METHOD AND SYSTEM FOR DETERMINING THE STRUCTURE OF AN ELECTRICITY TRANSMISSION GRID AND ASSOCIATED COMPUTER PROGRAM

Abstract

A method is provided for determining the structure of an electricity transmission grid, including a power supply station including one or more feeders for supplying electrical power to a plurality of electricity consumers, the method including acquiring first data relating to the electrical power consumed by each consumer during various time intervals; acquiring second data relating to the electrical power delivered by each feeder during the various time intervals; generating a plurality of different classes of data; calculating a first selection criteria for each class of data; selecting a first set of classes of data from the classes generated; and determining connection parameters.

Description

The present invention relates to a method implemented by computer, for determining the structure of an electricity transmission grid, an associated computer program and a system for determining the structure of the electricity transmission grid.

In the field of electricity transmission, it is known to provide electricity transmission to a plurality of electricity consumers via an electricity transmission grid comprising an electricity transformer station, the station comprising electrical feeders for supplying electrical power to electricity consumers. The transformer station is, for example, a high voltage/medium voltage (HV/MV) or medium voltage/low voltage (MV/LV) transformer station and the electrical feeders are medium or low voltage feeders.

High voltage corresponds to a voltage greater than 50 kVolts (kV), medium voltage corresponds to a voltage between 1 kV and 50 kV and low voltage corresponds to a voltage less than 1 kV.

In such electricity transmission grids, the consumers are supplied either in three-phase, or in single-phase, and the transformer station is configured for distributing the electrical power that it receives between the various electrical feeders.

However, the structure, and more precisely the mapping, of the electricity transmission grids supplying the electricity consumers is generally poorly known. But knowledge of the structure of such grids has become essential for meeting the constraints that national regulators impose on electricity transmission firms, notably in terms of quality of electricity supply.

Indeed, national regulators impose quality of service objectives on electricity transmission firms, and these must be measurable. Electricity transmission firms henceforth therefore need to be better acquainted with the structure of electricity transmission grids, notably for locating possible faults or failures on the grid or for identifying the feeders of the transformer station most contributing to electrical losses.

Thus a method is known from EP 2 458 340 A2 for determining the structure of an electricity transmission grid from specific calculation means and information relating to the electrical power consumed by each electricity consumer connected to the grid and relating to the electrical power delivered by each feeder from the transformer station.

However, the speed of execution, accuracy, and reliability of such a method remain to be improved.

The object of the invention is therefore to provide a robust method for determining the structure of an electricity transmission grid making it possible to determine the structure of the electricity transmission grid in a safer, more reliable and faster way. Such a method notably makes it possible to overcome errors, e.g. related to inaccuracies in measuring the power consumed and/or delivered or to the absence of some energy measurements.

For this purpose, the subject matter of the invention is a method for determining the structure of an electricity transmission grid, the method being implemented by computer, the transmission grid comprising a power supply station comprising one or more feeders for supplying electrical power to a plurality of electricity consumers and comprising the following steps:

• • a) acquiring, for each electricity consumer, first data relating to the electrical power consumed by the electricity consumer during various time intervals, measured via a first sensor for measuring power,
  • b) acquiring, for each feeder, second data relating to the electrical power delivered by the feeder during various time intervals, measured via a second sensor for measuring power,
  • c) generating, via a generating module, a plurality of different classes of data, each associated with one of the time intervals, each class of data comprising the first and second data associated with said time interval,

characterized in that the method further comprises the following steps:

• • d) calculating a first selection criterion for each class of data, the first selection criterion being a criterion chosen from an overall rate of electrical power losses between the feeder or feeders and the consumers and a difference in consumption of electrical power between the various consumers,
  • e) selecting, by a selection unit and on the basis of the first selection criteria calculated, a first set of classes of data from the classes generated,
  • f) determining, from the first selected set, connection parameters, said parameters comprising for each electricity consumer an identifier of the feeder to which it is connected.

Thanks to the invention, the steps of selecting the first set of classes of data on the basis of the first selection criteria calculated and of determining the connection parameters from the first selected set make it possible to determine the structure of the electricity transmission grid in a more reliable and faster way.

According to other advantageous aspects of the invention, the method of determination further comprises one or more of the following features, taken in isolation or according to all the technically acceptable combinations:

• • during the calculating step d), a second selection criterion is calculated for each class of data generated, the second criterion being the other criterion chosen from an overall rate of electrical power losses between the feeder or feeders and the consumers and a difference in consumption of electrical power between the various consumers,and, during the selecting step e), the first set is selected on the basis of the first and second criteria calculated;
  • following the selecting step e), the method comprises the following steps:
    • e1) calculating a second selection criterion for each class of data of the first set, the second criterion being the other criterion chosen from an overall rate of electrical power losses between the feeder or feeders and the consumers and a difference in consumption of electrical power between the various consumers,
    • e2) selecting on the basis of the second selection criteria calculated, a second set of classes of data from the first set,determining the connection parameters then being performed during the determining step f) on the basis of the second set selected from the first set;
  • during the determining step f), the connection parameters are determined separately for each feeder and independently of the second data relating to the electrical power acquired for the other feeders;
  • during the determining step f), an equation system to be solved is determined from a premise of conservation of electrical power for each feeder according to which the power delivered by the feeder is substantially equal to the sum of the power consumed by the electricity consumers connected to the feeder and the electrical losses;
  • during the determining step f), the equation system is defined from the following equation:

$E_{\mathrm{Dj}}\left(t_k\right)=\sum _{i=1}^n\left(a_{\mathrm{ij}}·E_{\mathrm{Ci}}\left(t_k\right)+\partial 1_{\mathrm{jk}}-\partial 2_{\mathrm{jk}}\right),$

where n is the number of electricity consumers, j is a feeder index varying from 1 to m, m being the number of feeders, i is an electricity consumer index varying from 1 to n,

• k is a class of data index of the first set varying from 1 to w, w being the number of classes of data of the first set, t_k is the time interval associated with the class of data of index k,
• a_if is the connection parameter indicating whether the consumer of index i is connected to the feeder of index j,
• E_Ci(t_k) is the electrical power consumed by the electricity consumer of index i during the time interval t_k,
• E_Dj(t_k) is the electrical power delivered by the feeder of index j during the time interval t_k, and
• ∂1_jk and ∂2_jk are adjustment variables representative of the electrical losses;
  • the equation system is written in matrix form as follows: A*z=B,
• where

