How to build a wind farm in a given space

Abstract

The present invention relates to a method for constructing a wind farm from a given number of wind turbines in a given space. The method comprises two discrete distributions (RD1, RD2), one for wind speed and one for wind direction, and a first discrete grid (RD3) in the given space. The method also includes a probability of occurrence (Prob) of each discrete wind speed value for each discrete wind direction. The method uses a first wind turbine arrangement in the given space, after which the positions of the wind turbines are changed one by one (i) by determining discrete positions (PDP_i) around the position to be changed. The selected positions (Pos_i) are those that maximize the annual energy generated by the wind farm. The present invention also relates to a computer program product implementing this method, and to a wind farm obtained with this method.

Description

[Technical Field]

[0001] The present invention relates to a method for constructing a wind power plant in a given space. [Background technology]

[0002] To solve environmental challenges, wind farms have emerged. These wind farms consist of several wind turbines placed apart from one another within a delimited space. This delimited space can be on land or at sea. A distinction is thus made between onshore wind farms and offshore wind farms, i.e. wind farms located at sea.

[0003] The wind turbines at these power plants are generally horizontal axis wind turbines with a system to orient their horizontal axis into the wind to maximize the energy captured by the turbine. In some cases, the wind turbines are designed to automatically orient themselves into the wind.

[0004] Vertical axis wind turbines are also known and have the advantage that they do not need to be oriented into the wind.

[0005] In wind farms, the wakes generated by turbines can have the effect of reducing wind speed downstream of the turbines, thereby reducing the energy recovered by other turbines, particularly those downstream of the turbines generating these wakes. The location of wind turbines within a farm is therefore important to maximise the energy recovered by the farm.

[0006] Furthermore, the location selected for locating a wind farm may exhibit varying local wind characteristics. Indeed, wind direction and wind speed are prominent parameters that may change over time at the location under consideration. In particular, these characteristics can be obtained by sensors placed at defined locations and maintained at those locations for months or years in order to obtain sufficient statistical data to characterize the wind resource at the selected location. These sensors may in particular be anemometers placed at a sufficient altitude (on the order of 100 m above the ground) to characterize the wind experienced by the turbine (i.e., wind at substantially the height of the rotor axis). Wind data can also be obtained using LiDAR (Light Detection and Ranging) technology.

[0007] Statistical knowledge of the wind at the site being considered for the power plant makes it possible to obtain, in particular, the wind speed distribution, the wind direction distribution and the probability of the simultaneous occurrence of certain wind speeds in a given direction.

[0008] Therefore, to maximize the annual energy produced by a power plant, the turbines must be positioned in an optimal manner at the planned site for the power plant.

[0009] "Annually produced energy" or "annual energy production" is understood to be the total average energy produced by a power plant and therefore by all wind turbines of the plant. This average energy, estimated taking into account statistical wind data (distribution of wind speeds and directions, probability of occurrence), is based on a period of one year, hence the term "annual", in order to avoid seasonal influences that may distort the results. In fact, the wind, whatever its speed and direction, varies greatly from season to season.

[0010] The mean energy is obtained by knowing the wind direction distribution, the wind speed distribution and the probability of occurrence of the wind speed in a given direction.

[0011] Of course, the annual energy produced will also depend on the wind turbine selected.

[0012] Several methods have been developed to determine the correct placement of wind turbines at the site selected for the power plant.

[0013] Patent application CN 105 119 320 relates to a method based on evolutionary algorithms. This type of algorithm is stochastic and uses random processes. This type of algorithm requires a large number of trials to obtain a result, which generates considerable calculation time and therefore requires a large amount of computer memory.

[0014] Patent applications CN 102 142 103A (Patent Document 2), CN 105 139 269 (Patent Document 3), and US 2016-171401 (Patent Document 4) relate to wind turbine positioning methods based on genetic algorithms. Genetic algorithms are a type of evolutionary algorithm. They require multiple evaluations from possible (continuous) actual data and crossover of these evaluations. Therefore, genetic algorithms are complex and require significant computation time and large amounts of computer memory.

[0015] The methods described in the following documents are also known.

[0016] "Wind farm layout optimization with a three-dimensional Gaussian wake model," Tao et al., Renewable Energy, June 9, 2020 (Non-Patent Document 1): This method requires the planned power plant location to be bounded by a rectangle. In other words, this method is not suitable for power plants with non-convex or non-connected spaces.

[0017] - "Continuous adjoint formulation for wind farm layout optimization," Antonini et al., Applied Energy, July 25, 2018 (Non-Patent Document 2): This method is based on analytical methods for evaluating wake effects. It is therefore a continuous optimization method based on CFD (computational fluid dynamics) calculations, which requires long calculation times and large amounts of computer memory.

[0018] - "A fast and effective local search algorithm for optimizing the placement of wind turbines," Wagner et al., Renewable Energy, October 10, 2012 (Non-Patent Document 3): This method requires that the locations are bounded by a rectangle. It is therefore not suitable for locations with non-convex or non-connected shapes.

[0019] "Solving the wind farm layout optimization problem using random search," Feng and Shen, Renewable Energy, January 20, 2015 (Non-Patent Document 4): This method requires that the locations are bounded by polyhedra, and is therefore not suitable for locations with non-connected shapes.

[0020] "Greedy robust wind farm layout optimization with feasibility guarantee," Quan and Kim, Engineering Optimization, September 6, 2018 (Non-Patent Document 5): This method is based on a greedy algorithm that considers all positions that are the minimum distance from wind turbines already placed at that location to determine the location of the next turbine. This method therefore requires long calculation times and large computer memories, especially when the location selected for the power plant is finely discretized, the space in which wind turbines must be placed is large, and / or the number of turbines to be placed is large. [Prior art documents] [Patent documents]

[0021] [Patent Document 1] Chinese Patent Application Publication No. 105 119 320 [Patent Document 2] Chinese Patent Application Publication No. 102 142 103 [Patent Document 3] Chinese Patent Application Publication No. 105 139 269 [Patent Document 4] US Patent Application Publication No. 2016 / 0171401 [Non-patent literature]

[0022] [Non-Patent Document 1] Tao et al., "Wind farm layout optimization with a three-dimensional Gaussian Wake model," Renewable Energy, June 9, 2020 [Non-patent document 2] Antonini et al., "Continuous adjoint formulation for wind farm layout optimization," Applied Energy, July 25, 2018 [Non-patent document 3] Wagner et al., A fast and effective local search algorithm for optimizing the placement of wind turbines, Renewable Energy, October 10, 2012 [Non-patent document 4] Feng and Shen, "Solving the wind farm layout optimization problem using random search," Renewable Energy, January 20, 2015. [Non-Patent Document 5] Quan and Kim, Greedy robust wind farm layout optimization with feasability guarantee, Engineering Optimization, September 6, 2018 [Non-patent document 6] Katic, Hojstrup and Jensen, "A simple model for Cluster Efficiency," EWEC 1986 [Non-Patent Document 7] Jensen, "A note on wind generator interaction," DTU, 1983 Summary of the Invention [Problem to be solved by the invention]

[0023] The technical problem to be solved by the present invention is therefore to develop a method for constructing wind farms that allows optimal positioning of wind turbines in locations with complex shapes, such as non-convex and / or disconnected zones, in order to maximize the total energy generated by the wind farm while minimizing the required computation time and computer memory. [Means for solving the problem]

[0024] The present invention therefore relates to a method for constructing a wind power plant (in English, wind turbine farm or wind farm) from a predetermined number of wind turbines in a predetermined space, the method comprising a first discrete wind speed distribution, a second discrete wind direction distribution, and occurrence probabilities of each wind direction value and each wind speed value of the first and second discrete distributions. Furthermore, the predetermined space is divided into a first discrete grid. The method comprises at least the following sequence of steps a) to c): a) determining a first wind turbine location in a first discrete grid of a predetermined space by a first location determination algorithm; b) determining a sequential order for correcting the wind turbine position determined by the first positioning algorithm, and then repeatedly performing at least the following steps b1) to b4) for each of the at least one wind turbine to be relocated in the determined sequential order; b1) for each wind turbine to be relocated, determining new possible discrete positions of the turbine to be relocated; b2) calculating an average annual energy production of a predetermined number of wind turbines for each possible discrete location of the wind turbines to be relocated from the first discrete wind speed distribution, the second discrete wind direction distribution and the occurrence probability; b3) selecting a position of the to-be-relocated wind turbine corresponding to the calculated maximum annual energy production and locating the to-be-relocated wind turbine there; b4) determining a new position of the wind turbine within the predetermined space, where only the turbine to be relocated corresponds to the position to which it has already been moved based on condition b3); c) determining a final fix position corresponding to the final configuration obtained;

[0025] The wind power plant (wind turbine farm or wind farm in English) is then constructed by erecting (installing / constructing) the wind turbines (a predetermined number of wind turbines) at determined locations in the final arrangement on a predetermined spatial physical site in order to generate energy, e.g. electricity, from the wind.

