CN117057257B - Interpolation calculation method, device and equipment for anemometer tower data and storage medium - Google Patents
Interpolation calculation method, device and equipment for anemometer tower data and storage medium Download PDFInfo
- Publication number
- CN117057257B CN117057257B CN202311312131.9A CN202311312131A CN117057257B CN 117057257 B CN117057257 B CN 117057257B CN 202311312131 A CN202311312131 A CN 202311312131A CN 117057257 B CN117057257 B CN 117057257B
- Authority
- CN
- China
- Prior art keywords
- wind speed
- data
- interpolation
- training
- sample
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000004364 calculation method Methods 0.000 title claims abstract description 24
- 238000003860 storage Methods 0.000 title claims abstract description 12
- 238000012549 training Methods 0.000 claims abstract description 116
- 238000000034 method Methods 0.000 claims abstract description 40
- 230000007547 defect Effects 0.000 claims abstract description 9
- 239000011159 matrix material Substances 0.000 claims description 35
- 230000006870 function Effects 0.000 claims description 19
- 238000013507 mapping Methods 0.000 claims description 9
- 230000004913 activation Effects 0.000 claims description 6
- 230000000007 visual effect Effects 0.000 claims description 4
- 238000013528 artificial neural network Methods 0.000 abstract description 9
- 238000004422 calculation algorithm Methods 0.000 abstract description 9
- 238000012545 processing Methods 0.000 abstract description 5
- 230000008569 process Effects 0.000 description 8
- 239000013598 vector Substances 0.000 description 6
- 230000006872 improvement Effects 0.000 description 5
- 238000010248 power generation Methods 0.000 description 5
- 238000011161 development Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 238000004891 communication Methods 0.000 description 3
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 238000005259 measurement Methods 0.000 description 3
- 230000001932 seasonal effect Effects 0.000 description 3
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000005431 greenhouse gas Substances 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 238000001338 self-assembly Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0499—Feedforward networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2113/00—Details relating to the application field
- G06F2113/08—Fluids
Abstract
The application discloses an interpolation calculation method, device, equipment and storage medium of anemometer tower data, and relates to the technical field of electric digital data processing, wherein the method comprises the steps of obtaining first wind speed data of a anemometer tower and second wind speed data of a reference tower; intercepting first wind speed data by taking a data missing point of the first wind speed data as a midpoint to form a sample to be interpolated, and intercepting second wind speed data to form a training sample; defining an extreme learning model according to the data dimension of the training sample, and training the extreme learning model through the training sample to obtain an interpolation model; and outputting the sample to be interpolated to an interpolation model to obtain the wind speed interpolation at the data defect point. According to the method, training is carried out through the extreme learning machine, and finally, wind speed interpolation corresponding to the data missing points is obtained through the trained model. Compared with the traditional neural network learning algorithm, the method and the device do not need to set a large number of network training parameters manually, and the possibility of generating a local optimal solution is avoided so as to ensure accurate results.
Description
Technical Field
The present disclosure relates to the field of digital data processing technologies, and in particular, to a method, an apparatus, a device, and a storage medium for interpolating wind tower data.
Background
Wind power generation is to convert kinetic energy of wind into electric energy, wherein the wind energy is clean and pollution-free renewable energy.
In recent years, along with the increase of the installed wind power plants, the new energy power generation ratio is gradually increased, and the development of new energy power generation, particularly wind power generation, is receiving a great deal of attention. The development of wind power becomes an important way for reducing the dependence of national economy on fossil energy, solving the contradiction between energy production and consumption and reducing greenhouse gas emission to keep ecological balance. Along with the continuous increase of the double pressure of resources and environment, the development of wind power generation has become the development direction of energy utilization in the future in China and even internationally.
In the running process of the wind power plant, the wind measuring tower is used as wind resource basic data detection equipment of the wind power plant, and has important significance for wind resource assessment of the wind power plant. However, in the actual use process of the wind measuring tower, the wind measuring tower is possibly affected by the conditions of abnormal photoelectric converters, sensor faults, communication faults of the optical fiber ring network, faults of a data acquisition interface, downtime of acquisition software and hardware and the like, so that the problem of incomplete data of the wind measuring tower is caused, and the wind resource assessment is inaccurate.
Disclosure of Invention
The main purpose of the application is to provide a method, a device, equipment and a storage medium for interpolating and calculating wind measuring tower data, so as to solve the problem of inaccurate wind resource assessment caused by incomplete wind measuring tower data in the prior art.
In order to achieve the above purpose, the present application provides the following technical solutions:
an interpolation calculation method of anemometer tower data, the anemometer tower data is derived from at least one anemometer tower and at least one reference tower located in a preset area, the interpolation calculation method of the anemometer tower data comprises:
acquiring first wind speed data of the wind measuring tower based on a first preset natural time period and second wind speed data of the reference tower based on the preset natural time period;
acquiring a data missing point in the first wind speed data;
intercepting the first wind speed data by taking the data missing point as a midpoint and taking a second preset natural time length as an intercepting length to form at least one sample to be interpolated, wherein the second preset natural time length is smaller than the first preset natural time length;
intercepting the second wind speed data based on the second preset natural time length to form at least one training sample, wherein the training sample does not contain missing points;
defining an extreme learning model according to the data dimension of the training sample and training the extreme learning model through the training sample to obtain an interpolation model;
and outputting the sample to be interpolated to the interpolation model to obtain the wind speed interpolation at the data missing point.