$\begin{array}{cc}z=\left[\begin{array}{c}{a}_{11}\\ ⋮\\ a_{n1}\\ ⋮\\ a_{1m}\\ ⋮\\ a_{\mathrm{nm}}\\ \partial 1_{11}\\ ⋮\\ \partial 1_{\mathrm{new}}\\ \partial 2_{11}\\ ⋮\\ \partial 2_{\mathrm{mw}}\end{array}\right],& \\ B=\left[\begin{array}{c}{E}_{D1}\left(t_1\right)\\ ⋮\\ E_{D1}\left(t_w\right)\\ E_{D2}\left(t_1\right)\\ ⋮\\ E_{D2}\left(t_w\right)\\ E_{\mathrm{Dm}}\left(t_1\right)\\ ⋮\\ E_{\mathrm{Dm}}\left(t_k\right)\end{array}\right]& \\ \mathrm{and}A=\left[\begin{array}{cccc}U& 0& \dots & 0\\ 0& \ddots & \ddots & ⋮\\ ⋮& \ddots & \ddots & 0\\ 0& \dots & 0& U\end{array}\right]& \end{array}$

with

$U=\left[\begin{array}{ccc}{E}_{C1}\left(t_1\right)& \dots & E_{\mathrm{Cn}}\left(t_1\right)\\ ⋮& \ddots & ⋮\\ E_{C1}\left(t_w\right)& \dots & E_{\mathrm{Cn}}\left(t_w\right)\end{array}\right]$

and A comprising m*n columns and m*w rows;

• • during the determining step, an optimization algorithm is used for solving the equation system, the optimization algorithm verifying:

$\underset{z}{\min }f^T·z$

such that

$\{\begin{array}{c}A·z=B\\ 1_z\le z\le u_z\end{array},$

where

$f^T·z=\begin{array}{ccc}\underset{\underset{m^{*}n}{}}{[0\dots 0}& \underset{\underset{w^{*}m}{}}{1\dots 1}& \underset{\underset{w^{*}m}{}}{1\dots 1]}\end{array}\times z,l_z=\left[\begin{array}{c}-\varepsilon \\ ⋮\\ -\varepsilon \\ 0\\ ⋮\\ 0\end{array}\right]$ $\mathrm{and}$ $u_z=\left[\begin{array}{c}1+\varepsilon \\ ⋮\\ 1+\varepsilon \\ +\infty \\ ⋮\\ +\infty \end{array}\right],$

with ε between 0 and 0.5;

• • each feeder comprises one or more electrical phase conductors, and, during the acquiring step b), the second data are acquired for each electrical phase conductor and relate to the electrical power delivered by each electrical phase conductor, and, during the determining step f), the connection parameters determined comprise for each electricity consumer an identifier of the electrical phase conductor or conductors to which it is connected;
  • during the determining step f), the method comprises the following steps:
    • f1) pseudo-random selecting of a third set of classes of data included in the first set,
    • f2) determining the connection parameters from the third set selected from the first set:
  • following the selecting step f1) and prior to the determining step f2), the method comprises the following steps:
    • f3) calculating a repetition parameter,
    • f4) determining, from the third selected set, intermediate connection indices comprising for each electricity consumer an identifier of the feeder to which it is connected,
    • f5) storing the intermediate connection indices, and
    • f6) comparing the repetition parameter with at least one predetermined repetition criterion,and, following the result of the comparing step f6), performing a step from, on the one hand, the repetition of the steps of selecting f1), calculating f3), determining f4), storing f5) and comparing f6) and, on the other hand, the step of determining f2) the connection parameters on the basis of the stored intermediate connection indices;
  • the method comprises the following step:
    • f7) calculating, for each consumer and for each feeder an assignment number relating to the number of times that the stored intermediate connection indices indicate that the consumer is connected to the feeder,and, during the determining step f2), the connection parameters are determined from the assignment numbers calculated.

The subject matter of the invention is also a computer program comprising software instructions, which when they are executed by a computer, implement the method as defined above.

The subject matter of the invention further relates to a system for determining the structure of an electricity transmission grid, the transmission grid comprising a power supply station comprising one or more feeders for supplying electrical power to a plurality of electricity consumers, the system comprising:

• • for each electricity consumer, a first sensor for measuring first data relating to the electrical power consumed by the consumer during various time intervals,
  • for each feeder, a second sensor for measuring second data relating to the electrical power delivered by the feeder during the various time intervals,
  • a module for acquiring the first and second data,
  • a module for generating a plurality of different classes of data, each associated with one of the time intervals, each class of data comprising the first and second data associated with said time interval,

characterized in that said determining system further comprises:

• • a module for calculating a first selection criterion for each class of data, the first selection criterion being a criterion chosen from an overall rate of electrical power losses between the feeder or feeders and the consumers and a difference in consumption of electrical power between the various consumers,
  • a module for selecting, on the basis of the first selection criteria calculated, a first set of classes of data from the classes generated, and
  • a module for determining, from the first selected set, connection parameters, said parameters comprising for each electricity consumer an identifier of the feeder to which it is connected.

The invention will be better understood and other advantages thereof will appear more clearly in the light of the following description, given solely as a non-restrictive example, and referring to the drawings, in which:

FIG. 1 is a schematic representation of an electricity transmission grid comprising a power supply station comprising a plurality of feeders for supplying electrical power to a plurality of electricity consumers;

FIG. 2 is a flow chart of a method for determining the structure of the electricity transmission grid in FIG. 1, according to a first embodiment of the invention;

FIG. 3 is a flow chart similar to that in FIG. 2, according to a second embodiment of the invention;

FIG. 4 is a flow chart similar to that in FIG. 2, according to a third embodiment of the invention; and

FIG. 5 is a flow chart similar to that in FIG. 2, according to a fourth embodiment of the invention.

