Impact of remote controlled switches on distribution grid recovering processes

(1)

Impact of remote controlled switches on

distribution grid recovering processes

KTH-School of Electrical Engineering

January 2010

(2)

(3)

I would like to thank my supervisor Jean MARTINON for his valuable teaching skills, for his guidance and for his consideration at every level of this work. I am also very grateful for the warm welcome as well as the support of Jean Louis LAPEYRE and of all the members of his department.

At last, I am thankful to Lennart SÖDER and Fredrik EDSTRÖM who accepted to supervise my thesis at KTH.

(4)

Abstract

(5)

I. Introduction

I.1. The French electricity market and ERDF

The French electricity market deregulation process:

In 2006 the French parliament transposed the 2003 European directive claiming that any new electricity supplier must get a free and non-discriminatory access to the grid. Thus in 2007 July 1st, the French electricity market was to be fully deregulated. This implied a complete transformation of the French market main actor: EDF (Electricité de France). Created in 1946 as a state monopoly, its first role was to achieve the electrification of France and to develop and maintain the electricity supply for every consumer from the production to the transmission and the distribution.

In a deregulated context EDF has then operated three major transformations:

1. 2004 and 2005: the group becomes a joint stock corporation and opens its capital

2. 2005 September 1st: RTE (Réseau de Transport d’Electricité), the Transmission System Operator, becomes a subsidiary company 100% owned by the group with the required independency

3. 2008 January 1st: ERDF (Electricité Réseau Distribution France) is founded as a subsidiary company 100% owned by the group

The French market is now structured as a classical deregulated market…

F

Figure 1: French electricity market organization

(7)

…With a huge predominant quasi-state-owned actor:

Figure 2: French EDF electricity production repartition (left) and French EDF production capacity (right)[1]

(8)

The French distribution grid and ERDF

In charge of 95% of the French distribution grid, ERDF (Electricité Réseau Distribution France) is a “Joint Stock Corporation with a Board of Directors and a Supervisory Board”. With 1.2 million km of high and low voltage lines and more than 32 millions of low voltage customers, ERDF develops and maintains the largest distribution grid in Europe.

The distribution grid tends to be homogeneous all over the country and is constituted by:

2,200 primary substations connected to the transmission grid (mainly 63, 90, 225 and 400kV)

MV network (from 10 to 50kV but mainly 20kV)

• 596,200 km, overhead: 61 %, underground: 39 %

• 100,000 connection points (subscribed power > 250 kVA) • 727,000 MV / LV substations

LV network (mainly 400V phase to phase)

• 669,300 km, underground: 36 %, overhead [bare: 17 % + twisted 47 %] • 310,000 connection points with 36 kVA < subscribed power < 250 kVA • 32,000,000 connection points with Ps < 36 kVA

(9)

Figure 4shows the limit of ERDF property which goes from primary substation (from high voltage disconnecters) to the low voltage counting systems in customers house. However this layout does not show how the distriution grid is interconnected : in a usual use it can be represented as on Figure 4 (radial use) but it is actually designed as a mesh (see III.1 Definition and utility). This point is one of the main differences between the distribution grid and the transmission grid which is always driven as a meshed system. In all the following text voltage levels will be defined as :

• HV: transmission grid voltage levels : mainly 63, 90, 225 and 400 kV (phase to phase)

• MV: high voltage part of the distribution grid : from 10 to 50kV but mainly 20kV (phase to phase)

• LV : low voltage grid : 400V phase to phase (230V phase to neutral)

Every year a large part of the French MV and LV consumption flows through this grid.

Figure 5: Annual energy flow through ERDF grid in 2008

Thus ERDF is a crucial actor in the reliability of the electrical system and has to maintain a proper and steady electricity quality.

346.4 TWh

ERDF

grid

(<50KV)

347.8 TWh

From the transmission grid (HTB > 50KV) Customers Distributed generation (wind farms, photovoltaic)

(10)

In this way public authorities impose several “public service missions” to distributors:

• Equal access to the distribution grid for suppliers • Public service and proper quality all over the country

• Development and enhancement of the public distribution grid

A state organisation, called “Commission de Régulation de l’Energie”, is in charge of the Electricity market actor control. Basically, It defines the regulated price of electricity transport in terms of how much the miantenance and the developpment of the grid costs. The official measurment of distributor performance is based on the so called B criterion (equivalent to SAIDI, see glossary and “IV.1 How reactivity is measured and controlled” for more information” ) and a sysem reawrding ERDF performance in term of an objective B criterion has been developped. This system makes of the B criterion the most important indicator of the service quality and lay more stress on an equal and non disrupted delivery service for every customer than on the financial efficiency of the grid.

I.2. Background of the project

To fulfil its missions, ERDF needs to launch large investment programs in the limit allowed by its regulated resources. Although tremendous progress has been realised for 30 years (see Figure 6), the B criterion has been regularly increasing since 2002 as the grid becomes older and investment programs were compressed (see Figure 8).

Moreover several strong weather events underlined the vulnerability of the grid as well as the increasing demand of the society. (Large storm in December 1999, heavy snow in December 2008, storms Klaus in January 2009 and Quinten in February 2009).

These two observations are obviously linked but the analysis of their causes requires to split out the problem into two axes:

• The “everyday’ driving operation of a distribution grid • The “crisis” driving operation of a distribution grid

Indeed the most reliable grid in an everyday situation (one outage after one, easy access to the lines…) could be extremely weak in case of strong weather conditions.

To be able to get these two kinds of focus ERDF commonly uses several criteria to sort unplanned outages. (These criteria are detailed further in the document in “IV.1 How reactivity is measured and controlled”)

Regarding the political impact of long outages due to strong weather events, more stress is laid on enhancing the strength of the grid regarding this aspect, with a big program aiming at burying 100 000km of overhead lines before 2025.

(11)

As a comparison the solution of burying overhead lines would have a bigger impact on the “everyday B criterion” (from around 6 incidents/100km for overhead lines to 2 incidents/100km for underground lines) but represents a huge investment (2M€ per feeder whereas a full RCS equipment costs around 100k€ per feeder). This basically shows that RCS are a cheap and efficient way to enhance every day delivery quality, by reducing each incident impact rather than reducing the incident number.

