Methods for determining wind speed components using laser remote sensors

Abstract

This invention relates to a method for determining wind speed components using a placed LiDAR sensor (1). In this method, the wind speed components are first approximated (APP) using signals from the LiDAR sensor, and these approximations are used in a wind signal model (MOD) and then in a non-static Kalman filter (KAL) to construct a filtered measurement signal. The filtered measurement signal is then used to reconstruct (REC) the wind speed components.

Description

TECHNICAL FIELD

[0001] The present invention relates to the field of determining wind speed components, in particular for the purpose of assessing the convenience of installing wind turbines at a site.

[0002] Before installing a wind turbine or a wind farm, it is necessary to assess the wind energy potential of the site. Indeed, the size, the class and the structure of the wind turbine depend on the characteristics of the wind, such as the average wind speed, the maximum wind speed, the wind turbulence intensity, which corresponds to the ratio of the standard deviation of the wind speed to the average wind speed, and the like. For example, the size of the wind turbine can be chosen according to the distribution of the average of the wind, while the class of the wind turbine can be chosen according to the turbulence intensity. Considering that replacing a wind turbine of one class to another one has significant costs, it is important to properly identify the characteristics of the wind before installing a wind turbine.

[0003] Moreover, this determination of the wind speed components is particularly critical as it also makes it possible to identify the resource that generates energy. This is important for a wind turbine project as it also defines the financial reliability of the wind turbine installation project. BACKGROUND

[0004] To make these measurements, the conventional technique is to install a measurement mast at the site of measurement. Such a measurement mast is equipped with a large number of sensors and requires a special installation of significant cost, and it is not easy to move it from one site to another because of its size.

[0005] In a second technique, it is possible to use a LiDAR sensor, which can also be understood as a remote laser sensing. LiDAR is a remote sensing or optical measurement technique based on the analysis of the properties of the light beam returning to its emitter. This method is particularly suitable for determining the distance to an object with the help of a pulsed laser. Unlike radar, which is based on a similar principle, LiDAR sensors use visible or infrared light, instead of radio waves.

[0006] In the field of wind turbines, LiDAR sensors are advertised as essential sensors for the proper operation of large wind turbines, especially as their size and power increase (5 MW today and soon offshore 12 MW). Such a sensor makes it possible to measure the wind at a distance, making it possible to initially calibrate the wind turbines so that they can deliver the maximum amount of power (optimization of the power curve). For this calibration step, the LiDAR sensor can be positioned on the ground and oriented vertically (profilograph), making it possible to measure the wind speed and its direction, as well as the wind gradient as a function of the height. This technique can be called placed LiDAR.

[0007] Such a technique is particularly described in patent applications EP3287810 and US2019293836.

[0008] However, it is important to provide the processing of the measurement signals in order to obtain the characteristics of the wind speed in an accurate, robust and reliable manner.

[0009] The radial measurement does not obtain a complete wind force measurement. It is the projection of the wind on the line of sight of the beam, filtered by the spatial transfer function inherent to the LiDAR measurement technique, and with the noise added by the measurement chain. In order to obtain information representative of the wind field passing above the placed LiDAR, it is therefore necessary to combine a plurality of radial measurements with each other and with a processing capable of reproducing the contribution of the wind in each measurement and from which the wind field or vector can be derived.

[0010] The standard reconstruction is based on the assumption that the radial measurements contain only what is consistent with the wind measured and that the wind field is homogeneous and homogeneous at a given height. SUMMARY

[0011] The present invention aims to determine wind speed components in an accurate, robust, reliable and inexpensive manner. To this end, the present invention relates to a method for determining wind speed components by means of a placed LiDAR sensor. For this method, the wind speed components are first approximated by means of the signals from the LiDAR sensor and these approximations are used in a wind signal model and then in a non-static Kalman filter to construct a filtered measurement signal. The filtered measurement signal is then used to reconstruct the wind speed components. Approximating the wind speed components makes it possible to build a reliable and robust wind signal model, allowing the wind speed components to be determined reliably and robustly. Indeed, by filtering the radial measurements so as to retain only the part of them that effectively corresponds to the contribution of the wind to the measurement, this method makes it possible to obtain an estimate of the wind (amplitude and direction) whose mean and standard deviation will be more representative and more realistic. This thus makes it possible to use a placed LiDAR as a replacement or complement to a measurement mast in order to perform an efficient, accurate and potentially cheaper site analysis.

