CN113323818B - Yaw error measurement method and device for multi-type fans - Google Patents
Yaw error measurement method and device for multi-type fans Download PDFInfo
- Publication number
- CN113323818B CN113323818B CN202110647957.5A CN202110647957A CN113323818B CN 113323818 B CN113323818 B CN 113323818B CN 202110647957 A CN202110647957 A CN 202110647957A CN 113323818 B CN113323818 B CN 113323818B
- Authority
- CN
- China
- Prior art keywords
- data
- equation
- wind speed
- formula
- density
- 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
- 238000000691 measurement method Methods 0.000 title claims description 16
- 238000000034 method Methods 0.000 claims abstract description 48
- 230000002159 abnormal effect Effects 0.000 claims abstract description 38
- 238000012545 processing Methods 0.000 claims abstract description 25
- 238000007781 pre-processing Methods 0.000 claims abstract description 8
- 238000004422 calculation algorithm Methods 0.000 claims description 18
- 238000010606 normalization Methods 0.000 claims description 15
- 238000005259 measurement Methods 0.000 claims description 13
- 238000013459 approach Methods 0.000 claims description 9
- 230000006870 function Effects 0.000 claims description 9
- 238000012549 training Methods 0.000 claims description 9
- 230000008569 process Effects 0.000 claims description 8
- 238000004140 cleaning Methods 0.000 claims description 5
- 238000004590 computer program Methods 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 claims description 3
- 238000012888 cubic function Methods 0.000 claims description 3
- 238000002955 isolation Methods 0.000 claims description 3
- 238000012935 Averaging Methods 0.000 claims 2
- 238000012546 transfer Methods 0.000 claims 1
- 230000008030 elimination Effects 0.000 abstract description 2
- 238000003379 elimination reaction Methods 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 10
- 230000003068 static effect Effects 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 3
- 230000005540 biological transmission Effects 0.000 description 2
- 238000011217 control strategy Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 235000014435 Mentha Nutrition 0.000 description 1
- 241001072983 Mentha Species 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- PCHJSUWPFVWCPO-UHFFFAOYSA-N gold Chemical compound [Au] PCHJSUWPFVWCPO-UHFFFAOYSA-N 0.000 description 1
- 239000010931 gold Substances 0.000 description 1
- 229910052737 gold Inorganic materials 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 235000014569 mints Nutrition 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000010248 power generation Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Classifications
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F03—MACHINES OR ENGINES FOR LIQUIDS; WIND, SPRING, OR WEIGHT MOTORS; PRODUCING MECHANICAL POWER OR A REACTIVE PROPULSIVE THRUST, NOT OTHERWISE PROVIDED FOR
- F03D—WIND MOTORS
- F03D17/00—Monitoring or testing of wind motors, e.g. diagnostics
Landscapes
- Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Sustainable Development (AREA)
- Sustainable Energy (AREA)
- Chemical & Material Sciences (AREA)
- Combustion & Propulsion (AREA)
- Mechanical Engineering (AREA)
- General Engineering & Computer Science (AREA)
- Wind Motors (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a yaw error measuring method and device for a multi-type fan, wherein the method comprises the following steps: acquiring multi-type fan data, and preprocessing the multi-type fan data; performing interval processing on the preprocessed multi-type fan data, performing polynomial fitting on wind speed-power scattered points of each interval by adopting a cubic spline method, selecting optimal wind speed-power curves of various types of fans, and acquiring yaw errors of the various types of fans according to the optimal wind speed-power curves. The method and the device can realize the simultaneous elimination of discrete type abnormal data and stacked type abnormal data in the data set.
Description
Technical Field
The invention relates to the technical field of computers, in particular to a yaw error measuring method and device for a multi-type fan.
Background
In the prior art, wind energy is used as the second most renewable energy source after water energy, the installed capacity space in the wind energy market is huge, the input of wind power is added in various countries, and the wind energy generator has the advantages of low cost, flexible application form and simple maintenance, has a mature technology and is widely researched and applied worldwide. According to statistics of the global wind power union in 2019, although the eight companies in China enter the front fifteen strong world fan manufacturers, the first two companies are occupied by Weist Tas and Siemens GeMean of European companies, and the gap between gold wind science and technology and distant energy sources of the fan manufacturers in China and other foreign manufacturers is needed to be considered. The current direction of solving the problem is to analyze the operation data of the multi-type fans by utilizing a big data technology to obtain fan manufacturers with excellent yaw performance. Other fans are guided by learning the yaw control strategy of the fan of the type, static deviation of a yaw system of each fan is corrected, and yaw performance of the unit is optimized.
Research is conducted on static deviation of a yaw control system, and in the normal operation power generation process of the wind turbine, a series of reasons such as vortex generated by rotation of a wind wheel or inaccurate installation of a anemoscope can possibly cause inaccuracy of wind speed and wind direction measurement, and accuracy of input data of a yaw controller is affected, so that yaw static errors are generated. In the prior art, an improved DBSCAN clustering algorithm is adopted to remove abnormal data, and a double-harmonic spline interpolation method and the like are utilized to correct the anemoscope, so that the generating capacity of the wind turbine is improved. However, the problem of inaccurate wind measurement by the wind vane still exists in the prior art.
Disclosure of Invention
The invention aims to provide a yaw error measuring method and device for a multi-type fan, and aims to solve the problems in the prior art.
The invention provides a yaw error measuring method of a multi-type fan, which comprises the following steps:
acquiring multi-type fan data, and preprocessing the multi-type fan data;
performing interval processing on the preprocessed multi-type fan data, performing polynomial fitting on wind speed-power scattered points of each interval by adopting a cubic spline method, selecting optimal wind speed-power curves of various types of fans, and acquiring yaw errors of the various types of fans according to the optimal wind speed-power curves.
