polluted insulators (WG D1.44).

Fig. 1: Switching impulse of positive polarity.

Range of m values as function of clearance.

Continuous curve: correction approach in old version of IEC 60.

Given my experience in the field, I offer several suggestions to optimize the new approach that will be proposed to IEC: 1. Influence of air density is generally a minor part of breakdown/flashover voltage: a small inaccuracy in measurement, in configuration simulation or in voltage parameters can lead to high inaccuracies in the parameter m when comparing results at different altitudes.

Comparative tests at various altitudes must therefore be designed and carried out accurately.

2. Best not to overlook the existing range of historical experimental data, using newly generated data to better integrate and implement them.

3. Many tests have been made and are still being made on basic configurations such as the rod plane under dry conditions, where influence of air density can be much different from that of actual insulator configurations. New data for actual configurations should be provided as much as possible.

4. One of the most important environmental conditions to be considered in design is performance under rain, which can dramatically reduce insulator strength depending on voltage, configuration and type of insulator. Since the relative influence of air density on insulator strength can change under rain, there is a need for more data to better understand how (e.g. by researching performance of insulators under DC voltage and rain).

Fig. 2: Switching impulse of positive polarity.

Discharge governed by streamers and leaders.

Extension of the streamer phase as function of air density (measurements by image converter)

Pigini Commentary

5. Since pollution is the governing design stress for DC systems, additional data is needed on the influence of air density on pollution flashover of hydrophilic as well as hydrophobic insulators as a function of their geometry.

6. Due to the complexity of the phenomenon and the many parameters involved, understanding the influence of air density can be made easier if accompanied by analysis of its impact on the physical processes leading to flashover, including its influence on the streamer and leader phases.

7. Because of the complexity referred to above, it does not seem possible to arrive at one approach that is both accurate and relatively simple. In the end, simplicity should be the goal for engineering applications and the required accuracy could be assessed by looking at the typical dispersion in experimental findings.

8. As much as possible the ‘formal’ approach should be the same for all the various standards in order to avoid the confusion of present standards that often express the same concept and give similar indications but use different language.

Alberto Pigini pigini@ieee.org

14

Since electricity has become the lifeblood of modern society, any interruption in supply has a huge impact on individuals, businesses and communities. While the transmission and distribution sector of the power industry makes great efforts to avoid such disruptions, the question must still be asked: is everything possible being done?

The electrical power business is going through another major transformation. Need for affordable and sustainable energy that still meets ever-increasing demand is driving growth in renewables (often remotely located) and also international electricity trading. That dynamic in turn has created the rise of so-called ‘super grids’ – extremely high voltage networks that transport electricity over much greater distances, often between countries and sometimes even between continents.

Is Enough Being Done to Assure No Major Power Disruptions?

Outsourcing practices by the power industry suggest that the system knowledge needed to avoid major power disruptions may increasingly be flowing away from network operators… and that leaves them at greater risk.

From the World of Testing

These large networks place added pressure on grid reliability. Losing power in a single region would be bad enough, but imagine what might happen if supply to half a continent were interrupted. Such a threat already attracts media headlines that speculate on when and where ‘the big one’ might strike.

Is there anything more that network operators can do to further reduce the risk of major power disruptions?

Consider how outages arise. The 2013 Eaton Blackout Tracker, for example, reported that bad weather was the leading cause of unscheduled disturbances in the U.S., accounting for 30% of major outages. Faulty equipment and human error together contributed a further 29%. Weather is obviously uncontrollable and, if current models of climate change prove accurate, will become even more extreme. Still, we can lessen the impact by ensuring localized problems do not escalate across the grid as well as by installing equipment that can withstand even extreme conditions.

Experience during the 2011 earthquake in eastern Japan demonstrates that such a strategy is feasible. A total of 848 pieces of key equipment at 200 high voltage substations were damaged by the temblor, causing the immediate shutdown of 80 stations. Clearly a major disruption. Yet electricity was fully restored in less than a week thanks to the stringent seismic specifications for all equipment installed in the region. The solution to reduce the impact of bad weather and other natural disasters is much the same, namely reducing equipment faults through higher quality components and better training of personnel on how to respond to these problems.

Most will agree that one of the most effective ways to ensure the quality of T&D components and reduce the frequency of major disturbances due to their failure is through independent testingbased certification. Such certification gives network operators confidence that equipment will perform correctly and safely under normal as well as under fault conditions. That not only helps reduce the frequency of failures but also the broader risk that any one failure might cascade across the network.

Certification must of course be based on international standards. But which? Present standards were developed largely based on experience from yesterday’s networks and service conditions.

With the power industry going through significant change due to the growth of renewables, increased grid utilization and longer transmission lines, these factors as well as climate change could affect the service conditions that components must routinely face in the future. In particular, more intelligence within networks raises the threat of cyber-attacks, adding yet another level of risk. Therefore, current standards need to be reviewed to ensure they will be suitable for future challenges as well.

That could mean adapting existing standards or creating entirely new ones, both for new types of components and for verifying power systems as a whole. Ensuring that the standards remain appropriate is a task that requires the combined expertise of the industry, working together through bodies such as IEC, IEEE, and CIGRE. While making efforts to improve quality, it’s also important to remember that modern components in the grid are not simply ‘plug-and-play’. Highly complex interactions can occur between them and these can greatly impact the stresses any single component might face. Hence, it’s essential that utility personnel fully understand their networks, not only at the system level but also in terms of the special stresses any equipment might face under ‘non-standard’ conditions.

Unfortunately, outsourcing practices by the power industry in recent years suggest that this kind of knowledge may increasingly be flowing away from network operators. That leaves them at risk. To combat the ‘brain drain’, more and more testing institutions (including DNV GL) offer training courses that provide the skills and knowledge engineers need for a more complete understanding of their system.

Such insight is essential for both planning and operating grids. In the planning stage, if you don’t fully understand all the conditions a component might face in your system, how can you define the correct specifications for buying it? At the same time, a systems-level understanding helps operating staff identify which unusual conditions represent a particular threat to the network.

Ongoing education is the best way to maintain such an understanding – and indeed even opens up the intriguing possibility of one day certifying people just as we currently certify components.

Jacob Fontijne jacob.fontijne@dnvgl.com 16

The 45th CIGRE Session was held in Paris during the last week of August. From the perspective of overhead line insulators, a number of events were scheduled and a range of relevant papers were accepted for the Technical Meetings of B2 and D1.

CIGRE’s unique format is based on a team of Special Reporters and has been in force now for many years. It sees specialists from across the globe (appointed by Chairmen of their respective Study Committees) review the papers relevant to each corresponding technical meeting. They then generate a Special Report that follows a pre-determined structure that includes: an introduction to the topic; extraction of the main points; and, most important, specific questions that focus subsequent discussions within the Technical Meeting.

Reviewing CIGRE’s Recent 2014 Session

All participants prepare their responses to these questions, typically using 3 or 4 Powerpoint slides sent to the Special Reporters some two weeks earlier. To shape and finalize the order of presentations as well as the contents of slides, contributors meet with the Chairmen and Special Reporters the day before the Group Meeting in order to review all planned contributions (e.g. for relevance, clarity, brevity). Spontaneous contributions are also permitted. All the Special Reports from this year’s General Session can be downloaded from the CIGRE web site (www.cigre.org).

In parallel to the Special Reporter system, a poster session was held on a different day from the Technical Meeting and allowed authors to present their paper content directly to interested visitors by means of posters.

A summary of some of the interesting papers related to line insulators at this past Session included: 1. Better utilization of existing transmission line corridors is increasingly important. For example, Paper B2-105 discussed use of existing infrastructure by means of a hybrid line concept whereby one circuit of an AC double circuit line is replaced by a bipolar DC system. In this same context, Paper D1-101 raised questions in terms of insulator behavior under such hybrid AC/DC stress.

Fig. 1: Comparison of different tower types, TWINNI as further development of Wintrack.

Other contributions in this area dealt with the concept of insulated cross-arms and solutions up to 765 kV (Papers B2106 & 107). Similarly, findings of initial operating experience with the Dutch Wintrack tower and line design were presented in Paper B2-110 while a further development, named TWINNI, with composite insulators, was discussed in Paper B2-111 (see Fig. 1).

2. A n example of one utility’s philosophy for insulator maintenance and inspection of composite insulators was presented in Paper B2-206 and then again for glass cap & pin insulators in Paper B2-209. From a specialized application point of view, a possible anti-icicle shed profile was discussed in Paper B2-204 (see Fig. 2) while the loss and recovery of hydrophobicity under severe climatic conditions was investigated in Paper B2-304.

Fig. 2: Reduced icicle bridging by varying insulator shed diameters.

Fig. 3: Flashover development at sea level (left) and at 4000 m.

Reporting from CIGRE 18

3. Two unusual papers on some insulation fundamentals were available in D1-212 & 213. In the former, experience with measurement of site severity using a Directional Dust Deposit Gauge (DDDG) was shown in different cases and compared to standard ESDD measurement (both as per IEC 60815). In the latter, the role of altitude on different flashover mechanisms was also discussed (see Fig. 3) as was of the impact of surface wettability.

Dr. Frank Schmuck frank.schmuck@sefag.ch

A recent project to measure resistance and lightning impulse impedance of a new transmission line has provided me another ‘education’ and reinforced the fact that while towers may appear identical above ground, foundations and surrounding soil conditions can vary considerably from one to another.

During the 1980s, the IEEE recognized the importance of this in estimating lightning performance in a practical way. Indeed, users of the FLASH program in IEEE 1243 (1997) are asked to prepare cumulative distributions of footing resistance based on 10% intervals rather than giving a mean or median value for the entire line. The software then uses each successive value to form a composite estimate of the line’s lightning trip-out rate.

‘Fine Tuning’ An Aesthetic Replacement Line

What I’ve noted recently is that most lines have statistical distributions of footing resistance that are ‘log-normal’ rather than normal. The raw values have high skew or lopsidedness as well as extra weight in the tails of the distribution (i.e. high kurtosis). Taking the logarithm of footing resistance allows a distribution that is normally distributed, with low skew and correct kurtosis. Indeed, the log-normal distribution is widely used to describe other lightning parameters (e.g. peak current and rate of rise) in up-to-date documents such as CIGRE Technical Brochure 549, “Lightning Parameters for Engineering Applications” (August 2013).

It seemed to me that this same log-normal distribution approach would be practical for grounding, partly because most lines with from 100 to 500 towers seem to share a common value of dispersion.

Whatever the median value of resistance (Rf), the standard deviation (σ) of the logarithms of individual values, lnRf to base e, is about 0.9. This should be divided by ln(10)=2.3026 to arrive at values of σ lg Rf of 0.4 when using logarithms to base 10. For context, the corresponding values of log standard deviation for peak first negative return stroke currents (I) in CIGRE TB 549 are smaller, i.e. in the range of σ lg I= 0.2 to 0.3 in most studies. This means that the statistical variation in footing resistance from tower to tower is considerably greater than the variation in peak return stroke current from flash to flash. Both aspects must therefore be considered when modeling lightning performance.

The project referred to above involved measuring transient grounding impedance on new monopole steel structures that were selected to replace lattice towers erected in the late 1950s. I admit it was sad to see long-serving structures cut apart so unceremoniously … lying in wait for a scrap dealer. This is partly a reflection of my age bias, since I’m known to turn my head whenever a 1950s era car drives by. At the same time, replacing the former lattice towers by steel poles also proved a pity from the perspective of the effect on surge impedance. A four-legged lattice tower presents an impedance of 130 to 180 ohms while the slim steel pole has about twice that value. Such high tower surge impedance will increase the voltage stress across line insulation whenever lightning hits either the groundwires or a tower peak.

The line’s former lattice towers had been painted green to better blend into the environment along the route, which traversed rock, backyards, a golf course and even a wetland habitat for turtles. Still, they had been deemed ‘technologically obsolete’. The new replacement line, for example, uses largediameter, steel-supported aluminum conductor engineered for a maximum operating temperature of 180°C and also incorporates optical fiber ground wires (OPGW) for protection and control.

Some of the new monopole towers, at intervals of about 3 km, have a cylindrical metal OPGW splice box and loop of cable tacked onto them – necessary to make, test and maintain a continuous optical circuit. My opinion is that this equipment and cable would be far better protected from lightning and other environmental damage if placed inside the pole. The adverse impact on the profile of the sleek tower is an even more obvious drawback.

Testing tower grounding with impulse currents as part of the safety process, I also measured the AC voltage from tower base to ground probes located only some 7 m away. It’s usual for grounded AC power system components to carry some neutral-to-earth voltage, almost always in the range of 1 to 10 Voltsrms. Tower base voltage varies with time and system configuration but is almost always elevated compared to the soil nearby.

One idea to improve the visual aspect of the unsightly OPGW splice box while also adding some monitoring capability for insulators would be to take advantage of the small amount of power that can be harvested from this neutral-to-earth voltage. Bringing an insulated wire from a ground rod near the tower up to the splice box could deliver continuous power of at least 1 mW to an electronic circuit. Ten years ago this may not have proven particularly helpful but the latest generations of micropower sensors require less than 100 microJoules to transmit a reading of temperature, relative humidity and insulator tilt or acceleration.

Transient Thoughts

Manufacturers of steel pole transmission towers could therefore look at pre-engineering canisters or access ports for optical fiber splicing and even possible Wi-Fi access. This would avoid the unsightly impact of tacked-on accessories detracting from the pole’s aesthetics. At the same time, attachment points for bottom-phase surge arresters and/or underbuilt ground wire (also described as aerial counterpoise) could be considered as part of an even more cohesive overall design.

INMR’s recent book, Inspiring Journey, captures a wide range of the impressive transmission structures now installed across the globe and recent articles at www.issuu.com/inmr describe such remarkable innovations, e.g. the Wintrack design used by the Dutch grid operator. With greater attention to these types of detail, utilities can better ensure that every tower along a line is as visually pleasing as these examples and make lines more acceptable to affected communities.

Dr. William A. Chisholm W.A.Chisholm@ieee.org 20

Lightning accounts for about 60% of all transmission line trippings in China and is closely linked to season as well as region. Based on the physical processes, lightning overvoltages can fall into two broad categories: those caused by direct strike to the tower, shield wire or conductor; and those where lightning strikes the ground near lines, generating an induced overvoltage on conductors. Experience demonstrates that the former imposes the far greater hazard and that the latter most threatens lines of 35 kV or below.

Based on where a line is hit, overvoltages from direct lightning strike are also of two types. In the first case, lightning strikes a point on the tower or shield wire and the resulting current causes its potential to ground to rise significantly. When this difference in potential between lightning strike point and conductor exceeds the lightning impulse discharge voltage of the line’s insulation, flashover or breakdown of the air gap occurs. By contrast, should the absolute value of the tower or shield wire potential at the moment of lightning strike be higher than that of the conductor, there is back flashover.

In the second case, overvoltages generated by lightning bypassing the shield wire and striking the conductor are shielding failures.

Evaluating Hazards of Lightning on Transmission Lines

The performance of a system’s lightning protection can be measured using two indices – lightning withstand level and lightning related tripping rate. The former refers to the maximum amplitude of lightning current (in kA) that the line can withstand without leading to insulation flashover under direct lightning strike. The higher this is, the better will be the line’s lightning performance. Lightning related tripping rate, by contrast, refers to the number of trips on a line (per 100km/year) under standard conditions or after being converted to the equivalent of 40 lightning days per year. Lightning tripping rate is therefore a comprehensive index of a line’s lightning protection performance.

Since shield wire is installed along the entire length of a line of 110 kV and higher, it has been widely assumed that lightning related trippings on such lines were due mainly to back flashover and incidence of shielding failure was lower. Recent statistics from China demonstrate the opposite. In 2011, the lightning related tripping rate on lines operated by the State Grid Corporation was 865, of which 592 (68.4%) were due to shielding failure, 269 to back flashover and 4 to other causes. Moreover, trippings of 750 kV and ±500 kV lines were all due to shielding failure. Indeed, comparative rates of shielding failure as a proportion of all lightning trippings on 66 kV, 110 kV, 229 kV, 330 kV and 500 kV lines in China were 1.6%, 58.4%, 76.1%, 80%, 95.1%, respectively, i.e. the higher the voltage, the higher the percentage of shielding failure related trippings. In 2012, lightning related trippings in China numbered 741, of which 520 (70.2%) were due to shielding failure, 209 to back flashover and 12 to other factors. Again, the higher the voltage, the higher was the percentage of shielding failure trippings.