[0012] The present invention relates to a method for determining wind speed components by means of a LiDAR sensor, said LiDAR sensor being oriented substantially vertically in order to make measurements in at least one substantially horizontal measurement plane. For this method, the following steps are implemented:

[0013] a) acquisition of measurement signals in said at least one measurement plane from said LiDAR sensor;

[0014] b) determination of approximations of the wind speed components in said at least one measurement plane by means of a geometric reconstruction of said wind speed components, based on the acquired measurement signals;

[0015] c) determining the mean wind speed and the standard deviation of the wind in the at least one measurement plane by means of the approximated wind speed component;

[0016] d) constructing a wind signal model using the sum of two first order filters, the wind signal model depending on two parameters;

[0017] e) determining the two parameters of the wind signal model by means of the determined standard deviation and the determined mean wind speed;

[0018] f) filtering the acquired measurement signal by means of a non-stationary Kalman filter, the wind signal model and the two determined parameters; and

[0019] g) determining the wind speed component in the at least one measurement plane by means of a geometric reconstruction of the wind speed component based on the filtered measurement signal.

[0020] According to one embodiment, the method comprises, for the step of approximating the wind speed component, the step of filtering the measurement signal using a first order low pass filter.

[0021] Advantageously, the wind signal model corresponds to the Kaimal model of the wind spectrum.

[0022] According to one aspect, the transfer function H of the wind signal model is written:

[0023] where a, b are the two parameters, s is the Laplace variable and τ is a time constant.

[0024] Preferably, the parameters a and b are obtained using the following equation: where f is the frequency vector, ω = 2πf, k is an index corresponding to the component in question, σ k is the standard deviation of the approximated wind speed, V hub is the approximated wind speed, L is the vector of the directional components of the measurement beam of the LiDAR sensor and S is the vector of the spectral components of the Kaimal model of the wind spectrum.

[0025] According to one implementation, the covariance matrix of the non-stationary Kalman filter is determined by minimizing a cost function taking into account the dispersion and the mean deviation of the measurement signal.

[0026] According to one embodiment, the reconstruction of the wind speed component by means of the measurement signal filtered by the non-stationary Kalman filter is achieved by means of a geometric reconstruction of the wind speed component based on the measurement signal filtered by the non-stationary Kalman filter.

[0027] According to one aspect, the method further comprises the steps of:

[0028] a) segmenting the determined wind speed component at predetermined time intervals; and

[0029] b) determining at least one property of the wind speed for said predetermined time intervals.

[0030] Advantageously, said predetermined time intervals are between 1 minute and 1 hour, and preferably between 5 minutes and 30 minutes.

[0031] Advantageously, said at least one wind property is selected from the group consisting of: mean wind speed, standard deviation of wind speed, maximum wind speed, mean wind direction, wind turbulence intensity, mean value of the vertical component of the wind speed, and standard deviation of the vertical component of the wind speed.

[0032] Further features and advantages of the method according to the present application will become apparent by reading the following description of non-limiting exemplary embodiments, with reference to the accompanying drawings described below. BRIEF DESCRIPTION OF DRAWINGS

[0033] Figure 1 A placed LiDAR sensor according to one embodiment of the present application is shown.

[0034] Figure 2 Steps of a method according to a first embodiment of the present application are shown.

[0035] Figure 3 Steps of a method according to a second embodiment of the present application are shown.

[0036] Figure 4 Steps of a method according to a third embodiment of the present application are shown.

[0037] Figure 5 Geometric parameterization of a measurement point is shown. DETAILED DESCRIPTION

[0038] The present application relates to a method for determining a wind speed component by means of a LiDAR sensor. A wind speed component is the name given to the projection of the wind speed in a reference system, in particular in an orthogonal reference system.

[0039] For the present invention, the LiDAR sensor is essentially oriented vertically; in other words, the measurements are made along an essentially vertical axis. For example, the LiDAR sensor can be placed on the ground and oriented vertically. According to the present invention, the LiDAR sensor makes it possible to measure the wind speed in at least one measurement plane. The measurement plane is essentially horizontal in view of the orientation of the LiDAR sensor. There are various types of LiDAR sensors, such as scanning LiDAR sensors, continuous LiDAR sensors or pulsed LiDAR sensors. Within the scope of the present invention, it is preferable to use a pulsed LiDAR. However, other LiDAR technologies can be used while remaining within the scope of the present invention.