The invention provides a yaw error measuring device of a multi-type fan, which comprises:
the acquisition module is used for acquiring the data of the multi-type fans and preprocessing the data of the multi-type fans;
the processing module is used for carrying out interval processing on the preprocessed multi-type fan data, carrying out polynomial fitting on wind speed-power scattered points of each interval by adopting a cubic spline method, selecting optimal wind speed-power curves of various types of fans, and acquiring yaw errors of various types of fans according to the optimal wind speed-power curves.
The embodiment of the invention also provides a yaw error measuring device of the multi-type fan, which comprises: the yaw error measurement method comprises the steps of a memory, a processor and a computer program stored in the memory and capable of running on the processor, wherein the computer program is executed by the processor to realize the yaw error measurement method of the multi-type fan.
The embodiment of the invention also provides a computer readable storage medium, wherein the computer readable storage medium stores an information transmission implementation program, and the program is executed by a processor to implement the steps of the yaw error measurement method of the multi-type fan.
By adopting the embodiment of the invention, the discrete type abnormal data and the stacked type abnormal data in the data set can be removed at the same time. In addition, the technical scheme of the embodiment of the invention is suitable for yaw error measurement of different types of units, and can realize comparison of yaw performances of different types of fans.
The foregoing description is only an overview of the present invention, and is intended to be implemented in accordance with the teachings of the present invention in order that the same may be more clearly understood and to make the same and other objects, features and advantages of the present invention more readily apparent.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are needed in the description of the embodiments or the prior art will be briefly described, and it is obvious that the drawings in the description below are some embodiments of the present invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of a yaw error measurement method of a multi-type wind turbine according to an embodiment of the present invention;
FIG. 2 is a detailed flow chart of a yaw error measurement method of a multi-type wind turbine according to an embodiment of the present invention;
FIG. 3a is a schematic diagram of the direct density reachability of the DBSCAN algorithm principle of an embodiment of the present invention;
FIG. 3b is a schematic diagram of density reachability of the DBSCAN algorithm principle of an embodiment of the present invention;
FIG. 3c is a schematic diagram of density-connected DBSCAN algorithm principles of an embodiment of the present invention;
FIG. 4 is a schematic diagram of outlier recognition results based on an isolated forest and DBSCAN according to an embodiment of the present invention;
FIG. 5 is a schematic diagram of a cubic spline fitted wind power curve of an embodiment of the present invention;
FIG. 6 is a schematic diagram of a class A fan compartmentalized power curve according to an embodiment of the present invention;
FIG. 7 is a schematic diagram of yaw performance comparisons for four types of fans according to an embodiment of the invention;
FIG. 8 is a schematic diagram of a yaw error measurement apparatus for a multi-type wind turbine according to a first embodiment of the present invention;
FIG. 9 is a schematic diagram of a yaw error measurement apparatus for a multi-type wind turbine according to a second embodiment of the present invention.
Detailed Description
Aiming at the problem that wind vane is inaccurate in wind measurement, the embodiment of the invention provides a yaw error measurement method of a multi-type fan. Firstly, cleaning abnormal data by adopting an isolated forest and DBSCAN clustering algorithm, carrying out normalization treatment, then carrying out interval treatment on the obtained data, and adopting a cubic spline method to fit wind speed-power scattered points of each interval, so as to select an optimal wind speed-power curve and obtain a yaw error. And comparing the optimal wind speed-power curves of the four types of fans to obtain the fan type with optimal yaw performance.
The technical solutions of the present invention will be clearly and completely described in connection with the embodiments, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", etc. indicate orientations or positional relationships based on the orientations or positional relationships shown in the drawings are merely for convenience in describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be configured and operated in a specific orientation, and thus should not be construed as limiting the present invention.
Furthermore, the terms "first," "second," and the like, are used for descriptive purposes only and are not to be construed as indicating or implying a relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defining "a first" or "a second" may explicitly or implicitly include one or more of the described features. In the description of the present invention, the meaning of "a plurality" is two or more, unless explicitly defined otherwise. Furthermore, the terms "mounted," "connected," "coupled," and "connected" are to be construed broadly, and may be, for example, fixedly connected, detachably connected, or integrally connected; can be mechanically or electrically connected; can be directly connected or indirectly connected through an intermediate medium, and can be communication between two elements. The specific meaning of the above terms in the present invention will be understood in specific cases by those of ordinary skill in the art.
Method embodiment
According to an embodiment of the present invention, there is provided a yaw error measurement method of a multi-type wind turbine, and fig. 1 is a flowchart of the yaw error measurement method of the multi-type wind turbine according to the embodiment of the present invention, as shown in fig. 1, the yaw error measurement method of the multi-type wind turbine according to the embodiment of the present invention specifically includes:
step 101, acquiring multi-type fan data, and preprocessing the multi-type fan data;
102, carrying out interval processing on the preprocessed multi-type fan data, carrying out polynomial fitting on wind speed-power scattered points of each interval by adopting a cubic spline method, selecting optimal wind speed-power curves of various types of fans, and acquiring yaw errors of the various types of fans according to the optimal wind speed-power curves.
In the embodiment of the invention, after yaw errors of various fans are obtained according to the optimal wind speed-power curve, the optimal wind speed-power curve of various fans can be compared to obtain the fan type with optimal yaw performance.
The above technical solutions of the embodiments of the present invention are described in detail below with reference to the accompanying drawings.
FIG. 2 is a detailed flow chart of a yaw error measurement method of a multi-type wind turbine according to an embodiment of the present invention, as shown in FIG. 2, specifically including the following processes:
s1: and acquiring data of multiple types of fans.
