• No results found

Test Grids for Reliability analysis : Analysing interruptions and developing test grids based on Mälarenergi Elnät distribution grid

N/A
N/A
Protected

Academic year: 2021

Share "Test Grids for Reliability analysis : Analysing interruptions and developing test grids based on Mälarenergi Elnät distribution grid"

Copied!
67
0
0

Loading.... (view fulltext now)

Full text

(1)

TEST GRIDS FOR RELIABILITY

ANALYSIS

Analysing interruptions and developing test grids based on Mälarenergi Elnät

distribution grid

JONAS ELD

JENS MELIN

School of Business, Society and Engineering Energy Engineering

Degree Project 30 hp

Master program in engineering, Energy System

Supervisor: Jan Sandberg Examiner: Maher Azaza

Customer: Stefan Svensson, Mälarenergi Elnät AB Date: 2017-06-08

(2)

ABSTRACT

The purpose of this degree project is to examine power outages that has occurred during the years 2009 and 2016 on Mälarenergi Elnät distribution grid. Interruptions that has occurred on 10 kV voltage level and on overhead lines or underground cable was examined. The examined interruptions are based on the DARWin data from Mälarenergi Elnät. The examined interruptions resulted in four different test grids. Test grid A are overhead line grid in rural area, test grid B are mixed grid in rural area, test grid C are underground cable grid in semi urban area and test grid D are underground cable grid in urban area. Test grid C and D which have the highest customer density were then used for reliability analysis. The reliability analysis was performed using Microsoft Excel. The main focus for the reliability analysis were the impact of different sectioning times. Another focus for the reliability analysis was introducing breakers at key-points in test grid C. The primarily cause of interruptions on overhead line grids are weather related events and on underground cable grids it is digging and fabrication or material errors. It is concluded that with quicker sectioning times, the reliability of the grid increases linearly. The introduction of breakers at key-points in the grid results in increased reliability.

Keywords: Sectioning time, disconnector, breaker, distribution system, reliability, underground cable, overhead line, test grid

Nyckelord: Sektioneringstid, frånskiljare, brytare, distributionssystem, tillförlitlighet, markkabel, friledning, typnät

(3)

PREFACE

This degree project in Master of Science energy system have been carried out at Mälarenergi Elnät AB. We would like to thank the people at Mälarenergi Elnät AB who has helped us with the project. Special thanks to our supervisor Stefan Svensson for providing support and technical knowledge throughout the degree project. We would also like to thank Johanna Rosenlind for the support throughout the project. Also special thanks to Marcus Nilsson for his expertise regarding the operations of the grid.

We would furthermore like to thank our supervisor Jan Sandberg at Mälardalen University who has contributed throughout this degree project.

Västerås in June 2017 Jonas Eld & Jens Melin

(4)

SUMMARY

Power distribution with a high reliability is necessary for the modern society. The distribution grid plays a fundamental role in order to supply reliable electricity to the consumers. Interruptions in the delivery of electricity comes with a high societal cost. Different sectors and customers have different costs for an interruption.

The purpose of this degree project were to examine the interruptions that have occurred on the voltage level of 10 kV affecting the overhead lines and underground cables within the Mälarenergi Elnät distribution grid. From the examined interruptions different test grids were developed. Further aspects that was considered were the cause of the outage and what day they occurred on.

DARWin data for the years 2009 to 2016 was used to conduct the degree project. The software Trimble NIS was used to examine the grid and collect the information of the cable types, substations, customers and the structure of the grid. Based on the data collected from DARWin and Trimble different test grids were created with different criteria’s. The criteria’s for the construction of the test grids were the amount of underground cable compared to the amount of overhead line and as sub criteria the customer density was used.

The test grids with the highest customer density was then used for reliability analysis. The cost for an interruption on these test grids was also calculated, based on the customer composition on these grids. Since the grids were structured as loops with multiple possible back-up feeds, different sectioning times were used to examine how the reliability indicators varied with shorter or longer sectioning times.

One additional scenario was also considered. This scenario consisted of installing breakers at key-points in the grid. For the base scenarios, breakers were only installed in the primary substation.

The examined components in the distribution grid, overhead line and underground cable, have the most number of interruptions. The cause of the interruptions are on overhead line mainly weather related and for underground cable it is digging and fabrications or material failures that are the most common causes. The spread on the numbers of interruptions during the week are quite even, except for the underground cable which has a clear drop during the weekends. The reason for the drop are that the interruptions that is caused by digging is not as frequent during the weekends. It is concluded that the interruption cost changes linearly with the sectioning time. Installing breakers at key-points in the grid increases the reliability.

(5)

SAMMANFATTNING

Eldistribution med en hög tillförlitlighet är en viktig aspekt för det moderna samhället. Kostnaderna för samhället vid strömavbrott är höga och olika sektorer samt kunder har olika kostnader vid ett eventuellt avbrott. Distributionsnätet spelar en stor roll i tillförlitligheten för distributionen av elektricitet.

Syftet med detta examensarbete är att undersöka de avbrott som har skett på 10 kV’s friledningar och markkablar inom Mälarenergi Elnäts distributionsnät. Sedan skapades typnät utifrån de mest drabbade nätsektionerna. Vidare så undersöktes också orsaken till avbrotten även vilken dag i veckan avbrotten inträffade.

DARWin data från åren 2009 till och med 2016 användes som grund för detta arbete. Programmet TRIMBLE NIS användes för att studera strukturen på distributionsnätet och för att sammanställa data om markkablar, friledningar, nätstationer och kunder. Typnäten som sedan konstruerades baserades på denna data. Typnäten delades in beroende på andelen markkabel och kundtätheten.

De typnäten med störst kundtäthet användes sedan för att genomföra tillförlighetsanalyser. Kundkostnaderna för ett eventuellt avbrott beräknades också för dessa typnät. Typnäten var uppbyggda som slingade nät med flera möjliga omkopplingar. Detta gjorde att sektioneringstiden för dessa omkopplingar användes som variabel för tillförlighetsanalysen. Ytterligare ett beräkningsscenario genomfördes. Detta scenario byggde på att installera extra brytare i vissa viktiga punkter i nätet. Som standard var endast mottagningsstationerna installerade med brytare.

De studerade komponenterna, friledning och markkabel, är de komponenterna i distributionsnätet som har flest avbrott. De vanligaste orsakerna för dessa avbrott är väderrelaterade på friledningar och avgrävning samt fabrikationsfel på markkablar. Avbrotten är jämt fördelade över veckan, förutom för markkablar där antalet avbrott minskar tydligt på helger. Anledningen till denna minskning på helger är att antalet grävningar och då också antalet avgrävningar minskar. Avbrottskostnaderna ändrades linjärt med sektioneringstiden. Att installera brytare in i viktiga punkter i nätet höjde tillförlitligheten.