[0040] The LiDAR sensor allows continuous measurement. The use of such a sensor thus allows continuous determination of the measurement signal. Furthermore, the LiDAR sensor can easily be moved from one location to another. For example, the sampling rate of the LiDAR sensor can be between 0.1 and 5 Hz (or even higher in the future) and can be 1 Hz. Furthermore, the LiDAR sensor makes it possible to obtain relative information in a plurality of measurement planes at a plurality of heights. The LiDAR sensor can thus be used to determine the wind speed components at a plurality of heights, which can be used in particular to determine the variation of the wind speed as a function of height.

[0041] Figure 1 A LiDAR sensor 1 placed in vertical orientation is shown schematically and not limitatively according to one embodiment of the method according to the present invention. The LiDAR sensor 1 is used to obtain at least one measurement signal in at least one measurement plane PM (only two measurement planes are shown). This figure also shows the x, y and z axes. The point of reference of this reference frame is the center of the LiDAR sensor. The x direction is a horizontal direction. The y direction, perpendicular to the x direction, is a second horizontal direction (the x, y directions form a horizontal plane). The z direction is a vertical direction (corresponding to the measurement direction of the LiDAR sensor 1), pointing upwards; the z axis is perpendicular to the x and y axes. The measurement plane PM is a plane formed by the x, y directions at a distance from the LiDAR sensor 1 (for a non-zero value of z). The measurement planes PM are parallel to each other.

[0042] As Figure 1 is shown, Figure 1 is one exemplary embodiment of a pulsed LiDAR sensor, the LiDAR sensor 1 used comprising four measurement beams or axes 2. The measurement beams 2 are tilted with respect to the vertical z axis. Without limitation, the method according to the present invention is also applicable to LiDAR sensors comprising any number of beams. The LiDAR sensor performs a one-off measurement at each measurement point b1, b2, b3, b4, which is the intersection of the measurement plane PM and the beam 2. These measurement points b1, b2, b3, b4 are in the measurement plane PM.Figure 1 The wind speed vector W is represented by a black circle.

[0043] This figure also shows the wind speed vector W at point bl only, and the three components Wx, Wy, Wz on the x, y and z axes respectively.

[0044] The method according to the application comprises the following steps:

[0045] 1 - Acquisition of the measurement signals

[0046] 2 - Approximation of the wind speed components

[0047] 3 - Determination of the mean speed and of the standard deviation

[0048] 4 - Construction of the wind signal frequency model

[0049] 5 - Determination of the parameters of the model

[0050] 6 - Non-stationary Kalman filter

[0051] 7 - Determination of the wind speed components

[0052] These steps will be described in detail in the remainder of the description. Steps 2 to 7 can be implemented by a computing device, in particular a computer. Steps 2 to 7 can be implemented offline after step 1.

[0053] Figure 2 The steps of the method according to the first embodiment of the application are illustrated schematically and not limitatively. First, the ACQ measurement signals M acq are acquired from the LiDAR sensor. Next, the APP wind speed components W app are approximated by geometric reconstruction of the measurement signals M acq . Then, the wind speed components W app are used to determine the mean value V app and the standard deviation σ app of the MOY wind speed. These data are used with the wind signal model MOD constructed to determine the parameters a and b of this wind signal model PAR. Next, the non-stationary Kalman filter KAL is applied in order to obtain filtered measurement signals M Kal . These measurement signals M Kal make it possible to reconstruct the REC wind speed components W est .

[0054] According to the second embodiment of the application, the method can comprise an additional step of filtering the acquired measurement signals before approximating the wind speed components, in order to remove outliers, thus making it possible to increase the robustness and reliability of the method.

[0055] Thus, the method according to the second embodiment of the application can comprise the following steps:

[0056] 1 - Acquisition of measurement signals

[0057] 1.2 - Filtering of measurement signals

[0058] 2 - Approximation of wind speed components

[0059] 3 - Determination of mean speed and standard deviation

[0060] 4 - Construction of a wind signal frequency model

[0061] 5 - Determination of parameters of the model

[0062] 6 - Non-stationary Kalman filter

[0063] 7 - Determination of wind speed components

[0064] These steps will be described in detail in the remainder of the description. Steps 1.2 to 7 can be implemented by computerized means, in particular a computer.