In the embodiment, the data of four types of fans with good states in a certain North China wind farm for half a year are selected, the time ranges from seconds to minutes, the smaller the time scale is, and the selected time range can be properly shortened. The data mainly comprises: time, angle to wind, wind speed, power, wind direction, etc. The data for the four types of fans are shown in table 1. The status codes of the four fans are acquired, and a foundation is laid for eliminating abnormal data points subsequently.
Table 1 four fan type data
S2: and (5) preprocessing data.
For the data acquired in the step S1, abnormal data need to be removed, and in order to compare the performances of different fans, normalization processing needs to be carried out on the data set.
S2.1: the isolated forest isolates the outliers for the characteristics of small outlier number and large difference between the outlier and the normal value. The main idea is to divide the data set equally until all data points are isolated, and judge whether the data points are abnormal points according to the score of each data point. The main process of isolated forests is as follows:
(1) An isolated forest is made up of t isolated trees, denoted as:
IF∈{t 1 ,...,t T } (1)
where T is the total number of orphaned trees.
(2) For each tree t, the number of iterations ht (x) required to isolate sample x can be calculated, and then the average number of steps required to isolate sample x in the forest is expressed as:
(3) The method can complete isolation of abnormal values by only a few steps. But the number of steps required to isolate the observations x is affected by the number of samples n. To eliminate this effect, a normalized outlier s (x, n) is defined:
wherein c (n) is represented as:
h (i) is a harmonic number, approximately equal to:
H(i)≈ln(i)+0.5772156649 (5)
it can be demonstrated that c (n) is the average number of steps required to separate one sample from the other n samples, which provides a normalization factor such that the s value is independent of the number of samples n.
(4) In the training process, a given training set is recursively divided, and training is stopped after abnormal data in a sample is isolated or reaches the height of a specified tree to generate a local model. The tree height h is approximately equal to the average tree height, and can be defined as:
h＝ceiling(log 2 ψ) (6)
where ψ represents the sub-sample size and CEILING represents the CEILING function. The probability that data points with path lengths shorter than the average path length fall in an abnormal region is greater, so that trees grow to the average tree height to effectively isolate abnormal data points.
The DBSCAN is used as an algorithm for processing abnormal data of the fan, density is used as a division basis, the number of clusters does not need to be determined in advance, but two important parameter cluster radius epsilon and a density threshold Mints in a radius area in the algorithm need to be set in advance, and if the set parameter values are not suitable for a data set, the clustering effect is not obvious. The setting of these two parameters is particularly important and remains unchanged throughout the operation. The basic idea is: and calculating the number of the neighborhood objects by taking epsilon as a radius, classifying the neighborhood objects into one type if the density threshold value is reached, and then communicating the high-density areas into clusters. Let the dataset be x= { p 1 ,p 2 ,L p n P is }, then i 、p j Distance between two points formula dist (p i ,p j ) The algorithm is defined as follows:
(1) Epsilon neighborhood: neighborhood N ε (p i ) The expression of (2) is N ε (p i )＝{p j ∈X|dist(p i ,p j ) ε is less than or equal to ε }, by p i Is the center of the circle and epsilon is the radius area.
(2) Core object: p is p i Neighborhood N ε (p i ) The number of midpoints is greater than or equal to MinPts, and is called p i Is a core object;
(3) The direct density can be achieved: if p j ∈N ε (p i ) And p is i Is the core point, then called p j Is from p i The direct density is achievable as shown in fig. 3 a.
(4) The density can be achieved: if the sequence { p } i ,p i+1 ,K,p j In } epsilon X, p i+1 From p i The direct density is up to the name p j Is from p i The density is achievable as shown in fig. 3 b.
(5) Density connection: point of presence p k E, X, if p i ，p j Can all be from p k The density is up to p i And p j Are density-connected as shown in figure 3 c.
S2.1.1: dividing data X screened by fan state codes into t groups, selecting current group data, randomly extracting psi samples, randomly selecting one sample as a node, randomly selecting a value in the characteristic value range, adopting a binary division method for the psi samples, and if the value of the extracted sample is smaller than the value, dividing the sample into a left branch, otherwise dividing the sample into a right branch.
S2.1.2: judging whether the current group data is inseparable or whether the height of the tree is larger than that of formula 6, and if the current group data does not meet the condition, returning to the step (1); if the condition is satisfied, a data score is calculated by equation 3, and when the score approaches 1, it is determined as abnormal. If the score approaches 0.5, it is not possible to determine whether the data is abnormal, and if the score approaches 0, it is determined that the data is normal.
S2.1.3: removing part of abnormal data by the step (2), inputting the rest data into a DBSCAN model, initializing parameters epsilon and MinPts, and selecting one data p from the obtained data set i And recording as read, and checking the number of neighborhoods of the marked data.
S2.1.4: if |N Eps (p i ) And if the I is more than or equal to MinPts, p is a core object, and then all data points connected with the density can be searched according to the density and marked and read.
S2.1.5: if |N is not satisfied Eps (p i ) And if the I is not less than or equal to MinPts, p is not a core object, the p point is marked as read, class label processing is not performed, data is not taken into consideration later, and when other data are subjected to density accessibility and density connection, processing is performed.
S2.1.6: and (3) repeating the steps (3) to (5), and treating the data sample without the class mark as abnormal data until all subsets are iterated.
The result of anomaly data identification using isolated forests and DBSCANs is shown in FIG. 4.
S2.2: data normalization
0-1 normalization, by quantizing the raw data to map it between [0,1], the calculation method is as follows:
x′ i as normalized value, x i X is the original data max And x min Is the maximum and minimum values before normalization.