(6)

TABLE OF CONTENT

1 INTRODUCTION ...1 1.1 Background ... 1 1.1.1 Mälarenergi Elnät AB ... 2 1.1.2 Previous research ... 2 1.2 Problem formulation... 3

1.3 Purpose of degree project ... 3

1.4 Research questions ... 4

1.5 Delimitation ... 4

2 METHOD ...5

2.1 The literature review ... 5

2.2 The empirical research ... 5

3 THE SWEDISH ELECTRICITY GRID ...6

3.1 Grid structure ... 6

3.2 The Swedish electricity market reform 1996 ... 8

3.2.1 Reliability of delivery ... 9

3.2.2 The impact of Gudrun ... 9

3.2.3 DARWin ...10

3.2.4 Interruption cost ...12

3.3 Overhead line and underground cable...13

3.3.1 Overhead line ...13

3.3.2 Underground cable ...15

3.4 Control system ...17

4 DISTRIBUTION SYSTEM RELIABILITY ... 19

4.1 DARWin-file ...19

4.1.1 Overhead line ...19

4.1.2 Underground cable ...19

(7)

4.2.1 Data collection ...20

4.2.2 Classification of the test grids ...20

4.2.3 Description of the structure of the test grids ...21

4.2.4 Specifications of the different test grids ...22

4.3 Analytical reliability analysis ...26

4.3.1 Failure rate ...26

4.3.2 Unavailability ...27

4.3.3 Reliability indicators ...27

4.3.4 Base scenario ...28

4.3.5 Different sectioning times ...28

4.3.6 Design suggestion for Test grid C ...28

4.3.7 Calculations ...29

5 RESULTS ... 31

5.1 Histograms of the interruptions ...31

5.2 Test grid A ...33

5.3 Test grid B ...34

5.4 Test grid C ...35

5.5 Test grid D ...37

5.6 Design suggestion for test grid C ...39

5.7 Sensitivity analysis ...40 6 DISCUSSION... 43 6.1 Empirical research ...43 6.2 Result ...43 7 CONCLUSIONS ... 45 8 FUTURE WORK ... 45 REFERENCES ... 46

(8)

APPENDIX 1 – CABLE DESIGNATION STANDARDS APPENDIX 2 – TEST GRID A

APPENDIX 3 – TEST GRID B APPENDIX 4 – TEST GRID C APPENDIX 5 – TEST GRID D

APPENDIX 6 – DESIGN SUGGESTION TEST GRID C APPENDIX 7 – DESCRIPTIONS OF CALCULATIONS

LIST OF FIGURES

Figure 1: Double cable structure. ... 7

Figure 2: Loop structure. ... 7

Figure 3: Radial structure. ... 8

Figure 4: Free overhead line (isolated and uninsulated) in use at the voltage level of 10 kV...14

Figure 5: Line street overhead line. ... 15

Figure 6: Underground cable types in use at the voltage level of 10kV in the MEE grid. ...16

Figure 7: Schematic over Test grid A. ... 23

Figure 8: Schematic over Test grid B. ... 23

Figure 9: Schematic over Test grid C. ... 24

Figure 10: Schematic over Test grid D. ... 24

Figure 11: Schematic over the design suggestion for Test grid C. ... 29

Figure 12: Interruptions on components at the voltage level of 10 kV during the years 2009-2016. ... 31

Figure 13: Interruption causes on overhead line and underground cable 10 kV 2009-2016. . 32

Figure 14: Number of interruptions per day at the voltage level of 10 kV and on overhead line and underground cable during the years 2009-2016. ... 32

Figure 15: Interruptions on overhead line sorted by day. ... 33

Figure 16: Interruptions on underground cable sorted by day. ... 33

Figure 17: Interruptions by category on Test grid A. ... 34

Figure 18: Interruptions by cause on Test grid A. ... 34

Figure 19: Interruptions by category on Test grid B. ... 35

Figure 20: Interruptions by cause on Test grid B. ... 35

Figure 21: Interruption by category on Test grid C. ... 36

Figure 22: Interruptions by cause on Test grid C. ... 36

Figure 23: Interruption costs Test grid C. ... 37

Figure 24: Interruptions by category on Test grid D. ... 38

Figure 25: Interruptions by cause on Test grid D. ... 38

(9)

Figure 27: SAIFI & SAIDI for the design suggestion for Test grid C, fault occurring in the east

feed. ... 39

Figure 28: SAIFI & SAIDI for the design suggestion for Test grid C, fault occurring in the west feed. ... 40

Figure 29: Interruption cost for the design suggestion for Test grid C. ... 40

Figure 30: SAIDI of the MEE grid, based on all the examined interruptions. ...41

Figure 31: SAIFI for MEE grid, based on all the examined interruptions. ...41

LIST OF TABLES

Table 1: The affected component causing the interruption. ... 11

Table 2: The cause of the interruption. ...12

Table 3: SNI-codes and interruption costs. ... 13

Table 4: Advantages and disadvantages with overhead line. ... 15

Table 5: Advantages and disadvantages with underground cables. ... 17

Table 6: K-factors and T-factors. ...21

Table 7: Technical data for the different test grids. ... 22

Table 8: Line and Cable composition of the different test grids. ... 22

Table 9: Number of customers of each category on each test grid. ... 25

Table 10: Annual electricity consumption per customer category and test grid. ... 25

Table 11: Length & number of interruptions on underground cables & overhead lines. ... 26

Table 12: Failure rates for overhead lines and underground cables. ... 26

Table 13: Sectioning times for the different scenarios. ... 28

Table 14: Example of calculations. ... 29

Table 15: Reliability indicators Test grid C. ... 36

Table 16: Reliability indicators Test grid D. ... 38

Table 17: Test grid C sensitivity analysis. ...41

Table 18: Test grid D sensitivity analysis. ... 42

ABBREVIATIONS AND NOTATIONS

Abbreviation/Notation Unit Description

ASAI Interruptions/Customer,

Year Average Service Availability Index

ASUI Hours/Customer, Year Average Service Unavailability

Index

CAIDI Hours/Interruption Customer Average Interruption

Duration Index

DSO - Distribution System Operator

Ei - Energy Market Inspectorate

(10)

Abbreviation/Notation Unit Description

MEE - Mälarenergi Elnät

r h Repair time

SAIDI - System Average Interruption

Duration Index

SAIFI h System Average Interruption

Frequency Index

U h Unavailability

(11)

1

INTRODUCTION

Power distribution with a high availability is necessary for the modern society and the future development of the modern world. (Wallnerström & Grahn, 2016)

In order to supply reliable electricity to the consumer, the power distribution grid plays a fundamental role (Cadini, et al., 2017). The baseline of this degree project is to examine the current most affected overhead line and underground cable grids, at 10 kV, and by different criteria construct them into several test grids. Then, the test grids could be used for analytical reliability analysis.

1.1 Background

Power outages generates a high societal cost. In the case of an outage, all the different parts of the society is affected. Sectors and customers are affected differently; some industries and services are affected independently of the length of the outage, and some are more affected by longer outages. Hospitals and other crucial public services often have backup generators, but the vulnerability of operation is significantly increased following power outages. (Wallnerström & Grahn, 2016)

The electricity act (SFS 1997:857) specifies that the transfer of electricity should be of good quality (3rd Chapter, 9§)

The Energy market inspectorate (Ei) in Sweden also specifies the requirements for power distribution quality. Those requirements include for instance, the maximum number of power outages per year, and voltage quality, among others. The Distribution System Operator (DSO) is responsible for the distribution grid in their area. (Morén, 2013)

Each year, the DSOs is responsible for submitting information about the power outages in their grid to Ei, through a system called DARWin. DARWin is a statistical database that makes it possible for the DSOs and Ei to steer investments and maintenance to those parts in the grid where the investments are needed mostly. (Wallnerström & Grahn, 2016)

The distribution grid is a local monopoly and therefore there is a need for some control on how much the customer will have to pay to the DSO to be connected to the grid. In the year 2009 the government in Sweden decided to change the electricity act (SFS 1997:857) so that the grid tariffs’ fairness should be examined beforehand in a so called ex-ante regulation. This regulation means that Ei in advance, decides the allowed revenue for each DSO depending on the reliability of the grid. (Ström, 2015)

(12)

The reason for this regulation is to counteract the potential risk of decreasing quality due to the monopoly situation for the DSOs. There is a risk that the reliability is decreasing in the grid in order to maximize the profit for the DSOs. (Ström, 2015)

To steer towards increased grid reliability cost effectively it is helpful to construct test grids, which can be a normalization of the actual grid for the current DSO. Using these test grids, it is possible to calculate the effects of different improvement efforts before implementing the new changes (Engblom & Ueda, 2007). One way to measure and compare the effects of power outages is to calculate the SAIFI and SAIDI for each outage. SAIFI is the number of customers affected and SAIDI is how long each customer is affected by the power outage.