[0026] The invention also relates to a computer program product implementing the method defined above and to a wind power plant obtained using the method defined above.

[0027] The present invention relates to a method for constructing a wind power plant, the method including a first discrete wind speed distribution, a second discrete wind direction distribution, and occurrence probabilities of each discrete wind speed value at each discrete wind direction value of the first and second discrete distributions, and the predetermined space is divided into a first discrete grid. In this method, at least the following series of steps a) to c) are performed: a) determining a first location of the wind turbines in a first discrete grid of the predetermined space by a first location determination algorithm; b) determining a sequential order for correcting the wind turbine position determined by the first positioning algorithm, and then repeatedly performing at least the following steps b1) to b4) for each of the at least one wind turbine to be relocated in the determined sequential order; b1) for each wind turbine to be relocated, determining new possible discrete locations of the turbine to be relocated by developing a grid of at least a first predetermined length, the grid being centered on the location of the turbine to be relocated, the grid being divided into a predetermined number of cells, the new possible discrete locations including cell intersections, the cell intersections being at a minimum distance from locations of other wind turbines within the predetermined space; b2) calculating an average annual energy production of a predetermined number of wind turbines for each possible discrete location of the wind turbine to be relocated from the first discrete wind speed distribution, the second discrete wind direction distribution, and the occurrence probability; b3) selecting a position of the wind turbine to be relocated that corresponds to the maximum annual energy production calculated in step b2); b4) determining a new arrangement of wind turbines within said predetermined space; c) determining a final determined location corresponding to the obtained final configuration for generating energy from wind power, and constructing said wind farm by erecting a predetermined number of wind turbines at the determined locations of the final configuration within said predetermined spatial physical site.

[0028] According to one implementation of the present invention, prior to step a), statistical wind data is collected by a collection means, preferably a LiDAR sensor, to determine the first and second discrete distributions and the occurrence probability of each wind speed for each wind direction in the first and second discrete distributions.

[0029] According to one configuration, the predetermined space is two-dimensional.

[0030] Preferably the sequential order is obtained randomly.

[0031] Preferably, step b2) is repeated several times, changing the sequence order for each iteration.

[0032] More preferably, for each iteration of step b), a first predetermined length is selected that is shorter than the first predetermined length of the previous iteration, said predetermined number being the same between iterations, and the iteration of step b) is stopped when the first predetermined length is less than a first threshold.

[0033] According to one implementation of the invention, the predetermined space comprises a non-bonding zone.

[0034] According to one variant of the invention, said predetermined space comprises a non-convex zone.

[0035] Advantageously, said first position determination algorithm comprises at least: arbitrarily determining a position of a first wind turbine; Then, for each wind turbine to be placed, successively performing the steps of: determining possible positions within the first discrete grid for the wind turbine to be placed; calculating annual energy production of the placed wind turbine and the turbine to be placed for the determined possible positions from the first discrete wind speed distribution, the second discrete wind direction distribution and the occurrence probability; and selecting a position of the wind turbine to be placed that corresponds to the calculated maximum annual energy production. determining the first configuration corresponding to positions of a predetermined number of wind turbines within the predetermined space; Run The possible locations include discrete locations of the first discrete grid located between a minimum distance and a maximum distance from all deployed turbines, and / or discrete locations on a boundary of a predetermined space located at the minimum distance from all deployed turbines.

[0036] According to one aspect of the invention, the arbitrary position corresponds to the maximum sum of the coordinates of the positions of said first discrete grid.

[0037] The invention also relates to a computer program product for carrying out the above-mentioned method from computer means.

[0038] The invention also relates to a wind power plant resulting from the method for constructing a wind power plant described above.

[0039] Other characteristics and advantages of the method and system according to the invention will become apparent from reading the following description of an embodiment given as a non-limiting example, with reference to the accompanying drawings, in which: [Brief explanation of the drawings]

[0040] [Figure 1] 1A-1D are various overviews of a position determination method according to the present invention; [Figure 2] 1A-1D show detailed views of the various steps of a first method for constructing a wind power plant according to the invention; [Figure 3] 3A-3C show detailed views of the various steps of a second method for constructing a wind power plant according to the invention; [Figure 4] 1 shows an example of a complex location for determining the position of a wind turbine according to the present invention; [Figure 5] FIG. 10 illustrates a variation of the first positioning algorithm according to the present invention. [Figure 6] FIG. 2 shows a first example of application of the first position determination algorithm according to the invention. [Figure 7] FIG. 2 illustrates a second example of application of the first position determination algorithm according to the present invention. [Figure 8] FIG. 1 illustrates a local search method for relocating turbines of a wind farm according to the present invention. [Figure 9] 1 illustrates the various steps of a local search method for relocating turbines of a wind farm according to the invention; [Figure 10] 1 illustrates the different steps of a method for building a wind farm on a complex location with non-connected and non-convex zones; [Figure 11] FIG. 11 shows a variation of the method for constructing a wind farm in the same complex location as in FIG. 10. DETAILED DESCRIPTION OF THE INVENTION

[0041] To facilitate the reading of this specification, some definitions are provided below.

[0042] A greedy algorithm is understood to be an algorithm that consists in developing a local optimum step by step, i.e., in the case of a wind farm, placing each turbine one after the other until all wind turbines are placed in the given space.

[0043] An evolutionary algorithm is understood to be a biologically inspired algorithm that consists in evolving a set of solutions to obtain better results. It is therefore stochastic and uses repeated random processes.

[0044] A genetic algorithm is understood to be an evolutionary algorithm that uses the concept of natural selection. In particular, this type of algorithm is able to combine or change some parameters of previous solutions in order to improve the result.

[0045] Non-connected zones are understood to be zones in which there are at least two points that cannot be connected by a continuous path that is entirely contained within the zone in question, whereas connected zones are zones in which every pair of points is connected by a continuous path that is entirely contained within the zone.

[0046] A convex zone is understood to be a zone in which all line segments connecting any two points in this zone are completely contained in this zone. For example, a circle, a square or a rectangle may delimit a convex zone.

[0047] Conversely, a non-convex zone is one in which there are at least two points connected by a line segment that is not entirely contained within the zone. For example, a zone bounded by concentric outer and inner circles is not convex.

[0048] The present invention relates to a method for arranging a predetermined number of wind turbines in a predetermined space or a method for constructing a wind farm, the method comprising: a first discrete wind speed distribution; a second discrete wind direction distribution; and a probability of occurrence of each discrete wind speed value at each discrete wind direction value of the first and second discrete distributions, wherein the predetermined space is divided into a first discrete grid. The discrete grid is understood to mean that the predetermined space is divided into segments, e.g., rectangular or square shaped, and the discrete grid is made up of points determined by this grid. For example, the first discrete grid can be made up of points defining the boundaries of the segments (e.g., four points on each rectangular or square segment) and / or intersections of the segments and / or centers of the grid cells (e.g., centers of the square or rectangular segments).

