Method for analyzing an electrical power distribution network aimed at reducing reactive power transmission and computer program implementing such a method

Abstract

The invention relates to a method for analyzing an electrical power distribution network aimed at reducing reactive power transmission within that network. The method comprises an initial step (S1) of processing cartographic and economic data to identify at least one overhead line in the network and a plurality of electrical loads associated with that overhead line. The method utilizes a prediction module capable of linking a description of an economic activity to reactive power consumption and, during a preparation step (S3), provides a distribution plan for reactive power compensation measures. Figure to be published with the abstract: None.

Description

Title of the invention: Method for analyzing an electrical power distribution network aimed at reducing reactive power transmission and Computer program implementing such a method. FIELD OF THE INVENTION

[0001] The present invention relates to the field of electrical power distribution networks, and more particularly to a method for analyzing and optimizing these networks aimed at reducing reactive power transmission. The invention falls within the framework of intelligent management of power distribution infrastructure, with the objective of minimizing losses related to reactive power transmission while improving network stability. TECHNOLOGICAL BACKGROUND OF THE INVENTION

[0002] The transmission of reactive power in electrical power distribution networks is a well-known problem for network operators. Reactive power, although necessary to maintain voltage in the network, contributes to significant energy losses and can lead to an overload of existing infrastructure, particularly overhead distribution lines.

[0003] The development of renewable energies is leading to a major transformation of electricity distribution networks, notably by reducing the distance between energy production and consumption sites. Unlike traditional power plants, often located far from consumption centers, renewable energy sources are often decentralized and installed closer to end users (i.e., the electrical loads to be supplied). This promotes the development of low-voltage (below 1000 volts) and medium-voltage (between 1000 volts and 50,000 volts) distribution networks. These networks are more sensitive to reactive power imbalances.

[0004] It should also be noted that reactive power compensation is becoming increasingly complex due to the progressive undergrounding of the electrical power distribution network, sometimes imposed for aesthetic and reliability reasons. Overhead lines of the distribution network, installed on pylons or poles, are inherently inductive. Conversely, underground lines, surrounded by insulating materials and buried in the ground, exhibit capacitive characteristics.

[0005] Also, in this context of transition to a decentralized network, in which buried, semi-buried or overhead lines coexist and which is dominated by low and medium voltage infrastructure, reactive power management becomes a crucial issue to ensure network stability, minimize energy losses, and guarantee a stable power quality for users.

[0006] To address these challenges, reactive power compensation methods are routinely deployed, such as capacitor banks or synchronous compensators. As an illustration, document EP2461454 describes an integrated control system within an electricity distribution network, designed to optimize the use of renewable energy sources and reactive power compensation in order to improve network performance. Document EP3245701, for its part, proposes a method for optimizing reactive power consumption in an electrical network by determining the type and configuration of compensation systems to be installed and, if necessary, activating these systems.

[0007] However, the installation of compensation means is generally opportunistic and based on a partial analysis of network needs, often limited to distribution infrastructures.

[0008] Today there is a growing demand for a more comprehensive approach to network needs, taking into account the nature and diversity of the network and promoting the installation of compensation means as close as possible to the electrical loads distributed on the network. SUBJECT OF THE INVENTION

[0009] One object of the invention is to propose a method for analyzing and / or optimizing an electrical power distribution network that provides a response to this need. BRIEF DESCRIPTION OF THE INVENTION

[0010] To achieve this goal, the object of the invention proposes a method for analyzing an electrical power distribution network aimed at reducing the transmission of reactive power in this network, the analysis method comprising: a. A step of exploiting cartographic and economic data to identify at least one overhead line of the electrical power distribution network and a plurality of electrical loads associated with at least one overhead line, the exploitation step producing a data structure associating with each electrical load, a description of an economic activity linked to this electrical load; b. A data structure processing step, the processing step aimed at completing the data structure to associate each electrical load with a reactive energy compensation requirement, the processing step exploiting a prediction module capable of linking a description of an economic activity to a reactive energy consumption; c. A step of preparing a reactive power compensation means distribution plan from the completed data structure, the distribution plan indicating the electrical loads likely to be associated with reactive power compensation means in order to reduce the reactive power transport on at least one overhead line.