B criterion (min) 0 50 100 150 200 250 300 350 400 450 198019821984198619881990199219941996199820002002200420062008201020122014201620182020

Figure 6: Evolution of the B-criterion (min) vs time. Only everyday outages due to the distributor are accounted on this graph

B criterion (min)

Incident number +1000 MV incident per year

(around 600 undergound and 400 overhead) 4M€ increase of fixing maintenance

each year

Average of 1,5 min of B criterion increase per year

B c rit e rio n (m in ) In cid e n t n u m b e r B criterion (min) Incident number +1000 MV incident per year

each year

B criterion (min)

Incident number +1000 MV incident per year

each year

B c rit e rio n (m in ) In cid e n t n u m b e r

(12)

Figure 8: Evolution of investment volume vs. time

I.3. Project definition and document structure

To achieve the previously described challenges, this project aims to define a theoretical optimal RCS equipment, accounting for specificities of each territories. This optimum is then to be compared with the current situation in order to highlight under-equipped areas, and to estimate investment volumes. To reach this result a model of the usual recovering process had to be developed and validated to derive the impact of RCS equipment on performance.

For a better understanding of the presented work, this document starts with a detailed review of ERDF organisation and of the technologies used to develop, maintain and drive the grid. Then, a development about the link between the number of installed RCS and the possible number of grid sections gives an overview on the structure of the French distribution grid and on the present state of the RCS equipment as well as the understanding of the RCS impact on the quality. A third part discusses a way to model the B criterion on a given grid and thus to forecast the impact of a RCS installation policy. This model is then used in a fourth part over all the French territory to compare different areas and to alight those where investment should be increased. This fourth part includes a discussion of the results and the optimisation used. Finally a conclusion will explain in which way the study could be used and how it could be developed.

(13)

II.ERDF organisation and tools

II.1 Today organisation

Originally the distribution grid management of EDF was based on a mesh of 100 “Centres” all over the country. Basically, each Centre included all the components of a distribution company. After the 10 past year of transformation some crafts were regrouped to lead to today organisation:

ERDF is constituted by a national direction leading 8 regions:

My work has been realised into the “Global Performance and Region support“ headed by Jean-Louis LAPEYRE and belonging to the “Network and Assets Direction”.

Each region is responsible for operational distribution business and the level of their performance. It relies on 4 key operational managers:

• Networks and « concessions »

• Customers management

• Human resources • Accounting

In addition, local distribution working groups are aggregated in intermediary performance organisation units (24-30) for each of the key distribution businesses: electricity networks, services to customers or suppliers. There are especially 30 Regional Operation Agencies (from 3 to 5 in each Region, see Figure 10) in charge of the everyday driving of the grid. Remote controlled switches, default detection and grid surveillance is performed from these agencies. Their experience is then of a great interest for the presented studies.

Customers &

Suppliers

8 Regions

President

of Directory

Network

& Assets

Operations

& Territoires

Finance

& Strategy

HR &

Communication

s

Top Executive

Corporate Secretary

(14)

Figure 10: The 30 operation agencies of ERDF and their controlled area

This complex organisation and history directly impacts this study: on the first hand the equipment diagnostic must be based on the previous “Centre” mesh as the grid was originally built on it. On the second hand the reactivity analysis must be done on the present organisation.

However this management structure should not hide electrical connexions : each agency is in charge of a part of the big interconnected electrical system that is ERDF grid. No part of the distribtuion grid is independent from the rest of the system and Region are for example to work together in case of major climatic events.

(15)

II.2. IT System

The huge grid that ERDF has to run requires complex IT system to be built and controlled. Only the description of the system useful parts regarding the presented study are presented here.

Originally, the grid is described in details by a database called “GDO” (“Gestion des Ouvrages”) giving a picture of the grid as it is built. This database is to be replaced by another called SIG (most of the Regions have achieved the replacement which started around 2007). This database is the heart of the IT system and is to be used by all the IT tools of the different crafts that constitute ERDF: Maintenance, development, driving…however this database is still not reliably updated.

For now, the most updated IT-tool is the one used to drive the grid: as the system is still to be finalized, the only way to be sure to have reality-related data is to use the database operated by people who “practice” the grid. This system has several levels.

A first level, called SIT-R, enables the agencies to get a real-time control on their operations: it shows the structure of the grid, which MV/LV substations are disconnected, which remote control switchers can be operated, maintenance operation…

All these operations and information are then kept in servers and can be accessed through the interface “EtaReso”.

Figure 11: Relation between the driving interface and SIG

SIG

(former GDO)

(16)

III. Remote Controlled Switches (RCS)

III.1 Definition and utility

RCS are strategically placed on the lines in order to sectionize the feeders. For a better understanding the primary circuit is detailed below (single-phase circuits), from primary substation to MV/LV substations, showing different structures encountered in the French grid. All the figures and explanation of this part come from [2]

Figure 12: Single phase circuit of a primary substation and of a feeder

MV-Distribution

grid

Primary Feeders

HV-Transmission grid

Rescue RCS

(open in normal use)

(17)

Figure 12 does not represent LV/MV substation and shows only the main line of the feeder. The following figures show different grid structure that includes LV/MV substation as well as secondary lines of primary feeders. The main characterisitc is that every feeder is connected to another one. This connexion, called “Rescue RCS” is left open in normal situation but enables a feeder to supply another one if needed.

• The most used overhead structure is the one called “from source to source”: the so called “main lines” of the feeder are linked to another feeder (from another or the same substation) through a “rescue RCS”. Some secondary lines lead the power flow from the main lines to LV/MV substations.

Figure 13: Source to source structure for overhead lines

• The same type of structure is used for underground networks. The difference is that LV/MV substations are directly placed on the main lines. RCS are in this case commonly integrated into the substations (Figure 14)

LV/MV substation

Primary substation (only MV lines are

(18)

• Many other structures can be encountered, as for example the so called “doubly shunt structure” that exists in some big and dense cities (Paris, Lyon…). This structure is to be very reliable but is very expensive (two feeders for each LV/MV substations which are all equipped with automatic switchers)

LV/MV substation with

remote control LV/MV substation with _{manual switch}

Rescue line

Work line

Circuit breaker station

Figure 15: Doubly shunt structure Figure 14: Underground

(19)

Doubly shunt structure is not very common and presents completely different issues than usual lines. It will then be obliterated in the rest of the text.

There are different types of remote controlled switches depending on what they are used for or where they are installed. This text focuses on particular type of remote controlled switches: MV remote controlled switches, which are used to break an under voltage three-phase circuit. They are different from circuit breakers as they require an external action to be triggered. They are also different from disconnecters as they can clos or break an under voltage circuit. The figures from Figure 16 to Figure 18 present some of these switches commonly used on the French grid, on overhead line as well as underground ones. It should be noticed that some of them are controlled with radio wave, and others through the existing telecom network.