As a further improvement of the present application, acquiring the data defect in the first wind speed data includes:
generating a rectangular coordinate system by taking natural time as a horizontal axis and taking a wind speed value as a vertical axis;
outputting the first wind speed data to the rectangular coordinate system to generate a wind speed curve;
acquiring a zero point of the wind speed curve, and acquiring a first wind speed value of a unit natural time before the zero point and a second wind speed value of a unit natural time after the zero point;
judging whether the first wind speed value and the second wind speed value are both different from zero;
if yes, acquiring a first slope and a second slope of the wind speed curve, which are respectively positioned at the first wind speed value and the second wind speed value;
judging whether the sum of the absolute value of the first slope and the absolute value of the second slope is larger than or equal to a preset threshold value;
if yes, judging the zero point as the data missing point.
As a further refinement of the present application, defining an extreme learning model according to the data dimension of the training sample and training the extreme learning model by the training sample to obtain an initial interpolation model includes:
defining the training samples asThe number of the training samples is +.>；
Wherein,，/>based on +.>Input sample data of a unit natural time length; />，/>For the second wind speed data +.>Self-assembly of unitsTraining sample data for a time duration;
defining the extreme learning model according to equation (1):
（1）；
wherein,，/>a first connection weight between an implicit layer and an output layer of the extreme learning model; />The number of the nodes of the hidden layer is the number; />For the output value of the hidden layer,；/>；/>is an activation function; />The input layer for the extreme learning model and the hidden layer +.>A second connection weight for the individual node;for the hidden layer->A threshold value for each node;
randomly generating the second connection weight and the threshold value in a feature space based on the extreme learning model through a nonlinear mapping function;
calculating to obtain an output value of the hidden layer through a second connection weight value and a threshold value which are randomly generated;
solving an optimal solution for the first connection weight by minimizing an approximate square difference according to equation (2):
（2）；
wherein,an output matrix for the hidden layer; />A target matrix for the training sample;
and obtaining the optimal solution, a second connection weight and a threshold corresponding to the optimal solution, and substituting the second connection weight and the threshold into the extreme learning model to obtain the initial interpolation model.
As a further refinement of the present application, the optimal solution is characterized by formula (3):
（3）；
wherein,an output matrix for said hidden layer +.>Is a mole-penrose generalized inverse matrix.
As a further improvement of the present application, the moore-penrose generalized inverse matrixCharacterized by formula (4):
（4）；
wherein,an output matrix for said hidden layer +.>Is a transposed matrix of (a).
As a further improvement of the present application, outputting the sample to be interpolated to the interpolation model to obtain a wind speed interpolation at the data defect point, and then, including:
outputting the wind speed interpolation to the first wind speed data to form interpolated wind speed data based on the anemometer tower;
and outputting the interpolated wind speed data to an external receiving end.
As a further improvement of the present application, outputting the first wind speed data to the rectangular coordinate system to generate a wind speed profile, and then, includes:
and outputting the wind speed curve to an external visual terminal.
In order to achieve the above purpose, the present application further provides the following technical solutions:
an interpolation calculation device of anemometer tower data, which is applied to the interpolation calculation method of anemometer tower data, the interpolation calculation device of anemometer tower data comprises:
the wind speed data acquisition module is used for acquiring first wind speed data of the wind measuring tower based on a first preset natural time length and second wind speed data of the reference tower based on the preset natural time length;
the data missing point acquisition module is used for acquiring the data missing point in the first wind speed data;
the sample to be interpolated intercepting module is used for intercepting the first wind speed data by taking the data missing point as a midpoint and taking a second preset natural time length as an intercepting length so as to form at least one sample to be interpolated, wherein the second preset natural time length is smaller than the first preset natural time length;
the training sample intercepting module is used for intercepting the second wind speed data based on the second preset natural time length to form at least one training sample, and the training sample does not contain missing points;
the interpolation model training module is used for defining an extreme learning model according to the data dimension of the training sample and training the extreme learning model through the training sample to obtain an interpolation model;
and the wind speed interpolation acquisition module is used for outputting the sample to be interpolated to the interpolation model so as to acquire wind speed interpolation at the data missing point.
In order to achieve the above purpose, the present application further provides the following technical solutions:
an electronic device comprising a processor, a memory coupled to the processor, the memory storing program instructions executable by the processor; and the processor realizes the interpolation calculation method of the wind measuring tower data when executing the program instructions stored in the memory.
In order to achieve the above purpose, the present application further provides the following technical solutions:
a storage medium having stored therein program instructions which, when executed by a processor, implement a method of performing interpolation calculations of anemometer tower data as described above.
According to the wind measuring method, first wind speed data of the wind measuring tower based on the first preset natural time length and second wind speed data of the reference tower based on the preset natural time length are obtained; acquiring a data missing point in the first wind speed data; intercepting first wind speed data by taking the data missing point as a midpoint and taking a second preset natural time length as an intercepting length to form at least one sample to be interpolated, wherein the second preset natural time length is smaller than the first preset natural time length; intercepting second wind speed data based on a second preset natural time length to form at least one training sample, wherein the training sample does not contain missing points; defining an extreme learning model according to the data dimension of the training sample, and training the extreme learning model through the training sample to obtain an interpolation model; and outputting the sample to be interpolated to an interpolation model to obtain the wind speed interpolation at the data defect point. According to the method, the data missing points are taken as midpoints, the data missing points are intercepted to serve as samples to be interpolated, the data of the reference tower are intercepted to serve as training samples in the same time dimension, the training is carried out through an extreme learning machine, and finally the wind speed interpolation corresponding to the data missing points is obtained through a trained model. Compared with the traditional neural network learning algorithm, the method and the device do not need to manually set a large number of network training parameters, the possibility of generating a local optimal solution is avoided to a large extent, the hidden layer node number of the extreme learning model is only required to be set, the connection weight and the threshold value of the model do not need to be adjusted in the algorithm execution process, and the method and the device have the advantages of being high in learning speed and good in generalization performance.