[0011] According to other advantageous and non-limiting features of the invention, taken alone or in any technically feasible combination: - the means of reactive power compensation are capacitor banks; - the analysis process includes a preliminary learning step of the prediction module, based on training data, each training data associating a description of an economic activity with a historical data of reactive energy consumption. - during the operation stage, based on cartographic and economic data, an overhead line is distinguished from an underground line, a line of a given length being considered overhead if an unburied portion of the line extends over less than 90% of the given length; - during the operational phase, at least one overhead line identified in the cartographic and economic data is chosen from among medium or low voltage lines of the electrical power distribution network; - the data structure is dynamically updated allowing continuous adaptation of the reactive power compensation means distribution plan; - the description of economic activity is a code from a pre-existing classification of economic activity.

[0012] According to another aspect, the invention also relates to a computer program containing instructions adapted to the implementation of each of the steps of the process just described, when the program is executed on a computer. BRIEF DESCRIPTION OF THE FIGURES

[0013] Other features and advantages of the invention will become apparent from the detailed description of the invention which will follow by reference to the single attached figure which schematically represents the different stages of a process according to the invention. DETAILED DESCRIPTION OF THE INVENTION

[0014] This description relates to a method for analyzing and optimizing an electrical power distribution network. This method aims to reduce reactive power transmission in this network by proposing a distribution plan for reactive power compensation means. These means could be capacitor banks, for example, or synchronous compensators.

[0015] The distribution plan specifies the location of the means of reactive energy compensation, at the level of the electrical loads presenting a need for compensation, these loads being all associated with an overhead line of the electrical energy distribution network.

[0016] The reactive power compensation distribution plan brings an overall benefit to the network by limiting the transit of reactive power along the power lines that constitute it. It is therefore not simply a matter of compensating locally, and opportunistically, at the level of a single individual user and solely for their benefit, the production of this power.

[0017] In the distribution plan, the compensation means are arranged "at the end of the line", at the level of the users who are likely to benefit from them and as close as possible to the electrical loads associated with this line.

[0018] It can be foreseen, within the framework of the present invention, that these means are activated and deactivated according to the instantaneous compensation needs, by a compensation management system, which brings an overall benefit by limiting the transit of reactive energy throughout the power line.

[0019] In addition to reducing losses due to Joule effect, this limits the stress on transformers used to stabilize network voltage.

[0020] The analysis and optimization process that is the subject of this description can take the form of a computer program containing instructions adapted to the implementation of each of the steps of this process, when the program is executed on a computer.

[0021] The method uses freely available cartographic and economic data to identify a line in a network and the electrical loads associated with it. These electrical loads correspond to network users, whose economic activities can also be described in the cartographic and economic data.

[0022] The intuition that led to the present invention is that it is possible to establish a link between an economic activity and the need for reactive power compensation. By way of illustration, the distribution by sector of activity of sites exhibiting reactive power consumption from a sample of 141 establishments supplied with medium voltage is presented in the following table.

[0023] Quarry operation 4

[0024] Agri-food 5

[0025] Industry 19

[0026] Waste Disposal Site 1

[0027] Construction 7

[0028] Commerce 16

[0029] Logistics and warehousing 6

[0030] Hotel chain 6

[0031] Restoration 4

[0032] Miscellaneous 9

[0033] Telecommunications 2

[0034] Bank 5

[0035] Real Estate 7

[0036] R&D 2

[0037] Public Administration 25

[0038] High School 4

[0039] Hospitals 13

[0040] Stage 6

[0041] An underlying idea of ​​the invention is to prepare a computer prediction module, this prediction module being capable of linking a description of an economic activity to a reactive energy consumption.

[0042] By way of illustration, in the extraction sector, a quarry is characterized by a well-established process: belt conveyor, crusher, grinder, washing pump, etc. This extraction process does not vary according to geographical location and is therefore essentially identical in other regions. Data collected in a particular region can thus be used to provide predictions in other regions. Additional data to the description of economic activity can be used to adjust the parameters of the prediction module for an estimate of the reactive energy consumption supplied, such as, for example, the age of the site, and / or its area, or other freely available cartographic and economic data.