The so-called IAT presented on Figure 16 is older (and cheaper) than the August of Figure 17. August is more reliable as its moving parts are operated into SF6 whereas IAT mechanism is moved in air, offering a lower and more vulnerable insulation.

Radio antenna motor

IAT

Opening mechanism

Voltage transformer Radio antenna motor

IAT

Opening mechanism

Voltage transformer

(20)

Figure 17: So-called “August” RCS with opening mechanisms moved in SF6.

Figure 18: So-called PSSB (« Poste au Sol Simplifié B ») used for underground lines, it is basically a LV-MV substation on which a RC switch have been installed.

MV-LV substation, with fuses, switches, transformer and LV feeders

(21)

III.2 Link between RCS and reactivity: notion of “grid bags”

In order to fully understand the role that RCS play in the way they are considered in the study, a view on their action in case of an incident is necessary.

The type of outage considered here are unplanned permanent incidents occurring on a MV line. In the whole following text, the word “incident” designs this type of outage if nothing else is stated before. These are the most B criterion-costing incidents as they induce a large number of disconnected clients (around 1000-1500 LV-customers per feeder). Low-voltage incidents induce only one MV/LV substation outages (from 1 to 100 clients) and temporary single-phase default current are commonly handled by a so-called shunt circuit breaker (no disconnection).

The restoring process after a MV-incident comprises 2 phases: localisation and repair. The first phase can be splited into two different actions: remote controlled actions (from the agency) and manual action (teams going along the line to precisely locate and identify the incident). This leads to three phases that are described below through an example of a tree falling on an overhead secondary line (Figure 19 to Figure 22).

1. In normal use, all the MV/LV substations are connected and the rescue RCS (marked by a flag) is open. (The arrow indicates the direction of another substation, the current is flowing in the other way on the feeder on the right)

Figure 19: Feeder configuration in normal use

Contiguous feeder

RC

Manual switches

LV/MV

(22)

2.

t=0s. A tree falls on a secondary MV line: a short circuit is created and the circuit

breaker placed at the output of the primary substation is automatically opened as a fault current is detected. Every customer previously connected on the feeder is disconnected:

Figure 20: Feeder configuration right after an incident

3.

t=T

RCS

~3min. In the following minutes, the operation agency in charge of the feeder

reacts and commands the RCS in order to isolate the damaged section of the grid. This is the first step of the localisation phase.

Figure 21: Feeder configuration after the RCS step of the localisation phase

RCS Contiguous feeder MS Disconnected customers Disconnected grid Open switch

The downstream main line is fed by a contiguous feeder through

(23)

Disconnected customers

4.

t=T

loc

~100min. Fixing teams are then sent to operate the manual switches and to

locate the incident. This is the second step of the localisation phase. After this phase disconnected customers have to wait for the line to be repaired (most of the cases) or an external generator to be installed.

Figure 22: Feeder configuration during repairing phase

5. Finally, when the tree is removed and the lines replaced, switchers are operated to recover the normal state (Figure 19)

The typical incident shown here corresponds to most of the cases. It has to be noticed that RCS could be installed on other places than main lines. They would be then useful in case of simultaneous incidents (on the contiguous feeder for example, which is common under strong weather conditions). But as announced in the introduction this is not the first focus of this study.

A crucial point is raised from this description : the time needed to process each phase aligths why RCS are of a higher interest than manual switches. Indeed, installing RCS enables to reconnect customers within 3 min (TRCS) instead of 100min (Tloc) whereas manual switches

require 100 min to be actioned and their benefit is then limited compared to RCS. If the difference was not so high, the supplementary cost of RCS might be dissuasive.

(24)

Figure 23:Limit of one grid bag on a source-to-source structure

This said a crucial data is the number of customer into a grid bag. As long as the context enables it, RCS are usually placed under the so called “P*L rule”[3].

This rule requires hypothesis to be used:

• In average, there is no reason that an incident is more likely to occur up or downstream a feeder. This means that the incident occurrence probability on a grid section is proportional to the length of line contained in this section. • The impact of an incident is the power that cannot be delivered

Then RCS have to be placed such as every grid bag on a given feeder should contain the same product “P*L”, meaning the subscribed power in a grid bag times the length of line contained in a grid bag. It should be noticed that this rule follows an “Energy non-served (ENS) view” which seems to be incoherent with today context that leads to consider B-criterion as the most important performance indicator (see “I. Introduction”). Indeed the impact of an outage is here measured by the subscribed power that is disconnected whereas the B-criterion assumes that the gravity of an outage depends on the number of disconnected customers (meaning that every cutomers are equally considered regardless their subscribed power). This incoherence is historical and is not discussed here but will be again raised in “V.1 Technical-economical approach”.

(25)

Considering that this P*L rule has been historically fulfilled, it is assumed that all RCS are placed under the P*L rule.

Beside, an even stronger hypothesis is here taken: customers are equally placed along a line (same subscribed power for a given length of line everywhere on a given feeder). This means that:

RCS are placed on a given feeder such as

Every grid bag contains the same number of customers

Conversely to the P*L rule, this approximation is not an official way of thinking at ERDF even if it is commonly used when a global and average view of the grid is needed as in this study. Thus, for a feeder with Ncdep customers on which Ngb grid bags can be isolated, every grid bag

contains Ncdep/ Ngb customers.

Figure 24: Two grid bags feeder outline

For instance, if the feeder of Figure 24 feeds 800 customers, both grid bags contain 400 customers: wherever an incident occurs, 400 customers can be reconnected within a few minutes thanks to RCS.

The consequence of this developpment is that the grid bag number per feeder is the most important data to evaluate the equipment state as it directly provide the number of customers that can be reconnected within TRCS.

III.3 Present equipment state evaluation

Regarding the simple feeder drawn on Figure 24, deriving the grid bag number appears to be simple: for a feeder equipped with one rescue point and NRCS RCS on its main line (NRCS does

not includes the rescue RCS here), Ngb can be derived:

Ngb = NRCS +1

However, regarding actual grid structures, this expression must be appended. A detailed examination of existing grid raises two main factors that influence Ngb:

(26)

Feeder with several rescue points:

The notion of “main line” and “secondary line” becomes inappropriate in this case. Considering the simple example of Figure 25 the definition of one main line seems to be arbitrary.

