Joint voltage regulation process in a medium voltage HTA branch and low voltage LV branches including a three-phase low voltage LV branch

Abstract

A method for jointly regulating a voltage in a medium voltage (MV) branch and in low voltage (LV) branches of a distribution network comprising energy generators, at least one low voltage branch being in three-phase mode, the method comprising: - a comparison: ° of the effective voltages of each low medium voltage (LV, HVA) branch with safety ranges, the effective voltage state of the HVA medium voltage branch having been obtained after a step of aggregating the data relating to the LV low voltage branches at their connection node, - if an effective voltage of the low voltage (LV) branch and / or the medium voltage (HTA) branch is outside the safety ranges: ° search for a set of joint commands, the joint regulation being preceded by: ° the construction of an admittance matrix for the three-phase LV low voltage branches. Abstract figure: Figure 2

Description

Title of the invention: Method for jointly regulating the voltage in a medium voltage HTA branch and low voltage LV branches including a three-phase low voltage LV branch Technical field

[0001] The present disclosure relates to the field of methods for jointly regulating a voltage in a medium voltage branch and in a plurality of low voltage branches, at least one of which is in three-phase mode, of an electricity distribution network connected thereto. Prior art

[0002] The purpose of managing a distribution network is to optimally control its operation in order to minimize losses, control voltage fluctuations, and thus guarantee continuity of service under reasonable economic conditions. However, the management of the electricity distribution network is set to change profoundly due to numerous factors. These factors include in particular the development of new decentralized means of producing carbon-free electricity, based on the use of renewable primary energy sources, the evolution of electricity uses, for example with the appearance of electric vehicles, the evolution of consumption patterns, for example by deferring certain consumption to reduce the energy bill or the appearance of new players on the energy market.

[0003] Decentralized production means, or "dispersed energy generators GED", can be managed by independent producers. Decentralized production means are characterized by the fact that they are of limited power and often connected to the distribution network. By being located as close as possible to the consumption points, these means make it possible to reduce the cost of energy transport, active power transits at the source substation and thus line losses on the transmission network. However, the connection of decentralized production means to the distribution network poses new operating problems, in particular the difficulty of maintaining the voltage plan, i.e. voltage values ​​within the regulatory ranges.

[0004] Thus, in the context of the energy transition, these developments induce difficulties leading to the establishment of new functions for controlling the medium and low voltage electricity distribution network. One of the main difficulties is voltage control and regulation. However, as distributed generators such as wind turbines and photovoltaics become increasingly widespread and As electric vehicles and energy storage systems develop, the growing number of connections to distribution networks leads to increasing voltage problems, due to the high variability of injections and withdrawals that this causes. Insufficient voltage control and regulation can affect the efficiency, capacity and performance of the distribution network.

[0005] Currently, methods are known for determining the voltage state in order to regulate the voltage of single-phase low-voltage branches of a distribution network, as described in document FR3084168. Methods are also known for determining the voltage state in order to regulate the voltage of medium-voltage branches of a distribution network, as described in document FR3006819.

[0006] Although the methods described in these documents make it possible to determine the respective voltage state of the low voltage LV branches and of a medium voltage HVA branch in a reliable manner, they do not take into account the electrical interactions between these branches.

[0007] Also known from document CN107196316B is a method for regulating the voltage in a distribution network, the use of which is limited to photovoltaic inverters, with coordinated control of the voltage regulator system in charge, the selection of capacitor banks and an adjustment of the active and reactive powers of the photovoltaic installations. However, this document proposes an approach that is imprecise and specific to photovoltaic generators.

[0008] Thus, existing methods present partial and limited solutions. They are in fact relative either to particular configurations, for example adapted to a single type of generator, or only to a low voltage branch or only to a medium voltage branch.

[0009] There is therefore a need for a method for regulating the voltage on a distribution network, a method which can be implemented on any distribution network comprising medium and low voltage branches, and taking into account the interactions between these branches, as well as the nature of these branches, between single-phase, balanced three-phase or unbalanced three-phase branches. Summary

[0010] The present disclosure improves the situation.

[0011] A method is proposed for jointly regulating a voltage at each node of a distribution network comprising a medium voltage, HTA, branch and at least one, and preferably a plurality of low voltage, LV, branches, the distribution network further comprising a plurality of energy generators, each low voltage branch being connected to the medium voltage branch at a connection node at an HTA / LV transformer, each low voltage branch further comprising a reference node, a plurality of controllable nodes and a plurality of non-controllable nodes, the medium voltage branch being supplied by a high voltage branch attached to the medium voltage branch at an attachment node at an HTB / HTA transformer with load regulator, the medium voltage branch further comprising a reference node, a plurality of controllable nodes and a plurality of non-controllable nodes, each node of the medium voltage and low voltage branches being subjected to energy withdrawals and / or injections, said at least one branch, or at least one among the plurality of low voltage branches being in three-phase mode, the method comprising: - a definition: °a target voltage at each node of each low voltage branch, °a first prescribed voltage range surrounding the target voltage of each low voltage branch; °of a target voltage at each node of the medium voltage branch, ° of a second prescribed voltage range surrounding the target voltage of the medium voltage branch; and the method comprising, at each step of a predetermined time: - a determination of a voltage state of the low voltage branches - a determination of a voltage state of the medium voltage branch taking into account electrical interactions of the low voltage branches attached to it - then, a comparison: °of the actual voltage state of each low voltage branch with the first prescribed voltage range ° of the actual voltage state of the medium voltage branch with the second prescribed voltage range; - if at least one voltage of the low voltage branch and / or the medium voltage branch is outside the first prescribed voltage range and / or the second prescribed voltage range respectively: - a search for a set of joint regulation commands allowing the effective voltage of the low voltage branch and the medium voltage branch to be included respectively in the first prescribed voltage range and in the second prescribed voltage range: - a determination and implementation of at least one optimal regulation command from among the set of joint regulation commands characterized in that the search for the set of joint regulation commands for each three-phase low voltage branch is preceded by a construction of a three-phase admittance matrix comprising, for each low voltage branch three-phase: ° creation of a list of nodes for each of the three phases of the three-phase low voltage branch by carrying out: *a tripling of each node of the low voltage branch, one node of each triplet of nodes being respectively associated with one of the three phases, * a correspondence of indices of the nodes of each of the three phases, ° obtaining three admittance sub-matrices corresponding respectively to an admittance matrix of each of the three phases of the LV branch, ° a reconstruction of a three-phase admittance matrix by a concatenation of the three admittance sub-matrices taking into account a corrective admittance illustrating losses at the level of the HTA / LV transformer of the LV branch considered.

[0012] The proposed regulation method applies to any distribution network, regardless of its tree topology, thus covering urban, semi-rural or rural networks. The three-phase nature of the low-voltage branches is taken into account so that the proposed method takes into account the realities of the field and makes it possible to obtain more precise and realistic results.

[0013] According to another aspect, there is provided a computer program comprising instructions for implementing all or part of a method as defined herein when this program is executed by a processor or several processors. According to another aspect, there is provided a non-transitory, computer-readable recording medium on which such a program is recorded.

[0014] The features set out in the following paragraphs may optionally be implemented. They may be implemented independently of one another or in combination with one another:

[0015] Determining said at least one optimal regulation command to be implemented from among the set of regulation commands comprises - obtaining a hierarchical list of regulation commands according to a decreasing order of priority of implementation, - a simulation of joint regulation of the HTA branch and the LV branches by activating a first regulation command from among the set of joint regulation commands, the first activated command corresponding to a first command on the hierarchical list, and, if the voltage state of the LV and HTA branches is not included in the first prescribed voltage range and / or in the second prescribed voltage range, ° iteratively carry out other joint regulation simulations by adding, at each iteration, the following regulation command in the hierarchical list, until the voltage states of the low voltage branches and the medium voltage branch are respectively included in the prescribed voltage prescribed voltage range and in the second safety range, - when the voltage states of the low voltage branches and the medium voltage branch are respectively within the first prescribed voltage range and the second prescribed voltage range, obtaining at the output of said at least one optimal regulation command to be implemented.

[0016] Thus, this step allows the selection of a minimum and sufficient number of optimal commands to be implemented. Indeed, as soon as the voltage states of the low voltage LV branches and the medium voltage HVA branch enter their respective prescribed voltage range, the iterative selection stops and only the selected commands are actually implemented. This step is particularly advantageous since there may be constraints that do not allow the activation of all the regulation commands.

[0017] The search for the set of joint regulation commands for each low voltage branch and for the HVA branch is preceded by a step of calculating regulation command values ​​comprising: - an iterative resolution, using a fixed point technique, of an optimization problem taking into account: ° constraints relating to the admissible domains of active and reactive powers at the controllable nodes, ° constraints relating to the three-phase configuration of the low-voltage branch, each iteration including a resolution of a linear program integrating a linearization of the Kirchhoff equations, the program returning values ​​of the regulation commands as output, the regulation commands including: ° an adjustment of a voltage value at the connection node of a low voltage branch to the HVA branch for an on-load adjustment of the HVA / LV transformer, when the adjustment is available, ° an adjustment of a voltage value at the connection node of the medium voltage branch for an on-load adjustment of the HTB / HTA transformers, °an active power adjustment (P), in medium voltage and in low voltage, °an reactive power adjustment, in medium voltage and in low voltage, and / or °an active and reactive power adjustment, in medium voltage and in low voltage.

[0018] The steps for determining the values ​​of the regulation commands therefore take into account the three-phase nature of certain low-voltage branches, which makes it possible to obtain more reliable regulation that takes into account the reality on the ground.

[0019] When an admissible range of active and reactive powers at the nodes com- commandable is not convex, the method includes - a convexification of the initial non-convex domain, - following the resolution of the optimization problem, a verification that the adjustment values ​​of the active and reactive powers (P, Q) are in the initial domain and, if this is not the case, * a substitution of the adjustment values ​​of the active and reactive powers by values ​​included in the initial domain, to have a solution respecting the initial domain of adjustment of the active and reactive powers.

[0020] This step makes it possible to broaden the application of the joint regulation method to any type of equipment, taking into account in particular the non-convexity of the adjustment domains of the active and reactive powers of this equipment. Among this equipment are in particular wind equipment.