The more precise the test grids can be, the more helpful they are for the DSO. In this degree project, different test grids are constructed based on the current grid at MEE. The DARWin report from the DSO Mälarenergi Elnät AB is used to search for certain areas in the grid where power outages are more or less clustered and moreover if there are any distinct similarities or differences between these areas.

1.1.1 Mälarenergi Elnät AB

Mälarenergi Elnät AB is the DSO for the grid in the municipalities of Arboga, Hallstahammar, Kungsör, Köping and Västerås. The grid stretches over the western part of Mälardalen. MEE distributes electricity to over 100 000 customers through the local distribution grid. As the DSO for this area, MEE is responsible for operation and maintenance of the distribution grid. According to the regulations stated by Ei, MEE is accountable for the quality of distribution. (Mälarenergi AB, 2015)

1.1.2 Previous research

Previous work has been done when constructing test grids that is representative for Sweden as a whole. The purpose of that project was to construct different test grids for the countryside and urban areas. These grids were used for reliability and interruption cost calculations. (Engblom & Ueda, 2007)

Other studies have created reliability test systems for learning purposes, based on statistics from other studies and then combined into a test system. The test system contains all the components found in a real system but it is small enough for students and others to perform calculations by hand. (Allan, et al., 1991)

The test system constructed by (Ueda, et al., 2009) describes the Swedish distribution grid as two different type grids, one rural and one urban. This is a simplification of Swedish grid. All these test systems have been created to fit many different grids due to an underlying purpose of generalisation. In this degree project test grids are constructed with the specific grid at MEE as a base system, due to a purpose of specificity.

(13)

1.2 Problem formulation

Overhead lines are often replaced with underground cables to get rid of weather-related power outages. Decreasing outages has a direct correlation to increased grid reliability. There are disadvantages with underground cables, compared to overhead lines. Underground cables are 2 – 10 times more expensive than overhead line and it is harder to localize faults along the length of the cable (Svensson, 2012).

Replacing overhead line with underground cable is a measure to decrease the number of outages. For the duration of an outage, the sectioning time is of importance. The sectioning time is the time it takes to isolate the fault and reconnect the customers. Longer sectioning time leads to an increased cost. The sectioning is mainly done manually by fitters on sight. The sectioning time often depends on the experience of the operators and fitters.

There will be both overhead lines and underground cables in the future electricity grid, hence creating test grids comprising both of them is of significant importance.

The most important aspect of using test grids, involves and understanding on how to tailor the model to suit the particular needs of the MEE grid. Previously-developed test grids could be considered too general and the underlying assumptions could be too restrictive, prohibiting a valuable analysis. In reality, the grid is a complex entity and looks different depending on aspects such as: amount of underground cables, overhead lines, rural and urban areas and the geographical location. All of this points to the fact that a tailored test grid, representable for the MEE grid, should be considered superior when it comes down to examining and analysing the grid reliability.

1.3 Purpose of degree project

The purpose of this degree project is to examine the power outages that has occurred on overhead lines and underground cables at the voltage level of 10 kV within the MEE grid. Further aspects that will be considered are: the cause of the power outages and the day of the outages. This degree project will be performed by using test grids constructed by relevant criteria with a relevant connection to the MEE grid. Then, the test grids are used for an investigation of the different types of outages and their corresponding consequence. The outcome of the analysis will be the interruption costs that occurs in the event of an interruption.

(14)

1.4 Research questions

The research questions in this degree project are stated in the following bulleted list.  What impact have the day of the week on the number of interruptions?

 What causes the interruptions that occurs in the different grid sections?

 How does the sectioning time affect the interruption cost in semi urban and urban grids?  How does more breakers in the grid influence the reliability?

1.5 Delimitation

The examined power outages in this degree project are the ones that occurs on underground cables and overhead lines. These two components are seen as the major criteria’s when creating test grids. The examined interruptions are spread throughout the MEE concession area. The voltage level examined is 10 kV. Planned outages are not considered in this degree project. The outage data that is examined originated from events occurring during the years 2009-2016, but the grid could, in some, locations be altered with new equipment and/or changes from overhead lines to underground cable could have occurred, for example. This factor has not been taken into consideration, and the grid “as is” right now is assumed to be the same as 2009. For the reliability and cost calculations, only one interruption at a time will be considered. The reliability and cost calculations will be performed on the test grids with the highest customer density.

(15)

2

METHOD

The method to achieve the objectives in this degree project is divided in to two parts: literature reviewing and empirical researching.

2.1 The literature review

A literature study was carried out by reading articles journals, books and degree projects within the subject. Electrical power distribution systems were examined. Theories about reliability was also examined. The Swedish electricity markets rules and regulations were examined. The article journals have been found on different databases through the university library at Mälardalen högskola. The construction of overhead lines and underground cables have been examined and is described in chapter 3. Construction of test grids has been examined from different articles and discussed with the supervisor at MEE.

2.2 The empirical research

The empirical research includes collecting information through discussions with employees at MEE and gathering data from historical records on power outages. The research is mainly on gathering data from historical records on power outages inside MEE concession area. Data comes from the database DARWin (managed by Energiföretagen Sverige). The empirical research include collecting information from discussions. The discussions have been performed with employees at MEE with extensive experience and technical knowledge on the electricity grid, within their own concession area.

In order to examine the structure of the grid along with geography, the software Trimble NIS is used. Using this software, the grid structure and different parts of the system could be examined. The geography of the area surrounding the grid could also be analysed in this software.

The ten most affected grid sections were examined and classified into different grid types (overhead line, underground cable or mixed) and customer density. For each grid section, certain data was collected.

The data collected during the empirical research were used for modelling the test grids and for reliability calculations. The analytical reliability analysis was conducted using Microsoft Excel.

(16)

3

THE SWEDISH ELECTRICITY GRID

The Swedish electricity grid is divided into three different levels depending on the voltage, the national grid, regional grid and the local grid (also called the distribution grid). The national grid is the backbone of the electricity grid in Sweden (Pettersson, 2017). It transfers the electricity from the large hydro power plants in the northern part of Sweden, to the south where most of the consumers are accommodated. The nuclear power plants are also connected to the national grid (Wallnerström & Grahn, 2016). The voltage in the national grid is 220kV or 400kV, the reason for the high voltage is to minimize the losses when transferring the power over long distances (Konsumenternas Energimarknadsbyrå, 2017).

The national grid is connected to the regional grid, where the voltage is between 40kV to 130kV (Konsumenternas Energimarknadsbyrå, 2017). Wind power plants and other smaller power plants are connected to the regional grid. The regional grid transfers the power from the national grid and the power plants directly connected to the regional grid to the local grid and also to the customers directly connected to the regional grid (Konsumenternas Energimarknadsbyrå, 2017; Wallnerström & Grahn, 2016).

The last level is the local grid (or the distribution grid), this is where most of the customers are connected. The distribution grid is divided into two different voltage levels, 10kV and 0,4kV. Industries and other high consuming customers are connected to the 10kV grid. The distribution grid is constructed in different ways depending on the density of the customers and the need for redundancy (Söder & Amelin, 2011).

3.1 Grid structure

The structure of the distribution grid is mainly dependent on economic aspects. Constructing redundant distribution grids is not economically viable in all places. In Sweden, the distribution grid is mainly structured in one of the three following ways (Engblom & Ueda, 2007).