In FIG. 1, an electricity transmission grid 10 is associated with a system 12 for determining the structure of the electricity transmission grid 10. The transmission grid 10 comprises a power supply station 14 for supplying electrical power to a plurality of electricity consumers 16_i. The power supply station 14 comprises electrical feeders 18_j for supplying electrical power to electricity consumers 16_i, with i an electricity consumer index and j a feeder index.

The electricity consumer index i varies from 1 to n, where n is the number of electricity consumers 16_i and the feeder index j varies from 1 to m, where m is the number of electrical feeders 18_j. In the example of FIG. 1, the number n of electricity consumers 16_i is equal to 3, and the number m of electrical feeders 18_j is equal to 2.

The determining system 12 comprises, for each electrical consumer 16_i, a first sensor 20_i for measuring power and a module 21, for transmitting the power measurements performed by the corresponding first sensor 20_i.

The determining system 12 comprises, for each feeder 18_j, a second sensor 22_j for measuring power.

The determining system 12 also comprises an information processing unit 24 formed, for example, from a processor 26 and a memory 28 associated with the processor 26.

The power supply station 14 is, for example, a medium voltage/low voltage transformer station connected between a medium voltage grid, not represented, and a low voltage grid 30 corresponding to the electricity consumers 16_i.

The power supply station 14 includes, at each feeder 18_j, the corresponding second sensor 22_j.

The electricity consumers 16_i are connected to the power supply station 14 via the feeders 18_j. More precisely, in the example of FIG. 1, the electricity consumers 16₁, 16₂ are connected to the feeder 18₁ and the electricity consumer 16₃ is connected to the feeder 18₂.

The electricity consumers 16_i are either three-phase and supplied with power by the corresponding feeder 18_j via four electrical conductors 32_j, 34_j, 36_j, 38_j, i.e. three electrical phase conductors 32_j, 34_j, 36_j and a neutral electrical conductor 38_j, or single-phase and supplied with power by the corresponding feeder 18_j via two electrical conductors: i.e., for example, a phase conductor 32_j, 34_j or 36_j, and the neutral conductor 38_j.

In the example in FIG. 1, the electricity consumer 16₁ is three-phase and the electricity consumers 16₂, 16₃ are single-phase.

Each electricity consumer 16_i includes the corresponding one of the first sensors 20_i and the corresponding one of the transmission modules 21_i.

Each electricity consumer 16_i is, for example, a communicating electricity consumption meter, appropriate for measuring the first data E_Ci relating to the electrical power consumed by the electricity consumer 16_i, via the corresponding first sensor 20_i, and transmitting the first data E_Ci to the processing unit 24 via the corresponding transmission module 21_i.

Each feeder 18_j is a three-phase feeder and comprises the corresponding three phase conductors 32_j, 34_j, 36_j and the corresponding neutral conductor 38_j. As a variant, not represented, the electrical feeders are single-phase and comprise one phase conductor and one neutral conductor. According to yet another variant, not represented, some feeders are single-phase and others are three-phase.

Each first sensor 20_i is appropriate for measuring the first data E_Ci(t_l) relating to the electrical power consumed by the corresponding electricity consumer 16_i, during various time intervals t_l.

More generally, the first sensors 20_i are configured for measuring the first data E_Ci(t_l) during the same time intervals t_l, the first data E_Ci(t_l) measured at each electricity consumer 16_i being measured synchronously.

Each transmission module 21_i is appropriate for transmitting the first data E_Ci(t_l) measured by the corresponding first sensor 20_i to the processing unit 24.

Advantageously, each transmission module 21_i is appropriate for transmitting with the first data E_Ci(t_l) a first item of information relating to the time interval during which the first data were measured.

Each second sensor 22_j is appropriate for measuring the second data E_Dj(t_l) relating to the electrical power delivered by the corresponding feeder 18_j, during the various time intervals t_l.

The first E_Ci(t_l) and second E_Dj(t_l) data measured are then synchronized in the direction in which they are measured during identical time intervals t_l.

Each second sensor 22_j is also configured for transmitting, via a respective electrical link 40_j, the second data E_Dj(t_l) that it measures to the processing unit 24.

Advantageously, each second sensor 22_j is configured for transmitting with the second data E_Dj(t_l) a second item of information relating to the time interval t_l during which the second data E_Dj(t_l) were measured.

The first E_Ci(t_l) and second E_Dj(t_l) data are, for example, active power measurements.

More precisely, in the rest of the description, it is considered that the first E_Ci(t_l) and second E_Dj(t_l) data are active power measurements.

As a variant, the first E_Ci(t_l) and second E_Dj(t_l) data are reactive energy measurements, apparent energy measurements, active power measurements, reactive power measurements, apparent power measurements or even current measurements.

The processor 26 is configured for running software comprised in the memory 28.

The memory 28 comprises software 41 for acquiring the first E_Ci(t_l) and second E_Dj(t_l) data, software 42 for generating a plurality of different classes of data Je_l from the first data E_Ci(t_l) and the second data E_Dj(t_l) acquired in a same time interval t_l and software 44 for calculating a first selection criterion C1, for each class of data Je_l.

The memory 28 also comprises software 46 for selecting a first set En₁ of classes of data Je_l, on the basis of the first selection criteria C1_l calculated by the calculation software 44 and software 48 for determining connection parameters a_ij, said connection parameters comprising for each electricity consumer 16_i an identifier of the feeder 18_j to which it is connected.

The acquisition 41, generation 42, calculation 44, selection 46 and determination 48 software correspond to software instructions and form a computer program capable of being run by a computer. The computer corresponds, for example, to the processing unit 24.

The acquisition software 41 is appropriate, for example, for transmitting to each electricity consumer 16_i and notably to each first sensor 20_i an order to measure first data E_Ci(t_l) and an order to transmit first data E_Ci(t_l), in order to retrieve the first data.

Similarly, the acquisition software 41 is, for example, configured for transmitting to each feeder 18_j and notably to each second sensor 22_j an order to measure second data E_Dj(t_l) and an order to transmit second data E_Dj(t_l), in order to retrieve the second data E_Dj(t_l).

Advantageously the orders to measure first E_Ci(t_l) and second E_Dj(t_l) data are transmitted simultaneously to all the feeders 18_j and all the consumers 16_i.