[0065] Figure 3 The steps of the method according to the second embodiment of the application are schematically and non-limitingly illustrated. First, ACQ measurement signals M acq are acquired from a LiDAR sensor. acq Second, the measurement signals M fil are filtered FIL so as to obtain filtered measurement signals M fil . Next, wind speed components W app are approximated APP by geometric reconstruction of the filtered measurement signals M app . Then, the wind speed components W app are used to determine MOY the mean value V app and the standard deviation σ of the wind speed. These data are used with a constructed wind signal model MOD to determine PAR the parameters a and b of this wind signal model. Next, a non-stationary Kalman filter KAL is applied so as to obtain filtered measurement signals M Kal . These measurement signals M Kal make it possible to reconstruct REC wind speed components W est .

[0066] According to a third embodiment of the application, the method can comprise additional steps in order to determine characteristics of the wind speed. For this third embodiment, the method can comprise the following steps:

[0067] 1 - Acquisition of measurement signals

[0068] 2 - Approximation of wind speed components

[0069] 3 - Determination of mean speed and standard deviation

[0070] 4 - Construction of a wind signal model

[0071] 5 - Determining parameters of the model

[0072] 6 - Non-stationary Kalman filter

[0073] 7 - Determining wind speed components

[0074] 8 - Segmentation of wind speed components

[0075] 9 - Determining characteristics of the wind speed

[0076] These steps will be described in detail in the remainder of the description. Steps 2 to 9 can be implemented by computerized means, in particular a computer.

[0077] Figure 4 The steps of the method according to the third embodiment of the application are schematically but non limitatively illustrated. First, ACQ measurement signals M acq are acquired from a LiDAR sensor. Next, APP wind speed components W acq are approximated by a geometric reconstruction of the measurement signals M app . Then, the wind speed components W app are used to determine the mean value V app and the standard deviation σ app of the MOY wind speed. These data are used with the constructed wind signal model MOD to determine the parameters a and b of this wind signal model PAR. Next, a non-stationary Kalman filter KAL is applied in order to obtain filtered measurement signals M Kal . These measurement signals M Kal make it possible to reconstruct REC wind speed components W est . These wind speed components W est are segmented SEG in time. Next, characteristics CAR of the wind speed are determined (for example the mean wind speed or the turbulence intensity).

[0078] The second and third embodiments can be combined.

[0079] 1 - Collecting measurement signals

[0080] This step comprises acquiring measurement signals of at least one measurement plane from a LiDAR sensor. In other words, for each measurement point of the at least one measurement plane, a measurement signal from the sensor is acquired. Advantageously, these measurement signals can be recorded, in particular in a computer memory, in order to be able to be processed by the computing means in the following steps.

[0081] In order to determine wind speed components in a plurality of measurement planes, this step can be performed on a plurality of measurement planes.

[0082] Advantageously, the measurement signal can be acquired over a long period of time, for example over a duration ranging from a few days to a year or even more.

[0083] 1.2 - Filtering measurement signals

[0084] It should be remembered that this is an optional step. This step consists in filtering the measurement signal, in particular in order to limit outliers for the purpose of making the method more reliable and robust.

[0085] According to one embodiment, this filtering can be implemented by means of a first-order low-pass filter in order to reproduce a continuous and true representation of the wind state measured. This can be a filter with a variable time constant. The older the last valid value passed to the first-order filter, the less the time constant of the filter decreases (in other words, the weight of the state stored in the filter is increasingly low compared to the weight of the next valid value t). This embodiment makes it possible to derive an instantaneous, low-frequency, denoised and true value of the wind state contained in the radial measurement.

[0086] 2 - Approximating wind speed components

[0087] This step consists in approximating the wind speed component on the basis of the measurement signal from the LiDAR sensor, or, where applicable, on the basis of the filtered measurement signal from the LiDAR sensor. This approximation serves to adapt the measurement signal model. This is only an intermediate approximation, and the value of the wind speed component determined in step 7 of the method is more accurate, reliable and robust.

[0088] For this step, the wind speed component is reconstructed by means of geometric reconstruction.