Measures to increase the lightning withstand level of a transmission line and reduce lightning related hazards include: decreasing grounding resistance of towers; adjusting protection angle of the shield wire; using a lightning protection differentiation design for double circuit towers; installing additional lightning rods or line surge arresters (TLSAs) in areas with highest risk. Moreover, using lightning monitoring equipment helps in locating and replacing damaged insulators in a timely manner.

Dealing with back flashover has a distinctly different focus than for shielding failures. Decreasing ground resistance of towers is highly effective in preventing back flash related trippings; the main measures to prevent shielding failure are decreasing the wire’s protection angle or installing more lightning rods and line arresters.

The lightning withstand levels of transmission lines in the event of shielding failure are low. Calculated lightning shielding failure withstand levels for 110 kV, 220 kV, 330 kV and 500 kV transmission lines are only 7 kA, 12 kA, 16 kA and 27 kA respectively. In most of China, the probability of a lightning current over 20 kA is 59% and that over 50 kA is 27%. As such, occurrences of shielding failure will typically lead to line tripping and therefore demand that more attention be given to preventing them, especially on higher voltage lines.

Line arresters can significantly reduce lightning related trippings and their practical application has proven remarkably effective. If their cost can be reduced, they would probably be used more widely.

At the same time, 1000 kV UHVAC lines have an additional shield wire installed above the center phase conductor on critical sections to increase withstand performance and reduce lightning shielding failures.

Scene From China

It’s also important to increase the lightning withstand level of insulators. Since composite insulators offer superior withstand to pollution flashover, their insulation distances are often shorter than for equivalent porcelain or glass strings. Moreover, a composite insulator’s grading ring’s lightning withstand level may be relatively low. Given that the lightning withstand level of 110 kV lines is already low, the net negative impact on lightning performance when composite insulators are specified can be significant. Chinese power companies have already noticed this problem and their specification of composite insulator lengths is no longer based solely on pollution flashover performance and takes into account the effect on lightning withstand level. Still, as long as the air gap between the grading rings of composite insulators is not less than that on porcelain or glass strings, the line’s lightning withstand level will not decrease.

Finally, any consideration of lightning related hazards to lines must consider not only lightning related tripping rate but also failure rate due to lightning. If reclosing is successful on a line struck by lightning, no interruption in service will result. Given this, installing parallel connected protection gaps on insulator strings helps keep the arc from a lightning flashover far enough from the surface to avoid interruption due to damaged insulators. The Electric Power Research Institute in Beijing has conducted extensive research on such gaps and this measure should be widely promoted in China as well as elsewhere.

Prof. Guan Zhicheng Tsinghua University, Shenzhen Campus guanzc@tsinghua.edu.cn 22

Product Range: Transmission Line Type: AC: 10 kV~1000 kV DC: 25 kV~1100 kV

Line Post Type: 10 kV~400 kV Station Post Type: 10 kV~230 kV

Station class arresters since 1980 have typically been based around various types of metal oxide varistors. Despite having been designed to withstand most power system surges, it’s still possible for these varistors to become overloaded or fail such that fault current will flow through the arrester’s core. But since this is an expected event over its life cycle, the arrester is designed to both conduct and withstand fault currents at significant levels and not fracture violently risking collateral damage to nearby equipment. In this column, I hope to advise readers how best to assess if their arresters are appropriately rated for the system to which they are being applied. Below are some basic concepts that must be noted beforehand:

Do Your Station Class Arresters Have Adequate Fault Current Withstand Capability?

1. Fault (or short circuit) current is not the same as lightning current but is rather the current that flows from the power system when there is a short circuit. If the short circuit is within the arrester, this is understood to be an arrester-conducting fault current.

2. Violent rupture of an arrester is not caused by lightning current but rather by gases created when fault current heats up and vaporizes materials within the housing. Lightning current can lead to fault current but, without fault current flowing, there would not be any rupture.

3. Arresters are designed to conduct lightning current multiple times over their service life and these operations do not generally result in failure or fault. If an arrester operation does lead to failure, it will likely become a very low impedance path to ground and result in fault current flowing.

4. Arresters are designed to conduct fault current one time at their end of life. Applying fault current to an arrester a second time, such as from reclosing a circuit, can result in its violent rupture.

5. When selecting an arrester for a new or old application, it’s important to check that the fault current available at the arrester’s location is compatible with the unit’s fault current withstand rating.

As with all critical arrester parameters, fault current withstand capability is verified by testing at high current laboratories (see Dr. Smeets’ Column in INMR Issue 104). The relevant test procedure is found in IEEE C62.11 as well as IEC 60099-4, last modified in 2009, and is one of the few arrester tests that are identical in both standards.

IEEE’s Application Guide C62.22 still makes no reference to assessing fault current withstand so, as part of the Working Group that updates this guide, I hope to address this omission. While IEC 60099-5 does cover fault current assessment as part of the arrester selection procedure, I feel that this guide needs to better clarify how to do it correctly.

To determine fault current withstand capability, the fundamental test procedure is to apply the specified fault current to the arrester, which is expected to take this current for the full duration without interruption. The arrester is permitted to fracture but parts cannot be expelled beyond a limit equal to a circular area with a diameter twice the unit’s height. Per both standards, all arresters must have a fault current or short circuit rating and must be tested at two or four current levels depending on their design.

Fault current withstand curve.

Arrester design A has substantial internal air volume while design B has no measureable internal air volume. The test currents most often used for design A station class arresters are: 63 kA, 25 kA and 12 kArms for 12 cycles and at 600 amps +/- 200 ampsrms for 1 second (peak levels are plotted in the graphic.) Design B would typically be tested only at 63 kA and 600 amps.

To help understand how to apply fault current test results, I have developed a chart referred to as ‘Arrester Fault Current Withstand Curve’ showing the currents at which arresters are tested in the standards. The graphic is divided into three zones – Elevated Risk of Violent Rupture, Low Risk of Violent Rupture and Unknown – that describe what might happen if an arrester is subjected to these fault current amplitudes and durations. End of life performance with respect to fault current can then be assessed by plotting expected fault current at the arrester’s location and analyzing the zones this passes through.

Woodworth on Arresters

The plot here shows the fault current experienced by two arresters – both certified to 63 kA short circuit capability – from two different circuits and compares how the arresters react to fault current based on location. The green curve represents the fault current experienced by a switchyard arrester with 17,000 peak amps that is interrupted in .0028 seconds; the blue line represents the fault current an arrester at a generator substation might conduct during an end of life event. The current is not interrupted in this case but decays as the generator slows down. For the switchyard arrester, it seems that the arrester should perform well and not rupture from the event. By contrast, the performance of the arrester at the generator station would be questionable and perhaps its supplier should be consulted since it passes into the zone of high rupture risk.

As seen from these two examples, what may seem a simple ‘yes or no’ proposition can at times prove more complex.

Jonathan Woodworth Jonathan.Woodworth@ArresterWorks.com 24

Outdoor terminations serve the important function of connecting cables to overhead lines, switchgear and transformers. In this role, they are exposed to a variety of service stresses, from electrical to mechanical, from weather to occasional interaction with wildlife.

First generation terminations used for XLPE and EPR type power cables were fluid-filled (mainly with insulating oil) and with a silicone or EPDM stress cone included inside their porcelain housings. This basic design concept has been used successfully now for many years and demonstrated excellent resistance against problems such as tracking, ageing due to UV radiation and damage due to bird pecking. However, in the event of catastrophic breakdown, violent shattering of the porcelain presents a serious risk to public safety. Another issue has been costly collateral damage to nearby equipment as a result of an explosive failure.

New Termination Designs for Polymeric HV Cables

For these reasons, replacement of porcelain housings with composite insulators on outdoor terminations has begun to accelerate over recent years. Such a design carries no risk of explosion and also brings several additional benefits such as superior pollution behavior, lower weight and greater resistance to seismic events. Still, one of the residual problems with this revised design concept, apart from higher cost, has been the continued presence of the internal insulating oil, which in the case of failure can leak and damage the environment.

A new generation of terminations has recently been developed that covers applications up to about 150 kV and can effectively overcome this environmental danger while offering other advantages as well (see article on page 86).

Such ‘dry-type’ terminations, first presented at Jicable 2011 and again at the 2013 INMR WORLD CONGRESS in Vancouver, are based on either pre-fabricated silicone slip-on or cold shrink technologies. The benefits for both types include easier installation and no need for any internal oil. But since this type of termination is not self-standing, some applications may require an additional support insulator. Extension of slip-on and cold shrink cable terminations to the EHV level may prove difficult but will certainly not be impossible.

Example of slip-on termination for 110 kV XLPE cable in combination with support insulator.

Focus On Cable Accessories

Replacing oil with gel in standard HV outdoor composite terminations promises to be an excellent solution to improve installation technology and to make these components more environmentally friendly.

Another novel design of termination for HV polymeric cables is the self-standing outdoor type, also presented at the 2013 INMR WORLD CONGRESS. Basically, this consists of a stress control cone as well as a composite hollow insulator filled with silicone gel. This gel compound allows for easier installation as one of its key benefits, with no requirement for special devices for injecting and de-foaming. Moreover, the electrical properties of the gel are said to be equal to or better than silicone oil (i.e. higher insulation performance, higher volume resistivity and lower power factor).

Indeed, electrical performance and long-term tests at 138 kV and 230 kV were carried out on such terminations in accordance with IEC 60840 and 62067. Results of the partial discharge test, AC voltage test, impulse test and heat cycle test were good at both voltages.

Replacing oil with gel in standard HV outdoor composite terminations promises to be an excellent solution to improve installation technology as well as to make these components more environmentally friendly. Perhaps the most important issue governing the future of this application at EHV – and possibly its extension to cable joints – will be the new gel’s long-term behavior.

Professor Klaus-Dieter Haim University of Applied Sciences Zittau/Görlitz, Germany KDHaim@hs-zigr.de 26

INSULATORS

Silicone Insulators Lessons from 30 Years Experience in China

30

S

ilicone composite insulators have undergone an extraordinary evolution in China since first being put into service on the country’s railways back in 1983. For example, one of the formidable challenges initially was overcoming the conservatism of the country’s power industry, which demanded years of operating experience before even considering this technology for use.

As it turned out, acceptance of silicone insulators across the country proved far more rapid and widespread than most could have guessed at the start. Indeed, the popularity

of composite insulators grew so quickly that they came to surpass both glass and porcelain as the dominant insulator technology used on domestic power lines. At the same time, China also became the first country to also broadly adopt composite line insulators on its UHVAC and UHVDC projects.

This article, contributed by Professor Liang Xidong and Mr. Wang Jiafu of the Department of Electrical Engineering at Tsinghua University in Beijing, reviews the development of composite insulators in China and proposes that this occurred over distinct phases, each with its own characteristics and decisionmaking drivers. Also discussed are factors such as growth in the numbers of units installed, progress in production technologies and causes of failures.

Photos: INMR Š

31

China’s Power Grid: 1983-2013

The rate of growth of HVDC has been equally striking, with 11 such lines now operating at ±400 kV, ±500 kV and ±660 kV. Their various lengths total over 10,500 km with combined rated power of 30.4 GW.

Developments in insulator technology in any country often link closely to development of the local power sector.

In the case of China, the electricity supply industry in that country underwent remarkable growth over the past 30 years and particularly so since the turn of the century.

On the UHV front, China began operating a 640 km long 1000 kV demonstration project in January 2009. Also, three ±800 kV UHVDC bi-pole projects, with combined line length of 5370 km and 18.6 GW total capacity, were recently put into operation – two of which have been in service for 3 years now without major incident. A second double circuit 1000 kV overhead line running 656 km is also being commissioned.

By the end of 2012, for example, installed power generation capacity reached 1145 GW while annual electricity generation had climbed to 4977 TWh. These figures represent multiples of 15 and 14 times the corresponding data from 1985.

Estimated generation capacity and annual electricity generation by the end of 2013 were 1230 GW and 5300 TWh respectively. Similarly, the total length of China’s overhead lines of 220 kV and above grew to about 500,000 km by the close of 2012 – about 10 times that in 1985.

Network expansion, however, is but one measure of how rapidly the Chinese power sector evolved since 1983. Another is the speed with which the grid climbed to ever-higher voltage levels. Before 1981, for example, the highest voltage lines in China operated at 330 kV. By contrast, today the length of 500 kV AC lines alone is some 140,000 km while by 2011 there were already over 10,000 km of 750 kV lines in service.

Composite suspension insulators have found broad application on Chinese EHV, UHV and HVDC lines.

Such a scale of new grid construction has naturally required huge volumes of insulators but also with much more demanding performance requirements. For example, rated mechanical strength of porcelain insulators on 110 kV and 220 kV lines in China has traditionally been from 70 kN to 120 kN. However, the majority of insulators now used at 500 kV are rated 160 kN to 210 kN while the mechanical strength of insulators for UHVAC and UHVDC is from 300-400 kN. Indeed, the success of composite insulators in meeting such requirements was among the

Photos: INMR ©

Acceptance of silicone insulators in China proved far more rapid and widespread than anyone could have imagined at the start.

Fig, 1: Installed power generation capacity and annual electricity generation in China.

32

Historical pollution flashover problems in China created a unique opportunity for silicone insulators to move rapidly from the testing phase in research laboratories to broad application in the field. factors that helped propel their growth – especially over the past 15 years (see Fig. 3).

Composite Insulators: Opportunities & Challenges

The rapid development of composite

insulators in China over the past 30 years took place over different phases, each marked by unique opportunities to overcome serious problems or find practical solutions to meet the needs of a fast expanding grid. One could even say that it might

have proven impossible to meet all the demands of rapid network growth given the domestic experience with porcelain and glass insulators. In this regard, the success of composite insulators in helping overcome network challenges accelerated their acceptance considerably. Fig. 4 illustrates three distinct ‘opportunity periods’ for silicone insulators in the case of China.

A. First Opportunity Period: Pollution Flashovers The first opportunity period for silicone insulators in China appeared during the late 1980s and early 1990s and could aptly be termed the ‘opportunity from pollution’.

China’s first 500 kV transmission line was put into service in 1981.

Local 500 kV grids in most provinces as well as the backbones for North China, Northeast China, Central China and East China then began to take shape over the following 10 years.

Photo: INMR ©

Fig. 2: Lengths of overhead transmission lines in China.

Silicone suspension insulators on ±400 kV line.

However, a series of severe pollution flashover events soon occurred, leading to widespread disruptions and outages. In January 1989, for example, five of the six 500 kV lines making up the Eastern China Grid experienced pollution flashovers during dense fog, with 12 trips recorded. Then, in December that year, two of these 500 kV lines again experienced pollution flashovers, while conductors on the 500 kV DouDu line dropped to the ground due to mechanical failure of porcelain cap & pin strings. This line remained out of service for almost a week.

In February 1990, dense fog covered North China for over a week and all four 500 kV lines on that grid lost service due to pollution flashovers on their porcelain insulators. The JingJin-Tang Grid surrounding Beijing and Tianjin also had widespread disruptions while conductors on the 500 kV Fang-Bei Line dropped due 33

Photo: INMR ©

Typical 1000 kV V-string on suspension tower.

to failure of porcelain cap & pin insulators. Five days were needed to restore power. Similarly, the 500 kV DA-Fang I Line remained out of service for 60 hours due to pollution flashovers.

All these problems led China’s Ministry of the Electric Power Industry to mandate annual conferences from 1990 to 1992 on the topic of finding solutions to combat pollution. Pollution flashover events across China had attracted attention at the highest levels of government amid worries about the safe operation of the country’s 500 kV backbone under polluted fog conditions.

Fig. 3: Quantity of silicone insulators in use in China.

34

Counteracting the impact of pollution on insulators at the time involved mostly traditional methods such as adding more discs to strings or increasing total creepage. Also, greater numbers of outages were scheduled to allow manual cleaning

the most attractive in China. In fact, silicone insulators soon became locally known as ‘anti-pollution insulators’ and alternative polymeric shed materials (e.g. EPDM or epoxy) were deemed unacceptable.

