• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:
    The object of the invention relates to a method of acquisition and modeling by a LiDAR sensor of an incident wind field. The acquisition and modeling includes a step of estimating the amplitudes and directions of the wind for a set of discretized points, as well as a step of reconstruction of the incident wind field in three dimensions and in real time. The invention also relates to a method for controlling and / or monitoring a wind turbine equipped with such a LiDAR sensor from the reconstructed incident wind field in three dimensions and in real time.
    公开号:FR3068139A1
    申请号:FR1755675
    申请日:2017-06-21
    公开日:2018-12-28
    发明作者:Hoai-Nam NGUYEN；Fabrice Guillemin
    申请人:IFP Energies Nouvelles IFPEN；
    IPC主号:
    专利说明:

    The present invention relates to the field of LiDAR (Light Detection And
    Ranging or light detection and localization) used as a means of remote sensing to measure wind speed. It also relates to the field of wind turbines equipped with LiDAR sensor, as well as the control thereof.
    The performance of LIDAR sensors in terms of precision, reliability and availability of measurements, makes it possible to develop estimates and predictions of a wind state for a target aerosol volume. However, LiDAR sensors have certain limitations in terms of accuracy and data availability. On the one hand they only provide a raw measurement of the wind, i.e. a projection of the wind on a measurement axis (otherwise called LASER beam for light amplification by stimulated emission of radiation or amplification of light by a stimulated emission radiation) and on the other hand they only allow access to a limited and noisy bandwidth of the spectral content of the wind. As the raw measurement is an indirect measurement of the wind which corresponds to the projection of the wind on the axis of a LASER beam, it is then necessary to combine several raw measurements of several beams (or measurement axes) of distinct directions, to obtain an accurate estimate of the wind vector.
    However, such estimates are not accessible in a trivial or direct manner, and require the design and development of precise and robust reconstruction algorithms related to the variable quality of the signal, the geometry of the sensor, and the wind conditions. .
    Most of the reconstruction methods developed so far are based on the assumption of a homogeneous and stable wind field over the entire surface swept by the rotor as described in the publication A tutorial on the dynamics and control of wind turbines and wind farms', In 2009 American Control Conference. IEEE. 2009, pp. 20762089.
    However, this assumption is neither representative nor realistic, given that the wind speed varies considerably depending on the altitude, within the atmospheric boundary layer, with very complex dynamics.
    The publication of “P Towers and B Ll Jones,‘ Real-time wind field reconstruction from LiDAR measurements using a dynamic wind model and State estimation ’, In Wind Energy 19.1 (2016), pp. 133-150 ”, offers an estimation algorithm to reconstruct a wind field. The approach consists in using an “unscented” Kalman filter incorporating a flow model based on the simplified Navier-Stocks equations. However, this technique provides a two-dimensional (2D) reconstruction of the wind field, at a fixed altitude. In addition, the technique, as described in this publication, is based on an unrealistic assumption that all LiDAR measurements are available for all beams at the same time.
    Finally, there is also known a reconstruction algorithm proposed by certain manufacturers of LiDAR sensors. In this case, the principle is to obtain an instantaneous estimate of the wind speed at unmeasured points in space, from interpolations on the measurements. However, in such cases it is only possible to obtain, in real time and online, an estimate of the wind component in the axis of the LiDAR. Longitudinal wind speed and direction are only obtained on the basis of a sliding average, and cannot be used for real-time applications.
    In the field of wind turbines, the productivity of wind turbines and maintenance costs depend strongly on the monitoring capacity of the system, and in particular on the capacity to exploit relevant wind information. Indeed, the main sources of damage to the structure and organs of the wind turbine are related to wind conditions involving extreme loading (strong turbulent wind, gusts) and to the fatigue of materials subjected to vibratory and oscillating phenomena. These are generated by the interactions between the wind turbine and the wind field, with in particular vibration problems exciting the eigen modes of the wind turbine. There are certain control strategies that are currently implemented but they do not have reliable wind information that can be integrated into the control loop to ensure the expected operating time. In some cases the rotor speed is regulated by the generating torque and the aerodynamic torque (via the orientation of the blades). In other cases there is no direct use of wind measurement in the control loop, which means that the speed regulation of the rotor is done in feedback. We can also have an alignment from an anemometric sensor located in turbulent zone (nacelle) and subjected to drifts which leads to having a wind turbine often misaligned.
    In all cases, this requires integrating constraints into the design of the wind turbine, with reinforced structures and an associated additional investment cost, and also with a loss of production and associated loading risks of the structure.
    In order to overcome the drawbacks mentioned above, a first aspect of the invention therefore consists in developing an improved method for estimating the speed and the direction of a three-dimensional (3D) wind field online, in real time, in a volume located upstream of a LiDAR sensor so as to have an estimate and a short-term forecast of the incident wind field on the LiDAR sensor. A second aspect of the invention aims to use this method and this LiDAR sensor in a wind turbine control strategy so as to have wind turbine rotor load forecasts, to detect gusts, turbulences, shears , etc.
    To this end, the invention relates to a method of acquisition and modeling by a LiDAR sensor of an incident wind field in a space located upstream of said LiDAR sensor. For the process, the following steps are carried out:
    a) a step of meshing the space situated upstream of said LiDAR sensor in which the meshing of the space is carried out by a set of discretized points positioned according to a predefined three-dimensional grid which comprises a set of meshes composed of estimation points and measurement points.