[0089] According to one embodiment of the application, the geometric reconstruction of the wind speed component can implement a pseudo-inverse "Moore-Penrose" operation applied to the measurement signal (or, where applicable, to the filtered measurement signal).

[0090] Figure 5 A geometric parameterization of the measurement signal from the LiDAR sensor is illustrated schematically but not limitingly. This figure shows a single beam 2 of the LiDAR sensor 1. This beam 2 is oriented along a measurement axis represented by the vector 1. This vector 1 is oriented by means of the angles θ and Φ with respect to a reference frame x, y, z (defined in the same way as the reference frame x, y, z of the wind speed vector W). Figure 1 The angle θ is defined in the plane (x, y) with respect to the x axis. The angle Φ is defined with respect to the z axis. This figure also shows the wind speed vector W and the components Wx, Wy, Wz at the measurement point bl.

[0091] Using the geometric projection, it is possible to write the following equations:

[0092]

[0093] where 1,..., i,..., n are the measurement points of the measurement plane, m1,..., m i n are the measurement signals of the measurement plane.

[0094] Using the pseudo-inverse "Moore-Penrose" operation, it is thus possible to obtain, with the help of the measurement signals, the estimated wind speed components in the measurement plane:

[0095]

[0096] According to one embodiment of the application, this step can comprise a step of filtering the estimated speed, in particular in order to limit outliers, to make the method more reliable and robust.

[0097] According to one embodiment, this filtering can be implemented with a first order low-pass filter, in order to reproduce a continuous and realistic representation of the measured wind state. This can be a filter with a variable time constant. The older the last valid value passed to the first order filter, the less the time constant of the filter decreases (in other words, the weight of the state stored in the filter is increasingly low compared to the weight of the next valid value t). This embodiment makes it possible to derive an instantaneous, low-frequency, denoised and realistic value of the wind state contained in the radial measurements.

[0098] 3 - Determining mean speed and standard deviation

[0099] This step comprises determining the mean speed and its standard deviation based on the approximation (possibly filtered) of the wind speed components determined in step 2. For this step, it is possible to implement a conventional mean and standard deviation calculation.

[0100] According to one embodiment of the application, it is possible to determine the mean speed and its standard deviation in a rolling time domain. For example, the rolling time domain of this embodiment can be between 10 minutes and several days.

[0101] 4 - Building a wind signal model

[0102] This step comprises building a wind signal model using the sum of two first order filters, said wind signal model being dependent on two parameters. These two parameters make it possible to adjust the contribution of each filter to the measurements provided by the LiDAR sensor.

[0103] ​According to one embodiment of the application, the wind signal model can comprise an integrator and a low pass filter. The integrator makes it possible to show the average speed and slow variations of the wind speed. The low pass filter makes it possible to simulate instantaneous turbulence, the time constant of which is approximated to the frequency response of the wind turbulence spectrum as defined in the IEC 61400-1 standard. The sum of the two filters makes it possible to simulate a power spectral density, PSD, signal that is consistent with the Kaimal model defined in the standard. The transfer function H corresponding to the PSD model can be given by the following equation (the function H is also called in the rest of the wind signal model description):

[0104]

[0105] where a and b are parameters of the wind signal model, s is the Laplace variable and τ is the time constant.

[0106] Thus, according to one embodiment of the application, the wind signal model can correspond to the Kaimal model, which can be written:

[0107]

[0108] where k is an index corresponding to the component considered (k varies between 1 and 3 and represents the z, y and x axes), f is the frequency, V hub is the average wind speed obtained in step 3, σ is the standard deviation component, S k is the spectrum of the speed component in the k direction, and L k is the integer scaling parameter of the speed component.

[0109] There also exists the following relationship:

[0110]

[0111] And, according to the component considered, the parameter σ k is associated with the standard deviation determined in the previous step. For example, σ1 can have the value of the standard deviation for the z axis, σ2 can be 0.8 times the standard deviation for the y axis, and σ3 can be 0.5 times the standard deviation for the x axis.

[0112] The parameter L can be associated with the scaling parameter.

[0113] As a variant, the method according to the application is applicable to other wind models, such as the "Von Karman" model.