 Double cable structure  Loop structure

 Radial structure

Normally, these three structures are used and combined into a single grid entity. Double cable structure is used in the city center where a high reliability level is required and economically viable. In this structure, each substation is connected to two parallel cables. This structure has a fast reconnection in case of a power outage, the reconnection from the faulty component is done automatically and therefore the average outage time is less than 1 minute (Engblom & Ueda, 2007). An illustration of a double cable structure is presented in Figure 1 below.

(17)

Figure 1: Double cable structure.

The looped structure is the most common way of structuring a distribution grid. The characteristic of a looped structure is that each substation has the possibility to be feed from a different primary substation (Engblom & Ueda, 2007). The looped grid is almost always run as a radial grid, which means that the loop is split into two radial sections by a disconnector, fed from two different slots from a primary substation. Sometimes the grid is fed from two different primary substations (Lexholm, 2016).

The breakers in the primary substation will, in case of a fault, disconnect all the substations fed from that slot. Before the customers can be reconnected, by closing the disconnector that splits the loop, the fault location needs to be found and this is often done by personnel on site (Lexholm, 2016). Also, the reconnection is done manually, this leads to an average down time equal to 0,5 to 2,5 hours (Engblom & Ueda, 2007). In Figure 2 below, an illustration of a looped structure is presented.

Figure 2: Loop structure.

The radial structure is most commonly used at the countryside, where the consequence of a power outage is limited as compared to more densely populated areas. In this structure, there is no redundancy in the grid. In the case of a power outage, each substation after the fault location will be out of power until the fault is fixed or until a mobile backup power is connected.

(18)

The average down time for a radial grid is approximately between 1 and 12 hours (Engblom & Ueda, 2007).

In a radial containing overhead lines, there is often an installed delayed reconnection on the breaker in the primary substation slot. In case of a transient fault, the customers will have their power back after a short period without any work done by personnel (Lexholm, 2016). Transient errors on overhead lines is often caused by tree branches touching the lines or lightning strikes (Torstensson & Bollen, 2012). An illustration of a radial structure is presented in Figure 3 below.

Figure 3: Radial structure.

3.2 The Swedish electricity market reform 1996

The electricity market was at the beginning a monopoly business, the DSOs was given an obligation to deliver electricity to its customers inside their concession area. Then, the DSOs had a sort of monopoly position of selling electricity to the connected customers (Banefelt & Larsson, 2006).

To make the electricity market more rational and effective, a new electricity market reform was made in 1996 (Banefelt & Larsson, 2006). In this reform, it is stated that production or selling of electricity cannot be conducted by a legal person that is also the concession owner (SFS 1997:857). The reform implies that the DSOs only have an obligation to distribute the electricity and not deliver to its connected customers. This allows the customer to choose where to buy their electricity. The purpose of the reform was to make the electricity market more competitive and to get rid of the monopoly business. Distribution of electricity is still conducted on a monopoly market, but the monitoring and control is done by a new authority (Banefelt & Larsson, 2006).

The supervision of the grid operations is done by the so called “Network authority”. The tasks of the network authority is in the hands of the Energy Market Inspectorate (Ei). There are two parts that the Ei is not in charge of, the safety of the grid - which is in the hands of the

(19)

“Elsäkerhetsverket”, and the competitive part (production and trade) – which is in the hands of the Swedish Competition Authority (Banefelt & Larsson, 2006).

3.2.1 Reliability of delivery

The electricity act (SFS 1997:857) specifies that the transfer of electricity should be of good quality (3rd Chapter, 9§). Quality, in this case, involves delivery reliability and voltage quality. The reliability of the delivery means the probability that the power will be distributed to the customer. The voltage quality of the delivery implies that the power can be distributed to the customer without disruption in the voltage (excluding power outages).

The distribution to low-voltage consumers, such as household, the delivery is of good quality if there are no more than three unannounced outages that exceeds three minutes per year. If the number of unannounced outages exceeds eleven per year, the quality is not of good quality (Morén, 2013).

The general voltage quality regulations are specified in the EIFS 2013:1 7th chapter (Morén, 2013).

The distribution companies have incentives for maintaining a good quality of the distribution provided by the Ei. These incentives were introduced in order to motivate the distribution companies to, on their own, increase the reliability of delivery and to make investments and conduct maintenance on the electricity grid, according to best practice. The delivery reliability affects the income frame for the company, increasing the reliability of the grid can lead to a bigger income frame for the company and vice versa (Wallnerström & Grahn, 2016).

3.2.2 The impact of Gudrun

In 2005, the southern part of Sweden was hit by the storm Gudrun. The storm gave rise to a numerous amount of outages that, not only affected the electricity grid, but also the railway network, telecommunication network and the mobile communication network (Banefelt & Larsson, 2006). For the electricity grid, the local grid was affected the most but the regional grid was also affected (Heden & Johansson, 2005). The impact of the storm was enormous, it left people power less for days, and for some, even weeks (Energimyndigheten, 2015). The direct costs for the DSOs, as a result of the storm, were also extremely high, the highest cost affected the DSO Sydkraft that had an estimated cost of 1 690 000 000 SEK (Heden & Johansson, 2005).

Gudrun was a tipping point when it comes to reliable electricity distribution, after the storm the Swedish government gave a mission to the Ei. The mission was to leave suggestions on how the reliability for the power distribution can be obtained, and regulations that keeps the grid companies to follow these suggestions (Banefelt & Larsson, 2006). The Ei came up with suggestions that could be added into the already consisted electricity act. Two of the suggestions that was added to the electricity act were (Engblom & Ueda, 2007):

(20)

 The duration of an outage should not exceed 24 hours.

 If the duration of an outage lasts between 12 and 24 hours, the DSO will pay a fee of 12,5 % of the costumers calculated yearly grid fee.

The compensation fee, mentioned above, for outages longer than 12 hours is a percentage of the yearly cost for the grid for the customer but no less than 900kr and it increases for every 24-hour period to a maximum of 300 percent of the annual cost for the customer (Engblom & Ueda, 2007).

It further states that the DSOs should every year conduct a risk and vulnerability analysis of its electrical distribution, as another suggestion of the Ei (Banefelt & Larsson, 2006). This will allow the DSOs to have an overview of the electrical distribution system and to see where investment could be allocated to achieve an increased reliability level. The Ei also announced that it will impose on the DSOs to report any outages that has a duration of 3 minutes or higher.

3.2.3 DARWin

During the years 1988 – 1990 a system for interruption and outage reporting was developed that was called DAR. DAR was used by approximately 40 DSOs, DAR was an upgrade from the previous system FAR-81. The new electricity reform in 1996 lead to higher demand for interruption and outage reporting (Jansson, 2000). This followed that the DAR system was upgraded to DARWin.

DARWin has 156 DSOs represented, which equals over 90 % of Sweden’s electricity customers (Tapper, 2016). DARWin is the database that the DSOs reports their yearly interruptions (>3 minutes). Energiföretagen Sverige is responsible for the database (Energiforsk, 2016). The DSOs that submits their data to DARWin reports the following data (Tapper, 2016):

• Interruption (announced or unannounced) • Date of interruption

• Time of interruption

• Number of affected customers • Breaking unit

• Voltage level of interruption

• Affected component of the interruption • Cause of interruption

This degree project focuses on the affected component and the cause of the interruption. The affected component of the interruption is divided into categories. In each of the different categories, there are subcategories where the affected component is further specified. This keeps the DSOs aware of the number of interruptions categorized by component in the electrical grid. The categories and the sub categories are presented in Table 1 (Jansson, 2000):

(21)

Table 1: The affected component causing the interruption.