S3: establishing a regression model;
in order to facilitate analysis of yaw static errors and fan performance of four fans of a yaw system, taking a class A fan as an example, by referring to technical documents of the class A fan, an existing control strategy is that yaw threshold time exceeds 160 seconds, and when a beta angle (an included angle between a nacelle and an average wind direction in the period of time) exceeds 8 degrees, the nacelle performs yaw, so alpha (-8, 8) intervals are divided into 16 intervals, and wind speed-power scattered points of each obtained interval are fitted.
And (3) performing polynomial fitting on each obtained wind speed-power scattered point to obtain 16 wind speed-power curves, and analyzing which curve has the largest power under certain wind speed, wherein the corresponding curve is the yaw error value.
Fitting the wind speed-power curve by adopting a cubic spline to construct a cubic function S k (x) In [ a, b ]]With continuous second derivative, selecting the second derivative value parameter S' on the node 3 (x i )＝M i (i=0, 1l, n), then the function S 3 (x i )＝y i (i=0, 1, l, n) at segment [ x i-1 ,x i ]The above needs are satisfied:
S 3 (x i-1 )＝y i-1 ,S 3 (x i )＝y i
S″ 3 (x i-1 )＝M i-1 ,S″ 3 (x i )＝M i (8)
interpolation using lagrangian has:
in the formula, h i ＝x i -x i-1 。
Twice the integration was performed on equation 9, and differentiation was performed to obtain:
will S 3 '(x i -0)＝S 3 '(x i +0) carry-over equation 10:
the above formula can be expressed as:
(1-a i )M i-1 +2M i +a i M i+1 ＝g i (i＝1,2,L n-1) (12)
the free boundary condition is the left boundary [ x ] at the origin 0 ,x 1 ]On, x is x 0 The time-introduced type 12 is as follows:
if S 3 ' As known from (0), x=x is taken from the formula 3.16 0 The method can obtain:
2M 0 +0×M 1 ＝2×S″ 3 (x 0 ) (14)
the boundary conditions for the origin are:
2M 0 +a 0 M 1 ＝g(0) (15)
wherein given y 0 When a is 0 ＝0,g 0 ＝2y″ 0 The method comprises the steps of carrying out a first treatment on the surface of the Given y' 0 In the time-course of which the first and second contact surfaces,
similarly, the boundary conditions of the endpoint are:
a n M n-1 +2M n ＝g n (16)
wherein given y n When a is n ＝0,g n ＝2y″ n The method comprises the steps of carrying out a first treatment on the surface of the Given y' n In the time-course of which the first and second contact surfaces,
the solution problem is constituted:
solving equation to obtain M 0 ,M 1 ,L M n The obtained power curve model is shown in fig. 6, wherein the obtained cubic spline fitting function formula is as follows:
f(x)＝0.3105x 3 +0.4633x 2 +0.1538x+0.009898 (18)
the wind power curve of the cubic spline fitting is shown in fig. 5.
S4: and (5) obtaining the static deviation of the yaw system.
And obtaining deviation values of the four types of fans based on fan data of four types of fans, namely an A type fan, a B type fan, a C type fan and a D type fan, in 2019. Taking a class A fan as an example, fitting the wind power scattered points in 16 intervals to obtain 16 wind speed-power curves, wherein the uppermost lines are [ -5, -4) and [ -4, -3) as shown in figure 6, so that the wind speed-power curves are taken as-4 degrees. And similarly, the yaw error of the fan of the type B is 7 degrees, the yaw error of the fan of the type C is 6 degrees, the yaw error of the fan of the type D is 6 degrees, and the yaw performance pairs of the four types of fans are shown in the figure 7.
In summary, by means of the technical scheme of the embodiment of the invention, for discrete abnormal data in the data set, the abnormal points can be well removed by adopting an isolated forest algorithm, and for stacked abnormal data points, the abnormal points can be well removed by adopting a DBSCAN clustering algorithm. The invention designs the combination recognition model of the isolated forest and the DBSCAN to process the abnormal data, and can realize the simultaneous elimination of discrete abnormal data and stacked abnormal data in the data set. In addition, the model of the embodiment of the invention is suitable for yaw error measurement of different types of units, and can realize comparison of yaw performances of different types of fans.
Device embodiment 1
According to an embodiment of the present invention, there is provided a yaw error measurement apparatus of a multi-type wind turbine, and fig. 8 is a schematic diagram of the yaw error measurement apparatus of the multi-type wind turbine according to the embodiment of the present invention, as shown in fig. 8, the yaw error measurement apparatus of the multi-type wind turbine according to the embodiment of the present invention specifically includes:
an acquisition module 80, configured to acquire multi-type fan data, and perform preprocessing on the multi-type fan data; the obtaining module 80 is specifically configured to: cleaning abnormal data by adopting an isolated forest and DBSCAN clustering algorithm, and carrying out normalization treatment:
the processing module 82 is configured to perform interval processing on the preprocessed multi-type fan data, perform polynomial fitting on wind speed-power scattered points of each interval by using a cubic spline method, select an optimal wind speed-power curve of each type of fan, and obtain yaw errors of each type of fan according to the optimal wind speed-power curve.
The above apparatus may further include:
and the comparison module is used for comparing the optimal wind speed-power curves of the fans of various types after the processing module acquires yaw errors of the fans of various types according to the optimal wind speed-power curves, so as to obtain the fan type with optimal yaw performance.
The embodiment of the present invention is an embodiment of a device corresponding to the embodiment of the method, and specific operations of each module may be understood by referring to descriptions of the embodiment of the method, which are not repeated herein.
Device example two
An embodiment of the present invention provides a yaw error measurement apparatus for a multi-type fan, as shown in fig. 9, including: memory 90, processor 92, and a computer program stored on the memory 90 and executable on the processor 92, which when executed by the processor 92, performs the steps as described in the method embodiments.
Device example III
Embodiments of the present invention provide a computer-readable storage medium having stored thereon a program for carrying out information transmission, which when executed by the processor 92, carries out the steps described in the method embodiments.