[0021] The search for the set of joint regulation commands for the low voltage branches and the HVA branch includes: - respectively, a definition of a set of regulation commands for each low voltage LV branch and the medium voltage HVA branch, each set of regulation commands comprising all possible regulation commands, - for each low voltage branch, a first step of simulating a regulation by implementing all the regulation commands of the low voltage branch considered, - a determination of the voltage state of each low voltage branch and, when the voltage state of each LV branch enters the first prescribed voltage range: ° an aggregation of data relating to each low voltage branch at the level of the respective connection node of each low voltage branch to the HVA branch, ° a simulation of regulation of the medium voltage HTA branch by activating all the regulation commands of the HTA branch, taking into account the aggregated data, ° a calculation of the voltage values ​​at the reference nodes of the aggregated low voltage branches, °and for each low voltage branch not having a load regulator: *the value of the voltage at the reference node of the low voltage branch is imposed on a value of the voltage at the connection node of the low voltage branch to the medium voltage branch, multiplied by a transformation ratio of the transformer, the method then comprising the iterative implementation of the following steps until a first stopping criterion is met:, H a modification of all the controls of each low voltage branch BT considered by removing a command corresponding to a load setting of the transformer, H another step of regulation simulation of each low voltage branch not having a regulator on load, by activating all the regulation commands of each low voltage LV branch considered, Hthen, obtaining each voltage state of the low voltage branches not having a regulator on load, H a return to the data aggregation stage,* if a low voltage branch has a regulator on load, the set of regulation commands comprises all possible regulation commands, - the method then comprising the calculation of the voltage states of the medium and low voltage branches (HTA, BT), following the respective activation of the sets of commands of each BT low voltage branch and of the HTA medium voltage branch, - and, when the voltage states of the medium and low voltage branches respectively fall within the first prescribed voltage range and the second prescribed voltage range, obtaining each set of regulation commands corresponding to each low voltage branch and for the medium voltage branch, said sets of commands forming the set of joint regulation commands.

[0022] This part of the method makes it possible to determine the sets of joint regulation commands, from the activation of all possible commands.Thus, in a first step, it is possible to determine whether there is a joint regulation solution by activating all possible regulation commands. Indeed, if the voltage states are not compliant even by activating all the commands, then there is no command allowing regulation at the given time step of implementation of the process. In a second step, this part makes it possible to discard the commands which are not available for regulation, in particular the regulator in charge of the HTA / LV transformer.

[0023] If, following the determination of the voltage state of each low voltage branch, at least one voltage state of a low voltage branch is outside the first prescribed voltage range, an alarm is activated for said at least one LV low voltage branch, and / or -If, following the calculation of the voltage states of the medium and low voltage branches (HTA, BT), the voltage state of at least one BT low voltage branch or the voltage state of the HTA medium voltage branch is outside the first prescribed voltage range and / or the second prescribed voltage range respectively, an alarm is activated.

[0024] The alarm is activated to indicate that there is no solution for joint regulation at the current time step. It is advantageous to determine as early as possible that such a solution does not exist. Thus, a first alarm can be triggered as soon as it is determined that at least one voltage state of a low voltage LV branch is not compliant following the activation of all the regulation commands.

[0025] Determining the voltage status of the low voltage LV branches and the medium voltage HVA branch includes: - a determination of the voltage status of each low voltage LV branch, - then, a step of aggregating data relating to each low voltage branch at the level of the respective connection node of each low voltage branch to the HVA branch in * calculating a pair of active and reactive power values ​​of each low voltage branch, each pair of active and reactive powers being respectively added to active and reactive power values ​​at the connection nodes of each LV branch to the HVA branch, and in Correcting each diagonal term of an admittance matrix of the HTA medium voltage branch, each diagonal term corresponding respectively to each node of connection of a LV branch to the HTA branch,: - following the aggregation step, obtaining the effective voltage state at each node of the HTA branch and the LV low voltage branches.

[0026] The strong interactions between the HTA medium voltage branch and the LV low voltage branches attached to it are taken into account using the aggregation step. The aggregation corresponds, in particular, to an aggregated representation of the electrical influence of the LV low voltage branches in the HTA medium voltage branch. Thus, the determination of the voltage state of the HTA medium voltage branch and the LV low voltage branches takes into account the electrical interactions between these branches, subject to known injections / withdrawals. The voltage state is made on each LV low voltage branch attached to the HTA medium voltage branch considered, then each LV low voltage branch is aggregated into an equivalent node representing the power at this node seen from the HTA medium voltage branch.Thus aggregated, the powers on all the nodes of the HTA medium voltage branch are known and the voltage state can be evaluated using any known technique, taking into account the interactions between the branches.

[0027] The effective voltage states at each node of the HTA branch and the LV branches are obtained by an iterative application of the following steps, until a first predefined stopping criterion is satisfied: °a determination of an effective voltage state of the medium voltage branch, taking into account the aggregated data of each low voltage branch, to obtain effective voltages at each node of the medium voltage branch, “then, a determination of an effective voltage state of each low branch voltage to obtain effective voltages at each node of each low voltage branch, from which new values ​​of the active and reactive power couples of each low voltage branch and the diagonal terms of the admittance matrix of the HTA branch are obtained - obtaining the effective voltage state at each node of the HTA branch and the LV branches when the first shutdown criterion is satisfied.

[0028] The iterative application of aggregation allows for a finer and more precise evaluation of the voltage state.

[0029] The method is implemented on a computer having a multi-core architecture.

[0030] In this way, when the networks considered are of large size, some Calculations can be implemented in parallel, so that regulation can be performed at each step of a predetermined time. For example, this time step is 10 minutes. Brief description of the drawings

[0031] Other characteristics, details and advantages will appear on reading the detailed description below, and on analyzing the attached drawings, in which: Fig.l

[0032] [Fig.l] shows a simplified diagram of a distribution network according to one embodiment. Fig. 2

[0033] [Fig.2] shows steps of a regulation method according to one embodiment. Fig. 3

[0034] [Fig.3] shows an example of authorized voltage variation ranges on a LV branch. Fig. 4

[0035] [Fig.4] shows another example of authorized variation ranges of the voltage on an HTA branch. Fig. 5

[0036] [Fig.5] shows an example of determining the voltage states of the low voltage LV branches and the medium voltage HVA branch. Fig. 6

[0037] [Fig.6] shows the construction of the linear program allowing the calculation of the regulation commands. Fig. 7

[0038] [Fig.7] shows the construction of an admittance matrix of a three-phase LV low voltage branch taking into account the losses of the HTA / LV transformer. Fig. 8

[0039] [Fig.8] shows a convex adjustment domain of the active and reactive powers of an installation. Fig. 9A

[0040] [Fig.9A] shows a non-convex adjustment domain of the active and reactive powers of an installation. Fig. 9B

[0041] [Fig.9B] shows the convexification of the adjustment domain of [Fig.9A]. Fig. 9C

[0042] [Fig.9C] shows obtaining a solution for the domains illustrated in Figures 9A and 9B.

[0043] Fig. 10

[0044] [Fig. 10] shows steps for obtaining a set of regulation commands for the low voltage LV branches and the medium voltage HVA branch. Fig. 11

[0045] [Fig. 11] shows steps for obtaining optimal and sufficient joint regulation commands. Fig. 12

[0046] [Fig.12] illustrates the effects of the steps illustrated in [Fig.11]. Fig. 13A

[0047] [Fig. 13A] illustrates an example of a single-core architecture of a machine on which the process is implemented. Fig. 13B

[0048] [Fig.l3B] illustrates an example of a multi-core architecture of a machine on which the method is implemented. Fig. 14

[0049] [Fig. 14] shows an example of application of joint regulation commands on a distribution network shown schematically according to one embodiment. Description of the embodiments

[0050] [Fig.l] illustrates an example of configuration of an electricity distribution network.

[0051] The distribution network may comprise a medium voltage HTA branch, connected to and supplied by a high voltage HTB branch at an HTB / HTA transformer which may comprise load regulators. The medium voltage branch comprises a plurality of controllable and / or non-controllable nodes. The medium voltage branch further comprises a reference node and at least one attachment node to which at least one low voltage branch is connected, and preferably a plurality of low voltage branches.

[0052] In the non-limiting example of [Fig. 1], the distribution network comprises two low voltage branches BTr, with r the index of the low voltage branch where, here, r = 1 and r = 2. Each low voltage branch comprises a plurality of controllable and / or non-controllable nodes as well as a reference node and a node for connection to the medium voltage branch HTA.

[0053] Several low voltage LV branches can be connected at the same connection node to the medium voltage HVA branch.

[0054] All controllable and non-controllable nodes are subject to injections and / or withdrawals of electricity, directly influencing the general voltage state of the distribution network but also the local voltage state of each low voltage and / or medium voltage branch.

[0055] The regulation method described below applies, advantageously, to any type of distribution network, whatever its tree topology, for example in urban, semi-rural or rural areas.

[0056] Furthermore, the regulation method applies, advantageously, to any balanced three-phase medium voltage branch, which can be modeled by a so-called “equivalent single-phase” diagram, and to single-phase, balanced three-phase and / or unbalanced three-phase low voltage LV branches.

[0057] Also, the regulation method makes it possible to cover all types of power injection equipment connectable to the medium and low voltage distribution network, in particular electricity generators of renewable or non-renewable origin, including wind turbines, hydro turbines, photovoltaics, biomass, cogeneration, hydraulics. The regulation method also covers all types of possible withdrawals from a medium and low voltage distribution network, including in particular usual consumption or even air conditioning and the recharging of electric vehicles. The method also covers all systems that can either inject or withdraw from the network, such as storage systems, including batteries, including electric vehicle batteries and flywheels.Finally, the method also covers all types of users, such as non-controllable single-phase consumers, controllable single-phase consumers, non-controllable three-phase consumers, controllable three-phase consumers, non-controllable single-phase producers, non-controllable three-phase producers, controllable single-phase producers and controllable three-phase producers.

[0058] [Fig.2] illustrates some of the main steps of a method for jointly regulating the voltage in a medium voltage HTA branch and the plurality of low voltage LV branches, of which at least one low voltage branch is three-phase.

[0059] More specifically, the joint voltage regulation method aims to maintain the voltage at each node of the network composed of a medium voltage HTA branch and its low voltage branches in prescribed voltage ranges. The prescribed voltage ranges can be chosen according to any criterion. The proposed method applies regardless of the values ​​of the prescribed voltage ranges.

[0060] Typically, in France, for medium voltage HTA branches, the target voltage to be maintained may be equal to 20 kV, while for low voltage LV branches, the target voltage may typically be equal to 230 V. However, these values ​​are given for information purposes because other target voltage values ​​may be defined contractually for each user of the distribution network. Similarly, these values ​​may vary from one country to another.

[0061] In addition, it is possible to process different contractual voltages among users, thus resulting in different target voltage values ​​for some of the nodes.

[0062] Since the target voltage cannot always be reached exactly, a percentage deviation to be respected can be established.

[0063] For example, the percentage deviation may be of the order of +10% and -10% around the target voltage value for LV low voltage branches. These values ​​are given as a non-limiting example. These values ​​may vary from one country to another.

[0064] Similarly, the percentage deviation can be of the order of +5% and -5% around the target voltage value for the HTA medium voltage branch. These values ​​are given as a non-limiting example. These values ​​may vary from one country to another.

