ANTTI KESKINEN STRATEGY FOR MEDIUM VOLTAGE CABLE … julkiset dtyot/Keskinen_Antti_julk.pdf · ERP...

ANTTI KESKINEN STRATEGY FOR MEDIUM VOLTAGE CABLE CONDITION MANAGEMENT Master Of Science Thesis

Examiner: Professor Pekka Verho The examiner and the subject was approved in the Faculty of Computing and Electrical Engineering council meeting on 9 March 2011

II

ABSTRACT TAMPERE UNIVERSITY OF TECHNOLOGY Master of Science Degree Programme in Electrical Engineering KESKINEN, ANTTI: ‘‘Strategy for medium voltage cable condition management” Master of Science Thesis, 83 pages, 4 Appendix pages March 2011 Major: Power Systems and Market Examiner: Professor Pekka Verho Keywords: Medium voltage cable, cable diagnostic, condition management, profitability, operation management In the year 2006 Vattenfall Verkko Ltd. (later VFV) started developing a weatherproof network. For this reason, cabling has become a prevailing construction method. The cabled medium voltage network has formed only a small part of the whole network earlier, but the situation will be changing quickly in the future. This creates the need for systematic condition management of the cable network. The purpose of this thesis has been to analyse and bring forth suitable methods, which can be used in the cable network’s condition management, from inside the Vattenfall concern and elsewhere in the world. The suitable methods have been examined by benefit calculations. The profitability analyses have been implemented through the network business result’s formation, when it has been possible to verify a straight connection between cable diagnostic and result of the network business. In the case of VFV, the distribution network is located within quite a large area in geographical aspect and this causes challenges for the development of the cable network condition management strategy. The geographical factor has been taken into account in the benefit calculations, and the analysis has been made for different type of areas. The division of these areas is based on the reliability criteria which can be regarded as a meter for the customized reliability and which give planning criteria in the future. The construction of the network, construction conditions, customers and their placing depends a lot on the area, as the condition management strategy establishment has to be based on the prevailing conditions of the each area. In rural areas, the construction of the network creates a challenge for the achievement of a long life time for the cable, which is due to the challenging construction conditions. For this reason, quality verification after the installation is emphasized in condition management in rural areas. Other condition management operations have to be tried to eliminate by the structure of the network because the cable diagnostic utilization is unprofitable in rural areas at the moment.

By cable diagnostic pecuniary advantages in urban and city areas can be achieved, whereby the condition management strategy differs in the case of rural areas. The strategy can be implemented in e.g. primary substation level. The operations which will be performed during the life-cycle are based on profitable analysis, structure and the age of the network. The operations management must be integrated in the network information system, so that the indirect costs can be minimized and operations become more profitable.

III

TIIVISTELMÄ TAMPEREEN TEKNILLINEN YLIOPISTO Sähkötekniikan koulutusohjelma KESKINEN, ANTTI: “Keskijännite kaapelin kunnonhallinta strategia” Diplomityö, 83 sivua, 4 liitesivua Maaliskuu 2011 Pääaine: Sähköverkot ja -markkinat Tarkastaja: Professori Pekka Verho Avainsanat: Maakaapeli, kaapelidiagnostiikka, kunnonhallinta, kannattavuus, toiminnanohjaus Vuonna 2006 Vattenfall Verkko Oy aloitti säävarman verkon rakentamisen. Tämän myötä verkon vallitsevaksi rakennustavaksi on tullut kaapelointi, eli aina kaapeloidaan kun se on mahdollista. Tällä hetkellä kaapeloitu keskijänniteverkko muodostaa vain pienen osan verkkomassasta, mutta tilanne tullee muuttumaan nopeasti, mikä puolestaan aiheuttaa tarpeen kaapeliverkon systemaattiselle kunnonhallinnalle. Työn tavoitteena on ollut tutkia ja tuoda kaapeliverkon kunnonhallintaan sopivia menetelmiä konsernin sisältä ja muualta maailmasta. Löydettyjä menetelmiä on tarkasteltu kannattavuuslaskennan avulla. Kannattavuustarkastelut on toteutettu verkkoliiketoiminnan tuloksen muodostumisen kautta, jolloin on voitu todentaa kaapelidiagnostiikan yhteys verkkoliiketoiminnan tulokseen. Vattenfall Verkko Oy:n jakeluverkko sijoittuu maantieteellisesti melko laajalle alueelle, jolloin haasteet keskijännitekaapeliverkon kunnonhallinta strategialle kasvavat. Kannattavuustarkasteluissa onkin otettu huomioon tämä maantieteellinen tekijä, jolloin tarkasteluita on tehty erityyppisillä alueilla. Tarkasteluissa käytetty aluejako perustuu toimitusvarmuuskriteerien mukaiseen aluejakoon, jota voidaan pitää asiakaskohtaisen toimitusvarmuuden mittarina ja josta saadaan verkon suunnittelukriteerit tulevaisuudessa. Verkon rakenne, rakentamisolosuhteet, asiakkaat ja asiakkaiden sijainti poikkeavat siis eri alueilla verkossa, jolloin kunnonhallinta strategian luominen täytyy perustua alueella vallitseviin olosuhteisiin. Haja-asutusalueella verkon rakentaminen verrattain vaikeissa olosuhteissa luo haasteen kaapelin pitkän elinkaaren saavuttamiselle. Tällöin kunnonhallinta on painotettava elinkaaren alussa asentamisen laadun varmistukseen. Muu elinkaaren aikainen kunnonhallinta on pyrittävä eliminoimaan verkon rakenteen avulla, koska kaapelidiagnostiikan hyödyntäminen ei ole taloudellisesti järkevää haja-asutusalueella ainakaan tällä hetkellä. Taajama- ja kaupunkialueilla kaapelidiagnostiikalla voidaan saavuttaa taloudellisia etuja, jolloin kunnonhallinta strategia muodostetaan eritavalla kuin haja-asutusalueen tapauksessa. Strategia voidaan toteuttaa esimerkiksi sähköasematasolla. Elinkaaren aikaiset toimenpiteet määräytyvät kannattavuustarkastelun, verkon rakenteen ja iän perusteella. Toiminnanohjaus tulee integroida verkkotietojärjestelmään, jolloin kunnonhallinnan välilliset kustannukset saadaan minimoitua ja toiminta entistä kannattavammaksi.

IV

PREFACE This work was one part of the bigger Smart Grids and Energy Markets (SGEM) project, where the Vattenfall Verkko Ltd. is involved. SGEM-project consists of different kinds of work packages, which this work belongs packages 2.1 Large scale cabling in distribution network and 4.5 Condition management as real-time process. The supervisor of this work has been worked professor Pekka Verho from Tampere University of Technology. I want to give thanks for him to interesting conversations, which I have to go through with him during this work. By these conversations I got tools to solve the hardest part in my work successfully.

From Vattenfall Verkko Ltd., where I got the possibility to carry out my thesis, I want to give thanks for all team of strategic network planning and “substation brothers”. Thanks belong to you from good working environment and expert comments. Special thanks belongs to Heikki Paananen who was the controller of my work. You gave me great opportunities to represent and to be involved in events, where rarely meet thesis workers. I also want to thanks my superior Sauli Antila for that you gave me possibility to work good team and to learn a lot from operations in the distribution network company.

I am also very grateful for my common-law wife Jenni from good, understanding and encouraging support which from you I got during my studying. Thanks also to my children Ella and Hilla which gave the counterbalance to the school.

Tampere 21.3.2011 Antti Keskinen

V

TABLE OF CONTENTS

1. Introduction ............................................................................................................... 1 2. The basic features and needs of the cable condition management ............................ 3

2.1. Construction strategy ........................................................................................ 4 2.2. Medium voltage cable ....................................................................................... 5 2.3. Medium voltage cables accessories .................................................................. 6 2.4. Installation and installation methods ................................................................. 8 2.5. Deterioration and damaging mechanisms ....................................................... 10

2.5.1. Mechanical and installation damage .................................................. 10 2.5.2. Operational damage ........................................................................... 11 2.5.3. Age-related deterioration ................................................................... 11

2.6. Statistics .......................................................................................................... 12 3. MV cable testing ..................................................................................................... 15

3.1. Test methods to check the quality of installation ............................................ 16 3.1.1. Insulation resistance test .................................................................... 16 3.1.2. Sheath integrity test ........................................................................... 16

3.2. Measuring tools to assess condition of MV cable and accessories ................. 18 3.2.1. Dielectric spectroscopy measurements .............................................. 19 3.2.2. Dissipation factor measurements ....................................................... 20 3.2.3. Partial discharge test (PD) ................................................................. 22

3.3. Strategies for field testing medium voltage cables ......................................... 25 3.3.1. Field tests after the installation .......................................................... 26 3.3.2. Field tests after fault .......................................................................... 28 3.3.3. Field tests and diagnostics before reinvestment ................................ 29

4. Life cycle costing .................................................................................................... 31 4.1. Determination of LCC .................................................................................... 31 4.2. Economic evaluation methods for LCC .......................................................... 31

4.2.1. Simple payback method ..................................................................... 32 4.2.2. Discounted payback method .............................................................. 32 4.2.3. Net present value method .................................................................. 32 4.2.4. Internal rate of return method ............................................................ 33

5. Actions of cable condition management in real-time process ................................ 34 5.1. Permanent continuous monitoring system ...................................................... 35 5.2. Portable partial discharge monitoring ............................................................. 36

5.2.1. Online condition monitoring process ................................................. 37 5.2.2. Monitoring process implementation in large distribution network ... 39 5.2.3. Sensors’ positions in distribution network ........................................ 42

6. Strategy of the MV cable life-cycle maintenance ................................................... 44 6.1. Definition of strategy ...................................................................................... 44 6.2. Maintenance .................................................................................................... 44

6.2.1. Run to breakdown maintenance (RTB) ............................................. 46

VI

6.2.2. Preventive maintenance (PV) ............................................................ 46 6.2.3. Time based maintenance (TBM) ....................................................... 46 6.2.4. Condition-based maintenance (CBM) ............................................... 47 6.2.5. Reliability-centred maintenance (RCM) ............................................ 47 6.2.6. Total productive maintenance (TPM) ................................................ 47

6.3. Optimal maintenance ...................................................................................... 47 7. Key figures and results formation in the network business .................................... 49

7.1. The reliability key figures in delivery ............................................................. 49 7.2. The interruption costs formation ..................................................................... 50 7.3. Cost of cable diagnostic .................................................................................. 51

7.3.1. Cost of investment ............................................................................. 51 7.3.2. Cost of operation ................................................................................ 51 7.3.3. Cost of interruption ............................................................................ 52

7.4. Allowed return on network business ............................................................... 52 8. Benefit calculations ................................................................................................. 55

8.1. Cost and profitability of the portable part time online diagnostic .................. 55 8.1.1. Suitability for Rural areas .................................................................. 57 8.1.2. Suitability for Urban areas ................................................................. 58 8.1.3. Suitability for City areas .................................................................... 60

8.2. Cost and profitability of the permanent online diagnostic .............................. 63 8.2.1. Life cycle cost of the permanent online diagnostic ........................... 63 8.2.2. The profitability of permanent online diagnostic .............................. 65

8.3. Cost and profitability of the off-line diagnostic .............................................. 66 8.3.1. The cost of off-line measurements .................................................... 67 8.3.2. The profitability of off-line measurements ........................................ 68

8.4. Conclusion of the benefit calculations ............................................................ 71 9. Operations management .......................................................................................... 73

9.1.1. Cable database ................................................................................... 74 9.1.2. Cable condition assessment ............................................................... 74 9.1.3. Prioritization with the economical aspect .......................................... 75

9.2. Planning of the maintenance operations in the cable network ........................ 76 9.2.1. Strategy for rural areas ....................................................................... 77 9.2.2. Strategy for urban and city areas ....................................................... 78

10. Conclusions ............................................................................................................. 80 10.1. Conclusions ............................................................................................... 80 10.2. Further study ............................................................................................. 81

References ....................................................................................................................... 84 Appendix 1 - ................................................................................................................... 89 Appendix 2 - ................................................................................................................... 91 Appendix 3 - .................................................................................................................... 92

VII

ABBREVIATIONS AND NOTATION AXALL-TT Three phase medium voltage cable CAIDI Customer Average Interruption Duration Index CBM Condition-based maintenance

DC Direct current DCO Disadvantage caused by electricity supply outages DEA Data envelopment analysis ERP Ethylene propylene rubber HFCT High frequency current transformers IEC The International Electrotechnical Commission IRR Internal rate of return LC Leakage current LCC Life cycle costing LLLP Low loss linear permittivity MAIFI Momentary Average Interruption Frequency Index MV cable Medium voltage cable NIS Network information system NPV Net present value PD Partial discharge PDIV Partial discharge inception voltage PDEV Partial discharge extinction voltage PE Polyethylene PILC Paper-insulated lead-covered cables PV Preventive maintenance RCM Reliability-centred maintenance RMU Ring main unit RNA Reliability based network analysis RTB Run to breakdown maintenance TBM Time based maintenance TLC Transition leakage current TPM Total productive maintenance SAIDI System Average Interruption Duration Index SAIFI System Average Interruption Frequency Index SFA Stochastic frontier analysis

VIII

VDP Voltage dependent permittivity VFV Vattenfall Verkko Oy VLF Very low frequency XLPE Cross-linked polyethylene ε’ Imaginary part of complex permittivity ε’’ Real part of complex permittivity

Loss angle

Probability distribution

C Capacitance C flow Cash flow during the year C investment Initial investment cost CLCC The sum of the costs of permanent online diagnostic system CINV The costs of control units and sensors CT The costs of training and installation CSYNC The costs of synchronization CM The cost of monitoring CRV The residual value of the system C savings Annual operating savings Ct The net cash flows in year t C uncovered Unrecovered cost at start of the year I Current IR The Resistive component of the current IC The Capacitive component of the current nj The number of the customers who experience the

interruptions i Ns The total amount of customers R Resistance r The rate of interest tij The time without electricity that customers j have to spend because of the interruptions i t Year T Last year of the period T before Year before full recovery T payback Payback time in years U Voltage U0 Nominal voltage

1

1. INTRODUCTION

During the last few years, the reliability of the electricity supply has become a subject of discussions at many different levels of society. The reason for this has been storms and snow loads caused by harsh winters, which have caused problems in the electricity distribution. On the other hand, the customers’ requirements of the quality of the electricity supply have increased and the control of the network business has become tougher. It has caused a need for the development of the operations and the search for new network solutions. In order to meet these challenges, Vattenfall Verkko Ltd. has chosen large scale cabling as its building strategy of the network which aims at weatherproof network and comprehensive long-term improvement of the network. At the moment, a big part of the medium voltage network consists of overhead lines and the share of cables is relatively small. In the case of the overhead lines, the need of the replacement investment will be voluminous in the near future and this will change the structure of the network quite fast. The need of replacements is due to the age structure and the mechanical condition of the overhead line network. Earlier Vattenfall Verkko Ltd. has carried out a master’s thesis, in which was determined the boundary conditions for the medium voltage cables which are now in use. The purpose of this thesis is to continue forward within the subject towards the condition management of the medium voltage cable network, which is the base for the networks’ asset management. It may have a huge influence on the result of the network business in the future, when the total lengths of the cable network increases. Many planning tasks in the network business are based on the technical and economic boundary conditions which under the optimization have to do. The development of the strategy for the medium voltage cable or cable network can also be considered to be included in these tasks. The planned operations forms’ safety, the structure of the network and its components determines the technical boundary conditions for the operations together with available cable diagnostic equipments, and the benefit calculations determines the economic criteria for the operations. The life time of the medium voltage cable is quite long. For this reason, the most important factors which may have an effect on the cable’s life time, and may cause maintenance operations, have been considered in this thesis. The factors which may have influence on the length of the life time of the cable network have been examined by statistical distributions. The examination has focused on the external and internal factors. After the identification the different types of measurement methods have been introduced, how suitable they are for the cable network and how they can be exploited for the conditions management in the cable network. The different types of field

2

measurement strategies for different situations have been presented in chapter 3. These strategies are partly based on the methods which are used by Vattenfall Service in Sweden. In addition to these methods, the worldwide knowledge and supply of the cable diagnostic is also attempted to be determined in this thesis. Few service provider solutions suitability has been examined as a part of cable network real time process. The profitability of the cable diagnostic has been examined by benefit calculations. The different types of measurement methods have been considered in present cases of the moment and cases of the future network. The profitability analysis has been implemented by using result formation in the network business. By these calculations it is possible to determine the most important factors which have the biggest effect on the profitability of the cable diagnostic and the risk levels which are willingly accepted if nothing is done or the contrary if the operations fail. The benefit calculations also give good alignment, of what kind of condition management strategy is suitable for different kinds of reliability criteria areas. The creation of the condition management strategy for the cable network and what it requires for a Network Company have been considered at the end of this study. In addition, the condition management strategies have been established for cable networks, which are located in different kinds of reliability criteria areas. The main purpose of these strategies is to give guidelines for the condition management. The model from comprehensive operations management strategy is also presented, which allows conditions management implementation in the case of a large cable network.

3

2. THE BASIC FEATURES AND NEEDS OF THE CABLE CONDITION MANAGEMENT

The life-cycle of a medium voltage cable is quite long and this might make medium voltage underground network condition management challenging. The condition management process can be regarded as starting when a strategic decision about the cabling of a medium voltage network is made. There are a lot different facilities which could affect the life-time, reliability and cost of an underground network. For example insulation properties of MV cable, installation methods, soil, loading rate, the structure of the network et cetera. The selection of a cable and its accessories can have great influence on the network’s life-cycle length and cost. Accurate and high-quality components reduce the need for maintenance and faults which are caused by manufacturing defects. In this case, the correct choice of materials can be regarded as one of the most important parts of the cabling process, which may also affect maintenance costs in the future.

When the strategic decision of cabling is made, and the components which are used are defined in a higher level, the next steps in the process are design and installation. Designing can affect the cable installation conditions, and thus affect the cable with harmful external factors. The design can also be a harmful influence for internal stresses of the cable, e.g. a situation where cable is under rated and the load is high. The contractor's competence and responsibility is also an important factor that may affect the future maintenance of the cable. Installation errors and errors which cause damage to the cables can shorten their life significantly.

Cabling process and its components have been examined in the following. Also what have been examined are the factors that may cause the need for maintenance in cables and accessories.