[0049] The use of discrete values ​​for wind speed and direction and discrete points in a first discrete grid in a given space simplifies the method, accelerates the calculation time, and limits the computer memory required compared to methods using continuous real data. Indeed, while continuous real data (continuous methods are discussed below) provide an infinite number of solutions, the discrete values ​​make it possible to limit the number of possible combinations (combinatorial methods are discussed below). Thus, by combining the first and second discrete distributions with the first discrete grid, good accuracy can be obtained in locating various wind turbines in a given space while limiting the calculation time required to determine their positions.

[0050] The probability of occurrence of each wind speed for each wind direction is used, inter alia, to calculate the average annual energy production. This probability can be obtained, inter alia, from a wind rose corresponding to a given space, which is well known to those skilled in the art. The wind rose can be generated, inter alia, using a wind speed and direction collecting means, such as an anemometer located on a mast at a sufficient altitude (e.g., between 80 m and 120 m, so as to be substantially at the hub height of the wind turbine), or a vertically oriented LiDAR (Light Detection and Ranging) sensor located close to the ground. This collecting means is maintained in place for an extended period of time, for several months, ideally for more than a year, so as to take into account seasonal variations in wind characteristics. In other words, the method can include a step of installing a collecting means, such as an anemometer, a measuring mast, or more advantageously a LiDAR sensor or any similar means, at a physical site in a given space for a given period of time to collect wind data at that physical site. To obtain data related to all four seasons, the given data collection period can be at least one year. Thus, a step of statistical wind data collection from at least one collection means may be provided.

[0051] In this method, at least the following series of steps a) to c) are performed.

[0052] a) determining a first arrangement of the wind turbines in a first discrete grid in the given space by a first positioning algorithm. Preferably, this first positioning algorithm can be an optimization algorithm that makes it possible to obtain a first distribution that allows for a large annual energy production. Advantageously, this first positioning algorithm can be a greedy algorithm that positions each wind turbine one after another so as to maximize the annual energy generated. The first wind turbine can be arbitrarily positioned in the given space (at discrete values ​​of the first discrete grid). The second wind turbine will be positioned at a discrete position of the first discrete grid so as to maximize the annual energy production of the two turbines. The position selected for the nth wind turbine corresponds to a discrete position of the first discrete grid that maximizes the annual energy production of the n turbines. Using a greedy algorithm makes it possible to obtain in a simple way a first arrangement of the wind turbines in the given space that can be used to initialize the local search optimization method of the following step b).

[0053] b) determining a sequential order for correcting the positions of the wind turbines determined by the first positioning algorithm, and then repeatedly performing at least steps b1) to b4) below for each of the at least one wind turbine (preferably all wind turbines of the predetermined number of turbines) to be relocated in the determined order.

[0054] b1) for each wind turbine to be relocated, determining new possible discrete positions of the turbine to be relocated by developing a grid of at least a first predetermined length (e.g., the grid may be a first square, the first predetermined length corresponding to a side of the first square). The grid is centered on the position of the turbine to be relocated (whereby the position of the turbine to be relocated corresponds, for example, to the center of the first square), the grid is divided into a predetermined number of cells (where the cells may be, for example, a second square), and the new possible discrete positions include intersections of cells (e.g., of the second square) that are located within the predetermined space and at a minimum distance from positions of other wind turbines. In other words, what is sought here are positions where the turbine can be installed that are close to the wind turbine to be relocated. Positions that are too close to other turbines or outside the predetermined space (outside the zone provided for installation of wind turbines) are therefore eliminated. Compared to using continuous actual positions, which would unnecessarily increase the calculation time and the load on the computer memory used, the use of a grid (e.g., a first square) allows to globally delimit the neighborhood of the turbine to be relocated, and dividing the grid (e.g., the first square) into cells allows to obtain discrete positions of this neighborhood. Furthermore, since the discretization of the cells can be advantageously finer than the discretization of the first discrete grid, a more accurate local precision can be obtained for relocating the wind turbine than the positioning obtained in the first placement based on the first discrete grid. Preferably, the minimum distance is greater than two times the diameter of the turbine rotor, more preferably greater than four times the diameter of the turbine rotor.

[0055] b2) calculating the average annual energy production of a predetermined number of wind turbines for each possible discrete position of the wind turbine to be relocated from the first discrete wind speed distribution, the second discrete wind direction distribution and the occurrence probability. Of course, to calculate the annual energy production, the characteristics of the wind turbines, in particular the swept area and power coefficient of the rotor, are required. This annual energy production also takes into account wake effects generated by the turbines, which may cause power losses in downstream or adjacent turbines.

[0056] b3) selecting (discrete) positions of the wind turbines to be relocated corresponding to the calculated maximum value of the annual energy production calculated in step b2).

[0057] b4) determining a new arrangement of the wind turbines within the defined space, the new arrangement including the changed positions of the relocated turbines and the changed positions of the turbines to be relocated;

[0058] c) determining a final location corresponding to the obtained final configuration to obtain a wind farm for generating energy from wind, and once the final configuration is determined, constructing the wind turbines according to the final configuration within the predetermined space (the physical site of the predetermined space) (in other words, the wind farm is constructed by installing the wind turbines in the predetermined space, i.e., within the physical site of the predetermined space, at the location of the final configuration, and said wind farm will generate energy, e.g., electricity, from wind.

[0059] In other words, step b) is a local search optimization method based on discrete values ​​around the initial position of the wind turbine to be relocated. Indeed, the turbine positions are modified one by one from the first configuration obtained in step b) by searching for discrete positions where the turbine could be located locally around the initial position (i.e., excluding discrete positions located at a smaller distance than other wind turbines and discrete positions outside the boundary or predetermined space). The local search optimization consists of evaluating the annual energy generated for each new possible configuration. That is, the local search optimization consists of maintaining the positions of all wind turbines other than the wind turbine to be relocated and considering possible discrete positions of the turbine to be relocated. The selected position is the one that maximizes the annual energy produced. Once a position is selected, this position is maintained for the next turbine to be relocated. In other words, to continue modifying the turbine position, a new configuration based on the already modified turbine position is used.

[0060] Step c) allows to construct a wind farm in which the turbine positions are optimized to maximize the total energy recovered by the wind turbines.

[0061] The method for placing wind turbines in a given space or for building a wind farm according to the present invention therefore uses optimization based on a pure combinatorial approach.

[0062] Steps a) and b) can be carried out by computer means, in particular by a computer, a mobile phone or a tablet.

[0063] According to one embodiment, the annual energy production can be estimated using the following formula:

[0064]

number

[0065] where aep is the annual energy production of the wind farm,

[0066]

number

[0067] is the wind speed w s Wind and direction p is the expected value of the total power generated by a wind farm (for a given number of wind turbines in a given space) for each wind direction w. The total power generated by a wind farm can therefore be calculated for each wind direction w, for example by a Weibull distribution. p Each wind speed w s Taking into account the statistical distribution of

[0068] The rotor of the wind turbine is facing the wind direction. p If the wind farm is systematically oriented in a plane perpendicular to the , the total power generated by the wind farm can be expressed as:

[0069]

number

[0070] where N is the predetermined number of wind turbines in a power plant within a given space, P f is each wind direction w p Each wind speed w s is the instantaneous power delivered by each turbine f in the power plant relative to

[0071] Some wind turbine rotors are moving in the direction of the wind p If the wind turbine is located in a plane that is offset from the plane perpendicular to the wind (i.e. the turbine is not positioned facing the wind but offset from it), a correction factor can be taken into account to account for the effect of the offset. This correction factor can be obtained in particular from CFD (Computational Fluid Dynamics) simulations.