    The meshing step makes it possible to discretize (or sample) the space upstream of the LiDAR sensor in a three-dimensional grid composed of discretized points and to be able to make these different discretized points coincide either in measurement points or in estimation points necessary for modeling process. It also makes it possible to relatively position the measurement and estimation points between them and to know the distances separating all of these discretized points.
    b) a step of measuring the amplitude and direction of the wind at the various measurement points located in the upstream space and positioned at least two separate distances from the LiDAR sensor, along at least three measurement axes ,
    The measurements carried out in this stage make it possible to obtain sufficient and reliable initial data to feed an algorithm intended to estimate the amplitude and the direction of the wind at the estimation points.
    c) a step of estimating the amplitude and the direction of the wind at any time over all of the estimation points and the estimation is carried out by means of optimization using a weighted recursive least squares method a cost function J which uses at least the data of the measured points, spatial coherence data of the wind speed, temporal coherence data of the wind speed, as well as data qualifying the quality of the measurements carried out on the measurement points.
    The taking into account of these different parameters in a cost function to be optimized is what will allow access to an estimate of the amplitude and direction of the wind at each point of estimation of the mesh.
    d) a step of reconstruction, in real time and in a defined reference frame, of the incident wind field in three dimensions (3D) from the amplitudes and directions of the wind estimated and measured for each point.
    This step allows the incident wind field to be reconstructed in 3D in the volume sampled by the three-dimensional grid. In this step, a history of LiDAR measurements is made, which makes it possible to know the past states of the wind field, and this is incorporated into the synthesis of the current and future estimates of the 3D wind field, which allows reconstruction in real time. .
    The advantage of using an optimization approach, using a weighted least squares recursive form, is to be able to determine a complete three-dimensional (3D) image of the incident wind propagating in the space located upstream of the LiDAR sensor.
    According to one aspect of the invention, the measurement m of the amplitude and direction of the wind at a measurement point is given by a relation of the form:
    where Vj, x (Jc), Vj, y (k), Vj, z (k) are values of the wind speed projected on a reference x, y, z at an initial time (k), and ag, bj , Cj with j = 0, 1, 2, 3, 4, are coefficients of measure, which are given as, dj = b ; = sin () cus (çy).
    where, (pj are respectively the zenith and the azimuth of the measurement axis in a spherical coordinate system.
    In this way the wind vector, at each sampled instant, for all the points in space is composed of the three components which will make it possible to determine the complete image in three dimensions. Furthermore, the choice of measurement coefficients makes it possible to depend only on the angles of the beams and is not a function of the measurement distances, which facilitates the computer programming of the cost function J.
    According to one aspect of the invention, the cost function J at any time (t) is written in the following form:
    J [t] = (-U0) - ilQ}) r P 0 -i U '( ( B -)) + £ Ulj] - ζΗ (f (j] -ωί / j - 1)) + + SU / Y 'jvu'i /) + y] (c /) - ui m i) - fiiirS ι) i, Î = 1 * J = i where ω is an ordered vector composed of all the components of the velocity of the points of the space where the wind is estimated, ώ (0) is the estimate of the wind speed at time 0, P o , Q,
    R s and R m are weighting matrices of appropriate dimension, and C s , C m are matrices which take into account wind speed and measurement noise.
    Using such a cost function, it is possible to estimate the wind speed at an estimation point. In addition, such a function allows a clear interpretation of the weighting matrices P o , Q, R s and R m .
    According to one aspect of the invention, the measurements of the amplitude and direction of the wind at the different measurement points are carried out at a sampling rate of at least 0.25 Hz. The use of such a sampling frequency range has the effect of obtaining several measurements simultaneously on the same measurement axis while having measurements which are reliable and precise.
    According to one aspect of the invention, the measurements of the amplitude and the direction of the wind at the different measurement points are taken at at least two different distances along the measurement axis. Measurements made at at least two distances make it possible to define a three-dimensional volume sufficient to encompass the blades of a wind turbine as will be described later.
    According to one aspect of the invention, the measurements of the amplitude and the direction of the wind are taken along at least three axes of measurement. Having at least three measurement axes allows a fine mesh of the space upstream and also allows to have a sufficient quantity of measurements for the step of estimating the wind speed.
    According to one aspect of the invention, the spatial coherence of the wind speed along x, y and z axes of a Cartesian coordinate system is estimated by a formula of the type:
    or :
    with
    o Ci characterizes the variation of the wind speed for an estimation domain along the longitudinal axis xet o Cf characterizes the variation of the wind speed for an estimation domain along the lateral axis y and o C v characterizes the variation of the wind speed for an estimation domain along the vertical axis z and
    Such a characterization has the effect of making possible the computer coding of such a function.
    According to one aspect of the invention, the spatial coherence of the wind speed along the x, y and z axes of the Cartesian coordinate system is estimated with the following assumptions:
    o The variation of the wind speed along the longitudinal axis x is small and the partial derivative dv x / dx is relatively small along the longitudinal axis, o the wind changes smoothly along the axis lateral y and the partial derivative dv x / dy is small along the lateral axis y, where the wind changes with a power law along the vertical axis z which is given by:
    El - t lr
    where a is an exponent of the power law, TY is the longitudinal wind at an altitude x above the ground, and z is a reference altitude.
    Such assumptions are realistic and allow reliable and accurate wind speed estimates.
    According to one aspect of the invention, the quality of the measurements carried out by the LiDAR is represented by a model of the form:
    where 6 m describes the measurement noise.
    The formulation of this type makes it possible to take into account the inaccuracies of the LiDAR measurements.
    According to one aspect of the invention, the estimation of the amplitudes and directions of the wind field at an instant (t) over all of the estimation points is given by the following formula:
    x [i) - - 1) 4- j - Cx t - 1))
    The previous formula has the advantage of linking the estimates of the wind speed over time for the estimation points.
    The invention also relates to a computer program product which comprises code instructions arranged to implement the steps of the acquisition and modeling method described above. The program is executed on a LiDAR processing unit.
    The invention also relates to a LiDAR sensor which includes in memory the code instructions of a computer program product as described above and which is arranged to execute such a computer program product.