2. The basic features and needs of cable condition management 4

2.1. Construction strategy

Vattenfall Verkko Ltd. has chosen large scale cabling as its building strategy of the network which aims at weatherproof network and comprehensive long-term improvement of the network. The new installation methods, cheaper prices of the components and installation work have brought the costs of cabling and overhead lines construction closer to each others. Figure 2.1 presents the cumulative length of the medium voltage cable network from the year 2005 to the year 2010, and in addition shows a prediction of the coming years’ developments in the case of Vattenfal Verkko Ltd.

0

500

1000

1500

2000

2500

3000

3500

4000

4500

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015

Year

Acc

umul

ated

cab

le le

ngth

(km

)

Figure 2.1 VFV’s accumulative medium cable length each year and estimation of the future.

The total length of the medium voltage cable network was about 1470 km at the end of year 2009, and it represents about 7 % of total length of the medium voltage network. A big part of that about 7 % was located in urban and city areas, where the density of the customers is quite high. Figure 2.1 shows that the structure of the network will change fast. If the total length of the network will remain almost constant and yearly replacement speed is 450 km per year, then the cabling rate has increased about 21 % in the end of year 2015. The strategic construction decision and the old cables’ location create the biggest input for the condition management strategy development for the medium voltage cable network.


2.2. Medium voltage cable

The one part of a cost effective and reliable distributions network cabling strategy is, to find a good medium voltage cable manufacturer. Manufacturer production methods, materials, construction solution and knowledge of cable testing might have a major impact on the quality of a cable network in a short-time and a long-time period. The cable also has to be easy to handle and highly durable. Other important features, which help cable installation, are low specific weight and high flexibility at low temperatures. Figure 2.2 represents the common MV underground cable construction used in Vattenfall Verkko Ltd.

Figure 2.2 AXAL-TT PRO 12/20(24) kV MV cable from Ericsson. [1]

The conductor is made of aluminium and it is watertight in longitudinal direction. The conductor shield and the insulation shield are semi conducting and extruded. Longitudinal water tightness has been created by using swelling powder and swelling yarn that prevent the spread of water in the whole design as cross sectional water tightness consists of aluminium tape which is glued to the outer sheath. AXAL-TT PRO cable’s insulation is dry cured cross-linked polyethylene, commonly abbreviated XLPE. Outer sheath has been made of black polyethylene composite with a hard outer layer and an impact absorbing inner layer. Outer sheath also includes meter marks and Kevlar tear threads to simplify installation and maintenance work. [1] Nowadays the most common insulate material in MV cables is XLPE, but there are also older cables especially in urban areas, where cable insulation material is paper and oil, known as PILC cables. At the moment only about 6,6 % of VFV’s distribution


network is cable and a very little percent of this is PILC cable. Therefore, it is possible to limit this examination only to XLPE insulated cables. XLPE cable’s benefits are that it is more cost efficient in operation than thermoplastic analogues, because it provides better reliability and higher operating temperature than the thermoplastic materials.[2] If compared to older impregnated paper systems, XLPE cable also has lower environmental and maintenance requirements.

The major problem associated with medium voltage XLPE insulated cables is deterioration by water trees, and it is sometimes the main reason for insulation failure in XLPE cables after a long service period.

2.3. Medium voltage cables accessories

In the aspect of cable networks condition management, joints and terminations have an important role. Joints and terminations are very often the weakest points in a cable network, because the stress of electric field is bigger than elsewhere. Joints and terminations are also likely sources of defects leading to cable system failures. Even if the manufacturing process of joints and terminations are high quality, their failure rate is still high, due to their manual assembling. [3] In figure 2.3 below are presented joints and terminations portion of MV cable network failures.

Figure 2.3 Joints and terminations portion of MV cable network failures. [4]

Figure 2.4 presents electric field distribution without field controlling. Figure 2.5 presents electric field distribution with field controlling in the end of the cable. Figure 2.4 shows, that the stress of electric field hitting a smaller area than figure 2.5. The bad electric field control is one of the main reasons why termination or joint fails.


Figure 2.4 The electric field distribution without field controlling. [5]

Figure 2.5 The electric field distribution with field controlling. [5]

Manufacturers offer wide range of different types of joints and terminations for different construction and maintenance situations. The joint is as reliable as normal cable, if the contractor’s competence to do this is excellent and the working environment is clean. Figure 2.6 illustrates weak places and typical partial discharge areas of termination and joint. As shown, the areas are the right ends of the cable where the stress of the electric field is the biggest.


Figure 2.6 The most common partial discharge sources in terminals and joints. [6] Consequently, improper stress control is the main reason which leads to failures in cable termination. Other risk factors may be tracking, erosion or weathering of external leakage insulation and ingress of moisture. The joints have almost the same factors as termination, but voids in insulation and inadequate insulation may be added to the list. [7]

2.4. Installation and installation methods

A large part of cable damages occur in the cables’ installation situation. Another risk factor is the challenging and variable soil in Finland. In the past there was only one installation method and it was normal digging. The advantage of digging is that backing sand can be used when filling the hole. In this way it is possible to avoid harmful rocks and the effect of earth freezing. It is also easy to control deepness of cable installation by using the digging method. The digging method has its own risks, especially when the soil contains big rocks. It is difficult to obtain smooth trench bottom, which may lead to a situation where water will rinse the back sand away from under the heavy filling earth. If the filling earth contains rocks, it may cause damage when the filling earth presses on the cable. [8] In figure 2.7 non electrical damage portion of MV cable network failures are presented.


Figure 2.7 Non electric damage portion of MV cable network failures. [4]

Trench digging by using an excavator is quite slow and the growth interest of cabling has caused the need to find new alternative installation methods. Faster installation and cost structure of cable networks construction, which are very contractor-oriented, have also influenced the interest in new methods of installation.

At the moment, the most widely used installation method is cable ploughing straight to ground and below the pilots experiment is new earths sawing method. Cable ploughing means that at the same time when the machine ploughs the ground, the cable is being directed in the channel. The ploughing method is a fast but a blind method. It is impossible to see how the cable settles in its channel. In Figure 2.8 cable ploughing and the cable plough are presented.

Figure 2.8 The cable ploughing and cable plough. [9] Today, about 50 % of VFV’s cable installation is made by ploughing. The success of ploughing depends a lot on the contractor’s competence. If the contractor has good ploughing skills, then the ploughing can be even a better installation method than digging. Ploughing requires precise environmental and pre-ploughing assessment from the contractor [8].


Figure 2.9 below shows two solutions which can be used in making the cable trench.

Figure 2.9 Earth saw [10] and earths cutter. [11] New solutions allow the achieving of a better cable trench than before. The earth saw is possible to be used throughout the year in cable trench making without loss of frost.

2.5. Deterioration and damaging mechanisms

The deterioration of dry cured XLPE insulated MV cables can be divided into three different categories, mechanical, chemical and electric deterioration. This kind of cables insulation does not reach the final state immediately after the manufacturing process. It can take years for the insulation structure to be stable and this exposes the XLPE cable to harmful effects which can be mechanical, chemical or electrical. Damaged cable insulation is weaker to resist partial discharge because the electric field stress is bigger than in normal state. Partial discharges are very harmful for XLPE insulation. The chemical deterioration weakens the polymer insulation. In chemical reaction temperature, oxygen and radiation are present. As a result of chemical reaction long polymer chains break which is known as depolymerisation. There can also be new cross-linking bridges formed and it makes isolation more brittle. [12] The electrical deterioration happens as a result of water trees which generate electrical trees and lead to partial discharge. The most common electrical deterioration is a local effect. Low electric field intensity and long development time are common for electrical degradation mechanisms in medium voltage cables [12].

2.5.1. Mechanical and installation damage

Components of MV cable network are very vulnerable during its operation chain, which includes manufacturing process, storing, transporting, handling and installation. Carelessness in cables different operation may lead to cuts, scrapes, too large sidewall


force and water penetration inside the cable. These undesirable damages may affect immediately and make cable energizing impossible or the damages affects long-time period in cable insulation and creates partial discharge which finally leads to failure. Damage to the outer sheath and cross sectional water tightening may allow water to enter the space between the outer sheath and the insulation shield. This may have a corrosive effect to the aluminium screen. If water has entered inside the cable structure and the internal pressure forces to carry it along the cable, the water may create conductive paths and this may lead to short circuit and failures. [3] Damage suffered by digging may also be included in this mechanical and installation damage category.

2.5.2. Operational damage

Overload, short circuit currents and extensive load changing are known as effects, which may cause operational damage. Extensive load changing causes temperatures changing and it leads to both expansion and contraction of the cable. This can cause conductor shields hardening and cracking. Extensive load changing has also harmful effect on cable system terminations and joints. In the interface of insulation and semi conducting conductor shield or insulation shield may arise voids and gaps, which can lead to partial discharges. Overloads and short circuit currents bring high temperatures, which are damaging for the cable systems components. These big temperature changes are more harmful than the changes of the load cycling. High temperatures create a risk that the cable system fails, it can damage the insulation and change its geometry. Insulation geometrical changes lead to a deterioration of dielectric strength. [3]

2.5.3. Age-related deterioration

The cable structure deteriorates over time. The Insulations and semi conductive shields attachment may deteriorate. This can lead to a situation where deterioration creates voids and gaps, which allows water penetration inside the cable. The moisture in the insulation may cause water treeing. The water tree generates when the moisture inside the insulation starts to move electric fields direction. There are two types of water trees which are called bow-tie and vented trees. Bow tie water trees initiate from impurities and voids within the bulk insulation and tend to grow in two directions. Bow tie trees reach a limiting length of some tens of μm, and do not have a significant effect on degradation at the low electric field stresses used in distribution cables [13]. A vented water tree arises at the interface between the semi conductive screens and insulation. It grows in only one direction and has more harmful effects on insulation than bow-tie water trees. If water trees grow enough they can change electric trees, which causes partial discharges and finally cable fails. [3]


2.6. Statistics

This chapter presents some MV cable statistics in Sweden. The medium voltage cables failure statistic has an important role in cable condition management. As has been revealed earlier, cables’ failing is the sum of many different factors. An individual explanatory factor is impossible to find and it makes condition management very difficult. Even determining a sinlge cable condition can be a difficult process, not to mention determining the bigger cable networks condition. However, historical information and cable statistics help to evaluate the condition of the cable network and this way supports decision making. On the other hand, it is to be understood that the materials which are used in cables and the manufacturing processes develop all the time, so direct connection cannot be established between new and old cables which are the basis of the historical data. However, rough estimates of possible behaviour of the cable can be made for certain insulation materials fault trends. Using these estimations, it enables preparing for maintenance, inspection or replacement operations. Following statistics are based on information collected in Vattenfall Distribution Sweden. The distribution networks large scale cabling has started earlier in Sweden than here in Finland. The statistics include construction volume and failure rates from Swedish cable networks. In this case, the main goal is to find some contexts which can help to understand MV cables failing and use it later in chapter five in determining strategy of medium voltage cable life-cycle maintenance. Figure 2.10 presents failure rate sorted by construction year for both PILC and XLPE cable. The red dots represent the total installed cable length for each year.

Figure 2.10 Failure deemed to be due to material or manufacturing defects, construction year1961-2000 Vattenfall Distribution Sweden. [14]


In figure 2.10 can be seen that the older cables failure rate is bigger than the cables which have been installed later than year 1994, but in some cases older cables are as good failure rate than later installed cables. Figure 2.11 below presents the absolute failure rate for XLPE and PILC cables.

Figure 2.11 Failure deemed to be due to material or manufacturing defects, number of failures by age of cable, Vattenfall Distribution Sweden. [14] The high absolute failure rate for XPLE cables with the age 0-3 years old are probably due to the huge amount of new cables laid during the years 2005-2008 [14]. If these youngest cables are left out of consideration, it seems that about 8-25 years old XLPE cables have the highest failure level in Sweden. Figure 2.12 presents different type of cables’ installation lengths and failure frequencies.

Figure 2.12 Failure deemed to be due to material or manufacturing, cable type, Vattenfall Distribution Sweden. [14]


It shows that there are few very critical cable types, e.g. AXKJ and FXKJ. The cables AXAL and AXLJ which are used in Finland have low failure frequencies, but the reason for this can be the cables short in-service time. It is difficult to find a straight context which explains cable failing, but when creating a reliability model for MV cable, the cable age and type are of some importance. Statistics before contained only the XLPE cables statistic from Sweden and failures deemed to be due to material or manufacturing, which contains the installation defects too. Figure 2.13 presents the whole XLPE cables underground systems failure profile.

Figure 2.13 Overall average underground distribution systems failures with XLPE cables. [7]

43 %

20 %

4 %

19 %

14 %

Physical damages Treeing in XLPE insulation Environment Fault cable termination and joints

Other (overload, short circuit effect surgevoltages)

15

3. MV CABLE TESTING

The medium voltage cable network forms a large part of the distribution company’s physical capital. The medium voltage network has a huge influence on interruption which customers suffer, because the biggest part of the interruptions is a result from medium voltage network faults. Defects cause harm to customers, and to the distribution network company. In the overhead lines network faults localization are much easier than in underground network. In some cases the overhead line network faults generation can be detected before the fault occurs. This can be achieved by using visual inspections, which can be done by a helicopter or walking and this way find the critical target e.g. trees over the line. Thus, trees can be removed before the fault happens. In the underground network causes of defects, the structure of network and the physical position are very different than the overhead lines network. It leads to the situation where normal overhead lines fault prevention methods are unusable. Also underground networks disconnector density is lower, and troubleshooting the calculated distance is more difficult to determine than overhead line network. The underground networks condition management is based on electric measurements, which are due to its physical location. The objective of the measurements is to provide information for the networks’ operator on the condition of the network. By means of measurement, it is attempted to determine e.g. the condition of the cables insulation, because the good condition of the insulation is a prerequisite for the functioning of the cable. This emphasizes the case of a three-core cable even more. Measurements are used to verify the condition of cables and aims to predict the remaining life-time. At the moment distribution network operators are very interested in diagnostic tools for cable systems condition management. Knowledge of the network’s condition helps to make the right maintenance and investment decisions.

3. Test methods to assess MV cable deterioration 16

3.1. Test methods to check the quality of installation

After the cable laying but before it is put into service, the quality of the installation has to be checked. At this moment VFV requires its contractor two quality checking tests; insulation resistance and sheath integrity test. Tests are required because the cables production, transportation, installation method or rocks and ground frost may cause damage in the cable structure. Upon installation, unnoticed defect may later become very costly, due to the cost of maintenance operation and supply interruption.

3.1.1. Insulation resistance test

Insulation resistance describes insulations magnitude between two phases or phase and earth. Effective insulation is the base of cable operating. Insulation resistance measurement can be performed only off-line. By measuring insulation resistance it is possible to detect short circuits, accidental earths and incorrectly installed or leaking joints.

The values obtained in the above tests should be recorded so that they are available for comparison purposes in the future. In failure situation insulation resistance test can be used indentifying the faulty conductors and fault classification. Faults can be divided into two categories; high and low resistive faults.

3.1.2. Sheath integrity test

By measuring sheath integrity, sheaths fault can be detected and localized before the bigger drawback. The sheath integrity test is recommended because the cables manufacturing, transportation and laying may cause defects. Sheath integrity measuring can be performed only off-line and it is suitable for polyethylene insulation cables. [15] Cable sheath tests are required for new cable installations to ensure that cable is continuous from end to end, and the cable is laid as planned and route cable joints are sound. Existing cables undergo sheath testing prior to returning to service following cable diversion or repair to ensure circuit integrity. [16] Sheath integrity test can be performed using the same test equipment as with insulation resistance test. Then the following interpretation rules which are presented in figure 3.1 can be used.


Figure 3.1 Cable sheath integrity test process and interpretation by using Megger. [15] Cable sheath testing voltage is 5 kV DC, if the voltage does not rise and stay there one minute there is a fault in outer sheath. If the cable does not pass the test, the fault has to be located and prepared.

Effective tool for sheath fault pre-location is SebaKMT Company’s product MFM 5-1 and fault location ESG-80 with pinpoint sticks. These products also allow performing sheath integrity test, but the difference is that this method measures leakage current. For faultless cable which has outer sheath of polyethylene, the leakage current value is about 1 μA/km and polyvinyl chloride sheath 1,5 μA/km 5 kV DC measuring voltages [17]. The table 3.1 presents interpretation rules for sheath integrity test leakages currents in cases where outer sheath is polyethylene.

Table 3.1. Leakage current values and interpretation rules 5 kV DC measuring voltages.

Measured leakage current x Interpretation

x > 1mA/km Sheath fault

10 μA/km < x < 1mA/km Sheath fault is possible

x < 10 μA/km Sheath Ok

Performing sheath integrity measurement the following things have to be noticed. The ends of the cable have to be dry and remain dry during the whole measurement, it cannot be performed earlier than seven days after the installation so that the ground gets settled down and the outer sheath over the AXAL-TT cables screen is semi conducting which means that it should not be in contact with earthing or other parts which could breakthrough during the measurement. In addition, some joints which include aluminium foil for cross sectional water tightness have been noticed to cause measurement errors.

Sheath

integrity test using by

Megger

Measured insulation

resistance

Insulation resistance

> 500 MΩ

Insulation

resistance < 5 MΩ

Sheath OK

Sheath

Fault

Repair


3.2. Measuring tools to assess condition of MV cable and accessories

XLPE cables have a relatively long life-time and the investment costs are high. For that reason different measuring methods have been studied and developed to assess the condition of medium voltage cables and accessories. All the measurement methods have the same goal, to detect weak points and evaluate and schedule future investments. The target of investigation has been how to measure and locate partial discharges and water trees, which generates deterioration of cable insulation, especially XLPE cable network. Partial discharges in cable insulation, joint and termination have bad influence. When partial discharge occurs, it will corrode the insulating material in such a way that carbonized path will be formed and grow, then the electrical treeing will take place and lead to breakdown.[18] This chapter introduces alternative measuring methods, which could support older methods, what information measuring could give and the possibilities of their performance. Figure 3.2 presents the partial discharge process.

Figure 3.2 Partial discharge development processes. [19]

Microscopic spaces may be formed in insulation systems due to water tree growth, aging, installation or manufacturing defects.

Continued stress and overvoltage can initiate PD in voids.

Heat and other forms of energy released by PD cause erosion of the internal surface of the voids

Continued erosion forms channels that develop into so-called electrical trees in the insulation

Continued PD produces further erosion until the electrical tree bridges the insulation

Insulation system failure


The alternative methods are very welcome especially later when cable network ages. Then the typical degradation processes have contributed the power cable insulation and its accessories. It is typical for aged and some cases for new joints and terminations that there arises an interface problem which finally leads to partial discharges. In accessories insulation bad hardening and conductor problems with overheating causes cracking which also leads to partial discharges. In figure 3.3 below all the most typical factors which cause partial discharges in MV cable network are presented.