Figure 25: Example of a feeder where two lines could be defined as "main"

For the focus of the study the notion that matters is the “rescuability” of a line: a RCS installed on a line that does not have any rescue point downstream is useless during most of the incidents:

Figure 26: Example of a RCS on a non-looped line

For any incident that would not occur downstream such a RCS, it would not be of any help. In any case the customers that are fed by the downstream grid can then never be reconnected thanks to remote controlled action.

These two lines could be defined as « main line » of

the feeder

MV/LV substation

Rescue RCS

This RCS can be used only for incidents occurring downstream itself : the downstream grid can in this case be isolated and the rest of

(27)

As a conclusion, any RCS installed on a derivation line that does not have any rescue point will be said to not creating any grid bag(see Figure 26). They will be thus considered of no interest for our study. Any line equipped with a rescue point will be called a “looped line” in the following text.

Multiple direction notes :

On Figure 25 three grid bags can be isolated on the feeder although only one node is equipped with a RCS. The RCS number then not only influences the number of grid bags. The type of node where they are installed as well as the number of “remote controlled directions” is also of a great importance.

Figure 27: Illustration of multiple direction node impact on the number of bags

In a general case (even if more than three directions with RCS is not common), for a node with n directions with looped lines, any direction equipped with a supplementary RCS creates one supplementary grid bag excepting for the last direction. In the example on Figure 27, the third RCS is indeed useless regarding the number of grid bags. Such equipment however exists as a complete remote control offers an easier driving.

Further investigations (thanks to a Python code reading GDO database) show that accounting for these two factors modifies the number of grid bags per feeder up to 0,5 on the Centre mesh point of view.

One three-directions nod with RCS with one RC direction : 2 grid bags

(28)

Regarding grid structure, and especially for underground lines, a last point must be investigated: how to deal with remote controlled switches installed into MV/LV substations. The question of the difference between the two structures shown on

Figure 28 is indeed raised.

Figure 28: Feeder with a MV/V substation equipped with RCS on a main line

Obviously two bags are formed in both cases. However on the first one there are customers connected in between the two RCS of one node. In the following calculation and results, these two cases will be equally considered. Actually the main difference is that when both directions of such a MV/LV substations are remote controlled, the customers that are served by this substation can be reconnected thanks to RCS wherever the incident occurs (used for priority structure as hospitals)

Other points have been investigated but their impact has been proved to be insignificant. (e.g. lines that are looped on their own feeder…)

These comments have now to be transposed to manage to count on a given grid how many grid bags are created by RCS. On a first step the GDO database from 2006 has been used to know the rate of RCS equipment on each Centre. GDO is indeed known to be reliable and 2006 is the last year where IT systems were homogeneous all over the country. This means that data from GDO is the latest and the easiest way to get a view of the grid at a national point of view. Of course the calculation made on GDO data are to be fitted up to the present IT system, especially to EtaReso data.

(29)

Figure 29: Microsoft Access table extracted from GDO database. The table describing the national grid contains 2,370,000 rows

It presents all the information needed to get a proper evaluation of the number of grid bags:

• Type of segment (looped line or not)

• Nature of nodes (some components and switches inside substation are described and have to be removed)

• Unique reference of nodes (in order to know how many directions a node have) • Action type (if the node is a switch it has to be known whether it is RC or manual) • Type of RCS (Rescue or usual: can be known in GDO thanks to the information

opened or closed in normal use)

• Feeder and Centre carrying the segment

Although all these data are provided, a method to count the number of bags per feeder on a given area described by a GDO table has to be used. The expression Eq.1 seems to be right for a random grid(see Figure 30). It has not been mathematically proved but appeared to be the most reliable one for any grid structure that have been encountered during the study (as soon as it has a source-to-source meshed structure).

N

gb

= N

RCdir

- N

RCnodes

+ N

feeders

- N

RescNodes

Where:

• Ngb is the number of grid bags on the focused area

• NRCdir is the number of remote controlled direction on the focused areas (including

rescue nodes)

• NRCnodes is the number of nodes equipped with remote controlled (including Rescue

nodes)

• Nfeeders is the number of feeders on the focused area

• NRescNodes is the number of Rescue nodes in the focused areas

The only thing to watch out in this expression is the case of a node where all the directions are not remote controlled and where the number of grid bags created must not be underestimated (case of a 3 direction nodes, all looped, with only 2 remote controlled directions: actually 3 grid bags whereas the above expression would give only 2. see Figure 27 ). This have been accounted during the algorithm conception.

The results for the Mediterranean Region are given on Table 1.

(30)

Figure 30: Grid bag number expression illustration

Nb total Average number per feeder Marseille 224 7,7 2 108 652 2,9 153 433 Provence 440 15,7 1 282 1 664 3,8 287 830 Alpes du Sud 198 26,6 992 679 3,4 164 513

Avignon Grand Delta 342 18,8 1 098 982 2,9 191 565

Aix 1 204 16,9 1 336 3 977 3,3 795 2 341

Var 385 18,8 1 813 1 496 3,9 281 981

Nice Alpes d'azur 255 11,5 1 935 796 3,1 158 431

Cannes 130 13,6 2 381 482 3,7 116 254 Toulon 770 15,5 1 949 2 774 3,6 555 1 666 Vallées d'Aude 168 33,7 1 515 900 5,4 181 556 Pyrénées Roussillon 171 20,8 1 945 637 3,7 133 409 Montpellier Hérault 389 16,2 1 605 1 522 3,9 365 730 Gard Cévennes 268 24,1 1 442 870 3,2 196 508 Nîmes 996 22,1 1 604 3 929 3,9 875 2 203 2 970 18,3 1 585 10 680 3,6 2 225 6 210

Centre

Rescue points number RCS equipped nodes number Méditéranée Feeder number Average feeder length Average customer number per feeder Grid Bags

Operation

Agency

Aix Toulon Nîmes (km)

Table 1: Mediterranean Region data

Several results are raised : the obvious feeder length variation shows that Centres are not homogeneous from terrain and density point of view. Whereas Marseille presents a mean length of 7.7 km (urban, high density) the rural Centre of “Vallée d’Aude” has feeders more than 4 times as much long. Besides the rescue point number per feeder varies a lot from a Centre to another. Although some Centres present as many rescue points as feeders (meaning two rescue points per feeder), others have a rescue point number just over half the feeder number (meaning that most of the feeders have only one rescue possibility). These two aspects, added with a historical different management yield differences between Centres.