Photo courtesy of China Southern Power Grid

Fig. 4: Illustration of opportunity and development stages of composite insulators against backdrop of growing electricity supply industry in China.

Impact of heavy pollution on 500 kV glass string.

insulators became one of the goals in China’s 7th five-year plan.

Two universities – Tsinghua and the Wuhan Institute of Hydraulic & Electrical Engineering – were designated as the focal points for this research and a range of 110 kV to 500 kV silicone insulators was developed and successfully tested. The technology for their production on a large scale was then disseminated to domestic insulator manufacturers, including information on ideal composition of shed material, best technique to connect end-fittings, etc.

This experience demonstrated that historical pollution flashover of polluted insulators and, to a lesser problems had created a unique opportunity for silicone insulators extent, there was some application to move rapidly from the research of silicone grease to line insulators.

Still, these remedial measures alone phase to broad application in the field. Indeed, from among the many proved insufficient. advantages of composite insulators, it was their superior performance It was against this background that under pollution that initially proved local development of composite

To illustrate just how successful the changeover in insulator technology was, between 1971 and 1999 more than 4000 pollution flashovers were recorded on Chinese overhead lines and more than 2000 pollution flashovers reported at substations from 35 kV to 500 kV AC system.

After broad application of silicone insulators on the network, however, only 34 pollution flashovers occurred in 2008 on the State Grid’s (SGCC) network of lines of 66 kV and above and only 11 outages resulted. These days, pollution flashover ranks last among all the factors that cause overhead line trips in China.

Incidentally, the rapid rise in application of silicone insulators in China starting around 1990 came in spite of a price disadvantage of some two to three times that of porcelain.

Still, local power utilities were willing to pay this premium due to perceived superior performance. Moreover, there was little concern about the risk of premature ageing given the widespread view that silicone rubber performed better in this respect than other all organic materials.

B. Second Opportunity Period: HVDC The second opportunity period in China for silicone insulators appeared in the years just before and after the turn of the century and could properly be termed the ‘opportunity from HVDC’.

Some 9000 km of 500 kV AC and ±500 kV HVDC lines were designed around the huge Three Gorges Project. The issue of pollution accumulation on DC insulators is much more severe than for AC and

Between 1971 and 1999 there were more than 4000 pollution flashovers recorded on Chinese overhead lines and more than 2000 pollution flashovers reported at substations from 35 kV to 500 kV AC system.

36

design institutes and power utilities to adopt greater numbers of silicone insulators for their HVDC lines.

Photo: INMR ©

Another factor favoring silicone insulators was the relatively high cost of porcelain as well as glass cap & pin insulators for these applications due to the need to prevent corrosion on caps and pins under DC field. In fact, unlike the case for HVAC, the cost of HVDC silicone insulators in China was only about 1/3 that of equivalent ceramic string insulators.

Unlike the case for HVAC, the cost of HVDC silicone insulators in China has been only about a third that of equivalent ceramic string insulators. therefore problems with pollution flashover are usually much worse (see article on page 46). Indeed reported pollution flashovers on the Pacific Intertie in the U.S. provided valuable insight into what could happen on Chinese HVDC lines as well.

For example, the 1200 MW ±500 kV Gezhouba to Shanghai (GE-Shang) Line, which runs 1045 km, was put into bi-pole operation in 1990. Even though estimated pollution severity was carefully considered during line design, there were still many problems from pollution affecting porcelain insulators. Therefore, particular attention was devoted to insulation on several subsequent ±500 kV HVDC lines from Three Gorges to East and South China, including the Long-Zheng, SanGuang and Jiang-Cheng lines. Each was designed to transmit 3000 MW over about 1000 km.

Work on drafting the electric power industry standard DL/T 810-2002 “Technical Specification for ±500 kV D.C. long rod composite insulators” began in 1998 and it was formally issued in 2002 and revised in 2012.

This was the first technical standard for HVDC composite insulators in China and proved helpful in terms of quality control for HVDC as well as HVAC silicone insulators. It also helped toward standardization of 38

UHV composite insulators a few years later.

The main concern by design institutes, however, was what would happen if silicone insulators on these lines all aged prematurely and needed to be replaced by either porcelain or glass types. This could prove an enormous challenge given the smaller tower dimensions made possible by shorter, superior pollution-performing silicone insulators. A related worry was the risk that if silicone insulators began to fail in large numbers, there would not be sufficient lead-time to order huge quantities of porcelain or glass replacements.

The new standard, along with successful operating experience for composite insulators on HVAC systems as well as improved shed materials, new stress-corrosion free The response to such concerns FRP rods and better compression end was that, in the event of such a fittings, all helped convince Chinese problem, new silicone replacement insulators could be ordered relatively quickly. Basically, the feeling among experts was that, due to the history of pollution flashover problems on HVDC lines in China, the era of porcelain insulators for this application was over.

Another factor that made silicone insulators such a good choice for HVDC was that any ageing mechanisms would never be uniform along the entire line route that crosses widely different service conditions.

Yet another factor that made silicone insulators seem a good choice for HVDC was that ageing mechanisms would never be uniform along the entire route given that these lines run over 1000 km across widely different service conditions. Any ageing due to specific local environmental factors (such as UV radiation or environmental attack from sustained wetting and heavy pollution when hydrophobicity can be lost and surface discharges result) would therefore be mostly localized and never require widespread insulator replacement over a short time.

The application of silicone composite insulators in China made additional gains based on the large demand for

Production of 1000 kV and ± 800 kV silicone insulators.

them coming from The Three Gorges HVDC project and lines such as the ±500 kV Ge-Shang, Tian-Guang, Long-Zheng and San-Guang – all of which adopted them as the dominant technology. Total number of silicone insulators used on these four HVDC lines alone reached 9750 by the summer of 2004, more than the sum of all HVDC composite insulators used in all other countries at the time.

In 2005, when the State Grid Corporation of China (SGCC) and China Southern Power Grid (CSG) planned 1000 kV UHV AC and ±800 kV UHV DC projects, doubts were expressed both by manufacturers and power utilities worldwide, particularly relating to the demands placed upon UHV insulator strings. These concerns were based mostly on pollution but also on other aspects such as mechanical

Photos: INMR ©

C. Third Opportunity Period: UHV The third opportunity period for silicone insulators in China appeared

at the time of construction of the country’s first UHVAC and UHVDC projects between 2006 and 2010.

China’s first 1000 kV UHVAC line has operated since January 2009.

performance and electric field stress control. Nevertheless, China’s first 1000 kV UHV AC and ±800kV UHV DC projects were approved by the National Development & Reform Commission in 2006.

Based on operating experience and results from artificial pollution testing with porcelain and glass insulators, string length would have to be about 10.5 m for a 1000 kV AC system in lightly polluted areas.

In medium to heavily polluted areas, this length would have to increase to between 12 and 15 m, maybe even longer. In the case of ±800 kV UHV DC, string lengths of porcelain or glass insulators would have to be 12.7 to 15.6 m in medium to heavy pollution areas.

Service experience in China since 2004, involving large numbers of silicone insulators, eventually provided a basis for them to be selected for UHV lines as well.

This experience included excellent behavior under pollution, high mechanical tensile strength, longterm creep characteristics for optimized compression-type end fittings and boron-free FRP rods to prevent risk of stress corrosion or fracture. The acquisition cost advantage of composite insulators at these voltages provided even more support for their eventual selection for UHV projects.

39

Photos: INMR ©

Example of classical brittle fracture of rod, with smooth planar surface as well as fractured strands.

Example of insulator that experienced growing brittleness in silicone material, making it easier for sheds to break off.

Insulator equipped with extra wide sheds designed to limit flashovers due to bird ‘streamers’.

1. Brittle Fracture The first recorded brittle fracture of a silicone insulator in China occurred in December 1994 and involved a 220 kV unit that had been in service for only about 4 years. This was a serious accident since the conductor fell to the ground and the event forced researchers, manufacturers and power utilities to devote attention to a failure mode that was already being reported in other countries. Another brittle fracture happened in 1998, this time with a 500 kV insulator and 4 additional brittle fractures occurred on 500 kV insulators in 1999, half from foreign suppliers. Then, in 2001, 6 more brittle

Challenges Ahead for Silicone Insulators

In spite of the success of silicone insulators in meeting the challenges associated with each of the ‘areas of opportunity’ for them in China, they are still a comparatively new technology. In this regard, there is an ever-growing body of knowledge on failures or unexpected damage. These represent the main challenges ahead in order to assure continued and even wider application in the future.

40

Photo courtesy of Tsinghua University

Notwithstanding these concerns, after much debate and testing a total of 7200 silicone insulators were applied to the country’s first 1000 kV UHVAC demonstration project. This line, which runs 640 km, has been in operation now for over 5 years. On the UHVDC side, the initial three ±800 kV projects included the YunGuang Line (1417 km/5000 MW), the Xiang-Shang Line (1895 km/6400 MW) and the Jin-Su Line (2059 km/7200 MW). These lines have been in operation now since June 2010, July 2010 and December 2012 respectively and more than two-thirds of all their insulator strings are silicone rubber type.

Photos: INMR ©

As in the case of UHV, there were initial concerns regarding possible premature ageing of silicone insulators on lines of such strategic importance to China. For example, some experts warned that this relatively new insulator technology should not be applied for projects that might challenge their technical limits.

Examples of sheds damaged by bird pecking.

In spite of the growth in application of composite insulators, many Chinese power companies still prefer glass or porcelain insulators in tension due to concerns about possible damage by maintenance crews stepping across silicone sheds to access conductor.

Photos: INMR Š

the brittle fracture process and better understand what was causing it.

Subsequently, boron-free FRP rods were developed and a stress corrosion test method and acceptance criteria became part of the standard, DL/T 810-2002. These days, the application of improved rods as well as the new technical specification for HVDC composite insulators has successfully prevented further brittle fracture incidents and resolved most concerns among users.

Decay-like fractures of composite insulator rods. fractures were reported on insulators at 500 kV, the majority of which had been imported.

Tsinghua University first began to devote attention to brittle fracture after 1994 and this work involved developing a test method to replicate

2. Broken/Damaged Sheds During the mid 1990s, some silicone insulators were found to exhibit ageing of their shed materials. For example, in some cases shed hardness increased so much that they became brittle and easy to break off or tear. Apart from creating misgivings among certain power companies, this experience also led to discussions within the industry about how to improve the composition of the silicone material in order to achieve the best balance between hydrophobicity transfer properties and tracking resistance.

Today, most manufacturers seem to have been able to resolve this problem.

A simulation experiment by Tsinghua University eventually made it clear that these types of flashover events were most often due to excrement from large birds. With this knowledge, some insulator manufacturers began offering units with extra wide sheds, intended to prevent or at least minimize this problem.

4. Decreasing Mechanical Strength During the 1990s, many composite insulators with circular wedge type fittings experienced a significant drop in mechanical strength after only a few years in service. For example, one sampling test for this problem found that only about 1/3 of these insulators could pass the 1 minute SML test. This test involved 44 units of 110 kV and 220 kV insulators with 100 kN SML that had been in operation for between 2 and 5 years.

Photo: INMR Š

Mechanical failing load for most of these units was found to be only 60% to 80% of the SML, with the worst case at only 50% SML. A second sample test involving 9 insulators with SMLs from 70 to 100 kN showed that only one of them passed the 1 min test, while the remainder failed at about 80% SML. Again, these circular wedge insulators had only been in service for only 1 to 5 years.

HV testing of composite insulators to assure compliance with domestic standards.

42

3. Unexplained Flashovers During the mid to late 1990s, many flashovers were reported on composite insulators but without any obvious cause. These flashovers, which affected mainly 110 kV and 220 kV insulators and which typically occurred in early morning, were usually re-closed successfully.

The electrical and mechanical performance of the insulators affected was tested and found to be normal, as was the condition of the shed material. Still, without any obvious cause, there was no means to apply countermeasures and this created uncertainty among those utilities that were considering broader application of silicone insulators.

Between 1997 and 1999, Tsinghua University studied crimped on end fittings and later disseminated a

Photos: INMR ©

Testing of shed material on insulators retrieved from the field after several years in service.

High altitude testing of composite insulators at UHV test station in Kunming.

superior crimping process technique to several insulator manufacturers.

Mechanical strength reached 750 MPa and 700 MPa for Φ18 mm and Φ24mm FRP core rods respectively.

Standard deviation of mechanical failing load was very small and mechanical creep characteristics were also excellent. For example, creep slope over the 600 day test was only -2.07% Mav per decade of log time in the load-time creep test, according to IEC 61109-1992 – a significant improvement over insulators with circular wedge type fittings.

Fortunately, timely development of optimized crimping helped dispel concerns by power companies just before the start of the ‘HVDC opportunity’ discussed earlier. Later, researchers at again co-operated with the domestic insulator industry to develop high SML (300 to 550 kN) composite insulators able to take advantage of the ‘UHV opportunity’.

5. Shed & Sheath Damage from Bird Pecking Damage to the sheds and sheaths of silicone insulators due to bird pecking has not been a widespread problem in China. Still, many such insulators installed on towers of UHV lines were discovered to be damaged even before the line was energized. Such damage was usually at the top of I-strings or the bottom of V strings. Some severely damaged insulators eventually had to be replaced before energization.

Although clearly not a threat to the future of this technology, bird pecking is one issue that the composite insulator industry still must resolve.

6. Decay-Like Fracture In recent years, a new type of failure mode of composite insulators has been discovered in China and elsewhere. Here, the failed rod resembles fracturing of dead wood,

Dye penetration test on FRP core rod.

without any of the smooth planar surfaces typical of classical brittle fractures. Chalking as well as separation of glass fiber and resin have also been reported on the rod.

Since 2008, 8 such failures have been reported in China at 500 kV and 5 of these had been supplied by foreign manufacturers.

This kind of failure has now been termed ‘decay-like fracture’ and some progress has already been made in the laboratory to understand its cause. Since such failures can occur on normal as well as on boronfree FRP rods, countermeasures employed to avoid brittle fracture are not effective. Rather, these fractures seem related to degradation and eventual failure of the rod-housing interface. Preventing their occurrence may therefore link closely to improving the bonding at the internal interface between the core rod and the silicone rubber housing.

43

Market Development Process for Composite Insulators in China Another way to look at the 30-year history of composite insulators in China could be to examine the various stages as they progressed from the R&D phase to becoming the dominant technology in use on the country’s transmission lines.

Each stage was marked by its own character and theme.

Stage 1: Research & Development This phase lasted about 10 years and saw composite insulators go mainly through laboratory testing by university researchers.

Required product attributes such as hydrophobicity, tracking and erosion resistance were all studied and then explained to engineers at the power utilities. Users were also educated about potential advantages when it came to superior flashover performance, high strength to weight ratio, easier transport and installation, etc. as a means to allow them to better appreciate what they could expect from composite versus traditional insulators.

were numerous reported accidents and failures. Batch type production became a common element of this stage as manufacturers tried to win customers through more advanced production lines. Cost control was another issue as suppliers tried to increase their competitive position in the market.

Stage 3: Quality & Standardization After 2000, composite insulators entered a stage of widespread application in China and were specified on several key 500 kV HVAC and ±500 kV HVDC projects.

Competition intensified, especially when it came to price and there was growing pressure to control

Photo: INMR ©

Stage 2: Promotion of Application From the early of 1990s to around 2000, growing numbers of composite insulators were installed on transmission lines but there

Composite insulators accounted for majority of new installations on Chinese transmission lines over last 5 years.

Composite insulators accounted for majority of new installations on Chinese transmission lines over last 5 years.

44

production costs. At the same time, there was a new emphasis on quality control so as to better ensure longterm performance. In fact, quality control and standardization became dominant themes in China after 2003.

Standards were viewed more as a way to screen and reject poor quality insulators than as a means to identify the best products since it was seen as difficult to predict an insulator’s long-term performance through only a few short-term tests. In this regard, the power industry standard DL/T 810-2002 introduced some basic material tests for insulators. For example, test methods and acceptance criteria were drafted in regard to shed material hydrophobicity transfer and recovery as well as tracking and erosion resistance under HVDC. Artificial pollution test methods for silicone insulators were also drafted as were test methods and acceptance criteria on required stress corrosion resistance of FRP rods.