Figure 3.3 Typical factors which cause partial discharges in MV cable network. From figure 3.3 can be seen that to manage cable networks’ condition, the factors which are presented in the figure have to be able to be diagnosed and localized. New supporting measuring methods, which are presented in the next three subchapters, can be divided into two parts on the bases of what is the target of the measurement. Dielectric spectroscopy and dissipation factor measurement are concentrated XLPE cables insulation condition and the partial discharge testing methods focus is in accessories.

3.2.1. Dielectric spectroscopy measurements

The dielectric response can be used to determine the condition of the insulation. It is a useful measurement technique to evaluate insulation deterioration caused by water trees in XLPE cables. This measurement must be performed off-line and it is non-destructive testing method, which means that the electric stress during the test does not reach the level higher than the normal service condition. Consequently, the risk of failure and accelerate of aging remains at a normal level.

Dielectric response can be categorized into four different groups. The first of which are the quite new and water tree free cables, which belong to LLLP group. The low loss linear permittivity response is characterized by an almost frequency independent capacitance. The loss tangent is low and has also very weak frequency

PD

Conductor’s

problems

Interface

problems

Bad

hardening

Cracking

Overheating

Local field

concentration

Electrical

trees

Delamination

Insulation

voids

Water

trees

XLPE insulation

Accessories


dependence. Both capacitance and loss tangent are independent of applied voltage, i.e. the insulation material is linear [20].

The second dielectric response type is the voltage dependent permittivity (VDP). This response type is the general type for cables, which contains water trees but the trees have not yet formed the whole insulation. The response not dependent frequency in this type and both capacitance, and loss tangent increase with increasing voltage.

The third response type is the transition leakage current (TLC). The leakage current occurs in high voltages levels. Water trees have reached the whole insulation and breakdown strength has decreased significantly. The last response type is leakage current (LC) response. This type has almost the same properties as TLC but the leakage currents already occurs in low voltage levels.[20] Following table 3.2 shows the typical values for different dielectric response types.

Table 3.2. Diagnostic criteria for XLPE cables using dielectric spectroscopy.

Assessment Dielectric response type

Permittivity values

Break down values

Strategy

Good LLLP ε’’<8*10-4 > 4*U0 Retest in 5-10 years

Aged VDP ε’≥8*10-4 Δε’’≥1*10-4 Δε’≥2*10-4

2,5*U0<BD<4*U0 Retest in 2-5 years

Significantly aged

TLC or LC <2,5*U0 Replace as soon as possible

3.2.2. Dissipation factor measurements

Dissipation factor, also known as loss tangent measurement, is a typical diagnostic method for XLPE cable. The measurement is a diagnostic method for testing cables to determine the quality of the cable insulation. The measurement also tries to predict the cables remaining life-time and in order to prioritize cable replacement and/or injection. By measuring dissipation factor, the possible need for other measurements can also be determined.

This measurement can only be performed off-line. It is able to detect water treeing in XLPE cables and give possibilities to evaluate aging and deterioration in the cable insulation. Tan-delta measurement is suitable especially for extruded types of insulation and when the cable length is not long. [3] As mentioned above, the measurement is based on determination of cables insulation condition. The cable which is free from impurities and defects recall almost perfect capacitor. In the cable case, the phase conductors and cable screens aluminium wires can be regarded as capacitors


whose surfaces are separated by insulation material. In such a case, the perfect capacitors current is purely capacitive, which means that it is 90 degrees ahead of phase voltage. The impurities in cable insulation increase the currents resistive component when the current is not purely capacitive anymore. [21] For this reason a loss angle is formed in between of the purely capacitive current and the real current of the starting situations. This angle will also often be used for measurement of tan delta measurement name. Below in figure 3.4 are shown schematic models of the insulation, the currents and the angle.

Figure 3.4 Schematic models of the insulation, and currents phase diagram. [22]

So, the more impurities and defects the insulation contains, the more resistive it comes and the loss angle increases. The cable which is in a bad condition has a big loss angle.

Most commonly the measurement is performed at 50 Hz or 0.1 Hz. When 0.1 Hz testing method, also known as Very Low Frequency (VLF), is used much smaller size voltage sources than normal 50 Hz test can be used. Another benefit compared to normal 50 Hz test is, that low frequency is more efficient. Differences in the values are greater at lower frequency, which makes interpretation of results much easier. VLF dissipation factor is measured at nominal voltage U0 and also at 2U0 and the differential loss tangent is calculated as:

(1)

Table 3.3 below presents XLPE cable condition evaluation criteria using dissipation factor diagnostics.

Table 3.3. XLPE cable condition evaluation criteria using dissipation factor diagnostics. [23]

02tan U tan Assessment

< 1,2*10^-3 < 0,6*10^-3 Good

≥ 1,2*10^-3 ≥ 0,6*10^-3 Aged

≥ 2,2*10^-3 ≥ 1*10^-3 Highly degraded

00 tan2tantan UU


In turn, this measurement method only tells us which cable is in good or aged condition, and requires testing of all cables. However, it can help a utility prioritize cable replacement. The dissipation factor measurement does not tell the location of the weak spot. It only gives an estimate of insulation condition of the measured part of the cable.

3.2.3. Partial discharge test (PD)

Partial discharge tests are one of the most commonly used test methods for medium voltage cable and its equipment. Partial discharge testing can be performed either on-line or off-line. Measuring tools small size, weight and costs are its advantages. Partial discharge test equipment is commercially available for many companies which produce measuring devices and applications. There are also a few companies which offer on-line partial discharge metering systems e.g. KEMA, HVPD Ltd. and KJ Dynatech Inc.

Partial discharge test is useful for extruded insulation types. XLPE cable insulation is one of the most sensitive to the destructive effects of electrical partial discharge activity, it is imperative that the network operator strives to operate their polymeric cable network discharge-free [24]. Partial discharges test provides a quick and quite a cheap solution for checking the quality of cable installation, because it can detect insulation defects that may have occurred during the cable laying. At the same time it is possible to find defects that may have occurred during the cable manufacturing process, but this is very rare because as all XLPE cable suppliers now must meet the requirements of < 5 pC level as per the IEC 60270 standard factory test. This means that any defects in the cable insulation such as voids or delaminating of the screens are detected during the quality control process in the factory which uses off-line partial discharge testing in a screened room in the factory’s laboratory. Partial discharge test is also used to detect the cable networks deterioration due to normal service operating conditions.

Partial discharge test is a predictive, non-intrusive and non-destructive testing method. Predictive means that the test indicates insulation degradation before the failure. Non-intrusive means, that it is possible to perform the test without interruption of service. The concept non-destructive is due to the fact that the test does not have destructive features such as over voltages or high voltage stress. Partial discharge activity is known to result in deterioration and erosion of the primary insulation in cables and most particularly at cable accessories such as joints and terminations. If partial discharge is not noticed and the cause of the partial discharge activity not repaired, it will result in failures, supply interruption, equipment damage and/or injury to personnel. When partial discharge measurements are performed off-line, the cable has to be disconnected from the grid and this is a huge disadvantage because the supply interruption is expensive for the Network Company. Off-line measurements advantages are that all disturbances which comes corona effect and partial discharge sources


originating outside the cable are eliminated [20]. In other words, the measured data interpretation is easier than on-line situation and also smaller defects can be detected. Another advantage compared to On-line is that off-line measurement enables to determine partial discharge inception and extinction voltages. In turn, determining the extinction voltage, the test voltage can reach as high as two and a half times the normal operating voltages and it increases the risk of failure. Partial discharge measurement off-line method should be used only after the installation or fault, because otherwise it causes supply interruption.

The new on-line technique for insulation condition testing of in-service cables was made possible by the development of inductive PD sensors, clip-on High Frequency Current Transformers (HFCT). These are attached to the earth bar/strap of the cable termination under test, with no outage required [25]. On-line measurement advantages are that during the test, cable or accessories operate under normal condition and the risk of failure does not increase as a result of high test voltages which are used in off-line methods.

A partial discharge does not occur all the time and for that reason the on-line monitoring possibility is a valuable diagnostic tool. Weeks’ or months’ monitoring allows the detection of partial discharge trends, which simplify decision making and its reliability. The biggest drawback of the on-line measuring is complicated interpretation of the results [20], but longer monitoring with partial discharge trends facilitates it. Even though any conclusions can be drawn from the results, they must be able to be interpreted in some way. As mentioned above, XLPE insulated cable has to be almost partial discharge free. It makes the measured data interpretation and partial discharge sources localization easier, because the cable itself is generally very reliable if it is properly produced and tested. Consequently, the vast majority of faults occur in manually installed cable accessories such as joints and terminations along the cable’s length.

As mentioned above, the biggest challenge in partial discharge measurement is the interpretation of measured data. For this reason, all information which the measurement can give has to try to be collected. The results of the partial discharge measurement can be divided into two groups. First group contains directly measured basic data e.g. PD level, PD inception voltage and PD extinction voltage. The results which belong to the second group are derived from results of the first group. Derived values are e.g. PD magnitude, intensity and mapping. By collecting all this information, it is possible to create so called fingerprint for cable section. This fingerprint can act as a reference value for a possible future condition monitoring cases or after the fault situation.

There are generally presented in some values for partial discharge measurement which facilitates interpretation of measured data and the failed component determining. The following table 3.4 presents some interpretation rules for partial discharge diagnostics for the cable.


Table 3.4. Interpretation rules for PD diagnostics on cables. [20]

PDIV < U0 > U0

PDEV < U0 > U0

PD magnitude < Typical values > Typical values

PD intensity Low (e.g.<5 pulses/period) High (e.g.>5 pulses/period)

PD pattern Less Harmful Harmful PD

PD location Cables accessories Cable insulation

PD mapping Scattered PD along cable PD concentrated at specific location

In addition, there are some apparent charge values for cables insulation, joints and terminations, which supports interpretation rules and facilitates results’ analysis. Acceptable apparent charge values are presented in table 3.5 below.

Table 3.5. Typical values of PD apparent charge in cables and accessories. [14]

PD Location Type Apparent charge

Cable insulation XLPE < 20 pC

Joints

Oil insulation 10 000 pC

Oil/ resin insulation 5000 pC

Silicone/ ERP insulation 500- 1000 pC

Terminations

Oil termination 6000 pC

Dry termination 3500 pC

Shrink/ slide-on termination 250 pC

As can be seen in table 3.5, the acceptable apparent charge depends on the type of the cable section component and its insulation material. Using measured results and the interpretation rules, it is possible to evaluate and categorize measured objects into three condition classes. The classes are introduced below in table 3.6.

Table 3.6. Condition class of cable section and the interpretation rules.

Condition class Interpretation

First class Not ok, need to replace or repair.

Second class Not ok, need monitoring.

Third class Cable section is ok

A decision diagram, which is presented in figure 3.5, can be used for determining condition classes.


Pd locate in accessories?

Pd located in accessories

Critical PD pattern/ intensity?

Critical PD pattern/ intensity?

First class Second class

No

No

No

No

NoNo

No

No

No

Yes

Yes

Yes

Yes

Yes

Yes

Yes

YesYes

PDIV/PDEV < operating

voltage?

Concentrated PD Mapping

Concentrated PD Mapping

Pd level > typical values for both cable

insulation and accessories

Pd level < typical values for both cable

insulation and accessories

Third class

Figure 3.5 Decision diagram for evaluation of cable condition using PD diagnostics. [20, 4]

3.3. Strategies for field testing medium voltage cables

The cable networks faults and quality control creates a need for the distribution Networks Company to have some kind of strategy for field testing. Clear goals and pre-made testing methods improve the efficiency of the operation in various situations. This reduces the time which the operation takes, and this way improves the quality of the service offered to customers. Good testing strategy also reduces the cost of faults repairing. The strategy which can be built depends on a lot of measuring tools and networks’ constructions. However, the main point of the field testing strategies are; the new cable network quality control, improving delivery performance in the existing network and support decision making in a reinvestment situation. The following three subchapters present three different strategies which are based on strategies used in Vattenfall Service Sweden.


3.3.1. Field tests after the installation

After laying the new cable, it should be tested. In this way, quality checking of installations can be made. Different parts of the cable network should be tested, including cable section, joints and terminations. Quality checking and testing ensures that the largest defects can be detected and prepared after the installation. Quality testing after the installation can be applied case by case. It is possible to perform either a strict type quality test or an extensive type quality test. At the moment, only strict type quality test is used here in Vattenfall Verkko Ltd. This strict quality test process ensures only the cable condition at the testing moment, but does not give any information which could be valuable for cable network condition management in future. Strict quality test contains an insulation resistance test and sheath integrity test. As mentioned before, insulation resistance test measures the magnitude of the insulation between two phases or between a phase and earth, and sheath integrity test is a test, which tells only, if the cable sheath is entire or broken. Both of these tests are valuable, but it is important for the life-time of the cable network to be as long as possible and this creates the base for extensive quality checking after the installation. However, there are a lot of commercial cable diagnostic measuring applications available, which can be used to extend the quality checking process at the moment. This new and extensive quality checking could contain at least the following measurements; off-line and on-line partial discharge and dissipation factors. By using these measuring methods, it would be possible to detect the defects, which are caused by installation mistakes even more effectively and determine the baseline values for a cable section. These values may be needed in condition management in the future. In addition, measurements confirm manufacturers’ promises of the quality e.g. almost partial discharge free XLPE cables. By using dissipation factor measurement, it gives an overview of the insulations’ condition. Later, this value could be compared with the new value and changes could be detected. Clear changes of the dissipation factor values can mean that the insulation is in a bad condition and this information could give reason for reinvestment. Another method is partial discharge measurement which allows creating the so called finger print for a cable section. At first, before connecting the new cable section in service, off-line partial discharge measurement could be performed. By using this method it could be possible to find even the smallest defects from the joints and terminations because the off-line methods’ accuracy is better than the on-line methods’. As the joints and the terminations are the most critical points in a cable section, the PD test is very useful in a cable network. Another partial discharge measurement could be performed after the cable is under service condition. On-line measurement would create the base for futures on-line PD measurements, which can be used in cable network condition management in the future. By using on-line partial discharge diagnostic methods, it is possible to find the networks’ weakest point and achieve condition based maintenance strategy. In the


next figure 3.6 the testing process after the installation is presented. Figure 3.6 presents the extensive quality test and the strict quality test which is used at the moment in Finland.

Extensive quality testingStrict quality testing

Insulationresistance test

Sheath integrity

test

Dissipationfactor

(Tan Delta)

Off-linePD

On-linePD

Figure 3.6 Quality testing process after installation. One option is that the extensive quality test is not performed after every installation. It could be performed in a situation where a new contractor, cable, joint or termination has been used. In this way, it is possible to find manufacturing or installation errors immediately. [26] Extensive quality testing should also be recommended to be used in situations where the strict measurement has revealed some problems. As mentioned, field tests after installation gives a lot of valuable information about the cable network and its condition. For this reason, all the information which comes from measurements should be collected. If this is not done, it could almost be said that the diagnostic measurements in the future are useless, because there are no base values. Collected cable network data could be saved into network information system or cable diagnostic data base, where it would be available later if problems arise in the network section. Figure 3.7 presents the information which could be collected from measurements, if both strict and extensive quality tests were performed.

On-linePD

measurement

Cable sections PD magnitude andintensity including

joints and terminations

Partial dischargeinception and

extinction voltages

Ohm valuesSomethingabnormal

revealed during the test )2(tan

)(tan

tan

0

0

U

U

Insulation resistancetest

Sheath integrity test Dissipation factormeasurement(Tan Delta)

Off-linePD

measurement

Cable sections finger print

Cable diagnostic data base

Figure 3.7 Information of the measured cable section.


Collected network data can be valuable, and it provides important information to support decision making in reinvestment situations at the end of the cable’s life-time. On the other hand, comprehensive information of networks condition could be modelled in a network information system. It makes networks condition management easier, because a weak cable section can be seen visually. The possibilities of networks information system and cable diagnostic data base integration are returned to later in chapter five.

3.3.2. Field tests after fault

As mentioned before, fault locating and repairing in the cable network is a much slower process than in the overhead lines network. Indication of the fault becomes harder because the cable network’s physical location is under the ground, so the network’s visual inspection is impossible without digging. Hopefully, a big part of the cable network’s faults are caused by some external factors, e.g. careless ground digging so then the fault location is known. Everything does not go this way and the fault has to be localized by using other methods. In practice, cable network’s fault repairing process starts by the fault’s distance determining. After this, the fault has to be found by using some other method. One possible solution for fault distance determining is the pulse echo meter, which sends a pulse straight to place of the fault where it reflects back and this allows the determining of fault distance by using mathematical methods. As the distance is known, the specific place of the fault can be searched with a cable canon. When the specific fault place has been found the cable is dug up and the damaged section cut off. The next step in the repair process can be interpreted to depend on fault’s instigator and the cable section’s earlier operating condition. If the instigator which caused the fault is some single factor e.g. ground digging and this can be confirmed and cable section operating history is good, then there is no need for any measurements and the fault can be repaired. On the other hand, when the reason which has led to cable failing is not so clear and there is a doubt that there are other faults or faults which are developing, the following shows a way in which to proceed after the cable is cut-off. It is recommended to perform dissipation factor and partial discharge measurements on the cable, starting from the fault position to both directions. The dissipation factor gives an overview of the insulation’s condition and comparing it to earlier results which were measured after the cable’s installation is now possible. In turn, by using partial discharge measurement, especially now when it is possible to be performed off-line, it gives possibilities to achieve better accuracy than with on-line methods and detect even the smallest defects. If these measurements do not reveal anything new, the cable can be repaired. But if the measurement shows that a cable section is in bad condition, it is possible to try and find the weakest point of the cable section using by VLF withstand test, which is known as a


pass or fail test. When the cable does not pass the test, a new fault point was born, which has to be localized and the fault repairing process starts again. This is repeated as long as the cable can be accounted to be in a satisfactory state. On the other hand, even if some single factor would have resulted in the cable failure e.g. excavation and operating history would be good, it is possible to exploit this interruption and perform some diagnostic measurement. Especially in a situation, where an alternative connection could be used and thus the client does not expire any harm. In figure 3.8 field process after the fault is presented.

Dissipation factor and PD test

VLF withstand test

Clear fault mechanism?

Determine fault distance

Cable identifying and fault finding

Found new fault?

START

Yes

Yes

Ok

No

No

Strict quality test

Repair the first fault or faults

Figure 3.8 Field testing process after the fault.