[0043] For ease of implementation, the description of the economic activity may correspond to a code from a pre-existing classification of economic activity. For example, this description may be the APE code (Main Activity Performed) extracted from the French Activity Classification (NAF). As the 1TNSEE website points out, this NAF classification is a classification of productive economic activities, primarily developed to facilitate the organization of economic and social information. It has the same structure as the European NACE classification of activities, itself derived from the international ISIC classification.

[0044] It should be noted that the APE code, preferably chosen for deploying a process according to the invention, is assigned to every company in France. Registers of this economic data are publicly available, making it easy to extract the required information (company identifier, address, APE activity code, etc.). Of course, for the implementation of the present invention, other data than the APE code may be chosen as a description of economic activity.

[0045] Returning to the description of the prediction module, it is therefore capable of predicting the reactive power consumption of a site based on the APE code corresponding to the economic activity associated with that site. The term "predicting reactive power consumption" does not necessarily mean that this prediction is expressed precisely in reactive power consumed, in units of reactive power (kilovolt-ampere-hour - kVARh), which of course depends on the site's level of activity. In the context of this description and according to certain implementation methods, it may involve providing, in arbitrary units, a relative quantity intended to determine whether a site is likely to produce reactive power. The larger this quantity, the greater the site's need for reactive power compensation. It will therefore be a prime location for installing reactive power compensation systems.

[0046] In certain implementations, however, this prediction can be quantitative and expressed, for example, in relation to the site's electricity consumption. For example, the prediction module can indicate that the reactive power consumption of a site identified by its APE code represents 40% of its total electricity consumption.

[0047] In other embodiments, the prediction can be expressed quantitatively in units of reactive energy over a given period, for example, over a period of one year. For this purpose, the prediction module can rely on historical energy consumption data for a site, to which a relative estimate of reactive energy consumption is applied in relation to total energy consumption. It should be noted that this historical consumption data is generally publicly available (for example, in France, this data is prepared and provided by the French Energy Network Operators Agency).

[0048] In addition to the description of economic activity, the prediction module may include other input elements, such as the geographical location of the site. Advantageously, these other input elements are also available in the form of freely accessible cartographic and / or economic data.

[0049] The prediction module can be configured in different ways, depending on specific needs and available data. Preferably, it is configured by a The learning process is supervised, meaning it uses pre-existing training data. This training data combines a description of an economic activity with historical reactive energy consumption data, allowing the module to build robust and relevant predictive models. As previously mentioned, it can incorporate input elements other than the description of the economic activity, such as the age of the site and / or its surface area.

[0050] More specifically, during the training phase, each training data point consists of a vector characterizing at least one economic activity descriptor (for example, the APE code) associated with historical quantitative or qualitative reactive energy consumption data. The module uses this data to extract a model of the underlying relationships between economic activity and reactive energy.

[0051] To improve its predictive capabilities, the module can implement an advanced machine learning technique, such as an artificial neural network. This approach makes it possible to capture complex nonlinear relationships that are difficult to model using traditional methods. The neural network is trained on historical data and uses backpropagation algorithms to adjust its internal parameters (synaptic weights) in order to generalize a reliable relationship from the training data.

[0052] In addition to, or as a replacement for, neural networks, other learning techniques can be used, such as random forests, Bayesian networks, or multivariate regression models, depending on the complexity of the data and the nature of the correlations observed between the different descriptors of economic activity, any other input data, and reactive energy. These techniques are configurable and adjusted according to the required performance, whether predictive accuracy, speed of execution, or adaptability to new data.

[0053] Regardless of how it was previously configured in a step prior to the implementation of a process according to the invention, for this implementation, there is a prediction module capable of linking a description of an economic activity to a consumption of reactive energy.

[0054] Continuing the very general description of the present invention, and as reported in the introduction to this application, a distribution plan for reactive power compensation means is relevant when these means are associated with an overhead line of the electrical power distribution network, that is to say, with a line having an inductive character.

[0055] It is for this reason that the invention proposes to distinguish, in cartographic and economic data, the overhead lines from the underground lines of the network. For the purposes of this description, and preferably, a line of a given length is considered "aerial" if at least 90% of its length is above ground. A different threshold than the 90% proposed here could, of course, be chosen.