[0065] The deviation percentages may be chosen to be greater or less than the examples presented above. The deviation percentages may even, possibly, have a value substantially equal to zero.

[0066] The percentages of deviation around the target voltage correspond to the prescribed voltage ranges.

[0067] In step SI, the fixed data relating to each low voltage LV branch attached to the medium voltage HVA branch are collected, as well as the fixed data relating to the medium voltage HVA branch.

[0068] Concerning the single-phase LV low-voltage branches, the fixed data include in particular the target values ​​of the voltage at the nodes of each low-voltage branch. These values ​​are real and positive. These are the ideal values ​​of the moduli of the voltages at the nodes, which must be achieved by the regulation. These values ​​are in the form of a vector Vtarget, of dimension equal to the number of nodes. This corresponds to the target voltage state.

[0069] In the remainder of the description, the terms “LV target voltage” or “LV target voltage vector” are used interchangeably to refer to the vector of target voltages corresponding to the target voltage values ​​of each low voltage branch.

[0070] The fixed data also includes an admittance matrix Y of each low voltage branch and the reference node, the “slack bus” node, of each low voltage branch. The reference node is the junction node of the low voltage LV branch to the medium voltage HVA branch and is designated by the reference “ref”.

[0071] The fixed data also includes the list of indices of the so-called “controllable” nodes. These controllable nodes are the nodes on which regulation can be carried out from the LV distribution network concentrator. These controllable nodes can be located at a smart meter.

[0072] On the contrary, non-controllable nodes are nodes to which the regulation method cannot be applied. These non-controllable nodes can be located at a conventional or smart electricity meter.

[0073] Furthermore, according to one embodiment, the fixed data comprise, for each controllable node of each low-voltage branch, the fixed variation domains permitted for the active and reactive power pairs corresponding to the controllable nodes. The active and reactive powers are counted positively if there is injection and negatively if there is withdrawal.

[0074] For each controllable node, a description of the admissible variation ranges can be provided. For example, the description can take the form of a diagram or a list of the extreme points of the domain.

[0075] Alternatively, the diagram illustrating the authorized variation ranges of the active powers Pi and the reactive powers Qi at the controllable nodes is variable over time. In particular, as will be described below, the method can be implemented at each step of a defined time, for example every 10 minutes. Thus, if the domains are variable over time at a time step less than the time step of implementation of the method, the collection of the data describing the authorized variation ranges of the active powers Pi and the reactive powers Qi is carried out during the collection of the variable data, as described below.

[0076] The regulation commands available at the reference node and at the controllable nodes of the LV low voltage branch are obtained. These regulation commands depend on the installed equipment and the terms of the connection contracts. The regulation commands include the setting of the setpoint voltage to regulate the voltage at the secondary of the HVA / LV transformer to which the LV low voltage branch is connected, the adjustment of the active power, the adjustment of the reactive power, and / or the adjustment of the active and reactive powers. This applies to the case where there is a regulator in charge of the HVA / LV transformer.

[0077] Finally, the fixed data includes a fixed list of available commands, prioritized in descending order of desired implementation. Optionally, this list may vary over time, and therefore be variable data, in order to meet contractual and / or material requirements.

[0078] The list can be obtained by means of an algorithm, or can be the result of a choice by the network manager.

[0079] Concerning the three-phase LV low voltage branches, the fixed data collected includes the list of nodes of the three-phase LV low voltage branch considered.

[0080] The fixed data also includes the admittance matrix Y describing the electrical behavior of each of the three phases, the three phases having an identical admittance matrix.

[0081] The fixed data further includes the value of the transformation ratio between the module of the voltage at the node of the HTA medium voltage branch to which the three-phase LV low voltage branch is connected, as well as the common value of the modules of the voltages on each of the three outputs of the HTA / LV transformer. The common value of the voltage can be deduced from the voltage at the node of connection of the three-phase LV low voltage branch to the HTA medium voltage branch by applying a transformation ratio.

[0082] In one embodiment, the HTA / LV transformer can be assumed to be perfect.

[0083] For greater realism, the losses of the HTA / LV transformer can be taken into account by considering that a known and given admittance yp is inserted between each output of the transformer assumed to be perfect and at each phase departure at the reference node of the LV low voltage branch.

[0084] The example presented below, as well as [Fig.7] illustrate the case where transformer losses are taken into account.

[0085] For each of the three phases of the three-phase low voltage LV branch, the following are collected: - the list of non-controllable nodes corresponding to consumers and / or producers operating in single-phase and connected to the phase considered, where for each of these nodes, the active and reactive powers injected and / or withdrawn are assumed to be given, - the list of controllable nodes corresponding to consumers and / or producers operating in single-phase on the phase considered, - the list of non-controllable nodes corresponding to consumers and / or producers operating in three-phase, where for each of these nodes, the active and reactive powers injected and / or withdrawn are assumed to be given, - the list of controllable nodes corresponding to consumers and / or producers operating in three-phase on the phase considered.

[0086] The regulation commands available at the three-phase controllable nodes are the same as those available for the controllable nodes of the single-phase LV low-voltage branches and depend on the technical characteristics of the node considered in the three-phase LV low-voltage branch.

[0087] A fixed list of available commands, prioritized in descending order of desired implementation, is collected. If the list varies over time, it can be collected with the variable data.

[0088] The list can be obtained by means of an algorithm, or can be the result of a choice by the network manager.

[0089] Finally, the fixed variation domains allowed for the active and reactive power pairs corresponding to the controllable nodes are collected. The active and reactive powers are counted positively if there is injection and negatively if there is withdrawal.

[0090] For each controllable node, a description of the permissible variation ranges of the active and reactive powers, in the form of a diagram, can be provided. The same variation range for the active and reactive power pairs is associated with each phase of the three-phase LV low voltage branch.

[0091] Alternatively, the diagram illustrating the authorized variation ranges of the active powers Pi and the reactive powers Qi at the controllable nodes is variable over time. In particular, as will be described below, the method can be implemented at each step of a defined time, for example every 10 minutes. Thus, if the domains are variable over time at a time step less than the time step of implementation of the method, the collection of the data describing the authorized variation ranges of the active powers Pi and the reactive powers Qi is carried out during the collection of the variable data, as described below.

[0092] The fixed data for the HTA medium voltage branch include in particular the target values ​​of the voltage at the nodes of the medium voltage branch. These values ​​are real and positive. These are the ideal values ​​of the voltage modules at the nodes, which must be achieved by regulation. These values ​​are in the form of a vector, of dimension equal to the number of nodes.

[0093] In the remainder of the description, the terms “HTA target voltage” or “HTA target voltage vector” are used interchangeably to refer to the target voltage vector corresponding to the target voltage state of the medium voltage branch.

[0094] Furthermore, the admittance matrix of the HTA medium voltage branch and the reference node of the HTA medium voltage branch are collected.

[0095] The list of indices of the controllable nodes is also collected. These controllable nodes are nodes on which regulation can be carried out from the site where the equipment used for regulating the medium branch is installed. high voltage.

[0096] For each medium and low voltage HTA BT branch, the nodes are identified by numbers. The numbering is independent in the different branches. Thus, the correspondence between the index of each reference node of the LV branches and the index of the node of the HTA branch to which each LV branch is attached is also collected.

[0097] The losses of the HTB / HTA transformer are also collected. They can be represented by a series admittance element which can be supplied directly.

[0098] The regulation commands available at the reference node and at the controllable nodes of the HTA medium voltage branch are also collected. The regulation commands may depend on the equipment installed and / or the terms of the connection contracts.

[0099] Finally, the fixed variation domains allowed for the active power pairs Pi and reactive power Qi corresponding to the controllable nodes of the HTA medium voltage branch can be collected. Alternatively, the active and reactive power pairs are variable over time and are collected during the steps S2 and S3 of collecting the variable data.

[0100] The regulation commands include the adjustment of the setpoint voltage of the regulator on load of the HTA / LV transformer, when this adjustment exists on the transformer, the adjustment of the active power, the adjustment of the reactive power, and / or the adjustment of the active and reactive power.

[0101] Advantageously, steps S2 to S12 of the method illustrated in [Fig.2] are implemented at each step of a predetermined time.

[0102] For example, the time step is equal to 10 minutes. The 10-minute time step can in particular be chosen in accordance with the IEC 61000-4-30 standard in force in France.

[0103] The time step to be chosen and modified according to the needs relating to the distribution network considered.

[0104] In step S2, the variable data relating to the LV low voltage branches are collected. The variable data of the single-phase and three-phase LV low voltage branches are the same.

[0105] For each LV low voltage branch, in addition to the diagram illustrating the authorized variation ranges of the active powers Pi and the reactive powers Qi at the controllable nodes if this is variable over time, the variable data may include the active powers Pi and reactive powers Qi injected or withdrawn at each node of the LV low voltage branch. These values ​​may be transmitted to the concentrator by means of a smart meter, for example in power line communication. In parallel, a voltage measurement may be made at the level of distribution substations. tribute by the concentrator.

[0106] Alternatively, a load estimation model or production and / or consumption profile may be used to determine the active and reactive powers produced.

[0107] The list of available commands can also be collected, ranked in descending order of desired implementation, if this list varies over time.

[0108] The list can be obtained by means of an algorithm, or can be the result of a choice by the network manager.

[0109] The variable data relating to the HTA medium voltage branch are collected in step S3.

[0110] The variable data, in addition to the diagram illustrating the authorized variation ranges of the active powers Pi and the reactive powers Qi at the controllable nodes if this is variable over time, include the values ​​of the active powers Pi and reactive powers Qi injected or withdrawn at each node of the HTA medium voltage branch. These values ​​can be transmitted to the concentrator by means of an electronic or digital meter providing the values ​​of electricity consumption and production.

[0111] The list of available commands can also be collected, ranked in descending order of desired implementation, if this list varies over time.

[0112] The list can be obtained by means of an algorithm, or can be the result of a choice by the network manager.

[0113] From the fixed and variable data, the voltage state of the network is determined in step S4.

[0114] More precisely, the voltage state of each low voltage branch and the voltage state of the medium voltage branch are determined. This step is described in more detail with reference to [Fig.5]. Typically, the voltage state of each low voltage branch and the voltage state of the medium voltage branch are determined according to any technique for solving the Kirchhoff system of equations, in other words the problem of calculating the distribution of loads, commonly called the "load flow" problem making it possible to determine the voltage at any node of the network by knowing the powers injected / withdrawn at each of the nodes of the network.

[0115] Determining the voltage state of each LV low voltage branch makes it possible to obtain a vector of so-called “effective” voltages, illustrating the effective voltage state of each node of each low voltage branch. By “effective voltage” is meant the real part of the complex voltage determined at a time t in a node of a low voltage branch. It is therefore a real state of the voltage at this node and at this time t. This real state of voltage in each low voltage branch is noted below and possibly in the references to the drawings “Vreal LV”.