3.3.3. Field tests and diagnostics before reinvestment

The timing of the cable reinvestment is a very difficult case, except in the cases where the cable is totally damaged and unusable. Most commonly a network owner has three different main parameters from which the most economical value is wanted to achieve. Firstly the network owner wants the existing network to achieve as long a life-time as possible. Prolonged life-cycle improves the profitability of the original investment, if this does not cause any additional costs. Secondly, the reinvestment cost is very high and if postponing the reinvestment is possible, it saves money. Thirdly is the cost of interruptions if the reinvestment is postponed. The purposes of the field tests are to optimize these three parameters. The field tests try to predict the cable sections remaining life-time or to prioritize the order of the replacement. For this purpose, appropriate measuring methods are dissipation factor and partial discharge measurements. As mentioned before the dissipation factor organizes


the cable into a right order, which is based on the insulation condition and partial discharge could give more specific information on an individual level. Figure 3.9 shows one possibility of how the diagnostic process could be performing before the reinvestment decision.

On-linePD scanning

On-linePD trending

Dissipation factor measurement

Not ok

Ok

Not ok

Not ok

Possibleinterruption

situationdo extensivequality test

Start

Repeat trendinge.g. after year

Ok

Not ok

Cable diagnostic data

Ok

Ok

OkReinvestment

Figure 3.9 Possible diagnostic process before reinvestment.

As can be seen, the diagnostic process starts with on-line partial discharge scanning. It is a fast method to find weak cable sections. A situation where PD scanning result is clear it may be considered to do the extensive quality test later during the possible interruption situation. If partial discharges which may be deemed to exceed the allowable limits are found, partial discharge trending is recommended. Trending can show the direction of partial discharge development. If there is no critical development, it can be concluded that the cable is in a good condition. In a situation where the development is fast, some off-line tests e.g. dissipation factor measurement are recommended to be done before the reinvestment decision. All data which the measurements offer, should be collected from every case irrespective of the result being acceptable or not, and especially in cases where the decision is to postpone the reinvestment.

31

4. LIFE CYCLE COSTING

4.1. Determination of LCC

In the publication [28] the concept of life cycle costing is defined as follows; “Life cycle costing is the process of collecting, interpreting and analysing data and applying quantitative tools and techniques to predict the future resources that will be required in any life cycle stage of a system of interest.” The life cycle costs which are the results of this analysis contains the costs of the acquisition and also other costs that are caused during the life time of the target product, service or process. The main purpose of the life cycle costing is to be a tool which supports managers’ decision making in analytical processes. However, it should be remembered that an investment decision cannot be solely based on the results of the life cycle costing. In order that LCC can be used, overall picture has to be created of the case which is under investigation. The whole life cycle costing process is presented in figure 4.1 below. Because the LCC is an analytical process it contains a lot of different types of data and assumptions in many cases. As there are different types of data usually used in management accounting, the input data has to be changed into the same format which allows calculation.

Life cycle costing result

Life cycle costingmodel

Input data changing

the same format

Existing data Risks

Assumptions

Figure 4.1 Life cycle costing process. [28]

After the input data formation and the risk levels determination LCC model can be created which gives LCC results. Final result contains cost estimation for object with assumptions and financial implications [28].

4.2. Economic evaluation methods for LCC

The most interesting and most important part of life cycle costing is the calculation phase. There exist many different methods for this purpose, which have their advantages

4. Life cycle costing 32

and disadvantages. All methods are not useful in all situations, because different calculation methods may give the result in a different unit. In the following calculation methods are presented which can be used when evaluating economic value for LCC.

4.2.1. Simple payback method

The simple payback method calculates the time which is required to return the initial investment. It allows comparing which investment has the shortest payback time and also gives rough estimation for investment profitability. The advantages of this method are that it is a quick and an easy calculation, and the results are easy to interpret. Disadvantages are that it does not take into account inflation, interest or cash flows which may come during the life time. If annual cash flows are equal, the payback period is found by dividing the initial investment by the annual savings. The equation of simple payback method is presented as following:

savings

investmentpayback C

CT (2)

Where paybackT is payback time in years, investmentC is initial investment cost and savingsC is

annual operating savings.

4.2.2. Discounted payback method

This method is almost the same as the simple payback method but it takes into account value of the time, which is an advantage of discounted payback method. This method should be used only as a help for the examination but the making of the decision should not be established through this method only. Discounted payback method does not take into account cash flows which come outside the payback period. The equation of discounted payback method can be written as following:

flow

uncoveredbeforepayback C

CTT (3)

Where paybackT is payback time in years, beforeT is year before full recovery, uncoveredC is

unrecovered cost at start of the year and flowC is cash flow during the year.

4.2.3. Net present value method

Net present value NPV is the result of the application of discount factors, based on a required rate of return to each year’s projected cash flow, both inflow and outflow, so

4. Life cycle costing 33

that the cash flows are discounted to present value. In general the positive net present value means a profitable investment. Consequently the best choice between two competing alternatives is the one with the biggest positive net value present. The advantage of this method is that it takes into account the time value of money and it uses all available data. Sometimes the amount of data can make it difficult to interpret. Its disadvantage also is that the method is not usable when the alternatives have different life lengths. The equation of net present value method can be written as following:

0

1

t

1C

r

CNPV

T

tt

(4)

Where t is year, 0C is initial investment cost, tC is the net cash flows in year t, r the

rate of interest and T is the last year of the period.

4.2.4. Internal rate of return method

The internal rate of return (IRR) calculates the discount rate value on what net present value is zero. In the [29] internal rate of return is defined as follows: “The internal rate of return is the actual rate of return expected from an investment”. This method is usable only if the investments will generate an income. The equation of internal rate of return method can be written as following:

01

01

t

Cr

CT

tt (5)

From equation (5), r is to be calculated by employing trial and error method, which equals as internal rate of return.

34

5. ACTIONS OF CABLE CONDITION MANAGEMENT IN REAL-TIME PROCESS

To achieve the best possible reliability a real-time condition management should be considered. As known, distribution network already includes a lot of automatics and information technology. Because the cabling rate has been very low so far, there has not been the exceptional need or stimulation to research condition management tools of cable real-time process. On the other hand, the use of cable networks have been considered now in urban areas, where there is almost always an alternative connection that can be used in a fault situation, but the situation changes dramatically when looking at this from the point of view of rural areas. Outages are expensive for distributions Network Company. Preparing for future faults could be done by using real-time process metering application. Faults detecting could be faster and more accurate than without controlling. Especially in area where there are no alternative connection possibilities, accurate fault localizing system is very useful. This chapter presents two different types of methods for two different purposes; what are commercially available now and what can be used as a part of the real-time process.

In principle, medium voltage cable online monitoring means that the cable operates normal voltage during the measurement. On-line monitoring is based on partial discharge measurement and it is only an application for condition assessment of the cable network, which exists and is commercially available at the moment. Its main purpose is to offer the possibility of the cable operations and condition monitoring without the electricity supply being interrupted or disrupted. To make this possible, the measuring equipment must be installed and monitoring carried out without interruption. In this way, the installation does not cause any harm and the measurement is not a disturbance to clients.

Disturbance free operation and additional information of the condition of the cable network improves the reliability and thereby reduces the cost due to interruptions. In the best case, real-time monitoring of the network is thus a tool for condition monitoring, which can be affected by the amount of unplanned interruptions and thus reduces down time of the network which always means the loss of income. By using real-time condition monitoring it is also possible to identify and prioritize the faulty cable sections which have to be repaired or replaced. In this way, it is possible to make target investments correctly and reduce costs of investments.

On-line monitoring can be divided in to two categories; part time and permanent monitoring. Generally, every online monitoring system that is used in medium voltages underground network is a part time application. Part time metering

5. Actions of cable condition management in real-time process 35

may take from ten seconds to three months and it depends a lot on the target of the metering.

5.1. Permanent continuous monitoring system

The permanent continuous monitoring system means a system which is used only for a certain object monitoring. The permanent monitoring system can be installed at the same time when the monitoring object is built or later. The monitoring system provides continuous real-time information about partial discharge levels of the monitoring object. By using continuous monitoring the changes in partial discharge levels of the cable can be seen better, because continuous monitoring allows the formation of partial discharge trends. Using the permanent continuous monitoring systems is only profitable in cases where monitoring object is very critical and the fault may cause unreasonable inconvenience to the customer, or the objects, where measurements performing is dangerous and the objects are at places where it is hard to get to perform the measurement. The permanent continuous monitoring system has disadvantages of its high price compared to the benefits which it offers. Nevertheless, if the solution is that the permanent continuous monitoring system will be installed, the most common installation place is primary substation. The place where the monitoring system is easy to install in the primary substation is the cable cellar. The construction of a new primary substation should take into account that the primary substation’s structure allows condition monitoring measurements performing as easily as possible in the future. In this case, all feeders can be measured from a single location which would reduce the cost of condition monitoring. Especially cases of permanent condition monitoring systems, the centralized monitoring reduces the number of monitoring equipment. At the moment, new light primary substations are built as compact as possible and so cost effective that there is no possibility to perform monitoring in the future, because there does not exist e.g. a spacious cable cellar. The only possibility to perform the monitoring in a light primary substation is to install permanent continuous monitoring system during the construction of the primary substation. Some kind of failure in monitoring equipment can be impossible to be repaired without supply interruption due to difficulty of access to the monitoring system. The permanent continuous monitoring system’s advantage compared to the portable partial discharge monitoring system is, that it allows the whole primary substation monitoring at the same time, because the permanent device has more input channels than a portable system. There are commercially available permanent devices, which have 16 input channels, and which can be expanded to 64 channels by using slave units [30]. The length of the monitored cable section can affect the usefulness of the permanent and centralized system, because the partial discharge detection accuracy gets


worse when the cable length increases. The cable insulation has also an effect on the distance which can be monitored. One partial discharge sensor manufacturer promises that sensors are capable of detecting partial discharge from as far as 8 kilometres in case of XLPE insulated cables and 4 kilometres in case of PILC insulated cables [31].

5.2. Portable partial discharge monitoring

Alternative technologies have been recently developed to perform partial discharge monitoring, which could replace the expensive permanently installed online systems. These kinds of monitoring units are known as portable partial discharge units. The development of these equipments has been motivated by high prices of the permanently installed systems. As the name of these equipments imply, the system is portable and it is easy to install in different places. As the target of the permanently installed system was very critical cables which had very high outage costs due to important customers or the high amount of customers, the area where portable PD monitoring can be used is entirely cabled distribution network. The basic idea is that the portable monitoring devices are rounded over the distribution network and short monitoring is performed in every point. The duration of the monitoring can be changed between one and two weeks. [32] However, this depends a lot on the capacity of monitoring devices and the amount of the targets. The acquisition cost of portable partial discharge monitoring devices is smaller compared to the acquisition cost of a permanently installed system. This is emphasized especially in the case, where permanently installed systems were installed on a large scale. Partial discharge monitoring implementation in entire distribution network is challenging. The one company which offers portable partial discharge monitoring equipment has developed a solution for this problem. They offer the measurement equipment which contains different kinds of portable partial discharge measurement devices for different purposes. The smallest device is known as “handheld device”. The purpose is to screen the entire network and find the cable section which needs extensive actions by using this device. According to the experiences of the producer of this device, there is about 20 % of screened population which needs extensive actions and 2 % needs monitoring. [24]


5.2.1. Online condition monitoring process

As mentioned earlier, the PD-screening of the network is started by using the handheld PD detector which is light by construction. It can be used for cables, switchgears and distribution transformers, PD-screening. The most versatile devices of this kind contain three different kinds of sensor technology which allows partial discharge detection. The suitable sensors which can be used for partial discharge testing are ultrasonic/acoustic sensors, high frequency current transformers and transient earth voltage sensors. The high frequency current transformer suits the best for partial discharge testing in the cable network. In the figure 5.1 below the handheld PD detector which is commercially available is presented.

Figure 5.1 PDSurveyor™ Handheld PD Detector made by HVPD Ltd. [30]

Pre-screening takes these kinds of detectors about 20 seconds per cable [30]. The detection is easy and quick to perform. These kinds of detectors could be used in some field measurement which was presented in chapter 3, especially in the case of quality checking after the installation, when the cable has been connected in service. Then the performed quality control can be extent if the PD-screening gives reason for it and the target are important. After the PD-screening, the cables which have partial discharge levels that are out of accepted limits go to the next level in the condition monitoring process. It is not necessary that the choice of these selected cables has to be based on only the screening levels. The decision that monitoring is needed can be also based on the operating history information of the cable e.g.; if there is something that was noticed earlier which differs from normal operating, or measurements for quality checking after the installation have been shown something, which refers some incipient fault but then the decision of the monitoring necessity has been negative. In the phase of monitoring the source, the aim is to identify location and development of partial discharge. The PD-screening devices are unsuitable for this purpose. Therefore a metering device must be used which allows


capturing the true PD signal wave shapes synchronously on up to four PD sensors and use this information to discriminate different PD sites and find the location of nearby PD sources [33]. When the source of partial discharges is identified from the target it can be taken under the continuous monitoring the duration of which can be days to weeks. [32] In this way it is possible to find and form the trends to partial discharge developing, which helps the decision making of the need of possible repairing operations. The decision can also be that the monitoring is continued longer. In figure 5.2 the whole online monitoring process is presented.

On-line PD pre-screening.

Duration about 30 second per

tested item.

On-line PDmonitoring.

Duration about From 1 day up to

2 weeks.

Continuous on-line PDmonitoring.

Duration > 2 weeks.

Cable diagnostic data base

Test date

The location of the discharge. Amplitude of

PD level and trend.

Diagnostic PD testing & site

location.Average duration about 45 minutes.

No

No

OkYes

Not ok

Need actions?

Ok

Yes

Start

Need actions?

Figure 5.2 On-line monitoring process. [32]

Considering the figure 5.2 above, the online condition monitoring process consists of three main phases which are; PD-screening, location and monitoring. The construction of condition monitoring process is not stationary, because the positions of the location and monitoring can be changed between each other. This change has been made in the following subchapter where the online condition monitoring process implementation is described in a larger distribution network. After these two phases changing, the determination as “online condition monitoring process” should be more meaningful. The process which is described in the figure 5.2 could be called partial discharge testing and monitoring process.


5.2.2. Monitoring process implementation in large distribution network

The basic idea of portable online monitoring devices is to offer affordable alternatives to permanently installed monitoring devices. Since acquisition the huge amount of monitoring devices is unprofitable, have strive to achieve as comprehensive result as possible by limited device capacity. To get the best result, it has to be taken into account that there are different factors which affect the partial discharge creation e.g. the season of the year and climate. Consider the larger scale online condition monitoring process implementation in primary substation level. The average amount of feeders which a primary substation contains is about eight in the case of Vattenfall Verkko Oy. It is also assumed that all feeders have been cabled. The monitoring in primary substation level can be now implemented e.g. by using HVPD Ltd.’s multi-monitoring device which contains 16 channels. In this way it is possible to monitor all feeders in a primary substation at the same time. This multi-monitoring device allows switchgears monitoring too. The manufacturer has announced that the monitoring distance of this device is 5 km or 6 ring main units in theory [32]. In reality, there are always branches on the feeders which have the attenuated influence on the partial discharge signals. So, if the rule of thumb is used, the monitoring distance will be about 2.5 km per feeder. If the suitability of the monitoring unit distribution network is considered, by using these theoretical monitoring distances the following can be stated: in the case of city and urban areas the distances are not usually limiting factors but the amount of ring main units can be, in the rural areas it is on the contrary. In the figure 5.3 is presented the partial discharge online monitoring in the primary substation level, by using a diagram of the primary substation.

Figure 5.3 Online partial discharge monitoring in the primary substation level.


If the feeder is longer than 5 km or contains more than 6 ring main units in this 5 km route, 4- channelled devices can be used which allow the monitoring distance to increase. This lighter four channel device is able to increase the monitoring distance about 2.5 kilometres or three ring main units. If the multi-monitor device and the four channels lighter device are used together, then the total length of monitored cable can be about 7.5 kilometres or 9 ring main units in theory. In table 5.1 are presented information of the feeders of the example network which is also presented in figure 5.4. Under consideration is the primary substation which feeds an urban area.

Table 5.1 Feeders’ information.

Colour of feeder

Length before 6th RMU (m)

Total length (m)

Number of RMUs

Blue 1523 3075 9

Light blue 3049 6888 11

Purple 2256 3125 7

Brown 4256 5253 7

Red 979 979 1

Darker red 2295 8856 9

Green 2420 2420 6

As from table 5.1 can be seen, any feeder which has a length before the sixth ring main unit would be over 5 kilometres long. The monitoring distances of the device which is located in the primary substation, are described in figure 5.4 and are highlighted with a white dotted line.


Substation 20/110 kV

Figure 5.4 MV cable network in urban area and metering distance of monitoring device, which is located in the primary substation. The circled areas in figure 5.4 represent the areas where the monitoring device which is located in the primary substation is not reached and needs additional devices. In this type of network the amount of monitoring devices that are needed can be determined according to the feeder which has the largest number of ring main units. The determining factor can also be the distance depending on the case. In the case of the described network, the determining feeder is highlighted with light blue in figure 5.4, which contains 11 ring main units. So that the entire distribution network can be monitored at the same time in the primary substation area, two extensive monitoring units are needed, due to the size of the area which is under the biggest circle. This is the only circle where two units are needed. These extensive two units will be rounded over the circled areas. If it is assumed that the monitoring period is for example 2 weeks per circled area, the one monitoring round takes about six weeks. Then the entire network has been monitored. This kind of monitoring can be performed as many times as wanted, the reliability of the results of monitoring increase when the number of monitoring rounds increase and seasonal affection of partial discharges can be eliminated in this way. This kind of monitoring process can be performed e.g. twice in a year, in the summer and the winter. Vattenfall Verkko Oy has 135 primary substation and if it is assumed that the entire medium voltage distribution network would be cabled for at least frame cables. In the example, the monitoring implementation took about two months when monitoring process was performed twice. Then the entire distribution network of Vattenfall Verkko Oy monitoring takes about 23 years if there was only one system. By using four systems, it allows the monitoring performing every sixth year and devices


installations can be done together with primary substation inspections. However, situation is very different at the moment as described above. There exist only few primary substations where almost all feeders are cabled.

5.2.3. Sensors’ positions in distribution network

There are two types of sensors for partial discharge measurements, capacitive and inductive sensors. The firstly mentioned can be used for voltage metering and the secondly mentioned for current metering. Usually the capacitive sensors have been used in the off-line measurements because they have caused reliability problems in the online measurements. For this reason, the sensors which are used in the online measurements are inductive high frequency current transformers. These kinds of sensors also have the characteristic that its core can be split, which allows the installation without interruption of the supply. In figure 5.5 below two sensors for partial discharge metering are presented which are commercially available.