The generation software 42 is configured for generating classes of data Je_l which are each associated with one of the time intervals t_l and which comprise the first E_Ci(t_l) and second E_Dj(t_l) data associated with said time interval t_l. In other words, the generation software 42 selects the first E_Ci(t_l) and second E_Dj(t_l) data measured during the various time intervals t_l, for creating the classes of data Je_l.

The calculation software 44 is suitable for calculating, for each class of data Je_l, the first selection criterion C1_l, which is chosen from an overall rate of electrical power losses between the feeders 18_j and the electricity consumers 16_i and a difference in consumption of electrical power between the various consumers 16_i. The electrical power losses include both ‘technical’ losses, e.g. losses linked to Joule losses during the flow of the current between the feeders 18_j and the consumers 16_i, and ‘non-technical’ losses, which are, for example, linked to electricity theft, owing to the fact that consumers are connected to the transmission grid 10 without the information processing unit 24 being informed thereof and to defective first sensors 20_i.

If the first selection criterion C1_l is, for example, the overall rate of electrical power losses between the feeders 18_j and the consumers 16_i, the first selection criterion C1_l is calculated from the following formula:

$\begin{array}{cc}C1_j=\frac{\Delta E}{E}\left(t_l\right)=\frac{\sum _{j=1}^mE_{\mathrm{Dj}}\left(t_l\right)-\sum _{i=1}^nE_{\mathrm{Ci}}\left(t_l\right)}{\sum _{j=1}^mE_{\mathrm{Dj}}\left(t_l\right)},& \left(1\right)\end{array}$

where l is a class of data index and varies from 1 to r, r being the number of classes of data generated by the generation software 42, t_l is the time interval associated with the class of data of index I, E_Ci(t_l) is the electrical power consumed by the electricity consumer of index i during the time interval t_l and E_Dj(t_l) is the electrical power delivered by the feeder of index j during the time interval t_l.

If the first selection criterion C1_l is, for example, the relative difference in consumption of electrical power between the various consumers 16_i, the first selection criterion C1_l is calculated from the following formula:

C1_l=Var(E_C1(tk), . . . , E_Cn(tk)),   (2)

where Var is the variance function.

As a variant, if the first selection criterion C1_l is, for example, the difference in consumption of electrical power between the various consumers, the first selection criterion C1_l is calculated from the following formula:

$\begin{array}{cc}C1_l=\frac{\mathrm{Var}\left(E_{C1}\left(\mathrm{tk}\right),\dots ,E_{\mathrm{Cn}}\left(\mathrm{tk}\right)\right)}{\mathrm{Moy}\left(E_{C1}\left(\mathrm{tk}\right),\dots ,E_{\mathrm{Cn}}\left(\mathrm{tk}\right)\right)},& \left(3\right)\end{array}$

where Moy is a mean function, such as an arithmetic, geometric, quadratic, etc. mean.

The selection software 46 is configured for selecting the first set En₁ of classes of data from the classes of data Je_l generated by the generation software 42, on the basis of the first selection criteria C1_l calculated. The selection software 46 is, for example, configured for comparing the first criteria C1_l with a first predetermined variable V1 and for selecting the classes of data Je_l for which the first criterion C1_l is less than the first predetermined variable V1.

The determination software 48 is configured for determining the connection parameters a_ij from the first selected set En₁.

The determination software 48 is, for example, configured for establishing or determining an equation system to be solved from a premise of conservation of electrical power for each feeder 18_j, according to which the power delivered by the feeder 18_j is substantially equal to the sum of the power consumed by the electricity consumers 16_i connected to the feeder 18_j and the electrical losses.

The equation system is, for example, defined from the following equation:

$\begin{array}{cc}{E}_{\mathrm{Dj}}\left(t_k\right)=\sum _{i=1}^n\left(a_{\mathrm{ij}}·E_{\mathrm{Ci}}\left(t_k\right)+{21}_{\mathrm{jk}}-\partial 2_{\mathrm{jk}}\right)& \left(4\right)\end{array}$

where n is the number of electricity consumers, j is a feeder index varying from 1 to m, m being the number of feeders, i is an electricity consumer index varying from 1 to n,

• k is a class of data index Je_k of the first set En₁ varying from 1 to w, w being the number of classes of data Je_k of the first set En₁, t_k is the time interval associated with the class of data Je_k of index k,
• a_ij is the connection parameter indicating whether the consumer of index i is connected to the feeder of index j,
• E_Ci(t_k) is the electrical power consumed by the electricity consumer of index i during the time interval t_k,
• E_Dj(t_k) is the electrical power delivered by the feeder of index j during the time interval t_k, and
• ∂1_jk and ∂2_jk are adjustment variables representative of the electrical losses, i.e. of the difference for each class of data Je_k and for each feeder 18_j between the electrical power delivered by the feeder 18_j and the electrical power consumed by the electricity consumers 16_i connected to the feeder of index j.

The equation system is then written, for example, in matrix form as follows:

A*z=B,   (5)

where

$\begin{array}{cc}z=\left[\begin{array}{c}{a}_{11}\\ ⋮\\ a_{n1}\\ ⋮\\ a_{1m}\\ ⋮\\ a_{\mathrm{nm}}\\ \partial 1_{11}\\ ⋮\\ \partial 1_{\mathrm{new}}\\ \partial 2_{11}\\ ⋮\\ \partial 2_{\mathrm{mw}}\end{array}\right],& \\ B=\left[\begin{array}{c}{E}_{D1}\left(t_1\right)\\ ⋮\\ E_{D1}\left(t_w\right)\\ E_{D2}\left(t_1\right)\\ ⋮\\ E_{D2}\left(t_w\right)\\ E_{\mathrm{Dm}}\left(t_1\right)\\ ⋮\\ E_{\mathrm{Dm}}\left(t_k\right)\end{array}\right]& \\ \mathrm{and}A=\left[\begin{array}{ccc}D& I_{w^{*}m}& -I_{w^{*}m}\end{array}\right],& \end{array}$

I_w*m representing the unity matrix of size w*m, with

$D=\left[\begin{array}{cccc}U& 0& \dots & 0\\ 0& \ddots & \ddots & ⋮\\ ⋮& \ddots & \ddots & 0\\ 0& \dots & 0& U\end{array}\right],U=\left[\begin{array}{ccc}{E}_{C1}\left(t_1\right)& \dots & E_{\mathrm{Cn}}\left(t_1\right)\\ ⋮& \ddots & ⋮\\ E_{C1}\left(t_w\right)& \dots & E_{\mathrm{Cn}}\left(t_w\right)\end{array}\right]$

and D comprising m*n columns and m*w rows.