Drawings
FIG. 1 is a schematic flow chart of steps of an embodiment of a method for interpolating wind tower data according to the present application;
FIG. 2 is a schematic diagram of a functional module of an embodiment of an interpolation computation device for anemometer tower data according to the present application;
FIG. 3 is a schematic structural diagram of one embodiment of an electronic device of the present application;
FIG. 4 is a schematic diagram illustrating the structure of one embodiment of a storage medium of the present application.
Detailed Description
The following description of the technical solutions in the embodiments of the present application will be made clearly and completely with reference to the accompanying drawings in the embodiments of the present application, and it is apparent that the described embodiments are only some embodiments of the present application, not all embodiments. All other embodiments, which can be made by one of ordinary skill in the art without undue burden from the present disclosure, are within the scope of the present disclosure.
The terms "first," "second," "third," and the like in this application are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defining "first," "second," and "third" may explicitly or implicitly include at least one such feature. In the description of the present application, the meaning of "plurality" means at least two, for example, two, three, etc., unless specifically defined otherwise. All directional indications (such as up, down, left, right, front, back … …) in the embodiments of the present application are merely used to explain the relative positional relationship, movement, etc. between the components in a particular gesture (as shown in the drawings), and if the particular gesture changes, the directional indication changes accordingly. Furthermore, the terms "comprise" and "have," as well as any variations thereof, are intended to cover a non-exclusive inclusion. For example, a process, method, system, article, or apparatus that comprises a list of steps or elements is not limited to only those listed steps or elements but may include other steps or elements not listed or inherent to such process, method, article, or apparatus.
Reference herein to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment may be included in at least one embodiment of the present application. The appearances of such phrases in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. Those of skill in the art will explicitly and implicitly appreciate that the embodiments described herein may be combined with other embodiments.
As shown in fig. 1, the present embodiment provides an embodiment of a method for calculating interpolation of anemometer tower data, in the present embodiment, the anemometer tower data is derived from at least one anemometer tower and at least one reference tower located in a preset area, and the method for calculating interpolation of anemometer tower data includes the following steps:
step S1, first wind speed data of a wind measuring tower based on a first preset natural time length and second wind speed data of a reference tower based on the preset natural time length are obtained.
Preferably, the first wind speed data of the anemometer tower and the second wind speed data of the reference tower can be directly detected and obtained, or directly obtained through a public channel.
Step S2, obtaining the data defect point in the first wind speed data.
And S3, intercepting the first wind speed data by taking the data missing point as a midpoint and taking the second preset natural time length as an intercepting length to form at least one sample to be interpolated, wherein the second preset natural time length is smaller than the first preset natural time length.
Preferably, the first preset natural time period may be set to one natural year, and the second preset natural time period may be set to one of one natural quarter, one natural month, one natural week, one natural day, and one natural hour.
And S4, intercepting second wind speed data based on a second preset natural time length to form at least one training sample, wherein the training sample does not contain missing points.
Preferably, since the wind speed and the wind direction have a certain rule, the sample to be interpolated and the training sample can be intercepted based on the same time stamp, so that the sample to be interpolated and the training sample are located under the same wind rule.
Preferably, the general rule of wind direction throughout the year is: the spring is mainly used for east wind, the summer is mainly used for southeast wind, the autumn is mainly used for west wind, and the winter is mainly used for northwest wind. Under normal conditions, wind always blows from a high air pressure region to a low air pressure region. The greater the differential air pressure, the greater the wind speed. Because of different seasons and different solar radiation energy, the earth surface is heated unevenly and the temperature is different. In winter, the northern near-stratum is air-cooled, high in density and high in air pressure, so that the northern wind is blown more in winter. In summer, the air in the near-south stratum is hot, low in density and low in air pressure, so that the south wind is blown more in summer. The wind direction forms a seasonal wind according to the change rule of seasons. In the eastern region of china, seasonal winds blow from the ocean to the inland, forming warm and moist summer monuments and cold winter monuments. In the southern areas of china, seasonal winds blow from inland to the ocean, forming dry summer and cold winter seasons.
And S5, defining an extreme learning model according to the data dimension of the training sample and training the extreme learning model through the training sample to obtain an interpolation model.
Preferably, the extreme learning machine is used for training a single hidden layer feedforward neural network, and unlike a traditional single hidden layer feedforward neural network training algorithm, the extreme learning machine randomly selects an input layer weight and a hidden layer bias, and an output layer weight is calculated and resolved according to a Moire-Penrose Moore-Penrose (MP) generalized inverse matrix theory by minimizing a loss function formed by a training error term and a regular term of an output layer weight norm. And even if the extreme learning machine randomly generates hidden layer nodes, the extreme learning machine still maintains the general approximation capability of the single hidden layer feedforward neural network.
And S6, outputting the sample to be interpolated to an interpolation model to acquire wind speed interpolation at the data missing point.
Preferably, one sample to be interpolated may contain only one data defect, which easily results in an excessive calculation effort during the actual simulation operation.
Further, the step S2 specifically includes the following steps:
in step S21, a rectangular coordinate system is generated with the natural time on the horizontal axis and the wind speed value on the vertical axis.
Step S22, outputting the first wind speed data to a rectangular coordinate system to generate a wind speed curve.
Step S23, obtaining a zero point of the wind speed curve, and obtaining a first wind speed value of a unit natural time before the zero point and a second wind speed value of a unit natural time after the zero point.
Step S24, judging whether the first wind speed value and the second wind speed value are both different from zero, and if the first wind speed value and the second wind speed value are both different from zero, executing step S25.
Step S25, a first slope and a second slope of the wind speed curve at the first wind speed value and the second wind speed value are obtained.
Step S26, judging whether the sum of the absolute value of the first slope and the absolute value of the second slope is larger than or equal to a preset threshold, if so, executing step S27.