Stage 4: Widespread Application as a Conventional Product After becoming the main type of insulator specified for Chinese UHV projects, composite insulators reached a phase of maturity where they came to be regarded as conventional products and no longer only as anti-pollution insulators.

Indeed, over the past 5 years composite insulator accounted for 43.7% of all new insulators installed in China – much higher than that of 26.4% for porcelain and 29.5% for glass insulators.

Looking to the future, the areas of application of composite insulator technology in China are now already expanding beyond mainly suspension insulators for transmission lines.

The most promising new areas in this regard include a variety of substation applications as well as for tension strings, which up to now are still mainly glass or porcelain.

Composite hollow core substation insulators in particular are attracting growing attention. Chinese line and substation insulator manufacturers will have to build their brands to best meet new market opportunities. 

You won’t want to miss this!

October 2015

SUNDAY

MONDAY

TUESDAY

WEDNESDAY THURSDAY

FRIDAY

SATURDAY

10 Mini-Conferences & 2 Workshops offering 120 practical technical presentations by world experts in design, application, monitoring & inspection of:

1 2 3 4 5 6 7 8 9 10 & substation 11 12 13 14 15 16 17 • line insulators surge arresters 18 19 20 21 22 23 24 • • cable joints & 25 26 27 28 29 30 31 terminations bushings •

Plus a PRODUCT & TECHNOLOGY Exhibition featuring industry leaders from across the globe.

Keep up-to-date on the expanding Conference Program by monitoring:

www.inmrworldcongress.com

INSULATORS

Latest Principles for Selection of DC Insulation S ince the dawn of transmission it became clear that pollution performance of insulators could have great impact on system reliability.

Insulators installed close to pollution sources soon become covered by a potentially conductive surface layer that, when wetted by events such as fog or dew, results in discharge activity, also known as dry band arcing. In severe cases, this process can culminate in flashover and power disruptions.

The IEC published its first guide (IEC 815) for selecting insulators with respect to pollution back in 1986. With this much-debated document, the now well-known concept of creepage distance first became standardized.

Although the guide quickly found

Photo: INMR ©

46

widespread use worldwide due to its relative simplicity, application experience did not always prove successful. In highly polluted environments, for example, simply adding more creepage was not a guarantee of satisfactory service performance. Rather, research demonstrated that a number of other insulator parameters, such as shape (profile) and material, also had to be taken into account.

By the early 1990s, it became clear that IEC 815 needed substantial revision to address such deficiencies and also to accommodate the newest insulator technologies. A large-scale project was therefore begun aimed at revising and updating this standard to cover AC and DC as well as all the different types of insulator technologies – ceramic, glass and polymer. In the process, IEC relied on Cigré working groups to provide sound technical support.

The work started in 1994 within SC 33 and the result was a reference document that would form the basis for future guidelines. This ‘review of current knowledge’ document proved sizeable, with 200 pages and 382 references, and was published in 2000 as Cigré Brochure 158.

Work subsequently continued with development of guidelines for selecting and dimensioning outdoor insulators for AC systems, published in 2008 as Cigré Brochure 361. The last task then became to develop similar guidelines for DC systems, finally issued in December 2012 as Cigré Brochure 518.

This article by industry expert Chris Engelbrecht, who was closely involved as Convener of the WG, reviews this latest brochure as well as some of its basic principles.

47

DC Versus AC Pollution Flashover Important differences between development of pollution flashover under DC versus AC energization can be summarized as follows:

Photo: INMR Š

1. DC energized insulators accumulate more pollution since, under low wind speeds, electrostatic attraction of pollution particles under unidirectional electric field dominates deposition by aerodynamic action. On AC energized insulation, by contrast, there is little to no attraction of pollutants by the alternating electric field. Indeed, research measurements suggest that the ratio of DC to AC pollution deposition in the same service environment can vary significantly and in some cases by as much as a factor of 10.

Detailed design process is necessary for DC systems because of great impact of either over-dimensioned or underperforming insulation.

Fig. 1: Relative AC versus DC pollution flashover stress (FOV) as function of pollution severity.

2. With the absence of voltage zeros, dry band arcing under DC is more likely to grow into flashover. This differs from the case under AC where dry band arcs must re-ignite

There is an urgent need to look at insulation design for HVDC systems at the very early stages of a project – certainly much sooner than customary for AC systems. after each voltage zero. DC dry band arcs are also more mobile and likely to leave the insulator surface to propagate through air.

This requires development of special DC insulator profiles to ensure the effectiveness of the creepage distance. Another result of this difference in flashover development is that the relative flashover strength of DC insulation deteriorates more, compared to AC insulation, for any given increase in pollution severity (illustrated conceptually in Fig. 1.)

Fig. 2: Comparison of indicative insulation distance requirements (glass and porcelain) for switching (blue) lightning (red) and pollution (green) for HVAC and HVDC systems.

48

3. Relative to other insulation stresses, the pollution performance of insulators is usually the parameter that dictates insulation design for HVDC systems. This

Fig. 2 also shows that extreme insulation lengths may be necessary in polluted areas, forcing system planners to reconsider any project and evaluate alternatives such as different line routes (i.e. avoiding polluted areas) or implementing cables or indoor switchyards – all in order to minimize the number of external insulation surfaces.

There is therefore an urgent need to look at insulation design for HVDC systems at the earliest stages of a project and much sooner than would be customary for AC systems.

It is also much more important to follow a detailed design process because of the potentially huge impact of either over-dimensioned or underperforming insulation.

Holistic Flow Chart Methodology

In the CigrĂŠ guidelines, designers are encouraged to follow an exhaustive approach with the aim of minimizing uncertainties in input data and the resulting impact on final design.

Dimensioning principles are introduced based on a flow chart (Fig. 3) to provide a holistic overview and context for each activity required.

Overall design strategy is shown in the vertical column of numbered blocks on the left. A number of alternatives for obtaining the relevant information are then presented such

Photo: INMR Š

differs from AC systems where insulation distances are typically determined by required lightning or switching performance of the line or substation (illustrated in Fig. 2 for EVH and UHV system voltages).

Research suggests that the ratio of DC to AC pollution deposition in the same service environment can vary by as much as a factor of 10.

that the further right one moves the less certain the results. For example, when determining site pollution severity, information collected from existing DC lines will provide far more accurate results than a purely qualitative estimation.

The various activities in the flowchart can be summarized as follows: 1. Identify Candidate Insulators Due to the special demands, some insulator types are specifically optimized for HVDC applications in terms of insulating materials and shed profiles and there is limited choice available. Nevertheless, choices still need to be made in regard to type, insulating material and profile for insulators that will be utilized at any particular location. Initial selection of a candidate insulator is usually based

on a simplified preliminary site assessment. This choice could then be revised over time as more detailed information about site conditions becomes available.

2. Assess Environmental & System Stresses Ideally, pollution deposition measurements over several years are needed to provide accurate data on the nature and severity of pollution at any given site. In the case of DC applications, it is especially important these measurements be performed on insulators that are energized to a representative DC stress. Since this is not always feasible, other techniques may have to be employed to obtain this information.

For example, if there is data on performance of HVAC installations in the area, it may be possible to

Fig. 3: Insulator selection flow chart.

49

is then estimated from published information or based on performance data summarized in guidelines with correction factors applied to correlate test conditions with actual site conditions. Such data are then used to make a rough selection of insulator type, material and dimensions.

4. Verify Design This is the last step in the process whereby the chosen insulation design is evaluated either by comparison with past experience or by testing.

Photo: INMR ©

For critical installations such as converter stations there is need for detailed site severity assessment, e.g. by setting up experimental stations at representative locations to obtain estimate of long-term pollution accumulation. ‘translate’ this to HVDC. But such a methodology is purely indicative and the designer must still make assumptions about differences in pollution accumulation on DC versus AC energized insulators. It is also possible to use a general environmental assessment to identify a comparable environment in a different location where an HVDC installation is already in operation.

Data from this installation could then be useful in insulation design and selection for the new project.

3. Determine Insulator Characteristics & Dimensions The most accurate way to select insulators for any new installation is to directly determine risk of flashover based on service experience of DC lines and substations in the same area or having similar environmental conditions. Flashover risk can also be obtained utilizing test stations where performance of a range of pre-selected insulators is monitored under DC voltage at locations considered representative of the new line and station corridor.

Where there is previous experience with DC lines in the same area, excellent data on insulator performance is available on which to base preliminary design. If there is a lead-time of a year or more, good data can also be obtained from installing energized insulators at field stations sited at representative locations along the line as well as at 50

Pollution deposition measurements over several years are needed to provide accurate data on the nature and severity of pollution at any given site.

Simplified Dimensioning Process

Besides the detailed dimensioning method discussed above, the Cigré WG also proposed a simplified method with the basic intent to: • provide useful orientation at the project’s start and identify a range of preliminary solutions; • analyze outage performance and the adequacy of insulation solutions on existing systems.

It is important to stress, however, that this simplified method has serious limitations which may result in either over or under-dimensioned insulation and is therefore not considered accurate enough for the final design process. On the other hand, it does provide insight into the various parameters that must be taken into account when dimensioning insulators for HVDC with respect to pollution.

The first part deals with determining site pollution severity, which is the equivalent value of ESDD at a 0.1 mg/cm2 reference value of NSDD. the converter station. Insulators at such field stations must be energized The second step is to adjust this value individually for each type of to representative stresses to take insulator being considered so that the account of the significant influence required USCD can be determined. of electrostatic field on pollution accumulation.

Determining Site DC Severity

Instead of determining risk of flashover directly, it is also possible to follow a simplified deterministic approach during design. In this simplified method, pollution stress (i.e. maximum pollution on the insulator) is determined from measurements and through site condition studies. Insulator strength

An overview of the process to determine DC severity is presented in Fig. 4. The aim of such a site assessment is to obtain an accurate picture of contamination severity of the area concerned based on data collected over a relatively long period. Initial assessment is usually based on:

Fig. 4: Part 1 of simplified dimensioning process – Determining site DC severity.

1. Collected performance data on existing lines or substations, preferably DC energized, although AC data could also be useful; 2. Identification of type (i.e. A or B, as defined in IEC 60815-1), and composition of pollutants (i.e. type of salts, non-soluble deposits etc.); 3. Measurement of quantity of pollution present; 4. Characterization of climate, specifically if there is a prolonged dry season; 5. Assessment of geographic, topological and geological features to identify possible contamination sources, and; 6. Survey of present and foreseeable future pollution sources and land use.

insulators. In such a case, it becomes necessary to estimate the contribution of electrostatic field on pollution accumulation on energized DC insulators. This is done with the DC/AC accumulation factor Kp (applicable to both ESDD and NSDD values). This factor can vary between 1 and 10 but, for the simplified dimensioning process, values between 1 and 3 are more typical.

Once site severity measurements become available, the maximum value of the average ESDDs measured on insulators is converted to an equivalent laboratory test severity. With this correction, it is recognized that artificial testing differs from natural pollution in a number of important aspects, namely:

usually subjects an insulator to a pollution layer with a non-soluble deposit density (NSDD) of 0.1 mg/cm2. In service, NSDD levels typically vary from 0.01 to 10 mg/cm2. The measured ESDD is normalized to an NSDD value of 0.1 mg/cm2 with a factor, Kn. These corrections result in an estimate of equivalent site severity at an NSDD=0.1 mg/cm2, defined as the DC site severity.

Determining Required Creepage Distance

The required insulator dimensions (notably creepage distance) are determined from available service experience or test results. If such data is not available, representative For critical installations such tests can be performed on candidate as converter stations, the above • Type of salt insulator types to determine their information may not prove Laboratory testing is generally statistical flashover properties. accurate enough and therefore performed with marine salt (NaCl) The term ‘representative test’ is there is a need for a more detailed whereas natural pollution layers understood to mean any laboratory assessment. Ideally, this includes often contain less soluble salts test designed to duplicate natural setting up experimental stations at such as gypsum (CaSO4). At contamination conditions as closely representative locations equipped present, there is still no widely as possible in terms of 1. pollution with selected different DC energized applicable methodology to quantify severity (i.e. ESDD and NSDD), 2. insulators so as to obtain an estimate this effect other than performing its composition (i.e. type of salt as of long-term pollution accumulation. specific flashover testing on well as non-soluble components), 3. insulators with natural pollution, uniformity of the deposit and 4. type An alternative approach, but so assume Kc=1. of wetting conditions. with increased uncertainty, is to base site severity assessment • Amount of non-soluble material Alternatively, standard laboratory on measurements taken on present in the pollution layer pollution test results can also AC energized or non-energized The standardized laboratory test be utilized but then a number of 52

20 YEARS OF SERVICE FOR INSULATOR FITTINGS Leading World Manufacturer of Insulator Fittings

HENAN JINGWEI ELECTRIC POWER TECHNOLOGY CO. LTD Baisha Industrial Park, Zhongmu, Zhengzhou, Henan, China P.C. 451464 Tel: + 86-371-62366506 Fax: + 86-371-62366505 Email: zzjw@zzjingwei.com Website: www.zzjingwei.com

• Best Delivery • Best Quality • Best Price • Best Service

Fig. 5: Part 2 of simplified dimensioning process – Determining required USCD. (This part of process performed separately for each insulator type).

be significantly higher than for uniformly polluted ones. This effect is corrected for with a factor Kcur.

• Insulator diameter Larger diameter insulators collect less contamination than small diameter ones. This effect is corrected for with a factor Kd. • Statistical considerations This correction factor, Ks, is chosen to obtain a sufficiently low risk of flashover.

Fig. 6: Median DC flashover characteristics (worst case) for hydrophobicity transfer (HTM) and non-hydrophobicity transfer material insulators as determined from solid layer laboratory test results.

Photo: INMR ©

adjustments to the DC severity become necessary. These adjustments depend on specific insulator type and are therefore done separately for each different insulator that will be utilized (see Fig. 5).

54

Briefly the adjustments can be described as follows: Non-uniform pollution deposition on ± 500 kV DC wall bushing can impact flashover voltage.

• Non-uniformity of contamination layer The flashover voltage of nonuniformly polluted insulators can

Application of these various correction factors results in a design DC severity that corresponds to the pollution severity at which representative laboratory tests can then be performed. At this stage, it is also possible to make a first estimate of leakage distance (unified specific creepage distance: USCD) required for the project, based on the graphs in Fig. 6. Two graphs are shown, one for hydrophobicity transfer materials (HTM) such as silicone rubber and the other for non-HTM insulators such as glass or porcelain.

These graphs are valid for line as well as substation insulators having relatively small diameter. Insulators with larger diameters generally have a lower flashover voltage than

those with smaller diameters and therefore require longer leakage distances. This effect is corrected for with factor Cd. At present, it is not deemed necessary to correct for the effect of diameter when it comes to hydrophobic insulators.

For installations at high altitude an additional correction factor can be considered to adjust leakage distance for the lower flashover voltage under low air density conditions. This is corrected for by a factor, Ca.

Verification

As confirmation of the applicability of the simplified method, the Cigré working group analyzed actual outage performances for various installations against the design curves presented. However, this proved difficult since detailed information on outage performance and corresponding pollution severities for HVDC systems in service are not generally published.

Still, some useful information was found and the results of this analysis are presented separately for substations (Fig. 7), lines (Fig. 8) and HTM insulators (Fig. 9). In these graphs a distinction is made as follows:

Fig. 7: DC stations: Collected field experience on non-HTM converter station insulators. Data normalized using the simplified method.

• Design values These data points represent minimum creepage distance values implemented on actual HVDC systems but for which performance data is not yet available.

• Good performance Indicates minimum creepage distance values implemented on actual HVDC systems for which positive service experience has been reported.

Fig. 8: DC overhead lines: Collected field experience on non-HTM insulators.

• Flashovers Indicates minimum creepage distance values implemented on actual HVDC systems for which negative service experience has been reported.

Data in these figures basically confirms the appropriateness of the simplified method recommended for obtaining a realistic first estimate of the insulation dimensions required at HVDC installations.  Fig. 9: HTM Insulators: Collected field experience.