[0072] The instantaneous power P of each wind turbine f in the wind farmf can be written as follows:

[0073]

number

[0074] where ρ is the density of air, S is the area swept by the rotor of the turbine f, and v f (w s ) is the wind speed at the rotor of turbine f, C pf is the wind speed at the rotor, v f (w s ) is the power coefficient of the turbine f, which depends on the power coefficient C pf is a characteristic of turbine f.

[0075] Indeed, the wake effects of wind turbines located upstream and / or to the side of turbine f can affect the energy production of turbine f. These wakes can cause wind speed reduction and / or wind turbulence at turbine f. The impact of these wake effects is due to the velocity v at the turbine rotor. f The speed w s no longer corresponds to the power coefficient C pf is also affected.

[0076] The influence of wake effects considered in the above equations can be based in particular on wake models. These wake models can show in particular that: - reduction of the wind speed upstream of the turbine f due to the wake from the upstream turbine. Such a model known to those skilled in the art is the Jensen wake model (described in the publication "A simple model for Cluster Efficiency", Katic, Hojstrup and Jensen, EWEC 1986, in particular paragraph 2.1); - Increased wind turbulence intensity; and / or the superposition of the wakes of several upstream turbines on a single turbine f, as described in the publication "A note on wind generator interaction", Jensen, DTU, 1983 (Non-Patent Document 7), which can combine the effects of the speed reduction and / or the turbulence intensity of the wind in several wakes, as in the aforementioned publication.

[0077] The wake model can also be determined from CFD calculations.

[0078] Preferably, the determined sequential order is different from the order in which the wind turbines were positioned to obtain the first positioning, for example via a greedy algorithm, thus improving the accuracy of the optimization. Step b) corresponds to a second algorithm for optimizing the position of each wind turbine within the given space.

[0079] c) A final arrangement is determined corresponding to the final arrangement obtained. The final arrangement corresponds to the positions of the wind turbines after all of the wind turbines have been relocated in step b). The final arrangement thus represents an optimization of the first arrangement to improve the annual energy produced by a given number of turbines in a given space.

[0080] Once the final location is found, the wind turbines can be constructed according to the final location within the given space to obtain a wind farm that allows maximizing the total energy generated and recovered from the wind within the given physical space.

[0081] Advantageously, the predetermined space can be two-dimensional, so that the height of the ground above sea level is not taken into account, i.e., variations in altitude when installing wind turbines are ignored, which simplifies the calculations while still ensuring sufficient accuracy.

[0082] Preferably, the sequential order can be obtained randomly: indeed, using a random function for the sequential order improves the quality of the optimization while avoiding optimization paths based on pre-established orders that may bias the optimization results.

[0083] According to one embodiment of the present invention, the determined number may be a square integer multiple (preferably, four square integer multiples). The discrete position of the wind turbine to be relocated is thus one of the possible discrete positions to be re-evaluated. Indeed, using four square integer multiples makes it possible to discretize the grid in two dimensions by having a first intersection line of the cells (e.g., vertical) corresponding to a first axis of symmetry of the (e.g., first square) grid and a second intersection line of the cells (e.g., horizontal) corresponding to a second axis of symmetry of the grid.

[0084] The pre-established number can therefore take values ​​such as 4, 16, 25, 64, 100, etc.

[0085] According to a preferred configuration of the invention, step b) can be repeated several times, thus performing several local search optimization loops to reposition the wind turbines and further improve the annual energy production.

[0086] Preferably, at each iteration of step b), the sequential order of the turbine position modifications can be modified, thus improving the local relocation search. When the sequential order is randomly obtained each time, the influence of several optimization paths (an optimization path is understood to be a gradual optimization of the relocation influenced by the determined sequential order) can be avoided. Thus, the annual energy produced can be increased.

[0087] According to an advantageous embodiment of the present invention, in each iteration of step b), a first predetermined length shorter than the first predetermined length in the previous iteration can be selected, the predetermined number being the same for each iteration, and the iteration of step b) can be stopped when the first predetermined length is less than a first threshold. Thus, in each iteration of step b), the local neighborhood in which it is desired to improve the position of the wind turbine to be relocated shrinks compared to the previous iteration. Furthermore, since the determined number remains constant for each iteration, the neighborhood becomes increasingly discretized with each iteration. Thus, the accuracy of the turbine position gradually improves. By defining the first threshold, the iterations can be stopped at a level of position accuracy sufficient for turbine installation, where no significant increase in annual energy production (or gains that would be included in the calculation uncertainty range) is expected. For example, the first threshold can be in the range of 1 m to 10 m, preferably 4 m to 6 m, to obtain a good compromise between calculation time, computer memory used, position accuracy, and annual energy production gain.

[0088] Advantageously, the predetermined space can include non-connected zones, and therefore can correspond to a real location where wind turbines are intended to be installed, for example a power plant planned at a location that includes two zones separated by a road of considerable width (a few meters or even tens of meters), such as a highway.

[0089] Preferably, the predetermined space may include non-convex zones, thus corresponding to real locations with complex shapes, such as spaces bounded by hills or steep cliffs, coastlines, streams, rivers or any other body of water, etc. Offshore, the predetermined space may include non-convex zones that may be specifically determined taking into account, for example, water depth, soil properties, borders with other countries, navigational channels, the laying of cables or pipelines.

[0090] According to one embodiment of the present invention, the first position determination algorithm may perform at least the following steps:

[0091] - arbitrarily determining the location of a first wind turbine, whereby the location of the first turbine is selected at a discrete value of a first discrete grid;

[0092] - then, for each wind turbine to be placed, successively (one by one) performing: - determining possible locations within the first discrete grid for locating wind turbines. The possible locations include discrete locations within the first discrete grid located between a minimum and a maximum distance from all located turbines (e.g., the maximum distance is at least four times the diameter of the wind turbine, preferably at least eight times the diameter of the turbine) and / or discrete locations located at a minimum distance from all located turbines. In other words, locations within the first discrete grid where the next turbine will be advantageously located are selected. Therefore, the number of operations and computer memory required to determine the selected locations are limited, thereby speeding up the calculations. Also, selecting discrete locations between a minimum and a maximum distance from already located wind turbines increases the likelihood of locating the maximum number of turbines within a given space. By locating as many wind turbines as possible within a given space, annual energy production can be improved. Selecting to locate turbines along the entire boundary at a minimum distance from the located turbines allows for the use of substantially more space than strictly locating turbines within the space excluding the boundary. Using a larger space allows for more energy to be generated: - calculating the annual energy production of the power plant from the deployed wind turbines and from various defined possible positions of the turbines to be deployed, such that a value of the annual energy production is associated with each defined possible position. Furthermore, the calculation of the annual energy production takes into account a first discrete wind speed distribution, a second discrete wind direction distribution and their occurrence probabilities. As is well known, this calculation includes wind turbine characteristics, i.e., in particular, the rotor swept area, the drag coefficient and / or the power coefficient; - selecting the positions of the wind turbines to be placed that correspond to the value of the maximum annual energy production calculated in the previous step, thereby maximizing the annual energy generated by the wind turbines positioned in the given space. These positions will serve as a basis for determining the positions of the next turbines, in particular for determining the possible positions of the first discrete grid.

[0093] - after all turbines have been positioned, determining a first arrangement corresponding to the location of a predetermined number of wind turbines within a predetermined space.

[0094] The first positioning algorithm is thus a greedy algorithm that includes a step of randomly positioning a first wind turbine, then iteratively positioning additional turbines in selected zones of the predetermined space until all of the predetermined number of wind turbines are positioned within the predetermined space. This greedy algorithm allows for accelerating the calculation time while carefully positioning the turbines by selecting possible positions. Such an algorithm allows for obtaining a first suitable position for step b) of the local positioning optimization for each turbine.