Preferably, the preset threshold value can be set according to the wind law of the region where the anemometer tower is located. For example, when the wind law at a specific time is that the wind speed is 5m/s and the wind direction is 30 degrees in the southeast direction, the preset threshold value of the area is set to be 8, and when the wind tower data at the time is zero, the wind tower data at the time is recognized to have zero value points, the wind speed data of the time in the previous hour and the next hour are acquired, if the wind speed data of the time in the previous hour and the next hour are respectively 5.5m/s, the wind direction is 28 degrees in the southeast direction and 4.5m/s, and the wind direction is 31 degrees in the southeast direction, the absolute value of the first slope (-5.5) and the absolute value of the second slope (4.5) at the time are respectively 5.5 and 4.5 (the transverse axis of the rectangular coordinate system is in the unit of hours and the vertical axis is in the unit of 1 m/s), the sum of the absolute values is 10, and is greater than 8, and the wind speed data at the time is judged to be missing if the judgment condition is met.
Preferably, the preset threshold is scaled according to the scale of the rectangular coordinate system, if the horizontal axis of the rectangular coordinate system is in hours and the vertical axis is in 5m/s, the density of the vertical axis is multiplied by 5, and at this time, the amplitude of the graph is reduced to one fifth, the slope is also reduced to one fifth, that is, the first slope-5.5 and the second slope-4.5 are respectively-1.1 and 0.9, and the preset threshold is also reduced to one fifth, that is, 1.6. At this point 1.1 plus 0.9 is still greater than 1.6.
In step S27, it is determined that the zero point is a data missing point.
Further, after step S22, the method further includes the following steps:
and S100, outputting a wind speed curve to an external visual terminal.
Further, the step S5 specifically includes the following steps:
step S51, defining training samples asThe number of training samples is +.>。
Wherein,，/>based on +.>Input sample data of a unit natural time length; />，/>For the second wind speed data->Training sample data for each unit of natural time length.
Step S52, defining an extreme learning model according to equation (1):
（1）；
wherein,，/>a first connection weight between an implicit layer and an output layer of the extreme learning model; />The number of nodes is an implicit layer; />As an output value of the hidden layer,；/>；/>is an activation function; />Input layer and hidden layer of extreme learning model +.>A second connection weight for the individual node; />Is the implicit layer->Threshold of individual nodes.
Preferably, the method comprises the steps of,may be set as an activation function such as Sigmoid function, gaussian function, etc.
Preferably, if the Sigmoid function is selected, thenCan be expressed as:
here +.>。
Step S53, randomly generating the second connection weight and the threshold value through a nonlinear mapping function in a feature space based on the extreme learning model.
And S54, calculating to obtain an output value of the hidden layer through the second connection weight value and the threshold value which are randomly generated.
Step S55, solving an optimal solution of the first connection weight by minimizing the approximate square difference according to equation (2):
（2）。
wherein,an output matrix that is an hidden layer; />To train the target matrix of the sample.
And step S56, obtaining an optimal solution, a second connection weight and a threshold corresponding to the optimal solution, and substituting the second connection weight and the threshold into the extreme learning model to obtain an initial interpolation model.
Further, the optimal solution in step S55 is characterized by formula (3):
（3）。
wherein,output matrix for hidden layer->Is a mole-penrose generalized inverse matrix.
Further, the above-mentioned Moire-Penrose generalized inverse matrixCharacterized by formula (4):
（4）。
wherein,output matrix for hidden layer->Is a transposed matrix of (a).
Preferably, the Moire-Penrose generalized inverse matrix can be calculated by singular value decomposition。
It can be understood that the training of the single hidden layer feedforward neural network by the extreme learning machine is divided into two main stages:
(1) random feature mapping:
the hidden layer parameters are randomly initialized and then the input data is mapped to a new feature space (called the extreme learning machine feature space) using some nonlinear mapping as an activation function. In short, the weights and the deviations on hidden nodes of the extreme learning machine are randomly generated. The random feature mapping phase differs from many existing learning algorithms, such as support vector machines that use kernel functions for feature mapping, the use of a constraint boltzmann machine in deep neural networks, automatic encoders/automatic decoders for feature learning, etc. The nonlinear mapping function in the extreme learning machine may be any nonlinear piecewise continuous function. In an extreme learning machine, hidden layer node parameters are randomly generated according to any continuous probability distribution, irrespective of training data.
(2) And solving linear parameters.
After the first stage, the second connection weight and the threshold are randomly generated. In the second stage, only the first connection weight is required to be solved. In order to obtain a first connection weight with good effect on the training sample set, it is necessary to ensure that its training error is minimal, by +.>Target matrix with training samples->And solving the minimum square error as an objective function for evaluating the training error, and solving the solution with the minimum objective function as the optimal solution.
Preferably, the feature space is a vector space that represents a particular property of a matrix. The feature space is a vector space made up of all feature vectors of the matrix. For example, for an n×n matrix a, if there is a scalar λ and a non-zero vector x such that ax=λx, then x is the eigenvector of a and the vector space in which vector x is located is the eigenvector of a.
Further, the step S6 further includes the following steps:
step S10, outputting wind speed interpolation to the first wind speed data to form interpolated wind speed data based on the anemometer tower.
And step S20, outputting the interpolated wind speed data to an external receiving end.
Preferably, all the above steps are realized by MATLAB.