Figure 5.5 Left side KEMA and right side HVPD Ltd. sensors which are suitable for partial discharge measurements. [34, 30]

In figure 5.6 different kinds of locations are presented where the partial discharge measurements can be done in the case of ring main unit. The normal and the simplest installation is numbered in the figure by number one. The other possible installation is numbered by number two. This installation is perceived as the best of these four. In this, high voltage current sensor will be installed around the cable neck and earth strap will go through the sensor twice which decreases the noise due to groundings of the cable. In other words, current in the earth strap cancel out each other and then the current of the conductor remains. On the other hand, this installation requires the length for the earth strap but usually polymer cables have this characteristic. [35]


Figure 5.6 Different locations where partial discharge measurements can be done.

When the measurement performing is done in the places numbered 1, 2 or 4, then the propagating of the partial discharge signals are measured in the shield to phase. Number three represents the situation where partial discharge propagating is measured in the phase to phase channel. The problem of this kind of installation is that the sensors’ installation is dangerous due to voltage. Figure 5.7 represents the sensors installation in practical terms in the primary substation.

Figure 5.7 The partial discharge sensors which have been installed into the switchgear of primary substation.

The sensors have been installed in primary substation inside the switchgear. The smaller commercially available sensors have been installed on the earth straps of the cable terminals and the biggest sensor [50], which has been developed in Technical university of Tampere, on the cable neck.

Transformer

1.

2.

3.

4.

44

6. STRATEGY OF THE MV CABLE LIFE-CYCLE MAINTENANCE

Nowadays electricity distribution business, and success in it, requires good distribution network condition management during the whole life-cycle. The knowledge of features and condition of network are the base for condition-based asset management which is one of the essential tools to enable the reliable and cost-effective life extension of existing cables to be achieved.

According to [36], “Of the studies carried out on large, public electricity utility networks, the largest savings are to be made in deferring the capital replacement programmes of the cable population. This is simply because the capital cost of replacement is so large compared to the other costs, that by saving even a small amount of this cost tends to ‘swamp’ the other costs.”

6.1. Definition of strategy

The word strategy has a very broad meaning. The strategy in corporate level is defined in [37] as follows: “Strategy is the direction and scope of an organisation over the long-term: which achieves advantage for the organisation through its configuration of resources within a challenging environment, to meet the needs of markets and to fulfil stakeholder expectations”. The corporation’s main strategy can be divided into sub strategies which try to lead them to the direction of corporate main strategy. The strategy of the MV cable life-cycle maintenance can be seen as an operational level strategy, which is one part of the whole network maintenance strategy. On the other hand, maintenance strategy is a part of a maintenance program, which is required for the distribution network operator. This is provided for Trade and Industry Ministry's (Finnish Kauppa- ja teollisuusministeiö) decision 517/1996. [38] For this reason, the maintenance strategy planning must also take into account the different instruments such as laws, regulations and recommendations that come from external stakeholders.

6.2. Maintenance

Maintenance is a way to act aimed at avoiding failures of equipments by taking care of them. The role of maintenance is important especially in a business, where results’ formation depends a lot on the usability of the assets and it has a key role in operations. [39] The electricity network business is a good example of this kind of business, where almost all operating is based on the ability of distribution network energy delivery.

6. Strategy of the MV cable life-cycle maintenance 45

Therefore maintenance is planned with operations, which purposes are to ensure the safeties, operational reliability, failure’s prevention and fast repairing of the occurred failures on optimum costs [38]. When the issues which are mentioned above are in order, it is possible to achieve a longer lifetime of the assets, lower operating costs and increase the reliability. Maintenance operations may be give important information of the condition of the assets and in this way the timing of the replacement investment can be improved. This is a valuable feature of the maintenance due to capital savings and it allows as long a service life for network assets as possible. The maintenance can be divided into preventive and run to break down maintenance. In the case of preventive maintenance, the creation of failures is tried to be prevented and the incipient fault detected before they occur in repairing operations. Further, the preventive maintenance can be divided into time based maintenance and condition based maintenance. [38] The distribution network consists of different kinds of components e.g. cables, overhead lines, distribution transformers and disconnectors etc. and all of these components have their own life time which has begun at some moment. This can make the maintenance very complicated. Some component suit time based maintenance and some of them suits condition based maintenance. But in the case of time based maintenance there always arises a scheduling problem due to different installation moments. The scheduling is not the only problem, there is also the problem of how valuable the maintain target is. It does not make sense that the same maintenance operations are applied to every target. For example the cable which feeds a big customer e.g. some industrial plant, it makes sense to apply much more intensive maintenance compared to some cable which feeds only few small customers in a rural area. In the following six subchapters different kinds of maintenance methods are shortly described. Some of these methods have been also included in the figure 6.1. The figure 6.1 represents the progressing of the failure of function of the time.


Time

(TPM)

Inspection interval Repairing time

Failure (RTB)

Time of reaction

The starting failure

can be recognized

(CBM)Allowable level of risk triggers (RCM)

Ope

ratio

nal a

bili

ty

Figure 6.1 Progressing of the failure of function of the time. [38]

6.2.1. Run to breakdown maintenance (RTB)

As the name of maintenance method implies, the maintenance targets are given run to breakdown before reaction. The repairing operations are performed only after the failure has occurred. This kind of maintenance method can be allowed only for targets which are less important, or the targets which cause so small outage costs that the preventive maintenance operations does not make sense in the economical aspect.

6.2.2. Preventive maintenance (PV)

In preventive maintenance the aim is to act before the failure occurs or total breakdown of the target has happened. In other words, the aim is to influence the probability that failure occurs or change the faulty component before it causes larger damage. As mentioned before the preventive maintenance can be divided into time based and condition based maintenance, which are presented in the following.

6.2.3. Time based maintenance (TBM)

Time based maintenance is the first part of preventive maintenance. As the name time based maintenance implies, the maintenance actions performing are time based in this case. The maintenance actions which are performed time based can be inspections as well as short or extensive maintenance operations. A big part of distribution networks’ components are filled within the scope of time based maintenance. This is due to requirements of the controlling authority as mentioned before.


6.2.4. Condition-based maintenance (CBM)

The condition-based maintenance is the second part of preventive maintenance. The basic idea in this method is that the maintenance operation performing is based on the current condition of the component. However, the actions have to be carried out in such a way that they have been done before a fault occurs. Usually this kind of maintenance method requires continuous monitoring of the target or good knowledge of the effect of aging on the target. In other words, the estimation for remaining life time of the target has to be very exact, so that the maintenance actions can be started at the right moment.

6.2.5. Reliability-centred maintenance (RCM)

Reliability-centred maintenance is also one preventive maintenance method. It is based on the risk level, which is allowed for the object of the maintenance. [38] Reliability-centred is used to develop scheduled maintenance plans that will provide an acceptable level of operability, with an acceptable level of risk, in an efficient and cost-effective manner.

6.2.6. Total productive maintenance (TPM)

The aim of this maintenance method is to improve the entire system more efficient, by integrating the maintenance thinking in each organizational level. By eliminating interference factors, it is possible to achieve better overall efficiency. [40] In the network business this kind of method can be concentrated, reducing the amount of interruptions, operational reliability improvement and in this way the entire network service quality improvement.

6.3. Optimal maintenance

Often it is thought that maintenance is only the cost, which is true in its simplest terms.

If optimal maintenance is considered in a more operational level, then the costs of the maintenance operations and planned outages should be always smaller than the cost of the unplanned outages, in order for it to be economically justified. However, in the network business the maintenance operation cannot be established only through the costing. The safety aspect has to be always taken into account and it is a determinant factor. Optimal maintenance strives to achieve the optimum process reliability, which is modelled in the following figure 6.2 below.


Tot

al c

osts

Amount of maintenanceOptimal OversizedUndersized

Cost of maintenance

Cost of unplanned and planned interruption

Safety impact. Optimal level in distribution network business

Figure 6.2 Optimal process reliability and behaviour of cost. [41]

In practical terms, the maintenance in the network company is some kind of a combination of the maintenance methods presented earlier, which should be optimized as cost efficient as possible with safety aspects. The authority provides a maintenance program for the network company which contains e.g. some time based maintenance actions. Alone this may not be the optimum maintenance strategy or a sufficient one. It can be seen in figure 6.2, it is possible that the total costs rise very sharply in the case where maintenance operations are undersized. The risk also exists in the case where the maintenance actions are oversized, and it can lead to situation of the excessive maintenance costs although the reliability of the system improves.

49

7. KEY FIGURES AND RESULTS FORMATION IN THE NETWORK BUSINESS

7.1. The reliability key figures in delivery

Reliability means that the system is able to perform the tasks which it is designed for under stated conditions for a certain period of time. For electricity network the term reliability of delivery means the system’s ability to meet its supply function from the customer point of view. According to IEEE 1366 standard reliability of delivery can be described with key figures introduced below [42]: • SAIDI (System Average Interruption Duration Index), which is the sum of durations of all interruptions of supply to individual customers during a time interval, divided by the total number of customers and the duration of that time interval.

sN

i jijt

SAIDI

(6)

• SAIFI (System Average Interruption Frequency Index), is the number of interruptions of supply to individual customers during a time interval, divided by the total number of customers and the duration of that time interval.

sN

jjn

SAIFI

(7)

• CAIDI (Customer Average Interruption Duration Index), sum of durations of all interruptions to individual customers during a time interval, divided by the number of these interruptions.

SAIFI

SAIDIt

CAIDI

jjn

i jij

(8)

7. Key figures and results formation in the network business 50

Where

NS the total number of customers nj is the number of the customers who experience the interruptions i tij is the duration of the ith interruption of supply to customer j

These equations can be used when the earlier behaviour of the network is under examination or when considering how a new investment will be affected by the network’s behaviour.

7.2. The interruption costs formation

In the electricity network business outage of the electricity distribution always causes harm to customers. This harm of the outage has been tried to be modelled by using DCO value and in this way changed to disadvantage the value of the monetary. Usually, the DCO value is calculated for one year period and it contains both unplanned and planned outages. It also takes reconnections into account. The equation for disadvantage caused by electricity delivery outages is presented in appendix one. In the table 7.1 the values for different types outages and reconnections are presented, which can be used when the cost of interruptions is calculated.

Table 7.1 Prices in the 2005 monetary value used when calculating the disadvantage caused by outages in electricity supply. Adapted from [43]

Unexpected outage

Planned outage Outage caused by

high-speed autorecloser

Outage caused by time-delayed autorecloser

€/kW €/kWh €/kW €/kWh €/kW €/kW

1.1 11.0 0.5 6.8 0.55 1.1

In addition that the outages cause disadvantage to customers the outages can cause disadvantage to the network company too. The control model of energy market authority has been created thus that it takes into account the DCOs in the calculation of the allowed return. Then the consideration of the actual interruption costs are compared to the reference level of interruption costs. The interruption costs can either increase or decrease the operating profit or operating loss. In this way, it is possible to examine the benefit of the cable diagnostic by the costs of using these outages. The equation for reference value of disadvantage caused by electricity delivery outages is presented in appendix two.


7.3. Cost of cable diagnostic

The costs of cable diagnostic can be divided into three main cost components which are the cost of diagnostic equipments, the cost of operations and the cost of outages in some cases due to off-line measurement methods. In the following the above-mentioned cost factors have been considered from the distribution network company’s point of view. In the case of Vattenfall Verkko Oy this means the point of view of the client organization, where all contractor services are outsourced and they acquired an external service provider.

7.3.1. Cost of investment

Cable diagnostic equipment acquisition causes the investment cost. Because the contractor services have been outsourced in Vattenfall Verkko Oy , there are no straigth cable diagnostic equipment investment costs which can be attributed directly to the network company but those costs will be reflected indirectly in the prices of services. On the other hand, there can be cases where customer is very important. Then a continuous online monitoring system is needed that can ensure the continuous supply of electricity. This kind of permanent monitoring system causes direct investment costs for the network company. In the case of permanent monitoring system, there are also other costs which can be considered to belong to the cost of investment e.g. cable data entry and remote synchronization and yearly monitoring subscription of the sensors. These costs, which are mentioned above are due to monitoring service. At the moment the monitoring model is such that a network company or someone else who owns the monitored cables buys the monitoring equipment, but the monitoring and data interpretation are bought as a service [44]. So, the investment costs of the permanent online monitoring system consist of monitoring equipment acquisition costs and monitoring service costs. With such an investment which contains acquisition cost and other yearly costs, it is sensible to look at the entire life cycle cost of the system.

7.3.2. Cost of operation

The operational costs of the cable diagnostic are caused by measurement performing and switching operations. It is quite difficult to estimate the operational cost of cable diagnostic precisely, because only off-line measurement services are available at the moment in Finland. There is no online diagnostic service available. Despite of missing online diagnostic service it is possible to make some estimation of its operational costs. The costs of online partial discharge measurements can be divided into three categories based on the time of measurement which it takes. The first category contains partial discharge screening measurement which is quite an easy and a fast operation to perform. It is also the cheapest measurement of these three


for its operational costs. The second category contains portable part time monitoring application, the operational costs of which are caused by installation and un-installation. The last category contains permanently installed online partial discharge monitoring system, which has the same operational costs as the second category. These installation and un-installation costs are realized only once during its life cycle.

7.3.3. Cost of interruption

The planned outage causes the costs for the network company. Sometimes measurement requires interruption so that the performing is possible. The reason for the interruption can be due to a characteristic of the measurement or electrical safety. The average costs of the planned outage can be determined by using an equation which is presented in appendix 1, prices in table 7.1 and yearly amount of the energy transferred to the users from the network operator’s network. The lastly mentioned is available in the network information system. However, there is almost always an alternative connection in the city and urban areas. By using these alternative connections and right switching arrangement, the costs of interruptions can sometimes be avoided.

7.4. Allowed return on network business

The electricity network business is a monopoly business and due to this reason it is under supervision of authority. The allowed return is especially under supervision. The supervision is also designed to encourage the network owners to develop operational efficiency and to keep the pricing of electricity transmission reasonable at the same time. The creation of the allowed return affects different kinds of factors; some of these factors are already presented in chapter 7.2. Let’s consider the creation of the allowed return from the point of view of cable diagnostics benefit calculation which is presented in the next chapter, and from which it is possible to determine the profitability of the entire network’s operational level. Generally allowed return is justified by using a supervising model which takes into account the costs presented earlier in chapter 7.3. In the figure 7.1 is presented a supervising model which is in use at the moment and includes the supervising period 2008-2011.


Cost of interruptions

Operational expenditures

Length of network

Amount of customers

Value of energy

Expansioninvestment

Replacement investment

Replacement price

Straight-line depreciation

Current replacement cost

Efficiency measurement(DEA/SFA)

Straight-line depreciation

Controllableoperational

expenditures

Referencelevel of

outage costs

Actual outage costs

Quality bonus

Quality sanction

Efficiency-improvement objective

Allowedreturn

Figure 7.1 Supervising model 2008-2011. [49]

For every network company efficiency improvement requirements are formed in the supervising model. It requires improving the efficiency of the operations in the network company. If the network company succeeds improving its operations efficiency then it is entitled to a bigger allowed return. If the requirements to improve efficiency are not achieved, then the allowed return is lower. Thus from the basis of the quality of operations either a quality bonus or a quality sanction can follow. Considering the cable diagnostic, the sections of the supervising model which the cable diagnostic have an impact on can be easily identified. The measurements performing and the installation of the monitoring systems require operations which cause operational expenditures. Some of these operations, especially off-line diagnostic need planned outages which have an influence on the costs of interruptions. The normal failures also have an influence on the actual outage costs. The profitability of the cable diagnostic can be also considered from the point of view of the principle of calculating actual adjusted return. The Energy market authority calculates the actual adjusted return for the network companies every year. In the table 7.2 below the calculation of the actual adjusted return after the computational entity taxes have been described [45].


Table 7.2 The principle of calculating actual adjusted return. Adapted from [45]

Operating profit (or loss) + 0.5* Actual outage costs + Actual controllable operational costs + Paid network rents + Depreciation according to plan from goodwill + Depreciation according to plan from the electricity network + All net change in the accrued entered refundable connection charges (additions- refunds)

- 0.5* reference level of outage costs - Controllable operational costs in accordance with the efficiency target - Annual straight-line depreciation calculated from the network replacement value - The Cost of the financial assets needed to ensure the operation of the network

= Adjusted operating profit/loss

= Imputed profit +/- Other adjustment items

= Profit before taxes - Imputed corporation tax for the enterprise

= Actual profit (adjusted)

As can be seen from the principle of calculating actual adjusted return, the half of the actual and the reference level of the outages costs have been taken into account. In addition, there are also included the actual controllable operational costs and as a benchmark, controllable operational costs in accordance with the efficiency target. Thus from the operational aspect, it is profitable to try to reduce the actual adjusted income, when the risk for exceeding the allowable return decreases.

There is also published submission of new supervising model, which contains guidelines for the next supervising period 2012-2015. [51] The biggest change in the new principle of calculating actual adjusted return will be the annual straight-line depreciation and how take into account of these. The above-mentioned and other changes have no influence on the profitability analyse of the cable diagnostic.

55

8. BENEFIT CALCULATIONS

The aim of this chapter is to consider benefit calculations from the point of view of the cable diagnostic and go through three different cases where three different cable diagnostic methods are under consideration. By using the results of the benefit calculations the determination of the strategy for medium voltage cable network condition management can begin. The results of how different cable diagnostic methods and their utilizations in different situations can be affected on the network company’s operations in economical aspect will be shown. The considerations are divided into three different individual parts; the consideration of the permanent online diagnostic system, off-line diagnostic and portable part time online diagnostic process. In practical terms the strategy for medium voltage cable condition management will consist of some kind of combination of these three diagnostic parts. However, by using benefits calculation the correct types of objects can be correctly directed in the most appropriate measurement procedures and auxiliary costs of the operation can be avoided.