[0116] In the remainder of the description, the terms “LV effective voltage” or “LV effective voltage vector” are used interchangeably to refer to the effective voltage vector corresponding to the actual voltage state of each low voltage branch, mentioned above (LV real).

[0117] Similarly, determining the voltage state of the HTA medium voltage branch makes it possible to obtain a vector of so-called “effective” voltages, illustrating the effective voltage state of each node of the medium voltage branch. “Effective voltage” means the real part of the complex voltage determined at a time t in a node of the medium voltage branch. It is therefore a real state of the voltage at this node and at this time t. This real state of voltage in the medium voltage branch is noted below and possibly in the references to the drawings “Vreal HTA”.

[0118] In the remainder of the description, the terms “effective HTA voltage” or “effective HTA voltage vector” are used interchangeably to refer to the effective voltage vector corresponding to the actual voltage state of the aforementioned medium voltage branch (Vreal HTA).

[0119] In step S5, it is checked whether voltage regulation is necessary.

[0120] For this, it is determined whether the effective voltage state Vreal of each low voltage branch LV is indeed within a first prescribed voltage range.

[0121] To determine whether the actual voltage state Vreal BT is within the first prescribed voltage range, the deviations between the actual voltage and the target voltage are determined. The deviations between the actual voltage and the target voltage are the absolute value deviations between the target voltage Vtarget and the modulus of the actual voltage.

[0122] [Fig.3] illustrates examples of prescribed ranges for LV low voltage branches.

[0123] Similarly for the medium voltage branch, it is determined whether the effective voltage state Vreal HTA is indeed included in a second prescribed voltage range.

[0124] [Fig.4] illustrates an example of the prescribed voltage range for HTA low voltage branches.

[0125] If the voltage state of each branch is within the first and second prescribed voltage ranges defined respectively for the low and medium voltage branches, regulation is not necessary and the method waits for the next time step in step S6, then implements step S2 again.

[0126] If the voltage state of at least one low voltage LV branch and / or the medium voltage HVA branch is outside, respectively, the first or second prescribed voltage range, regulation must be implemented.

[0127] In step S7, regulation command values ​​are separately calculated for the HVA medium voltage branch and for each LV low voltage branch. This calculation will be more particularly described with reference to [Fig.6].

[0128] The regulation command values ​​will then be used for the joint regulation of the HTA medium voltage branch and the LV low voltage branches.

[0129] Then, step S8 comprises verifying the existence of a solution ensuring the smallest difference between the actual voltages and the target voltages of the HTA medium voltage branch and the LV low voltage branches.

[0130] In order to carry out this verification, a regulation simulation is carried out, in which it is considered that there are no constraints limiting the effective use of all possible regulation commands at each 10-minute time step.

[0131] Thus, during this simulation, all possible regulation commands are activated, the commands having as values ​​the values ​​obtained in step S7.

[0132] This step has the property of ensuring that there is no solution in which the deviation from the target voltage could be reduced in one of the HTA or LV branches without increasing the deviation from the target voltage in another branch.

[0133] In other words, if, following this simulation, the voltage states of the HTA and LV branches enter their respective prescribed voltage range, it is assured that there is indeed a solution allowing joint regulation of the distribution network.

[0134] On the other hand, if, following this simulation, the voltage states of at least one of the HTA and LV branches do not fall within their respective prescribed voltage range, it is assured that there is no solution allowing joint regulation of the distribution network at the time step considered.

[0135] During step S9 it is determined whether such a solution exists.

[0136] In the remainder of the description, this solution is called “set of joint regulation commands” or “set of possible regulation commands” or “set of possible commands”. This solution groups together all the possible regulation command values, the values ​​having been calculated during step S7.

[0137] The set of possible commands includes the voltage value at the reference node of the branches which corresponds to the voltage setpoint value to be given to the regulator in charge of the HTA / LV transformer, the adjustment of the active power, the adjustment of the reactive power and the adjustment of the active and reactive powers.

[0138] If this is not the case, an alarm is triggered in step S10.

[0139] In step SI 1, the method continues towards the selection of sufficient commands allowing the effective joint regulation of the distribution network.

[0140] The solution identified during step S8 makes it possible to ensure the smallest difference between the actual voltage state and the target voltage state for the HTA and LV branches.

[0141] This solution includes the activation of all possible commands. However, the use of all possible commands may be subject to constraints, linked to the network or contractual or even to the availability of the hardware. It is therefore to construct a voltage-compliant regulation solution comprising the minimum and sufficient number of regulation commands to be effectively implemented during the voltage regulation stage so that the voltage at any node of the network complies with the prescribed range.

[0142] In the remainder of the description, the regulation solution is called “set of optimal regulation commands” or “set of optimal commands” or even “optimal commands”.

[0143] In some cases, there are no particular constraints on the implementation of commands, so that step S1 1 may be optional and all possible commands determined in step S9 may be effectively implemented for the joint regulation of the voltage.

[0144] The choice of the optimal set of commands for regulation is based on the hierarchical lists of commands in descending order of implementation preference. The hierarchical lists are assumed to be given, either as variable data obtained at each step of the implementation time, or as fixed data, for the HTA medium voltage branch and the LV low voltage branches.

[0145] Once the optimal commands have been chosen, they are implemented jointly for the joint regulation of the voltage of the HTA and LV branches in step S12.

[0146] With reference to [Fig.5], the calculations of the effective voltage state of the HTA medium voltage branch and the LV low voltage branches are now described in more detail.

[0147] The determination of the voltage state of the HTA medium voltage branch and the LV low voltage branches is carried out iteratively until a stopping criterion is satisfied. The stopping criterion may in particular be a maximum number of iterations, typically two or three iterations, or even a small difference between the results obtained between two successive iterations.

[0148] The determination of the voltage state of the HTA medium voltage branch alone, or of a LV low voltage branch alone, can be carried out by solving the so-called “load flow” problem, that is to say by solving the Kirchhoff system of equations making it possible to calculate the voltage at any node of the network by knowing the voltage at the reference node and the power injected / drawn at each node. The voltage state of the HTA medium voltage branch alone, or of a LV low voltage branch alone can be determined by any “load flow” resolution technique, a non-limiting example of which is described below.

[0149] In the context of the joint regulation of a medium voltage HTA branch and the low voltage LV branches connected to it, it may be interesting to take into account the electrical interactions between the medium voltage HTA branch and the low voltage LV branches.

[0150] Thus, in one embodiment, the voltage state is determined for each LV low voltage branch, then the data relating to each LV low voltage branch are aggregated into an equivalent node representing the power at this node seen from the HVA medium voltage branch.

[0151] Thus aggregated, the powers on all the nodes of the HTA medium voltage branch are known and it is possible to determine the voltage state according to any technique for solving the “load flow” problem by taking into account the interactions between the LV low voltage branches and the HTA medium voltage branch.

[0152] Initially, the voltage state of each low voltage LV branch alone is determined in step S20, by any technique for solving the “load flow” problem, and by taking as the voltage at the reference node of the low voltage LV branch the voltage imposed at the reference node of the medium voltage HVA branch, taking into account the ratios of the HVA / LV transformer.

[0153] Then in steps S21 and S22, the data relating to each low voltage LV branch are aggregated at their node of attachment to the medium voltage HVA branch.

[0154] Generally, the data relating to each LV low voltage branch are aggregated at the level of a node connecting the LV low voltage branch considered to the HVA medium voltage branch by calculating a pair of active and reactive power values ​​of the low voltage branch considered, each pair of active and reactive powers being respectively added to the active and reactive power values ​​at the nodes connecting the LV branch to the HVA branch, and by correcting the diagonal term of the admittance matrix of the HVA medium voltage branch corresponding to the node connecting the LV branch considered to the HVA medium voltage branch.

[0155] More precisely, considering a distribution network comprising a medium voltage HTA branch to which a low voltage LV branch is attached, the distribution network can be described by: the admittance matrix Yade the HTA medium voltage branch with the index node ref as the reference node; the admittance matrix Yb of the low voltage LV branch with the node of index 1 as the reference node (this index is given for information); the connection node between the low voltage and medium voltage HTA branch, here for example node 1 can be connected to any node r of the medium voltage HTA branch, r being different from ref; the vector of complex powers Sa injected / drawn at all nodes of the HTA medium voltage branch; the vector of complex powers Sh injected / drawn at all nodes of the low voltage LV branch.

[0156] Considering the HTA medium voltage branch alone, the voltage state Va of the HTA branch can be determined as being a solution to the following Kirchhoff system. [0i57] = v

[0158] Where the “*” sign indicates complex conjugation.

[0159] When a low voltage branch is connected to the HTA medium voltage branch, due to the injections / withdrawals, described by the vector Sb, in the low voltage branch, the voltage state of the medium voltage branch is modified and becomes Va.

[0160] The voltage state Va can thus be considered as solving a system of Kirchhoff equations, different from formula (3) but which can be deduced directly by adding: to the diagonal term of the matrix Ya a corrective term noted correctly Ya(r,r) to the component r of the vector Sa a corrective term noted correctly Sa(r).

[0161] The corrective terms below form an aggregated representation of the influence of the low voltage branch on the HTA medium voltage branch, at its connection node r.

[0162] More precisely, the corrective terms can be written as follows: ___.'J correetSB = -<(1) ■ with nb the number of nodes in the LV branch, K the subset of indices corresponding to the nb-1 nodes of the LV low voltage branch other than the reference node, Sb(K) the vector of complex powers injected / withdrawn at the nodes of the branch other than the reference node, Vb the vector of complex voltages at the nodes of the LV low voltage branch, Vb(K) the vector of size nb-1 obtained by removing the first component of Vb, Yb(l, K) the row vector deduced from the first row of the admittance matrix Yb after removing the first component, Yh*(l, K) the complex conjugate row vector of Yh(l, K), Yh(K, K) the submatrix Yb obtained by removing the first row and the first column of Yb, Y b(K, K) being an invertible matrix with its inverse denoted Yh(K, K)1 and Yh*(K, K) the conjugate matrix of Yh(K, K).

[0163] For the case where more than one low voltage LV branch is attached to the medium voltage branch, formulas (2) and (3) can be applied simultaneously to all the low voltage branches attached to the medium voltage branch, so that each of the diagonal terms of the admittance matrix of the medium voltage HTA branch corresponding to a node of attachment to a low voltage LV branch are corrected, and each component of the power vector Sa corresponding to a connection node to a low voltage LV branch are corrected.

[0164] Then, in step S23, the voltage state of the HTA medium voltage branch is determined by using the voltage imposed on the reference node of the HTA medium voltage branch as an initialization value and by taking into account the aggregated data of the LV low voltage branches.

[0165] The voltage state of the HTA medium voltage branch is determined by solving the “load flow” problem, using any known resolution technique.