[0114] 5 - Determining parameters of the model

[0115] This step comprises determining the parameters of the wind signal model by means of the standard deviation and the mean wind speed determined in step 3. Thus, at the end of this step, the wind signal model is adapted to the measurements performed by the LiDAR sensor, making it possible to determine the wind speed components accurately.

[0116] According to the above embodiment of the transfer function H, the parameters a and b of the transfer function are determined. The parameters a and b can be determined by means of the following equation:

[0117]

[0118] where the vector L is a vector of components of the direction of the measurement beam in question, and S is a vector of spectral components of the Kaimal model, which depends on known parameters, in particular both the mean wind speed and the standard deviation determined in step 3.

[0119] Thus, the parameters a and b can be estimated using the following equation:

[0120]

[0121] 6 - Non-stationary Kalman filter

[0122] This step comprises filtering the measurement signal obtained in step 1 by means of the non-stationary Kalman filter, the wind signal model constructed in step 4, and the parameters of the wind signal model determined in step 5. Thus, this gives a filtered measurement signal which is suitable for determining the wind speed components accurately in a robust manner. Recall that a Kalman filter is an infinite impulse response filter that estimates the state of a dynamic system based on a series of incomplete or noisy measurements. The filter is called non-stationary because the model embedded in the Kalman filter (explained in particular in steps 4 and 5) describes a signal with a priori variable variance and mean value, which depends on the measurement conditions and the characteristics of the wind being measured.

[0123] To this end, it is possible to formulate the transfer function H in a continuous state space as follows:

[0124]

[0125] where and

[0126] x comprises the states of the transfer function H; and

[0127] Y comprises an estimate of the contribution of the wind to the raw and noisy radial measurements. Thus, Y is a filtered radial measurement, in other words a measurement signal filtered by the non-stationary Kalman filter.

[0128] By discretizing this transfer function, it is possible to write:

[0129] X k+1 = A. X k + B. U k

[0130] Y k = C. X k + D. U k

[0131] where and

[0132] where Ts is the discretized sampling time.

[0133] For this step, the following recursive equations can also be implemented:

[0134] At time k, the prediction of the state at time k+1 is denoted as x k+1|k and the predicted output is denoted as y k+1|k where x k+1|k = Ax k|k y k+1|k = CAx k|k .

[0135] Corresponding to the error covariance in the state, the confidence in the prediction is as follows: P k+1|k = AP k|k A T + BQ k B T where Q is the covariance matrix.

[0136] Therefore, this prediction gives the most likely value of the state and output.

[0137] At time k+1, the prediction is updated with the measurement:

[0138] x k+1|k+1 = x k+1|k + K k+1 (y k+1 - y k+1|k )

[0139] The gain K can be given by the following formula:

[0140] K = P k+1|k * C T * (Ck+1|kC T + DQ k+1 D T + V k+1 )

[0141] Then there is the possibility to update the prediction matrix:

[0142] P k+1|k+1= (I - K k+1 P k+1|k C) P k+1|k

[0143] According to one embodiment of the present application, the covariance matrix of the non-stationary Kalman filter can be determined by minimizing a cost function that takes into account the dispersion and the mean bias of the measurement signal. The principle of adjusting the covariance matrix can be summarized as follows: the more complex the site, the more uncertain the model, and the higher the "method noise" (corresponding to the covariance matrix).

[0144] For example, the following operations can be implemented:

[0145] • Prerequisite: having measurement data from a LiDAR sensor and so-called reference data, for example data from a wind measurement mast. The placed LiDAR sensor can be placed close enough to the measurement mast so that the respective measurements are related.

[0146] • Principle:

[0147] Develop a cost function that takes into account the dispersion and the bias of the reconstructed turbulence intensity TI compared to the TI measured from the "measurement mast"

[0148] Minimize the cost function using a suitable optimization algorithm by adjusting the "multiplicative gain of the covariance matrix Q" parameter.

[0149] Preferably, optimization constraints as a function of acceptable error constraints can be added; for example, this acceptable error constraint can be defined according to the range of average wind speed.

[0150] 7 - Determining wind speed components

[0151] This step comprises determining the wind speed components in the at least one measurement plane by means of the measurement signal filtered by the non-stationary Kalman filter obtained in step 6. For this step, the reconstruction of the wind speed components is implemented on the basis of the measurement signal filtered by the non-stationary Kalman filter. This step thus makes it possible to determine the wind speed components in a robust and precise manner.