    In this way, a LiDAR sensor that executes such a computer program product will return reliable information from an incident wind field in three dimensions in real time.
    An object of the invention also relates to a wind turbine which comprises a sensor
    LiDAR as previously described.
    According to one aspect of the invention, the LiDAR sensor is placed on the nacelle of said wind turbine.
    Finally, the invention also relates to a method for controlling and / or monitoring a wind turbine equipped with a LiDAR sensor and a control automaton, and the method comprises the following steps:
    a) A step in the development of a control strategy in anticipation of said wind turbine by exploiting the reconstruction of the incident wind field in three dimensions and in real time,
    b) A piloting step integrating the elaborate control strategy which consists in piloting the angle of the blades or the orientation of the nacelle.
    In this way the provision of sufficiently robust and precise information on the incident wind state when approaching the rotor allows a new control approach, with the integration of a dynamic and preventive pre-positioning term. In addition, the ability to reconstruct online, in real time, an incident wind field on approach to the rotor plane opens up many operational perspectives: quantification of wind turbine misalignment, power curve, nacelle transfer function , burst detection, monitoring and diagnosis of loading and fatigue risks, optimization of preventive maintenance, resource analysis, production optimization. This then increases the efficiency of the wind turbines, reduces maintenance costs, increases the service life of the components and reduces investment costs by optimizing the design.
    Brief presentation of the figures
    Other characteristics and advantages of the method according to the invention will appear on reading the following description of a nonlimiting exemplary embodiment, with reference to the appended figures and described below.
    FIG. 1 illustrates a wind turbine equipped with a LiDAR sensor according to the invention.
    FIG. 2 illustrates the steps of the acquisition and modeling process by the LiDAR sensor according to the invention.
    Figure 3 is a front view of the mesh of the space according to the invention.
    Figure 4 is a perspective view of the mesh of the space according to the invention.
    Figure 5 illustrates a 3D wind field reconstructed from LiDAR measurements in a particular case.
    FIG. 6 illustrates the steps of the method of piloting the wind turbine according to the invention.
    Detailed description of the invention
    notations
    During the description, the following notations are used:
    x, y, z: directions of the three-dimensional coordinate system, with z the vertical axis and x the main wind direction.
    Θ and φ: orientation angles of said LiDAR sensor. These angles are explained in Figure 1: the angle Θ is the angle made by the projection of the measurement axis of LiDAR in the plane (y, z), and φ is the angle made by the projection of l measurement axis of LiDAR in a plane consisting of the x axis and the projection of the measurement axis of LiDAR in the plane (y, z).
    m (t): measurement of the LiDAR sensor at a measurement point.
    - Vjy (Æ), Vj, z (k): projections of the wind speed on x, y, z.
    - ω: ordered vector composed of all the components of the wind speed at points in space where the wind is estimated on the x, y and z axes of the three-dimensional coordinate system.
    - i5 (t): estimate of ω (ί) at time t.
    P (t): time-varying auxiliary matrix, which can be obtained at time t.
    P o , Q, Fl s and Rm are weighting matrices of appropriate dimension.
    In the following description, the term “LiDAR” is used to designate a “LiDAR” sensor.
    The invention relates firstly to a method of acquisition and modeling by an LiDAR sensor of an incident wind field with the aim of estimating the wind speed and direction for a wind field approaching and upstream of the LiDAR and this in the most reliable way possible. This estimate must be made online, in real time, for a sampled 3D wind field.
    FIG. 2 represents the different stages of the acquisition and modeling process according to the invention:
    1. Mesh (MA) of the space located upstream of said LiDAR sensor, the mesh comprises estimation points (PE) and measurement points (PM).
    2. Measurement (MES) of the amplitude and direction of the wind at the different measurement points (PM).
    3. Estimation (EST) of the amplitude and direction of the wind at any time (t) for all the estimation points (PE).
    4. Reconstruction (MOD 3D) of the incident wind field in three dimensions (3D) and in real time on all discretized points.
    FIG. 1 represents a wind turbine 1 equipped with a LiDAR sensor 2. The LiDAR sensor is used to measure the wind speed at a given distance at a PM measurement point. A priori knowledge of the wind measurement allows a priori to give a lot of information.
    There are several types of LiDAR sensor, for example scanned LiDAR sensors,
    Continuous LiDAR or pulsed LiDAR. In the context of the invention, a pulsed LiDAR is preferably used. However, the other LiDAR technologies can be used while remaining within the scope of the invention. As shown in Figure 1, which is an example of an embodiment, the LiDAR used comprises 5 beams or measurement axes (bO, b1, b2, b3, b4). Without limitation, the acquisition and modeling process also works with a LiDAR comprising three or more beams. The pulsed 5-beam LiDAR sensor is mounted on a nacelle 3 of wind turbine 1.
    Conventionally a wind turbine 1 makes it possible to transform the kinetic energy of the wind into electrical or mechanical energy. For the conversion of wind into electrical energy, it consists of the following elements:
    - a mast 4 making it possible to place a rotor (not shown) at a height sufficient to allow its movement (necessary for wind turbines with horizontal axis) or to place this rotor at a height allowing it to be driven by a stronger wind and regular than at ground level 6. Mast 4 generally houses part of the electrical and electronic components (modulator, control, multiplier, generator, etc.);
    - A nacelle 3 mounted at the top of the mast 4, housing mechanical, pneumatic components, certain electrical and electronic components (not shown), necessary for the operation of the machine. The nacelle 3 can rotate to orient the machine in the right direction;
    - The rotor, fixed to the nacelle, comprising several blades 7 (generally three) and the nose of the wind turbine. The rotor is driven by wind energy, it is connected by a mechanical shaft directly or indirectly (via a gearbox and mechanical shaft system) to an electric machine (electric generator ...) (not shown) which converts the energy collected into electrical energy. The rotor is potentially fitted with control systems such as variable angle blades or aerodynamic brakes;
    - a transmission, composed of two axes (mechanical shaft of the rotor and mechanical shaft of the electric machine) connected by a transmission (gearbox) (not shown).