[0166] In step S24, new voltage values ​​at the connection nodes of the LV low voltage branches to the HVA medium voltage branch are obtained, from the voltage state of the HVA medium voltage branch determined in the previous step.

[0167] In step S25, the value of the voltage at the connection node of the low voltage LV branches to the medium voltage HVA branch is imposed on the value of the reference node, multiplied by the HVA / LV transformation ratio, if there is no regulator on load at the HVA / LV transformer.

[0168] The voltage state of each low voltage LV branch is then calculated again in step S26, by solving the “load flow” problem.

[0169] New values ​​of the terms Correct and CorrectY are then obtained (step S27).

[0170] It is then determined in step S28 whether the stopping criterion is satisfied. If it is not, steps S22 to S28 are implemented again. If the stopping criterion is satisfied, the actual voltage state of the HTA medium voltage branch and the LV low voltage branches is obtained. In other words, the voltage is known at each node of the HTA medium voltage branch and the LV low voltage branches.

[0171] Step S7 of [Fig.2], during which initial regulation commands of the medium voltage HTA branch and each low voltage LV branch are calculated separately, is described more precisely with reference to [Fig.6].

[0172] Concerning the three-phase LV low voltage branches, an additional step must be carried out upstream. During this step, the admittance matrix of each three-phase LV low voltage branch connected to the HVA medium voltage branch is constructed upstream.

[0173] In a first case, the admittance matrix of each three-phase low voltage LV branch can be constructed taking into account the losses linked to the HVA / LV transformer.

[0174] In this first case, an additional node is added to each phase of the three-phase LV low voltage branch, the additional node representing the output of a transformer assumed to be perfect and with the same transformation ratio p as the imperfect real transformer. The three additional nodes thus introduced are noted rl, r2 and r3. Each of the three nodes is connected to the reference node of the low voltage LV branch corresponding to the associated phase by a branch formed by putting in series a resistance R and a self-inductance L (of low values), which corresponds to an admittance noted yt (yt = l / (R+jLco)), the index t indicating that it is an admittance relative to the transformer.

[0175] The values ​​of R and L are data which are determined such that the difference between the complex voltages at the two ends of this branch is a good approximation of the losses and phase shift induced by the imperfections of the transformer.

[0176] For example, for transformers with a nominal power between 160 and 250 kVA, typical values ​​of R and Lco are in the intervals [8.e-3, 2O.e-3] Ohm for R and [25.e-3, 50.e-3] Henry-Ohm for Lco.

[0177] Since the moduli of the output voltages of the three secondary windings of the assumed perfect transformer are the same and equal to 1 / p times the modulus of the voltage on the HTA branch side, the three nodes rl, r2, r3 can be grouped into a single node corresponding to the new reference node of the three-phase LV low voltage branch, represented in a developed form. This new node is numbered 1. This convention allows the construction of the three-phase admittance matrix taking into account the transformer losses.

[0178] The admittance matrix of each three-phase LV low voltage branch with losses in its developed form is noted YTP and is composed of the data of the three phases and the losses linked to the transformer. The YTP matrix is ​​of size (3n + 1, 3n + 1).

[0179] The list of nodes of the low voltage branch LV, other than the reference node re / is called ind.

[0180] Thus, to maintain the same numbering for the 3n-2 common nodes, the three nodes corresponding to the three respective attachment nodes of the three phases are numbered 3n-l, 3n and 3n+l respectively. In this way, the three rows and columns corresponding to these nodes in the YTP matrix are the last three rows and the last three columns.

[0181] A numbering of the nodes of the three-phase low voltage LV branch in its developed form is chosen such that: The first node is the reference node of the three-phase LV low voltage branch in its developed form; this node is the reference node common to each of the three phases; Nodes other than the reference node ref corresponding to the first phase are numbered from 2 to n (in the order they appear in the list of node indices ind); Nodes other than the reference node ref corresponding to the second phase are numbered from n+1 to 2n-l (in the order they appear in the list of node indices ind); Nodes other than the reference node ref corresponding to the third phase are numbered from 2n to 3n-2 (in the order they appear in the list of indy node indices. The three nodes corresponding to the three connecting nodes of the three phases are numbered 3n-l, 3n and 3n+l respectively. In this way, the three rows and columns corresponding to these nodes in the YTP admittance matrix are the last three rows and the last three columns.

[0182] The YTP admittance matrix of the three-phase LV low voltage branch in its developed form and taking into account the losses is constructed as follows: A first term equal to Y(ref, ref), followed by the n-1 terms of the line Y(ref, ind), followed again by the n-1 terms of the line Y(ref, ind), and followed, finally, by the n-1 terms of the line Y(ref, ind). The first column of the admittance matrix YTP is constructed in a similar way, by concatenating the column Y(ind, ref) 3 times following the diagonal term Y(l,l). The first line has the following unaffected coefficients: YTP(1,1) = -3*yt; YTP(1, 3*nl) = yt; YTP(1, 3*n) = yt; YTP(1, 3*n+l) = yt; The first column is the transpose of the first row: YTP(:,1) = (YTP(1,:))' ; The submatrix of YTP formed by rows and columns numbered 2 to 3*n-2 is the same as the submatrix of YTP formed by rows and columns numbered 2 to 3*n-2: YTP(2:3*n-2, 2:3*n-2) = YT(2:3*n-2, 2:3*n-2); The 3n-la line for unaffected coefficients: YTP(3*nl,l) = yt; YTP(3*nl,3*nl) = Y(ref, ref)-yt; YTP(3*nl, 2:n) = Y(ref,ind) Column 3n-la for unaffected coefficients: YTP(l,3*nl) = yt; YTP(2:n,3*nl) = Y(ind,ref); (Y(ind, ref) denotes the column of Y corresponding to the ref node, deprived of its ref component) Line 3n has the following unaffected coefficients: YTP(3*n,l) = yt; YTP(3*n,3*n) = Y(ref, ref)-yt ; YTP(3*n, n+l:2*n) = Y(ref,ind) ; where Y(ref,ind) denotes the line of Y corresponding to the reference node ref, deprived of the component ref). Column 3n has the following unaffected coefficients: YTP(l,3*n) = yt; YTP(n+l:2*n,3*n) = Y(ind,ref); The line 3n+la for unaffected coefficients: YTP(3*n,l) = yt; YTP(3*n,3*n) = Y(ref, ref)-yt; YTP(3*n, n+l:2*n) = Y(ref,ind); Column 3n+la for unaffected coefficients: YTP(l,3*n) = yt; YTP(n+l:2*n,3*n) = Y(ind,ref);.

[0183] The YTP admittance matrix of a three-phase LV low voltage branch in its developed form, taking into account the losses, thus constructed is presented in [Fig.7].

[0184] In other words, to construct the YTP admittance matrix of a three-phase LV low-voltage branch in its expanded form taking into account the losses related to the HTA / LV transformer, the list of all the nodes of the LV low-voltage branch is obtained. Then, a new list of nodes is created by considering that each node of the LV low-voltage branch is tripled, one node per phase, and by establishing an index correspondence between the nodes of the phases. Then, admittance sub-matrices for each of the three phases are obtained. The YTP admittance matrix is ​​reconstructed from the concatenation of the three admittance sub-matrices taking into account the corrective admittance illustrating the transformer losses, when the transformer losses are taken into account.

[0185] In a second case, a YT admittance matrix can be constructed for the low voltage LV branches without taking into account the losses linked to the HTA / LV transformer.

[0186] In this second case, and unlike the YTP matrix, a single reference node is defined for the three phases of the low voltage LV branch considered. The YT matrix is ​​therefore a matrix of size (3*n-2)x(3*n-2).

[0187] When constructing the YT matrix, the transformer losses yt are not taken into account. More precisely, in the first case with loss consideration, three reference nodes are created for each of the three phases. The three reference nodes are connected to each phase by three independent electrical connections, each with admittance yt, where yt is assumed to be known and given. Transformer losses occur in these three additional electrical connections. The currents flowing through the three electrical connections are generally different and the loss values ​​may also be different, as may the voltages at the reference nodes of the three phases.

[0188] Referring again to [Fig.6], in the case where the voltage state of one of the low voltage LV branches or the medium voltage HVA branch is not compliant, and therefore regulation is to be carried out, possible regulation command values ​​are determined.

[0189] For each low voltage LV branch and the medium voltage HVA branch, the regulation commands include an adjustment of the set voltage at the transformer (Vref) if the transformer has such an adjustment, and / or an adjustment of the active and / or reactive powers at all the controllable nodes of this network (P_i, Q_i), with i the index of a node.

[0190] The values ​​of the regulation commands are determined by an iterative resolution of the optimization problem consisting of minimizing, for each low voltage LV branch and the medium voltage HVA branch, the maximum of the deviations in absolute value between the target voltage Vtarget(i) and the real part of the voltage at node i, real(V(i)), calculated on all the n nodes of the network, and respecting: - Kirchhoff's equations which establish a non-linear relationship between the values ​​of the charges and the productions and the tensions, which are functions of the unknowns Vref and (P_i, QJ), - The admissible domains of active and reactive powers at the controllable nodes, - the possible voltage values ​​that can be imposed at certain nodes, and for three-phase low voltage LV branches, the equalities to be imposed on the decision variables (P_i, Q_i) corresponding to controllable producers operating in three-phase.

[0191] The optimization problem is not linear, due to the Kirchhoff equations. The resolution of the optimization problem is carried out iteratively, in which at each iteration, a linearization of the Kirchhoff equations is carried out in order to approximate the calculation of the regulation commands as a solution to the optimization problem of the form:

[0192] t (mm cx X SC « 4x < / 3 Ex = /

[0193] Where x denotes the unknown vector of the regulation commands to be determined, cTx the objective function and Ax<=b and Ex = f represent the set of inequality and equality constraints to be respected. The resolution of a linear program uses known and efficient algorithms.

[0194] In the case where not all the regulation commands are available, for example if the load adjustment is not available, the subset of the commands is part of the same model. To do this, the values ​​of the variables corresponding to the unavailable commands are fixed and optimized in relation to the other variables.