According to the embodiment, first wind speed data of the wind measuring tower based on the first preset natural time length and second wind speed data of the reference tower based on the preset natural time length are obtained; acquiring a data missing point in the first wind speed data; intercepting first wind speed data by taking the data missing point as a midpoint and taking a second preset natural time length as an intercepting length to form at least one sample to be interpolated, wherein the second preset natural time length is smaller than the first preset natural time length; intercepting second wind speed data based on a second preset natural time length to form at least one training sample, wherein the training sample does not contain missing points; defining an extreme learning model according to the data dimension of the training sample, and training the extreme learning model through the training sample to obtain an interpolation model; and outputting the sample to be interpolated to an interpolation model to obtain the wind speed interpolation at the data defect point. According to the embodiment, the data missing points are taken as the middle points, the data missing points are intercepted to serve as samples to be interpolated, the data of the reference tower are intercepted to serve as training samples in the same time dimension, the training is carried out through the extreme learning machine, and finally the wind speed interpolation corresponding to the data missing points is obtained through the trained model. Compared with the traditional neural network learning algorithm, the method and the device have the advantages that a large number of network training parameters are not required to be set manually, the possibility of generating a local optimal solution is avoided to a large extent, the hidden layer node number of the extreme learning model is only required to be set, the connection weight and the threshold value of the model are not required to be adjusted in the algorithm execution process, and the method and the device have the advantages of high learning speed and good generalization performance.
As shown in fig. 2, this embodiment provides an embodiment of an interpolation calculation device for wind measurement tower data, where the interpolation calculation device for wind measurement tower data is applied to the interpolation calculation method for wind measurement tower data in the above embodiment, and the device includes a wind speed data acquisition module 1, a data missing point acquisition module 2, a sample interception module 3 to be interpolated, a training sample interception module 4, an interpolation model training module 5, and a wind speed interpolation acquisition module 6 that are electrically connected in sequence.
The wind speed data acquisition module 1 is used for acquiring first wind speed data of a wind measuring tower based on a first preset natural time length and second wind speed data of a reference tower based on the preset natural time length; the data missing point acquisition module 2 is used for acquiring data missing points in the first wind speed data; the sample to be interpolated intercepting module 3 is configured to intercept the first wind speed data with the data missing point as a midpoint and with a second preset natural time length as an intercepting length, so as to form at least one sample to be interpolated, where the second preset natural time length is smaller than the first preset natural time length; the training sample intercepting module 4 is used for intercepting second wind speed data based on a second preset natural time length to form at least one training sample, wherein the training sample does not contain missing points; the interpolation model training module 5 is used for defining an extreme learning model according to the data dimension of the training sample and training the extreme learning model through the training sample to obtain an interpolation model; the wind speed interpolation acquisition module 6 is used for outputting a sample to be interpolated to the interpolation model so as to acquire wind speed interpolation at the data missing point.
Further, the data missing point obtaining module comprises a first data missing point obtaining sub-module, a second data missing point obtaining sub-module, a third data missing point obtaining sub-module, a fourth data missing point obtaining sub-module, a fifth data missing point obtaining sub-module, a sixth data missing point obtaining sub-module and a seventh data missing point obtaining sub-module which are electrically connected in sequence; the first data missing point acquisition sub-module is electrically connected with the wind speed data acquisition module, and the seventh data missing point acquisition sub-module is electrically connected with the sample interception module to be interpolated.
The first data missing point acquisition submodule is used for generating a rectangular coordinate system by taking natural time as a horizontal axis and taking a wind speed value as a vertical axis; the second data missing point acquisition submodule is used for outputting the first wind speed data to a rectangular coordinate system so as to generate a wind speed curve; the third data missing point obtaining submodule is used for obtaining zero points of the wind speed curve and obtaining a first wind speed value of a unit of natural time before the zero points and a second wind speed value of a unit of natural time after the zero points; the fourth data missing point acquisition submodule is used for judging whether the first wind speed value and the second wind speed value are both different from zero; the fifth data missing point obtaining submodule is used for obtaining a first slope and a second slope of the wind speed curve respectively located at the first wind speed value and the second wind speed value if the first wind speed value and the second wind speed value are not zero; the sixth data missing point obtaining submodule is used for judging whether the sum of the absolute value of the first slope and the absolute value of the second slope is larger than or equal to a preset threshold value; the seventh data missing point obtaining submodule is used for judging that the zero point is a data missing point if the sum of the absolute value of the first slope and the absolute value of the second slope is larger than or equal to a preset threshold value.
Further, the interpolation calculation device of the anemometer tower data further comprises a wind speed curve output module electrically connected to the second data missing point acquisition sub-module, and the wind speed curve output module is used for outputting a wind speed curve to an external visual terminal.
Further, the interpolation model training module comprises a first interpolation model training sub-module, a second interpolation model training sub-module, a third interpolation model training sub-module, a fourth interpolation model training sub-module, a fifth interpolation model training sub-module and a sixth interpolation model training sub-module which are electrically connected in sequence; the first interpolation model training submodule is electrically connected with the training sample interception module, and the sixth interpolation model training submodule is electrically connected with the wind speed interpolation acquisition module.
Wherein the first interpolation model training submodule is used for defining training samples asThe number of training samples is。
Wherein,，/>based on +.>Input sample data of a unit natural time length; />，/>For the second wind speed data->Training sample data for each unit of natural time length.
The second interpolation model training sub-module is for defining an extreme learning model according to equation (1):
（1）；
wherein,，/>a first connection weight between an implicit layer and an output layer of the extreme learning model; />The number of nodes is an implicit layer; />As an output value of the hidden layer,；/>；/>is an activation function; />Input layer and hidden layer of extreme learning model +.>A second connection weight for the individual node; />Is the implicit layer->Threshold of individual nodes.
The third interpolation model training sub-module is used for randomly generating the second connection weight and the threshold value through a nonlinear mapping function in a feature space based on the extreme learning model.
And the fourth interpolation model training sub-module is used for calculating the output value of the hidden layer through the second connection weight value and the threshold value which are randomly generated.