[0152] Various reconstruction techniques based on stronger or weaker assumptions of wind homogeneity, spatial coherence and / or fixed turbulence propagation can be used for this step.

[0153] According to one embodiment, the reconstruction of the wind speed component by means of the filtered measurement signal can be implemented by means of a geometrical reconstruction of said wind speed component based on said filtered measurement signal. In other words, for this embodiment, it is possible to implement the method for reconstructing the wind speed component implemented in step 2, that is to say, the geometrical reconstruction of the wind speed component implements a pseudo-inverse "Moore-Penrose" operation applied to the measurement signal filtered by the non-static Kalman filter.

[0154] 8 - Splitting wind speed components

[0155] It will be recalled that this step is optional. This step can be implemented in order to determine at least one characteristic of the wind speed based on the wind speed component. This step comprises segmenting in time the wind speed component obtained in the preceding step by means of a predetermined time interval. In other words, a data set of the wind speed component is generated for a predetermined time interval. Said predetermined time interval is the time interval on which it is desired to determine the characteristic of the wind speed.

[0156] According to one embodiment of the application, the predetermined time interval can be between 1 minute and 1 hour, preferably between 5 minutes and 30 minutes, and can for example be 10 minutes. These time intervals make it possible to have significant characteristics of the wind speed in order to determine the possibility of installing a wind turbine at the measurement location.

[0157] 9 - Determining characteristics of wind speed

[0158] It will be recalled that this step is optional and follows step 8. This step comprises determining at least one characteristic of the wind speed over the predetermined time interval for at least one measurement plane.

[0159] According to one embodiment of the application, the wind characteristic is chosen from the following: the average wind speed, the standard deviation of the wind speed, the maximum wind speed, the average wind direction, the wind turbulence intensity, the average of the vertical component of the wind speed and the standard deviation of the vertical component of the wind speed.

[0160] Preferably, in this step, it is possible to determine at least the wind turbulence intensity over the predetermined time interval, which corresponds to the ratio of the standard deviation of the wind speed to the wind speed. Indeed, this characteristic makes it possible to determine the category of wind turbine.

[0161] Preferably, it is possible to determine at least the wind turbulence intensity, the average wind speed and the average wind direction over the predetermined time interval. Indeed, these characteristics make it possible to determine the category of wind turbine, its location and its size.

[0162] The application also relates to a method of installing a wind turbine, in which the following steps are implemented:

[0163] - determining, by means of the method according to any one of the above variants or variant combinations, a wind speed component and / or at least one property of the wind speed at at least one site,

[0164] - installing a wind turbine at the site according to the wind speed component and / or the at least one property of the wind speed.

[0165] In the installation step, the installed wind turbine can be determined according to its size, class and structure, and also its orientation and its control according to the wind speed component and / or the at least one property of the wind speed.

[0166] According to one embodiment of the application, the first step can be repeated at a plurality of sites. Next, the site most suitable for installing a wind turbine is determined according to the wind speed component and / or the at least one property of the wind speed. This can be, in particular, the site at which the wind speed is in an operating range suitable for the recovery of energy by means of a wind turbine.

Claims

1. A method for determining a wind speed component by means of a LiDAR sensor (1), which is oriented essentially vertically in order to carry out a measurement in at least one essentially horizontal measurement plane (PM), characterized in that, The following steps are implemented: a) acquiring measurement signals in the at least one measurement plane (PM) from the LiDAR sensor (1); b) determining an approximation (APP) of the wind speed component by means of a geometrical reconstruction of the wind speed component in the at least one measurement plane (PM) based on the acquired measurement signals; c) determining a mean velocity and a standard deviation (MOY) of the wind in the at least one measurement plane (PM) by means of the approximated wind speed component; d) constructing a wind signal model (MOD) using a sum of two first order filters, the sum being a sum of an integrator and a low pass filter, the wind signal model being dependent on two parameters (a, b); e) determining the two parameters (a, b) of the wind signal model (MOD) by means of the determined standard deviation and the determined mean wind velocity; f) filtering the acquired measurement signals by means of a non-stationary Kalman filter (KAL), the wind signal model and the two determined parameters; and g) determining a wind speed component in the at least one measurement plane by means of a geometrical reconstruction (REC) of the wind speed component based on the filtered measurement signals.