The computer readable storage medium of the present embodiment includes, but is not limited to: ROM, RAM, magnetic or optical disks, etc.
It should be noted that, in the present specification, the embodiment about the storage medium and the embodiment about the blockchain-based service providing method in the present specification are based on the same inventive concept, so the specific implementation of this embodiment may refer to the implementation of the yaw error measurement method of the corresponding multi-type fan, and the repetition is omitted.
The foregoing describes specific embodiments of the present disclosure. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims can be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments. In particular, for system embodiments, since they are substantially similar to method embodiments, the description is relatively simple, as relevant to see a section of the description of method embodiments.
The foregoing description is by way of example only and is not intended to limit the present disclosure. Various modifications and changes may occur to those skilled in the art. Any modifications, equivalent substitutions, improvements, etc. that fall within the spirit and principles of the present document are intended to be included within the scope of the claims of the present document.
Claims (9)
1. A yaw error measurement method for a multi-type wind turbine, comprising:
acquiring multi-type fan data, and preprocessing the multi-type fan data, wherein the method specifically comprises the following steps of:
cleaning abnormal data by adopting a combined recognition model of an isolated forest and a DBSCAN clustering algorithm, and carrying out normalization treatment;
performing interval processing on the preprocessed multi-type fan data, performing polynomial fitting on wind speed-power scattered points of each interval by adopting a cubic spline method, selecting optimal wind speed-power curves of various types of fans, and acquiring yaw errors of the various types of fans according to the optimal wind speed-power curves.
2. The method of claim 1, wherein after obtaining yaw errors for each type of wind turbine from the optimal wind speed-power curve, the method further comprises:
and comparing the optimal wind speed-power curves of the fans of the plurality of types to obtain the fan type with optimal yaw performance.
3. The method of claim 1, wherein the cleaning and normalizing the anomaly data using a combined recognition model of an isolated forest and DBSCAN clustering algorithm specifically comprises:
step 1, representing an isolated forest consisting of t isolated trees according to formula 1:
IF∈{t 1 ,...,t T equation 1;
wherein T is the total number of the isolated trees;
for each tree t, the number of iterations h required to isolate the observation x is calculated t (x) The average number of steps required to isolate sample x in an isolated forest is then expressed according to equation 2:
normalized outliers s (x, n) are defined according to equations 3 and 4:
where n is the number of samples, x is the isolation observation, H (i) is the harmonic number, H (i) +.ln (i) +0.5772156649, c (n) is the number of averaging steps required to separate one sample from the other n samples, which provides a normalization factor such that the s value is independent of the number of samples n
In the training process, a given training set is recursively divided, and training is stopped after abnormal data in a sample is isolated or reaches the height of a specified tree to generate a local model, and the tree height h is determined to be approximately equal to the average tree height according to a formula 5:
h＝ceiling(log 2 ψ) equation 5;
wherein, ψ represents the sub-sample size, and CEILING represents the CEILING function;
let the data set of the DBSCAN algorithm be x= { p 1 ,p 2 ,…,p n P is }, then i 、p j Distance between two points formula dist (p i ,p j ) The DBSCAN algorithm is defined as follows: neighborhood N ε (p i ) The expression of (2) is N ε (p i )＝{p j ∈X|dist(p i ,p j ) ε is less than or equal to ε }, by p i Is the center of a circle, epsilon is the area of radius, p i Neighborhood N ε (p i ) The number of midpoints is greater than or equal to MinPts, and is called p i As core object, if p j ∈N ε (p i ) And p is i Is the core point, then called p j Is from p i The direct density is reachable if the sequence { p } i ,p i+1 ,…,p j In } epsilon X, p i+1 From p i The direct density is up to the name p j Is from p i The density is up to, the existence point p k E, X, if p i ，p j Can all be from p k The density is up to p i And p j Are density-connected;
step 2, dividing the data X screened by the fan state code into t groups, selecting current group data, randomly extracting psi samples, randomly selecting one sample as a node, randomly selecting a value in the characteristic value range, adopting a binary division method for the psi samples, and if the value of the extracted sample is smaller than the value, dividing the sample into a left branch, otherwise dividing the sample into a right branch;
step 3, judging whether the current group data is inseparable or whether the height of the tree is larger than that of the formula 5, and if the current group data is not satisfied, returning to the step 2; if the condition is satisfied, calculating a data score by a formula 3, judging that the data is abnormal when the score approaches 1, judging that the data is abnormal when the score approaches 0.5, and judging that the data is normal when the score approaches 0;
step 4, removing part of abnormal data through the step 3, inputting the rest data into a DBSCAN model, initializing parameters epsilon and MinPts, and selecting one data p from the obtained data set i And recording as read, and checking the number of neighborhoods of the marked data;
step 5, if |N Eps (p i ) If the I is more than or equal to MinPts, p is a core object, and then all data points connected with the density can be searched according to the density, and marking is read;
step 6, if |N is not satisfied Eps (p i ) I is not less than MinPts, p is notThe core object marks the p point as read, does not perform class label processing, and does not take data into consideration later, and when other data are subjected to density accessibility and density connection, the processing is performed;
step 7, repeating the steps 4-6, and treating the data sample without class marks as abnormal data until all subsets are iterated;
step 8, 0-1 normalization is performed according to formula 6, and the original data is quantized to be mapped between [0,1], and the calculation method is as follows:
wherein x' i As normalized value, x i X is the original data max And x min Is the maximum and minimum values before normalization.