The fifth interpolation model training submodule is used for solving an optimal solution of the first connection weight by minimizing an approximate square difference according to the equation (2):
（2）。
wherein,an output matrix that is an hidden layer; />To train the target matrix of the sample.
The sixth interpolation model training sub-module is used for obtaining an optimal solution, a second connection weight and a threshold value corresponding to the optimal solution, and substituting the second connection weight and the threshold value into the extreme learning model to obtain an initial interpolation model.
Further, the fifth interpolation model training sub-module is further equipped with a formula (3) for representing an optimal solution:
（3）。
wherein,output matrix for hidden layer->Is a mole-penrose generalized inverse matrix.
Further, the fifth interpolation model training sub-module is further equipped with a matrix for representing the above-mentioned Moire-Penrose generalized inverse matrixFormula (4):
（4）。
wherein,output matrix for hidden layer->Is a transposed matrix of (a).
Further, the interpolation calculation device of the anemometer tower data also comprises a wind speed interpolation output module and an interpolated wind speed data output module which are electrically connected in sequence; the wind speed interpolation output module is electrically connected with the wind speed interpolation acquisition module.
The wind speed interpolation output module is used for outputting wind speed interpolation to first wind speed data so as to form interpolated wind speed data based on a anemometer tower; the post-interpolation wind speed data output module is used for outputting post-interpolation wind speed data to an external receiving end.
It should be noted that, the embodiment is an apparatus embodiment based on the foregoing method embodiment, and the preferred, expanded, limited, exemplified, and principle description portions of the embodiment may refer to the foregoing embodiments, which are not repeated herein.
According to the embodiment, first wind speed data of the wind measuring tower based on the first preset natural time length and second wind speed data of the reference tower based on the preset natural time length are obtained; acquiring a data missing point in the first wind speed data; intercepting first wind speed data by taking the data missing point as a midpoint and taking a second preset natural time length as an intercepting length to form at least one sample to be interpolated, wherein the second preset natural time length is smaller than the first preset natural time length; intercepting second wind speed data based on a second preset natural time length to form at least one training sample, wherein the training sample does not contain missing points; defining an extreme learning model according to the data dimension of the training sample, and training the extreme learning model through the training sample to obtain an interpolation model; and outputting the sample to be interpolated to an interpolation model to obtain the wind speed interpolation at the data defect point. According to the embodiment, the data missing points are taken as the middle points, the data missing points are intercepted to serve as samples to be interpolated, the data of the reference tower are intercepted to serve as training samples in the same time dimension, the training is carried out through the extreme learning machine, and finally the wind speed interpolation corresponding to the data missing points is obtained through the trained model. Compared with the traditional neural network learning algorithm, the method and the device have the advantages that a large number of network training parameters are not required to be set manually, the possibility of generating a local optimal solution is avoided to a large extent, the hidden layer node number of the extreme learning model is only required to be set, the connection weight and the threshold value of the model are not required to be adjusted in the algorithm execution process, and the method and the device have the advantages of high learning speed and good generalization performance.
As shown in fig. 3, the present embodiment provides an embodiment of the electronic device, and in the present embodiment, the electronic device 7 includes a processor 71 and a memory 72 coupled to the processor 71.
The memory 72 stores program instructions for implementing the method of interpolation calculation of anemometer tower data according to any of the embodiments described above.
The processor 71 is configured to execute program instructions stored in the memory 72 to perform interpolation calculations of the anemometer tower data.
The processor 71 may also be referred to as a CPU (Central Processing Unit ). The processor 71 may be an integrated circuit chip with signal processing capabilities. Processor 71 may also be a general purpose processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
Further, fig. 4 is a schematic structural diagram of a storage medium according to an embodiment of the present application, where the storage medium 8 of the embodiment of the present application stores a program instruction 81 capable of implementing all the methods described above, where the program instruction 81 may be stored in the storage medium in the form of a software product, and includes several instructions to cause a computer device (which may be a personal computer, a server, or a network device) or a processor (processor) to execute all or part of the steps of the methods described in various embodiments of the present application. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, an optical disk, or other various media capable of storing program codes, or a terminal device such as a computer, a server, a mobile phone, a tablet, or the like.
In the several embodiments provided in the present application, it should be understood that the disclosed systems, apparatuses, and methods may be implemented in other manners. For example, the apparatus embodiments described above are merely illustrative, e.g., the division of elements is merely a logical functional division, and there may be additional divisions of actual implementation, e.g., multiple elements or components may be combined or integrated into another system, or some features may be omitted, or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be an indirect coupling or communication connection via some interfaces, devices or units, which may be in electrical, mechanical or other forms.
In addition, each functional unit in each embodiment of the present application may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units. The foregoing is only the embodiments of the present application, and is not intended to limit the scope of the patent application, and all equivalent structures or equivalent processes using the contents of the specification and drawings of the present application or directly or indirectly applied to other related technical fields are included in the scope of the patent protection of the present application.
The embodiments of the present application have been described in detail above, but they are merely examples, and the present application is not limited to the above-described embodiments. It will be apparent to those skilled in the art that any equivalent modifications or substitutions for this invention are within the scope of this application, and therefore, such equivalent changes and modifications, improvements, etc. are intended to be within the scope of this application without departing from the spirit and principles of this application.