    In the description set out below, the acquisition and modeling process described is theoretical and operates independently of the wind turbine 1. However, the various examples and developments are given in the case of a LiDAR mounted on the nacelle 3 of the wind turbine 1 so as to carry out the various stages of the acquisition and modeling process shown in FIG. 2 at a certain altitude relative to the ground 6.
    In this part, the different stages of the acquisition and modeling process according to the invention are described:
    1. Mesh (MA) of the space located upstream of said LiDAR sensor
    In this first step, the space upstream of the LiDAR sensor is defined according to a mesh as visible in FIGS. 1, 3 and 4. In this step a coordinate system in which the Lidar performs measurements is defined. The defined coordinate system is the direct trihedron illustrated in Figures 1 and 3. The origins x - y of this system are at the position of the LiDAR on the nacelle 3, and the origin z is at ground level 6.
    The x-axis points horizontally in the wind direction, the z-axis points vertically upward and the y-axis is perpendicular to form a direct three-dimensional coordinate system (according to the right hand rule).
    In this step, the space mesh includes a set of discretized points placed upstream and which define a three-dimensional grid. For each distance x fixed, the plane y - z is divided into cells without overlap as visible in Figure 3. The mesh includes measurement points (PM) and estimation points (PE) of the wind speed.
    In connection with this grid of space, we also define underlying variables, called optimization variables, necessary for the estimation step described below. In order to allow a clever and efficient implementation of the optimization algorithm described below, all the optimization variables are gathered in an ordered vector, noted ω. The order determined for these optimization variables is a crucial engineering element for the feasibility and performance of a coding algorithm for this process.
    A vector ω is defined for each point of the discretized space and it is composed of all the components 14 of the points of space (PE) where the wind is estimated, followed respectively by the components v y and v z . The estimation of the wind speed in n points involves the construction of a vector ω of size 3n, with Wi containing all the v x , W „+ 1 containing all the v y , and HLn + 1 containing all the v z .
    The following example is given for the components v x of the wind speed, it being understood that the method is identical for v y and v z . As was done in the initial step, and as visible in Figure 3, the space is discretized in x, y and z with n x points in x, n y points in y and n z points in z.
    In this configuration we have:
    We define by v, _ M the component of the wind speed v x , whose coordinate is (x ,, y, z k ). The index / of VVz, where the corresponding estimate is located, is obtained as follows:
    l = - i} n y n z -h (k - 1 -rj
    For example, if i = n x , k = 1 and j = 1, then l = - (k - l) r y - d - 1
    This corresponds to the upper left corner of the estimation domain, at the furthest distance upstream from the rotor plane, as illustrated in Figure 4.
    2. Measurement (MES) of the amplitude and direction of the wind at different measurement points
    Secondly, the LiDAR sensor performs a measurement m (t) relating to the wind speed at a measurement point (PM) located upstream of the wind turbine 1. This measurement m (t) corresponds to the signal received by the sensor from the measurement point (PM) in response to the signal from the LiDAR sensor. Indeed, by interferometry and Doppler effect, part of the Laser signal emitted by the LiDAR sensor is reflected by the air molecules at the measurement point and also by aerosols (dust and microparticles in suspension). The measurement point is defined by the characteristics of the LiDAR sensor, in particular the focal distance, as well as by its orientation. This measurement, dependent on the wind speed, is a time and depends on the orientation of the LiDAR sensor.
    For the case of pulsed LiDAR, the measurements are obtained successively according to the mesh defined in the previous step, starting with the longitudinal beam bO, then the oblique beam b1, up to the beam b4. An interesting feature of this system is that it allows to measure the projection of the wind speed at several distances, simultaneously, for a given beam. It is thus possible to obtain, for example, 10 successive distances between 50m and 400m, at a sampling rate of 0.25Hz or 1 Hz. It is of course possible to limit oneself to two measurements, which are sufficient to reconstruct a three-dimensional model. At each sampling time, only the measurements of the current beam selected are refreshed.
    In a particular case, in accordance with FIG. 4, the measurements are made at seven distances and in particular at x = [50 80 120 160 200 240 280] m for the five beams. Thus for each x fixed, the plane y - z is divided into cells as follows:
    • The first four points (PM) correspond to the y - z coordinates of the measurement points for the beams 1,2, 3, 4 of the distance x = 280m.
    • The four second points (PM1) correspond to the y - z coordinates of the measurement points for the beams 1,2, 3, 4 of the distance x = 240m.
    • The four third points (PM2) correspond to the y - z coordinates of the measurement points for the beams 1,2, 3, 4 of the distance x = 200m.
    • The four fourth points (PM3) correspond to the y - z coordinates of the measurement points for the beams 1,2, 3, 4 of the distance x = 160m.
    • The four fifth points (PM4) correspond to the y - z coordinates of the measurement points for the beams 1,2, 3, 4 of the distance x = 120m.
    • The four sixth points (PM5) correspond to the y - z coordinates of the measurement points for the beams 1,2, 3, 4 of the distance x = 80m.
    • The four seventh points (PM6) correspond to the y - z coordinates of the measurement points for the beams 1,2, 3, 4 of the distance x = 50m.
    • The central point (PM7) corresponds to the y - z coordinates of the measurement points for beam 0 for all distances.
    The measures m (k) of LiDAR for the beams j = 0, 1,2, 3, 4 at the distance x meters, and at the instant k are given by the formula mj, x (k), with j = 0 , 1,2, 3, 4.