3rd_{case : 16 RC directions, 7 RC-equipped}

nodes, 2 feeders, 2 RC rescue points

14-7+2-2=7 GB

2nd_{case : 16 RC directions, 7 RC-equipped}

14-7+3-2=8 GB

1st_{case : 16 RC directions, 8 RC-equipped}

16-8+3-3=8 Grid Bags

14-7+2-2=7 GB

14-7+3-2=8 GB

16-8+3-3=8 Grid Bags

(31)

In this part Reactivity is often associated to B-criterion which is the key data of the sutdy. However reactivity concerns the impact of each incident, meaning the outage time per customer for a random incident whereas B criterion expresses in addition the number of incident and the reliability of the grid. This is not a problem as long as the grid reliability is not discussed here but this difference should not be forgotten.

IV.1 How reactivity is measured and controlled

When an incident occurs, it is solved by the concerned operation agency. Agents can see where a fault current is detected, and reacts with remote controlled action or by alerting fixing teams (located in exploitation agencies). All these actions and indications are memorised through SIT-R. The information of which MV-LV substations are disconnected at any time, as well as the list of actions operated on the lines are then known. Therefore, one can for example access curves showing the number of customers disconnected vs. time on any period and any mesh (feeder, primary substation, all the area covered by the agency).

Fixing phase

Localisation phase

T

_loc

T

_RCS

Fixing phase

Localisation phase

T

_loc

T

_RCS

Figure 31: Disconnected customer number vs. Time. Graph extracted from EtaReso for one incident (underground line) in Aix-en-Provence agency

(32)

As announced in introduction, there are two analysis levels for these data: the everyday situation and the crisis one (when a usual driving of the system is impossible).

For each incident sent by SIT-R, ERDF has defined several criteria to classify it in terms of these two levels.

• Some incidents are stated to be “exceptional”: if an incident stems from exceptional events but which are no related to ERDF action(such as war, terrorist attack…), from external action that cannot be managed (fire, aircraft crash…), from outages due to unavailable production capacity (above a certain limit settled in the TSO contract), from outages imposed by public authorities independent from any action of the distributor or for big climatic events it is said to be exceptional. This last point is the most important regarding the number of concerned incident as well as for the study. “Big climatic events” are defined at the Centre mesh; if the weather leads to more than 100 000 LV-customers disconnected inside one Centre and if the climatic event has a yearly occurrence probability under 5% this event will be qualified as a “big climatic event” and all the related disconnections will be classified as exceptional.

• The notion of exceptional incident appears to be not sufficient to get an actual “everyday situation” view. Therefore another selection criterion called “filtering” has been defined. For each Region a threshold daily number of events has been settled. For every day during which the number of events overtakes this threshold the daily B criterion will be replaced by the mean daily B criterion of the year. For example, if 5 days are filtered in a given Region, the B criterion of these days will be substituted by the average daily B criterion on the 360 other days. The so-called “filtered B criterion” is then:

∑

⋅

=

days filtered non d nf

B

N

B

Where :

• N is the total number of days in the year

• Nnf is the number of non-filtered days of the year • Bd is the daily B criterion

The threshold of each Region is based on a historical observation and is to eliminate every day that presents an incident number larger than the 98% percent days of a mean year. Around 7 days a year are then usually filtered by this method. France is split into 8 Regions, and the threshold varies between 21 (Paris area) and 53 (South East).

Another important notion of a big impact is the difference between so-called “short outages” and “long outages”. Every outage that is solved in a shorter time than 3 minutes is said to be short and is not accounted. The B criterion and the calculations of this document are then only related to “long outages”.

If no other specifications are given, all incidents and B values used in the following text will concern filtered and non-exceptional incidents.

(33)

IV.2 B criterion modeling

Investment programs require reactivity to be modeled, so that the impact of each material can be planned. This part is to define a B-criterion model, to show that the model is correct for actual data and that it can therefore be used.

The three-phases model leads to the following recovering curve modeling:

Disconnected LV-customers number

Time Fault current occurence End of remote

controlled action Customer number per

feeder

End of localisation End of recovering procass

Customer number between two manual switches Customer number per grid

bags

T

loc

_T

_fix

T

RCS B loc B fix B RCS Ratio : 1 / Ngb

Figure 32: Recovering curves model

This area shows the gain brought

(34)

The comparison with a typical real curve (Figure 31) seems to corroborate this model. The value of each phase-related B criterion can be derived (Eq. 3 to 6)

fix loc RCS f e inc fix fix f inc gb RCS loc loc f inc RCS RCS

B

N

T

B

N

T

B

N

T

B

+

=

−

=

.

)

(

.

Where:

• BRCS is the value of the B criterion related to the remote controlled action phase • Bloc is the value of the B criterion related to the manual action phase (2nd step of the

localisation phase)

• Bfix is the value of the B criterion related to the fixing phase

• TRCS is the average time between the occurrence of the incident (when the feeder

circuit breaker is triggered) and the end of the remote controlled actions (the last remote controlled action needed to isolate the grid bag carrying the fault)

• Tloc is the localisation time, that is to say the time between the occurrence of the

incident and the last manual action needed to isolate the fault

• Tfix is the time needed to repair the damages if needed. It is the time required

between the localisation of the fault and the total recovering of the grid. • Ninc is the annual number of incidents on the observation area

• Nf is the feeder number on the observation area • Ngb is the grid bag number per feeder

• Ne is the number of customers between two manual switches.

Bfix is not included in the focus of this study. It should however be noticed that the grid bag

analysis could be transposed for manual switches.

These values of B are directly proportional to the highlighted areas of Figure 32. For one incident, these areas are indeed equal to the produuct “disconnected customer number times the outage time”. This value divided by the customer number included in the observation area gives the incident-related B criterion. This multiplied by the annual number of incident yields the total B-criterion over a year.

This model has been validated by corroborating the value of the actual filtered Bloc , and the value obtained from the model with actual average values of Ngb, Tloc , Tomt, Nf and Ninc.

Ngb is known thanks to the algorithm presented in “III.3 Present equipment state evaluation”

Nf is known.

Eq. 6

(35)

The actual value of Bloc+BRCS as well as Tloc has been known for several years on the Centre mesh. Besides, the value of TRCS has only been memorised since June 2008 thanks to

the development of EtaReso database. The period used to validate the model is then from May 2008 to 31st august 2009. However this period is too short regarding the evolution of Ninc. The reliability rate of the distribution grid is indeed around 1 to 10

incidents/100km/year, which leads to an annual incident rate per feeder around 1 (average of 25 km of line per feeder). Ninc value has then been derived from data over 5 years (with a

selection fitting with the specification given above on the incident type)

The numerical values of Mediterranean area are given on Table 1 and Table 2.