The determination software 48 is then configured for solving the equation system from an optimization algorithm verifying:

$\underset{z}{\min }f^T·z$

such that

$\begin{array}{cc}\{\begin{array}{c}A·z=B\\ 1_z\le z\le u_z\end{array},& \left(6\right)\end{array}$

where

$f^T·z=\begin{array}{ccc}\underset{\underset{m^{*}n}{}}{[0\dots 0}& \underset{\underset{w^{*}m}{}}{1\dots 1}& \underset{\underset{w^{*}m}{}}{1\dots 1]}\end{array}\times z,$

f^T corresponding to a cost function of the optimization algorithm and z to an objective vector defined above and where,

$l_z=\left[\begin{array}{c}-\varepsilon \\ ⋮\\ -\varepsilon \\ 0\\ ⋮\\ 0\end{array}\right],$

I_z being of dimension m*n+2*w*m and comprising m*n times the value −ε and 2*w*m times the value 0 and

$u_z=\left[\begin{array}{c}1+\varepsilon \\ ⋮\\ 1+\varepsilon \\ +\infty \\ ⋮\\ +\infty \end{array}\right],$

u_z being of dimension m*n+2*w*m and comprising m*n times the value (1+ε) and 2*w*m times the value +∞, with ε between 0 and 0.5, preferably between 0 and 0.1, preferably still equal to 0.05.

The optimization algorithm is appropriate for determining the objective vector z and therefore the connection parameters for which the function f^T.z is minimized. More precisely, the determination software 48 is configured for starting from an initial vector solution

$X=\left[\begin{array}{c}{a}_{11}\\ ⋮\\ a_{nm}\end{array}\right]$

chosen randomly and for iteratively converging towards a solution. At each iteration, the adjustment variables ∂1_jk, ∂2_jk and the function f^T.z are calculated and can be used to decide the next vector solution X.

More precisely, at each iteration, the optimization algorithm balances the power differences with the adjustment variables ∂1_jk, ∂2_jk so that the equalities of the equation system are verified. Thus, if the power delivered on a feeder of index j shows a surplus, the corresponding adjustment variable ∂1_jk is increased and if this power shows a deficit, the corresponding variable ∂2_jk is increased. Thus, the more the connection parameters a_ij comply with the principle of conservation of power, the lower the adjustment variables are. The objective is to minimize the adjustment variables, resulting in the objective function

$f^T·z=\sum _{j=1}^m\sum _{k=1}^w\left(\partial 1_{\mathrm{jk}}+\partial 2_{\mathrm{jk}}\right).$

The matrices l_z and u_z bound the connection parameters a_ij between −ε and 1+ε and the adjustment variables ∂1_jk and ∂2_jk between 0 and +∞. More precisely, during the application of the optimization algorithm, the connection parameters a_ij are real numbers allowing the constraints to be relaxed. Then, following the application of the optimization algorithm, the determination software 48 is configured for setting the values of the connection parameters a_ij to 0 or to 1 on the basis of their value following the application of the optimization algorithm. The value 0 indicates a non-connection of the consumer of index i to the feeder of index j, while the value 1 indicates a connection of the consumer of index i to the feeder of index j.

The determination software 48 is, for example, configured for determining the values of the connection parameters following the application of the optimization algorithm via the following equation:

$\begin{array}{cc}{a}_{\mathrm{ij}}=\{\begin{array}{cc}1& \mathrm{if}a_{\mathrm{ij}}=\underset{j}{\max }\left(a_{\mathrm{ij}}\right)\\ 0& \mathrm{else}\end{array}& \left(7\right)\end{array}$

Thus, according to equation (7) above, each consumer 16_i is connected to a single feeder 18_j.

In addition, the processing unit 24 is configured for identifying, on the basis of the connection parameters determined and from, for example, identification software, not represented, comprised in the memory 28, subsets of consumers, with each subset of consumers that corresponds to the set of consumers 16_i connected to a same feeder 18_j.

As a variant, the determination software 48 is configured for determining the connection parameters separately for each feeder and independently of the second data relating to the electrical power acquired for the other feeders.

According to this variant, an optimal equation system is determined for each feeder, and the optimization algorithm is applied to each optimal equation system. Thus m optimal equation systems are obtained, solved independently via the optimization algorithm.

Thus, for the feeder of index 1, the corresponding optimal equation system verifies:

A1*z1=B1,   (8)

where

$z1=\left[\begin{array}{c}{a}_{11}\\ ⋮\\ a_{n1}\\ \partial 1_{11}\\ ⋮\\ \partial 1_{1w}\\ \partial 2_{11}\\ ⋮\\ \partial 2_{1w}\end{array}\right],B1=\left[\begin{array}{c}{E}_{D1}\left(t_1\right)\\ ⋮\\ E_{D1}\left(t_w\right)\end{array}\right]\mathrm{and}A1=\left[\begin{array}{ccc}U& I_w& -I_w\end{array}\right],$

with

$U=\left[\begin{array}{ccc}{E}_{C1}\left(t_1\right)& \dots & E_{\mathrm{Cn}}\left(t_1\right)\\ ⋮& \ddots & ⋮\\ E_{C1}\left(t_w\right)& \dots & E_{\mathrm{Cn}}\left(t_w\right)\end{array}\right].$

Then, according to this variant, the determination software 48 is, for example, appropriate for determining the values of the connection parameters via equation (7).