Stations Regional station Distribution station Switching station Rectifier station Substations Pole station

Concrete station (Indoor operated) Concrete station (Outdoor operated) Sheet metal station

Satellite station Transformer Housed station Other station type

Overhead line Free overhead line (Uninsulated) Free overhead line (Isolated) Hang line

Hang spiral line Other type of line Fuse box

Underground cable Cable in ground Cable in water Cable cabinet Fuse box

The affected component examined in this degree project are free overhead line (uninsulated and isolated), hereinafter referred to as overhead line, and cable in ground, hereinafter referred to as underground cable.

The categories are associated with the cause of the interruption. In the categories, the sub categories specify the cause furthermore. The different categories are a list from weather-based to, unknown. The different categories and subcategories are presented in Table 2 below (Jansson, 2000):

(22)

Table 2: The cause of the interruption.

Weather Falling tree, snow Falling tree, wind Rain Salt Snow Thunder Weather Wind Damage Animals Damage Digging Sabotage Traffic Wood cutting

Material/method Manufacturing or material failure Inaccurate method

Insufficient maintenance Material/method Sizing failure Personal Inaccurate montage

Incorrect operation Personal

Testing

Other Fuse failure

Other Overload Returning load Unknown

3.2.4 Interruption cost

When calculating the cost of a power outage it is common to calculate it from the approximated costs for the affected customer. Different customers have different costs for an outage. The figures presented in Table 3 comes from an update of a previous survey conducted by the University of Gothenburg in 2005, SINTEF was responsible for this update upon the request of Ei (Vefsnmo & Kjolle, 2015). The costs are adjusted to the consumer price index of 2015 (Wallnerström & Grahn, 2016). The cost “SEK/kW” is for every interruption independent of the duration, while “SEK/kWh” depends on the duration of the interruption.

(23)

Table 3: SNI-codes and interruption costs.

Unannounced interruptions SNI code

Customer category SEK/kW SEK/kWh Range 1 Range 2

Agriculture 8 44 01110 - 03220

Industry 23 71 05100 - 43999

Businesses & services 62 148 45110 - 82990 94111 - 96090

Public sector 5 39 84111 - 93290

Household 1 2 97000 - 98200 111111

To calculate the cost for an interruption the following equation is used. Were C is the customer category, P is the total power for that customer category and t is the duration of the interruption.

𝐶𝑜𝑠𝑡 = ∑𝑆𝐸𝐾/𝑘𝑊𝐶𝑖∗ 𝑃𝑖+ ∑𝑆𝐸𝐾/𝑘𝑊ℎ𝐶𝑖∗ 𝑃𝑖∗ 𝑡 Equation 3.1

The different customer categories are divided according to the standard called “Svensk Näringsgrensindelning” (SNI). The SNI-code for household is developed by Ei (Ström, 2015).

3.3 Overhead line and underground cable

For the transfer of electricity in the distribution grid, the two most common techniques are overhead lines and underground cables. This section covers the differences between overhead lines and underground cables. Furthermore, it covers the techniques for constructing both overhead lines and underground cable, and pros and cons with overhead line and underground cable.

3.3.1 Overhead line

Poles support overhead lines. The design of each pole differ and it is dependent on the voltage level. The most common poles are the one made out of impregnated wood. At the highest voltage levels, the poles are made from galvanized steel (Hildenwall & Westberg, 1999). General the overhead lines on the voltage level 10 kV are the types: Free overhead-line (isolated or uninsulated), hang line and hang spiral line. The most common on the 10 kV voltage level are the free overhead line (isolated or uninsulated). A free overhead line is constructed by metal-lines hanging free from each other (Svenska Kraftnät, 2014). The metal lines are made from aluminium alloy or copper, were the aluminium alloy is the more common type. The line is either isolated or uninsulated. Please regard the following examples of overhead lines:

• FeAl 62/0 12 kV

o Uninsulated overhead line with an iron-aluminium alloy, the cross sectional area is 62 mm2.

(24)

The isolated overhead line is less sensitive, than the uninsulated. For example, any contact on the uninsulated line could cause an outage (Svenska Kraftnät, 2014). The different types of overhead lines within MEE concession areas are presented in Figure 4.

Figure 4: Free overhead line (isolated and uninsulated) in use at the voltage level of 10 kV.

The spacing area needed for the overhead line depends on the voltage level. The higher the voltage level, the more area is needed. The area surrounding the overhead line is called a forest street. At the end of the forest street the side area starts. The forest street plus the side area is called a line street. The forest street should be constructed so the tree closest to the line cannot fall on it hence it is dependent on the expected height of the trees. The forest street should be kept maintained during the overhead lines lifetime (Svenska Kraftnät, 2014). An illustration of the spacing area is presented in Figure 5 below (Svensson, 2012):

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% AlMgSi BLX CU FeAl

(25)

Figure 5: Line street overhead line.

Overhead lines affects the surroundings visually and the agriculture and forestry during its lifetime. The distance between the poles is very flexible. This makes the overhead lines less sensitive to the type of construction area (Svensson, 2012).

The advantages and disadvantages with overhead line are presented in Table 4 (Svensson, 2012).

Table 4: Advantages and disadvantages with overhead line.

Advantages Robust and well-tested technique that is rarely damaged if the surrounding is tree safe.

Easy to identify damage

Less expensive technique compared to underground cable Disadvantages Big effect on the landscape

Affects the agriculture and forestry more than underground cable

3.3.2 Underground cable

Underground cables are most common on the lower voltage levels, but they do occur on the higher voltage levels as well. Underground cables are either one phase cable or a three-phase cable. The underground cable consists of a conductor in the middle, and an isolation material surrounding it. Underground cables consists of different types of conductor and isolation materials, to keep the different sorts of cables apart there is a designation system (Svenska Kraftnät, 2014).

The different cables are defined by different designations that consists of letters (Svenska Kraftnät, 2014). Power Cables, control cables and installation cables are defined by Swedish standard by two different systems (Svenska Elektriska Komissionen, 2003):

(26)

o This consists of a designation system that has been long used in Sweden and is applied on national cable types. It applies also on CENELEC harmonized cable types that is not covered in SS 424 07 02.

• SS 424 17 01, Edition 3

o This which fully reflects the designation system of CENELEC HD 361 S3 (System for cable designation). It applies on CENELEC harmonized cable types with a rated voltage level up to 450/750 V.

For a complete cable designation on a cable, the number of conductors, conductor area and in some cases the area of the concentric conductor should be defined. The area of the concentric conductor is separated from the conductor by a slash (/). Below, an example of the cable designation be demonstrated (Svenska Elektriska Komissionen, 2003).

• AKKJ 3x150/41

o The cable have three isolated aluminium phase-conductor and one concentric conductor with the area 41 mm. The isolation and the concentric conductor consists of PVC.

The cable design is defined by different letters that range from A-Z, each letter has different meanings, the letters also have different meaning depending on order (Svenska Elektriska Komissionen, 2003). See Appendix 1 for the cable standard. The different type of underground cable on MEE concession area are presented in Figure 6.

Figure 6: Underground cable types in use at the voltage level of 10kV in the MEE grid.

0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50%

Underground cable

(27)

The advantages and disadvantages with underground cable are presented in Table 5 (Svensson, 2012):

Table 5: Advantages and disadvantages with underground cables.

Advantages Less exposed to weather Less impact on the landscape

Disadvantages Expensive technique compared to overhead line Harder to localize faults

3.4 Control system

The demand for reliable distribution have increased continuously since the start of the electrical era. In order to increase the reliability of the grid, new technologies were invented and implemented to the grid. This technology goes under the name SCADA, “Supervisory Control and Data Acquisition”. SCADA enables remote monitoring of the grid and the different components, such as substations. SCADA can also be used for remote operations of breakers and disconnectors (Nordell, 2008).