[0095] According to a variant of the invention, any position can correspond to the maximum sum of the coordinates of the positions of the first discrete grid. Thus, the wind turbine position is located at one end of the given space. This option allows more turbines to be placed within the given space, thereby allowing more possible options for the position of a given number of turbines within the given space. Therefore, there is a lot of room to optimize the annual energy generated by the wind farm.

[0096] The present invention also relates to a computer program product for implementing the above-mentioned method from a computing means such as a computer, a mobile phone or a tablet. The computer program product is downloadable from a communications network and / or stored on a computer-readable medium and / or executable by a processor or a server, and contains program code instructions for implementing a method according to one of the above-mentioned features when the program is run on a computer or a mobile phone. Indeed, the above-mentioned method is particularly suitable for use on a computing means, and can thus be implemented in a simple manner and results can be obtained quickly.

[0097] The invention further relates to a wind power plant resulting from the method for arranging a predetermined number of wind turbines (or constructing a wind power plant) in a predetermined space as described above. Indeed, once the locations of the wind turbines are known, the various turbines can be physically constructed (or installed) in the positions determined by this method on the physical site of the predetermined space. In this way, a power plant with an optimized total annual energy production is obtained.

[0098] FIG. 1 shows, by way of non-limiting example, a general overview of a method for arranging a given number of wind turbines in a given space (or for building a wind farm) according to the invention.

[0099] The method includes, in particular, a first location algorithm Alg1, the input data of which are at least a first discrete distribution RD1 of wind speeds, a second discrete distribution RD2 of wind directions, a first discrete grid RD3 of a given space on which a given number of wind turbines can be installed, and the occurrence probability Prob of each wind speed for each wind direction. In particular, the characteristics of the wind turbines, which are required for determining the annual energy production, may also be used.

[0100] In particular, wind speed and direction data and the probability of occurrence of each speed value in each direction can be obtained from the collection means in a preliminary step, which data can in particular be used to obtain a wind rose as known to those skilled in the art.

[0101] The first positioning algorithm Alg1 determines a first arrangement Disp1 of a given number of wind turbines in a given space. This first arrangement Disp1 can be perfect, but of sufficient quality to allow local optimization in the next step. This first positioning algorithm Alg1 can be a greedy algorithm.

[0102] This first configuration Disp1, obtained quickly using a greedy algorithm, is used as input data in a second algorithm Alg2, which is an optimization algorithm that modifies one by one the position of at least one wind turbine of the first configuration Disp1, preferably all wind turbines of the first configuration Disp1 in order to increase the annual energy generated by the power plant by testing various possible discrete positions of each turbine around its initial position (or its last defined position).

[0103] To determine the annual energy production, the second algorithm uses, inter alia, the first discrete wind speed distribution RD1 and the second discrete wind direction distribution RD2.

[0104] Once all the wind turbines have been relocated, a final location DispF for a given number of wind turbines in a given space is obtained. Then, the location (construction / installation / assembly) Pos of the turbines on the physical site corresponding to the given space can be performed.

[0105] FIG. 2 shows, by way of non-limiting example, a detailed view of a first embodiment of a method for positioning a given number of wind turbines in a given space (or for building a wind farm) according to the invention.

[0106] The method includes, inter alia, a first location algorithm Alg1. The input data of the first location algorithm Alg1 are at least a first discrete distribution RD1 of wind speeds, a second discrete distribution RD2 of wind directions, a first discrete grid RD3 of a given space on which a given number of wind turbines can be installed, and the occurrence probability Prob of each wind speed for each wind direction. In particular, the characteristics of the wind turbines, which are required for determining the annual energy production, may also be used. A wake model is used to determine the annual energy production.

[0107] In particular, wind speed and direction data and the probability of occurrence of each speed value in each direction can be obtained from the collection means in a preliminary step, which data can in particular be used to obtain a wind rose diagram as known to those skilled in the art.

[0108] The first positioning algorithm Alg1 determines a first arrangement Disp1 of a given number of wind turbines in a given space. This first arrangement Disp1 can be perfect, but of sufficient quality to allow local optimization in the next step. This first positioning algorithm Alg1 can be a greedy algorithm.

[0109] This first constellation Disp1, obtained quickly using a greedy algorithm, is used as input data in a second algorithm Alg2, which is an optimization algorithm that modifies the positions of the various wind turbines of the first constellation Disp1 one by one in order to increase the annual energy produced by the power plant by testing various possible discrete positions of each turbine around its initial position.

[0110] More specifically, the second algorithm Alg2 includes the following steps:

[0111] - determining a successive order OS for changing the positions of the turbines one by one, this successive order OS being obtainable in particular by means of a random function;

[0112] - then iteratively correcting the position of at least one wind turbine i, preferably each turbine i, by performing the following sub-steps: * Determining possible discrete positions PDP_i of wind turbine i. These possible discrete positions PDP_i are in the vicinity of the position of turbine i to be relocated (other wind turbines remain in their positions, either their initial positions from the first placement or their already relocated positions). For example, a certain perimeter can be defined around the position of the wind turbine to be relocated (a grid with one or two predetermined lengths, a square with sides of a certain length, a circle with a certain diameter, etc.). Discrete positions are determined within this perimeter, for example, by dividing this perimeter into cells. Discrete positions are thus cell centers and / or cell intersections and / or points defining cells, and discrete positions that are too close to other wind turbines are excluded from this perimeter. In other words, discrete positions that are at a distance shorter than a required minimum distance are removed from this perimeter. The required minimum distance can be, for example, at least twice the diameter of the turbine rotor. Positions outside the predetermined space are also excluded from this perimeter: * A subsequent sub-step of evaluating the annual energy production Eval_i for each possible arrangement (for each possible discrete position PDP_i of the wind turbine i to be relocated, the other turbines remaining in their last assigned positions). The annual energy production takes into account the first and second discrete distributions RD1, RD2 of wind speed and direction, as well as the probability Prob of occurrence of each wind speed for each wind direction. Thus, at the end of this evaluation step, the annual energy production corresponds to each possible discrete position PDP_i of the wind turbine i to be relocated: * Substep of selecting as position Pos_i of wind turbine i the possible discrete position PDP_i corresponding to the maximum value of annual energy production at step Eval_i: * Thus, a substep of obtaining a new configuration Disp_N for the wind turbines in the given space, this new configuration comprising the last positions of the turbines already positioned and the new position Pos_i of wind turbine i.

[0113] Then, a loop B1 allows the selection F1 of the next wind turbine in the defined sequential order (i going to i+1) in order to carry out the same procedure for the next wind turbine.

[0114] Once all the wind turbines have been relocated, a final location DispF of the predetermined number of wind turbines in the predetermined space is obtained, and then the location Pos (i.e., physical construction, installation, or erection) of the turbines on the physical site corresponding to the predetermined space can be performed.

[0115] Figure 3 shows, by way of non-limiting example, a schematic variation of Figure 2. The same reference numerals as in Figure 2 correspond to the same elements, and therefore they will not be described in detail again here.

[0116] FIG. 3 differs from FIG. 2 by the addition of a second loop B2 in the second algorithm Alg2.

[0117] In fact, when all the wind turbines have been relocated by loop B1, it is now planned to modify the positions of these turbines several times. The determination of the successive order OS and the steps of loop B1 are therefore repeated several times, one for each wind turbine. Loop B1 comprises determining, for each turbine i to be relocated, the possible discrete positions PDP_i, carrying out an evaluation of the annual energy production Eval_i for each possible discrete position, selecting the position Pos_i of the wind turbine i to be relocated, and defining the new position Disp_N.

[0118] These steps can be repeated several times to refine the positioning of the wind turbines within a given space and improve the annual energy produced by the plant. By changing the sequence each time (e.g., by random drawing each time), the annual energy production can be further improved.