8.1. Cost and profitability of the portable part time online diagnostic

The profitability of the online cable diagnostic can be examined by comparing the operational costs which it causes and the benefit which it offers. The online cable diagnostic has the advantage that part of the cable faults can be avoided. These avoided faults and their saved costs of outages forms the economical component which can be used for comparison purposes as they can be compared to operational costs. The costs of operations and the costs of interruptions are not directly comparable with each other. For this reason the comparison has to be made by using the model of actual adjusted return in the network company, which was presented in chapter 7. The examination has been implemented in the primary substation level. The portable part time online diagnostic process has been performed which consists of two types of measurements; partial discharge screening measurements and monitoring measurements which have a longer duration. The review will be examined in three different areas in which regional division is based on the reliability criteria. In these areas, the profitability of the portable part time online diagnostic process has been examined in the case of the current network and the case of the network of the future, where the entire network has been cabled. In principle, the portable part time online diagnostic process is carried out in such a way that the PD-screening measurements are implemented throughout the entire cable

8. Benefit calculations 56

network in the primary substation level in order for the most critical cable sections to be identified. These critical cable sections are supposed to be in calculation about 10 % of the screened network. It is also assumed that these 10 % needs monitoring after the PD-screening. The number of PD-screening or monitoring measurements that are needed for the implementation of the portable part time online diagnostic process is determined by using a diagnostic device, following the manufacturer’s rule of thumb; one measurement can be measured in 2.5 km-long cable section so that the accuracy of the measurement remains [33]. So, for example, a 10-km-long cable requires 4 pieces PD-screening measurements. The cost of online measurements which are due to operations cannot be determined precisely because such a service is not available in Finland at the moment. The following types of assumptions of operational costs have been made. The cost of PD-screening measurement is about 25€ per piece when it is performed by secondary substation inspection. The cost of normal secondary substation inspection is about 50€ per piece at the moment. This has led to the assumption that one worker’s hourly rate is 50€ in the case of inspection, and its performing takes one hour. If the rule of thumb is now used, what is obtained is that the cost of PD-screening measurement is 10€ per cable kilometer. In the case of monitoring the costs are caused by the exporting of the monitoring equipment to the target, installation and un-installation. With these kinds of operations the cost is assumed to be about 100€ per monitoring target. In the calculations it is also assumed that the portable part time online diagnostic process is carried out two times during 40 years. In addition, it has been considered how the profitability behaves if the cost of operations increase due to added processes rounds. It is assumed that by using the portable part time online diagnostic process it is possible to catch 50 % of the faults before the failure is realized, excluding faults which are caused externally which factors about 30 % of all faults. Further, it is assumed that half of the observed 50% can be repaired without interruption and the other half causes work interruptions. The fault frequency in the examination is assumed to be 0.75 failure per 100 km,a during the first 20 years. Subsequently, the fault frequency is assumed to grow 0.02 per year. According to this e.g. the failure frequency for the cable is 0.77 failure per 100 km,a when it is 21 years old. In this way, for the network the number of faults in a deferred 40-years life cycle which is under consideration has been reduced. For each of these cases which have been examined, the average costs of unplanned outages and planned outages have been determined by using reliability network analysis calculation. These average costs are determined for one hour of interruption. It is also assumed that operational repairing costs are to be about 1000€ per one failure of the medium voltage network. This cost does not include the costs of materials. This has led


to the assumption that the working group that includes two men spent about 5 hours in troubleshooting and repairing and their hourly rate is € 100 per hour per man. The cost of medium voltage failure is estimated to be about a third of the previous for the case of threatened failure, because the contraction of the stock and labour as well as other preparatory of the work can be handled in advance and controlled.

8.1.1. Suitability for Rural areas

Firstly the portable part time online diagnostic process and its profitability are considered in a rural area. The input data and some technical information of the network which is under consideration are presented in the table 8.1. The examination is performed by using total cable length which exists at the moment and the case of future where the whole network which the primary substation feeds is cabled. The case of the future situation is also simulated in such a way that the portable part time online diagnostic process will be performed only with cables which are the most critical in terms of fault interruption costs of the entire network. In these cases of the future, the cables which belongs population of the 20 % of the most critical cables or the 10 % of the most critical cables have been taken into account. The reliability network analysis calculation which is included in the network information system combined with MS-Excel allows this kind of prioritizations implementation in a quite an easy way.

Table 8.1 The input data for the calculations of different network cases.

RURAL

Current Future 100% Future 20% Future 10%

Total cable length (km): 6 123 24 12

Number of faults during 40 years: 2 42 9 5

Pd-screening 3 50 10 5

Pd-monitoring 1 5 1 1

As can be seen in table 8.1 the total length of the cable network is quite short in the situation of current network and the number of faults is almost nonexistent. In the network of the future the situation is different, the total length of the network is long and the probability of failure that occurs in the network area has increased. In the case of the future, the number of faults means one failure per year in theory. The average values of the disadvantage which is caused by unplanned and planned outage for one hour interruption are presented in the table 8.2. The values are calculated for each case.


Table 8.2 The average costs of outages in different network cases.

RURAL


Average costs of unplanned outage 1 901 € 2 282 € 3 468 € 3 859 €

Average costs of planned outage 101 € 202 € 450 € 536 €

The costs of unplanned outages are very small which is normal in rural areas, because the number of customers is small. In the next table 8.3 the results of the benefit calculation are presented.

Table 8.3 The cable diagnostic affection on the DCO, operational costs and adjusted return.

RURAL

Current 100% Future 100% Future 20% Future 10%

DCO before diagnostic 3802 € 95842 € 31216 € 19293 €

DCO after diagnostic 1901 € 77586 € 19141 € 12112 €

Difference 1901 € 18256 € 12075 € 7181 €

Difference with building cost index 2167 € 20812 € 13766 € 8186 €

Operational costs before diagnostic 2000 € 42000 € 9000 € 5000 €

Operational costs after diagnostic 1683 € 35495 € 7032 € 3783 €

Difference 317 € 6505 € 1968 € 1217 €

Influence on adjusted return -1401 € -16911 € -8851 € -5310 €

The portable part time online process has positive influence on the disadvantage caused by outages where the amount and costs decrease. The positive influence can be seen in the costs of operations which decrease too. This is due to the fact that repairing operations can be done in a controlled way. In every case, the results are positive in the point of view of the network company, but there are also the costs which have not been taken into account in the calculation. For example, the cost which arises from the organization of the operation and these costs consume the generated savings. For this reason, can be said, that the portable part time online process is not profitable in rural areas, for at least from an economic perspective.

8.1.2. Suitability for Urban areas

In principle, profitability analysis in urban areas is quite similar to the rural areas. Now the degree of cabling is considerably higher in the case of current network. The input data and some technical information of the network which is under consideration are presented in table 8.4. The examination has been carried out in the same way as in the previous case of rural area, i.e. situation of current network and the case of future where


the whole network which the primary substation feeds is cabled. Both cases of future are included where the cables critical aspect has been taken into account and the process has been performed only over the most critical cables.

Table 8.4 The input data for the calculations of different network cases.

URBAN


Total cable length (km): 45 180 26 10




From the table 8.4 can be seen that the amount of the cable is 25 % of the current network which the primary substation feeds at the moment. In the different cases both the average costs of unplanned and planned outages of the network under considerations are presented in table 8.5. In the case of current network, the average costs of unplanned outages in urban area are almost 15 times bigger compared to the case of current network in the rural area.

Table 8.5 The average costs of outages in different network cases.

URBAN



Average costs of planned outage 852 € 1 021 € 1 553 € 1 268 €

Table 8.6 shows that the portable part time online diagnostic process has a profitable impact on the results of disadvantage caused by outages and operational costs. The most profitable results calculation will provide the network situation of future, where the entire network is under the part time online diagnostic process. The realized benefit of this option indicates the level which may be covered by the indirect costs of the operations.


URBAN


DCO before diagnostic 447490 € 1397869 € 450423 € 268191 €

DCO after diagnostic 282236 € 913083 € 100094 € 75099 €

Difference 165254 € 484786 € 350329 € 193092 €

Difference with building cost index 188390 € 552656 € 399375 € 220125 €

Operational costs before diagnostic 16000 € 62000 € 9000 € 4000 €

Operational costs after diagnostic 14098 € 49177 € 7482 € 3116 €

Difference 1902 € 12823 € 1518 € 884 €

Influence on adjusted return -96097 € -289151 € -201206 € -110946 €


In table 8.7 the results of the sensitivity analysis are presented. In this examination the amount of the process rounds have been increased in a way that the process is performed every four years. The probability that the incipient fault is detected has been also increased from 50 % to 80 %, excluding the failures caused by external factors. Under the examination is the network of the future situation where the entire network is under the part time online diagnostic process. From the results in the table 8.7 can be seen that the increased operational costs, which are due to the increased portable part time online diagnostic process rounds, has very small influence on the profitability of cable diagnostic.

Table 8.7 The results of the sensitivity analysis and the cable diagnostic affection on the DCO, operational costs and adjusted return.

URBAN

Future 100%

DCO before diagnostic 1397869 €

DCO after diagnostic 626108 €

Difference 771761 €

Difference with building cost index 879808 €

Operational costs before diagnostic 62000 €

Operational costs after diagnostic 72905 €

Difference -10905 €

Influence on adjusted return -428999 €

The dominant factors which are affecting the profitability of the portable part time online diagnostic process are the total length of the cable, the failure frequency and the volume of outages costs.

8.1.3. Suitability for City areas

Finally under review are two primary substations which are intermediated in city areas from the point of view of regional division based on the reliability of the supply program criteria. In the tables 8.8 and 8.9 are presented the input information of the networks which are used in benefit calculation. As can be seen from the tables, the total length of the cable is almost the same as in the case of the future network. The biggest difference compared to considered network of the rural area is the total length of the cable and this has an impact on the theoretical amount of failures.


Table 8.8 The input data for the calculations of different network case, substation 1.

CITY_Luolaja


Total cable length (km): 37 40 7 2,5




Table 8.9 The input data for the calculations of different network cases, substation 2.

CITY_Suosaari


Total cable length (km): 56 57 8 -

Number of faults during 40 years: 19 20 3 -

Pd-screening 23 23 4 -

Pd-monitoring 3 3 1 -

The tables 8.10 and 8.11 shows that the average cost of unplanned outages in the case of future network, where the portable part time online process has been carried out the entire network, is about 1.5 times higher compared to the same network case in urban area.

Table 8.10 The average costs of outages in different network cases, substation 1.

CITY_Luolaja



Average costs of planned outage 804 € 828 € 1 134 € 976 €

Table 8.11 The average costs of outages in different network cases, substation 2.

CITY_Suosaari


Average costs of unplanned outage 33 797 € 33 697 € 61 915 € -

Average costs of planned outage 947 € 943 € 2 096 € -

In the tables 8.12 and 8.13 the results of the benefit calculation are presented. The results of both primary substations areas are profitable in the point of view of network business if the indirect costs of operations are not taken into account.



CITY_Luolaja


DCO before diagnostic 461720 € 473181 € 168236 €

DCO after diagnostic 285743 € 304188 € 112157 €

Difference 175977 € 168993 € 56079 € 0

Difference with building cost index 200614 € 192652 € 63930 € 0

Operational costs before diagnostic 13000 € 14000 € 3000 €

Operational costs after diagnostic 11215 € 12265 € 2683 €

Difference 1785 € 1735 € 317 € 0

Influence on adjusted return -102092 € -98061 € -32282 € x


CITY_Suosaari


DCO before diagnostic 642146 € 673949 € 185744 €

DCO after diagnostic 408408 € 440897 € 123829 €

Difference 233738 € 233052 € 61915 € 0

Difference with building cost index 266461 € 265679 € 70583 € 0

Operational costs before diagnostic 19000 € 20000 € 3000 €

Operational costs after diagnostic 16081 € 17081 € 2733 €

Difference 2919 € 2919 € 267 € 0

Influence on adjusted return -136150 € -135759 € -35559 € x

However, the profitability of the portable part time online diagnostic process is less than half compared to the case of urban network. This is due to the difference between the total lengths of the network that seems to be the dominant factor in the profitability of the review. When reviewing the results what must be observed is that the number of faults in the network is calculated at the beginning when the network is completely new and defects arise as it is a new cable. In practical terms, the considered cases city and urban are almost at the end of their theoretic life time. Then the actual failure frequency may be substantially higher than it has been assumed. Thus the amount of failures increase which increases the profitability of the portable part time online diagnostic process.


8.2. Cost and profitability of the permanent online diagnostic

8.2.1. Life cycle cost of the permanent online diagnostic

The life cycle cost of permanently installed online monitoring system can be described as the following equation:

RVMSYNCTINVLCC CCCCCC (9)

Where

LCCC is the sum of the costs of permanent online diagnostic system

INVC is the costs of control units and sensors

TC is the costs of training and installation

SYNCC is the costs of synchronization

MC is the cost of monitoring

RVC is the residual value of the system.

Case 1 In this short study there is one system and one cable that are monitored. The system consists of two control units and two sensors which are permanently installed on one cable. In the table 8.14 below, the costs which the system causes during the first year are presented.

Table 8.14 The first year costs of permanent online diagnostic system.

Cost Number (pc) Total cost Control unit/year 909 € 2 1818 €Sensor/year 522 € 2 1044 €Monitoring Subscription per sensor/year 600 € 2 1200 €Cable Data Entry and Remote synchronization 320 € 1 320 €Total cost exl. training installation travel etc. 4382 €

The investment costs of the control units and sensors are divided evenly for the first four years and after that investment are done for these parts. In other words four years straight-line depreciation has been used. There are also costs which occur only once at the beginning of a life time e.g. the costs of system synchronization and installation. The costs of training and installation are not taken into account in this calculation. The discount rate is assumed to be 7 % and the observation period 40 years. The results are presented in the table 8.15.


Table 8.15 The life cycle cost of permanent online diagnostic system.

Year 1 Year 2 - 4 Year 5 - 40 Cost of control unit and sensor: 2 861 € 2 861 € 0 €Cable Data Entry and Remote synchronization: 320 € 0 € 0 €Monitoring Subscription of the sensors: 1 200 € 1 200 € 1 200 €Total costs: 4380.5 4060.5 1 200 €Present value of the life cycle cost: 25 986 €

Case 2 In this short study there is one system and four cables that are monitored. In other words when one system is rotated over the four cables, then the monitoring period for each cable is 3 months per year. As the table 8.16 shows in this case there are some additional costs due to the rotating of the monitoring system. In addition, the cost of cable data entry and synchronization increases because it has to perform every cable section which amount is now larger than in case 1.

Table 8.16 The first year costs of permanent online diagnostic system.

Cost Number (pc) Total cost Control unit/year 909 € 2 1818 €Sensor/year 522 € 2 1044 €Monitoring Subscription per sensor/year 600 € 2 1200 €Cable Data Entry and Remote synchronization 320 € 4 1280 €Additional costs due to control units rotating/year 50 € 4 200 €Total Price exl. training installation travel etc 5542 €

The investment of the control units and sensors is done in the same way as in case 1, by using four years straight-line depreciation. Other assumptions are also the same as in case 1 and the observation period is 40 years. The results and present value of the life cycle costs for permanent monitoring system are presented in the table 8.17 below.

Table 8.17 The life cycle cost of permanent online diagnostic system.

Year 1 Year 2 - 4 Year 5 - 40 Cost of control unit and sensor: 2 861 € 2 861 € 0 €Cable Data Entry and Remote synchronization: 1 280 € 0 € 0 €Monitoring Subscription of the sensors: 1 200 € 1 200 € 1 200 €Additional costs due to control units rotating: 200 € 200 € 200 €Total costs: 5 541 € 4 261 € 1 400 €Present value of the life cycle cost: 29 550 €

The calculation shows that the biggest cost component is yearly monitoring subscription of the sensors.


8.2.2. The profitability of permanent online diagnostic

The previous chapter 8.2.1 determined the life cycle costs for a permanently installed online diagnostic system in two different cases. The profitability analyses of the investment can be used with e.g. net present value method which has been presented in chapter 4. In order for the investment to be a profitable one, the net present value needs to be at least zero or positive. The objective is to determine the limit value for the cost of unplanned outages and part of this which affects the actual adjusted return, the way that the above-mentioned condition is fulfilled. In this review, the costs of interruptions can be perceived as positive revenue flows. It is assumed in the calculation that the average time which the boundary of the cable fault takes is 45 minutes. The fault frequency of the cable is the same as that presented in chapter three, 0.75 failure per 100 km,a. In theory this means, that with a cable three kilometers long one failure will happen and with a six kilometer long cable two failures will happen during its deferred life time. In the following tables 8.18 and 8.19 the results of the calculation are presented. These values given in the tables can be used as the limit values, when the monitoring objects are selected. In theory a three kilometre long cable is required so that a computational failure occurs during the life cycle, and in this case the occurred interruption should cause disadvantage, the cost of which is 74 147€. In the longer cable the probability and the amount of failures increases, which reduces the required value of the occurred disadvantage. The limit value of cost of interruption is calculated as following:

lim3

4*2 NPVDCOcase (10)

Where

caseDCO is the limit value for the cost of unplanned outage in case

which makes investment profitable

limNPV is the limit value that net present value of the investment is

zero factor 4/3 represents the factor which fixes 45 minutes disconnection

time to one hour outage factor 2 have to be used because only half of outage costs can be taken

into account when adjusted return is calculated.


Table 8.18 The limit values for profitability in the case where only one cable is monitored.

CASE 1 Minimum length of the cable

Limit value for the target which gives positive NPV

Limit value for the cost of outage in case 1

One fault during 40 years period: 3 km 27 805 € 74 147 € Two fault during 40 years period: 6 km 13 903 € 37 073 €

Case 1 can be seen as a case where only one cable is monitored and the required cost of interruption needs to be very great so that the investment would be a profitable.

Table 8.19 The limit values for profitability in the case where four cables are monitored.

CASE 2

Minimum length of the monitored

cable section

Limit value for the target which gives positive NPV

Limit value for the cost of outage in case 2

One fault during 40 years period: 3 km 31 619 € 21 076 € Two fault during 40 years period: 6 km 15 809 € 10 538 €

In the example case two, the current value of the cost of interruption needs to be greater than in case one so that the net present value of investment would be zero or positive. However, in this case the monitoring equipment is rotated over the four cables which increase the total length of the monitored cable. Therefore it is possible to consider the cases where four three kilometer length cables or four six kilometer length cables are monitored. In this way the total length of monitored cables are 12 km or 24 km. If the theoretical amount of failures is calculated during the 40 years period in these cable lengths, the results are 4 failures for the shorter section and 8 failures for the longer section. Because the amount of occurring failures increases, the limit value of the cost of interruption which an individual failure causes can decrease.

8.3. Cost and profitability of the off-line diagnostic

Consider the off-line cable diagnostic as well as its costs and the profitability through the example cases. In the following three chapters the costs of off-line diagnostic have been considered which are cost of measurement operations and cost of planned outages which the measurement performing requires. These costs can be compared to the costs


which arise only when a fault has occurred and does not have a reacted in advance. The figure 8.1 below represents the configuration of the costs and how they are compared.