When a fault occurs on the 10 kV voltage level the breaker that detects the fault will break. The installed SCADA system will register this event and alert the grid operator. Fault occurring on lower voltage levels will not be detected by the SCADA system and therefore the DSO is dependent on the customer to call in the interruption.

The length of an interruption depends mainly on three events, the sectioning time, the time it takes to find the exact location of the fault, and the repair time (Wallnerström, et al., 2012). In a looped structured grid, when a fault occurs on an underground cable or overhead line, the duration of the outage is only dependent on the sectioning time.

Sectioning shortens the interruption time for some of the affected customers. Isolating the fault in smallest possible area, will decrease the interruption time for many customers but some might still be affected. There are three main ways of sectioning, breakers, remote-controlled disconnectors and manually operated disconnectors.

Remote breakers can be used to automatically disconnect a part of the grid if an interruption occur. Customers closer to the primary substation will not be affected by the interruptions. However, the fault tolerance on the remote breaker must be correct, it must be lower than the tolerance on the primary substation breaker, but too low and it might break even though there is no fault (Wallnerström, et al., 2012).

Remote-controlled disconnectors are used to isolate the fault. In case of an interruption the operator opens all the disconnectors. The disconnectors are then closed one at a time until the power goes out again. When the faulted segment is found the operator connects the grid through a secondary feed. Then the repair crew is sent to the fault location to repair the fault (Wallnerström, et al., 2012).

(28)

Manual disconnectors are operated in a similar way as remote-controlled disconnectors described in the section above. However, since these disconnectors are manually operated a fitter needs to be sent out directly to the affected grid section. Once on sight the fitter can start to open and close disconnectors until the fault is located. Skilled and experienced fitters will find the fault location quicker (Wallnerström, et al., 2012).

(29)

4

DISTRIBUTION SYSTEM RELIABILITY

4.1 DARWin-file

The original DARWin-file consisted of over 5000 interruptions, both announced and unannounced. Excluding the interruptions stated by the constraints in the limitation, the number of interruptions to examine was over 1000.

4.1.1 Overhead line

Approximately 750 interruptions occurred on both uninsulated and isolated overhead line. Except for the information given in the DARWin-file was the following data collected using the software Trimble NIS.

• Total length of overhead line and underground cable • Start location of interruption (substation or disconnector) • Overhead line o Type o Length • Underground cable o Type o Length

• End location of interruption (substation or disconnector)

Grids containing overhead line often includes underground cable as well.

4.1.2 Underground cable

The remaining 300 outages occurred on underground cables. The description of the fault never specified exactly which part of the cable that was the cause of the interruption. A cable section between two substations usually consisted of several different types of underground cable. The following data were collected with Trimble NIS.

• Total length

• Start location of the interruption (substation or disconnector) • Number of joints

• Underground cable o Type o Length

(30)

4.1.3 Evaluation of DARWin-file

Histograms over the interruptions on the examined components were made to see if there was any relationship between the time of the day and the day of the week and the number of interruptions.

4.2 Modelling of test grids

The primary substation that had the most number of interruptions reported for overhead line and underground cables were examined. The most representative feeds when it comes to reported outages were analysed for those primary substations.

4.2.1 Data collection

For constructing the test grids input data is needed, the input data is collected from the software Trimble NIS. The following data were collected:

• Structure o Radial o Looped o Double cable

• Overhead line and or underground cable o Type

o Length • Substation

o Type

o Number of customers o Annual energy delivered • Sectioning points

o Number of possible back-up feeds • Interruptions

o Number of interruptions o Cause of interruption o Duration of interruption o Affected customers

4.2.2 Classification of the test grids

To construct the test grids, the examined feeds are classified by two criteria’s. The main criteria is the K-factor, and the sub criteria is the T-factor.

The K-factor states the composition of the grid. 𝐾 =𝑇𝑜𝑡𝑎𝑙 𝑙𝑒𝑛𝑔𝑡ℎ 𝑜𝑓 𝑢𝑛𝑑𝑒𝑟𝑔𝑟𝑜𝑢𝑛𝑑 𝑐𝑎𝑏𝑙𝑒

𝑇𝑜𝑡𝑎𝑙 𝑙𝑒𝑛𝑔𝑡ℎ 𝑜𝑓 𝑜𝑣𝑒𝑟ℎ𝑒𝑎𝑑 𝑙𝑖𝑛𝑒 [−] Equation 4.1

(31)

𝑇 =𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑠

𝑇𝑜𝑡𝑎𝑙 𝑙𝑒𝑛𝑔𝑡ℎ 𝑜𝑓 𝑔𝑟𝑖𝑑 [

𝐶𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑠

𝑘𝑚 ] Equation 4.2

The terms for the K-factor and the T-factor are displayed in Table 6 below.

Table 6: K-factors and T-factors.

K- Factor T-Factor

K < 0,1 Overhead line grid T << 10 Sparsely populated area K ≈ 1 Mixed grid T < 10 Rural area

K > 10 Underground cable grid 10 < T < 20 Semi Urban area T > 20 Urban area

To construct the test grids, the ten primary substation feeders with the highest frequency of interruptions were examined and classified according to the above-mentioned K and T-factors. After the classification, there were two overhead line grids in rural area, three mixed grids in rural area, two underground cable grids in semi urban area and three underground cable grids in urban area. From these classified feeds, four different test grids were constructed.

• Test grid A: Overhead line grid, Rural area • Test grid B: Mixed grid, Rural area

• Test grid C: Underground cable, Semi urban area • Test grid D: Underground cable, Urban area

4.2.3 Description of the structure of the test grids

The structure for the different test grids have been decided from observations of the different feeds in Trimble NIS.

The two overhead line grids in rural area follows the same structure. The main feed in both of the grids contains overhead line, except a short section closest to the primary substation, which is underground cable. Along the main feed there are branches which in turn has branches and substations connected to it. In both grids, the structure is looped with several manually operated disconnectors connected to possible back-up feeds at different branches. The breaker in the primary substation are remotely operated. The structure of test grid A are presented in Figure 7 and Appendix 2.

The three mixed grids in rural area have all different structures. One of the three grids have a looped structure with branches along the main feed, the grid has three manually operated disconnectors connected to the branches these disconnectors are used for back-up feeds. The main feed contains of overhead line, except a short section closest to the primary substation. The next grid has almost the same structure except for one difference. The main feed goes as a loop, and on this loop, there are branches and substations connected to it. The last grid has the same structure as the previous grid, but in this case, the main feed is divided into two loops. The structure of test grid B is presented in Figure 8 and Appendix 3.

For the grids that were classified as underground cable in semi-urban area, two of them were very similar when it comes to size and structure, the last grid is much larger when it comes to

(32)

the length of cable and the number of substations. The first two mentioned grids are more similar and the structure from these two will decide the structure of the test grid. The sectional points in this grid consists of manually operated disconnectors and also a remote-controlled breaker in the primary substation. The structure of test grid C is presented in Figure 9 and Appendix 4.

The two grids that were classified as an underground cable grid in urban areas, had differences in the structure. In one of the grids, the main supply is split into two sections. The other grid has only one main feed. The structure for the test grid in this degree project will consist of one main feed. The primary substation are fitted with remote controlled breakers and the other substations are fitted with manually operated disconnectors. The structure of test grid D is presented in Figure 10 and Appendix 5.

All the test grids are considered as looped structured grids, but they are operated as radial grids.