[0056] For the same reasons of efficiency, the invention is proposed to preferentially focus on medium voltage or low voltage lines.

[0057] Another underlying idea of ​​the invention is to use mapping data of the public electricity grid to locate an overhead line of interest and, where appropriate, to install reactive power compensation devices at the most suitable sites associated with that line. These most suitable sites are those of users, operating an electrical load associated with the analyzed line, who have the most obvious need for reactive power compensation. It should be noted that the sites associated with the analyzed line can also be identified in the mapping data of the electricity grid or inferred from this mapping data. It should also be noted that this mapping data makes it possible to identify the location of the sites along the line, and therefore the distance of energy transmission from its point of generation.Losses, and in particular Joule effect losses, are directly related to this transport distance, so this information can be used to preferentially seek compensation for sites where this distance is relatively large.

[0058] Within the scope of the invention, several data sources can naturally be used to collect the cartographic and economic data necessary for the envisaged processing. For example, data relating to the location of power lines can be extracted from a first database, and data relating to the location of sites and their respective economic activities extracted from a second, separate database. This information can be cross-referenced, and based on the geographical proximity of a site and a power line, it can be assumed that the site is indeed electrically connected to the line in question.

[0059] With reference to [Fig. 1], a detailed description of the different stages of the process of analysis and optimization of the electrical power distribution network aimed at reducing the transport of reactive power in this network is now provided.

[0060] During a first SI step, called the "exploitation" step, the cartographic and economic data that have been collected are used to identify: - at least one overhead line of the electrical power distribution network; - a plurality of electrical loads associated with at least one overhead line.

[0061] As previously indicated, this processing step may involve cross-referencing data from separate databases. It may be assisted by an operator, for example, to select, among several overhead power lines, the line or plurality of lines that will be analyzed in the following steps of the process.

[0062] Importantly, during the SI operation stage, aerial lines are distinguished from underground lines based on cartographic and economic data because, to be relevant, the analysis is carried out on the aerial lines.

[0063] Also importantly, the overhead line identified in the cartographic and economic data is chosen from among medium or low voltage lines of the electrical power distribution network.

[0064] This IS processing step leads to the production of a data structure associating, with each electrical load linked to the analyzed overhead line, a description of an economic activity related to that electrical load. This data structure can take any suitable form, for example as a list, a table, a collection of vectors, or even as a collection of pointers, as is perfectly understood by a person skilled in the art.

[0065] As already mentioned, this data structure can include other information extracted from or derived from cartographic and economic databases. In particular, it is possible to supplement the structure, for each entry, with the surface area of ​​the corresponding site, its age, the transport distance separating the electrical load from the energy production site, etc.

[0066] In a second step S2 called "processing", which follows the first step SI, the data structure provided by this first step SI is processed, using the prediction module presented in an earlier passage of this description.

[0067] The S2 processing step aims to complete the data structure and associate, with each electrical load at a site, a reactive power compensation requirement. To this end, the prediction module is applied to the data of the structure, and this structure is supplemented, for each electrical load at a site, with the compensation requirement returned by this module.

[0068] As already presented, this compensation requirement can be provided in the form of qualitative information (according to a predetermined needs scale), in a relative quantitative form (such as a percentage of the electrical energy consumed by the site's electrical load), or even in an absolute quantitative form. In the latter case, the prediction module can rely on a database containing a history of electrical energy consumption.

[0069] In a subsequent step S3, called the "preparation" step, the reactive power compensation means distribution plan is prepared from the data structure completed during the processing step. As previously explained, this distribution plan indicates the electrical loads likely to be associated with means of reactive power compensation with a view to reducing the transmission of reactive power on at least one overhead line.

[0070] In its simplest implementation, this preparation step can thus identify, within the data structure produced at the end of the S2 processing step, the inputs associated with the most significant reactive power compensation needs. The distribution plan then consists of a list of sites where the identified electrical loads are located, and the capacity of the compensation means to be implemented at these sites.