Cost of planned outage

Cost of off-line diagnostic

Cost of planned operationCost of unplanned outage

Cost of unplanned operation

Figure 8.1 The costs configuration of the off-line diagnostic in the profitability analyses.

The left side in the figure represents the total costs of off-line diagnostic and right side is its contrary, which also represents the accepted situation in many network companies at the moment.

8.3.1. The cost of off-line measurements

As was mentioned earlier, the off-line measurements require that the measured cable has to be disconnected from the network during the measurement; this causes interruption of electricity distribution which influences the customers in the disconnected area. However, in this review the focus has been on the cables which can be disconnected without causing harm to the customers. It requires that replacement of energy feeding is possible to implement in some other way. Let’s consider the costs which the measurement process causes. The costs are the results of measurement which is purchased from a service provider, and switching arrangements that have to be designed and partly carried out. In the table 8.20 different kinds of factors are produced, which cause the costs in off-line measurements.


Table 8.20 The costs of the off-line measurement.

Direct unit costs of the measurement process Measurement operational costs/cable in Sweden: 500 € Measurement operational costs/cable in Finland: 275 € Travelling expenses: - Costs of switching operation and planning: 80 € Total costs of measurement process in Finland: 1 455 € Total costs of measurement process in Sweden: 2 500 €

In Sweden the cost of off-line measurement is about 250€ per hour, when PD- and Tan delta measurements are performed. The duration of the one cable measurement is about two hours, when the cost of one cable measurement is 500€. During one day, it is possible to measure about 4 cables if preparatory work is well done [14]. In the table above it is assumed that five cables are measured during the day, when the total cost of day is 2500€. In Finland measurement service is purchased from an outside service provider. Then the costs of measurement consist of measurement device charging and labour costs. In this case the operational cost of measurement will be 275€ per cable, thereby the total operational costs of the one day measurements are 1375€. [46] When the costs of switching operation and planning have been taken into account the one day cost reached the value of 1455€.

8.3.2. The profitability of off-line measurements

The profitability of off-line measurement is based on the replacement investment postponing as well as the advantage which it gets and controlled implementation. Consider the performance of off-line measurement over the one feeder in the urban area, which is presented in the following figure 8.2.


Figure 8.2 The feeder in urban area and cable sections where are possible to perform off-line measurements. The five most critical cable sections where the planned outage can be implemented without harm to customers are numbered in figure 8.2 above. This determination is possible to carry out by using reliability analysis in the network information system. The cable sections which will be measured have been numbered and highlighted blue in the figure above. These cables are oil paper insulated and over 30 years old, which means that they are almost at the end of their computational life time. In the table 8.21 the technical information of cable sections are presented and the costs of unplanned and planned outages determined.

Table 8.21 The technical information and the costs of outages of the cable sections.

Example targets

Cable type Startingbranch

Ending branch

Cost of unplanned outage (€)

Cost of unplanned

outage/h (€) Length (m)

1. APY185 1 2 0 56120 933

2.

APY185 90 91 0 37944 81

APY185 91 92 0 45463 303

3.

HPL120 108 109 0 38428 176

APY185 109 110 0 38428 179

4.

AHXW185 139 140 0 38428 108

PYL120 140 141 0 38428 292

5.

APY185 144 145 0 38428 186

APY185 145 146 0 38428 268


Further, the average length of a cable section and the cost for it can be determined, if it has to be replaced by a new cable. This can be done by using year 2011 unit prices for components and work of electricity network in Finland, which were presented in appendix 3. In addition, what can be calculated is the average cost for unplanned outages and 45 minute periods which the troubleshooting and disconnection is assumed to take. When this has been done it is possible to calculate that part of the previous cost which has the affection on the result formation of the network company. In the table 8.22 the results of calculations in the case of example network are presented.

Table 8.22 The average costs of outages and replacement.

Average costs of unplanned outage which affect adjusted return 20561 € Average costs of planned outage which affect adjusted return 0 € Average replacement costs of example target 17 777 €

If the investment is made immediately, its cost is as the table shows about 17777€. By postponing the investment later it is possible to achieve savings, because the present value of the investment is decreasing during the time. The figure 8.3 below presents how the investment postponing brings savings in this case. The figure also shows what kind of influence the off-line measurements have on the savings which are due to postponing.

-20 000 €

-15 000 €

-10 000 €

-5 000 €

0 €

5 000 €

10 000 €

15 000 €

20 000 €

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29

Savings without diagnostic

Savings with diagnostic

Outage's affection on savings

Years

Figure 8.3 The behaviour of the savings and the cable diagnostic affection on it.

In this case, the off-line measurement is assumed to be carried out every four years, which is shown on the figure above. The off-line measurement consumes the savings every time when it is performed. It is also assumed that four year measurement cycle


gives a sufficient amount of information about the condition of the cable that the unplanned interruptions can be avoided. By this type of analysis it is also possible to determine the optimal moment for replacement investment. In the figure shows that the optimal moment for replacement investment is just before the fifth measurement. There is also plotted the currently average outage costs level and its affection on savings. It is assumed that it valuation remains constant over the time. From the figure 8.3 can be determined the point, where the savings which are due to investment postponing achieved the cost which one unplanned outages caused. In other words, the cable should withstand about 28 years more without failure so that interruption costs should be possible to compensate for the savings due to investment postponing. In this examination, the cost of interruption had been assumed to remain constant, but its appreciation may be greater in the future. Then the point where the savings compensate the outage costs moves further away.

8.4. Conclusion of the benefit calculations

In the following table 8.23 contains a conclusion of the results of the benefit calculations which have been classified in the different areas of the reliability criteria. The table shows that any diagnostic method exploitation is unprofitable in the rural areas. However, there might be some cases where an important customer is located in a rural area, and then the profitability can be considered case by case.

Table 8.23 The conclusion of the benefit calculations.

At the moment Rural Urban City Portable part time online diagnostic - ? x

Permanent online diagnostic ? x x Off-line diagnostic - x x

Future Portable part time online diagnostic - x x

Permanent online diagnostic ? x x Off-line diagnostic - x x

x = Profitable, - = Unprofitable, ? = Can be considered case by case In the Urban and City areas calculations gave positive results and shows that the cable diagnostics may have positive influence on the result of the network company. As always, the calculations which are directed far in the future contain a lot of assumptions. It is important to take into account the reliability and the affections of these assumptions when the decisions are made. The versatile tool which could help the interpretation of the results is known as “nine boxes”. Its purpose is to help indentifying


different kinds of factors which can change the profitability. The nine boxes where the cable diagnostic is integrated are presented in the figure 8.4.

The change in the number ofincipient faults

The developmentof cable diagnostic

The change of profit

The change of profitability

The change of diagnostic capacity

The change of costs

The change of operative expenditures

The change of cost ratio

The changeof outages

costs

Figure 8.4 The change factors of the profitability in the case of cable diagnostic [47].

Lets’ consider the figure 8.4. From there can be seen that the factors which have the biggest influence on the cable diagnostic profitability are located on the upper right and left corners. It should be also noticed that both of these factors contains uncertainty. However, the calculation can be regarded as relatively reliable, because probability of the incipient fault detecting is assumed quite low in the benefit calculation. On the other hand, one important factor which may be affected on the profitability of the cable diagnostic is the behaviour of the cable failure frequency in the future. The ageing of cable network in urban and city areas, and increased length of constructed cable network together with new contractors and installation methods may change the failure frequency. The determining factor in the benefit calculations consisted of outages costs, but these costs valuation of the future are partly unknown. The estimation is that outage costs valuation increases which makes the cable diagnostic more profitable than earlier. The second valuable result, which the benefit calculation gave, was that the operational expenditures have quite small influence on the profitability, especially in the case of the portable part time online diagnostic.

73

9. OPERATIONS MANAGEMENT

When it is wanted to move from run to breakdown maintenance in the direction of the condition based maintenance, the condition of the cables has to be able to be evaluated. In order for the maintenance actions to be directed correctly, the cables need to be able to be set in the right order based on their condition and interruption risks. In the figure 9.1 one model of the evaluation process is presented, which could provide tools for directing of cables maintenance actions.

Condition assessmentalgorithm

Prioritization Actions

RNA simulation

Cost of interruption

Key figures of

IEEE 1366 standard

• Technical information

• Installation environment

• Contractor

• After laying measurement condition

• Measured data after laying

• Monitoring data

• Fault history

• Monitoring

• Repairing

• Reinvestment

• New data to a database

• Some calculation which based on ms-excel

• Giving out condition classes which based on e.g. failure probability within the next year and remaining life-time estimation

• Take account of economical aspect

Input data Technical allocation ResultsEconomical allocation

Cable database

Figure 9.1 The maintenance actions allocation.

In this model, maintenance operations have tried to be presented as condition based as possible, which aims to take into account different factors which may have an effect on the need of maintenance actions. Commonly the model suits condition management of the cables, but it does not take into account the cases where failure has already occurred. It follows, that the model is a preventive oriented maintenance tool. The main purpose of the system is to behave as a tool, which could be carried out on the whole cable networks monitoring at fixed intervals, performing calculations and simulations. However, this requires an updated database, condition assessment algorithm and reliability based network analysis integration.

9. Operations management 74

9.1.1. Cable database

The first step towards comprehensive condition based condition management in the underground cable network is the establishment of the cable database. There are a lot of factors which can affect the condition of a cable and its remaining life time. The development of cable database has to identify the most critical factors which have the biggest influence on the condition assessment. These kinds of factors are presented in figure 9.1 above. These factors can be perceived as the main categories which contain the lower state factors. These represent more individualized information and are presented in the following table 9.1.

Table 9.1 Example of cable database and its contents.

Cable database

Technical information Installation environment Fault history

Cable type

Length

Installation year

Number of joints

Joint type

Location

Installation method

Soil

Contractor

The time of the year when installed

Line interruption data (fault type, location and duration)

After laying measurement condition

Measured data Monitoring data

The time of the year

Temperature

Insulations resistance

Sheath integrity

Dissipation factor

PD

Data from terminals inspections

On-line PD

The construction of this kind of cable database requires time. On the other hand, a big part of this data is available after the installation in the case of new cables. Then only the fault history, inspection data from terminals and data from partial discharge monitoring are missing. The point of view of the cable data base construction is very essential so that the information e.g. inspections and measurements, on the field comes to the network information system. If it fails, the picture of the cable network condition has distorted.

9.1.2. Cable condition assessment

The cable condition assessment represents the second part of allocation of the maintenance actions. It could be based on some kind of MS-excel calculations. The input information flows from the cable database to assessment algorithm, which gives out condition classes of the cables. These condition classes can be based on e.g. the failure probability within the next few years or the estimation of remaining life-time.


9.1.3. Prioritization with the economical aspect

The technical allocation and classification of the cables is not a sufficient way to direct maintenance operation in a sensible way. The economical aspect is also needed which supports the planning of the maintenance operations. If the economical prioritization is not carried out, it could lead e.g. to the following situations:

The contractor resources are directed to less valuable targets, which can cause undesirable costs.

Funding problem of the repairing/replacement operations, yearly budgets exceeding, valuable targets replacement have to be postponed in the future.

The contractors availability and competence to perform work In order for effective and economically sensible condition management and further maintenance operations to be achieved, prioritization has to be performed from the economical aspect. The prioritization can be made by considering the factors which have influence on the network operations in the economical aspect. Some factors are listed below which could allow the implementation of prioritization.

The strategic location of the cable in the distribution network: this cable forms the frame or branch. Is the cable frame line between two primary substation and forms reserve connection in a fault situation?

Actual disadvantage (outage cost) caused by electricity supply outages to the customers (DCO): If the cable section fails, which is the cost of unplanned outages? In other words, how large is the area where the fault in this cable causes harm. What kind of customers are there? What is the possibility to limit the fault and its duration?

The costs of forthcoming operations: The cost of organization and planned outages? Can this be done without interruption of the supply?

The costs of unplanned and planned outages are the most suitable for economical prioritization purposes of these which have been presented above, because it’s also partly taking into account two other factors. For these outages costs calculation and prioritization of the cable sections the reliability network analysis tool can be used in the network information system, which allows identifying the most critical cable sections in an economic perspective. After the technical and economical allocation, it is possible to organize the critical cable sections in the risk matrix, which is presented in the figure 9.2. The closer


the upper right corner the cable sections are located in the matrix, the greater is the need of their maintenance.

HighLow

High Criticality of the cable

Good condition of the cable

High interruption costs

Good condition of the cable

Low interruption costs

Bad condition of the cable

High interruption costs

Bad condition of the cable

Low interruption costs

Economical allocation

Tec

hnic

alal

loca

tion

Figure 9.2 Risk matrix for cables. [48]

If the cable section is located in the left bottom corner of the risk matrix then the need of maintenance does not exist due to good condition of the cable and low interruption costs. When the age of the cable network increases, so does the probability of the failure. In other words, the points which represent the cable in the risk matrix move upwards all the time. The cable diagnostic and maintenance operations can drop these points lower in the risk matrix. The cross-sectional movement of the points cannot be influenced by cable diagnostic or maintenance operations. It depends on the physical location of the cable. This can be only influenced by the constructional changes of the cable network.

9.2. Planning of the maintenance operations in the cable network

As mentioned in chapter 3, there exist different types of field measurements which can be divided into three main categories. The first category contained measurements after the installation and their purpose was the quality checking of the installation. The second category contained the measurements which could be performed after the fault. The last category contained the field measurements before replacement investment. In these three categories a fourth category can be added as measurements for condition monitoring. The partial discharge measurements which have been presented in chapter 5 can be included in this category.


There does not exist a one and only guideline for the condition management strategy which can be applied for every cable. For this reason, it is impossible to create one operation model for the maintenance of the cables. Instead of one strategy, different kinds of strategies have to be created, which are based on for example the area division of the reliability criteria. In this way, it is possible to develop main condition management strategies for every area where the characteristics of the network are different. For the development of these main condition management strategies the results of benefit calculations from chapter 8 can be exploited.

9.2.1. Strategy for rural areas

As the benefit calculations showed, the condition management operations are unprofitable in the rural areas. The benefits from the economical and the network business point of view are so small that there is no sense to invest in them. However, in the case of new cables, the condition management should be based on measurements performed after laying, which have been performed carefully. The other condition management measurements which can be performed during the life cycle of the cable should be left out in the rural areas. This also applies to existing cables. Due to the lack of profitability of the cable diagnostic in rural areas, the methods which can affect the cost of unplanned outages in failure situation have to be discovered in some other way. Besides careful construction and after laying measurements the one important factor is network planning. In figure 9.3 the example structure of the network is presented where the planning phase of the network should be aimed in rural areas.

Primary

substation

Primary

substation

Figure 9.3 Example structure of the network.

By the structure of the network the duration of fault detecting, limiting and separation can be influenced. In this way it is also possible to affect the life cycle costs of the


cable. In principle can be said that the structure of the network should be meshed, at least in the case of frame lines and long branches. For the branches, which the meshed structure is not sensible from the economic point of view should be considered constructional solutions which allows branches operation when the frame line is damaged. The branches operation without frame line and alternative connection requires its disconnection possibility to the frame line and an easy way to connect reserve power.

9.2.2. Strategy for urban and city areas

The strategy of the cable network condition management in city and urban areas differs from the strategy of rural areas. The reason for this comes from the results of the benefit calculations, which shows that the cable diagnostic may have positive influence on the network business in these areas. On the other hand, the network structure is different in these areas which allows more versatile condition management without any harm to customers, due to supply interruptions which some condition management measurement requires. If the cable condition management is considered during its deferred life cycle, and the maintenance operations are thought to be only time-based, then the measurement schedule can be as presented in the figure 9.4.

Bath curve

Time in years

Com

mis

sio

ning

test

s

PD

-Scr

een

ing

Off

-lin

e m

easu

rem

ent

s or

P

orta

ble

onlin

e m

oni

torin

g

10 20 30

End

of g

uara

nte

epe

riod

Figure 9.4 The maintenance operations of the cable during its life cycle.

The aim of the commissioning tests is to avoid failures at the beginning of the life cycle of the cable which is due to bad installation and material defects. The aim of the other tests is to manage the condition of the cable, which means the avoidance of the outages costs and the time of replacement investment prediction.


The measurements which are performed near at the end of the cable life cycle depend on the characteristics of the cable. Let’s considering two feeders, which are presented in the figure 9.5. Feeders are divided into different sections, which are based on the suitability of measurement performing.

Figure 9.5 Measurement strategies for condition management in the old cable network.

The off-line measurements are suitable, if the cable disconnection is possible without any harm to customers, these cable sections are highlighted on green in figure 9.5. If disconnection is not possible without outages costs, the online monitoring should be considered. Before the operations, the outages costs which the cable causes in the failure situation should be taken into account and compared to the limit costs of the profitability which were determined in the benefit calculations, these cable sections are highlighted in yellow. If the outages cost of the cable does not exceed the limit costs of online monitoring and the off-line measurements are also impossible, then one can consider PD-screening or make a decision to do nothing, these kinds of sections are highlighted in red in the figure 9.5.

80

10. CONCLUSIONS

10.1. Conclusions

At the moment, the cable network part of the total length of the medium voltage network is about 7 % in the case of Vattenfall Verkko Ltd., but the change through the higher cabling rate is quite fast. This is due to continued replacement investment, which the age structure of the network and the mechanical condition causes. In addition, the need of replacement investment causes the objectives, which are set in reliability criteria. All these replacement investments will be implemented by cabling which is the main construction method in Vattenfall Verkko Ltd. Almost 500 kilometres yearly cabling length will change the character of the network quite clearly in the near future. In this case, the asset management of the network and the operations which it requires have to change to be more suitable. A single approach of the condition management strategy for the cable network is not possible to establish, because the condition of the cable network may be affected by many different factors, and the operations must be economically justified. Mainly can be said that the geographical location and the role of the cable section determines the profitability of the operations. The complexity of the condition management strategy is also caused by the fact, that different types of cables require different kinds of measurements. In addition, by different measurements different phenomena can be measured in the cable network, e.g. quality of the installation or general condition. In this thesis, the measurement methods were studied around the world and more intensive examination were performed for the most suitable methods. These examinations produced useful methods for the purpose of the profitability analysis. By these methods is possible to identify the change factors which have an impact on the profitability.