4.2.4 Specifications of the different test grids

The specifications for the test grids are presented below. The data for the test grids are mean values from the examined feeds in the same classifications. Technical data as well as line and cable composition and customers in those areas are presented. The technical data for all the test grids are presented in Table 7 below.

Table 7: Technical data for the different test grids.

Test grid A Test grid B Test grid C Test grid D Length of line & cable [m] 52 722 49 067 13 108 5 077

K-factor 0,25 4,4 >>10 >>10

T-factor 6,3 6,6 15,5 91,2

Structure of the grid Looped Looped Looped Looped

Possible back-up feeds 3 2 4 4

Customers 335 325 204 464

Electricity consumption [MWh/year] 4 919,2 4 774,9 7 872,5 3 656,1

Substations 53 46 16 5

The line and cable composition for all the test grids are presented in Table 8 below.

Table 8: Line and Cable composition of the different test grids.

Test grid A Test grid B Test grid C Test grid D

Overhead line 80 % 50 % 0 % 10 %

Underground cable 20 % 50 % 100 % 90 %

The schematics over the different test grids are presented in Figure 7 to 10. Test grid C and test grid D are more detailed because those two test grids were used for the reliability analysis.

(33)

Figure 7: Schematic over Test grid A.

(34)

Figure 9: Schematic over Test grid C.

(35)

The number of customers of each category are presented in Table 9 below.

Table 9: Number of customers of each category on each test grid.

Customer category

Number of customers

Test grid A Test grid B Test grid C Test grid D

Agriculture 16 10 1 0

Industry 7 6 27 3

Businesses & services 17 17 65 30

Public sector 13 10 23 14

Household 282 282 88 417

Total number of customers 335 325 204 464

The annual electricity consumption for the different customer categories in each test grid are presented in Table 10.

Table 10: Annual electricity consumption per customer category and test grid.

Customer category

Annual electricity consumption [kWh]

Test grid A Test grid B Test grid C Test grid D

Agriculture 3 566 823 3 402 207 83 972 0

Industry 244 993 206 701 3 236 793 47 043

Businesses & services 561 395 405 285 3 160 606 1 275 671

Public sector 304 095 556 980 677 525 270 863

(36)

4.3 Analytical reliability analysis

Test grids C and D will be used for the analytical reliability analysis. Test grid C and D has a higher customer density than test grid A and B, this is the reason why they were chosen. Analytical reliability analysis method is applicable when there is a high frequency of faults, which makes it preferable on reliability calculations for the electrical grid (Wallnerström, 2011).

4.3.1 Failure rate

The number of interruptions on overhead line and underground cable between the years 2009-2016 is taken from the DARWin-file. The length of the cables and overhead lines within the MEE grid is also taken from the DARWin file. The data for calculation of the failure rate for both the overhead line and underground cable are presented in Table 11 below.

Table 11: Length & number of interruptions on underground cables & overhead lines.

Year

Length [km] Number of interruptions Cables OH-lines Cables OH-line

2009 1 358,1 466,5 39 118 2010 1 554,6 432,7 43 104 2011 1 565,8 418,0 48 114 2012 1 585,2 415,8 31 93 2013 1 616,2 398,7 44 84 2014 1 635,8 375,0 32 88 2015 1 720,2 347,4 33 67 2016 1 742,1 333,8 43 70

The average failure rate for underground cable and overhead line during the examined year are calculated with equation 4.3 & 4.4. According to the data in Table 11 above, F is the number of interruptions on each component for one year, and L is the total length for the same year. 𝜆𝐶𝑎𝑏𝑙𝑒= ∑ (𝐹𝑌,𝐶𝑎𝑏𝑙𝑒 𝐿𝑌,𝐶𝑎𝑏𝑙𝑒) 𝑌 ∑ 𝑌𝑒𝑎𝑟𝑠 [ 𝐹𝑎𝑖𝑙𝑢𝑟𝑒𝑠 𝑘𝑚,𝑌𝑒𝑎𝑟] Equation 4.3 𝜆𝑂𝐻 𝐿𝑖𝑛𝑒= ∑ (𝐹𝑌,𝑂𝐻 𝐿𝑖𝑛𝑒 𝐿𝑌,𝑂𝐻 𝐿𝑖𝑛𝑒) 𝑌 ∑ 𝑌𝑒𝑎𝑟𝑠 [ 𝐹𝑎𝑖𝑙𝑢𝑟𝑒𝑠 𝑘𝑚,𝑌𝑒𝑎𝑟] Equation 4.4

The calculated failure rates are presented in Table 12.

Table 12: Failure rates for overhead lines and underground cables.

Overhead line Underground cable

Failure rate 0,2297 0,0247

(37)

4.3.2 Unavailability

The average repair time, 𝑟𝑖, for overhead line is according to Allan, et al, (1991), 5 hours and

for underground cable it is 30 hours. The average unavailability, 𝑈𝑖, in case of a failure on either

of these components is calculated using Equation 4.5. 𝑈𝑖 = 𝜆𝑖∗ 𝑟𝑖[

𝐻𝑜𝑢𝑟𝑠

𝑦𝑒𝑎𝑟] Equation 4.5

Since this degree project only handles the interruptions on overhead line and underground cable the unavailability, 𝑈𝑖, can also be calculated with Equation 4.6, if the grid has the

possibility to be fed from a different primary substation feed. The unavailability time is then dependent on the time it takes to isolate the faulted line and section the grid.

𝑈𝑖 = 𝜆𝑖∗ 𝑆𝑒𝑐𝑡𝑖𝑜𝑛𝑖𝑛𝑔 𝑡𝑖𝑚𝑒 [ 𝐻𝑜𝑢𝑟𝑠

𝑦𝑒𝑎𝑟] Equation 4.6

4.3.3 Reliability indicators

Reliability indicators are used for comparing the effect of interruptions between different grids or grid sections. These reliability indicators are either weighted on number of customers affected or the electric power affected. (Wallnerström, 2011)

The reliability indicators used in this degree project are presented in Equation 4.7 – 4.12 below, where 𝑁𝑖 is the number of affected customers in each substation, 𝑈𝑖 is the unavailability

calculated from Equation 4.6.

System Average Interruption Frequency Index 𝑆𝐴𝐼𝐹𝐼 =∑ 𝑁𝑖 𝑖𝜆𝑖

∑ 𝑁𝑖 𝑖 [

𝐼𝑛𝑡𝑒𝑟𝑟𝑢𝑝𝑡𝑖𝑜𝑛𝑠

𝐶𝑢𝑠𝑡𝑜𝑚𝑒𝑟,𝑌𝑒𝑎𝑟] Equation 4.7

System Average Interruption Duration Index 𝑆𝐴𝐼𝐷𝐼 =∑ 𝑈𝑖 𝑖𝑁𝑖

∑ 𝑁𝑖 𝑖 [

𝐻𝑜𝑢𝑟𝑠

𝐶𝑢𝑡𝑜𝑚𝑒𝑟,𝑌𝑒𝑎𝑟] Equation 4.8

Customer Average Interruption Duration Index 𝐶𝐴𝐼𝐷𝐼 =𝑆𝐴𝐼𝐷𝐼

𝑆𝐴𝐼𝐹𝐼[

𝐻𝑜𝑢𝑟𝑠

𝐼𝑛𝑡𝑒𝑟𝑟𝑢𝑝𝑡𝑖𝑜𝑛] Equation 4.9

Average Service Availability Index 𝐴𝑆𝐴𝐼 =8760∗∑ 𝑁𝑖 𝑖−∑ 𝑁𝑖 𝑖∗𝑈𝑖

8760∗∑ 𝑁𝑖 𝑖 [𝑃𝑟𝑜𝑏𝑎𝑏𝑖𝑙𝑖𝑡𝑦] Equation 4.10 Average Service Unavailability Index

𝐴𝑆𝑈𝐼 = 1 − 𝐴𝑆𝐴𝐼 [𝑃𝑟𝑜𝑏𝑎𝑏𝑖𝑙𝑖𝑡𝑦] Equation 4.11

(38)