[0119] Loop B2 may terminate, for example, when the annual energy increase relative to the previous iteration falls below a certain value (eg, below 0.5%).

[0120] When the loop B2 is completed, the obtained last configuration Disp_N becomes the final configuration DispF, and then the configuration Pos (construction, erection, installation) of the wind turbines to the planned positions based on the final configuration can be performed to obtain a wind farm.

[0121] FIG. 4 shows, by way of non-limiting example, a schematic representation of a given space suitable for carrying out the location method according to the invention (or the wind farm construction method according to the invention).

[0122] The given space can include, in particular, a first zone Z1 and a second zone Z2, represented by vertical hatching. These zones Z1 and Z2 are disjoint. In fact, there is a non-zero minimum distance D between zone Z1 and zone Z2. Furthermore, zone Z1 is rectangular and therefore convex. Zone Z2 has a complex, non-convex shape. In fact, considering two points A and B, it can be observed that part of the line segment Seg connecting points A and B lies outside zone Z2.

[0123] A Boolean matrix can be used to identify zones Z1, Z2 of a given space within a third, larger zone ZE that encompasses these two zones Z1, Z2. The third zone ZE is rectangular and easier to digitally process than non-connected and / or non-convex zones such as Z1, Z2. The Boolean matrix associates, for each discrete value (discrete position) of the zone ZE, a value equal to 1 if the discrete position is inside zone Z1 or Z2, and a value of 0 if it is outside zones Z1 and Z2. This Boolean matrix allows the definition of the given space used in the method according to the invention. From this Boolean matrix, one or more boundaries of the given space can be determined. Indeed, a point is considered to belong to a boundary if it has a value of 1 in the Boolean matrix and at least one of its direct neighbors has a value of 0 in the Boolean matrix.

[0124] FIG. 5 shows, by way of non-limiting example, a schematic representation of a first position determination algorithm according to the invention.

[0125] The first location algorithm is a greedy algorithm.

[0126] From a first discrete grid RD3 in space, the position P_E1 of the first wind turbine is for example arbitrarily determined.

[0127] Next, for each wind turbine j, the location that maximizes annual energy production is sought.

[0128] Thus, for each wind turbine, possible positions PE_Ej are iteratively defined based on F3, one by one, for the turbine j to be positioned. The possible positions PE_Ej are defined by a first discrete grid RD3 of the predetermined space. For example, the possible positions may correspond to positions of the first discrete grid RD3 located between a minimum and a maximum distance from wind turbines already positioned in the predetermined space, and / or corresponding positions at the boundary of the predetermined space.

[0129] Once these possible positions PE_Ej for the wind turbine j to be placed have been determined, an evaluation (Eval_AEP) is made for each of these possible positions PE_Ej of the annual energy production of the turbines already placed and of the turbine j to be placed in a given space. This evaluation Eval_AEP makes use in particular of the first and second distributions RD1, RD2 of wind speeds and directions as well as the probability of occurrence Prob of each wind speed for each wind direction. The characteristics and wake effects of the wind turbines described above can be used in a known manner.

[0130] The position Pos_j of wind turbine j can then be selected that corresponds to the maximum annual energy production from the previous step.

[0131] Then, a new arrangement Disp_Ej of the wind turbines (including turbine j) located in the given space can be defined, which will be used in the next iteration to determine the possible positions PE_Ej of the new wind turbines to be located and to evaluate the annual energy production Eval_AEP.

[0132] A loop F3 is executed, where j ranges from 1 to N-1, N being a predetermined number of wind turbines in a given space, and considering that the first turbine is sequentially placed at P_E1.

[0133] Once all wind turbines have been placed (i.e., when j=N-1), the last found placement Disp_Ej corresponds to the first placement Disp1. Thus, a wind farm can be constructed by installing (erecting / constructing) wind turbines at the location of the last found placement corresponding to the final placement so as to generate energy from wind at a given spatial physical site.

[0134] Figure 6 shows, by way of non-limiting example, a schematic representation of the steps for determining the positions of the first three wind turbines by a greedy positioning algorithm, with diagrams a), b) and c) corresponding to the positioning steps for the first, second and third turbines, respectively.

[0135] The predetermined space Esp defined here is rectangular. The first wind turbine E1 is arbitrarily placed at the bottom right corner of the rectangle of the predetermined space Esp. Therefore, the first wind turbine E1 is located on the boundary of the predetermined space Esp.

[0136] Once the position of this wind turbine is determined, the possible positions for determining the position of the second wind turbine PE_E2 correspond to discrete values ​​located on the boundary of the predetermined space Esp and at a distance from the first wind turbine E1 that is greater than a minimum distance, and discrete values ​​located within the predetermined space Esp and at a distance ranging between a minimum distance and a maximum distance, where the minimum distance is two times the diameter of the turbine rotor and the maximum distance is four times the diameter of the turbine rotor.

[0137] Once the annual energy production for each position PE_E2 has been calculated, the position of the second wind turbine E2 is determined at a position on diagram b), so that the second wind turbine is on the boundary of the predetermined space Esp.

[0138] These positions of the first two wind turbines can define possible positions PE_E3 for the third turbine to be placed, including discrete positions located within the predetermined space between a minimum and a maximum distance from the wind turbines already placed (the first and second turbines E1, E2), and discrete positions located at the boundary of the predetermined space at a distance greater than the minimum distance from the wind turbines E1, E2.

[0139] Once the annual energy production has been calculated for each position PE_E3, the position of the third wind turbine E3 is determined at a position on diagram c), which is thus also on the boundary of the predetermined space Esp.

[0140] These positions of the first three wind turbines allow for the definition of possible positions PE_E4 for the fourth wind turbine to be placed. These possible positions PE_E4 include discrete positions located within the predetermined space between the minimum and maximum distance from the turbines already placed (turbines E1, E2, E3). Note that there are no discrete positions outside the already defined zone on the boundary of the predetermined space Esp that are at a distance greater than the minimum distance from wind turbines E1, E2, E3, since all other values ​​of the boundary no longer satisfy the condition of a minimum distance from wind turbines E1-E3.

[0141] 7 shows, by way of non-limiting example, a schematic representation of a variant of the step of determining the positions of the first three wind turbines by a first greedy positioning algorithm, where diagrams a), b) and c) correspond to the steps of determining the positions of the first, second and third turbines, respectively.

[0142] Figure 7 differs from Figure 6 in that the possible positions are defined only by discrete values ​​in space located between a minimum and maximum distance from the already placed wind turbines. Obviously, discrete values ​​located on the boundaries of the given space at distances greater than the maximum distance from the already placed turbines are not considered additionally.

[0143] The predetermined space Esp defined here is rectangular. The first wind turbine is arbitrarily placed at the bottom right corner of the rectangle of the predetermined space Esp. Thus, the first wind turbine E1 is located on the boundary of the predetermined space Esp.

[0144] Once this wind turbine is positioned, the possible positions for determining the location of the second wind turbine PE_E2 correspond to discrete values ​​located in a predetermined space Esp (boundaries included) at distances ranging between a minimum distance and a maximum distance, where the minimum distance is two times the diameter of the turbine rotor and the maximum distance is four times the diameter of the turbine rotor.

[0145] Once the annual energy production has been calculated for each location PE_E2, the location of the second wind turbine E2 is determined to be within diagram b), so the predetermined space is the same, the first and second discrete distributions and the first discrete grid are the same, the occurrence probabilities are the same, and the location of the first wind turbine E1 is the same, but the second wind turbine is not on the boundary of the predetermined space Esp, unlike the solution of diagram b) in Figure 6.

[0146] These positions of the first two wind turbines make it possible to define possible positions PE_E3 for the third turbine to be placed, which include discrete positions located within a predetermined space between a minimum and a maximum distance from the turbines already placed (the first and second turbines E1 and E2).

[0147] Once the annual energy production has been calculated for each location PE_E3, the location of the third wind turbine E3 is determined at the location in diagram c).