4. The method according to claim 1, wherein the steps of performing interval processing on the preprocessed multi-type fan data, performing polynomial fitting on wind speed-power scattered points of each interval by adopting a cubic spline method, selecting optimal wind speed-power curves of various types of fans, and acquiring yaw errors of various types of fans according to the optimal wind speed-power curves comprise:
performing interval processing on the preprocessed multi-type fan data, and fitting a wind speed-power curve by adopting a cubic spline to construct a cubic function S k (x) In [ a, b ]]With continuous second derivative, selecting the second derivative value parameter S' on the node " 3 (x i )＝M i (i=0, 1, …, n), then the function S 3 (x i )＝y i (i=0, 1, …, n) at segment [ x i-1 ,x i ]The above requirement satisfies equation 7:
S 3 (x i-1 )＝y i-1 ,S 3 (x i )＝y i
S” 3 (x i-1 )＝M i-1 ,S” 3 (x i )＝M i equation 7;
interpolation is performed using lagrangian according to equation 8:
wherein h is i ＝x i -x i-1 ；
According to equation 9, equation 8 is integrated twice and differentiated:
will S 3 '(x i -0)＝S 3 '(x i +0) into equation 9 to yield equation 10:
equation 10 is expressed as equation 12:
(1-a i )M i-1 +2M i +a i M i+1 ＝g i (i=1, 2, … n-1) equation 11;
let the free boundary condition be the left boundary at the origin [ x ] 0 ,x 1 ]On, x is x 0 When formula 11 is imported, formula 12 is obtained:
if S' 3 (0) Taking x=x is known 0 Equation 13 is obtained:
2M 0 +0×M 1 ＝2×S” 3 (x 0 ) Equation 13;
the boundary condition for determining the starting point according to equation 13 is equation 14:
2M 0 +a 0 M 1 =g (0) equation 14;
wherein given y' 0 When a is 0 ＝0,g 0 ＝2y” 0 The method comprises the steps of carrying out a first treatment on the surface of the Given y' 0 In the time-course of which the first and second contact surfaces,
similarly, the boundary condition for the endpoint is determined according to equation 15:
a n M n-1 +2M n ＝g n equation 15;
wherein given y' n When a is n ＝0,g n ＝2y” n The method comprises the steps of carrying out a first treatment on the surface of the Given y' n In the time-course of which the first and second contact surfaces,
the solution problem is constructed according to equation 16:
solving equation 16 to obtain M 0 ,M 1 ,…M n Obtaining a power curve model, wherein the obtained cubic spline fitting function formula is formula 17:
f(x)＝0.3105x 3 +0.4633x 2 +0.1538x+0.009898 equation 17.
5. A yaw error measurement apparatus for a multi-type wind turbine, comprising:
the acquisition module is used for acquiring the data of the multi-type fans, preprocessing the data of the multi-type fans and specifically used for:
cleaning abnormal data by adopting a combined recognition model of an isolated forest and a DBSCAN clustering algorithm, and carrying out normalization treatment:
the processing module is used for carrying out interval processing on the preprocessed multi-type fan data, carrying out polynomial fitting on wind speed-power scattered points of each interval by adopting a cubic spline method, selecting optimal wind speed-power curves of various types of fans, and acquiring yaw errors of various types of fans according to the optimal wind speed-power curves.
6. The apparatus of claim 5, wherein the apparatus further comprises:
and the comparison module is used for comparing the optimal wind speed-power curves of the fans of various types after the processing module acquires yaw errors of the fans of various types according to the optimal wind speed-power curves, so as to obtain the fan type with optimal yaw performance.
7. The apparatus of claim 5, wherein the acquisition module is specifically configured to:
step 1, representing an isolated forest consisting of t isolated trees according to formula 1:
IF∈{t 1 ,...,t T equation 1;
wherein T is the total number of the isolated trees;
for each tree t, the number of iterations h required to isolate the observation x is calculated t (x) The average number of steps required to isolate sample x in an isolated forest is then expressed according to equation 2:
normalized outliers s (x, n) are defined according to equations 3 and 4:
where n is the number of samples, x is the isolation observation, H (i) is the harmonic number, H (i) +.ln (i) +0.5772156649, c (n) is the number of averaging steps required to separate one sample from the other n samples, which provides a normalization factor such that the s value is independent of the number of samples n
In the training process, a given training set is recursively divided, and training is stopped after abnormal data in a sample is isolated or reaches the height of a specified tree to generate a local model, and the tree height h is determined to be approximately equal to the average tree height according to a formula 5:
h＝ceiling(log 2 ψ) equation 5;
wherein, ψ represents the sub-sample size, and CEILING represents the CEILING function;
let the data set of the DBSCAN algorithm be x= { p 1 ,p 2 ,…,p n P is }, then i 、p j Distance between two points formula dist (p i ,p j ) The DBSCAN algorithm is defined as follows: neighborhood N ε (p i ) The expression of (2) is N ε (p i )＝{p j ∈X|dist(p i ,p j ) ε is less than or equal to ε }, by p i Is the center of a circle, epsilon is the area of radius, p i Neighborhood N ε (p i ) The number of midpoints is greater than or equal to MinPts, and is called p i As core object, if p j ∈N ε (p i ) And p is i Is the core point, then called p j Is from p i The direct density is reachable if the sequence { p } i ,p i+1 ,…,p j In } epsilon X, p i+1 From p i The direct density is up to the name p j Is from p i The density is up to, the existence point p k E, X, if p i ，p j Can all be from p k The density is up to p i And p j Are density-connected;
step 2, dividing the data X screened by the fan state code into t groups, selecting current group data, randomly extracting psi samples, randomly selecting one sample as a node, randomly selecting a value in the characteristic value range, adopting a binary division method for the psi samples, and if the value of the extracted sample is smaller than the value, dividing the sample into a left branch, otherwise dividing the sample into a right branch;