    For example, m 050 (1) is the measure of LiDAR for the beam j = 0 at the distance x = 50 meters and at the instantaneous instant k = 1. In the context of the invention, the measure LiDAR is then given by a formula of the type:
    where Vj, y (k), Vj, z (k) are values of the wind speed projected on a reference given at the initial time (k), and O , bj, Cj, with j = 0, 1, 2, 3, 4 are coefficients of measure, which are given as,
    HJ = CûS (. B 1 - î cos! /,).
    = ^ in (fl,! siiif j f !
    where, (pj, with j = 0, 1,2, 3, 4 are respectively the zenith and the azimuth of the measurement axis in a spherical coordinate system.
    The advantage of defining the LiDAR measurement equation in the previously defined coordinate system, with the choice of spatial discretization chosen, is that it can be used directly, since the coordinates of the measurement point coincide with a particular point of the discretized space.
    3. Estimate (EST) of the amplitude and direction of the wind at any time (t) on all the discretized points
    This step consists in obtaining a wind value on the estimation points (PE) of the mesh.
    For this purpose the estimation is carried out by means of the optimization by a method of least recursive squares weighted with a cost function which uses the measured data m (k) of the LiDAR, but also spatial coherence data of the speed of the wind, data of temporal variation of the wind speed, as well as data qualifying the quality of the measurements m (k) of the LiDAR. This is what is explained below.
    3.1 Spatial differences
    These subsections aim to define the spatial coherence data of the wind in the context of the invention and more particularly in the context of a LiDAR mounted on the nacelle 3 of a wind turbine 1.
    In this step, we consider the components of the wind speed on the x, y and z axes of the previously defined benchmark.
    During this estimation step, it is accepted that the wind speed changes relatively little in space, and that the wind has a strong spatial coherence in a small volume of space. The following presentation is made here for the components v x , that is to say for the first n variables of ω with an estimation domain represented in FIG. 4 (The approach is similar for the components v y and v z ) and taking n x = n y = n z = 3.
    3.1.1 Longitudinal difference
    The longitudinal difference corresponds to the change in v x along the x axis and this changes slowly according to the invention. In this case the partial derivative dv x / dx is relatively small. In other words, "I i" 4 '[ij' w τ I
    The previous equation can be written in a compact vector form like:
    or
    U j IJ -1
    U ... 11 o -i ... o
    U U ... -J- i ll li ... -1
    It should be noted that each line of C x / contains a +1 and a -1.
    Similarly, we can calculate the variation of v y and v z along the longitudinal axis as:
    j (. -t 11, ί C- -c. 'Pi 0'
    where Cyi, C zl are, coefficients matrices, which contain only a +1 and a -1 on each line.
    By defining :
    G / C.,]} Gi we obtain the equation which characterizes the variation of the wind speed for the domain of estimation along the longitudinal axis.
    3.1.2 Lateral difference _The lateral difference is the change of v x along the y axis. Similarly, since the wind changes smoothly, the partial derivative dv x / dy is relatively small. In other words, rf _>'• J | ju, | -, Α, j ~ u - '-v'.} 0
    G
    LJ ΐγ I!
    We can write the previous equation in a compact vector form like n
    L V
    OR
    ... 0 0 '
    ... 0 0
    i -i
    Each line of C a contains a +1 and a -1.
    Similarly, the variation of v y and v z along the lateral axis can be calculated as,
    where Cyt, Czt are matrices of coefficients which contain only a +1 and a -1 for each line.
    By defining :
    It is clear that the equation:
    characterizes the variation in wind speed for the range of estimation along the lateral axis.
    3.1.3 Vertical difference
    The vertical profile of the wind speed is given by a power law which allows a much more precise description of the wind speed component v x at different heights.
    The vertical wind speed profile describes the evolution of the longitudinal wind speed as a function of the relative altitude on the ground. The power law of the wind speed profile is generally used to estimate the longitudinal wind speed v, at an altitude above the ground z, taking into account the longitudinal wind speed v / r at a reference altitude z r , using the equation,
    where alpha is the power law exponent, which is usually specified in terms of stability.
    The constant value alpha = 1/7 is commonly used, consistent with a relatively low wind shear hypothesis 5. However, it should be noted that to consider constant alpha is to disregard the roughness of the ground surface, the interactions of the wind with possible obstacles, and the stability of the atmosphere.
    Using this power law, we have a vertical difference in wind given by:
    '
    - <- '2 - } •' C Æ θ <.....
    u. r 2j - (êj 1 ) 0 where Zj is the height of ω, a is the power law exponent, which is assumed to be 1/7.
    We can write the previous equation in a compact vector form like:
    L 0; O; ©
    G 0 0 +1 fl li - (- t or + i n © - (ή-) π ci o + i ο ϋ - (ri) o
    GGGG ΟΙ 5 Similarly, we can quantify the variation of v y and v z along the vertical axis as
    0
    VS ; (L ’RC 0
    However, as the power law of the wind profile only applies to the longitudinal wind speed, C yv , C zv contain only a +1 and a -1 for each line.
    By defining :
    We get the equation:
    C,. = Which characterizes the variation of the wind speed for the estimation domain along the vertical axis.
    Finally, using:
    and
    we have :
    where equivalently, which is the equation that characterizes the variation of the total wind speed along the x, y and z axis.
    With:
    G
    G c ,.
    3.2 LiDAR measurements
    For calculation purposes, it is important to rewrite the measurement equation in vector form of w. In the previous example of a LiDAR with five beams and for seven measurements per beam, we add aj = 0, 1,2, 3, 4, and x = [50, 80, 120, 160, 200, 240, 280] , = L 0
    C-U = iü
    0îl : = c
    OU = Ci
    By combining with:
    m (k) = Hj i y T (k) - h j v 7 , J /.