Marseille 0,627 2,8 53,1 12 13 11%

Provence 1,150 2,8 70,9 25 25 -1%

Alpes du Sud 1,303 2,8 99,0 36 42 16%

Avignon Grand Delta 1,304 2,8 70,7 27 36 32%

Aix 1,122 2,8 70,7 23 27 17%

Var 1,685 2,9 61,3 31 31 3%

Nice Alpes d'azur 0,908 2,9 61,5 16 21 30%

Cannes 1,282 2,9 68,3 23 27 17% Toulon 1,359 2,9 62,7 25 28 11% Vallées d'Aude 1,381 3,5 73,1 23 24 5% Pyrénées Roussillon 0,746 3,5 72,5 18 17 -4% Montpellier Hérault 0,860 3,5 54,3 11 15 38% Gard Cévennes 1,036 3,5 70,8 24 26 11% Nîmes 0,976 3,5 63,6 16 20 20% 1,134 3,1 66,0 21 25 17% Nîmes Méditéranée Annual incident number per feeder TRCS May 2008-31 August 2009 (min)

Operation

Agency

Centre

Aix Toulon Tloc-TRCS May 2008-31 August 2009 (min) Actual Bloc May 2008-31 August 2009 (min) theorical Bloc May 2008-31 August 2009 (min) Relative deviation

Table 2: Reactivity periods data and reliability of the model for Mediterranean Region

(36)

V. RCS number optimisation

V.1 Technical-economical approach

The main purpose of the study is to alight former Centre where RCS equipment should be enhanced, and to be able to valid Region investment programs. In this way the three-phases reactivity model has been used (as it has been shown to be valid in the previous part) such as the number of grid bags varies whereas actual other data are kept to account for differences between each area (for instance the terrain can have an influence on Tloc; Ninc

varies with the age and the type of grid…)

An approach historically used at EDF to priorise and control investments is the so-called “technical-economical” approach. It is based on an optimised balance between the cost of a policy and the energy that could be distributed if investments are made. In other words, the specificity of this method is to state that the benefit of an investment can be measured by the “energy non-served” that could be spared. In this way a “social” cost of an outage has been historically settled to 9.2 €/kWh of energy non-served (ENS)[4]. It originally stems from the total cost for EDF of one kWh that is not distributed when every step of the electricity market was integrated in one company (cost of supplementary energy sources, non sold energy,…) . This value seems huge compared to usual electricity market prices (usually around 100€/MWh) and this approach can be discussed. Even a comparison with the ratio GDP/Energy consumption, which is more significant, shows that 9.2 €/kwh is a high estimation : in 2008 French GDP was of 6 058 billions$ whereas the total Energy consumption has been estimated around 6058PJ . With a change of 0,7€/$ this leads to only 1.2€ of GDP per consumed kWh.

On the other hand this high value reflects the high demand and impact of outage for authorities and from users of ERDF public service point of view. Anyway, this approach is used in a first step to prioritise different territories and therefore satisfies the requirements.

A second statement classically made in this approach is that an investment should foster profit within ten years with a discount rate of 8% (meaning that the gain made one year after the investment should be 8% above the investment for the operation to be balanced)

As in “III.2 Link between RCS and reactivity: notion of “grid bags””, it might appear incihenrent to use ENS to evaluate the gain of a policy when the context lays more stress on the B-criterion. But the two views are here the same as the hypothesis of an equal consummed power for each customer is taken (Pc=1.1kW, [6]). The B-criterion is then directly proportional to ENS. Beside, on a more general situation, this ENS approach tends to be only used to priorise investments and not really to plan their volumes

(37)

Energy non-served gain related to RCS policy

The effect of a RCS policy is to create grid bags and to split the localisation phase into 2 phases. The B criterion that is spared thanks to RCS can be derived from the 3 phases model, from the area that is indicated on Figure 32:

f inc gb RCS loc spared

N

)

T

(T

B

.

1

1 .













−

=

The annual energy non-served related to this B criterion is:

c c spared spared

B

N

P

ENS

=

.

Where:

• Nc is the total customer number of the area

• Pc is the mean consumed power per customer which is of 1.1kW

This leads to a gain (spared cost would be more appropriate) related to a RCS policy of:

ENS c c f inc gb RCS loc RCS

N

P

p

N

)

T

(T

GAIN

.

1

1 .

.

























−

=

This model can be used on one feeder as well as on one Centre. However the observation area must be relevant: the number of grid bags per feeder should not be too spread around the mean value (the result would then hide some over or under equipped area and the approximation that takes the inverse of the average grid bag number per feeder instead of the average of 1/Ngb would be totally wrong).

A parameter that has not been raised for now is the reliability of RCS. This information is to be very soon officially collected in ERDF but data are not really reliable yet. However including a reliability rate in this model could be of a big interest to evaluate the impact of a maintenance policy. A track to be developed is proposed here, and accounts for the increasing number of remote controlled action with the increasing number of grid bags, which thus yields an increasing probability of failure:

(38)

Where τ_RCSf _{is the failure rate of RCS which will be settled to 5% in calculations.}

Assuming that no parameters change with time, reporting this gain over 10 years with a discount rate τ of 8% will lead to append a factor of [5]:

(

)

7 .

71

1

11 10 0

=













+

−













+

−

=

+

=

∑

=

τ

i i disc

X

Cost of a RCS policy:

The evaluation of the required investment to reach a given number of grid bags involves two issues:

• The installation cost of one RCS

• The link between the number of installed RCS and the number of created grid bags The installation cost of one RCS has been based on experience. Three cases have been selected (see “III.1 Definition and utility” for more information on materials):

• Transformation of a PSSA to a PSSB (a PSSA is a PSSB without any MV-switcher) • Installation of one Remote controlled function on a MV-LV substation (two direction

node)

• Installation of 2 remote controlled function on a MV-LV substation (three direction node)

For each of these operations the execution rate as well as the mean cost have then be evaluated:

Execution rate cost (€)

50% 11000 36% 6000 14% 8000 100% 8780 "mean" RCS PSSB 1RC LV-MV substation 2RC LV-MV substation

Table 3: Data on RCS (estimations)

The link between RCS number and created grid bags number has already been explored in part “III.2 Link between RCS and reactivity: notion of “grid bags””. Three results should be reminded:

• Each of the above presented operations yields a different number of grid bags: one PSSB as well as one 1RC LV-MV substation lead to one supplementary grid bag, whereas a 2 RC LV-MV substation creates 2 supplementary grid bags. Regarding the so-called execution rates, this gives an average of (0.5+0.36)*1 + 0.14*2 = 1.14 created grid bags per installed RCS.