step 3, judging whether the current group data is inseparable or whether the height of the tree is larger than that of the formula 5, and if the current group data is not satisfied, returning to the step 2; if the condition is satisfied, calculating a data score by a formula 3, judging that the data is abnormal when the score approaches 1, judging that the data is abnormal when the score approaches 0.5, and judging that the data is normal when the score approaches 0;
step 4, removing part of abnormal data through the step 3, inputting the rest data into a DBSCAN model, initializing parameters epsilon and MinPts, and selecting one data p from the obtained data set i And recording as read, and checking the number of neighborhoods of the marked data;
step 5, if |N Eps (p i ) If the I is more than or equal to MinPts, p is a core object, and then all data points connected with the density can be searched according to the density, and marking is read;
step 6, if |N is not satisfied Eps (p i ) If the I is not less than or equal to MinPts, p is not a core object, the p point is marked as read, class label processing is not performed, data is not considered later, and when other data are subjected to density accessibility and density connection, processing is performed;
step 7, repeating the steps 4-6, and treating the data sample without class marks as abnormal data until all subsets are iterated;
step 8, 0-1 normalization is performed according to formula 6, and the original data is quantized to be mapped between [0,1], and the calculation method is as follows:
wherein x' i As normalized value, x i X is the original data max And x min Maximum and minimum values before normalization;
the processing module is specifically configured to:
performing interval processing on the preprocessed multi-type fan data, and fitting a wind speed-power curve by adopting a cubic spline to construct a cubic function S k (x) In [ a, b ]]With continuous second derivative, selecting the second derivative value parameter S' on the node " 3 (x i )＝M i (i=0, 1, …, n), then the function S 3 (x i )＝y i (i=0, 1, …, n) at segment [ x i-1 ,x i ]The above requirement satisfies equation 7:
S 3 (x i-1 )＝y i-1 ,S 3 (x i )＝y i
S” 3 (x i-1 )＝M i-1 ,S” 3 (x i )＝M i equation 7;
interpolation is performed using lagrangian according to equation 8:
wherein h is i ＝x i -x i-1 ；
According to equation 9, equation 8 is integrated twice and differentiated:
will S' 3 (x i -0)＝S' 3 (x i +0) into equation 9 to yield equation 10:
equation 10 is expressed as equation 12:
(1-a i )M i-1 +2M i +a i M i+1 ＝g i (i=1, 2, … n-1) equation 11;
assuming free edgesThe boundary condition is the left boundary [ x ] at the start point 0 ,x 1 ]On, x is x 0 When formula 11 is imported, formula 12 is obtained:
if S' 3 (0) Taking x=x is known 0 Equation 13 is obtained:
2M 0 +0×M 1 ＝2×S” 3 (x 0 ) Equation 13;
the boundary condition for determining the starting point according to equation 13 is equation 14:
2M 0 +a 0 M 1 =g (0) equation 14;
wherein given y' 0 When a is 0 ＝0,g 0 ＝2y” 0 The method comprises the steps of carrying out a first treatment on the surface of the Given y' 0 In the time-course of which the first and second contact surfaces,
similarly, the boundary condition for the endpoint is determined according to equation 15:
a n M n-1 +2M n ＝g n equation 15;
wherein given y' n When a is n ＝0,g n ＝2y” n The method comprises the steps of carrying out a first treatment on the surface of the Given y' n When a is n ＝1,
The solution problem is constructed according to equation 16:
solving equation 16 to obtain M 0 ,M 1 ,…M n Obtaining a power curve modelThe resulting cubic spline fitting function formula is formula 17:
f(x)＝0.3105x 3 +0.4633x 2 +0.1538x+0.009898 equation 17.
8. A yaw error measurement apparatus for a multi-type wind turbine, comprising: memory, a processor and a computer program stored on the memory and executable on the processor, which when executed by the processor, performs the steps of the yaw error measurement method of a multi-type fan according to any one of claims 1 to 4.
9. A computer-readable storage medium, wherein a program for realizing information transfer is stored on the computer-readable storage medium, which program, when executed by a processor, realizes the steps of the yaw error measurement method of the multi-type fan according to any one of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110647957.5A CN113323818B (en) | 2021-06-10 | 2021-06-10 | Yaw error measurement method and device for multi-type fans |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110647957.5A CN113323818B (en) | 2021-06-10 | 2021-06-10 | Yaw error measurement method and device for multi-type fans |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113323818A CN113323818A (en) | 2021-08-31 |
| CN113323818B true CN113323818B (en) | 2024-01-19 |
Family
ID=77420465
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110647957.5A Active CN113323818B (en) | 2021-06-10 | 2021-06-10 | Yaw error measurement method and device for multi-type fans |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113323818B (en) |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2013152776A1 (en) * | 2012-04-13 | 2013-10-17 | Kk-Electronic A/S | A configuration system for a wind turbine control system |
| CN103440405A (en) * | 2013-08-08 | 2013-12-11 | 广东电网公司电力科学研究院 | Method and system for steady-state modeling of wind power plant based on measured data |
| CN109139363A (en) * | 2017-06-15 | 2019-01-04 | 南京工程学院 | A kind of maximum power point-tracing control method promoting multi-model wind mill performance |
| CN109488528A (en) * | 2018-11-23 | 2019-03-19 | 华润电力技术研究院有限公司 | A kind of method of adjustment and device of fan yaw system |
| CN109667727A (en) * | 2018-11-18 | 2019-04-23 | 浙江大学 | A kind of identification of Wind turbines yaw error inherent variability and compensation method based on power curve analysis |