According to another variant, the determination software 48 is configured for determining the connection parameters for each electrical phase conductor 32_j, 34_j, 36_j and not simply for each feeder 18_j. According to this other variant, the equation system then comprises as many equations as electrical phase conductors 32_j, 34_j, 36_j and the variables described above and relating to a specific feeder j then relate to a specific phase conductor. Thus, the second sensors 22_j measure the electrical power delivered by each phase conductor 32_j, 34_j, 36_j and not each feeder 18_j, the connection parameters a_ij are determined for each phase conductor 32_j, 34_j, 36_j and the adjustment variables are determined for each phase conductor 32_j, 34_j, 36_j. According to this other variant, each electrical conductor is identified, for example, by an index and the variable j, set out in the equations above, then corresponds to an electrical phase conductor index varying from 1 to u, with u the number of electrical phase conductors which is equal to 3*m, or 6 in the example of FIG. 1.

Several embodiments of a method for determining the structure of the electricity transmission grid 10, implemented via the processing unit 24, and more generally via the determination system 12, will now be described.

According to a first embodiment described below with reference to FIG. 2, the method comprises an initial step 100 of acquiring main data for the transmission grid 10. The main data comprise, for example, the total number n of consumers 16_i, the total number m of feeders 18_j, the first data E_Ci(t_l) measured, the second data E_Dj(t_l) measured and the various time intervals t_l associated with the first and second data measured. Indeed, during the initial step 100, the acquisition software 41 controls, for example, the measurement, by each first 20_j and each second 22_j sensor, of the first E_Ci(t_l) and second E_Dj(t_l) data during the time intervals t_l and the transmission of the first E_Ci(t_l) and second E_Dj(t_l) data which are then associated with the time interval t_l during which they were measured, via, for example, the first and second information.

Subsequently, during a generating step 102, the generation software 42 generates a plurality of different classes of data Je_l, each associated with one of the time intervals t_l and which comprise the first E_Ci(t_l) and second E_Dj(t_l) data associated with said time interval t_l.

Then, during a calculating step 104, the calculation software 44 calculates the first selection criterion C1_l for each class of data Je_l. The first selection criterion C1_l is, for example, the overall rate of electrical power losses between the feeders 18_j and the consumers 16_i.

Subsequently, during a selecting step 106, the selection software 46 selects the first set En₁ on the basis of the first selection criteria C1_l calculated. The first set En₁ is selected from the classes of data generated in step 102. In the course of the selecting step 106, the selection software 46 compares, for example, the first criteria C1_l with the first predetermined variable V1, the value of the first predetermined variable V1 being defined, for example, during the acquiring step 100. The selection software 46 then selects the classes of data Je_l for which the first criterion C1_l is less than the first predetermined variable V1.

Finally, during a determining step 108, the connection parameters are determined via the determination software 48 and from the first selected set En₁.

More precisely, the determination software 48 determines the equation system to be solved from, for example, equation (4), as set out above. Then, as set out above during the description of the determining system 12 and equation (5), the optimization algorithm is applied to the equation system in order to determine the connection parameters a_ij.

As a variant, during the determining step, the determination software 48 determines the optimal equation systems, as set out above via equation (8), and applies the optimization algorithm to each optimal equation system. The speed of determining the connection parameters a_ij is then improved, since the optimal equation systems comprise a limited number of equations.

In the first embodiment, the fact of selecting classes of data Je_k eliminates the classes of data Je_l for which the losses are the greatest, since these classes of data risk leading to an erroneous determination of the connection parameters a_ij.

Thus, according to the first embodiment, the determined connection parameters a_ij, are determined with more certainty compared with the methods known in the prior art and the reliability of the method of determination is improved.

In addition, the optimization algorithm used enables any number of classes of data Je_k to be used during the determining step 108, even if it is preferable that the number of classes of data of the first set is greater than or equal to the total number n of consumers 16_i.

According to a second embodiment of the invention described below with reference to FIG. 3, the method comprises steps 200, 202, 204, 206, 208 respectively identical to steps 100, 102, 104, 106, 108 of the first embodiment and, during step 204, a second selection criterion C2_l is calculated for each class of data generated. The second selection criterion C2_l is different from the first criterion C1_l and is chosen from an overall rate of electrical power losses between the feeder or feeders and the consumers and a difference in consumption of electrical power between the various consumers.

Thus, during step 204, each first criterion C1_l is, for example, the overall rate of loss relating to the corresponding class of data Je_l and each second criterion C2_l is the difference in electrical consumption relating to the corresponding class of data Je_l, and is calculated via equation (2) or equation (3).

Then, in the course of the selecting step 206, the selection software 46 selects the first set En₁ on the basis of the first selection criteria C1_l and the second selection criteria C2_l calculated. The first set En₁ is selected from the classes of data generated in step 202. In the course of the selecting step 206, the selection software 46 compares, for example, the first criteria C1_l with the first variable V1 and the second criteria C2_l with a second predetermined variable V2. The value of the second predetermined variable V2 is, for example, defined during the acquiring step 200. The selection software 46 then selects the classes of data for which the first criterion C1_l is less than the first predetermined variable V1 and the second criterion C2_l is less than the second predetermined variable V2.

Finally, in the course of the determining step 208 the connection parameters are determined via the determination software 48 and from the first selected set En₁.

The second embodiment makes it possible to refine the selection of the classes of data compared with the first embodiment and thus to select classes of data from which the risk of error in determining the connection parameters a_ij is minimized. The accuracy, speed and reliability of the method of determination are thus improved.

According to a third embodiment of the invention described below with reference to FIG. 4, the method comprises steps 300, 302, 304, 306, 310 respectively identical to steps 100, 102, 104, 106, 108 of the first embodiment and, following step 306 and prior to step 310, the method comprises a step 307 of calculating a second selection criterion C2_k for each class of data of the first set En₁, the second criterion C2_k being different from the first criterion and being chosen from an overall rate of electrical power losses between the feeder or feeders 18_j and the consumers 16_l and a difference in electrical power consumption between the different consumers 16_i.

In the third embodiment, each first criterion C1_l is, for example, the overall rate of loss relating to the corresponding class of data generated in step 302 and each second criterion C2_k is the difference in electrical consumption relating to the corresponding class of data of the first set En₁.

Following the calculating step 307, during a selecting step 308, a second set En₂ of classes of data is selected from the first set En₁. Then, the first set En₁ is, for example, set equal to the second selected set En₂ for performing the determining step 310 on the basis of the second selected set En₂.