    GG
    J /, we get, m
    Μ '= C
    J, T- 4 '
    WHERE (-jT - ( JJ
    G Ct G i which can be rewritten in a compact vector form (Γ ΐη Ή · m m or
    Cn vi ! Su C p; - m î ûaij . G -> ij
    To take measurement noise into account, a more realistic model for Lidar measurements can be introduced as follows, where £ m describes the measurement noise.
    3.3 The weighted recursive least squares method
    It is recognized that the wind speed changes little not only in space, but also in time. In the following, we provide a way to take this information into account in the optimization approach. ω (0) is the estimate of the wind speed at time 0. At each instant, the optimization problem is as follows:
    min J i f) with
    J (t'i = u (0) - ^! N) i T Gr L UHH j — 1 • L ·. ! JTï - j. 11)) j - | - Σ i j- j / 1 - j. · Ij - 1 jj lj G j. ' ijj - i / - 1 π η ( = i (G hi + Σ îf i - i T ! C „, r () 1 - îîg.gi} i
    7 == 1
    There are four terms in the previous cost function.
    • The first term penalizes knowledge of the initial wind speed ω (0).
    • The second term penalizes the variation of the wind speed over time.
    • The third term penalizes the variation of the wind speed in space.
    • The fourth term penalizes the Lidar measurement quality.
    Using the previous formula, we can have a clear interpretation of the weighting matrices P o , Q, R s and R m . So :
    • If the wind speed ω (ί) at time t = 0 is well known, then ω (0) = ω (0), then P o is small. Otherwise P o is large.
    • If there are many variations in wind speed over time, then Q is large. Otherwise Q is small.
    • If the wind speed changes quickly, then R s is large. Otherwise R s is small.
    • If there is a lot of noise in the Lidar measurements, then R m is large. Otherwise, R m is small.
    In the case where the following three limiting cases are considered:
    • No information on the initial wind speed is available. Therefore P o is very large. The term :
    can thus be neglected in the cost function.
    There is no relationship between the wind speed at time t and the wind speed at time t-1. In this case, we can choose O very large. The following term can be overlooked:
    J = 1 • The variation in wind speed in space is very small. In this case, we can take R s very small. The following term is important in the cost function:
    We define :
    C- H
    J s 0 0
    The weighted recursive least squares method used to solve the optimization problem is presented as follows:
    • We initialize the optimization variables as follows:
    f ucut = λn i.
    I Pî û! = Pi • At any time t:
    we define :
    I) where 0 is a zero vector of appropriate size.
    We compute an auxiliary matrix K such that
    K = {P {t - 1) + Q} C (C T (P [f - 1 ΐ - O ii '+ 7 ! 1
    We calculate the matrix P (t) such that
    P (t ') = (I- KC) P {t - 1) where I is an identity matrix of appropriate dimension.
    The wind speed at time t is then estimated as follows:
    = upf - 1S K [tj (f} - Capt - 1})
    4. Reconstruction of the incident wind field in three dimensions (3D) and in real time
    In this step, a processor integrated into the LiDAR sensor collects all the wind amplitude and direction data measured and estimated during the previous steps. The recovery of this data is done in real time for each measurement point (PM) and estimation point (PE) defined previously. The LiDAR sensor is therefore able to reconstruct the entire incident wind field on the LiDAR as shown in Figure 5.
    In the same figure 5, a reconstructed wind field is represented for a time at 68 seconds. On the ordinate, the altitude relative to the ground is represented (in m) and on the abscissa it is represented the distance to the nacelle (in m) and the relative lateral positions at LiDAR (in m).
    The invention secondly relates to a method for controlling and / or monitoring a wind turbine equipped with a LiDAR sensor as described above and an associated control automaton 10 which comprises the following steps:
    i) A step of developing a control strategy (CON) in anticipation of said wind turbine 1 by exploiting the reconstruction of the incident wind field in three dimensions and in real time obtained by the method according to the invention, ii) A piloting stage (PIL) integrating the control strategy developed which notably consists in piloting the angle of the blades 7 or the orientation of the nacelle 3.
    FIG. 6 represents the overall operation of such a wind turbine 1. The wind turbine 1 comprises for this purpose a LiDAR sensor 2 according to the invention, and its processing unit, a computer device comprising a software solution for 3D reconstruction of the field of wind, a control automaton integrating the control strategy and a device for piloting the blades and / or the nacelle of the wind turbine. In connection with Figure 6, the invention applied to a wind turbine works as follows:
    First, the LiDAR performs the step of acquiring and modeling the incident wind field as described above so as to reconstruct a 3D incident wind field (steps ME, MA, EST, MOD 3D in FIG. 6), • Secondly, the control robot 10 draws up the control strategy (CON) and carries out the control (PIL) of the components of the wind turbine 1 taking account of the control strategy developed.
    This method according to the invention makes it possible to analyze the incident wind in real time or detect gusts, power curves and turbulence intensities which can be used to regulate or supervise the wind turbine so as to obtain better alignment of the wind turbine, which leads to an optimization of production and a minimization of loads and fatigue.
    权利要求:
    Claims (15)
    [1" id="c-fr-0001]
    claims
    1) Method of acquisition and modeling by a LiDAR sensor of an incident wind field in a space located upstream of said LiDAR sensor characterized in that the method comprises:
    a) a step of meshing (MA) of the space located upstream of said LiDAR sensor in which the meshing of space is carried out by a set of discretized points positioned according to a predefined three-dimensional grid which comprises a set of meshes composed of points estimation and measurement points (PM),
    b) a measurement step (MES) of the amplitude and direction of the wind at the different measurement points (PM) located in the upstream space and positioned at least two distinct distances from the LiDAR sensor, along at least three measurement axes,
    c) an estimation step (EST) of the amplitude and direction of the wind at any time (t) over all of the estimation points and the estimation is carried out by means of optimization by a recursive weighted least squares method of a cost function which uses at least the data of the measurement points (PM), spatial coherence data of the wind speed, temporal coherence data of the wind speed, as well as data qualifying the quality of the measurements carried out on the measurement points,
    d) a reconstruction step (MOD 3D), in real time and in a defined reference frame, of the incident wind field in three dimensions (3D) from the amplitudes and directions of the wind estimated and measured for each point of said mesh (MA ).