55

INSULATORS

Resolving External Insulation Problems at HVDC Converter Stations

58

Along

with China’s rapid economic growth over the past three decades has come a major revamping of that country’s energy and load planning strategies. In particular, DC transmission has begun to take on an increasingly important role since 1990 due to its known advantages for HV & UHV lines running great distances (see article on page 46).

Starting with the ±500 kV of Ge-Nan project, first commissioned in 1989, and the ±800 kV Xiang-Shang and Chu-Sui lines completed in 2010, the framework of a huge DC backbone has now taken shape across the country. Indeed, by the end of 2012 China was already operating three ±800 kV lines, one ±660 kV line, ten ±500 kV lines, one ±400 kV line as well as three back-toback DC connection projects. More DC lines are now in planning, under construction or recently commissioned and some will operate as high as ±1100 kV. At the same time, the source of DC transmission technology used in China has undergone a complete changeover – from only foreign turnkey projects at the start, to foreign-dominated projects, to foreign cooperation on projects, to Chinese-dominated projects and finally to total Chinese self-reliance. For example, at ±500 kV DC, China has achieved almost full local production capability.

During development of these various DC projects, there has been considerable experience with failures of external insulation on certain apparatus. At the beginning, there were a relatively large number of such faults. But after applying suitable countermeasures, these have progressively been reduced to the current situation of mostly trouble-free operation.

This article, contributed by experts Wie Jie and Su Zhiyi of the Electric Power Research Institute in Beijing, reviews past research conducted by the State Grid and the China Southern Grid on external insulation faults at ±500 kV converter stations, using voltage dividers as the reference for experience with a range of equipment.

Photos: INMR ©

59

Overview of ±500 kV Point-to-Point DC Transmission in China

At the start of 2012, China already had a total of 10 operating ±500 kV point-to-point DC systems with a combined transmission capacity of some 28,800 MW (see Table 1).

Table 1: ± 500 kV DC Transmission Systems in China as of 2012 Transmission System Ge-Nan (before being modified to double circuit) Ge-Nan (after modification to double circuit) Long-Zheng Jiang-Cheng Yi-Hua Tian-Guang Gao-Zhao Xing-An De-Bao Yi-Mu Lin-Feng (second circuit of Three Gorges to Shanghai line)

Pole

Line Date of Capacity Length Commissioning (MW) (km)

I

1989.09

600

II

1990.08

600

I II I II I II I II I II I II I II I II I II I II

2011.04 2003.06 2004.06 2006.12 2000.12 2001.06 2004.09 2004.05 2007.12 2007.06 2010.04 2010.09 2011.05

1500 1500 1500 1500 1500 1500 1500 1500 900 900 1500 1500 1500 1500 1500 1500 1500 1500 1500 1500

1046

1118 860 941 1049 963 891 1194 574 908 977

Ge-Nan, China’s first long-distance DC transmission line, ran from the Gezhouba Station in Hubei to Nanqiao in Shanghai and its equipment and design came entirely from suppliers based in Europe.

About 2009 to 2010, this initial system was retired as part of a comprehensive retrofit lasting some 6 months and during which capacity was increased to 5.786 billion kWh. The Jiang-Cheng, Long-Zheng and Yi-Hua DC lines, transporting up to 9000 MW from the Three Gorges to Guangdong and Huadong, were designed jointly by Chinese and western-based suppliers. During 2010, the availability of these systems was high (respectively 17.23, 13.89 and 14.91 billion kWh) and together they accounted for 55% of the total capacity generated at that time at the Three Gorges.

Similarly, China Southern Power Grid’s Tian-Guang, Gao-Zhao and Xing-An DC transmission systems were designed jointly by Chinese and German suppliers and had a combined capacity of close to 7800 MW. By early 2010, the Tian-Guang system’s transmission capacity 60

Photo courtesy of China EPRI

Photos: INMR ©

Outdoor ±800 kV Suidong DC field (top) and indoor DC field at ±400 kV Geermu Converter Stations.

Fig. 2: DC voltage divider at Tianshengqiao Converter Station after adding booster sheds and coating with RTV material.

Photo courtesy of China EPRI

Fig. 1: DC voltage dividers and grading rings at Longquan (left), Jiangling (center) and Yidu Converter Stations.

Table 2: Key Parameters of DC Voltage Divider Bushings at Longquan, Jiangling and Yidu Converter Stations

62

Longquan

Jiangling

Yidu

Rated Voltage

±500 kV

±500 kV

±500 kV

Material

Complex

Complex

Complex

Shed form

Big-small

Big-small

Big-small

Insulation height (mm)

≥5720

5950

5950

Creepage distance (mm)

≥22789

22789

22789

Rod diameter (mm)

510

553

553

Structure height (mm)

6515

6545

6545

Shed width (mm)

70-50

70-50

70-50

Shed distance (mm)

60

60

60

Units of high voltage insulation (kV/m)

86

87

87

Creepage (mm/kV)

≥44

44

44

was up to 5 billion kWh while those of the Gao-Zhao and Xing-An lines were 12.25 and 13.38 billion kWh respectively.

The ±500 kV De-Bao, Yi-Mu and Lin-Feng DC systems, commissioned after 2010, were entirely designed in China and, aside from certain key technologies and equipment, everything was locally produced.

Up to now, these lines have had mostly stable operation, apart from individual discharge faults on the line side flat wave reactor, found to be due to a design defect.

External Insulation Faults Affecting ±500 kV DC Voltage Dividers Faults at JiangLing, LongQuan & YiDu Converter Stations Flashovers of external insulation on ±500 kV DC voltage dividers at Jiangling, Yidu and Longquan Converter Stations were reported late in 2004, early 2009 and early 2010. These flashovers occurred under similar conditions of more

Table 3: Configuration of External Insulation for Main ±500kV DC Voltage Dividers (PT) Converter Station

Voltage Class

Outdoor/ Indoor

Creepage Distance (mm/kV)

Housing

Operating Condition

Gezhouba

±500 kV

Outdoor

≥40

Porcelain

Experienced discharges

Coated with RTV silicone

Current Status

±500 kV

Outdoor

≥40

Porcelain

Experienceddischarges

Coated with RTV silicone and with booster sheds added

±500 kV

Outdoor

≥40

Porcelain

Experienceddischarges

Coated with RTV silicone and with booster sheds added

Longquan

±500 kV

Outdoor

46

Composite material

Flashed over

Coated with RTV silicone

Zhengping

±500 kV

Indoor

19

Composite material

No problems

Coated with RTV silicone

Jiangling

±500 kV

Outdoor

46

Composite material

Flashed over

Covered with RTV silicone

Yidu

±500 kV

Outdoor

46

Composite material

Experienced discharges

Covered with RTV silicone

Huaxin

±500 kV

Outdoor

46

Composite material

No problems

Covered with RTV silicone

Deyang

±500 kV

Outdoor

59

Composite material

No problems

No change

Baoji

±500 kV

Outdoor

59

Composite material

No problems

No change

Photo courtesy of China EPRI

Nanqiao Tianshengqiao

(a) Gezhouba

(b) Nanqiao

(c) Longquan/ Jiangling

(d) Zhengping

(e) Yidu/Huaxin

(f) Deyang/ Baoji

(g) Tianshengqiao

Fig 3: Profiles of ±500 kV DC voltage dividers (PTs) used in China.

than 2 weeks of persistent rain, including 2 to 3 days of heavy fog during which visibility was reduced to only about 5 m.

Suppliers of the affected equipment were contacted and confirmed that the same silicone rubber material was used on all bushings of the affected DC voltage dividers. Table 2 lists the technical parameters of these units, which are shown in Fig. 1.

Housings on the DC voltage dividers at Longquan, Jiangling and Yidu were all silicone rubber type with basically the same section size. The grading rings on the equipment at Jiangling and Longquan stations were also the same although the voltage divider at Yidu had a different single ring structure. In 2009/2010, staff at the State Grid decided to apply RTV coatings to the composite housings of all three and there have been no further reports of problems since then.

Fault at Tianshengqiao Converter Station In May 2005, again during a period of heavy rain, a flashover occurred on the porcelain housing of the voltage divider on Pole I at Tianshengqiao and white traces remained visible afterwards on the high voltage portion of the equipment. To prevent a recurrence, operating staff at the station decided to add booster sheds to the porcelain and also applied an RTV silicone coating (see Fig. 2).

There have been no further flashover incidents with this apparatus.

63

diminished performance under Different External Insulation for Main ±500 kV DC Voltage Dividers conditions of high pollution and

Photo courtesy of China EPRI

Fig. 4: Examples of poor hydrophobicity of two indoor standby voltage dividers in DC fields at Yidu and Huaxin Converter Stations.

Research has demonstrated that hydrophobicity has such a significant influence that the pollution flashover voltage of composite insulators under good hydrophobicity is 2.4 times higher per unit length than when this property has been lost.

64

The housings of the voltage dividers used in the Ge-Nan and Tian-Guang DC projects were made from porcelain with relatively low specific creepage distance. However, once coated with RTV, creepage distance was seen as equivalent to that of a compositehoused voltage divider. Moreover, at the Tianshengqiao and Nanqiao Converter Stations, the creepage distance of porcelain-housed voltage dividers, once coated with RTV and with added booster sheds, exceeded that of composite-housed units.

In addition, the design of the DC field busbar voltage dividers used on the Yi-Hua and De-Bao systems has been improved by reducing the size of their grading rings at the high voltage end. This permits more efficient use of the composite housing’s creepage.

The external insulation on these different types of ±500 kV DC voltage dividers is described in Table 3 and shown in Fig. 3.

Failure Countermeasures

wetting.

While water dripping from the sheds of a saturated insulator due to gravity takes some deposited surface salts with it (effectively cleaning the surface), a decrease in hydrophobicity means that a continuous film of conductive moisture can more easily form on the surface of polluted insulators.

The result is a serious net reduction in dielectric strength. Indeed, research has demonstrated that hydrophobicity has such a significant influence that the pollution flashover voltage of composite insulators under good hydrophobicity is 2.4 times higher per unit length than when this property has been lost. For example, under pollution conditions of ESDD = 0.10 mg/cm2 and NSDD = 0.60 mg/cm2, pollution withstand of composite insulators whose surface hydrophobicity is lost is approximately 47 mm/kV.

In 2009/2010, ESDD levels on post insulators at the Longquan After analyzing past insulation Converter Station reached 0.08 mg/ faults affecting ±500 kV DC voltage 2 dividers in China, the causes of these cm and in the case of the Yidu Conveter Station were between various problems could be grouped 0.11 and 0.13 mg/cm2. Given the into the following categories: 22,789 mm total creepage of the three DC dividers, their pollution a. Failures Due to Insufficient Creepage withstand voltage under moderate The minimum specific creepage of contamination was only about 485 equipment used on early ±500 kV kV – less than the normal operating DC projects in China (e.g. Ge-Nan voltage. The probability of flashover and Tian-Guang) was only about became high simply because there 40 to 41 mm/kV. However, this was no longer any insulation margin. value proved insufficient and indeed almost all equipment used PTs used in the Long-Zheng, Jiangin the DC field, including voltage Cheng and Yi-Hua DC projects were dividers, experienced varying all supplied by a European-based degrees of discharge phenomena multinational and investigation at – some even when operating at Huaxin and Yidu converter stations lower voltages. External insulation revealed that even unused spares failures caused by lack of sufficient from this manufacturer had poor creepage were found best resolved hydrophobicity on shed surfaces, i.e. by adding booster sheds or applying only HC5 to HC6. Figure 4 shows RTV coatings to increase external examples of the hydrophobicity of insulation strength. these standby indoor DC voltage dividers. b. Failures Due to Decline in Hydrophobicity Composite insulators with diminished Flashovers of external insulation surface hydrophobicity are more on ±500 kV DC voltage dividers likely to experience arc burns as well at Jiangling, Longquan and Yidu as tracking and erosion caused by Converter Stations under persistent partial discharges or corona activity rain and fog were traced mainly to decline in the hydrophobicity of their under service conditions marked by UV, temperature change, rain, fog, composite housings and resulting

High quality glass insulators with 50 years experience for diverse applications Nanjing Electric is China’s only glass insulator supplier with 20 years export experience.

As the world’s largest producer of glass insulators, we have grown to become global leaders, supplying line projects in over 60 countries.

Make us your first choice to assure the very best performance possible, whatever the application.

AC insulators up to 1000 kV, from 70 to 830 kN DC insulators up to ±800 kV, from 70 to 830 kN

Nanjing Electric (Group) Co., Ltd. Founded 1936

NO. 63, TaiXin Road, Nanjing, China 210038 Tel: +86-25-85372050/2052/2053/2336/2199 sales@glass-insulator.com www.glass-insulator.com

snow, etc. Loss of hydrophobicity by a composite insulator usually means unsatisfactory service performance since, even if it does not experience pollution flashover or serious discharge phenomena, it will still require greater scrutiny by maintenance staff.

Moreover, premature ageing of silicone rubber sheds and housings due to a variety of service stresses might cause performance to decrease as the insulator mcould experience discoloration, loss of gloss, hardening, deformation, cracking or increased brittleness.

In such cases, one solution would be to coat the unit with RTV material to restore hydrophobicity and increase pollution flashover voltage.

At the same time, this experience has shown that it is a good idea to regularly monitor silicone insulators – even on spare equipment – for any possible decrease in hydrophobicity. c. Problems Due to Design Defects Another cause of flashovers on ±500 kV voltage dividers at Jiangling and Longquan Converter Stations under conditions of fog or drizzle was discovered to be poorly designed grading rings and insulator sheds.

The external insulation distance of the DC voltage dividers at the two

gap between the grading ring and the nearest shed on the bushing has to withstand very high voltage, especially when the temperature near the top of the divider (i.e. within range of the grading ring) is relatively high. Once the air gap breaks down, all voltage is taken on by the lower portion of the insulation and voltage per unit of creepage for this section increases by 20% relative to that along the whole bushing. This can induce flashover.

Fig. 5: Dimensions of ±500 kV DC voltage dividers at Jiangling and Longquan Converter Stations.

affected stations was 5600 mm.

Their circular grading ring at the high voltage end had a diameter of 1580 mm while tube diameter was 240 mm. Given the grading ring’s shield depth of about 1100 mm, the air gap between it and the nearest bushing shed was only about 265 mm – clearly too short.

Basically, because some grading rings have such high shield depth, effective creepage distance of composite bushings can be shortened by as much as 20%. When the surface becomes wetted, the air

In addition, spiral-shaped voltage divider housings with big-small alternating shed configurations do not seem ideal when it comes to inhibiting rapid increases in surface leakage currents. This is because, should hydrophobicity be reduced or lost, water on the sheds can more easily flow down along the spiral surface and accelerate formation of a full film of moisture.

The geometry of the composite bushings on faulted voltage dividers was basically the same as that of the porcelain bushings used on DC voltage dividers at the Gezhouba Station that never flashed over (even though they did experience minor discharge phenomena on their surface). The explanation here may be that while the composite housing has alternating sheds with spiral structure, the alternating big-small

Table 4: Discharge Countermeasures for DC Equipment Discharge Classification

Porcelain Insulating Surface Discharge Conditions

II

Transient partial arcs simultaneously on bottom surface of most Transient Short-term along the insulating partial arcs simultaneously on bottom surface; More than half of sheds surface of most insulator sheds; No arcing; Accompanied by prolonged or arc development from bottom surface intense noise. to upper surface; Accompanied by continuous noise.

III Instantaneous local arcing on bottom surface of several or individual sheds; Local arc may develop from bottom surface to upper surface; Arc accompanied by intense intermittent noise.

Instantaneous local arcing along insulator; Partial short arc on whole shed; Accompanied by intense discharge noise.

Intermittent transient partial arc simultaneously on lower surface of insulator shed; Accompanied by intermittent discharge noise.

Instantaneous blue discharge flashing on bottom surface of several or individual sheds; Intermittent, occasional noise.

Need for Countermeasures

Mandatory

Enhanced inspection

No effect

Types of Countermeasures

(a) Decrease operational voltage (b) Coat with RTV or add booster sheds during shutdown.