The third embodiment makes it possible to refine the selection of the classes of data compared with the first embodiment and thus to select classes of data from which the risk of error in determining connection parameters a_ij is minimized. The accuracy, speed and reliability of the method of determination are thus improved.

According to a fourth embodiment of the invention described below with reference to FIG. 5, the method comprises steps 400, 402, 404, 406 respectively identical to steps 100, 102, 104, 106 of the first embodiment.

Then, following the selecting step 406, the method comprises a step 408 for determining the connection parameters. More precisely, during step 408, the method comprises a first substep 408A for pseudo-randomly selecting a third set En₃ of classes of data included in the first set En₁.

Subsequently, during a calculating substep 408B, a repetition parameter R1 is calculated. The repetition parameter R1 corresponds to a number of iterations of the selecting substep 408A.

Then, during a determining substep 408C, intermediate connection indices b_ij, indicating for each consumer 16_i the feeder 18_j to which it is connected, are determined from the third selected set En₃. More generally, the intermediate indices b_ij comprise for each electrical consumer 16_i an identifier of the feeder 18j to which it is connected. The determining substep 408C is similar to the determining step 108 of the first embodiment, but is performed from the third set En₃.

Subsequently, during a storing substep 408D, the intermediate connection indices b_ij are stored by the memory 28.

Then, in the course of a calculating substep 408E, performed for each consumer 16_i and for each feeder 18_j, an assignment number NA_ij relating to the number of times that the stored intermediate connection indices indicate that the consumer 16_i is connected to the feeder 18_j is calculated.

The assignment number NA_ij is, for example, calculated via the following formula:

${\mathrm{NA}}_{\mathrm{ij}}=\frac{N\left(b_{\mathrm{ij}}=1\right)}{N_T},$

where N(b_ij=1) corresponds to the number of times that the stored intermediate indices b_ij are equal to 1 for the consumer 16_i of index i and the feeder 18_j of index j and N_T corresponds to a total number of iterations of the selecting substep 408 A. Advantageously, the total number of iterations N_T is initialized to 0 prior to the execution of the selecting substep 408A and incremented by 1 for each execution of the selecting substep.

It should be noted that the number of intermediate connection indices stored for a given electricity consumer 16_i and a given feeder 18_j is equal to the number of iterations of the selecting substep 408A.

Advantageously, the assignment numbers calculated are stored following the calculating substep 408E.

Then, during a comparing substep 408F, the repetition parameter R1 is compared to a predetermined repetition criterion CR1. The predetermined repetition criterion CR1 is, for example, initialized during the acquiring step 400.

The repetition criterion CR1 is, for example, a minimum number of iterations of the selecting substep 408A.

If during the comparing substep 408F, the repetition parameter is less than the repetition criterion, then the steps of selecting 408A, calculating 408B, determining 408C, storing 408D, calculating 408E and comparing 408F are repeated.

If during the comparing substep 408F, the repetition parameter R1 is greater than the repetition criterion CR1, then a substep 408G is performed for determining the connection parameters a_ij.

During the determining substep 408G, the connection parameters a_ij are determined from the third selected set or sets En₃, and more precisely on the basis of the intermediate connection indices b_ij stored during the storing substep 408D, and more precisely on the basis of the assignment numbers NA_ij calculated at the last iteration of the substep 408E. More precisely, for each electricity consumer 16_i, the feeder index j corresponding to the largest assignment number NA_ij is identified, and the connection parameter a_ij corresponding to said consumer 16_i and to said feeder 18_j is set equal to 1, the other connection parameters a_ij relating to said consumer being set equal to 0.

Advantageously, if during the determining substep 408G all the assignment numbers relating to a consumer 16_i are less than a first predetermined threshold S1, e.g. equal to 0.6, then an error in identifying the connection for said consumer 16_i is identified.

As a variant, if the number of iterations of the substep 408A is greater than 2, then, during the calculating substep 408E, and at each iteration of the substep 408E, a first mean of the assignment numbers NA_ij, calculated at each iteration of the substep 408E, for each consumer 16_i and for each feeder 18_j, is calculated. Then, a second mean of the assignment numbers NA_ij, calculated at the last iteration, for each consumer 16_i and for each feeder 18_j, is calculated. Subsequently, during the substep 408 E, a difference between the first mean and the second mean is calculated.

The first and the second mean are, for example, arithmetic, geometric, or quadratic, etc. means.

According to this variant, during the comparing substep 408F, the absolute value of the last difference calculated is compared to a second predetermined threshold S2, e.g. equal to 0.1. Then, if during the comparing substep 408F, the repetition parameter R1 is greater than the repetition criterion CR1 and the absolute value of the last difference calculated is less than the second threshold S2, then the substep 408G is performed. Else, steps 408A, 408B, 408C, 408D, 408E and 408F are reiterated.

The fourth embodiment makes it possible, notably when the number of iterations of the substep 408A is greater than 2, to determine the connection parameters a_ij from third sets En₃ of different classes of data. Thus, the accuracy and reliability of the method of determination are improved.

In addition, the fourth embodiment advantageously makes it possible to identify each consumer 16_i for which the associated feeder 18_j is determined with a good index of confidence and each consumer 16_i for which the associated feeder 18_j is undetermined or determined with a poor index of confidence. Indeed, as described above, if during the determining substep 408G all the assignment numbers relating to a consumer 16_i are less than the first predetermined threshold S1, e.g. equal to 0.6, then an error in identifying the connection for said consumer 16_i is identified and the feeder 18_j to which the consumer 16_i is connected is undetermined. In the same way, if during the determining substep 408G an assignment number relating to a consumer 16_i is greater than a third predetermined threshold S3, e.g. equal to 0.95, then the connection of the consumer 16_i to the corresponding feeder 18_j is identified with a good index of confidence.

Advantageously again, the fourth embodiment makes it possible to associate, with each connection parameter a_ij, set equal to 1 during the determining substep 408G, an index of confidence representing the probability that the determined connection is correct. The index of confidence is, for example, equal to the corresponding assignment number NA_ij.

The embodiments and variants envisaged above are appropriate for being combined with each other, according to all technically acceptable combinations, for giving rise to other embodiments of the invention.