Claims (9)
1. An interpolation calculation method of anemometer tower data, wherein the anemometer tower data is derived from at least one anemometer tower and at least one reference tower which are located in a preset area, and the interpolation calculation method of the anemometer tower data comprises the following steps:
acquiring first wind speed data of the wind measuring tower based on a first preset natural time period and second wind speed data of the reference tower based on the preset natural time period;
acquiring a data missing point in the first wind speed data;
intercepting the first wind speed data by taking the data missing point as a midpoint and taking a second preset natural time length as an intercepting length to form at least one sample to be interpolated, wherein the second preset natural time length is smaller than the first preset natural time length;
intercepting the second wind speed data based on the second preset natural time length to form at least one training sample, wherein the training sample does not contain missing points;
defining an extreme learning model according to the data dimension of the training sample and training the extreme learning model through the training sample to obtain an interpolation model;
outputting the sample to be interpolated to the interpolation model to obtain wind speed interpolation at the data missing point;
acquiring the data defect point in the first wind speed data comprises the following steps:
generating a rectangular coordinate system by taking natural time as a horizontal axis and taking a wind speed value as a vertical axis;
outputting the first wind speed data to the rectangular coordinate system to generate a wind speed curve;
acquiring a zero point of the wind speed curve, and acquiring a first wind speed value of a unit natural time before the zero point and a second wind speed value of a unit natural time after the zero point;
judging whether the first wind speed value and the second wind speed value are both different from zero;
if yes, acquiring a first slope and a second slope of the wind speed curve, which are respectively positioned at the first wind speed value and the second wind speed value;
judging whether the sum of the absolute value of the first slope and the absolute value of the second slope is larger than or equal to a preset threshold value;
if yes, judging the zero point as the data missing point.
2. The method according to claim 1, wherein defining an extreme learning model based on the data dimension of the training sample and training the extreme learning model with the training sample to obtain an initial interpolation model, comprises:
defining the training samples asThe number of the training samples is +.>；
Wherein,，/>based on +.>Input sample data of a unit natural time length; />，/>For the second wind speed data +.>Training sample data of a unit natural time length;
defining the extreme learning model according to equation (1):
（1）；
wherein,，/>a first connection weight between an implicit layer and an output layer of the extreme learning model; />The number of the nodes of the hidden layer is the number; />For the output value of the hidden layer,；/>；/>is an activation function; />The input layer for the extreme learning model and the hidden layer +.>A second connection weight for the individual node;for the hidden layer->A threshold value for each node;
randomly generating the second connection weight and the threshold value in a feature space based on the extreme learning model through a nonlinear mapping function;
calculating to obtain an output value of the hidden layer through a second connection weight value and a threshold value which are randomly generated;
solving an optimal solution for the first connection weight by minimizing an approximate square difference according to equation (2):
（2）；
wherein,an output matrix for the hidden layer; />A target matrix for the training sample;
and obtaining the optimal solution, a second connection weight and a threshold corresponding to the optimal solution, and substituting the second connection weight and the threshold into the extreme learning model to obtain the initial interpolation model.
3. The method of interpolation computation of anemometer tower data according to claim 2, wherein the optimal solution is characterized by formula (3):
（3）；
wherein,an output matrix for said hidden layer +.>Is a mole-penrose generalized inverse matrix.
4. A method of interpolating wind tower data according to claim 3, wherein said moore-penrose generalized inverse matrixCharacterized by formula (4):
（4）；
wherein,an output matrix for said hidden layer +.>Is a transposed matrix of (a).
5. The interpolation calculation method of anemometer tower data according to claim 1, wherein outputting the sample to be interpolated to the interpolation model to obtain a wind speed interpolation at the data missing point, and thereafter, comprising:
outputting the wind speed interpolation to the first wind speed data to form interpolated wind speed data based on the anemometer tower;
and outputting the interpolated wind speed data to an external receiving end.
6. The method according to claim 1, wherein outputting the first wind speed data to the rectangular coordinate system to generate a wind speed profile, and then comprising:
and outputting the wind speed curve to an external visual terminal.
7. An interpolation calculation device for anemometer tower data, which is applied to the interpolation calculation method for anemometer tower data according to any one of claims 1 to 6, wherein the interpolation calculation device for anemometer tower data comprises:
the wind speed data acquisition module is used for acquiring first wind speed data of the wind measuring tower based on a first preset natural time length and second wind speed data of the reference tower based on the preset natural time length;
the data missing point acquisition module is used for acquiring the data missing point in the first wind speed data;
the sample to be interpolated intercepting module is used for intercepting the first wind speed data by taking the data missing point as a midpoint and taking a second preset natural time length as an intercepting length so as to form at least one sample to be interpolated, wherein the second preset natural time length is smaller than the first preset natural time length;
the training sample intercepting module is used for intercepting the second wind speed data based on the second preset natural time length to form at least one training sample, and the training sample does not contain missing points;
the interpolation model training module is used for defining an extreme learning model according to the data dimension of the training sample and training the extreme learning model through the training sample to obtain an interpolation model;
and the wind speed interpolation acquisition module is used for outputting the sample to be interpolated to the interpolation model so as to acquire wind speed interpolation at the data missing point.
8. An electronic device comprising a processor, and a memory coupled to the processor, the memory storing program instructions executable by the processor; the processor, when executing the program instructions stored in the memory, implements the method for interpolating wind tower data according to any one of claims 1 to 6.