(a) Decrease operational voltage when serious. (b) Clean or coat with RTV during shutdown.

Enhanced inspection needed.

Composite Insulating Surface

66

I

sheds on the porcelain bushing are concentric.

By contrast, the reason that the porcelain-housed DC voltage divider at the Tianshenqiao Station experienced flashover during rain was that the geometry of its sheds was poorly designed. In this case, the spacing between the large sheds was 65 mm while width of sheds was 65 mm to 70 mm, meaning a ratio of close to 1. Moreover, the difference in the widths of the small and big sheds was just 20 mm. A DC arc is more stable than an AC arc and therefore any arc-shortening phenomenon between sheds becomes more serious since it reduces effective utilization of the entire creepage.

Loss of hydrophobicity by a composite insulator usually means unsatisfactory service performance since, even if it does not experience pollution flashover or serious discharge phenomena, it will still require greater scrutiny by maintenance staff.

It has been found that the pollution and rain flashover performance of any DC device that has relatively small shed separation distances on its external insulation can be improved by installing silicone rubber booster sheds. These not only control shorting of adjacent sheds by an arc but also improve the condition of the insulation when wetted. Utilization are appropriate to deal with such rate of the creepage distance is problems. For example, equipment improved as well. should be taken out of service and either coated with RTV silicone Classification of Alternative or have booster sheds added if Flashover Countermeasures experiencing level II or level III China’s Electric Power Research discharges. Moreover, if there are Institute analyzed localized arc corona discharges, better grading generation and its impact on various should be installed to improve field configurations and types of external insulation on DC devices experiencing distribution and reduce or eliminate such phenomena. different forms of discharge (i.e. severe, moderate and slight).

At the same time, there should be greater monitoring of silicone The goal was to facilitate evaluation composite bushings as well as of of discharge phenomena under the hydrophobicity on the surface various wetting conditions such as of RTV coatings at a converter fog, rain, snow and ice as well as to station. In particular, it is important judge the comparative reliability of to inspect for any sustained loss of different insulation based on field hydrophobicity during several days and laboratory experience. At the same time, this analysis would assist of consecutive humid conditions as well as the rate of hydrophobicity classifying the most appropriate recovery once the weather improves. countermeasures to deal with discharges in each case (see Table 4). It is also recommended to test for hydrophobicity changes due to the existing surface pollution layer and Based on the existing operating prevailing weather conditions, with situation at each converter station measurements best carried out when studied, discharges in the DC there are several days of continuous field during fog or rainy weather sun. The ideal interval between rewere either partial arcs along the coating with RTV silicone should be surface of the external insulation or determined according to the station’s corona discharges at the insulator specific operating environment. fittings. Different countermeasures

The surface of composite external insulation does not normally need to be cleaned or washed since this can result in short-term damage to the hydrophobicity of the silicone rubber material. The only exception is when the pollution layer has become so thick that the low molecular weight species in the bulk rubber that are responsible for transfer of hydrophobicity can no longer effectively migrate to the surface.

Summary

At the end of 2012, China had already built ten ±500 kV pointto-point DC transmission circuits, transporting a total capacity of 28,800MW. These ±500 kV transmission lines have now taken on an important role in interconnecting the country’s regional power grids.

Typical external insulation failures of DC dividers at ±500 kV converter stations include flashover of both composite and porcelain bushings, typically occurring under severe wet weather such as fog, sleet or heavy rain. The main causes of failures on potential transformers have been found to be inadequate external insulation configuration, reduction or loss of surface hydrophobicity of the composite housing and design defects in the equipment.

The main measures to improve the level of external insulation in such cases have been found to include adding a coating of RTV silicone to the porcelain or to the composite insulators as well as installing booster sheds. Periodic monitoring of surface hydrophobicity of the silicone housing or of the RTV coating should also be increased.

Problems due to inappropriate grading ring design and/or shed geometry should be identified and solved as quickly as possible.

Discharges affecting equipment in the DC field of converter stations can fall into one of three classifications: severe, moderate or slight, and different measures are required to deal with each. Ideally, equipment should be coated with RTV material or have booster sheds added should either level II or level III discharges be present.  67

TESTING

Innovations in Type & Commissioning Testing of High Voltage Cables

O

At present, there are two primary IEC standards and one ICEA standard that prescribe the methods and requirements for type testing as well as commissioning testing of transmission class cables. This article, contributed by industry expert John Kuffel as well as Mark Fenger and other testing staff from Kinectrics in Toronto, reviews the procedures which form the basis for these standards with a view to performing type test qualification to IEC and ICEA concurrently. It also proposes novel techniques intended to make these tests more efficient and economical and reviews the experience in performing them.

68

Photos: INMR Š

ver the past two decades, the use of extruded XLPE cables in underground transmission systems has increased steadily to the point that they now account for the vast majority of all new installations. This near total shift away from oil and paper insulation has made it worthwhile to look for improved methodologies for testing solid dielectric cables.

Standards for Laboratory Type Tests

Type tests for solid insulation HV cables are described in two IEC standards and one ICEA standard: • IEC 60840 Edition 4.0 2011-11 “Power cables with extruded insulation and their accessories for rated voltages above 30 kV (Um = 36 kV) up to 150 kV (Um = 170 kV) – Test methods and requirements” • IEC 62067 Edition 2.0 2011-11 “Power cables with extruded insulation and their accessories for rated voltages above 150 kV (Um = 170 kV) up to 500 kV (Um = 550 kV) – Test methods and requirements” • ICEA S-108-720-2012 “Standard for Extruded Insulation Power Cables above 46 through 345 kV”.

The electrical type test requirements for these standards are similar but not the same (see flowcharts in Fig. 1). As evident, the procedures mandated are different only in that cables rated for system voltages of 300 kV and higher require a hot switching impulse test while those rated below do not. The procedures outlined in the ICEA standard differ more significantly, as follows: 1. Different sequence of tests; 2. Higher required elevated test temperatures; 3. Requirement that cable loop is installed in a pipe;

4. No requirement that cable systems rated 300 kV and above be subjected to hot switching impulse test; 5. Requirement that hot impulse test be completed by breakdown (or that cable system is tested to limit of test equipment); 6. PD test performed at higher test voltages; 7. Requirement for 2 hour AC withstand test at 2.5Uo. Typically, cable manufacturers and users prefer a single test program to satisfy the requirements of both standards and this can be done by utilizing the procedure outlined in Table 1 and applicable to cables systems rated for systems up to 150 kV. Since this procedure combines the most severe factors from each specification, passing this program results in qualification to both standards.

Developments in Laboratory Type Testing

In order to improve the accuracy and integrity of cable type test procedures, several innovations have now been developed. These include a method of continuous heating throughout the impulse testing as well as eliminating the need for the dummy loop traditionally used during thermal cycling tests.

Impulse Test Heating During the impulse testing segment of type testing, conductor

temperature must be maintained within given limits linked to the maximum operating temperature of the cable system. Application of the specified 10 positive and 10 negative impulses takes time since impulses are typically applied every 2 minutes.

This means that about 30 minutes are required for each set of required positive and negative impulses when taking into account the need for conditioning impulses at reduced levels.

Concerns about damaging the heating system during possible impulse breakdown have lead some test laboratories to heat the cable to the required temperature before application of the impulse sequence test. Then, upon initiation of the impulse applications, the heating source is disconnected.

If the cable is not heated during the 30 min period required for each set of positive and negative impulse applications, its temperature can drop significantly below the required limits. To address this issue, a method of heating the cable between successive impulse applications has been developed which ensures that the temperature limits required by the standards are maintained throughout impulse testing.

Eliminating the Dummy Loop The protocol used during thermal cycling testing requires that the cable system undergo a heat cycling

IEC 60840

IEC 62067

ICEA S-108-720

Cable bending test

Cable bending test

Cable bending test

PD at ambient temperature following test loop construction Measurement at 1.5Uo

PD at ambient temperature following test loop construction Measurement at 1.5Uo

Thermal cycling, 20 cycles, in pipe, T=TCEM Constant voltage of 2Uo

Hot Tan δ Measurement at Uo

Hot Tan δ Measurement at Uo

Hot impulse test, T=TCEM+0/-5ºC 15 min ac withstand

Thermal cycling, 20 cycles TCMAX+5ºC≤T≤TCMAX+10ºC Constant voltage of 2Uo

Thermal cycling, 20 cycles TCMAX+5ºC≤T≤TCMAX+10ºC Constant voltage of 2Uo

AC Voltage Withstand test 2.5 Uo for 2 hours

Ambient temp and hot PD Measurement at 1.5Uo

Ambient temp and hot PD Measurement at 1.5Uo

Ambient temp PD Measurement at 2Uo

Hot lightning impulse test 15 min ac withstand

Hot switching impulse test for cables rated 300 kV and above Hot lightning impulse test 15 min ac withstand

Hot Tan δ, T=TCEM+0/-5ºC Measurement at Uo

Dissection and analysis of sample

Dissection and analysis of sample

Dissection and analysis of sample

Fig. 1: Flowcharts of electrical type test requirements for IEC & ICEA standards.

69

Table 1: Sequence of Electrical Type Tests for 138 kV Cable System in Accordance with IEC & ICEA Standards Item

Test

Clause(s)

1

Bending test

IEC 12.4.3 ICEA 10.1.2

2

Partial discharge test at ambient temperature at 160 kV

IEC 12.4.4 ICEA 10.1.6

3

Tan delta measurement at 95 to 100°C & 100 to 105°C, both at 80 kV

IEC 12.4.5

4

Thermal cycling – 20 heating cycles at 100 to 105°C & 160 kV

ICEA 10.1.3

5

Tan delta measurement at 100 to 105°C & 80 kV

ICEA 10.1.7

6

Lightning impulse voltage test at 100 to 105°C & ±650 kV

ICEA 10.1.4 IEC 12.4.7

7

Ac voltage withstand test at 200 kV for 2 hours

ICEA 10.1.5

8

Partial discharge test at ambient temperature, at 95 to 100°C and at 100 to 105°C, all at 160 kV

IEC 12.4.4 ICEA 10.1.6

9

Examination

IEC 12.4.8

10

Dissection & analysis of test specimens

ICEA 10.1.8

voltage test over the relatively long period of 20 days. The test involves heating, soaking and cooling the cable system for 20 cycles while the system is energized according to the specific voltage class of the cable and accessories. Each cycle lasts 24 hours while heating is maintained for 8 hours. During the first 6 hours,

utilizing an identical cable. This ‘dummy’ loop is heated in the same manner as the test loop and the temperature of its sheath and conductor are continuously recorded.

The only difference between the loops is that the dummy loop is not energized and therefore thermocouples can be directly attached to the conductor in order to measure its temperature.

the cable conductor must reach a specified temperature and this must be maintained within a 5°C limit over the next 2 hours. The cable is then allowed to cool naturally for 16 hours.

The need for such a dummy loop has effectively been eliminated by implementing a mechanism for transmitting data under voltage using a wireless data logging transmitting system. Basically, conductor temperature is measured by means of a smart link telemetry system installed on a length of the same cable being tested. Thermocouples are attached directly to the surface of the conductor as well as to the sheath of the control cable and connected to a wireless transmitter nearby. The control cable is then installed between the outdoor terminations in series with the test loop such that the conductor in this length of cable carries the same current as the test loop itself.

When performing this test, the standards suggest a control loop

Use of such a ‘smart link’ allows for electrically isolated temperature

Thermal-controlling element (smart link)

Location of Shielded Wireless Transmitter

Location of Temperature Sensors

Photos courtesy of Kinectrics

Location of Thermocouples

Cable system test setup showing control cable piece and telemetry system installation point.

70

Cable system test setup showing control cable and fiber-optic temperature monitoring and control system installation.

Test Equipment

Fig. 2: Schematic of high voltage power supply.

measuring points directly on the conductor and on its sheath. At the moment, existing monitoring equipment at Kinectrics during heat cycle voltage testing cannot be used to automatically control the test loop’s heating cycles since it is unable to continuously transfer data under high voltage.

However, new fibre-optic based technologies for monitoring temperature under voltage are available and one has already been identified that has the potential to perform well during the heating cycle voltage test described. The advantage of such a system lies in the fact that the fiber-optic cables are isolated and can be attached directly to the energized conductor.

Such a set-up allows temperature reading of the smart loop’s conductor to be captured on a continuous basis, thereby enabling automatic control of the test loop heating current. These minor modifications to the temperature sensing elements combined with custom programming allow the system to be tailored to the application.

cables have been applied. However, possible problems caused by space charge injection during such testing resulted in DC tests being abandoned for cables insulated with solid extruded polymers. Lack of alternative forms of external energization has led to the so-called ‘soak test’ under which the newly installed cable is put on potential for a 24-hour period. Unfortunately, there have been cases reported of subsequent cable and accessory failures within only a short time after the tested cable system was placed into service.

Availability of variable frequency AC power supplies has since enabled after-laying high voltage testing of XLPE-insulated transmission cables up to 400 kV and over 20 km in length. In this regard, application of AC overpotential testing in conjunction with partial discharge (PD) testing can aid in assuring reliability of a new cable installation.

High Voltage Power Supply The high voltage power supply employed is a 260 kV, 83A variable frequency resonant test set (RTS) that complies with IEC standards 60840 and 62067 and operates within the frequency range 20-300 Hz. A schematic of the test setup is illustrated in Figure 2. As can be seen, a blocking impedance is placed between the power supply and the high voltage connection to the cable under test with two goals in mind: firstly, the blocking impedance protects the RTS in the unlikely event of cable failure; secondly, it effectively filters any high frequency noise originating from the RTS and improves the signal-tonoise ratio of the power supply when performing PD measurements.

A capacitive voltage divider provides for a voltage reference for the control unit of the power supply. The common point grounding of the entire test circuit is connected to station ground.

A 6 inch (15 cm) wide copper foil provides a high frequency ground path while a stranded, insulated aluminium conductor positioned directly on top of the copper sheath constitutes the power frequency ground. Signal coupling is provided by attaching a High Frequency Current Transformer (HFCT) sensor around the ground link from the cable joint toward the link-box. High frequency currents induced as a result of any partial discharge activity in the joint or in the cable section will be coupled to the HFCT sensor and measured by a conventional partial discharge monitor.

A fiber-optic temperature monitoring and control system as described above has already been successfully tested in monitoring mode during a 132 kV cable system type test.

Moreover, it is now being deployed as the primary monitoring and control system for type testing of 240 kV cable.

Traditionally, due to lack of AC high voltage power supplies capable of energizing several kilometers of cable, DC overpotential tests similar to those used for fluid-filled

Photos courtesy of Kinectrics

Commissioning Tests for HV XLPE Cable Systems

Typical test set up in the field.

71

Photos courtesy of Kinectrics

HFCT sensor installed on joint cross bonding lead.

The commercially available partial discharge monitor used has a 350 kHz to 800 MHz bandwidth and measures the amplitude (in mV) as well as phase angle of any signal detected. A pulse count rate for various categories of magnitudes and phase-angles is also generated. The phase angle reference is provided by a low-frequency winding embedded in the HFCT sensor.

Test Method While IEC 60840 and 62067 provide basic guidance on the waveform, frequency and prescribed voltage to be employed during the overvoltage test, there is still no standard prescribed procedure for PD measurements. As such, there is some variation in the measurement procedures employed by various service providers in this field.

The measurement protocol followed in the tests described here consists of the following: 1. Upon tuning the high voltage power supply to the appropriate resonant frequency, a relatively low voltage (on the order of 30 to 40 kV) is applied for two minutes during which various diagnostic parameters are checked to ensure that the system is functioning properly.

2. Voltage is increased to the nominal line-to-ground potential (U0) and held for a further two minutes while diagnostic parameters are confirmed as normal.

72

Parallel setup of 2 resonant test systems for testing 10 km 220 kV cable installation.

3. The voltage is raised to the prescribed level specified in the IEC standards for a period of one hour.

In order to enable performing PD measurements on all accessories during the limited 1-hour hi-pot test, PD detectors can be installed at each of the accessories in the circuit. Signals from these devices are then fed back to a remote test operator for display and analysis.

This approach requires the availability of a communication path between the individual joints and the location of the test operator – something that is now increasingly possible as more and more installations of HV cable incorporate fiber communication links into the initial work. This approach, while expensive, has significant advantages in that each sensor point can be observed simultaneously in real-time.

Where such communication networks are not available, the PD must be recorded sequentially at each individual accessory. For significant cable lengths, the time required to carry out these measurements usually exceeds the one-hour hi-pot test duration. In these cases, the PD levels at the maximum possible number of joints are recorded during the one-hour hi-pot test, while the remaining accessories are PD tested at an applied voltage of between Uo and the specified one-hour hi-pot level. This level must be agreed to by

the parties involved and has usually been in the range of 1.2 Uo. Field Experience Numerous tests have been performed in various locations in the Middle East and across North America involving XLPE-insulated cables from 66 through 380 kV and with circuit lengths ranging from 3 to 30 km. Typically, those circuits rated at 220 kV and higher have required use of two resonant test systems operating in parallel.

The majority of circuits subjected to such overpotential testing successfully withstood application of the prescribed voltage for one hour. However, in a small number of cases, there were dielectric breakdowns and these failures occurred almost exclusively either in or close to a joint or termination.

Typically, defects in the accessories were attributed to problems with installation procedures rather than to deficiencies in design or materials.

For example, there was one instance of cable failure where mechanical damage during transport or storage was suspected as root cause.

It may seem attractive to perform PD measurements at the terminals of the cable circuit rather than at individual joints. However, it is well known that problems due to signal attenuation and dispersion limit the lengths of cable where this method could be successfully applied.

Detection of PD-related phenomena

occurring several kilometres from the detection point requires that the measurement bandwidth be relatively low. Such techniques therefore suffer from problems due to relatively high background electrical interference associated with field measurements. Moreover, although various steps are taken to mitigate the effect of switching noise from the high voltage power supply, these signals will also be present and add to the difficulty of separating PD signals from background noise.

phase-to-ground voltage. These are classic locations for phase-to-ground dependent partial discharge data.

The left side of Fig. 3 displays data acquired during the off-line test whereas the right side shows data from the on-line test – both conducted at a voltage of U0. PD activity measured during the off-line test has a frequency content ranging from 7 MHz to 8 MHz whereas minimum, mean and maximum PD magnitudes for the off-line data are 5, 7 and 24 mV respectively. In addition, the phase resolved PD plot for the off-line data shows clusters of negative and positive polarity pulses centered near 45° and 225° phase angle with reference to the

One key point emerging from the testing performed to date is that all of the circuits that successfully withstood the high voltage tests and on which no PD was detected have operated without incident since commissioning. Circuits which have only been hi-pot tested without the inclusion of PD testing have experienced some failures.

Consequently, the combination of overvoltage and PD testing appears to provide the highest level of confidence in the reliability of the cable, accessories and installation method.

Coordinated development of existing standards for high voltage cable systems allows their effective type test qualification to both IEC and ICEA simultaneously. At the same time, innovations in techniques used to perform standard type testing of HV cables can improve accuracy and efficiency, while also lowering costs.

Numerous solid dielectric transmission cable circuits have been subjected to high voltage and PD testing using this methodology as part of the commissioning process following installation. The majority of cable systems tested successfully met the test criteria.

However, a small number suffered dielectric breakdown during the overvoltage test with failures generally located in or in close proximity to accessories. Breakdown of the cable itself is extremely rare.

Where post-failure analysis has been performed, the results are consistent with installation issues or damage to the cable during transport or storage. None of the breakdowns have been attributed to design, materials or processing issues. While there is still ongoing debate regarding the technical and cost issues associated with optimizing PD test procedures for on-site transmission cable testing, experience shows that the combination of high voltage and PD testing of transmission cables is necessary to satisfy industry demands for high reliability. 

Photos courtesy of Kinectrics

A critical question when making PD measurements relates to establishing pass/fail criteria for their magnitudes. The simplest and most conservative answer is that any detectable level of PD is too high but many practical issues need to be addressed. Among these are the costs and delays needed to replace any accessory or section of cable.

One practical example of attempting to define PD-based acceptance criteria for commissioning testing is illustrated in Figure 3.

However, during on-line testing, measured signal activity is quite different, i.e. frequency content of the signals ranges from circa 800 kHz to circa 2 MHz. In addition, the phase resolved PD plot shows the pulses measured to be located throughout the AC cycle.

Consequently, the data acquired during the on-line test relates to electrical noise. This figure demonstrates that no PD signals were present during on-line tests since, if present, these would have been detected simultaneously with the noise signals and would have been identifiable in the frequency domain plot of Figure 3. While there are a number of factors that may account for the differences observed above, a decision was made to put the cable in-service in spite of detectable PD during the off-line test. This circuit is currently operating without incident.

Conclusions

Parallel setup of 2 resonant test systems for testing an 11 km 380 kV cable installation.

Fig. 3: Comparison of off and on-line PD test results.

73

MAINTENANCE

Monitoring

System Detects Critical Insulator Contamination on Overhead Lines

Photos: INMR Š

74

Monitoring the build of contamination on insulators operating in polluted service areas is key to ensuring the reliability of any overhead network.

If not washed in a timely way, contaminated insulators present an elevated risk of flashover, with potentially costly voltage sags and supply interruptions.

Various on-line systems have been developed to monitor leakage current with a view to alerting utilities when there is increased risk of pollution flashover. Specialized cameras are also used widely to detect corona emanating from excessive partial discharge activity on heavily contaminated insulators.

INMR visits Metrycom Communications, which has developed its own technology to anticipate heightened risk of pollution flashover and allow maintenance staff time to undertake remedial action. This system is already in place along vulnerable sections of 160 kV and 400 kV lines in Israel and Russia with additional installations now planned in other countries as well.

75

Dynamics of ‘Dry Banding’

Insulators on overhead lines can become contaminated by a variety of pollution sources, from sandstorms to maritime winds, from industrial emissions to road salt. Moreover, contamination deposition is usually not uniform and depends on insulator geometry. Edges of sheds typically accumulate more contamination while smooth surfaces usually retain less contamination due to more efficient natural cleaning.

Photos: INMR ©

When insulators are wetted by fog, rain or dew, salts present in the pollution layer dissolve and surface conductivity can rise sharply – sometimes to hundreds of mA depending on the levels of contamination and wetting.

Metrycom founder and CEO, Liron Frenkel, reviews the dynamics of dry band arc formation on contaminated insulators under wetting conditions.

Since an insulator's surface area is not uniform, he explains, current density along a string will also change such that different regions heat up differently. Those areas where current density is highest dry up soonest and the voltage drop across them increases to a point at which there is breakdown of the air gap. An arc or discharge ensues.

“Our system detects dry band arcing by means of special sensors and a data processing device connected to them that allows analyzing the situation from afar.”

Frenkel points out that while the dynamics of pollution flashover have been studied for many years, traditional solutions to alert maintenance staff of an elevated risk are not always reliable or costeffective. For example, he claims that one of the problems with relying on leakage current measurement is that this parameter correlates with contamination level only if relative humidity exceeds 85%. It therefore cannot always accurately predict pollution flashover. Moreover, evaluating insulator contamination using this methodology requires installing sensor devices on many strings.

Corona cameras, observes Frenkel, also face limitations when it comes to alerting maintenance staff to excessive contamination levels on insulators. These include the high cost of helicopters to efficiently cover line sections, variance of output of each inspection according to prevailing weather conditions (such as humidity and wind speed) and the operator skill required to interpret readings that are subjective and often not supported by a reliable past database. Also, since contamination can build up

Photos courtesy of Israel Electric

“But this phenomenon does not in itself constitute insulator failure,” emphasizes Frenkel, “since surface currents dry the moisture and the arc extinguishes. But then, as the dried surface again becomes humid, the process restarts. Arcing can

Pollution accumulation tends to be greatest along edges of insulator sheds.

grow to cover one or more insulators along the string and, depending on level of contamination and amount of wetting, there may even be a flashover. Our system was designed to detect such dry band arcing by means of special sensors connected to a data processing device that allows analyzing the situation from afar.”

Partial discharge on polluted insulator strings. Discharge activity does not interrupt normal line operation but, when there is severe contamination as well as high ambient humidity, arcs cover larger areas of insulator string. Surface leakage current increases while resistance of arcs decreases, leading to eventual flashover.

76

Looking to identify ideal suppliers of network components & hardware?

WE CAN HELP www.inmrbuyersguide.com Independent T&D Information Resource BUYER’S GUIDE

“Our central goal was to use dry band arcing to monitor flashover risk and to know when it is time to conduct washing along any line segment.”

quickly from events such as dust storms or road salting a corona camera is not that effective a tool when it comes to monitoring for sudden changes in insulator contamination levels.

“Our approach is different,” he remarks, “and relies instead on single sensors that can each cover some 2 kilometers of line. Indeed, our initial development concept was that the system had to be easy to install in the field using wireless sensors that could be deployed over large areas.”

Sensors installed on shield wire measure electromagnetic pulses from dry band arcing.

Frenkel explains that Metrycom sensors located on transmission towers measure the high frequency signals generated by partial discharges over contaminated insulator strings. Each such unit includes a wireless modem that communicates with multiple adjacent sensor units thus creating a meshed IP sensor network.

Data measured by any unit is routed from sensor to sensor along the line until it reaches a gateway, usually located at the substation. The network is also selfhealing such that, if any sensor stops functioning, the others find a new routing through more distant units.

The gateway then sends the data received to a control server where it is stored, converted to a comprehensible format and dispatched to local maintenance departments responsible for insulator washing.

According to Frenkel, this technology evaluates risk of flashover based on a combination of contamination as well as other relevant factors – not simply contamination level observed. Indeed, in many cases insulator strings that look heavily contaminated will not develop flashovers while insulators with high levels of dry band arcing have a high probability of flashing over.

Measurement Technology Photos courtesy of Metricom

78

Dry band activity generates electromagnetic pulses that travel over the ground wire and phase conductors.

Lengths of arcs from dry band discharges can vary from only a few mm to some tens of cm, with arc head speeds of 50 m/sec or higher. The resulting pulses have amplitudes up to tens of kV with a front of between tens of nanoseconds and several microseconds. Along with radio interference, acoustic noise and visible light are produced in a spectrum of up to hundreds of MHz. The dry band arcing phenomenon depends on level and type of contamination but also on ambient

Wireless data acquisition unit transmits findings to substation gateway for subsequent analysis by system software.

are taken with maximal pulse amplitude measured for each onesecond interval. The dynamic range of these measurements is then divided into 32 equal levels ranging from 0.7 V to 32.7 V in order to obtain a set of pulses with corresponding amplitudes. Apart from the sensor, remote terminal unit and gateway, the other key component of the system is the software necessary to analyze the data.

Photo courtesy of Metrycom

humidity, wind speed and direction, temperature and so on.

Radio interference travels along a transmission line and is detected by a wide band sensor that can be located as far as 1 km from a contaminated insulator. The sensor, which measures amplitude of these pulses, is placed on the shield wire of the line section being monitored and has a frequency range of 0.1 to 100 MHz. This signal from the sensor then enters a specially designed data acquisition unit where it is filtered, analyzed and transmitted by modem to the utility’s operations and maintenance center.

Here, information from measurements made 8 times each day can be processed using statistical methods.

Sampling and analysis takes 5 minutes during which 300 samples

Amplitudes of pulses that govern flashover risk are usually different for lines having different voltage ratings and even for lines of the same voltage class but with different types of insulators.

Field Experience

Head of the Transmission Team at Israel Electric (IECo), Alex Levinzon, reports that several Metrycom systems for insulator contamination monitoring have already been installed in Israel and are operating on selected 160 kV and 400 kV lines. Among the objectives behind these installations was to provide IECo maintenance teams real-time information about line segments that are at higher risk of flashovers and thereby reduce the costs linked to consequent short circuits and power sags. Other goals included lowering total washing expenses by optimizing cleaning schedules as well as choice of which line segments most needed cleaning.

Levinzon agrees with Frenkel that there is not much value in monitoring contamination level alone as an indicator of when it has

Photos courtesy of Metrycom

Sensors being deployed on selected 161 kV lines in Israel.

79

become necessary to wash lines since this parameter does not always correlate with risk of flashover. Says Levinzon, “maintenance staff at power companies such as ours want to be in a position to anticipate risk of flashover due to insulator contamination. So, we look instead to measuring the amplitude of partial discharge pulses on insulator surfaces and also observe which portion of the string has flashed-over.

For example, we have found that in the case of porcelain insulators, when dry band arcing occurs over more than two discs of the string, there is increased risk of flashover.”

Levinzon goes on to state that because of the large number of factors that can affect measurements on existing lines, statistical processing of long-term measurements is needed in order to obtain reliable data. He also points out that the amplitudes of pulses that govern flashover risk are usually different for lines having different voltage ratings and even for lines of the same voltage class but equipped with different types of insulators, i.e. silicone versus ceramic. Since the system provides daily information online, users get a complete picture of the level of partial discharge pulses under different humidity conditions, based on measurements over many days.

One example of field application of Metrycom’s sensor system involved a 160 kV line running through a polluted construction area in the south of Israel. Results of preliminary measurements on this line revealed that considerable contamination was present when pulse amplitudes were higher than 3.5 V (Yellow alert). Moreover, flashover became a serious risk when amplitudes exceeded 6.5 V (Red alert), in which case washing would be recommended (see Fig. 1 top).

Levinzon reports that experience with such systems installed in Israel has so far proven successful and demonstrated that more than half of past washing operations were not required since there was little risk of pollution flashover. At the same time, critical levels of agricultural pollution affecting insulators on 80

Courtesy of Israel Electric

Fig. 1: (top) Example of interruption of power due to pollution flashover on 160 kV line in southern Israel that was predicted by system but not acted upon. (bottom) Example of unnecessary washing operation performed on 400 kV line even though system did not signal such a need. certain line segments created an urgent need for washing. He adds that a new sensor installation will monitor a coastal line equipped with silicone insulators where unexpected problems with pollution flashovers have been known to occur during the early morning hours.

“Helicopter washing costs us between $4500 and $6000 per kilometer of line based on about 3 to 5 minutes for each insulator string,” remarks Levinzon. “If you consider how much we have saved by not washing lines that are not at risk of pollution flashover, the return on investment of this sensor system is probably less than 2 years.”

For his part, Frenkel points out that while there has so far been less than 5 years field experience with this system, the basic technology is backed up by more than 15 years of research. Says Frenkel, “much field as well as laboratory testing has allowed us to set accurate green, yellow and red thresholds when it comes to the need for washing any particular section of line.

Moreover, our growing database of types of insulators and transmission line designs now allows us to set these levels without having to perform specific calibration for every new installation. At the same time, when necessary, we have developed laboratory procedures for calibrating the system for new types of insulator strings. For example, laboratory testing we have performed has shown that there are different physics when it comes to the development of flashovers on porcelain insulators versus glass insulators and therefore we use different parameters for these.”

Apart from its system of partial discharge pulse sensors, Metrycom has also developed other monitoring systems intended for grid operators, including one that helps locate ground faults. This system has been targeted for MV systems where a wire on the ground with less than 2 A current is hard to detect by existing protection systems but still represents a safety issue in rural areas. 

WILDLIFE PROTECTION

Preventing Wildlife Outages at Substations

Cornwall Seaway News: At around 6:30 p.m. power was cut to some 2,500 customers in Cornwall, Ontario (Canada), including many downtown. Utility officials say wildlife got into a substation on Adulphus Street, which led to the outage.

82

Outages triggered by interaction of wildlife with equipment at substations are hardly a new problem.

In places like the United States and Canada animals are said to regularly account for between 5 and 10 percent of all power outages – placing them among the 3 leading single causes. Depending on geography and climate, this proportion could be even higher in other countries.

Apart from the heavy financial losses suffered by industrial, commercial and residential customers, the direct cost of recovering from wildlife outages can

also be substantial. For power utilities in the U.S., for example, this figure has been estimated at up to US$ 20 million a year.

The wildlife problem at substations involves birds as well as climbing animals. Indeed, the two are related since bird nests attract scavengers such as raccoons, squirrels, feral cats and snakes. At the same time, substations are also inviting to a range of other creatures because they are a source of heat and vibration.

INMR talks with TransGard Systems, a manufacturer of a patented fencing designed to protect assets at substations against wildlife intrusion.

Since fencing around substations is usually focused on keeping people out, gaps may be large enough to allow foraging creatures to enter.

Photos: INMR Š

83

M

ost ideas are born of necessity. In the case of U.S. based TransGard Systems, the necessity behind the company’s main line of products came from a local power utility where the yearly toll of substation outages linked to squirrel incursions had become intolerable. The affected utility, based in upstate New York, approached a nearby manufacturer of electrified fences for livestock with the goal of developing a permanent solution based on the same principle.

The final product was the result of successive design refinements by both producer and user.

most effective solutions in each case may therefore not always be the same.

Says Reichard, who came to TransGard shortly before its facilities were re-located to Pennsylvania, “One of the wildlife protection devices on the market today consists of insulating covers installed over the live ends of equipment such as insulators or bushings. While these may prove effective some of the time, snakes and squirrels still manage to bypass them and cause outages.”

Moreover, security fencing installed around the perimeters of substations is intended mainly to prevent public access and not fully effective in preventing entry by all forms of wildlife. “Rocks from the base of substations are sometimes washed out by heavy rain,” explains Reichard. “Perimeter fences are not designed to prevent animal incursion. Gaps in the fencing, such as those near doors, make it easy for squirrels and snakes to enter a substation. Climbing animals can simply hoist themselves over a chain link perimeter fence.”

Photos courtesy of TransGard Systems

According to TransGard General Manager, Bill Reichard, the topic of wildlife outages must be looked at from different perspectives since there is no universal answer to every problem. For example, protecting live components from contact by birds roosting in a substation is one issue; keeping climbing or foraging animals away from live equipment may prove another problem altogether. The

Specialized additional fencing installed inside perimeter fence to protect vulnerable distribution assets at transmission substations.

84

In any case, Reichard emphasizes that TransGard panels are not intended to replace perimeter fencing. Rather, they increase the margin of safety by surrounding and protecting the equipment deemed especially vulnerable to wildlifeinduced outages. Because of the typical insulation distances involved, this applies mainly to voltage levels of 34.5 kV and below. However, in regions where long snakes have been known to make their way onto buswork, systems have been installed at 115 kV and even 138 kV. Most of the time, there is sufficient clearance for a second interior installation that is far enough away from the main perimeter fence that animals cannot jump over it on their way in. Aside from meaning that there is no need to replace existing perimeter fencing, this also offers the advantage of being able to further isolate distribution assets located within larger transmission substations. Notes Reichard, “our typical installations involve only perhaps 150 ft (45 m) of specialized added fencing versus the alternative of looking to somehow reinforce an entire substation enclosure that could run ten times longer.”

“Preventing wildlife outages has to be looked at from different perspectives since there is no universal solution to all potential problems.” panels that Reichard claims can be assembled at any location within a substation and in only a matter of hours. These modular panels are electrified with a DC pulse” that delivers a mild shock to any creature coming into contact with them. The shock is not enough to kill the creature but, given learned behavior, sufficient to discourage it from returning. With a current of less than 1mA along the interconnected panels, the system’s energy requirement is very low and, according to Reichard, amounts to less than a 1A draw on the station’s power supply.

The TransGard product concept is based on lightweight interlocking

Reichard contrasts the concept of specialized electrified fencing with insulated cover-ups, which he claims are never 100 percent reliable and tend to become less effective over time. They also suffer from the disadvantage that a substation typically has to be de-energized for their installation and this could involve the expense of bringing in a mobile temporary unit. At the same time, Reichard points out that it is sometimes impossible to cover devices such as the blades of switches while certain types of insulating covers may prevent heat from dissipating from components such as bushings, possibly accelerating ageing. “The system we offer,” he says, “can be installed in a matter of hours and there's never a need to take the station out of service.”

Apart from its specialized fencing, TransGard also markets devices intended to prevent squirrels from climbing poles located along the perimeter of substations or from gaining access along incoming conductors. These include aluminum pole wraps as well as devices that clamp onto conductors to establish a secure barrier against passage by squirrels.

Photos: INMR ©

For the moment, TransGard has concentrated its business within the U.S. and Canada, where there are now reportedly more than 2500 installations. Reichard is interested to expand internationally, especially into South America, but estimates that his level of domestic market penetration is still quite low. That leaves lots of opportunities to grow locally.

“When it comes to the problem of wildlife at substations,” says Reichard, “some utilities do not react until they have outages while others do it proactively. At the same time, birds and land creatures present much different challenges with different ideal solutions. While there does not yet seem to be a single best solution when it comes to birds, we feel that nothing is more effective than fine mesh electrified fence to keep everything else out, including snakes.”  85

CABLE ACCESSORIES

Photos: INMR Š

New Design of Dry Type Cable Terminations up to 170 kV

The traditional design of most outdoor cable terminations at the higher voltages has been based on a hollow core insulator made from either porcelain or a composite tube covered by silicone sheds. The insulators in both cases are equipped with a field control element made from silicone rubber and are filled with an insulating liquid or gas.

Completely dry terminations were first developed during the 1990s, driven by easier field installation, more stability in service and improved safety in the event of failure. This article, contributed by Matthias Freilinger of Pfisterer Ixosil in Switzerland, discusses new developments in dry cable terminations when it comes to materials, production technologies and customized designs.

86

Traditional cable terminations are based on self-supporting porcelain or composite insulator housing.

Electrical Design Requirements

The electrical design of the new generation of dry terminations was based on existing field control devices and therefore relies on the same type of semi-conducting deflector found in traditional outdoor terminations. This ensures that all internal standards governing distribution of electric field inside the silicone and at its interface with the XLPE or EPR cable would be satisfied – including maximum and minimum defined contact pressure of the silicone onto the cable.

Similarly, the outer sheds had to be designed according to the principles of IEC 60815-3 (2011) while taking into account possible changes in shed geometry due to expansion of the termination during the process of slipping it onto the cable. Moreover, all additional required standards, including IEC 60840 (High voltage cables and accessories up to 170 kV) and IEC 60071-1 (Insulation coordination), also had to be met.

Photo courtesy of Pfisterer Ixosil

Finding the right silicone material was essential to allowing such a product concept to be realized. For example, it was important that both the semi-conducting silicone used in the deflectors and the silicone used for insulation could be bonded together without adding any special agent to prevent de-lamination. The mechanical properties and flexibility of the cross-linked silicone were also important parameters given the different diameters of cable insulation. In addition, the liquid silicone material had to have the proper viscosity to achieve perfect filling of the mold during production.

Finally, to ensure the electrical performance required, inclined plane tests had to be performed on material samples according to IEC 60587 (3rd Edition, Electrical insulating materials used under severe ambient conditions – Test methods for evaluating resistance to tracking and erosion). For example, applying a constant voltage of 4.5 kV, the performance criteria of the material slabs included not exceeding 60 mA of leakage current as well as no tracking within 25 mm of the bottom electrode.

Finally, to help justify commercialization of this new concept, a decision was made that deflectors for the new generation of terminations had to be capable of being manufactured using existing molding and dosing equipment as well as the current production technology for terminations made from liquid silicone rubber (LSR).

Examples of flexible as well as supported dry terminations up to 170 kV.

Photo courtesy of Pfisterer Ixosil

The end result of this electrical design process was a flexible, dry type outdoor termination but without the self-supporting function.

Material Requirements

Photos: INMR ©

Among the main goals behind the development of an entirely new generation of dry type terminations has been better meeting customer requirements for easy installation and greater cost efficiencies while also maintaining all the advantages of the outdoor terminations currently in use. For this to become a reality, however, it became necessary to look at separating the mechanical requirements of the termination from its basic electrical design.

Fig. 1: Schematic of flexible dry-type outdoor termination.

Fig. 2: Mechanical design of self-supported, dry-type outdoor termination.

Fig. 3: Design of 360° rotating support on dry type termination.

Cantilever load test on post insulator selected for new dry type terminations.

87

Mechanical Requirements

Replacing standard outdoor cable terminations by the new dry types made it necessary to look at different mechanical aspects of their design.

In particular, the mechanical support function offered by either porcelain or a composite tube in standard terminations would have to be taken on by an adjoining post insulator. Such a solution, based on an FRP core with silicone sheds molded on, is a well-accepted insulator technology used widely on overhead networks. In this case, the post insulator must be capable of withstanding cantilever forces up to 4 or 5 kN to deal with maximum short circuit forces at minimum phase-to-phase clearances.

New dry type terminations lifted onto tower for final installation.

Photos courtesy of Pfisterer Ixosil

Lifting 170 kV dry type outdoor termination with no need for scaffold on tower.

Dry type flexible outdoor terminations for railway applications.

88

If all process parameters remained the same, this would mean longer vulcanization time and reduction in manufacturing capacity of each clamping unit.

To shorten the vulcanization process and maximize output with existing equipment, a new process was tested and set-up using proprietary curing technology. The result is that vulcanization time has been significantly decreased, with a corresponding increase in machine capacity and there has been higher stability in quality as well.

Advantages of Dry Type Outdoor Terminations

Dry type cable terminations offer a range of potential benefits in terms of installation, safety and reduced To guarantee a perfectly straight overall project costs. For example, termination under all installation they are designed for convenient and service conditions as well installation on the ground. Related as to compensate for all thermal line de-energization time is also movements of the conductor, it was minimized. Moreover, their stable determined that the cable had to be clamped at the bottom. Moreover, the construction allows them to be easily top fitting is equipped with a special and safely lifted onto towers.

At the same time, the 360° rotating spring design. This combination of clamping and spring assembly offers support construction compensates for cable torsion and ensures rapid fixing the benefit of being able to move of the lifted terminations. Since the complete cable system after scaffolding is no longer necessary installation. with these dry type designs, there are also appreciable savings in The cable screen is connected to construction and installation costs. the unit’s base plate to achieve a typical cable termination design.

In regard to performance, all the This base plate has 360° rotating parts to ensure that, if necessary, the usual risks in service are minimized with a dry type termination post support insulator and the top design since normal monitoring fitting can all be turned around the requirements linked to the presence termination. This makes it possible of internal insulating oil (e.g. to easily lift the terminations onto leakage, moisture behavior and towers. impact of ambient temperatures) are no longer needed. With no internal Manufacturing Facilities volume of oil or gas, there is no risk Production of the new design of of over-pressure caused by thermal dry type cable termination requires expansion of the fluid during an some modifications to the usual internal failure such as short circuit. equipment set-up. For example, the Dry type cable terminations therefore semi-conducting deflector used for field control must be embedded into fulfill the requirement of being a fully functional silicone termination ‘explosion proof’. and mounted on the core before Apart from all these advantages for filling the mold with pre-heated common XLPE power cables, dry silicone. This meant making certain type termination designs also offer a adjustments to the typical layout of solution for EPR type cables. clamping machine and dosing unit.

Another manufacturing consideration was curing time. The new design of dry type termination has increased thickness and therefore a greater volume of silicone material is used.

Applications

Dry type outdoor terminations can be grouped into two broad categories – flexible and self-supported – and cover a range of applications.

170 kV dry type outdoor cable termination installed on tower.

Photos courtesy of Pfisterer Ixosil

Movable preassembled energy back-up solutions for easy and quick installation to bypass transmission lines.

• Flexible Dry Type Outdoor Terminations In the 52 kV to 72 kV range and up to conductor cross-sections of 1200 mm2, flexible dry type outdoor terminations without added support function can be used in both power industry and railway systems.

Here, fixing of the termination is generally done by cable clamps and lugs instead of using head fittings.

From 72 kV to 170 kV and for conductor crosssections up to 2500 mm2, flexible terminations can be used for movable testing cables or as movable energy supply systems for building work stations on complete cable drums. In particular, installed movable systems with EPR cables consisting of flexible Class 5 conductor and dry type outdoor terminations are known for their flexible yet robust design.

• Self-Supported Dry Type Outdoor Terminations Self-supported dry type outdoor terminations are available from 72 kV up to 170 kV and for conductor cross-sections up to 2500 mm2 for power systems or for movable energy back-up solutions.

Robust construction under high cantilever forces, explosion-proof design and capability for convenient on-ground installation at towers have together resulted in rapid acceptance of this design concept to replace conventional outdoor terminations that use hollow core composite insulators in the lower high voltage range up to 170 kV.  89

FITTINGS

Manufacturer Expands Production of Insulator Fittings

P

ower engineers focus on ensuring the proper design of insulators so that they can offer reliable performance under a wide range of different operating conditions. While shed geometry and material are critical factors in achieving the necessary electrical characteristics, the insulator’s metallic end fittings allow it to endure the long-term mechanical stresses, even when operating in corrosive service environments.

INMR visits Jinan Meide Casting, a foundry based near the northeastern Chinese city of Jinan and which in recent years has diversified its product base to include insulator fittings.

90

Examples of deteriorated fittings from exposure to coastal pollution in Crete (left) and Florida.

Jinan Meide’s history goes back more than 50 years to when the firm first started manufacturing hardware for plumbing. By the early 1990s, the company began to supply its products overseas and the export sector grew so quickly from then on that today it accounts for nearly two-thirds of the 340,000-ton annual output.

“Meide diversified into electrical fittings in 2007,” explains International Sales Director, Su Zhaoxia, “with one of our first customers in the field being a large European insulator manufacturer. Since that time we have focused more and more on growing this business by applying our production expertise and technologies to make fittings more efficiently and also more economically.”

When it comes to castings such as fittings, about half of the typical production cost is spent on energy while the other half comes from the materials. Achieving gains in both areas, notes Su, has required substantial R&D work. For example, Meide’s latest generation of castings is designed for lower temperature production

and that has helped not only reduce energy costs but also improved yield. Su reports that developments of this sort now save the equivalent of about 10,000 tons of raw materials each year.

According to Su, Meide’s strategy to control the quality and dimensional consistency of its insulator fittings has been to manufacture all the necessary molds internally. These are either designed to common industry standards such as IEC or ANSI but can also be customized to meet any special requirements. To accomplish this, nearly 40 CNC machining centers make the molds, each of which offers a typical life of from 3 to 6 months, depending on usage.

“A fully automated production line has been designed in order to shorten the normal turnaround time for fittings.”

Says Su, “Since our entire production of insulator fittings is done in-house, we can take full responsibility over final quality. At the same time we

Automated production allows shortening normal production turnaround times.

91

have designed what we feel is a unique automated line that allows us to shorten normal turnaround times. We see these as our main competitive strengths within this industry.”

Depending on application, Meide castings for insulator fittings such as caps, bases and ball & sockets are made from a range of different materials, from malleable to ductile iron to the newer low temperature grade of ductile iron. These are then offered with or without a galvanized zinc collar depending on whether they will be used for AC or DC. The hot dip, galvanizing layer is available in three different classes of thickness from the typical 87 to 107 µm to as much as 140 µm for especially severe environments. Su explains that these thickness levels are guaranteed by inspection of every unit produced while the iron used in all fittings is continuously monitored by spectrometers to ensure consistency of composition. 

Rubber Injection Moulding Machine for Long Rod Insulators & Cable Accessories

92

The Best & Most Practical Solution for Top-Level Insulator, Arrester & Cable Accessory Producers LSR Injection Moulding Machine for Hollow Core Insulators

Dekuma is now the world leader in providing total equipment solutions for first level insulator manufacturers.

• Longest history as first rubber injection moulding machine for insulators in China

• Largest rubber injection machine in the world

• Most reliable service in the industry • World leader in mould design and manufacturing

• Suitable for HTV & LSR material • German-designed dosing and metering accessories provided for LSR insulator production

• Injection volume up to 100,000 cc; Clamping force up to 2,800 tons

• Mould platen size designed as per customer needs

Dekuma Rubber & Plastic Technology (Dongguan) Ltd. (Member of Hong-Kong based, Cosmos Machinery Ltd.) Yinzhu Road, Zhouwu Industrial Park, Dongcheng District, Dongguan, Guangdong Province, China P.C. 523118 Tel.: +86-769-22667183 Telefax.: +86-769-22667227 Email: info@dekuma.com www.dekuma.com