| CN110046665A (en) * | 2019-04-17 | 2019-07-23 | 成都信息工程大学 | Based on isolated two abnormal classification point detecting method of forest, information data processing terminal |
| KR20190087095A (en) * | 2018-01-16 | 2019-07-24 | 군산대학교산학협력단 | Method and system for yaw control of WindTurbin |
| CN110067708A (en) * | 2019-05-13 | 2019-07-30 | 北京天泽智云科技有限公司 | A method of it is not positive to wind using power curve identification yaw |
| EP3569859A1 (en) * | 2018-05-14 | 2019-11-20 | Laborelec CVBA | Wind turbine efficiency improvement methods related to wind sensor misalignment |
| CN111563543A (en) * | 2020-04-26 | 2020-08-21 | 国网冀北电力有限公司电力科学研究院 | Method and device for cleaning wind speed-generated power data of wind turbine generator |
| CN112818604A (en) * | 2021-02-05 | 2021-05-18 | 河海大学 | Wind turbine generator risk degree assessment method based on wind power prediction |
| CN112905583A (en) * | 2021-04-01 | 2021-06-04 | 辽宁工程技术大学 | High-dimensional big data outlier detection method |
-
2021
- 2021-06-10 CN CN202110647957.5A patent/CN113323818B/en active Active
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2013152776A1 (en) * | 2012-04-13 | 2013-10-17 | Kk-Electronic A/S | A configuration system for a wind turbine control system |
| CN103440405A (en) * | 2013-08-08 | 2013-12-11 | 广东电网公司电力科学研究院 | Method and system for steady-state modeling of wind power plant based on measured data |
| CN109139363A (en) * | 2017-06-15 | 2019-01-04 | 南京工程学院 | A kind of maximum power point-tracing control method promoting multi-model wind mill performance |
| KR20190087095A (en) * | 2018-01-16 | 2019-07-24 | 군산대학교산학협력단 | Method and system for yaw control of WindTurbin |
| EP3569859A1 (en) * | 2018-05-14 | 2019-11-20 | Laborelec CVBA | Wind turbine efficiency improvement methods related to wind sensor misalignment |
| CN109667727A (en) * | 2018-11-18 | 2019-04-23 | 浙江大学 | A kind of identification of Wind turbines yaw error inherent variability and compensation method based on power curve analysis |
| CN109488528A (en) * | 2018-11-23 | 2019-03-19 | 华润电力技术研究院有限公司 | A kind of method of adjustment and device of fan yaw system |
| CN110046665A (en) * | 2019-04-17 | 2019-07-23 | 成都信息工程大学 | Based on isolated two abnormal classification point detecting method of forest, information data processing terminal |
| CN110067708A (en) * | 2019-05-13 | 2019-07-30 | 北京天泽智云科技有限公司 | A method of it is not positive to wind using power curve identification yaw |
| CN111563543A (en) * | 2020-04-26 | 2020-08-21 | 国网冀北电力有限公司电力科学研究院 | Method and device for cleaning wind speed-generated power data of wind turbine generator |
| CN112818604A (en) * | 2021-02-05 | 2021-05-18 | 河海大学 | Wind turbine generator risk degree assessment method based on wind power prediction |
| CN112905583A (en) * | 2021-04-01 | 2021-06-04 | 辽宁工程技术大学 | High-dimensional big data outlier detection method |
Non-Patent Citations (1)
| Title |
|---|
| 基于模糊聚类和孤立森林的用电数据异常检测;赵嫚 等;《陕西理工大学学报（自然科学版）》;第36卷(第4期);第38-43页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113323818A (en) | 2021-08-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107944622B (en) | Wind power prediction method based on continuous time interval clustering | |
| Pandit et al. | Comparison of advanced non‐parametric models for wind turbine power curves | |
| US10593422B2 (en) | Interaction network inference from vector representation of words | |
| CN111091233A (en) | Wind power plant short-term wind power prediction modeling method based on wavelet analysis and multi-model AdaBoost depth network | |
| CN114021799A (en) | Day-ahead wind power prediction method and system for wind power plant | |
| CN113822418A (en) | Wind power plant power prediction method, system, device and storage medium | |
| CN115511227B (en) | Wind power generation power prediction method and device based on stable learning | |
| CN116911806B (en) | Internet + based power enterprise energy information management system | |
| Li et al. | An improved feature selection method based on random forest algorithm for wind turbine condition monitoring | |
| CN111274532B (en) | Short-term wind power prediction method and device based on CEEMD-LZC and manifold learning | |
| Elvira-Ortiz et al. | Power quality monitoring strategy based on an optimized Multi-domain feature selection for the detection and classification of disturbances in wind generators | |
| CN114781520A (en) | Natural gas behavior abnormity detection method and system based on improved LOF model | |
| CN114781779A (en) | Unsupervised energy consumption abnormity detection method and device and storage medium | |
| CN113323818B (en) | Yaw error measurement method and device for multi-type fans | |
| Wang et al. | A novel multi-layer stacking ensemble wind power prediction model under Tensorflow deep learning framework considering feature enhancement and data hierarchy processing | |
| CN116910592B (en) | Log detection method and device, electronic equipment and storage medium | |
| CN115628910B (en) | Embedded bearing fault diagnosis device and equipment based on transfer learning | |
| CN108898157B (en) | Classification method for radar chart representation of numerical data based on convolutional neural network | |
| CN111062509A (en) | Wind power medium-short term prediction method and system based on improved neural network | |
| CN115801152A (en) | WiFi action identification method based on hierarchical transform model | |
| Rashedul Islam et al. | Maximum class separability-based discriminant feature selection using a GA for reliable fault diagnosis of induction motors | |
| CN113554203B (en) | Wind power prediction method and device based on high-dimensional meshing and LightGBM | |
| CN114118416A (en) | Variational graph automatic encoder method based on multi-task learning | |
| CN110175639B (en) | Short-term wind power prediction method based on feature selection | |
| CN114330881A (en) | Data-driven fan blade icing prediction method and device |
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 |