[0195] ""

[0196]

[0197]

[0198]

[0199]

[0200]

[0201]

[0202]

[0203]

[0204]

[0205]

[0206]

[0207] It should be noted that in the case of a three-phase LV low voltage branch using the YTP admittance matrix, the voltage at the reference node (numbered 1) does not have to be taken into account in the calculation of the deviation from the target voltage Vtarget. The objective function aims to ensure compliance with regulatory voltage ranges at each network node. The proposed objective function is the largest absolute value deviation between the target voltage Vtarget(i) and the real part of the voltage at node i, real(V(i)), calculated across all network nodes. Such an objective function makes the solution as insensitive as possible to possible variations in the values ​​of active and reactive powers and thus guarantees a certain robustness with respect to uncertainties in injections and withdrawals. Kirchhoff's laws are taken into account by making a linear approximation of the relationship between the voltages and the injected powers. Various linear approximations exist (for example, the approximation of the real part of the voltage vector by a linear function of the real components and the imaginary components of the injected powers). Any suitable technique for linearizing Kirchhoff's equations allows them to be rewritten in the form of linear constraints of the type: Kx = S Initialization is done by choosing the target voltage as the linearization point of the Kirchhoff equations, between the target voltage of the LV low voltage branches or the target voltage of the HVA medium voltage branch, depending on whether the commands are calculated for a LV low voltage branch or for the HVA medium voltage branch. For the construction of the linear program, several constraints must be taken into account. A constraint corresponds to the admissible domains of active and reactive powers at the controllable nodes, which are fixed data obtained in step S30. At each controllable node, the adjustment of the active and reactive power values ​​must be carried out in a domain specific to the equipment installed at the node in question. The first case is that in which the adjustment domain of the active and reactive powers is convex. This domain is described by a polyhedron of R2 (convex domain of / ^2j in which it is possible to choose the pair of active and reactive powers (P_i,Q_i). This constraint is of type: Cx < d An example of an application in which the adjustment domain of active and reactive powers is convex is on photovoltaic equipment present at the level

[0208]

[0209]

[0210]

[0211]

[0212]

[0213]

[0214]

[0215]

[0216]

[0217]

[0218]

[0219]

[0220]

[0221]

[0222] of a controllable node. [Fig.8] illustrates such an example. In [Fig.8], the adjustment domain of the active and reactive powers is convex and delimited between the active power Pi, min and Pi, max. For a phase shift between active and reactive powers of q>, the admissibility domain, shaded in [Fig.8], is defined by the inequalities: 4an(cp) Pr<-Qr<-tan(^ Eüm- - ' P i with "tan" the tangent function applied to q>. These two inequalities can be rewritten in the form: So the constraint is indeed of the general form Cx <= d. The case where the domain requires a higher number of linear inequalities falls under the same type of constraints. The second case is that in which the adjustment domain of the active and reactive powers is not convex. It is then a question of approximating the domain by a polyhedron of R2 to have a linear constraint of the same type, that is to say of the type Cx <= d. To do this, it is possible to perform a convexification of the initial non-convex domain by determining a convex hull. This can be done by any known algorithm, for example, Jarvis's March algorithm, Chan's algorithm, Voronoi diagram, etc. Then, the constraint system on the convexified domain is constructed. The solution is obtained after solving the linear program and obtaining the regulation commands at the last step in [Fig.7]. More precisely, the solution obtained is a regulation command corresponding to the adjustment of active and / or reactive powers. It is then checked whether the adjustment values ​​of the active and / or reactive powers obtained belong to the operating domain of the equipment considered. In other words, it is checked that the adjustment values ​​of the active and / or reactive powers belong to the initial domain which is non-convex. When the solution obtained does not respect the constraints of the domain, it is possible to substitute a point from the initial domain for the solution. According to one embodiment, the substitution is done by a chosen projection technique. For example, it is possible to choose the Euclidean projection.

[0224] If the solution respects the domain constraints, there is no correction to be made.

[0225] An example of an application in which the adjustment domain of the active and reactive powers is not convex is a wind turbine equipment present at a controllable node. Figures 9A to 9C illustrate such an example.

[0226] In [Fig.9A], the non-convex active and reactive power adjustment domain is shown in hatching.

[0227] The convexification of the domain is represented in [Fig.9B], with the initial non-convex domain represented by the hatching, and the additional domain having given rise to the convexification represented by the crosshairs.

[0228] When verifying whether the solution belongs to the initial domain, it is determined whether the solution obtained is indeed in the initial domain, i.e. whether it is indeed in the hatched part of [Fig.9B].

[0229] If the solution obtained is in the initial domain, then there is no correction to be made.

[0230] If the obtained solution is not in the initial domain, as illustrated in [Fig.9C] by the cross-shaped point, an adjustment is necessary. The adjustment can be made by substituting a point in the initial domain for the obtained solution using any chosen technique. An example of a technique may be a substitution by a projection according to a chosen norm.

[0231] [Fig.9C] illustrates a non-limiting example of two possible substitutions.

[0232] The first substitution, illustrated by the round point, is carried out by the orthogonal projection of the solution obtained on the initial domain, by means of the Euclidean norm.

[0233] The second substitution, illustrated by the triangular point, is carried out by the oblique projection of the solution obtained on the initial domain, to limit the loss of active power.

[0234] The constraints relating to the configuration of a three-phase low voltage LV branch can also be taken into account.

[0235] In particular, when a low voltage LV branch is three-phase, the additional equality constraints to be imposed on the different decision variables for the active and reactive powers corresponding to producers operating in three-phase at the level of controllable nodes are taken into account.

[0236] These constraints are of the form -.Tx=p.

[0237] The addition of the two constraints linking the voltage at the reference node and the value of the objective in the linear program are made inactive due to the fact that, for a three-phase LV low voltage branch using the YTP admittance matrix, the voltage at the reference node (numbered 1) does not have to be taken into account in calculating the deviation from the target voltage.

[0238] To obtain the complete formulation in linear programming (LP) of the regulation problem, it remains to take into account the fact that the components of the vector of powers relating to injections and withdrawals at non-controllable nodes are fixed and known. They are therefore not variables of the problem. These constraints are of the form Sx = v.

[0239] Finally, the optimization problem taking into account all the constraints and fixed data (step S30, S31 of [Fig.6]), can be written:

[0240] toc O ™ s Tx ~ p Sr— F Cx <

[0241] The optimization problem can be solved by known solvers.

[0242] The optimization problem can be rewritten as a linear program:

[0243] min X ■VX SX -z < ôT(xj< +z Kx™s Tx ss p F ex < d I

[0244] Rewriting the optimization problem in the form of the above linear program makes it possible, in particular, to process large distribution networks more efficiently.

[0245] This corresponds to step S33.

[0246] In step S34, the linear program can be solved using several techniques known from the prior art, and in particular using solvers.

[0247] Once the linear program is solved, a first optimized value of the voltage at the reference node and the values ​​of the active reactive powers at the controllable nodes are obtained in step S35.

[0248] The active powers P(i) and reactive powers Q(i) thus obtained are used together as part of the initial regulation commands. The reactive powers Q(i) alone thus obtained can also be used.

[0249] The voltage at the reference node can however be improved in step S36. More pre- Specifically, step S36 aims to increase the accuracy of this first optimized value.

[0250] Indeed, the first optimized value is obtained following a linearization of the Kirchhoff system of equations. Thus, the value obtained with respect to the effective voltage state of the network, that is to say which would result from the exact resolution of the Kirchhoff system of equations for the vector S of the powers injected to the different nodes of the network, where, for each controllable node i, the real part (S(i)) and the imaginary part (S(i)) are the values ​​taken by the decision variables corresponding to node i at the optimum of the linear program.

[0251] However, the accuracy of the voltage values ​​at the reference node Vref and of the deviation from the target voltage at the reference node Vcible can be significantly improved by determining the voltage state minimizing the deviation from the target voltage at the reference node Vcible from the optimal values ​​of the powers injected to the controllable nodes. For this, an additional calculation is performed immediately after solving the linear program.

[0252] Firstly, a dichotomous search on the voltage at the reference node can be carried out using a linear approximation of the variations of the voltage state (the Kirchhoff system solution) as a function of the voltage at the reference node, and from the knowledge of two given Kirchhoff solutions, Vsoll and Vsol2. These solutions correspond to chosen values, framing the value of the target voltage at the reference node Vcible(ref). Thus, for example, the two solutions can be between 0.9xVcible(ref) and l.lxVcible(ref), Vsoll being less than the value of the target voltage at the reference node, and Vsol2 being greater than the value of the target voltage at the reference node. Preferably, the solutions Vsoll and Vsol2 correspond respectively to the values ​​0.95xVcible(ref) and 1.05xVcible(ref) of the target voltage at the reference node.The result of this first step leads to identifying a framework of the unknown final optimal value of the voltage at the reference node between two values ​​noted Vreflbis and Vref2bis.

[0253] Then, the Kirchhoff system of equations is solved for each of the two values ​​Vreflbis and Vref2bis. This allows two new solutions Vsoll' and Vsol2' to be determined. With these two new solutions, the first step is implemented again.

[0254] This allows high precision on the voltage state to be obtained.

[0255] In step S37, the accuracy of the optimal voltage at the reference node can still be improved. For this, the voltage value obtained at the output of step S36 is used as the linearization point of the linear program. In fact, this value is injected into the linear program construction step (S32).

[0256] Thus, steps S30 to S37 can be implemented iteratively, up to convergence of the solution. The number of iterations is between 1 and 15, and preferably between 1 and 5, and preferably between 2 and 4. At the output, an optimal value of the voltage at the reference node is obtained.

[0257] This value of the voltage at the reference node is notably part of the regulation commands.

[0258] [Fig. 10] illustrates more precisely step S8 of [Fig. 2], during which the verification of the existence of a solution ensuring the smallest difference between the actual voltages and the target voltages of the HVA medium voltage branch and the LV low voltage branches is carried out.

[0259] In other words, it is a question of determining whether there is a solution allowing regulation, or whether there is no solution to trigger an alarm.

[0260] An initialization is first carried out in step S40, during which a regulation simulation is carried out for each low voltage LV branch by activating all possible regulation commands, even if certain commands are not available.

[0261] More precisely, a regulation simulation including the activation of all possible regulation commands is carried out for each low voltage LV branch.

[0262] Each low voltage LV branch is associated with a set of possible regulation commands.

[0263] The regulation simulation is notably carried out using the values ​​of the commands obtained at the output of the method described with reference to [Fig.6].

[0264] A test is then performed in step S41, to determine whether the voltage states of all the LV low voltage branches are consistent with the target voltage state. For this, the voltage states of each LV low voltage branch are calculated following the regulation simulation performed in step S40.

[0265] If the voltage status of the LV low voltage branches is not compliant, an alarm is activated for each non-compliant LV low voltage branch (step S53).

[0266] Then in step S42, the correction values ​​Corrects and CorrectY are calculated. These values ​​are calculated in the same manner as that described with reference to [Fig.5].

[0267] The values ​​Corrects and CorrectY are then used to carry out the step of aggregating the data relating to each low voltage LV branch at the level of their connection node with the medium voltage HVA branch, in step S43.

[0268] Steps S42 and S43 correspond to steps S21 and S22 described with reference to [Fig.5].

[0269] Then, a simulation of regulation of the HTA medium voltage branch is carried out in step S44, taking into account the aggregated data of the LV low voltage branches, and therefore the electrical interactions between the HTA medium voltage branch and the low voltage LV branches attached to it.

[0270] As for the low voltage LV branches, a set of regulation controls is associated with the medium voltage HVA branch.

[0271] The voltage regulation simulation on the HTA medium voltage branch is carried out by activating all the possible regulation commands of the set of regulation commands associated with the HTA medium voltage branch, even if certain commands are not available, the values ​​of which were obtained at the output of the method described with reference to [Fig.6]. The voltage state of the HTA medium voltage branch can also be obtained.

[0272] Then, in step S45, the voltage values ​​at the connection nodes of each low voltage LV branch to the medium voltage HVA branch are calculated.

[0273] This step is advantageous since it allows the voltage values ​​at the connection nodes to be adjusted, depending on the voltage state of the HTA medium voltage branch obtained following the regulation simulation.

[0274] It is then determined in step S46 whether each low voltage LV branch has a load regulator on their reference node.

[0275] When a low voltage LV branch does not have a load regulator, the voltage at the reference node of the low voltage LV branch is not controllable and is therefore not part of the available controls that can be activated during the actual regulation of the voltage. For these low voltage LV branches, the voltage at the reference node is imposed on the value of the voltage at the connection node of the low voltage LV branch to the medium voltage HVA branch, multiplied by the HVA / LV transformation ratio (step S47).

[0276] The set of regulation commands associated with each low voltage LV branch not having an on-load regulator is modified to remove the regulation commands corresponding to the on-load adjustment and the adjustment of the voltage at the reference node.

[0277] The sets of regulation commands associated with each low voltage LV branch not including a load regulator therefore include the load adjustment, the value of the voltage at the reference node and the adjustment of the active and / or reactive powers.

[0278] In other words, there is no constraint on the value of the voltage at the reference node of the LV low voltage branch. The regulation can be done as if the regulation of the HVA medium voltage branch and the regulation of the LV low voltage branch considered were independent and the value of the voltage at the reference node can be an activatable lever for the regulation of the voltage (step S48).

[0279] When a low voltage LV branch has a load regulator, the voltage at the reference node of the low voltage LV branch is controllable. In other words, all possible commands are available and can be activated during the voltage regulation stage. The sets of regulation commands associated with each LV low voltage branch with a load regulator are not modified and include all possible commands.

[0280] The method continues at step S49, during which a new simulation of regulation of the LV low voltage branches is carried out by activating the sets of regulation commands associated respectively with each LV low voltage branch comprising or not comprising a load regulator.

[0281] Thus, for the LV low voltage branches not having a load regulator, the regulation simulation is carried out taking into account the fact that the value of the voltage at the reference node of the LV low voltage branch is imposed on the value of the voltage at the connection node with the HVA medium voltage branch. In other words, all the commands activated for this simulation do not include the value of the voltage at the reference node. The available commands therefore include the adjustment of the active and / or reactive powers at the nodes of the LV low voltage branch considered.

[0282] For LV low voltage branches with a load regulator, the regulation simulation is carried out with all possible commands. The activated regulation commands therefore include the adjustment of the voltage value at the reference node as well as the adjustment of the active and / or reactive powers at the nodes of the LV low voltage branch considered.

[0283] Then, in step S50, during which new values ​​of the corrective terms Corrects and CorrectY are calculated following the regulation simulation on the low voltage LV branches.

[0284] In step S51 it is determined whether the second stopping criterion is satisfied. This stopping criterion may in particular correspond to a fixed maximum number of iterations, a difference between the results obtained between two successive iterations, in particular if this difference is less than a predefined threshold or even compliance with a predefined criterion.

[0285] If the stopping criterion is not respected, the method continues with the iterative implementation of steps S43 to S51, starting with the aggregation of the data relating to the low voltage LV branches, obtained in the previous step (i.e. in step S50), which corresponds to step S43.

[0286] If the stopping criterion is satisfied in step S51, it is determined whether the voltage states of each LV low voltage branch and of the HVA medium voltage branch, obtained following the regulation simulations carried out respectively in steps S44 and S49, comply with the regulatory ranges defined respectively for the LV low voltage branches and the HVA medium voltage branch (step S52).

[0287] If the respective voltage state of each low voltage LV branch and of the medium voltage HVA branch does not fall within the first and second regulatory range, an alarm is activated for each branch whose voltage state does not comply with step S53.

[0288] This means in particular that the joint regulation of the HTA medium voltage branch and the LV low voltage branches cannot be carried out. Indeed, if by activating all possible commands certain voltages cannot be regulated, it follows that the branches concerned cannot be regulated at the time step considered.

[0289] If all the voltage states are compliant, the sets of regulation commands associated with each LV low voltage branch and each HVA medium voltage branch together form the set of joint regulation commands that can be implemented for the effective regulation of the LV low voltage branches and the HVA medium voltage branch are obtained.

[0290] Unlike the set of possible commands activated in the initialization step S40, the method returns in step S54 a set of “available” commands. The set of available commands makes it possible in particular to take into account the presence or absence of a load regulator on the HTA / LV transformer.

[0291] The process then continues to [Fig. 1 1]. The steps implemented and illustrated by [Fig. 11] make it possible to choose, then implement, the optimal regulation commands to enable the joint regulation of the HTA medium voltage branch and the LV low voltage branches connected to it.

[0292] Optimal regulation commands include commands sufficient to enable regulation.

[0293] Thus, in some cases, the available commands, obtained in step S54, are not all part of the optimal commands. The optimal commands represent a subset of the available commands.

[0294] In step S60, the hierarchical lists in descending order of preference for implementing the commands are obtained for the low voltage LV branches and the medium voltage HVA branch.

[0295] The hierarchy of orders is predefined and takes into account in particular the connection contracts of the producers, or even the availability of the production entities.

[0296] In step S61, the first command in the list is selected, then a simulation of joint regulation of the voltage on the low voltage LV branches and the medium voltage HVA branch is carried out.

[0297] During this simulation, only the first command of the hierarchical lists is activated.

[0298] Then, in step S62, the voltage state of the low voltage LV branches and the medium voltage HVA branch is calculated by solving the “load flow” problem under the effect of the command selected in step S61.

[0299] It is then determined whether the respective voltage states of the low voltage LV branches and the medium voltage HVA branch are compliant thanks to the activation of the first command (step S63).

[0300] If the voltage states are not compliant, the next command on the hierarchical lists is selected in addition to the previously selected command, and a regulation simulation is again carried out in step S64 with the cumulative commands.

[0301] The method resumes again at step S62 by calculating the voltage states of the low voltage LV branches and the medium voltage HVA branch.

[0302] If it is determined that the voltage states are in accordance with step S63, the selected commands which made it possible to regulate the voltage of the low voltage LV branches and the medium voltage HVA branch correspond to the sufficient commands and are the optimal commands actually implemented on the distribution network and allowing the joint regulation of the low voltage LV branches and the medium voltage HVA branch.

[0303] [Fig. 12] shows the effects of hierarchical activation of regulation commands to obtain the optimal and sufficient regulation commands to be implemented for joint regulation of the voltage within the prescribed ranges.

[0304] According to this example, five different regulation commands can be activated. The lists therefore include commands C1 to C5, with command C1 being the first command in the list and command C5 the last.

[0305] It is visible in the figure that the activation of the first command C1 alone does not allow a voltage state to be obtained within the prescribed voltage ranges.

[0306] On the other hand, activating the first two commands in the list C1 and C2 makes it possible to obtain a compliant voltage state, i.e. lower than the prescribed voltage ranges. The first two commands C1 and C2 are therefore the optimal regulation commands which will actually be implemented and which will allow the joint regulation of the HVA medium voltage branch and the LV low voltage branches.

[0307] The activation of commands C1 to C3, C1 to C4 and C1 to C5 are illustrated as an example. As can be seen in the figure, the activation of these commands would also allow joint regulation of the voltage. However, only commands C1 and C2 will be activated since these are the commands sufficient for the voltage state on the network to be compliant. The activation of commands C3 to C5 therefore becomes superfluous.

[0308] Figures 13A and 13B illustrate examples of machine architecture on which the method can be implemented.

[0309] The computer used can be either a conventional computer, comprising a single processor, and called “single-core”, or a computer comprising several processors and called “multi-core” or “cluster” in English.

[0310] In implementing the method, certain calculations can be carried out independently of each other, and can therefore be processed in parallel, in particular the calculation of the voltage state of a single LV low voltage branch.

[0311] For other calculations, additional knowledge is required so that these calculations are carried out sequentially, in particular for example the calculation of the voltage state of the HTA medium voltage branch following the step of aggregating the data from the LV low voltage branches.

[0312] For a distribution network comprising a large number of low voltage LV branches, the calculation of the optimal joint regulation commands can take a certain amount of time, which would not be compatible with implementation every ten minutes.

[0313] Also, the regulation commands can be calculated preferentially on a “multi-core” computer. This makes it possible to carry out in parallel the calculations of the voltage state and the regulation commands for each LV low voltage branch, as well as the aggregation of the data relating to the LV low voltage branches, independently of each other and to reduce the calculation times.

[0314] [Fig. 14] illustrates a mode of application of the regulation process on a schematic distribution network.

[0315] The distribution network comprises a medium voltage HVA branch, supplied by a high voltage HTB branch (not shown) at an HTB / HTA transformer comprising an HTA load regulator. The distribution network further comprises a plurality of low voltage LV branches, supplied by the medium voltage HVA branch at an HTA / BT transformer which may comprise an HTA / BT load regulator.

[0316] For the sake of clarity, only a low voltage LV branch including a load regulator is shown in [Fig. 14].

[0317] Each of the medium and low voltage branches comprises a plurality of non-controllable nodes, represented as empty circles on the diagram, and a plurality of controllable nodes, represented as solid circles on the diagram.

[0318] Each of the medium and low voltage branches comprises a consumer whose consumption is respectively measured by an HTA meter for the HTA medium voltage branch and by a conventional LV meter or a smart LV meter. for the low voltage branch.

[0319] From these meters, the real active and reactive powers (P, Q), i.e. the active and reactive powers actually measured in the medium or low voltage branches, are transmitted to the regional control station via the concentrator for the LV low voltage branches. For the HVA medium voltage branch, the transmission can be made via a switched telephone network (PSTN) or equivalent and / or a GPRS type cellular network for example.

[0320] The distribution network also includes in each medium and low voltage branch at least one producer-consumer, whose consumption and production are respectively transmitted to an HTA smart meter for the HTA medium voltage branch and to a LV smart meter for the LV low voltage branch. The smart meters can be located at a controllable node.

[0321] The real active and reactive powers (P, Q), i.e. the active and reactive powers actually measured in the medium or low voltage branches at the producer-consumer level, are also respectively transmitted to the regional control station via the concentrator for the low voltage LV branches. For the medium voltage HVA branch, the transmission can be made via a switched telephone network (PSTN) or equivalent and / or cellular network such as GPRS for example.

[0322] Each HTB / HTA and HTA / BT load regulator also receives the corresponding Vreal voltage, i.e. the effective voltage in each medium or low voltage branch.

[0323] The regulation method as described with reference, in particular, to figures 2, 5, 6, 10 and 11 is implemented at the level of a regional control station (not shown).

[0324] The optimal regulation commands are sent to the relevant equipment and devices.

[0325] More precisely, the optimal value of the Vopt voltage to be applied to the medium voltage branch is sent to the HTB / HTA load regulator. When the load adjustment of the HTA / BT transformer is available, the optimal value of the Vopt voltage to be applied to the low voltage branch is sent to the HTA / BT load regulator.

[0326] On the other hand, the optimal values ​​of the P(i)opt and Q(i)opt pairs at the controllable nodes, as determined by the regulation method, are transmitted to the smart meters and the control-command boxes of the controllable nodes. In practice, each producer-consumer at a controllable node will have to produce or absorb reactive (voltage too high or too low respectively) according to the optimal values ​​of the P(i)opt and Q(i)opt pairs calculated. The operational implementation of this adjustment, “downstream of the meter” and falling under the responsibility of the user, therefore includes a series of actions carried out by a control-command at the user's premises such as for example a clipping of production, storage in any means of storage of excess production, management of flexibilities in particular for the recharging of electric vehicles, activation of heating, etc., and the adjustment of active and reactive powers on an inverter.

Claims

1. Claims Method for jointly regulating a voltage at each node of a distribution network comprising a medium voltage, HTA, branch and at least one, and preferably a plurality of low voltage, LV, branches, the distribution network further comprising a plurality of energy generators, each low voltage branch being connected to the medium voltage branch at an attachment node at an HTA / LV transformer, each low voltage branch further comprising a reference node, a plurality of controllable nodes and a plurality of non-controllable nodes, the medium voltage branch being supplied by a high voltage branch connected to the medium voltage branch at an attachment node at an HTB / HTA transformer with load regulator, the medium voltage branch further comprising a reference node, a plurality of controllable nodes and a plurality of non-controllable nodes,each node of the medium voltage and low voltage branches being subjected to energy withdrawals and / or injections, said at least one branch, or at least one among the plurality of low voltage branches being in three-phase mode, the method comprising: - a definition:, °a target voltage at each node of each low voltage branch, °a first prescribed voltage range surrounding the target voltage of each low voltage branch; ° of a target voltage at each node of the medium voltage branch, ° of a second prescribed voltage range surrounding the target voltage of the medium voltage branch; and the method comprising, at each step of a predetermined time: - a determination of a voltage state of the low voltage branches - a determination of a voltage state of the medium voltage branch taking into account electrical interactions of the low voltage branches attached to it -then, a comparison: °of the actual voltage state of each low voltage branch with the first prescribed voltage range ° of the actual voltage state of the medium voltage branch with the second prescribed voltage range; - if at least one voltage of the low voltage branch and / or the branch

2. medium voltage is outside the first prescribed voltage range and / or the second prescribed voltage range respectively: - a search for a set of joint regulation commands allowing the effective voltage of the low voltage branch and the medium voltage branch to be included respectively in the first prescribed voltage range and in the second prescribed voltage range: - a determination and implementation of at least one optimal regulation command from among the set of joint regulation commands characterized in that the search for the set of joint regulation commands for each three-phase low voltage branch is preceded by a construction of a three-phase admittance matrix comprising, for each three-phase low voltage branch: ° creation of a list of nodes for each of the three phases of the three-phase low voltage branch by carrying out: *a tripling of each node of the low voltage branch, one node of each triplet of nodes being respectively associated with one of the three phases, * a correspondence of indices of the nodes of each of the three phases, ° an obtention of three admittance sub-matrices corresponding respectively to an admittance matrix of each of the three phases of the low voltage branch, ° a reconstruction of a three-phase admittance matrix by a concatenation of the three admittance sub-matrices taking into account a corrective admittance illustrating losses at the level of the HTA / LV transformer of the low voltage branch considered. The method of claim 1, wherein: determining said at least one optimal regulation command to be implemented from among the set of regulation commands comprises - obtaining a hierarchical list of regulation commands according to a decreasing order of priority of implementation, - a simulation of joint regulation of the medium voltage branch and the low voltage branches by activating a first regulation command from among the set of joint regulation commands, the first activated command corresponding to a first command on the hierarchical list, and, if the voltage state of the branches low voltage and medium voltage is not included in the first prescribed voltage range and / or in the second prescribed voltage range, ° iteratively carry out other joint regulation simulations by adding, at each iteration, the following regulation command in the hierarchical list, until the voltage states of the low voltage branches and of the medium voltage branch are respectively included in the prescribed voltage prescribed voltage range and in the second safety range, - when the voltage states of the low voltage branches and of the medium voltage branch are respectively included in the first prescribed voltage range and in the second prescribed voltage range, obtaining at the output of said at least one optimal regulation command to be implemented.

3. Method according to one of the preceding claims, in which the search for the set of joint regulation commands for each low voltage branch and for the medium voltage branch is preceded by a step of calculating regulation command values ​​comprising: - an iterative resolution, by a fixed point technique, of an optimization problem taking into account: ° constraints relating to the admissible domains of the active and reactive powers at the controllable nodes, ° constraints relating to the three-phase configuration of the low voltage branch, each iteration comprising a resolution of a linear program integrating a linearization of the Kirchhoff equations, the program returning values ​​of the regulation commands as output,regulation commands comprising: ° an adjustment of a voltage value at the connection node of a low voltage branch to the medium voltage branch for an on-load adjustment of the HTA / LV transformer, when the adjustment is available, ° an adjustment of a voltage value at the connection node of the medium voltage branch for an on-load adjustment of the HTB / HTA transformers, ° an adjustment of active power (P), in medium voltage and in low voltage, °a reactive power adjustment (Q), in medium voltage and in low voltage, and / or °an adjustment of active and reactive power (P, Q), in medium voltage and in low voltage.

4. Method according to the preceding claim, in which when an admissible domain of the active and reactive powers at the controllable nodes is not convex, the method comprises - a convexification of the initial non-convex domain, - following the resolution of the optimization problem, a verification that the adjustment values ​​of the active and reactive powers (P, Q) are in the initial domain and, if this is not the case, * a substitution of the adjustment values ​​of the active and reactive powers (P, Q) by values ​​included in the initial domain, to have a solution respecting the initial adjustment domain of the active and reactive powers.

5. Method according to one of the preceding claims, in which the search for the set of joint regulation commands for the low voltage branches and the medium voltage branch comprises - respectively, a definition of a set of regulation commands for each low voltage branch and the medium voltage branch, each set of regulation commands comprising all the possible regulation commands, - for each low voltage branch, a first step of simulating a regulation by an implementation of the set of regulation commands of the low voltage branch considered, - a determination of the voltage state of each low voltage branch, and, when the voltage state of each low voltage branch falls within the first prescribed voltage range: ° an aggregation of data relating to each low voltage branch at the level of the respective connection node of each low voltage branch to the medium voltage branch, ° a simulation of regulation of the medium voltage branch by activating all the regulation commands of the medium voltage branch, taking into account the aggregated data, ° a calculation of the voltage values ​​at the reference nodes of the aggregated low voltage branches, °and for each low voltage branch not having a load regulator:

6. *the value of the voltage at the reference node of the low voltage branch is imposed on a value of the voltage at the connection node of the low voltage branch to the medium voltage branch, multiplied by a transformation ratio of the transformer, the method then comprising the iterative implementation of the following steps until a first stopping criterion is met:, H a modification of all the commands of each low voltage branch considered by removing a command corresponding to a load setting of the transformer, H another step of simulation of regulation of each low voltage branch not having a regulator in charge, by activating all the regulation commands of each low voltage branch considered, Hthen, obtaining each voltage state of the low voltage branches not having a load regulator, H a return to the data aggregation step, * if a low voltage branch has a load regulator, the set of regulation commands includes all possible regulation commands, - the method then comprising the calculation of the voltage states of the medium and low voltage branches, following the respective activation of the control sets of each low voltage branch and of the medium voltage branch, - and, when the voltage states of the medium and low voltage branches respectively fall within the first prescribed voltage range and the second prescribed voltage range, obtaining each set of regulation commands corresponding to each low voltage branch and for the medium voltage branch, said sets of commands forming the set of joint regulation commands. Method according to the preceding claim, in which: - if, following the determination of the voltage state of each low voltage branch, at least one voltage state of a low voltage branch is outside the first prescribed voltage range, an alarm is activated for said at least one low voltage branch, and / or - if, following the calculation of the voltage states of the medium and low voltage branches, the voltage state of at least one low voltage branch or the voltage state of the medium voltage branch is outside the first prescribed voltage range and / or the second prescribed voltage range respectively; prescribed voltages, an alarm is activated.

7. Method according to one of the preceding claims, in which the determination of the voltage state of the low voltage branches and of the medium voltage branch comprises: - a determination of the voltage state of each low voltage branch, - then, a step of aggregating data relating to each low voltage branch at the level of the respective connection node of each low voltage branch to the medium voltage branch by * calculating a pair of active and reactive power values ​​of each low voltage branch, each pair of active and reactive powers being respectively added to active and reactive power values ​​at the connection nodes of each low voltage branch to the medium voltage branch, and by correcting each diagonal term of an admittance matrix of the medium voltage branch,each diagonal term corresponding respectively to each connection node of a low voltage branch to the medium voltage branch:, - following the aggregation step, obtaining the effective voltage state at each node of the medium voltage branch and the low voltage branches.

8. Method according to the preceding claim, in which the effective voltage states at each node of the medium voltage branch and the low voltage branches are obtained by an iterative application of the following steps, until a first predefined stopping criterion is satisfied: °a determination of an effective voltage state of the medium voltage branch, taking into account the aggregated data of each low voltage branch, to obtain effective voltages at each node of the medium voltage branch, “then, a determination of an effective voltage state of each low voltage branch to obtain effective voltages at each node of each low voltage branch, from which new values ​​of the active and reactive power pairs of each low voltage branch and of the diagonal terms of the admittance matrix of the medium voltage branch are obtained - obtaining the effective voltage state at each node of the medium voltage branch and the low voltage branches when the first stopping criterion is satisfied.

9. Method according to one of the preceding claims, characterized in that it is implemented on a computer having a multi-core architecture.

10. Computer program comprising instructions for implementing the method according to one of claims 1 to 9 when this program is executed by one or more processor(s).

11. Non-transitory recording medium readable by a computer on which is recorded a program for implementing the method according to one of claims 1 to 9 when this program is executed by one or more processor(s).

Images