    [2" id="c-fr-0002]
    2) Method according to claim 1, characterized in that the measurement m of the amplitude and direction of the wind at a measurement point (PM) is given by a relation of the form:
    1) = Nj i'j.T 1 / ') ~ Pj.i) i) 4 “bj Pj.zf) where Vy, x (Æ), Vjj / Jt), Vj, z (k) are values of the wind speed projected on a reference given at the initial time (k), and CLj, bj, Cj with j = 0, 1, 2, 3, 4 are measurement coefficients, which are given as, (7, = CUS h / 'i.
    f = sni! 0, î (u.-ççç).
    u, = sm (0, i sin (j where θ , φϊ, j = 0, 1, 2, 3, 4 are respectively the zenith and the azimuth of the measurement axis in a spherical coordinate system.
    [3" id="c-fr-0003]
    3) Method according to one of the preceding claims, characterized in that the cost function J at any time (t) is written in the following form:
    in which ω is an ordered vector composed of all the components of the speed of the points in space where the wind is estimated, ώ (0) is the estimate of the wind speed at time 0, P o , Q, R s and R m are weighting matrices of appropriate dimension, and C s , C m are matrices which take into account wind speed and measurement noise.
    [4" id="c-fr-0004]
    4) Method according to one of the preceding claims, characterized in that the measurements of the amplitude and the direction of the wind at the different measurement points (PM) are carried out at a sampling rate of at least 0.25Hz .
    [5" id="c-fr-0005]
    5) Method according to one of the preceding claims, characterized in that the measurements of the amplitude and the direction of the wind at the different measurement points (PM) are taken at at least two different distances along the axis of measured.
    [6" id="c-fr-0006]
    6) Method according to one of the preceding claims, characterized in that the measurements of the amplitude and direction of the wind are taken along at least three axes of measurement.
    [7" id="c-fr-0007]
    7) Method according to one of the preceding claims, characterized in that the spatial coherence of the wind speed along the x, y and z axes of a Cartesian coordinate system is estimated by a formula of the type:
    C, / 0 with Γ C;
    Ci = Ct .G.
    or :
    o C / characterizes the variation of the wind speed for an estimation domain along the longitudinal axis x and o C t characterizes the variation of the wind speed for an estimation domain along the lateral axis y and o C v characterizes the variation of the wind speed for an estimation domain along the vertical axis z and o the vector ω is an ordered vector composed of all the components of the wind speed at the points of l space where the wind is estimated.
    [8" id="c-fr-0008]
    8) Method according to the preceding claim, characterized in that the spatial coherence of the wind speed along the x, y and z axes of the Cartesian coordinate system is estimated with the following assumptions:
    o The variation of the wind speed along the longitudinal axis x is small and the partial derivative dv x / dx is relatively small along the longitudinal axis, o the wind changes smoothly along the axis lateral y and the partial derivative dv x / dy is small along the lateral axis y, where the wind changes with a power law along the vertical axis z which is given by:
    where alpha is an exponent of the power law, v, is the longitudinal wind at an altitude z above the ground, and z r a reference altitude.
    [9" id="c-fr-0009]
    9) Method according to one of the preceding claims, characterized in that the quality of the measurements made by the LiDAR sensor is represented by a model of the form:
    C ît ^ îa- - Dim ~ t - L
    Where 6 m describes the measurement noise.
    [10" id="c-fr-0010]
    10) Method according to one of the preceding claims, characterized in that the estimation of the amplitudes and directions of the wind field at a time (t) over all of the estimation points is given by the following formula:
    = χ-ff - 1) + K {y (f] - Cxjf - 1))
    [11" id="c-fr-0011]
    11) Computer program product characterized in that it comprises code instructions arranged to implement the steps of a method of acquisition and modeling by a LiDAR sensor of an incident wind field according to one of the preceding claims, when said program is executed on a processing unit of said LiDAR sensor.
    [12" id="c-fr-0012]
    12) LiDAR sensor characterized in that it includes in memory the code instructions of a computer program product according to the preceding claim and arranged to execute such a computer program product.
    [13" id="c-fr-0013]
    13) Wind turbine 1 characterized in that said wind turbine 1 comprises a LiDAR 2 sensor according to the preceding claim.
    [14" id="c-fr-0014]
    14) Wind turbine 1 according to the preceding claim characterized in that said LiDAR sensor is disposed on the nacelle of said wind turbine.
    [15" id="c-fr-0015]
    15) Method for controlling and / or monitoring a wind turbine 1 equipped with a LiDAR sensor 2 and a control automaton, characterized in that the following steps are carried out
    i) A step of developing a control strategy (CON) in anticipation of said wind turbine by exploiting the reconstruction of the incident wind field in three dimensions and in real time obtained by the acquisition and modeling process by a sensor LiDAR of an incident wind field according to one of claims 1 to 10, ii) A piloting step (PIL) integrating the developed control strategy which consists in piloting the blade angle 7 or the orientation of a basket 3.