Research also shows that it is possible to measure many different variables in the cable network, but the biggest drawback is the results’ interpretation. In order to evolve the interpretation of the results, one solution could be to aim to use the same type of cables and accessories of the cable network and in this way accelerate the learning process. However, this kind of operation model creates the risks of type faults which have to be taken into account in the selection of the components. By a successful choice the cable network condition management can be simplified and the maintenance costs reduced in the future. The life time length of the cable network can be influenced a lot by the installation moment, as outlined in chapter one, where the reasons for failures were presented. In the case of today’s cables, the challenges are mainly in the joints, terminations and the defects which are caused by external factors. Cable installation of high quality requires good installation work and methods which it can ensure. The

81

solutions for ensuring problems may be a partial discharge measurement introduction in the context of commissioning testing, when the biggest defects would be possible to eliminate especially from joints and terminations. On the other hand the commissioning tests create the base for the measurements of the future, which can be performed from the condition management point of view. Due to this perspective and the physical location of the cable network, it is very important that in the construction phase after the installation all essential information is collected from the constructed network, in order for condition management to be implemented in the future. In the overhead line network information collection is possible afterwards, but impossible in the case of cable network. The information should be saved into the network information system, where its exploitation for condition management purposes would be as easy as possible. The life time of the cable network progresses, and measurement of condition management purposes can be performed. These measurements can be regarded as forming one part of the network’s condition management strategy. In the case of rural areas, there are no economic justifications for these kinds of measurements. Therefore it can be said that it is almost impossible to affect the life-cycle costs of the cable, by the cable diagnostic. Then the only way to affect these costs is the structure of the network which means in simplest terms that the structure of the network should be meshed, at least in the case of frame lines and long branches. In this way, the duration of fault detecting, limiting and separation can be influenced.

In the urban and city areas the situation is different, when the condition management strategy can be implemented the way as described in chapter 9. The measurements which are performed in the condition management purposes based on the limit values of the profitability and the structure of the network. As the final result of the study can be concluded, that by successfully implemented cable network condition management and cable diagnostic exploitation relating to it, the results of the network business can be influence in a positive way.

10.2. Further study

The study showed that the condition management of the medium voltage network and its related operations are relatively unknown in Finland among the network companies, contractors and measurement service providers. The Technical University of Tampere has made some research in this area earlier. At the moment, the methods which are used to check the quality of the installation can be noticed as the condition of the outer sheath, the adequacy of the insulation and correct installation. However, these methods do not give indication of smaller installation defects which can be located in the joints and terminations. The reason for these kinds of defects can be the consequence of badly implemented electricity field direction, when the stress increases in the joint or termination and finally causes breakdown. In order for these kinds of defects to be avoided, it is

82

recommended that the commissioning test is expanded as in chapter 3 was presented. The partial testing could be implemented e.g. as in chapter 5 in the figure 5.2, dissipation factor measurements should be considered case by case. This is highly recommended especially in areas where the other measurements during the life-cycle are unprofitable and the required quality of the installation is significant or the installation environment is challenging. On the other hand, if the costs of expansion can be pressed low enough then the expansion could be taken in use for all new installations. In this way it is possible to get so called finger prints from every cable which can be used later in the urban and city areas for condition management purposes. In order for this expansion to be possible, negotiations with contractors or service providers should be started and the real costs and availability of the operations clarified. At the moment, the larger existing cable networks located urban and city areas have some parts which are at the end of their computational life-cycle. From these areas would be advisable to find the targets where the off-line measurements performing are possible without any harm to customers. The risk level which would be willingly accepted should then be decided, by using the benefit calculation method which was presented in chapter 8.3.2. For the most critical cables the measurements can be performed and the operations should be expanded for the less critical cables when the learning process of the interpretation of the results is improving. The cable sections where the off-line measurements are not suitable, but where the customer is very important should be determined. The most critical of these objectives would be considering the continuous monitoring, which can be rotated over few cable as in the chapter 8.2 was presented. In terms of cable diagnostic and its profitability, it is important that the indirect costs, which are due to organisation of the measurements and results’ interpretations, can be minimized. In order to achieve this it requires some kind of an operation management system that allows process operations. The main characters of this kind of system were presented in chapter 9 in the figure 9.1. It is based on the cable database where analyses can be done which describe the condition of the network and in this way the maintenance operations can be directed condition based. There is some kind of model in the concern level, but to modify it compatible for Vattenfall Verkko Ltd. and measurement service which is available in Finland requires development. In practical terms, this requires that documentation of the cable network must be done more carefully and saved into the network information system where the final analyses could be implemented. The advantage of the online partial discharge measurement was that its performing does not affect any disturbance to normal operation. However, the measurement performing needs to be developed in operational level, due to electrical safety aspects. The sensors’ installation can be dangerous. Therefore, the development of operation should be to develop methods so that this risk can be eliminated. In practical terms, it means that some kind of voltage tool has to be found, which allows the installation without getting

83

too close to parts where there is voltage. Another solution to the problem could be e.g. constructional modification of the secondary substation such as that the installation of sensors on the earth straps of terminations is safety. The problems mentioned above require cooperation with measurement equipment and network component manufacturers.

84

REFERENCES [1] Ericsson company, Product information, [WWW] Available at: http://www.ericsson.com/ourportfolio/products/medium- voltagecables?nav=fgb_101_275%7Cfgb_101_271 [2] Hampton, Hartlein, Lennertsson, Orton and Ramachandran, Long-life XLPE insulated power cable, [WWW] Available at: www.neetrac.gatech.edu/publications/jicable07_C_5_1_5.pdf [3] Vahlstrom, W., "Strategies for field testing medium voltage cables," Electrical Insulation Magazine, IEEE, vol.25, no.5, pp.7-17, Sept.-Oct. 2009 doi:10.1109/MEI.2009.5276074 URL: http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=5276074&isnu mber=5271883 [4] Gulski, E., Wester, F., J., & all. Knowledge rules support for CBM of powercable circuits, 2002, pp. 12, [WWW] available at: ieeexplore.ieee.org/iel5/10834/34152/01627160.pdf [5] Lähdeaho T., Maaseudun keskijännitekaapeli, Diplomityö, 2007 [6] Baur, The high voltage company presentation, subject Cable diagnostic and preventive maintenance, 11/2009 [7] Hirji A. General Manager, EPD RAYCHEM RPG LTD, Reliability aspects of power cables and accessories, [WWW] Available at: http://www.scribd.com/doc/29421861/Reliability-Aspects-of-MV-XLPE-Cable- Joints-and-Terminations [8] Mäkiranta, M. Vattenfall Verkko Oy, Finland, Interview 28.7.2010 [9] Koneurakointi Kaski Oy,[WWW] Available at: http://www.koneurakointikaski.fi/index.php?pinc=2 [10] Maasaha Oy, [WWW] Available at: http://www.maasaha.fi/albumi/kuviatyomailta/661485?back=history [11] Marais Grouppe, [WWW] Available at: http://www.samarais.com/index.php?id=68&L=1&tx_obladygestionparc_pi1[pro duct]=21&cHash=802f98540a]

85

[12] Hyvönen, P., Prediction of insulation degradation of distribution power cables based on chemical analysis and electrical measurements, Doctoral dissertation, Espoo 2008, [WWW] available at: http://lib.tkk.fi/Diss/2008/isbn9789512294039/ [13] Densley, J., "Ageing mechanisms and diagnostics for power cables – an overview," Electrical Insulation Magazine, IEEE , vol.17, no.1, pp.14-22, Jan.- Feb.2001 doi:10.1109/57.901613 URL: http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=901613&isnu ber=19500 [14] Brändström, Torkildsson, Elfving, Rüter, Feasibility study- Estimating remaining lifetime of cables, Vattenfall research and development AB, 2009, unpublished material [15] SLO Oy, presentation, subject Materiaali perehdytys Vattenfall, Jarmo Lindberg [16] Cablejoints.uk.com, [WWW] Available at: http://www.cablejoints.co.uk/sub- product-details/cable-sheath-test-testing [17] Larsson, A., Boo, C., Buhr, A., Tekniska riktlinjer max 24 kv inom Vattenfall eldistribution AB, 2009, pp. 57, unpublished material [18] Halim, H.S.A., Ghosh, P., "Condition Assessment of Medium Voltage Underground PILC Cables Using Partial Discharge Mapping and Polarization Index Test Results," Electrical Insulation, 2008. ISEI 2008. Conference Record of the 2008 IEEE International Symposium on , vol., no., pp.32-35, 9-12 June

2008 doi:10.1109/ELINSL.2008.4570270 URL: http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=4570270&isn mber=4570258

[19] Emerson process management company, [WWW] Available at: http://www2.emersonprocess.com/ENUS/BRANDS/ELECTRICALRELIABILIT YSERVICES/PLCS/PREDICTIVDIAGNOSTICS/Pages/PartialDischargeTesting Monitoring.aspx [20] Dubickas, V., Brief literature study on methods for assessing condition of cables and accessories, 2009, unpublished material

86

[21] High voltage Inc., Tan δ (DELTA) Cable Testing overview and answers, Available at: www.hvinc.com/downloads/Tan_Delta_FAQ.pdf [22] Brändström, F., Test och tillståndskontroll av kabelnät, Vattenfall Eldistribution AB, 2007, pp. 37, unpublished material [23] “IEEE Guide for field testing and evaluation of the insulation of shielded power

cable systems,” IEEE Std 400-2001 (Revision of IEEE Std 400-1991), 2002. [24] HVPD Company UK, Better Asset Management through: On-line Partial Discharge Testing and Monitoring of MV and HV Networks, [WWW] available at: http://www.hvpd.co.uk/products/ [25] Renforth, L., Mackinlay, R., Seltzer-Grant, M., Deployment of distributed on-line partial discharge monitoring devices on medium voltage electricity networks, CIRED, Prague, 8-11 June 2009 [26] Svensson, M., Smedbo, F., Presentation Quality control of cable networks, Vattenfall Service Sweden, 03/2010 [27] Energiamarkkinavirasto, Verkkokomponentit ja indeksikorjatut yksikköhinnat vuodelle 2011 (alv 0%), [Network components and index corrected unit prices in year 2011], 30.9.2010, [WWW] Available at: http://www.energiamarkkinavirasto.fi/files/Sahkoverkkokomponenttien_yksikkoh intataulukko_2011.xls [28] North Atlantic Treaty Organisation NATO, Research and Technology Organisation RTO, Publication SAS-069, Code of practice for life cycle costing, pp.64, 2009, [WWW] Available at: www.rta.nato.int/pubs/rdp.asp?RDP=RTO- TR-SAS-069 [29] Atkinson, Banker, Kaplan, Young, Management accounting, Third edition, 2004, pp.595. [30] HVPD Ltd. UK, [WWW] Available at: http://www.hvpd.co.uk/products/ [31] KEMA, [WWW] Available at: http://www.kema.com/services/consulting/reliability/online_pd/Default.aspx [32] HVPD Ltd. UK, Proposal and quotation, 27.07.2010, unpublished material.

87

[33] Oates, R., HVPD Ltd. UK, Interview by email 27.8.2010 [34] KEMA, [WWW] Available at: http://www.kema.com/Default.aspx [35] HVPD Ltd. UK, 2009, HVPD-LongshotTM Diagnostic Partial Discharge (PD) Spot Tester & Short-Term Monitor, Summary User Guide, May 2009, Available at: www.hvpd.co.uk [36] HVPD Ltd. UK, MV Network Asset Management through On-line Condition Monitoring, 2009, [WWW] Available at: http://www.hvpd.co.uk [37] Johnson, Gerry, Exploring corporate strategy: text & cases, 8th ed., 2008, pp. 878. [38] Paananen, H., Kunnossapito-ohjelma, Vattenfall Verkko Ltd., 2010, pp.15,

unpublished material [39] Slack, Nigel, Operations management, 4th ed., 2004, pp.794. [40] Opetushallitus, Kunnossapito, Tuottava kunnossapito, [WWW] Available at: http://www03.edu.fi/oppimateriaalit/kunnossapito/perusteet_54_tuottava_ kunnosapito.html [41] Willems, J., RCM vai järkevä kunnossapito ja optimaalinen luotettavuus, Kunnossapitolehi 7/2006, [WWW] Available at: http://www.promaint.net/lehti [42] Electropedia, [WWW] IEC, International Electrotechnical Commission, [Cited1.11.2010] Available at: http://www.electropedia.org/ [43] Energiamarkkinavirasto, [WWW] Available at: http://www.emvi.fi/data.asp?articleid=1699&pgid=133 [44] Maurer, E., KEMA Netherland B.V., Phone interview 16.9.2010 [45] Energiamarkkinavirasto, Methods of determining the return DSOs during the regulatory period 2008-2011, [WWW] Available at: http://www.emvi.fi/data.asp?articleid=1699&pgid=133 [46] Keränen, J., Dektra industrial Oy, Phone interview 7.12.2010

88

[47] Uusi-Rauva, E., Ohjauksen tunnusluvut ja suoritusten mittaus, Second edition, Tampere 1994, pp.76 [48] Pylvänäinen, J., Reliability analysis for distribution networks combined with transformer condition assessment, Doctoral dissertation, Tampere, 2010 [49] Järventausta, P., Sähkövoimatekniikan kurssi, luentokalvot, TTY, Tampere, 2010, [WWW] Available at: http://webhotel2.tut.fi/units/set/opetus/kurssit/Materiaalisivut/SVT_3310/Luotetta vuus_2010_verkkoon.pdf [50] Nevalainen, P., Nousiainen, K., Rogowski coil in partial discharge measurements on MV networks, NORD-IS03, 2003, Tampere, Finland [51] Energiamarkkinavirasto, Sähkön jakeluverkkotoiminnan ja suurjännitteisen jakeluverkkotoiminnan hinnoittelun kohtuullisuuden valvontamenetelmien suuntaviivat vuosille 2012-2015, 14.1.2011, [WWW], Available at: http://www.emvi.fi/files/Sahko_jakeluverkko_suurjannitteinen_jakeluverkkotoim inta_%20suuntaviivat%202012-2015_luonnos_14012011.pdf

89

APPENDIX 1 - EQUATION FOR DISADVANTAGE CAUSED BY ELECTRICITY DELIVERY OUTAGES

kt

t

hsatTDAt

planwtplanplanEtplan

unwtununEtun

kt BCIT

W

hHSAhTDA

hONhOT

hONhOT

DCO

1,,,,

exp,exp,exp,exp,

,

Where

DCOt,k = Actual disadvantage (outage cost) caused by electricity supply outages to the customers of the network operator in the year t in the value of money of the year k.

OTunexp,t = Customer’s average annual outage time weighted by annual

energies, caused by unexpected outages in the 1–70 kV network in the year t, hour.

hE,unexp = Price of disadvantage caused by unexpected outages to the

customer, EUR/kWh, in the 2005 value of money.

ONunex,t = Customer’s average annual number of outages weighted by annual energies, caused by unexpected outages in the 1–70 kV network in the year t, number.

hW,unexp = Price of disadvantage caused by unexpected outages to the

customer, EUR/kW, in the 2005 value of money.

OTplan,t = Customer’s average annual outage time weighted by annual energies, caused by planned outages in the 1–70 kV network in the year t, hour.

hE,plan = Price of disadvantage caused by planned outages to the

customer, EUR/kWh, in the 2005 value of money.

ONplan,t = Customer’s average annual number of outages weighted by annual energies, caused by planned outages in the 1–70 kV network in the year t, number.

hW,plan = Price of disadvantage caused by planned outages to the

customer, EUR/kW, in the 2005 value of money.

90

TDAt = Customer’s average annual outage number weighted by annual energies, caused by time-delayed autoreclosers in the 1–70 kV network in the year t, number.

hTDA = Price of disadvantage caused by time-delayed autoreclosers to

the customer, EUR/kW, in the 2005 value of money.

HSAt = Customer’s average annual outage number weighted by annual energies, caused by high-speed autoreclosers in the 1– 70 kV network in the year t, number.

hHSA = Price of disadvantage caused by high-speed autoreclosers to

the customer, EUR/kW, in the 2005 value of money.

Wt = The amount of energy transferred to the users from the network operator’s network with a voltage of 0.4 kV and 1–70 kV in the year t, kWh.

Tt = Number of hours in the year t.

kBCI = Change in the building cost index for the year k.

The change in the building cost index in the above formula for calculating outage costs for the year k will be calculated with the following formula:

12004

1

BCI

BCIBCI k

k

Where

kBCI = Change in the building cost index for the year k.

kBCI = The average of the building cost index (1995=100) index

figures for April–June in the year k. Adapted from [43].

91

APPENDIX 2 - EQUATION FOR REFERENCE VALUE OF DISADVANTAGE CAUSED BY ELECTRICITY DELIVERY OUTAGES During the supervising period 2008 – 2010 reference value for year k is calculated as following:

4

2008

2005,

,

t t

kkt

kref

W

WDCO

DCO

Where DCOref ,k = Actual disadvantage (outage cost) caused by electricity supply

outages to the customers of the network operator reference value of the year k.

DCOt,k = Actual disadvantage (outage cost) caused by electricity supply

outages to the customers of the network operator in the year t in the value of money of the year k.

Wk = The amount of energy transferred to the users from the network

operator’s network with a voltage of 0.4 kV and 1–70 kV in the year k, kWh.

Wt = The amount of energy transferred to the users from the network

operator’s network with a voltage of 0.4 kV and 1–70 kV in the year t, kWh.

Adapted from [43].

92

APPENDIX 3 – YEAR 2011 UNIT PRICES FOR COMPONENTS AND WORK OF THE ELECTRICITY NETWORK

The unit prices of the year 2011 are presented in tables 1 and 2 which are used in this thesis.

Table 1 Year 2011 unit prices for components and work of electricity network in Finland. Adapted from [27]

20 kV underground cable installation Unit Unit price (€)

Cross-sectional area ≤ 70 mm km 25130

Cross-sectional area 95-120 mm km 34410



Cross-sectional area ≤ 70 mm (installation in water)

km 57910

Joint piece 1 260

Terminal piece 2 630

Switchgear terminal piece 2 480

Table 2 Additional cost which is caused by the installation environment. Adapted from [27]

Installation environment Unit Unit price (€)

Rural area km 10 130

Urban area km 21 580

City km 66 730

The most commonly used cable sizes are 95 mm and 150 mm. These sizes are used in urban and city installation environments. The smaller size of these two can be also used in the case of frame lines in rural areas. Usually, the size of cross-sectional area of the cable is smaller than 70 mm in rural areas.

ANTTI KESKINEN STRATEGY FOR MEDIUM VOLTAGE CABLE … julkiset dtyot/Keskinen_Antti_julk.pdf · ERP...

Documents

Transcript of ANTTI KESKINEN STRATEGY FOR MEDIUM VOLTAGE CABLE … julkiset dtyot/Keskinen_Antti_julk.pdf · ERP...