2. The method of claim 1, wherein, The method comprises, for the step of approximating the wind speed component, a step of filtering (FIL) the measurement signals using a first order low pass filter.

3. The method according to any of the preceding claims, characterized in that, The wind signal model (MOD) corresponds to the Kaimal model of the wind spectrum.

4. The method of claim 1, wherein, The transfer function H of the wind signal model (MOD) is written as: where a, b are the two parameters, s is the Laplace variable, and τ is the time constant.

5. The method of claim 4, wherein, The parameters a and b are obtained using the following equations: where f is the frequency vector, ω = 2πf, k is the index corresponding to the component in question, σ k is the standard deviation of the approximated wind speed, V hub is the approximated wind speed, L is the vector of components of the direction of the measuring beam of the LiDAR sensor, and S is the vector of spectral components of the Kaimal model of the wind spectrum.

6. The method of claim 1, wherein, The covariance matrix of the non-stationary Kalman filter (KAL) is determined by minimizing a cost function which takes into account the dispersion and the mean deviation of the measurement signals.

7. The method of claim 1, wherein, The reconstruction (REC) of the wind speed component by means of the measurement signals filtered by the non-stationary Kalman filter (KAL) is achieved by means of a geometrical reconstruction of the wind speed component based on the measurement signals filtered by the non-stationary Kalman filter.

8. The method of claim 1, wherein, The following steps are also included: a) segmenting (SEG) the determined wind speed component at predetermined time intervals; and b) determining at least one characteristic (CAR) of the wind speed for the predetermined time intervals.

9. The method of claim 8, wherein, The predetermined time intervals are between 1 minute and 1 hour.

10. The method of claim 9, wherein, The predetermined time intervals are between 5 minutes and 30 minutes.

11. The method of claim 8, wherein, The at least one wind characteristic is chosen from the following: mean wind speed, standard deviation of the wind speed, maximum wind speed, mean wind direction, wind turbulence intensity, mean value of the vertical component of the wind speed and standard deviation of the vertical component of the wind speed. The following steps are implemented: a) acquiring measurement signals in the at least one measurement plane (PM) from the LiDAR sensor (1); b) determining an approximation (APP) of the wind speed component by means of a geometrical reconstruction of the wind speed component in the at least one measurement plane (PM) based on the acquired measurement signals; c) determining a mean velocity and a standard deviation (MOY) of the wind in the at least one measurement plane (PM) by means of the approximated wind speed component; d) constructing a wind signal model (MOD) using a sum of two first order filters, the sum being a sum of an integrator and a low pass filter, the wind signal model being dependent on two parameters (a, b); e) determining the two parameters (a, b) of the wind signal model (MOD) by means of the determined standard deviation and the determined mean wind velocity; f) filtering the acquired measurement signals by means of a non-stationary Kalman filter (KAL), the wind signal model and the two determined parameters; and g) determining a wind speed component in the at least one measurement plane by means of a geometrical reconstruction (REC) of the wind speed component based on the filtered measurement signals. The method comprises, for the step of approximating the wind speed component, a step of filtering (FIL) the measurement signals using a first order low pass filter. The wind signal model (MOD) corresponds to the Kaimal model of the wind spectrum. The transfer function H of the wind signal model (MOD) is written as: The covariance matrix of the non-stationary Kalman filter (KAL) is determined by minimizing a cost function which takes into account the dispersion and the mean deviation of the measurement signals. The reconstruction (REC) of the wind speed component by means of the measurement signals filtered by the non-stationary Kalman filter (KAL) is achieved by means of a geometrical reconstruction of the wind speed component based on the measurement signals filtered by the non-stationary Kalman filter. The following steps are also included: a) segmenting (SEG) the determined wind speed component at predetermined time intervals; and b) determining at least one characteristic (CAR) of the wind speed for the predetermined time intervals. The predetermined time intervals are between 1 minute and 1 hour. The predetermined time intervals are between 5 minutes and 30 minutes. The at least one wind characteristic is chosen from the following: mean wind speed, standard deviation of the wind speed, maximum wind speed, mean wind direction, wind turbulence intensity, mean value of the vertical component of the wind speed and standard deviation of the vertical component of the wind speed.

Images