(39)

• For a given feeder, the installation of rescue points does not create any grid bag but its cost has to be included. On the other hand, when rescue RCS are installed without other RCS, any supplementary “mean RCS” will create 1+1.14 grid bag.

This yields:

Where NRCStot is the total number of RCS (including rescue RCS) as understood in Table 3. It

should then be interpreted as RCS equipped nodes. The same comments can be made for NrescuRCS which is the number of rescue nodes.

The cost of a RCS policy on a given area can then be derived:

Where CRCS is the mean cost of one RCS (settled to 8780€ considering Table 3).

V.2 Results

Every parameter needed is known on the former Centre-mesh. The principle of the calculation is to define an optimum number of grid bags per feeder in each Centre considering that their organisation and grid structure do note change (especially reactivity periods and number of rescue points per feeder). This optimum will then be compared to the actual grid bag number.

The optimisation has been realised by the numerical solving program available on Microsoft Excel 2000 and aims at finding the grid bag number that yields the maximum balance “Gain-Cost”. The use of Excel is possible as the problem presents only one variable.

Moreover, regarding the steady and simple behavior of the functions involved (both terms are always increasing and the balance has only one maximum), the non linear aspect is not a problem and a simple no-constraint Newton method (available in Excel parameters) enables to get quick and reliable results (around 15 s are required to get a value on all the 97 Centres).

For each Centre, the objective function f to be maximised is derived from equations 10,11 and 13 :

COST

GAIN

X

N

f

_Centre

(

_gb

)

=

_disc

⋅

_RCS

−

The results for the Mediterranean Region are presented on Table 4.

(40)

Marseille 3,5 2,9

Provence 4,3 3,8

Alpes du Sud 4,7 3,4

Avignon Grand Delta 4,2 2,9

Var 5,7 3,9

Nice Alpes d'azur 4,3 3,1

Cannes 6,0 3,7 Vallées d'Aude 5,1 5,4 Pyrénées Roussillon 4,3 3,7 Montpellier Hérault 3,6 3,9 Gard Cévennes 4,3 3,2 Aix Toulon Nîmes Optimized grid bag number per feeder

Actual grid bag number per

feeder

Operation

Agency

Centre

Table 4: Comparison between actual and optimised grid bag number.

Big differences are again raised between former Centres. Here they stem from different grid structures (customer and rescue point number per feeder, reliability) and reactivity period (terrain, overhead or underground…)

Based on a scenario where stress would be laid on every under-equipped Centre to reach the optimum, investment as well as benefit (NDE spared, B criterion decreased) can be derived and yields the results shown in Table 5.

Marseille 1,9 1 197 220 996 684

Provence 2,5 1 845 583 1 633 307

Alpes du Sud 10,6 2 718 008 1 963 227

Avignon Grand Delta 10,6 5 165 563 3 507 293

Aix 5,2 10 926 375 8 100 511

Var 8,6 7 787 475 5 283 172

Nice Alpes d'azur 5,2 3 337 913 2 388 281

Cannes 9,2 3 692 178 2 262 206 Toulon 7,6 14 817 565 9 933 658 Vallées d'Aude 0 0 Pyrénées Roussillon 1,9 839 742 729 198 Montpellier Hérault 0 0 Gard Cévennes 5,7 2 860 264 2 153 901 Nîmes 1,8 3 700 007 2 883 099 4,8 29 443 947 20 917 268 Aix Toulon Nîmes Méditéranée B criterion decrease if grid bag number

was optimized NDE spared over 10 years (discount rate included) (€) Investment needed to reach optimum (€)

Operation

Agency

Centre

Table 5: Impact of a policy aiming at reaching the optimum from present situation

(41)

The most relevant value is the impact on the B criterion (as it does not account for assumptions on NDE price and RCS cost). Comparing the results to the actual B criterion shows that RCS enables tremendous progress. Especially in the most well-equipped Centre the benefits of a strong RCS policy appear to be very interesting. These results should however not obliterate that the efficiency of RCS regarding the B criterion increases when the annual incident number decreases, which should obviously not be regarded as a benefit.

To conclude it should be reminded that even if most of these results have been corroborated by Region experience and own studies, some strong hypothesis have been taken:

• Every grid bag of a feeder contains the same customer number • Each Centre is homogeneously equipped

• RCS failures have been almost neglected

(42)

V.3 Discussion on technical economical approach

Again, this approach was used to priorise investments in terms of areas and to account for their specificity. This purpose was successfully achieved and results have been corroborated by Region teams experience. However the model can be used for wider application, and it could be suitable to overtake the previously defined optimum. Indeed, even if it is not a long-term solution, RCS will always be a cheaper way to enhance reactivity and to counter the age-related loss of reliability. The example shown on Figure 33 has been chosen for two reasons: the actual equipment is close to the optimum and it overtakes the optimum. Indeed, whereas the technical-economical approach puts this Centre away from the study focus, RCS investment is still a suitable solution.

RCS policy impact on "Vallée d'Aude"

4,7 5,0 -5,0 0,0 5,0 10,0 15,0 20,0 25,0 30,0 1,1 1,5 1,8 2,2 2,5 2,9 3,2 3,6 3,9 4,3 4,6 5,0 5,3 5,7 6,0 6,4 6,7 7,1 7,4 7,8 8,1 8,5 8,8 Investment volume (M€) 0,0 10,0 20,0 30,0 40,0 50,0 60,0 70,0 80,0 90,0 100,0 1,1 1,5 1,8 2,2 2,5 2,9 3,2 3,6 3,9 4,3 4,6 5,0 5,3 5,7 6,0 6,4 6,7 7,1 7,4 7,8 8,1 8,5 8,8 RCS number per feeder

Balance over 10 years (M€) B criterion decrease (min) technical economical optimun present situation gain (M€) Balance maximum

Balance is not negative even for high equippment level B-criterion gain is limited

RCS policy impact on "Vallée d'Aude"

4,7 5,0 -5,0 0,0 5,0 10,0 15,0 20,0 25,0 30,0 1,1 1,5 1,8 2,2 2,5 2,9 3,2 3,6 3,9 4,3 4,6 5,0 5,3 5,7 6,0 6,4 6,7 7,1 7,4 7,8 8,1 8,5 8,8 Investment volume (M€) 0,0 10,0 20,0 30,0 40,0 50,0 60,0 70,0 80,0 90,0 100,0 1,1 1,5 1,8 2,2 2,5 2,9 3,2 3,6 3,9 4,3 4,6 5,0 5,3 5,7 6,0 6,4 6,7 7,1 7,4 7,8 8,1 8,5 8,8 RCS number per feeder