[0071] Other, more complex implementation methods are of course possible. For example, the distribution distance associated with each input of the data structure can be taken into account to estimate the Joule losses occurring in the analyzed line, and to identify the electrical loads which, once equipped with compensation means, will allow for the best possible reduction of reactive power transport on at least one overhead line.

[0072] Other constraints can be added during this distribution plan preparation stage, such as sites or types of sites (identified, for example, by their APE code) that are not to be equipped with compensation systems or that have a compensation limit. These may be sites or types of sites for which the quality of the delivered energy is critical and which could be degraded by the presence of compensation systems. For example, it could be required that "data center" type sites be excluded from processing during this distribution plan preparation stage.

[0073] More generally, the distribution plan may recommend a type of compensation method best suited to the nature of the site (capacitor banks or synchronous compensators, for example), or even a specific model. During this preparation stage, an installation and / or operating cost may also be associated with the compensation methods proposed by the distribution plan.

[0074] The preparation stage can then establish a distribution plan aimed at optimizing separately or jointly the overall cost of installation and / or operation of the compensation means (overall cost which we seek of course to minimize) and the losses avoided, in particular Joule losses, by the installation of the compensation means (losses avoided which we seek of course to maximize).

[0075] It is therefore understood that, in a general approach, this step of preparing the distribution plan can implement complex processing of the data structure, for example a multi-criteria optimization.

[0076] In addition to the distribution plan, the preparation stage may provide other data, such as losses avoided through the installation of compensation means in the plan. proposed distribution, or the cost associated with the deployment of this plan and / or its operation.

[0077] The analysis method can be implemented on an ad hoc basis to develop a distribution plan for reactive power compensation means, in order to plan its deployment. In this case, the execution of the method is followed by a step of installing the compensation means at the electrical loads recommended by the distribution plan.

[0078] However, it is also possible to update the data structure dynamically, thus enabling continuous adaptation of the reactive power compensation distribution plan. In this way, network performance can be optimized over time and in response to changes in the network and its uses.

[0079] Of course the invention is not limited to the modes of implementation described and alternative embodiments can be made without departing from the scope of the invention as defined by the claims.

[0080] Thus, during and at the end of each step of the process, the data being processed can be displayed in text, graph, or mixed form, allowing the user to access this data. In certain implementations, the user can interact with the presented data, for example, to add or remove a site from an electrical load or to choose the power line to analyze, if the mapping data allows for the analysis of several lines.

Claims

Demands

1. A method for analyzing an electrical power distribution network to reduce reactive power transmission in that network, the analysis method comprising: a. A data processing step (S2) of cartographic and economic data to identify at least one overhead line of the electrical power distribution network and a plurality of electrical loads associated with at least one overhead line, the processing step producing a data structure associating each electrical load with a description of an economic activity linked to that electrical load; b. A data structure processing step (S2), the processing step aimed at completing the data structure to associate each electrical load with a reactive power compensation requirement, the processing step using a prediction module capable of linking a description of an economic activity to a reactive power consumption; c.A preparation step (S3) of a reactive power compensation means distribution plan from the completed data structure, the distribution plan indicating the electrical loads likely to be associated with reactive power compensation means in order to reduce the reactive power transport on at least one overhead line.

2. An analysis method according to the preceding claim in which the reactive power compensation means are capacitor banks.

3. An analysis method according to any one of the preceding claims comprising a preliminary learning step of the prediction module, based on training data, each training data associating a description of an economic activity with a historical reactive energy consumption data.

4. An analysis method according to any one of the preceding claims, wherein during the exploitation step (IS), an overhead line is distinguished from cartographic and economic data of an underground line, a line having a given length being considered aerial if an unburied portion of the line extends over at least 90% of the given length.

5. An analysis method according to any one of the preceding claims, wherein during the operating step (SI), at least one overhead line identified in the mapping and economic data is selected from among medium or low voltage lines of the electrical power distribution network.

6. An analysis method according to any one of the preceding claims, wherein the data structure is dynamically updated, allowing continuous adaptation of the reactive power compensation means distribution plan.

7. An analysis method according to any one of the preceding claims, wherein the economic activity description is a code from a pre-existing economic activity classification.

8. Computer program containing instructions adapted to carry out each of the steps of a process according to any one of their claims 1 to 7, when the program is executed on a computer.