[0148] These positions of the first three wind turbines E1-E3 make it possible to define possible positions PE_E4 of the fourth wind turbine to be placed, which include discrete positions located in a predetermined space between a minimum and a maximum distance from the already placed wind turbines (turbines E1, E2, E3).

[0149] Note that the positions of the three wind turbines E1, E2, E3 in Figure 7 are very different from those in Figure 6, but all the parameters are the same. The distinction only relates to whether the definition of possible positions includes (Figure 6) or does not include (Figure 7) discrete positions on the boundary of the given space Esp that are located at a distance greater than the maximum distance of all wind turbines already deployed.

[0150] An advantage of the solution of Figure 7 over Figure 6 is that by placing wind turbines one after the other, the turbines can be placed closer to each other, thereby allowing more wind turbines to be placed in a given space. This is particularly advantageous when there is a fairly large number of wind turbines (proportional to the number of wind turbines to be placed) in a limited given space.

[0151] FIG. 8 illustrates, by way of non-limiting example, the application of a local optimization model to determine wind turbine positions from a first constellation.

[0152] The figure shows four diagrams a), b), c), and d) illustrating different turbine relocation procedures.

[0153] The method of wind turbine location (or wind farm construction) applied here corresponds to FIG.

[0154] Diagram a) shows a first arrangement of five wind turbines embodied by small circles within a given space Esp defined by a grey rectangle.

[0155] The position of the turbine Eo1 needs to be improved. Therefore, a first grid Car1 is defined, which grid is centered on the turbine Eo1 and is a square with side L1.

[0156] In diagram b), the grid Car1 is discretized into 16 (but may be a different number, in particular 4, 25, 100, etc.) cells Car2, where the cells Car2 are second squares with smaller sides than the grid Car1 in order to discretize the grid Car1. Possible discrete positions PDP_i are defined as intersections of the cells Car2. Possible discrete positions PDP_i are embodied by crosses at intersections of cells in the given space Esp and at a sufficient distance from other wind turbines. In diagram b), three discrete positions PDP_i are possible. Thus, after local optimization of the annual energy at the three possible positions PDP_i, the position of the wind turbine Eol changes from position P_init to position P_modif1.

[0157] Once all wind turbines have been relocated, step b) of the location method (or wind farm construction method) can be repeated.

[0158] Next, a new grid (new first square) with side L2 smaller than side L1 used in the previous iteration is redefined. This new grid, centered at position P_modif1, is again discretized with a pre-established number of cells (here 16 second squares) to define new possible discrete positions. The new possible discrete positions correspond to crosses placed at intersections of cells located within the given space, but at a sufficient distance from other wind turbines. Note that intersections Pexc are located too close to turbine Eol2 (the distance to turbine Eol2 is less than the minimum distance). Therefore, these intersections are not considered in the possible discrete positions for relocating turbine Eol. After this new iteration, the position of turbine Eol becomes position P_modif2.

[0159] Once all wind turbines have been relocated, step b) of the location method (or wind farm construction method) can be repeated again.

[0160] In diagram d), a new local search for relocating the wind turbine Eol is performed. A new grid (a new first square) is redefined with a side L3 smaller than the side L2 and centered at the position P_modif2. The new grid is discretized into a pre-established number of cells (here 16 second squares) (the same as the grid of the previous iteration). In these 16 cells, 16 intersections are located within the predetermined space Esp at a sufficient distance from other wind turbines (greater than the minimum distance from other turbines). These intersections then serve as possible discrete positions, embodied by crosses, for relocating the wind turbine Eol.

[0161] Calculation of the annual energy based on the position of the wind turbine Eol at these different possible discrete locations allows determining the position P_modif3 that corresponds to the maximum annual energy production.

[0162] It should be noted that although the positions of the wind turbines other than Eol have not been changed with respect to diagrams c) and d), considering that diagrams c) and d) correspond to a new iteration of step b), in which a new turbine arrangement is determined as input, the positions of the wind turbines other than Eol may also be changed.

[0163] FIG. 9 shows, by way of non-limiting example, some steps of a method for the positioning of a pre-established number of wind turbines (or the construction of a wind farm) within a given space Esp.

[0164] Here, the predetermined space Esp is rectangular.

[0165] Diagram a) shows a first arrangement of five wind turbines, represented by dark grey dots, within a given space Esp. This first arrangement is obtained from a first greedy algorithm, such as the algorithm of FIG. 7.

[0166] In diagram b), it is desired to improve the location of wind turbine El1 in order to increase its annual energy production. This is achieved by determining a grid (first rectangle) centered on the location of turbine El1. This grid is then discretized into a predefined number of cells (e.g., second rectangles). The intersections of these cells correspond to possible discrete positions, embodied by small grey dots around wind turbine El1, located within a given space Esp and at a sufficient distance from other turbines (greater than the minimum distance from other turbines, e.g., twice the turbine rotor).

[0167] Diagram c) corresponds to a new iteration of step b), in which the wind turbines have been repositioned. Indeed, at the end of the iteration of step b), the wind turbines have been repositioned one after the other.

[0168] In diagram c), the relocation of wind turbine E12 is sought. This is achieved by determining a grid (first square) centered on the location of turbine E12, which is then discretized into a predetermined number of cells (here, second squares). The intersections of these cells, located within the given space Esp at a distance greater than the minimum distance to other turbines (e.g., twice the diameter of the turbine rotor), correspond to the possible discrete positions of wind turbine E12. These possible discrete positions correspond to the small light grey dots around turbine E12.

[0169] FIG. 10 shows, by way of non-limiting example, a schematic application of the method for determining the position of wind turbines (or for constructing wind farms) in a complex given space.

[0170] Here, the given space consists of two non-connected zones Es1 and Es2. Furthermore, considering that the triangular portion of the boundary Ar faces inward into zone Es2, zone Es2 is non-convex.

[0171] In diagram a), the grey dots represent the boundaries of a given space and correspond to discrete locations where the first wind turbine can be installed.

[0172] In diagram b), a single wind turbine EE1 is placed at the top of zone Es1 and is embodied as an isolated grey point. A set of dark grey points T2 represent possible locations for the next turbine to be placed in space. They correspond to discrete points on the boundaries of zones Es1 and Es2 that are located at a sufficient distance from turbine EE1 (i.e., greater than the minimum distance, which may be twice the turbine diameter), and to discrete positions within zone Es1 that are located at a distance ranging between the minimum and maximum distance from turbine EE1.

[0173] Diagram c) corresponds to the stage where 20 wind turbines, embodied by isolated grey points, have been located by the first greedy location algorithm. The set of dark grey points T21 corresponds to the possible positions of the 21st wind turbine to be located by the first greedy location algorithm.

[0174] Diagram d) corresponds to the stage where 50 wind turbines, embodied by grey isolated points, have been located by the first greedy location algorithm. The set of dark grey points T51 corresponds to the possible positions of the 51st wind turbine to be located by the first greedy location algorithm.

[0175] Diagram e) corresponds to the stage where the position of wind turbine ELi is improved by local optimization of its position. The placement of the 52 wind turbines was predetermined by a first greedy positioning algorithm, and the current position of each of the 52 turbines corresponds to the large grey isolated dots in diagram e).

[0176] The small grey dots arranged around wind turbine ELi are possible discrete positions PDP_ELi for relocating turbine ELi. The annual energy production is predicted based on the positions of the wind turbines other than turbine ELi in the arrangement of Figure e) and the various possible discrete positions PDP_ELi of turbine ELi.

[0177] FIG. 11 shows, by way of non-limiting example, a schematic variation of the method of determining the location (or constructing a wind farm) for the same complex given space as FIG.

[0178] The given space includes two non-connected zones Es1 and Es2. Furthermore, considering that the triangular portion of the boundary Ar faces inward into zone Es2, zone Es2 is non-convex.