Thus, the second embodiment is appropriate for being combined with the fourth embodiment and the third embodiment is also appropriate for being combined with the fourth embodiment.

Claims

1. A method for determining the structure of an electricity transmission grid, the method being implemented by computer, the transmission grid comprising a power supply station comprising one or more feeders for supplying electrical power to a plurality of electricity consumers, the method comprising: a) acquiring, for each electricity consumer, first data (ECi(tl)) relating to the electrical power consumed by the electricity consumer during various time intervals (tl), measured via a first sensor for measuring power;b) acquiring, for each feeder, second data (EDj(tl) relating to the electrical power delivered by the feeder during the various time intervals (tl), measured via a second sensor for measuring power;c) generating, via a generating module, a plurality of different classes of data (Jel), each associated with one of the time intervals, each class of data (Jel) comprising the first (ECi(tl)) and second (EDj(tl)) data associated with said time interval;d) calculating a first selection criterion (C1l) for each class of data, the first selection criterion (C1l) being a criterion chosen from an overall rate of electrical power losses between the feeder or feeders and the consumers and a difference in consumption of electrical power between the various consumers;e) selecting, by a selection unit and on the basis of the first selection criteria (C1l) calculated, a first set (En1) of classes of data (Jek) from the classes generated (Jel); andf) determining, from the first selected set (En1), connection parameters (aij), said parameters (aij) comprising for each electricity consumer an identifier of the feeder to which it is connected.

2. The method according to claim 1, wherein, during the calculating step d), a second selection criterion (C2l) is calculated for each class of data (Jel) generated, the second criterion (C2l) being the other criterion chosen from an overall rate of electrical power losses between the feeder or feeders and the consumers and a difference in consumption of electrical power between the various consumers, andwherein, during the selecting step e), the first set (En1) is selected on the basis of the first (C1l) and second (C2l) criteria calculated.

3. The method according to claim 1, wherein following the selecting step e), the method comprises: e1) calculating a second selection criterion (C2l) for each class of data (Jek) of the first set (En1), the second criterion (C2l) being the other criterion chosen from an overall rate of electrical power losses between the feeder or feeders and the consumers and a difference in consumption of electrical power between the various consumers; ande2) selecting on the basis of the second selection criteria (C2l) calculated, a second set (En2) of classes of data from the first set (En1), determining connection parameters (aij) then being performed during the determining step f) on the basis of the second set (En2) selected from the first set (En1).

4. The method according to claim 1, wherein, during the determining step f), the connection parameters (aij) are determined separately for each feeder and independently of the second data (EDj(tl)) relating to the electrical power acquired for the other feeders.

5. The method according to claim 1, wherein, during the determining step f), an equation system to be solved is determined from a premise of conservation of electrical power for each feeder, according to which the power delivered by the feeder is substantially equal to the sum of the power consumed by the electricity consumers connected to the feeder and the electrical losses.

6. The method according to claim 5, wherein, during the determining step f), the equation system is defined from the following equation:

7. The method according to claim 6, wherein the equation system is written in matrix form as follows: A*z=B, where

8. The method according to claim 7, wherein, during the determining step f), an optimization algorithm is used for solving the equation system, the optimization algorithm verifying:

9. The method according to claim 1, wherein each feeder comprises one or more electrical phase conductors, and wherein, during the acquiring step b), the second data (EDj(tl)) are acquired for each electrical phase conductor and relate to the electrical power delivered by each electrical phase conductor and, during the determining step f), the connection parameters (aij) determined comprise for each electricity consumer an identifier of the electrical phase conductor or conductors to which it is connected.

10. The method according to claim 1, wherein, during the determining step f), the method comprises: f1) pseudo-random selecting of a third set (En3) of classes of data included in the first set (En1); andf2) determining connection parameters (aij) from the third set (En3) selected from the first set (En1).

11. The method according to claim 10, wherein, following the selecting step f1) and prior to the determining step f2), the method comprisesf3) calculating a repetition parameter;f4) determining, from the third selected set (En3), intermediate connection indices (bij) comprising for each electricity consumer an identifier of the feeder to which it is connected;f5) storing the intermediate connection indices (bij);f6) comparing the repetition parameter with at least one predetermined repetition criterion (CR1); andfollowing the result of the comparing step f6), performing a step from, on the one hand, the repetition of the steps of selecting f1), calculating f3), determining f4), storing f5) and comparing f6) and, on the other hand, the step of determining f2) the connection parameters (aij) on the basis of the stored intermediate connection indices (bij).

12. The method according to claim 11, wherein the method further comprises:f7) calculating, for each consumer and for each feeder, an assignment number (NAij) relating to the number of times that the stored intermediate connection indices (bij) indicate that the consumer is connected to the feeder, andwherein, during the determining step f2), the connection parameters (aij) are determined from the assignment numbers (NAij) calculated.

13. A computer program comprising software instructions, which when executed by a computer, implement the method according to claim 1.

14. A system for determining the structure of an electricity transmission grid, the transmission grid comprising a power supply station comprising one or more feeders for supplying electrical power to a plurality of electricity consumers, the system comprising: for each electricity consumer, a first sensor for measuring first data (ECi(tl)) relating to the electrical power consumed by the consumer during various time intervals (tl);for each feeder, a second sensor for measuring second data (EDj(tl)) relating to the electrical power delivered by the feeder during the various time intervals (tl);a module for acquiring first (ECi(tl)) and second (EDj(tl)) data;a module for generating a plurality of different classes of data (Jel), each associated with one of the time intervals (tl), each class of data (Jel) comprising the first (ECi(tl)) and second (EDj(tl) data associated with said time interval (tl);a module for calculating a first selection criterion (C1l) for each class (Jel) of data, the first selection criterion (C1l) being a criterion chosen from an overall rate of electrical power losses between the feeder or feeders and the consumers and a difference in consumption of electrical power between the various consumers;a module for selecting, on the basis of the first selection criteria (C1l) calculated, a first set (En1) of classes of data (Jek) from the classes generated; anda module for determining, from the first selected set (En1), connection parameters (aij), said parameters (aij) comprising for each electricity consumer an identifier of the feeder to which it is connected.