𝐸𝑁𝑆 =∑ 𝐴𝑛𝑛𝑢𝑎𝑙 𝑒𝑛𝑒𝑟𝑔𝑦𝑖 𝑖∗𝑈𝑖

8760 [ 𝑘𝑊ℎ

𝑦𝑒𝑎𝑟] Equation 4.12

4.3.4 Base scenario

In the base scenario, there is only one breaker in the grid. This breaker is installed at the primary substation that feeds the grid. All the substations have manually operated disconnectors. In event of a failure in the grid, the breaker in the primary substation will break, and all the customers in the grid will have a power outage initially. The SCADA system will alert the grid operator that a fault has occurred. The fitter is called to the grid section and decides, based on information from SCADA and experience, where to start the sectioning. The fitter then manually operates the disconnectors until the fault location is isolated and the customers have their power back on. The sectioning time for an interruption is set to 2 hours according to section 3.1.

4.3.5 Different sectioning times

The sectioning time at MEE is dependent on the experience of the operators and the fitters. Therefore, the sectioning time is varied between 50 and 150 % of the original sectioning time. This is done to see how much the sectioning time affects the cost and the reliability indicators. The sectioning time for the different scenarios are presented in Table 13 below.

Table 13: Sectioning times for the different scenarios.

Scenario 1 (Base) 100% 2 hours

Scenario 2 50% 1 hour

Scenario 3 75% 1,5 hours Scenario 4 125% 2,5 hours Scenario 5 150% 3 hours

4.3.6 Design suggestion for Test grid C

There are several techniques and technologies to minimize the number of customers affected by an interruption, SAIFI, as mentioned before removing overhead line and replacing it with underground cable is one way of decreasing SAIFI. One other technology is to replace the disconnectors in some of the substations with breakers that will break automatically in case of an interruption. In this design suggestion, the substation LP3 in Test grid C is fitted with breakers on the outgoing feeds. The purpose of the suggested design is to increase the overall reliability of the test grid. The reason for the chosen design is because the grid splits into two sections at LP3, and therefore, the potential for a reliability increase is assumed to be the biggest at this point. In Figure 11, the schematics for this design are presented.

(39)

Figure 11: Schematic over the design suggestion for Test grid C.

A failure on line L1 – L3, will still cause a power outage on all the connected substations LP1 – LP16. However, a failure on line section L4 – L10, or the “west” part of the test grid, will only affect substation LP4 – LP10, because of the breaker in LP3. In the same way, if a failure occurs on line L11 – L16, or the “east” part of the test grid, substation LP11 – LP16 will experience an interruption.

4.3.7 Calculations

The first step of the calculations were to distribute the cables and lines between the substations. This was done by randomization, with the original feeds as background for reasonable distribution. The customers were distributed as demonstrated in Table 9. Since the test grids is operated radially but with several possible back-up feeds, the failure rate for the whole grid is the sum of each cable and line failure rate. This includes the unavailability as well. To calculate the specific failure rate for one underground cable or overhead line section, the length of that section is multiplied with the average failure rate for the cables or overhead lines. The unavailability is calculated by multiplying the specific failure rate with the sectioning time. An example is shown in Table 14 for substation n in test grid D. Were L1-L5 is cables except L2 which is overhead line.

Table 14: Example of calculations.

Substation n Test grid D

λ [F/Y] Sectioning time[h] U [h]

L1 0,032 2,0 0,064

(40)

L4 0,025 2,0 0,049

L5 0,028 2,0 0,057

Σ 0,242 0,483

From the above presented example, the total failure rate, λ, and unavailability, U, is now known. The reliability indicators were then calculated according to Equation 4.7 – 4.12. The cost for an interruption in the test grids was calculated according to Equation 3.1 and the annual average power, P, mentioned in Equation 3.1 was calculated from the annual average electricity from Table 10. The duration, t, in Equation 3.1 is the sectioning time.

(41)

5

RESULTS

The results from this degree project are presented in the sections below. In section 5.1 different histograms of the examined interruptions are presented. Sections 5.2 to 5.5 presents the results from the test grids and calculations.

5.1 Histograms of the interruptions

Free overhead line, uninsulated, and underground cable are the two components that are affected by the most numbers of interruptions. Free overhead line is undoubtedly the most dominating part when it comes to the number of interruptions. The other parts have fairly the same amount of interruptions, except for pole station and distribution station. The interruptions are presented in Figure 12.

Figure 12: Interruptions on components at the voltage level of 10 kV during the years 2009-2016.

Fabrications or material failure is the most common cause for an interruption on overhead line and underground cable. For free overhead line, uninsulated, falling trees due to wind, and thunder are the most common cause for interruptions. Fabrication or material failure and digging causes the most interruptions on underground cable. The causes for an interruption are presented in Figure 13.

0 100 200 300 400 500 600 700 800 N u m b er o f in terru p tio n s

Interruptions on components at the voltage level of 10 kV

during the years 2009-2016

(42)

Figure 13: Interruption causes on overhead line and underground cable 10 kV 2009-2016.

The most interruptions on both overhead line and underground cables occurs on Mondays. Other days that have almost the same amount of interruptions are Wednesday and Friday. Fewest interruptions occurs on Tuesday and on the weekends. The interruptions during the week are presented in Figure 14.

Figure 14: Number of interruptions per day at the voltage level of 10 kV and on overhead line and underground cable during the years 2009-2016.

0 20 40 60 80 100 120 140 160 180 200 N u m b er o f in terru p tio n s

The cause of the interruptions on the studied components

Overhead line & Underground cable Underground cable Overhead line uninsulated Overhead line isolated

0 20 40 60 80 100 120 140 160 180 200

Monday Tuesday Wednesday Thursday Friday Saturday Sunday

N u m b er o f in terru p tio n s

Figure

Figure 2: Loop structure.
Figure 3: Radial structure.
Table 1: The affected component causing the interruption.
Table 2: The cause of the interruption.
+7

References

Related documents

46 Konkreta exempel skulle kunna vara främjandeinsatser för affärsänglar/affärsängelnätverk, skapa arenor där aktörer från utbuds- och efterfrågesidan kan mötas eller

where r i,t − r f ,t is the excess return of the each firm’s stock return over the risk-free inter- est rate, ( r m,t − r f ,t ) is the excess return of the market portfolio, SMB i,t

Parallellmarknader innebär dock inte en drivkraft för en grön omställning Ökad andel direktförsäljning räddar många lokala producenter och kan tyckas utgöra en drivkraft

Närmare 90 procent av de statliga medlen (intäkter och utgifter) för näringslivets klimatomställning går till generella styrmedel, det vill säga styrmedel som påverkar

I dag uppgår denna del av befolkningen till knappt 4 200 personer och år 2030 beräknas det finnas drygt 4 800 personer i Gällivare kommun som är 65 år eller äldre i

Den förbättrade tillgängligheten berör framför allt boende i områden med en mycket hög eller hög tillgänglighet till tätorter, men även antalet personer med längre än

På många små orter i gles- och landsbygder, där varken några nya apotek eller försälj- ningsställen för receptfria läkemedel har tillkommit, är nätet av

Det har inte varit möjligt att skapa en tydlig överblick över hur FoI-verksamheten på Energimyndigheten bidrar till målet, det vill säga hur målen påverkar resursprioriteringar