[0179] In positioning wind turbines according to the first greedy positioning algorithm, the possible positions correspond to discrete positions located within the given space at distances ranging between a minimum and a maximum distance from already positioned wind turbines. Boundary points located at distances greater than the maximum distance from already positioned turbines are not considered for maximizing the number of turbines to be installed within the given space, but are considered in the first positioning algorithm of Figure 10 by the possible position T2 of the second turbine, as can be seen in diagram b) of Figure 10.

[0180] Two wind turbines EE1, EE2 are arranged in diagram a) of Figure 11. Possible positions for the third turbine T3 to be arranged correspond to the set of grey dots.

[0181] In diagram b) of Fig. 11, ten wind turbines are deployed, embodied by isolated grey dots. The possible locations of the eleventh turbine T11 to be deployed correspond to the set of grey dots.

[0182] example A first example consists of placing wind turbines in a given space in Figure 7 according to the prior art continuous method (using continuous real data instead of discrete values) and the method of Figure 3 of the present invention. When following the method of Figure 3, possible discrete locations are defined from a grid (first square) centered on the location of the wind turbine to be relocated. The grid is discretized into a predefined number of cells (here second square). The intersections of cells located within the given space and at a sufficient distance (greater than a minimum distance) from other wind turbines form possible discrete locations.

[0183] The wind turbines to be installed are 2.5 MW turbines, with a minimum distance between turbines of four times the diameter of the turbine rotor.

[0184] In the method according to the invention, the maximum distance between wind turbines used in the first greedy positioning algorithm is 8 times the diameter of the turbine rotors, the length of the first square in the first iteration is 500 m, the threshold enabling stopping the iteration of step b) is 5 m, the predetermined number is 4, and the ratio of the length of the first square to the length of the first square of the previous iteration of step b) is 0.8.

[0185] The prior art method is implemented in FarmShadow™ software (IFP Energies nouvelles, France).

[0186] The first wind speed distribution has discrete values ​​at 0.5 m / sec intervals, and the second wind direction distribution has discrete values ​​at 1° intervals.

[0187] The method according to the invention makes it possible to estimate an increase of about 2.2% in annual energy production compared to the prior art method, which shows the advantage of the method for improving the energy recovered by the power plant while reducing the calculation time and the computer memory used.

[0188] The second application example is related to Figure 10. Firstly, the continuation method, which must be applied to convex and connected zones, which are generally rectangular, cannot be applied to this application, which involves a given volume with non-connected and non-convex zones. Fifty-two 10 MW wind turbines are placed in this given volume.

[0189] The increase in annual energy production obtained with the final wind turbine placement is about 2.84% compared to the first placement obtained from the first greedy location algorithm, which on the one hand proves the superiority of the first placement obtained from the first algorithm, and on the other hand proves the efficiency of the local search method in optimizing the position of each wind turbine.

[0190] The method for placing wind turbines (or constructing a wind farm) within a given space allows for fast and simple calculations while providing accurate and optimized positions of turbines within the given space, which may be non-connected and / or non-convex, to maximize the total annual energy generated by the wind farm.

Claims

1. A construction method for constructing a wind power plant in a predetermined space (Esp) from a predetermined number of wind turbines, the construction method comprising a first discrete wind speed distribution (RD1), a second discrete wind direction distribution (RD2), and occurrence probabilities (Prob) of each discrete wind speed value at each discrete wind direction value of the first and second discrete distributions (RD1, RD2), the predetermined space (Esp) being divided into a first discrete grid (RD3), and at least the following series of steps a) to c) being carried out: a) determining a first location (Disp1) of said wind turbines in a first discrete grid (RD3) of said predetermined space (Esp) by a first location determination algorithm (Alg1); b) determining a sequential order (OS) for correcting the wind turbine position determined by the first positioning algorithm (Alg1), and then repeatedly performing at least the following steps b1) to b4) for each of the at least one wind turbine to be relocated according to said determined sequential order (OS); b1) for each wind turbine to be relocated, determining new possible discrete positions (PDP_i) of the turbine to be relocated by developing a grid of at least a first predetermined length, the grid being centered on the position of the turbine to be relocated, the grid being divided into a predetermined number of cells, the new possible discrete positions including intersections of the cells, the intersections of the cells being located within the predetermined space (Esp) at a minimum distance from positions of other wind turbines; b2) calculating the average annual energy production (Eval_i) of the predetermined number of wind turbines for each possible discrete position of the wind turbine to be relocated from the first discrete wind speed distribution (RD1), the second discrete wind direction distribution (RD2) and the occurrence probability (Prob); b3) selecting the position (Pos_i) of the wind turbine to be relocated that corresponds to the maximum value of the annual energy production calculated in step b2); b4) determining a new location of the wind turbines (Disp_N) within the predetermined space (Esp); c) determining a final determined configuration (DispF) corresponding to the obtained final configuration for generating energy from wind power, and constructing the wind farm by erecting the wind turbines at the positions of the final determined configuration within the predetermined spatial site.

2. 2. The wind power plant construction method according to claim 1, wherein prior to step a), statistical wind data is collected in a predetermined space (Esp) by a collection means, preferably a LiDAR sensor, to determine the first and second discrete distributions (RD1, RD2) and the occurrence probability (Prob) of each wind speed in each wind direction of the first and second discrete distributions (RD1, RD2).

3. 3. The method of claim 1, wherein the predetermined space (Esp) is two-dimensional.

4. A method according to any one of claims 1 to 3, wherein said successive orders (OS) are obtained randomly.

5. Method for constructing a wind power plant according to any one of claims 1 to 4, wherein step b) is repeated (B2) several times, preferably with a different succession order (OS) for each iteration.

6. 6. The wind farm construction method according to claim 5, wherein for each iteration (B2) of step b) a first predetermined length is selected that is shorter than the first predetermined length of the previous iteration, the predetermined number being the same between iterations, and the iterations of step b) are stopped when the first predetermined length is less than a first threshold.

7. 7. A wind power plant construction method according to any one of claims 1 to 6, wherein the predetermined space (Esp) comprises non-connected zones (Es1, Es2).

8. Method for constructing a wind power plant according to any one of claims 1 to 7, wherein said predetermined space (Esp) comprises a non-convex zone (Z2, Es2).

9. The first location algorithm (Alg1) comprises at least - arbitrarily determining the position (P_E1) of a first wind turbine (EE1); Then, for each wind turbine to be placed, the steps of successively determining possible positions (PE_Ej) in the first discrete grid (RD3) for the wind turbine to be placed, calculating annual energy productions (Eval_i) of the placed wind turbine and of the turbine to be placed for the determined possible positions (PE_Ej) from the first discrete wind speed distribution (RD1), the second discrete wind direction distribution (RD2) and the occurrence probability (Prob), and selecting the position (Pos_j) of the wind turbine to be placed that corresponds to the calculated maximum value of annual energy production, determining the first disposition (Disp_Ej) corresponding to positions of the predetermined number of wind turbines within the predetermined space (Esp); Run 9. The wind farm construction method according to any one of claims 1 to 8, wherein the possible positions (PE_Ej) comprise discrete positions of the first discrete grid (RD3) located between a minimum and a maximum distance from all of the placed turbines and / or discrete positions of the boundary of the predetermined space (Esp) located at the minimum distance from all of the placed turbines.

10. 10. A method for constructing a wind farm according to claim 9, wherein said arbitrary position corresponds to the maximum sum of the coordinates of the positions of said first discrete grid (RD3).

11. A computer program downloadable from a communications network and / or recorded on a computer readable medium and / or executable by a processor or a server, the computer program comprising program code instructions which, when executed on a computer or mobile phone, perform the method according to any one of claims 1 to 10.

12. A wind power plant obtained by the wind power plant construction method according to any one of claims 1 to 10.

Images