    类似技术:
    公开号 | 公开日 | 专利标题
    FR3068139A1|2018-12-28|METHOD OF ACQUIRING AND MODELING BY A LIDAR SENSOR OF AN INCIDENT WIND FIELD
    Mirocha et al.2014|Implementation of a generalized actuator disk wind turbine model into the weather research and forecasting model for large-eddy simulation applications
    Matzarakis et al.2011|Sky view factor as a parameter in applied climatology-rapid estimation by the SkyHelios model
    EP2876302B1|2019-10-23|Method for controlling and monitoring a wind turbine by estimating wind speed using a lidar sensor
    Giebel et al.2006|Short-term forecasting using advanced physical modelling: The results of the anemos project
    Herges et al.2019|Robust Lidar data processing and quality control methods developed for the SWiFT wake steering experiment
    EP3657217A1|2020-05-27|Method for acquiring and modelling an ambient wind field by a lidar sensor
    Held et al.2019|Lidar estimation of rotor-effective wind speed–an experimental comparison
    EP3754341A1|2020-12-23|Method for determining the vertical profile of the wind speed upstream of a wind turbine provided with a laser remote detection sensor
    EP3650687B1|2021-06-16|Method for determining an induction factor for a wind turbine provided with a laser remote detection sensor
    EP3712621A1|2020-09-23|Method for predicting the wind speed in the rotor plane of a wind turbine provided with a sensor for remote laser detection
    Held2019|Inflow Measurements by Nacelle Mounted Lidars for Wind Turbine and Farm Control
    EP3862560A1|2021-08-11|Method for determining the wind speed in the rotor plane of a wind turbine
    FR3095246A1|2020-10-23|Method and control system of a size of a wind turbine by choice of controller by automatic learning
    CA3108187A1|2021-08-10|Method for determining wind direction with a laser remote sensing sensor
    FR3111991A1|2021-12-31|Method of determining the components of wind speed by means of a laser remote sensing sensor and by means of temporal coherence
    Cantero et al.2022|On the measurement of stability parameter over complex mountainous terrain
    Marinšek et al.2020|Towards efficient wind energy monitoring: Learning more from open source data
    EP3380947A1|2018-10-03|Method and device for forecasting cloudiness by statistical processing of data selected by spatial analysis
    Angelou et al.2022|Wind lidars reveal turbulence transport mechanism in the wake of a tree
    Graña2018|LiDAR Applications for Energy Industry
    FR3111196A1|2021-12-10|Method for determining the components of wind speed by means of a laser remote sensing sensor
    FR3065804A1|2018-11-02|METHOD FOR EVALUATING THE ACOUSTIC IMPACT OF A VIRTUAL WIND FARM
    Bayon et al.2014|Investigating LiDAR Sensing Errors for Wind Turbines Loads Reduction
    Indasi2015|Windfield investigations at lake Turkana wind farm, Kenya
    同族专利:
    公开号 | 公开日
    EP3642647A1|2020-04-29|
    CN110832351A|2020-02-21|
    US11248585B2|2022-02-15|
    CA3065892A1|2018-12-27|
    FR3068139B1|2019-12-20|
    US20200124026A1|2020-04-23|
    WO2018234409A1|2018-12-27|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    EP3712621A1|2019-03-18|2020-09-23|IFP Energies nouvelles|Method for predicting the wind speed in the rotor plane of a wind turbine provided with a sensor for remote laser detection|
    EP3754341A1|2019-06-19|2020-12-23|IFP Energies nouvelles|Method for determining the vertical profile of the wind speed upstream of a wind turbine provided with a laser remote detection sensor|
    EP3862563A1|2020-02-10|2021-08-11|IFP Energies nouvelles|Method for determining the direction of the wind by means of a laser remote detection sensor|US9086488B2|2010-04-20|2015-07-21|Michigan Aerospace Corporation|Atmospheric measurement system and method|
    US9234506B2|2011-10-14|2016-01-12|Vestas Wind Systems A/S|Estimation of wind properties using a light detection and ranging device|
    FR3088971B1|2018-11-26|2021-02-19|Ifp Energies Now|method of acquisition and modeling by a LIDAR sensor of an incident wind field|CN111472930B|2020-03-23|2021-07-23|浙江大学|Evolvable wind speed calculation method and feedforward unified variable pitch control method based on evolvable wind speed calculation method|
    CN111505596A|2020-04-16|2020-08-07|北京理工大学重庆创新中心|Three-dimensional wind field inversion method based on non-uniform sampling correction VAD technology|
    法律状态:
    2018-12-28| PLSC| Publication of the preliminary search report|Effective date: 20181228 |
    2019-06-25| PLFP| Fee payment|Year of fee payment: 3 |
    2020-06-26| PLFP| Fee payment|Year of fee payment: 4 |
    2021-06-25| PLFP| Fee payment|Year of fee payment: 5 |
    优先权:
    申请号 | 申请日 | 专利标题
    FR1755675|2017-06-21|
    FR1755675A|FR3068139B1|2017-06-21|2017-06-21|PROCESS FOR ACQUISITION AND MODELING BY AN LIDAR SENSOR OF AN INCIDENT WIND FIELD|FR1755675A| FR3068139B1|2017-06-21|2017-06-21|PROCESS FOR ACQUISITION AND MODELING BY AN LIDAR SENSOR OF AN INCIDENT WIND FIELD|
    EP18731117.0A| EP3642647A1|2017-06-21|2018-06-20|Method for acquiring and modelling an incident wind field by means of a lidar sensor|
    US16/621,450| US11248585B2|2017-06-21|2018-06-20|Method for acquiring and modelling an incident wind field by means of a LiDAR sensor|
    CN201880041899.8A| CN110832351A|2017-06-21|2018-06-20|Method for capturing and modeling an incident wind field by means of a LIDAR sensor|
    CA3065892A| CA3065892A1|2017-06-21|2018-06-20|Method for acquiring and modelling an incident wind field by means of a lidar sensor|
    PCT/EP2018/066478| WO2018234409A1|2017-06-21|2018-06-20|Method for acquiring and modelling an incident wind field by means of a lidar sensor|
    [返回顶部]