Balance over 10 years (M€) B criterion decrease (min) technical economical optimun present situation gain (M€) Balance maximum

Balance is not negative even for high equippment level B-criterion gain is limited

(43)

Investment Efficiency in "Vallée d'Aude" 0,00 5,00 10,00 15,00 20,00 1,1 1,5 1,8 2,2 2,5 2,9 3,2 3,6 3,9 4,3 4,6 5,0 5,3 5,7 6,0 6,4 6,7 7,1 7,4 7,8 8,1 8,5 8,8

RCS number per feeder

B criterion decrease per invested M€ (min/M€) Technical economical optimumo present situation

Figure 34: Investment Efficiency in Vallée d'Aude Centre

Three curves are shown on Figure 33 and Figure 34. The blue one is the technical optimical balance (NDE over 10 years-investment cost) and defines the technical-economical optimum. The most important curve is the red one, showing the impact of RCS equipment on the B-criterion. This curve is of course limited by an asymptote (corresponding to the hypothetical equipment enabling to reconnect every customers under the time TRCS). The

(44)

VI. Conclusion

The results of this document show that today context where CAPEX are compressed whereas the grid becomes older is to lead to a further development of RCS in the future. Especially as IT-systems enable to simultaneously deal with a larger and larger number of actions, RCS remain an efficient way to enhance reactivity and to limit the impact of grid reliability loss as long as maintenance can be handled.

However, RCS do not reduce the incident number: if B criterion can be tremendously decreased, customers still have to face an increasing number of outages (even if they are shorter for many of them). Especially as ERDF is developing automatic controls of RCS in its Operation agencies, which will tends to reduces TRCS under 3 minutes, this aspect should be

stressed: all the customers reconnected within less than 3 minutes will not be counted in B-criterion.

On the second hand, RCS efficiency tends to be reduced in case of a localised major climatic event, where feeder-rescuing possibilities are limited.

These two axes should be accounted in a deeper study based on the modelling approach developed here.

(45)

Liste of Figures

Figure 1: French electricity market organization... 6

Figure 2: French EDF electricity production repartition (left) and French EDF production capacity (right)[1] ... 7

Figure 3: EDF Shareholders repartition [1] ... 7

Figure 4: Distribution grid structure ... 8

Figure 5: Annual energy flow through ERDF grid in 2008... 9

Figure 6: Evolution of the B-criterion (min) vs time. Only everyday outages due to the distributor are accounted on this graph... 11

Figure 7: B criterion increasing and reliability deterioration over the past 10 years. Only the “everyday” and MV-related outages are accounted on this graph ... 11

Figure 8: Evolution of investment volume vs. time ... 12

Figure 10: The 30 operation agencies of ERDF and their controlled area ... 14

Figure 11: Relation between the driving interface and SIG ... 15

Figure 12: Single phase circuit of a primary substation and of a feeder ... 16

Figure 13: Source to source structure for overhead lines ... 17

Figure 16: Classical historical overhead RCS radio controlled with opening mechanisms moved in air... 19

Figure 17: So-called “August” RCS with opening mechanisms moved in SF6... 20

Figure 18: So-called PSSB (« Poste au Sol Simplifié B ») used for underground lines, it is basically a LV-MV substation on which a RC switch have been installed. ... 20

Figure 19: Feeder configuration in normal use... 21

Figure 20: Feeder configuration right after an incident ... 22

Figure 21: Feeder configuration after the RCS step of the localisation phase ... 22

Figure 22: Feeder configuration during repairing phase ... 23

Figure 23:Limit of one grid bag on a source-to-source structure ... 24

Figure 24: Two grid bags feeder outline ... 25

Figure 25: Example of a feeder where two lines could be defined as "main" ... 26

Figure 26: Example of a RCS on a non-looped line ... 26

Figure 27: Illustration of multiple direction node impact on the number of bags... 27

Figure 28: Feeder with a MV/V substation equipped with RCS on a main line ... 28

Figure 29: Microsoft Access table extracted from GDO database. The table describing the national grid contains 2,370,000 rows... 29

Figure 30: Grid bag number expression illustration ... 30

Figure 31: Disconnected customer number vs. Time. Graph extracted from EtaReso for one incident (underground line) in Aix-en-Provence agency ... 31

Figure 32: Recovering curves model ... 33

Figure 33: Impact of RCS policy, figures from “Vallee d’Aude” Centre ... 42

Figure 34: Investment Efficiency in Vallée d'Aude Centre ... 43

List of Tables

Table 1: Mediterranean Region data ... 30

Table 2: Reactivity periods data and reliability of the model for Mediterranean Region... 35

Table 3: Data on RCS (estimations) ... 38

Table 4: Comparison between actual and optimised grid bag number. ... 40

(46)

References

[1] Rapport d’activité 2008, Groupe EDF, www.edf.fr

[2] « Structure de réseau », ERDF internal formation session, août 2009 [3] ERDF experts from Network and Assets Direction

[4] “Données économiques de référence”, internal ERDF note D.1.3-01 [5] so-called « COIFFIER » method, Network and Assets Direction expert

[6] Network and Assets Direction expert : mean value of the power flow from the TSO over the customer number.

Glossary

B-criterion:

The most important indicator of the service quality. It is approximately equal to the System Average Interruption Duration Index (SAIDI): it expresses the outage mean time for each low voltage customer and is thus directly related to the delivery quality from the society point of view. See part “IV.1 How reactivity is measured and controlled” for more information.

RCS:

Remote Controlled Switches. If no other indication is given this notion is restricted to MV remote controlled switches that can be actioned under voltage. Rescue RCS is a RCS that is left open in normal used and whoch is only used to connect a feeder to another.

HV:

High Voltage: represents the voltage level on the transmission grid, that is to say 63, 90, 225 and 400kV and all votlage that would be aove 50kV.

MV:

Medium Voltage: represents the highest voltage leveles of the distribution grid, that is to say mainly 20kV (but can varies from 10 to 50kV)

LV:

Low Voltage: represents the lowest voltage levels of the distribution grid, that is to say mainly 400V or all voltage level that would be below 1kV.

Gir Bag:

Impact of remote controlled switches on distribution grid recovering processes