9. A storage medium having stored therein program instructions which, when executed by a processor, implement a method of interpolation computation capable of implementing anemometer tower data according to any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311312131.9A CN117057257B (en) | 2023-10-11 | 2023-10-11 | Interpolation calculation method, device and equipment for anemometer tower data and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311312131.9A CN117057257B (en) | 2023-10-11 | 2023-10-11 | Interpolation calculation method, device and equipment for anemometer tower data and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117057257A CN117057257A (en) | 2023-11-14 |
| CN117057257B true CN117057257B (en) | 2024-01-26 |
Family
ID=88669601
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311312131.9A Active CN117057257B (en) | 2023-10-11 | 2023-10-11 | Interpolation calculation method, device and equipment for anemometer tower data and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117057257B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6024449A (en) * | 1998-07-13 | 2000-02-15 | Smith; Robert F. | High speed topography measurement of semi-diffuse objects |
| CN109255477A (en) * | 2018-08-24 | 2019-01-22 | 国电联合动力技术有限公司 | A kind of wind speed forecasting method and its system and unit based on depth limit learning machine |
| CN110135618A (en) * | 2019-04-01 | 2019-08-16 | 北京观详光电技术有限公司 | A kind of Wind outline data predication method |
| CN110909810A (en) * | 2019-11-26 | 2020-03-24 | 国网山东省电力公司电力科学研究院 | Renewable energy short-term prediction method based on data mining and variational modal decomposition |
| CN112832959A (en) * | 2021-03-31 | 2021-05-25 | 北京金风慧能技术有限公司 | Method and system for monitoring foundation safety condition of wind generating set |
| CN113449464A (en) * | 2021-06-11 | 2021-09-28 | 淮阴工学院 | Wind power prediction method based on improved depth extreme learning machine |
| CN114781709A (en) * | 2022-04-11 | 2022-07-22 | 上海电气风电集团股份有限公司 | Wind measurement data interpolation method and device and computer readable storage medium |
| WO2023179593A1 (en) * | 2022-03-23 | 2023-09-28 | 华为技术有限公司 | Data processing method and device |
-
2023
- 2023-10-11 CN CN202311312131.9A patent/CN117057257B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6024449A (en) * | 1998-07-13 | 2000-02-15 | Smith; Robert F. | High speed topography measurement of semi-diffuse objects |
| CN109255477A (en) * | 2018-08-24 | 2019-01-22 | 国电联合动力技术有限公司 | A kind of wind speed forecasting method and its system and unit based on depth limit learning machine |
| CN110135618A (en) * | 2019-04-01 | 2019-08-16 | 北京观详光电技术有限公司 | A kind of Wind outline data predication method |
| CN110909810A (en) * | 2019-11-26 | 2020-03-24 | 国网山东省电力公司电力科学研究院 | Renewable energy short-term prediction method based on data mining and variational modal decomposition |
| CN112832959A (en) * | 2021-03-31 | 2021-05-25 | 北京金风慧能技术有限公司 | Method and system for monitoring foundation safety condition of wind generating set |
| CN113449464A (en) * | 2021-06-11 | 2021-09-28 | 淮阴工学院 | Wind power prediction method based on improved depth extreme learning machine |
| WO2023179593A1 (en) * | 2022-03-23 | 2023-09-28 | 华为技术有限公司 | Data processing method and device |
| CN114781709A (en) * | 2022-04-11 | 2022-07-22 | 上海电气风电集团股份有限公司 | Wind measurement data interpolation method and device and computer readable storage medium |
Non-Patent Citations (3)
| Title |
|---|
| 基于加权正则极限学习机的短期风速预测;袁;戚佳金;王文霞;黄南天;;水电能源科学;第35卷(第05期);208-211 * |
| 基于极限学习机的脉动风速快速预测方法;李春祥;迟恩楠;李正农;;上海交通大学学报;第50卷(第11期);1719-1723 * |
| 极限学习机与支持向量机在储层渗透率预测中的对比研究;潘华贤;程国建;蔡磊;;计算机工程与科学;第32卷(第02期);131-134 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117057257A (en) | 2023-11-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Hamid et al. | Solar cell parameters identification using hybrid Nelder-Mead and modified particle swarm optimization | |
| Olauson et al. | Modelling the Swedish wind power production using MERRA reanalysis data | |
| Saeed et al. | An optimal approach of wind power assessment using Chebyshev metric for determining the Weibull distribution parameters | |
| Emeksiz et al. | In case study: Investigation of tower shadow disturbance and wind shear variations effects on energy production, wind speed and power characteristics | |
| JP2017187371A (en) | Weather prediction device and wind power generation | |
| CN104901309B (en) | Electric power system static security assessment method considering wind speed correlation | |
| Singh et al. | Wind power estimation using artificial neural network | |
| Doubrawa et al. | Benchmarks for model validation based on lidar wake measurements | |
| CN112100921A (en) | Method for acquiring wind resource and wind speed based on WRF and random forest | |
| Rasool et al. | Quantifying the reduction in power variability of co-located offshore wind-wave farms | |
| CN111311021A (en) | Theoretical power prediction method, device, equipment and storage medium for wind power plant | |
| Phan et al. | Short-term Solar Power Forecasting Using XGBoost with Numerical Weather Prediction | |
| Yazdan et al. | Forecasting Energy Consumption Time Series Using Recurrent Neural Network in Tensorflow | |
| Nasery et al. | Yaw-adjusted wind power curve modeling: A local regression approach | |
| CN110020801B (en) | Method for evaluating influence loss of wind power plant by external wake flow | |
| Jiang et al. | Time series analysis and forecasting for wind speeds using support vector regression coupled with artificial intelligent algorithms | |
| Guo et al. | Wind power assessment based on a WRF wind simulation with developed power curve modeling methods | |
| CN107657116A (en) | A kind of method of the affine modeling of wind power curve | |
| CN117057257B (en) | Interpolation calculation method, device and equipment for anemometer tower data and storage medium | |
| CN107947201B (en) | Method for judging stability of small disturbance of electric power system caused by wind power fluctuation | |
| CN114744623B (en) | New energy power generation prediction method and system based on deep learning | |
| Chabaud | Synthetic turbulence modelling for offshore wind farm engineering models using coherence aggregation | |
| Aimeur et al. | Short-term solar irradiance forecasting and photovoltaic system management using octonion neural networks | |
| Chen et al. | Probabilistic load flow calculation method based on polynomial normal transformation and extended Latin Hypercube | |
| Penfield | Ultra Long-Term Wind Farm Generation Forecast by Combining Numerical Weather Prediction with Gated Recurrent Units |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |