Method, device and computer equipment for cleaning outliers of fan characteristic data

Through the combination of the kernel density estimation method and the main peak threshold, the problem of inconsistent cleaning assumptions of fan characteristics data in the prior art is solved, and the modeling accuracy and cleaning effect of fan characteristics data are improved.

CN114647941BActive Publication Date: 2024-11-19TSINGHUA UNIVERSITY
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
CN202210300150.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-03-25
Publication Date
2024-11-19
Estimated Expiration
2042-03-25

AI Technical Summary

Technical Problem

The existing outlier cleaning method of fan characteristic data assumes that the data is normally distributed, resulting in low cleaning accuracy and poor cleaning effect, which affects the modeling accuracy of fan characteristic data.

Method used

The probability density distribution curve of fan characteristic data is determined by the kernel density estimation method, and the main peak interval is determined according to the main peak threshold for cleaning. The cleaning results of each interval are combined to improve the cleaning accuracy of fan characteristic data.

Benefits of technology

By determining the actual distribution of fan characteristic data, the accuracy and effect of outliers cleaning are improved, the number of outliers after cleaning is reduced, thereby improving the modeling accuracy of fan characteristic data.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114647941B_ABST
    Figure CN114647941B_ABST
Patent Text Reader

Abstract

The present application relates to a method, device, computer equipment, storage medium and computer program product for cleaning outliers of fan characteristic data. The method comprises: for each first fan parameter interval, according to the fan characteristic data corresponding to the first fan parameter interval, using the kernel density estimation method to determine the fan probability density distribution curve corresponding to the second fan parameter in the first fan parameter interval; according to the fan probability density distribution curve and the preset main peak threshold, determining the main peak interval corresponding to the first fan parameter interval, and according to the fan characteristic data and the main peak interval corresponding to the first fan parameter interval, determining the cleaning result corresponding to the first fan parameter interval; merging the cleaning results corresponding to each first fan parameter interval to obtain the cleaning result of the fan characteristic data. The use of this method can improve the modeling accuracy of fan characteristic data.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of data analysis and processing, and in particular to an outlier cleaning method, device, computer equipment, storage medium and computer program product. Background Art

[0002] As one of the main renewable energy sources, the importance of wind energy cannot be ignored. At present, wind farms collect wind speed, power and other state parameters through the Supervisory Control and Data Acquisition (SCADA) system. However, there are always a large number of abnormal data in the operating data collected by the system, resulting in many outliers in the scatter plots of wind turbine characteristic curves such as wind speed-power, speed-power and wind speed-speed. An effective way to reduce the adverse effects of these outliers is to use appropriate data cleaning technology to identify and remove outliers.

[0003] The current outlier cleaning method for wind turbine characteristic data realizes the cleaning of abnormal data in wind turbine characteristic data by assuming that the distribution of data points in the data interval is a normal distribution.

[0004] However, in actual situations, the distribution of data points within the data interval of fan characteristic data is not normally distributed, that is, the current outlier cleaning method of fan characteristic data does not conform to the actual situation of fan characteristic data. Therefore, the current outlier cleaning method of fan characteristic data has low cleaning accuracy and poor cleaning effect. After cleaning, there are still many outliers, resulting in low modeling accuracy of fan characteristic data. Summary of the invention

[0005] Based on this, it is necessary to provide an outlier cleaning method, device, computer equipment, computer-readable storage medium and computer program product that can improve the modeling accuracy of fan characteristic data in response to the above technical problems.

[0006] In a first aspect, the present application provides a method for cleaning outliers in wind turbine characteristic data. The method comprises:

[0007] Acquire fan characteristic data; the fan characteristic data is two-dimensional data including a first fan parameter and a second fan parameter;

[0008] Determining the fan characteristic data corresponding to each first fan parameter interval according to the fan characteristic data and the number of preset first fan parameter intervals;

[0009] For each first fan parameter interval, according to the fan characteristic data corresponding to the first fan parameter interval, a kernel density estimation method is used to determine a fan probability density distribution curve corresponding to the second fan parameter within the first fan parameter interval;

[0010] Determine the main peak interval corresponding to the first fan parameter interval according to the fan probability density distribution curve and the preset main peak threshold, and determine the cleaning result corresponding to the first fan parameter interval according to the fan characteristic data corresponding to the first fan parameter interval and the main peak interval; the main peak interval is the second fan parameter interval;

[0011] The cleaning results corresponding to the first fan parameter intervals are combined to obtain the cleaning results of the fan characteristic data.

[0012] In one embodiment, determining the main peak interval corresponding to the first wind turbine parameter interval according to the wind turbine probability density distribution curve and a preset main peak threshold value includes:

[0013] Determine a peak-valley interval corresponding to the first wind turbine parameter interval according to the wind turbine probability density distribution curve and a preset main peak threshold;

[0014] A main peak interval corresponding to the first wind turbine parameter interval is determined according to the wind turbine probability density distribution curve, the peak-valley interval and a preset boundary threshold.

[0015] In one embodiment, determining the main peak interval corresponding to the first wind turbine parameter interval according to the wind turbine probability density distribution curve, the peak-valley interval and a preset boundary threshold comprises:

[0016] Determining a boundary interval corresponding to the first wind turbine parameter interval according to the wind turbine probability density distribution curve, the peak-valley interval and a preset boundary threshold;

[0017] A main peak interval corresponding to the first wind turbine parameter interval is determined according to the boundary interval and a preset extension coefficient.

[0018] In one embodiment, determining the peak-valley interval corresponding to the first wind turbine parameter interval according to the wind turbine probability density distribution curve and a preset main peak threshold value includes:

[0019] Determine, according to the wind turbine probability density distribution curve, a first set of first wind turbine parameters corresponding to a maximum value of the wind turbine probability density distribution curve and a second set of first wind turbine parameters corresponding to a minimum value;

[0020] Determine a peak second fan parameter and a trough second fan parameter according to the first set, the second set and a preset main peak threshold;

[0021] A peak-valley interval corresponding to the first fan parameter interval is determined according to the peak second fan parameter and the trough second fan parameter.

[0022] In one embodiment, determining the fan characteristic data corresponding to each first fan parameter interval according to the fan characteristic data and the number of preset first fan parameter intervals includes:

[0023] According to the number of preset first fan parameter intervals, the fan characteristic data is divided into fan characteristic data corresponding to each first fan parameter interval;

[0024] For each first fan parameter interval, calculating a mean and a standard deviation of a second fan parameter in the first fan parameter interval;

[0025] An assumed distribution interval is determined according to the mean value and the standard deviation, and the fan characteristic data within the assumed distribution interval is used as the fan characteristic data corresponding to the first fan parameter interval.

[0026] In one of the embodiments, before determining the fan characteristic data corresponding to each first fan parameter interval according to the fan characteristic data and the number of preset first fan parameter intervals, the method further includes:

[0027] Determining design parameters of the target device according to the target device corresponding to the wind turbine characteristic data;

[0028] Determine a value range of the fan characteristic data according to the design parameters, and use the fan characteristic data within the value range as the fan characteristic data;

[0029] According to the fault record of the target device, the fault point is determined, and the fan characteristic data excluding each of the fault points is used as the fan characteristic data.

[0030] In a second aspect, the present application also provides a device for cleaning outliers in fan characteristic data. The device comprises:

[0031] An acquisition module, used to acquire fan characteristic data; the fan characteristic data is two-dimensional data including a first fan parameter and a second fan parameter;

[0032] A first determination module, configured to determine the fan characteristic data corresponding to each first fan parameter interval according to the fan characteristic data and the number of preset first fan parameter intervals;

[0033] A second determination module is used to determine, for each first fan parameter interval, a fan probability density distribution curve corresponding to the second fan parameter within the first fan parameter interval by using a kernel density estimation method according to the fan characteristic data corresponding to the first fan parameter interval;

[0034] a third determination module, configured to determine a main peak interval corresponding to the first fan parameter interval according to the fan probability density distribution curve and a preset main peak threshold, and determine a cleaning result corresponding to the first fan parameter interval according to the fan characteristic data corresponding to the first fan parameter interval and the main peak interval; the main peak interval is a second fan parameter interval;

[0035] A merging module is used to merge the cleaning results corresponding to the first fan parameter intervals to obtain the cleaning results of the fan characteristic data.

[0036] In one embodiment, the third determination module is specifically used to:

[0037] Determine a peak-valley interval corresponding to the first wind turbine parameter interval according to the wind turbine probability density distribution curve and a preset main peak threshold;

[0038] A main peak interval corresponding to the first wind turbine parameter interval is determined according to the wind turbine probability density distribution curve, the peak-valley interval and a preset boundary threshold.

[0039] In one embodiment, the third determination module is specifically used to:

[0040] Determining a boundary interval corresponding to the first wind turbine parameter interval according to the wind turbine probability density distribution curve, the peak-valley interval and a preset boundary threshold;

[0041] A main peak interval corresponding to the first wind turbine parameter interval is determined according to the boundary interval and a preset extension coefficient.

[0042] In one embodiment, the third determination module is specifically used to:

[0043] Determine, according to the wind turbine probability density distribution curve, a first set of first wind turbine parameters corresponding to a maximum value of the wind turbine probability density distribution curve and a second set of first wind turbine parameters corresponding to a minimum value;

[0044] Determine a peak second fan parameter and a trough second fan parameter according to the first set, the second set and a preset main peak threshold;

[0045] A peak-valley interval corresponding to the first fan parameter interval is determined according to the peak second fan parameter and the trough second fan parameter.

[0046] In one embodiment, the first determining module is specifically configured to:

[0047] According to the number of preset first fan parameter intervals, the fan characteristic data is divided into fan characteristic data corresponding to each first fan parameter interval;

[0048] For each first fan parameter interval, calculating a mean and a standard deviation of a second fan parameter in the first fan parameter interval;

[0049] An assumed distribution interval is determined according to the mean value and the standard deviation, and the fan characteristic data within the assumed distribution interval is used as the fan characteristic data corresponding to the first fan parameter interval.

[0050] In one embodiment, the device further comprises:

[0051] A fourth determination module, configured to determine design parameters of the target device according to the target device corresponding to the wind turbine characteristic data;

[0052] a fifth determination module, configured to determine a value interval of the fan characteristic data according to the design parameters, and use the fan characteristic data within the value interval as the fan characteristic data;

[0053] The sixth determination module is used to determine the fault point according to the fault record of the target device, and use the fan characteristic data without each of the fault points as the fan characteristic data.

[0054] In a third aspect, the present application further provides a computer device, wherein the computer device comprises a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps described in the first aspect are implemented.

[0055] In a fourth aspect, the present application further provides a computer-readable storage medium, wherein a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the steps described in the first aspect are implemented.

[0056] In a fifth aspect, the present application further provides a computer program product, wherein the computer program product comprises a computer program, and when the computer program is executed by a processor, the steps described in the first aspect are implemented.

[0057] The above-mentioned outlier cleaning method, device, computer equipment, storage medium and computer program product of fan characteristic data obtain fan characteristic data; the fan characteristic data is two-dimensional data including a first fan parameter and a second fan parameter; according to the fan characteristic data and the number of preset first fan parameter intervals, the fan characteristic data corresponding to each first fan parameter interval is determined; for each first fan parameter interval, according to the fan characteristic data corresponding to the first fan parameter interval, the kernel density estimation method is used to determine the fan probability density distribution curve corresponding to the second fan parameter in the first fan parameter interval; according to the fan probability density distribution curve and the preset main peak threshold, the main peak interval corresponding to the first fan parameter interval is determined, and according to the fan characteristic data corresponding to the first fan parameter interval and the main peak interval, the cleaning result corresponding to the first fan parameter interval is determined; the main peak interval is the second fan parameter interval; the cleaning results corresponding to each of the first fan parameter intervals are merged to obtain the cleaning result of the fan characteristic data. In this way, by determining the fan probability density distribution curve, the actual distribution of the fan characteristic data can be obtained, which is consistent with the actual situation. On the basis of obtaining the actual distribution of the fan characteristic data, the main peak of the fan characteristic data distribution is determined through the main peak threshold, so that the actual main distribution of the fan characteristic data can be obtained, so that the outlier cleaning of the fan characteristic data can be achieved, and the outlier cleaning accuracy of the fan characteristic data can be improved. The cleaning effect is good, and there are fewer outliers after cleaning, thereby improving the modeling accuracy of the fan characteristic data. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 A schematic flow chart of a method for cleaning outliers of wind turbine characteristic data in one embodiment;

[0059] Figure 2 It is a flowchart of a step of determining a main peak interval corresponding to a first wind turbine parameter interval according to a wind turbine probability density distribution curve and a preset main peak threshold value in one embodiment;

[0060] Figure 3 It is a flowchart of a step of determining a main peak interval corresponding to a first wind turbine parameter interval according to a wind turbine probability density distribution curve, a peak-valley interval and a preset boundary threshold in one embodiment;

[0061] Figure 4 A schematic diagram of a flow chart of a step of determining a peak-valley interval corresponding to a first wind turbine parameter interval in an embodiment;

[0062] Figure 5 A schematic flow chart of a step of determining fan characteristic data corresponding to each first fan parameter interval in an embodiment;

[0063] Figure 6A schematic flow chart of a method for cleaning outliers of wind turbine characteristic data in another embodiment;

[0064] Figure 7 is a schematic diagram of characteristic data of a fan to be cleaned in one embodiment;

[0065] Figure 8 It is a schematic diagram of fan characteristic data obtained by taking fan characteristic data within a value range as fan characteristic data in one embodiment;

[0066] Fig. 9 It is a schematic diagram of fan characteristic data obtained after removing fan characteristic data of each fault point as fan characteristic data in an embodiment;

[0067] Fig.10 Schematic diagram of fan characteristic data obtained by using the fan characteristic data within the assumed distribution interval as the fan characteristic data corresponding to the wind speed interval in an embodiment; wherein (d-1) and (d-2) are both wind speed intervals;

[0068] Fig.11 for Fig.10 Probability density distribution curve of wind turbines in the (d-1) wind speed interval; the interval between the two dotted lines is the main peak interval of the (d-1) wind speed interval;

[0069] Fig.12 for Fig.10 Probability density distribution curve of wind turbines in the (d-2) wind speed interval; the interval between the two dotted lines is the main peak interval of the (d-2) wind speed interval;

[0070] Fig.13 A schematic diagram of a cleaning result of fan characteristic data in one embodiment;

[0071] Fig.14 A structural block diagram of an outlier cleaning device for fan characteristic data in one embodiment;

[0072] Fig.15 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0073] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0074] In one embodiment, Figure 1As shown, a method for cleaning outliers of fan characteristic data is provided. This embodiment uses the method applied to a terminal as an example for illustration. It is understandable that the method can also be applied to a server, and can also be applied to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server. Among them, the terminal can be but is not limited to various personal computers, laptops, smart phones, tablet computers, Internet of Things devices and portable wearable devices. The Internet of Things devices can be smart speakers, smart TVs, smart air conditioners, smart car-mounted devices, etc. Portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc. The server can be implemented as an independent server or a server cluster composed of multiple servers. In this embodiment, the method includes the following steps:

[0075] Step 101, obtaining fan characteristic data.

[0076] The fan characteristic data is two-dimensional data including a first fan parameter and a second fan parameter.

[0077] In an embodiment of the present application, the first wind turbine parameter and the second wind turbine parameter are state parameters of the wind turbine set. The first wind turbine parameter includes: wind speed, rotation speed and power. The second wind turbine parameter includes: wind speed, rotation speed and power. The terminal may first determine the type of wind turbine characteristic data. Among them, the type of wind turbine characteristic data is used to indicate the type of operating data of the wind turbine set. The types of wind turbine characteristic data include: wind speed-power type, rotation speed-power type and wind speed-rotation speed type. Then, the terminal may obtain the wind turbine characteristic data from the pre-collected operating data of the wind turbine set according to the type of the wind turbine characteristic data.

[0078] Step 102: determining the fan characteristic data corresponding to each first fan parameter interval according to the fan characteristic data and the number of preset first fan parameter intervals.

[0079] In an embodiment of the present application, the terminal may pre-set the number of first fan parameter intervals. The number of first fan parameter intervals is related to the fan characteristic data. The denser the fan characteristic data, the greater the number of first fan parameter intervals. In this way, the range of the first fan parameter interval is small enough so that the number of fan characteristic data in each first fan parameter interval is appropriate and not too much, which can improve the accuracy and cleaning effect of outlier cleaning of the fan characteristic data, and there are fewer outliers after cleaning, thereby improving the modeling accuracy of the fan characteristic data. When the fan characteristic data remains unchanged, the number of first fan parameter intervals should be less than the threshold value of the number of first fan parameter intervals corresponding to the fan characteristic data. In this way, the number of first fan parameter intervals is appropriate and not too much, which avoids the situation where the cleaning process cannot be completed due to too few fan characteristic data in each first fan parameter interval, and ensures that the outlier cleaning of the fan characteristic data is carried out smoothly, thereby improving the modeling accuracy of the fan characteristic data.

[0080] Then, the terminal may divide the fan characteristic data into fan characteristic data corresponding to each first fan parameter interval according to the number of preset first fan parameter intervals.

[0081] Step 103 : for each first fan parameter interval, according to the fan characteristic data corresponding to the first fan parameter interval, a kernel density estimation method is used to determine a fan probability density distribution curve corresponding to the second fan parameter in the first fan parameter interval.

[0082] In an embodiment of the present application, for each first fan parameter interval, the terminal may calculate the fan probability density corresponding to each second fan parameter in the first fan parameter interval by using a kernel density estimation method based on the fan characteristic data corresponding to the first fan parameter interval.

[0083] In an example, the terminal may calculate the fan probability density corresponding to each second fan parameter in the first fan parameter interval by using a kernel density estimation method according to the fan characteristic data corresponding to the first fan parameter interval, which may be expressed as:

[0084]

[0085] where x 1 ,x 2 ,x 3 …x n is the fan characteristic data corresponding to the first fan parameter interval, n is the number of fan characteristic data corresponding to the first fan parameter interval, is the probability density at x, x is the second wind turbine parameter point whose probability density needs to be estimated, h is the window width, and K(·) is the kernel function, which is the Gaussian kernel function here.

[0086] Then, the terminal can normalize the fan probability density corresponding to each second fan parameter in the first fan parameter interval to obtain the normalized fan probability density corresponding to each second fan parameter in the first fan parameter interval. Then, the terminal can determine the fan probability density distribution curve corresponding to the second fan parameter in the first fan parameter interval based on the normalized fan probability density corresponding to each second fan parameter in the first fan parameter interval. Specifically, the terminal can use cubic spline interpolation to convert the normalized fan probability density corresponding to each second fan parameter in the first fan parameter interval into the fan probability density distribution curve corresponding to the second fan parameter in the first fan parameter interval.

[0087] Step 104, determining the main peak interval corresponding to the first fan parameter interval according to the fan probability density distribution curve and a preset main peak threshold, and determining the cleaning result corresponding to the first fan parameter interval according to the fan characteristic data and the main peak interval corresponding to the first fan parameter interval.

[0088] Among them, the main peak interval is the second wind turbine parameter interval.

[0089] In an embodiment of the present application, the terminal may pre-set a main peak threshold. The main peak threshold is a probability density threshold. The main peak threshold is used to screen out the peak-trough combination with the highest distribution density of the second wind turbine parameter within the first wind turbine parameter interval, that is, the main peak of the second wind turbine parameter distribution. The main peak threshold is related to the wind turbine group corresponding to the wind turbine characteristic data.

[0090] Then, the terminal can determine the peak-valley interval corresponding to the first fan parameter interval according to the fan probability density distribution curve and the preset main peak threshold. Then, the terminal can determine the main peak interval corresponding to the first fan parameter interval according to the peak-valley interval corresponding to the first fan parameter interval. Among them, the peak-valley interval is the second fan parameter interval. The peak-valley interval includes the second fan parameter corresponding to the peak of the main peak in the fan probability density distribution curve and the second fan parameter corresponding to the trough.

[0091] In an example, the terminal may use the peak-valley interval corresponding to the first wind turbine parameter interval as the main peak interval corresponding to the first wind turbine parameter interval.

[0092] Then, the terminal uses the fan characteristic data corresponding to the first fan parameter interval within the main peak interval as the cleaning fan characteristic data corresponding to the first fan parameter interval, and obtains the cleaning result corresponding to the first fan parameter interval.

[0093] Step 105 , merging the cleaning results corresponding to the first fan parameter intervals to obtain a cleaning result of the fan characteristic data.

[0094] In an embodiment of the present application, the terminal can splice together the cleaning fan characteristic data corresponding to each first fan parameter interval according to the original connection order of each first fan parameter interval to obtain the cleaning fan characteristic data, that is, obtain the cleaning result of the fan characteristic data.

[0095] In one embodiment, the number of the first fan parameter intervals is 2, which are the first fan parameter interval 1 and the first fan parameter interval 2. The cleaning result corresponding to the first fan parameter interval 1 is (a 1 ,a 2 ,a 3 ,…,a n ), the cleaning result corresponding to the first fan parameter interval 2 is (b 1 ,b 2 ,b 3 ,…,b n ). Among them, a 1 ,a 2 ,a 3 ,…,a n and b 1 ,b 2 ,b 3 ,…,b n The terminal can splice together the cleaning fan characteristic data corresponding to each first fan parameter interval according to the original connection order of each first fan parameter interval to obtain the cleaning result of the fan characteristic data (a 1 ,a 2 ,a 3 ,…,a n ,b 1 ,b 2 ,b 3 ,…,b n ).

[0096] In the above-mentioned outlier cleaning method of fan characteristic data, the terminal obtains fan characteristic data; then, according to the fan characteristic data and the number of preset first fan parameter intervals, the fan characteristic data corresponding to each first fan parameter interval is determined; then, for each first fan parameter interval, according to the fan characteristic data corresponding to the first fan parameter interval, the kernel density estimation method is used to determine the fan probability density distribution curve corresponding to the second fan parameter in the first fan parameter interval; then, according to the fan probability density distribution curve and the preset main peak threshold, the main peak interval corresponding to the first fan parameter interval is determined, and according to the fan characteristic data and the main peak interval corresponding to the first fan parameter interval, the cleaning result corresponding to the first fan parameter interval is determined; then, the cleaning results corresponding to each first fan parameter interval are merged to obtain the cleaning result of the fan characteristic data. In this way, by determining the fan probability density distribution curve, the actual distribution of the fan characteristic data can be obtained, which is consistent with the actual situation. Moreover, on the basis of obtaining the actual distribution of the fan characteristic data, the main peak of the fan characteristic data distribution is determined through the main peak threshold, and the actual main distribution of the fan characteristic data can be obtained, so that the outlier cleaning of the fan characteristic data can be achieved, and the outlier cleaning accuracy of the fan characteristic data can be improved, and the cleaning effect is good. There are fewer outliers after cleaning, thereby improving the modeling accuracy of the fan characteristic data. In addition, the fan characteristic data is divided into fan characteristic data corresponding to a plurality of first fan parameter intervals of equal size, so that outlier cleaning is performed for the fan characteristic data corresponding to each first fan parameter interval, and then the cleaning results corresponding to each first fan parameter interval are merged to obtain the cleaning result of the fan characteristic data, which can further improve the outlier cleaning accuracy of the fan characteristic data, and the outliers existing after cleaning are further reduced, thereby further improving the modeling accuracy of the fan characteristic data.

[0097] In one embodiment, Figure 2 As shown, according to the probability density distribution curve of the wind turbine and the preset main peak threshold, the specific process of determining the main peak interval corresponding to the first wind turbine parameter interval includes the following steps:

[0098] Step 201 : determining a peak-valley interval corresponding to the first wind turbine parameter interval according to a wind turbine probability density distribution curve and a preset main peak threshold.

[0099] In an embodiment of the present application, the terminal can determine the peak second fan parameter and the trough second fan parameter according to the fan probability density distribution curve and the preset main peak threshold. Among them, the peak second fan parameter is the second fan parameter corresponding to the peak of the main peak in the fan probability density distribution curve. The trough second fan parameter is the second fan parameter corresponding to the trough of the main peak in the fan probability density distribution curve. There is a main peak in the fan probability density distribution curve corresponding to each first fan parameter interval.

[0100] Then, the terminal may determine the peak-valley interval corresponding to the first fan parameter interval according to the peak second fan parameter and the trough second fan parameter.

[0101] Step 202 : determining the main peak interval corresponding to the first wind turbine parameter interval according to the wind turbine probability density distribution curve, the peak-valley interval and a preset boundary threshold.

[0102] In an embodiment of the present application, the terminal may pre-set a boundary threshold. The boundary threshold is a probability density threshold. The boundary threshold is used to divide the boundary of the peak-trough combination with the highest distribution density of the second wind turbine parameter in the first wind turbine parameter interval, that is, the second wind turbine parameter distribution main peak dividing boundary. The boundary threshold is related to the probability density distribution curve in the peak-trough interval.

[0103] Then, the terminal can determine the second fan parameter of the main peak peak value according to the fan probability density distribution curve and the peak-to-valley interval. At the same time, the terminal can determine the second fan parameter of the main peak boundary according to the fan probability density distribution curve and the preset boundary threshold. Then, the terminal can determine the boundary interval corresponding to the first fan parameter interval according to the second fan parameter of the main peak peak value and the second fan parameter of the main peak boundary. Among them, the boundary interval second fan parameter interval. The boundary interval is the boundary of the main peak in the fan probability density distribution curve.

[0104] Then, the terminal determines the main peak interval corresponding to the first wind turbine parameter interval according to the boundary interval corresponding to the first wind turbine parameter interval.

[0105] In an example, the terminal may use the boundary interval corresponding to the first wind turbine parameter interval as the main peak interval corresponding to the first wind turbine parameter interval.

[0106] In the above-mentioned outlier cleaning method of fan characteristic data, the terminal determines the peak-valley interval corresponding to the first fan parameter interval according to the fan probability density distribution curve and the preset main peak threshold; and then determines the main peak interval corresponding to the first fan parameter interval according to the fan probability density distribution curve, the peak-valley interval and the preset boundary threshold. In this way, by setting the boundary threshold and defining the boundary for the main peak of the fan characteristic data distribution, the fan characteristic data that is located in the peak-valley interval but has a large degree of deviation can be removed, further improving the accuracy and cleaning effect of the outlier cleaning of the fan characteristic data, and there are fewer outliers after cleaning, thereby further improving the modeling accuracy of the fan characteristic data.

[0107] In one embodiment, Figure 3 As shown, according to the probability density distribution curve of the wind turbine, the peak-valley interval and the preset boundary threshold, the specific process of determining the main peak interval corresponding to the first wind turbine parameter interval includes the following steps:

[0108] Step 301 : determining a boundary interval corresponding to a first wind turbine parameter interval according to a wind turbine probability density distribution curve, a peak-valley interval, and a preset boundary threshold.

[0109] In the embodiment of the present application, the terminal can determine the second wind turbine parameter of the main peak value according to the wind turbine probability density distribution curve and the peak-valley interval.

[0110] In one example, the terminal can determine the fan probability density corresponding to each second fan parameter in the peak-valley interval based on the fan probability density distribution curve in the peak-valley interval. Then, the terminal can compare the fan probability density corresponding to each second fan parameter in the peak-valley interval, and use the second fan parameter corresponding to the largest fan probability density as the second fan parameter of the main peak peak value.

[0111] In one example, the terminal can determine the fan probability density corresponding to each second fan parameter in the peak-valley interval according to the fan probability density distribution curve in the peak-valley interval. Then, the terminal can normalize the fan probability density corresponding to each second fan parameter in the peak-valley interval. Then, the terminal uses the normalized fan probability density as the fan probability density corresponding to each second fan parameter in the peak-valley interval. Then, the terminal uses the second fan parameter with a fan probability density of 1 in the peak-valley interval as the second fan parameter of the main peak peak value. In this way, by normalizing the fan probability density corresponding to each second fan parameter in the peak-valley interval and determining the second fan parameter of the main peak peak value, it can be ensured that in the case where the highest point of the fan probability density in the first fan parameter interval is not in the main peak of the data, the same standard can still be used to determine the second fan parameter of the main peak peak value, that is, it can provide a unified standard for determining the second fan parameter of the main peak peak value for various types of data main peaks. Moreover, in this way, the main peaks of various types of data in each first fan parameter interval only need a unified boundary threshold, thereby improving the outlier cleaning efficiency of the fan characteristic data.

[0112] In one example, the terminal can determine the fan probability density corresponding to each second fan parameter in the peak-valley interval based on the fan probability density distribution curve in the peak-valley interval. Then, the terminal can normalize the fan probability density corresponding to each second fan parameter in the peak-valley interval. Then, the terminal cubes the normalized fan probability density respectively. Then, the terminal uses the cubed fan probability density as the fan probability density corresponding to each second fan parameter in the peak-valley interval. Then, the terminal uses the second fan parameter with a fan probability density of 1 in the peak-valley interval as the second fan parameter of the main peak peak value. In this way, by normalizing and cubing the fan probability density corresponding to each second fan parameter in the peak-valley interval, the main peak can be made more prominent, and the accuracy and efficiency of determining the second fan parameter of the main peak peak value can be improved, thereby further improving the outlier cleaning efficiency of the fan characteristic data.

[0113] Then, the terminal can determine the fan probability density corresponding to each second fan parameter according to the fan probability density distribution curve. Then, the terminal can compare the fan probability density corresponding to each second fan parameter with the preset boundary threshold value. At the same time, the terminal compares the fan probability density corresponding to each second fan parameter. Then, the terminal can use the second fan parameter corresponding to the fan probability density that is less than or equal to the boundary threshold and closest to the boundary threshold as the second fan parameter of the main peak boundary.

[0114] Then, the terminal can calculate the boundary interval limit of the first fan parameter interval according to the second fan parameter of the main peak peak value and the second fan parameter of the main peak boundary. Then, the terminal determines the boundary interval corresponding to the first fan parameter interval according to the boundary interval limit of the first fan parameter interval. Among them, the boundary interval limit is the second fan parameter. The boundary interval limit includes: the boundary interval upper limit and the boundary interval lower limit.

[0115] In one embodiment, the terminal can calculate the peak value of the second wind turbine parameter P according to the main peak value. c and the second wind turbine parameter P at the main peak boundary b , calculate the lower limit p of the boundary interval of the first fan parameter interval 21 and the upper limit of the boundary interval p 22 , which can be expressed as:

[0116] P 2l =P c -abs(P c -P b )

[0117] P 22 =p c +abS(p c -P h )

[0118] Wherein, abs(·) is an absolute value function. Then, the terminal calculates the lower limit p of the boundary interval of the first wind turbine parameter interval according to the lower limit p of the boundary interval of the first wind turbine parameter interval. 21 and the upper limit of the boundary interval p 22 , determine the boundary interval corresponding to the first fan parameter interval as [p 21 ,p 22 ].

[0119] Step 302: Determine the main peak interval corresponding to the first wind turbine parameter interval according to the boundary interval and the preset extension coefficient.

[0120] In an embodiment of the present application, the terminal may pre-set an extension coefficient. The extension coefficient is a dimensionless constant. The extension coefficient is used to extend the boundary of the peak-trough combination with the highest distribution density of the second fan parameter in the first fan parameter interval screened out, that is, to extend the boundary of the main peak of the second fan parameter distribution. The extension coefficient is related to the wind turbine group corresponding to the fan characteristic data.

[0121] Then, the terminal can calculate the extension interval limit of the first wind turbine parameter interval according to the boundary interval and the preset extension coefficient. The extension interval limit is the second wind turbine parameter. The extension interval limit includes: an extension interval upper limit and an extension interval lower limit.

[0122] In one example, the terminal may calculate the upper limit of the extension interval according to the preset extension coefficient and the boundary interval upper limit of the boundary interval. Meanwhile, the terminal may calculate the lower limit of the extension interval according to the preset extension coefficient and the boundary interval lower limit of the boundary interval.

[0123] Then, the terminal determines the extended interval corresponding to the first fan parameter interval according to the extended interval limit of the first fan parameter interval. Then, the terminal can use the extended interval corresponding to the first fan parameter interval as the main peak interval corresponding to the first fan parameter interval.

[0124] In one embodiment, the terminal can use a preset extension coefficient λ 3 and the upper bound of the boundary interval p 22 , calculate the upper limit p of the extended interval 32 It can be expressed as:

[0125] P 32 =(1+λ 3 ) 22

[0126] At the same time, the terminal can be based on the preset extension coefficient λ 3 and the lower limit p of the boundary interval 21 , calculate the lower limit p of the extended interval 31 , which can be expressed as:

[0127] P 31 =(1-λ 3 ) 21

[0128] Then, the terminal determines the lower limit p of the extended interval of the first wind turbine parameter interval. 31 and the upper limit of the extended interval p 32 , determine the extension interval corresponding to the first fan parameter interval as [p 31 ,p 32Then, the terminal can store the extended interval [p 31 ,p 32 ] as the main peak interval corresponding to the first fan parameter interval.

[0129] In the above-mentioned outlier cleaning method of fan characteristic data, the terminal determines the boundary interval corresponding to the first fan parameter interval according to the fan probability density distribution curve, the peak-valley interval and the preset boundary threshold; and then determines the main peak interval corresponding to the first fan parameter interval according to the fan probability density distribution curve, the peak-valley interval and the preset boundary threshold. In this way, by setting the extension coefficient, the boundary of the main peak of the fan characteristic data distribution is extended, and the boundary of the main peak of the fan characteristic data distribution is appropriately relaxed, so that the outlier cleaning of the fan characteristic data can achieve a more reasonable cleaning effect, thereby further improving the modeling accuracy of the fan characteristic data.

[0130] In one embodiment, Figure 4 As shown, according to the probability density distribution curve of the wind turbine and the preset main peak threshold, the specific process of determining the peak-valley interval corresponding to the first wind turbine parameter interval includes the following steps:

[0131] Step 401 : determining, according to the wind turbine probability density distribution curve, a first set of first wind turbine parameters corresponding to the maximum value of the wind turbine probability density distribution curve and a second set of first wind turbine parameters corresponding to the minimum value of the wind turbine probability density distribution curve.

[0132] In an embodiment of the present application, the terminal can determine the maximum and minimum values ​​of the wind turbine probability density distribution curve based on the wind turbine probability density distribution curve. Then, the terminal determines the first wind turbine parameters corresponding to the respective maximum values ​​of the wind turbine probability density distribution curve based on the respective maximum values ​​of the wind turbine probability density distribution curve. Then, the terminal forms a first set of the first wind turbine parameters corresponding to the respective maximum values. At the same time, the terminal determines the first wind turbine parameters corresponding to the respective minimum values ​​of the wind turbine probability density distribution curve based on the respective minimum values ​​of the wind turbine probability density distribution curve. Then, the terminal forms a second set of the first wind turbine parameters corresponding to the respective minimum values.

[0133] Step 402: Determine the peak second fan parameter and the trough second fan parameter according to the first set, the second set and the preset main peak threshold.

[0134] In an embodiment of the present application, the terminal may sort the first fan parameters corresponding to each maximum value in the first set in order from large to small. Then, the terminal may match the first fan parameters corresponding to the minimum values ​​with the first fan parameters corresponding to the maximum values ​​in turn according to the obtained sorting results. If the first fan parameter corresponding to a maximum value and the first fan parameter corresponding to a minimum value meet the preset matching conditions, the first fan parameter corresponding to the maximum value and the first fan parameter corresponding to the minimum value are successfully matched. Among them, the matching conditions include a preset main peak threshold. The matching conditions are used to screen out the peak-trough combination with the highest distribution density of the second fan parameter in the first fan parameter interval, that is, the main peak of the second fan parameter distribution.

[0135] In one example, the preset matching conditions include: matching condition one and matching condition two. Matching condition one is that the difference between the probability density of the second fan parameter corresponding to the maximum value and the probability density of the second fan parameter corresponding to the minimum value is greater than the preset main peak threshold. Matching condition two is that the first fan parameter corresponding to the maximum value and the first fan parameter corresponding to the minimum value are closest to each other. When a first fan parameter corresponding to a maximum value and a first fan parameter corresponding to a minimum value satisfy both matching condition one and matching condition two, the first fan parameter corresponding to the maximum value and the first fan parameter corresponding to the minimum value satisfy the preset matching condition, and the two are matched successfully.

[0136] Then, the terminal uses the first fan parameter corresponding to the maximum value and the first fan parameter corresponding to the minimum value of the first pair of successful matches as the first fan parameter at the peak and the first fan parameter at the trough, respectively. Then, the terminal can determine the second fan parameter at the peak corresponding to the first fan parameter at the peak based on the first fan parameter at the peak. At the same time, the terminal can determine the second fan parameter at the trough corresponding to the first fan parameter at the trough based on the first fan parameter at the trough. Among them, the second fan parameter at the peak is the second fan parameter corresponding to the first fan parameter at the peak when the matching condition 1 is met. The second fan parameter at the trough is the second fan parameter corresponding to the first fan parameter at the trough when the matching condition 1 is met.

[0137] Step 403: Determine the peak-valley interval corresponding to the first fan parameter interval according to the peak second fan parameter and the trough second fan parameter.

[0138] In an embodiment of the present application, the terminal can calculate the peak-valley interval limit of the first fan parameter interval based on the peak second fan parameter and the trough second fan parameter. Wherein, the peak-valley interval limit is the second fan parameter. The peak-valley interval limit includes: a peak-valley interval upper limit and a peak-valley interval lower limit. Then, the terminal determines the peak-valley interval corresponding to the first fan parameter interval based on the peak-valley interval limit of the first fan parameter interval.

[0139] In one embodiment, the wind turbine characteristic data type is a wind speed-power type, the first wind turbine parameter is wind speed, and the second wind turbine parameter is power. The terminal determines a first set of wind speeds corresponding to the maximum value of the wind turbine probability density distribution curve according to the wind turbine probability density distribution curve A={a 1 ,a 2 ,…} and the second set of wind speeds corresponding to the minimum value B={b 1 ,b 2 ,…}. Then, the terminal can sort the wind speeds a in the first set A in descending order. i Then, the terminal can sort the wind speed a corresponding to each maximum value according to the sorting results. i Match the wind speed b corresponding to the minimum value j The preset matching conditions include: matching condition 1 and matching condition 2. Matching condition 1 can be expressed as Among them, λ 1 is the main peak threshold, and The wind speed a i and b j The probability density of the corresponding power point. The second matching condition is the wind speed a corresponding to the maximum value i The wind speed b corresponding to the minimum value j When the wind speed is a i With wind speed b j At the same time, the matching condition 1 and the matching condition 2 are met, and the wind speed a i With wind speed b j The preset matching conditions are met and the matching is successful. Then, the terminal will i With wind speed b j , as the peak wind speed and trough wind speed respectively. Then, the terminal can determine the peak power corresponding to the peak wind speed based on the peak wind speed. The peak power is the power corresponding to the peak wind speed when the matching condition is met. At the same time, the terminal can determine the trough power corresponding to the trough wind speed according to the trough wind speed. The trough power is the power corresponding to the trough wind speed when the matching condition is met. Then, the terminal can calculate the peak power according to the peak power. and trough power Calculate the peak-to-valley interval upper limit p of the first fan parameter interval 12 and the lower limit of the peak-to-valley interval p 11 , which can be expressed as:

[0140]

[0141]

[0142] Then, the peak-to-valley interval upper limit p of the first wind turbine parameter interval is 12 and the lower limit of the peak-to-valley interval p 11 , determine the peak-valley interval corresponding to the first fan parameter interval [p 11 ,p 12 ].

[0143] In the above-mentioned outlier cleaning method for fan characteristic data, the terminal determines the first set of first fan parameters corresponding to the maximum value of the fan probability density distribution curve and the second set of first fan parameters corresponding to the minimum value according to the fan probability density distribution curve; then determines the peak second fan parameter and the trough second fan parameter according to the first set, the second set and the preset main peak threshold; then determines the peak-valley interval corresponding to the first fan parameter interval according to the peak second fan parameter and the trough second fan parameter. In this way, through the main peak threshold, only one set of peak second fan parameters and trough second fan parameters are obtained in each first fan parameter interval, and only one peak-valley interval is obtained, which can determine the main peak of the fan characteristic data distribution and obtain the actual main distribution of the fan characteristic data. In this way, the outlier cleaning of the fan characteristic data is realized, which can improve the accuracy of the outlier cleaning of the fan characteristic data, and the cleaning effect is good. There are fewer outliers after cleaning, thereby improving the modeling accuracy of the fan characteristic data.

[0144] In one embodiment, Figure 5 As shown, according to the fan characteristic data and the number of preset first fan parameter intervals, the specific process of determining the fan characteristic data corresponding to each first fan parameter interval includes the following steps:

[0145] Step 501 : dividing the fan characteristic data into fan characteristic data corresponding to each first fan parameter interval according to the number of preset first fan parameter intervals.

[0146] In the embodiment of the present application, the terminal can divide the first fan parameter interval corresponding to the fan characteristic data into a plurality of first fan parameter intervals of equal size according to the number of preset first fan parameter intervals. Then, the terminal divides the fan characteristic data into fan characteristic data corresponding to each first fan parameter interval.

[0147] Step 502: For each first fan parameter interval, calculate the mean and standard deviation of the second fan parameter in the first fan parameter interval.

[0148] In the embodiment of the present application, for each first fan parameter interval, the terminal may calculate the mean value of the second fan parameter in the first fan parameter interval according to the fan characteristic data corresponding to the first fan parameter interval. Then, the terminal may calculate the standard deviation of the second fan parameter in the first fan parameter interval according to the fan characteristic data corresponding to the first fan parameter interval and the mean value of the second fan parameter in the first fan parameter interval.

[0149] Step 503: determine a hypothetical distribution interval according to the mean and the standard deviation, and use the fan characteristic data within the hypothetical distribution interval as the fan characteristic data corresponding to the first fan parameter interval.

[0150] In an embodiment of the present application, the terminal can calculate the limit of the assumed distribution interval based on the 3σ principle, the mean and the standard deviation. Wherein, the limit of the assumed distribution interval is the second wind turbine parameter. The limit of the assumed distribution interval includes: the upper limit of the assumed distribution interval and the lower limit of the assumed distribution interval. Then, the terminal can determine the assumed distribution interval based on the limit of the assumed distribution interval. Then, the terminal can use the fan characteristic data within the assumed distribution interval as the fan characteristic data corresponding to the first wind turbine parameter interval.

[0151] In one embodiment, the mean value of the second fan parameter in a first fan parameter interval is μ 0 , with a standard deviation of σ. The terminal can calculate the upper limit of the assumed distribution interval as μ based on the 3σ principle, mean and standard deviation. 0 +3σ, assuming the lower limit of the distribution interval is μ 0 -3σ. Then, the terminal determines the assumed distribution interval as [μ 0 -3σ,μ 0 +3σ]. Then, the terminal can convert the interval [μ 0 -3σ,μ 0 The fan characteristic data within the range of [+3σ] is used as the fan characteristic data corresponding to the first fan parameter interval.

[0152] In the above-mentioned outlier cleaning method of fan characteristic data, the terminal divides the fan characteristic data into fan characteristic data corresponding to each first fan parameter interval according to the number of preset first fan parameter intervals; then, for each first fan parameter interval, the mean and standard deviation of the second fan parameter in the first fan parameter interval are calculated; then, based on the mean and standard deviation, the assumed distribution interval is determined, and the fan characteristic data within the assumed distribution interval is used as the fan characteristic data corresponding to the first fan parameter interval. In this way, by dividing the fan characteristic data into fan characteristic data corresponding to a plurality of first fan parameter intervals of equal size, outlier cleaning is performed for the fan characteristic data corresponding to each first fan parameter interval, and then the cleaning results corresponding to each first fan parameter interval are merged to obtain the cleaning results of the fan characteristic data, the accuracy of outlier cleaning of the fan characteristic data can be further improved, and the outliers existing after cleaning are further reduced, thereby further improving the modeling accuracy of the fan characteristic data. In addition, by determining the assumed distribution interval through the mean and standard deviation, and performing outlier cleaning on the fan characteristic data based on the assumed distribution interval, outliers that obviously deviate from normal data can be removed, thereby achieving multiple cleanings of the fan characteristic data, and further improving the accuracy of outlier cleaning of the fan characteristic data. The number of outliers after cleaning is further reduced, thereby further improving the modeling accuracy of the fan characteristic data.

[0153] In one embodiment, Figure 6 As shown, the specific process before determining the fan characteristic data corresponding to each first fan parameter interval according to the fan characteristic data and the number of preset first fan parameter intervals also includes the following steps:

[0154] Step 601, determining design parameters of the target device according to the target device corresponding to the wind turbine characteristic data.

[0155] In an embodiment of the present application, the terminal can determine the target device corresponding to the wind turbine characteristic data according to the wind turbine characteristic data. Then, the terminal can determine the design parameters of the target device according to the target device. The target device is a target wind turbine. The design parameters include: rated power, rated wind speed, rated speed and cut-out wind speed.

[0156] Step 602, determining a value range of the fan characteristic data according to the design parameters, and taking the fan characteristic data within the value range as the fan characteristic data.

[0157] In an embodiment of the present application, the terminal may first determine the type of wind turbine characteristic data. The type of wind turbine characteristic data is used to indicate the type of operating data of the wind turbine group. The types of wind turbine characteristic data include: wind speed-power type, speed-power type and wind speed-speed type. Then, the terminal may obtain the target design parameters in the design parameters according to the type of wind turbine characteristic data. The target design parameters include: a first target design parameter and a second target design parameter. The target design parameter is the design parameter corresponding to the type of wind turbine characteristic data, that is, the first target design parameter is the design parameter corresponding to the first wind turbine parameter, and the second target design parameter is the design parameter corresponding to the second wind turbine parameter. For example, the type of wind turbine characteristic data is the wind speed-power type, the first wind turbine parameter is the wind speed, the second wind turbine parameter is the power, the first target design parameter is the cut-out wind speed, and the second target design parameter is the rated power.

[0158] Then, the terminal determines the first fan value interval of the fan characteristic data according to the first target design parameter. At the same time, the terminal determines the second fan value interval of the fan characteristic data according to the second target design parameter. Among them, the first fan value interval is the interval of the first fan parameter. The second fan value interval is the interval of the second fan parameter. The first fan value interval and the second fan value interval constitute a value interval.

[0159] In one example, the terminal uses the interval from 0 to the first target design parameter as the first fan value interval of the fan characteristic data. At the same time, the terminal multiplies the second target design parameter by a preset interval coefficient. Then, the terminal uses the interval from 0 to the product of the second target design parameter and the preset interval coefficient as the second fan value interval of the fan characteristic data. Among them, the interval coefficient can be a constant between 1-1.5. For example, the interval coefficient can be 1.2. The fan characteristic data obtained in this way is more reasonable, can improve the accuracy of outlier cleaning of the fan characteristic data, and has a better cleaning effect, thereby further improving the modeling accuracy of the fan characteristic data.

[0160] Then, the terminal uses the fan characteristic data within the value range as the fan characteristic data.

[0161] In one embodiment, the type of wind turbine characteristic data is a wind speed-power type, the first wind turbine parameter is wind speed, the second wind turbine parameter is power, the first target design parameter is a cut-out wind speed of 20 m / s, and the second target design parameter is a rated power of 1500 kW. The terminal uses the interval from 0 to 20 m / s as the first wind turbine value interval of the wind turbine characteristic data. At the same time, the terminal multiplies the second target design parameter 1500 kW by a preset interval coefficient of 1.2. Then, the terminal uses the interval from 0 to 1800 kW, which is the product of the second target design parameter and the preset interval coefficient, as the second wind turbine value interval of the wind turbine characteristic data. Then, the terminal uses the wind turbine characteristic data in the interval from 0 to 20 m / s and in the interval from 0 to 1800 kW as the wind turbine characteristic data.

[0162] Step 603, determining the fault point according to the fault record of the target device, and removing the fan characteristic data of each fault point as the fan characteristic data.

[0163] In the embodiment of the present application, the terminal can determine the fault point according to the fault record of the target device. Then, the terminal removes each fault point from the fan characteristic data, and uses the fan characteristic data without each fault point as the fan characteristic data.

[0164] In the above-mentioned outlier cleaning method of fan characteristic data, the terminal determines the design parameters of the target device according to the target device corresponding to the fan characteristic data; then determines the value range of the fan characteristic data according to the design parameters, and uses the fan characteristic data within the value range as the fan characteristic data; then determines the fault point according to the fault record of the target device, and uses the fan characteristic data without each fault point as the fan characteristic data. In this way, by using the value range and fault point of the fan characteristic data to clean the fan characteristic data, the range of normal fan characteristic data can be determined first, and the fault point can be removed, so as to achieve multiple pre-cleaning of the fan characteristic data, further improve the accuracy of the outlier cleaning of the fan characteristic data, and further reduce the outliers existing after cleaning, thereby further improving the modeling accuracy of the fan characteristic data. At the same time, the use of multiple pre-cleaning of the fan characteristic data with short time consumption can reduce the time of subsequent outlier cleaning, thereby reducing the total time of outlier cleaning of the fan characteristic data and improving the efficiency of outlier cleaning of the fan characteristic data.

[0165] In one embodiment, the wind turbine characteristic data is the wind speed-power data of a certain wind farm, the first wind turbine parameter is the 1-second average wind speed V, the second wind turbine parameter is the active power P, and the wind turbine characteristic data is as follows: Figure 7As shown. The terminal determines the design parameters of the target device according to the target device corresponding to the fan characteristic data, including: fan rated power 1500KW, cut-in wind speed 3m / s, rated wind speed 9.5m / s, cut-out wind speed 20m / s. Then, the terminal determines the first fan value interval of the fan characteristic data as [0,20] (unit: m / s) and the first fan value interval as [0,1800] (unit: KW) according to the design parameters. Then, the terminal uses the fan characteristic data that is located in the first fan value interval and in the second fan value interval as the fan characteristic data, as shown in FIG. Figure 8 Then, the terminal determines the fault point according to the fault record of the target device, and removes the fan characteristic data of each fault point as the fan characteristic data, as shown in Fig. 9 As shown in the figure, the black dots are the fan characteristic data, and the gray crosses are the outlier data that have been cleaned. Then, the terminal divides the fan characteristic data into 40 wind speed intervals according to the preset number of wind speed intervals, that is, the wind speed intervals [0, 0.5), [0.5, 1), …, [19.5, 20] corresponding to the fan characteristic data. Then, for each wind speed interval, the terminal calculates the mean value of the power in the wind speed interval as μ 0 , with a standard deviation of σ. Then, the terminal determines the assumed distribution interval as [μ 0 -3σ,μ 0 +3σ]. Then, the terminal uses the fan characteristic data within the assumed distribution interval as the fan characteristic data corresponding to the wind speed interval, such as Fig.10 As shown, the black dots are the wind turbine characteristic data, and the gray crosses are the outlier data that have been cleaned. Then, the terminal determines the peak-valley interval corresponding to the wind speed interval based on the wind turbine probability density distribution curve and the preset main peak threshold. Then, the terminal determines the boundary interval corresponding to the wind speed interval based on the wind turbine probability density distribution curve, the peak-valley interval, and the preset boundary threshold. Then, the terminal determines the main peak interval corresponding to the wind speed interval based on the boundary interval and the preset extension coefficient, so that the boundary of the main peak of the wind turbine characteristic data distribution is appropriately relaxed. As shown in Fig.10 and Fig.11 As shown, Fig.11 for Fig.10 Probability density distribution curve of wind turbines in the (d-1) wind speed range, Fig.11 The interval between the two dotted lines in is the main peak interval of the (d-1) wind speed interval. Fig.10 and Fig.12 As shown, Fig.12 for Fig.10 Probability density distribution curve of wind turbines in the middle (d-2) wind speed range, Fig.12The interval between the two dotted lines in is the main peak interval of the (d-2) wind speed interval. Then, the terminal determines the cleaning result corresponding to the wind speed interval based on the fan characteristic data and the main peak interval corresponding to the wind speed interval. Then, the terminal combines the cleaning results corresponding to each wind speed interval to obtain the cleaning result of the fan characteristic data, such as Fig.13 As shown, the black points are the fan characteristic data, and the gray crosses are the outlier data that have been cleaned.

[0166] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0167] It is understandable that the method can also be applied to outlier cleaning of characteristic data of other mechanical equipment, and the embodiments of the present application are not limited thereto.

[0168] Based on the same inventive concept, the embodiment of the present application also provides an outlier cleaning device for fan characteristic data for implementing the outlier cleaning method for fan characteristic data involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in the embodiments of one or more outlier cleaning devices for fan characteristic data provided below can refer to the limitations of the outlier cleaning method for fan characteristic data above, and will not be repeated here.

[0169] In one embodiment, Fig.14 As shown, a device 1400 for cleaning outliers of fan characteristic data is provided, comprising: an acquisition module 1410, a first determination module 1420, a second determination module 1430, a third determination module 1440 and a merging module 1450, wherein:

[0170] The acquisition module 1410 is used to acquire fan characteristic data; the fan characteristic data is two-dimensional data including a first fan parameter and a second fan parameter;

[0171] A first determination module 1420, configured to determine the fan characteristic data corresponding to each first fan parameter interval according to the fan characteristic data and the number of preset first fan parameter intervals;

[0172] A second determination module 1430 is used to determine, for each first fan parameter interval, a fan probability density distribution curve corresponding to the second fan parameter within the first fan parameter interval by using a kernel density estimation method according to the fan characteristic data corresponding to the first fan parameter interval;

[0173] The third determination module 1440 is used to determine the main peak interval corresponding to the first fan parameter interval according to the fan probability density distribution curve and the preset main peak threshold, and determine the cleaning result corresponding to the first fan parameter interval according to the fan characteristic data corresponding to the first fan parameter interval and the main peak interval; the main peak interval is the second fan parameter interval;

[0174] The merging module 1450 is used to merge the cleaning results corresponding to the first fan parameter intervals to obtain the cleaning results of the fan characteristic data.

[0175] Optionally, the third determining module 1440 is specifically configured to:

[0176] Determine a peak-valley interval corresponding to the first wind turbine parameter interval according to the wind turbine probability density distribution curve and a preset main peak threshold;

[0177] A main peak interval corresponding to the first wind turbine parameter interval is determined according to the wind turbine probability density distribution curve, the peak-valley interval and a preset boundary threshold.

[0178] Optionally, the third determining module 1440 is specifically configured to:

[0179] Determining a boundary interval corresponding to the first wind turbine parameter interval according to the wind turbine probability density distribution curve, the peak-valley interval and a preset boundary threshold;

[0180] A main peak interval corresponding to the first wind turbine parameter interval is determined according to the boundary interval and a preset extension coefficient.

[0181] Optionally, the third determining module 1440 is specifically configured to:

[0182] Determine, according to the wind turbine probability density distribution curve, a first set of first wind turbine parameters corresponding to a maximum value of the wind turbine probability density distribution curve and a second set of first wind turbine parameters corresponding to a minimum value;

[0183] Determine a peak second fan parameter and a trough second fan parameter according to the first set, the second set and a preset main peak threshold;

[0184] A peak-valley interval corresponding to the first fan parameter interval is determined according to the peak second fan parameter and the trough second fan parameter.

[0185] Optionally, the first determining module 1420 is specifically configured to:

[0186] According to the number of preset first fan parameter intervals, the fan characteristic data is divided into fan characteristic data corresponding to each first fan parameter interval;

[0187] For each first fan parameter interval, calculating a mean and a standard deviation of a second fan parameter in the first fan parameter interval;

[0188] An assumed distribution interval is determined according to the mean value and the standard deviation, and the fan characteristic data within the assumed distribution interval is used as the fan characteristic data corresponding to the first fan parameter interval.

[0189] Optionally, the device 1400 further includes:

[0190] A fourth determination module, configured to determine design parameters of the target device according to the target device corresponding to the wind turbine characteristic data;

[0191] a fifth determination module, configured to determine a value interval of the fan characteristic data according to the design parameters, and use the fan characteristic data within the value interval as the fan characteristic data;

[0192] The sixth determination module is used to determine the fault point according to the fault record of the target device, and use the fan characteristic data without each of the fault points as the fan characteristic data.

[0193] Each module in the outlier cleaning device for fan characteristic data can be implemented in whole or in part by software, hardware, or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each module.

[0194] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Fig.15As shown. The computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a method for cleaning outliers of fan characteristic data is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse, etc.

[0195] Those skilled in the art will understand that Fig.15 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0196] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0197] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0198] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0199] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0200] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0201] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0202] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for cleaning outliers of fan characteristic data, characterized in that: The method comprises: Acquire fan characteristic data; the fan characteristic data is two-dimensional data including a first fan parameter and a second fan parameter; Determining the fan characteristic data corresponding to each first fan parameter interval according to the fan characteristic data and the number of preset first fan parameter intervals; For each first fan parameter interval, according to the fan characteristic data corresponding to the first fan parameter interval, a kernel density estimation method is used to determine a fan probability density distribution curve corresponding to the second fan parameter within the first fan parameter interval; Determine the main peak interval corresponding to the first fan parameter interval according to the fan probability density distribution curve and the preset main peak threshold, and determine the cleaning result corresponding to the first fan parameter interval according to the fan characteristic data corresponding to the first fan parameter interval and the main peak interval; the main peak interval is the second fan parameter interval; Merging the cleaning results corresponding to the first fan parameter intervals to obtain a cleaning result of the fan characteristic data; Wherein, the determining the main peak interval corresponding to the first fan parameter interval according to the fan probability density distribution curve and the preset main peak threshold value comprises: determining the peak-valley interval corresponding to the first fan parameter interval according to the fan probability density distribution curve and the preset main peak threshold value; determining the main peak interval corresponding to the first fan parameter interval according to the fan probability density distribution curve, the peak-valley interval and the preset boundary threshold value; Wherein, the determining the main peak interval corresponding to the first fan parameter interval according to the fan probability density distribution curve, the peak-valley interval and the preset boundary threshold comprises: determining the boundary interval corresponding to the first fan parameter interval according to the fan probability density distribution curve, the peak-valley interval and the preset boundary threshold; determining the main peak interval corresponding to the first fan parameter interval according to the boundary interval and the preset extension coefficient; Among them, determining the peak-valley interval corresponding to the first fan parameter interval according to the wind turbine probability density distribution curve and a preset main peak threshold value includes: determining a first set of first fan parameters corresponding to the maximum value and a second set of first fan parameters corresponding to the minimum value of the fan probability density distribution curve according to the wind turbine probability density distribution curve; determining a peak second fan parameter and a trough second fan parameter according to the first set, the second set and the preset main peak threshold value; determining the peak-valley interval corresponding to the first fan parameter interval according to the peak second fan parameter and the trough second fan parameter.

2. The method according to claim 1, characterized in that The determining, according to the fan characteristic data and the number of preset first fan parameter intervals, the fan characteristic data corresponding to each first fan parameter interval comprises: According to the number of preset first fan parameter intervals, the fan characteristic data is divided into fan characteristic data corresponding to each first fan parameter interval; For each first fan parameter interval, calculating a mean and a standard deviation of a second fan parameter in the first fan parameter interval; An assumed distribution interval is determined according to the mean value and the standard deviation, and the fan characteristic data within the assumed distribution interval is used as the fan characteristic data corresponding to the first fan parameter interval.

3. The method according to claim 1, characterized in that: Before determining the fan characteristic data corresponding to each first fan parameter interval according to the fan characteristic data and the number of preset first fan parameter intervals, the method further includes: Determining design parameters of the target device according to the target device corresponding to the wind turbine characteristic data; Determine a value range of the fan characteristic data according to the design parameters, and use the fan characteristic data within the value range as the fan characteristic data; According to the fault record of the target device, the fault point is determined, and the fan characteristic data excluding each of the fault points is used as the fan characteristic data.

4. The method according to claim 1, characterized in that: Determining the fan probability density distribution curve corresponding to the second fan parameter within the first fan parameter interval by using a kernel density estimation method according to the fan characteristic data corresponding to the first fan parameter interval includes: Calculate the fan probability density corresponding to each of the second fan parameters within the first fan parameter interval using a kernel density estimation method according to the fan characteristic data corresponding to the first fan parameter interval; Normalizing the fan probability density corresponding to each of the second fan parameters within the first fan parameter interval to obtain a normalized fan probability density corresponding to each of the second fan parameters within the first fan parameter interval; A fan probability density distribution curve corresponding to the second fan parameter in the first fan parameter interval is determined according to the normalized fan probability densities corresponding to each of the second fan parameters in the first fan parameter interval.

5. A device for cleaning outliers of fan characteristic data, characterized in that: The device comprises: An acquisition module, used to acquire fan characteristic data; the fan characteristic data is two-dimensional data including a first fan parameter and a second fan parameter; A first determination module, configured to determine the fan characteristic data corresponding to each first fan parameter interval according to the fan characteristic data and the number of preset first fan parameter intervals; A second determination module is used to determine, for each first fan parameter interval, a fan probability density distribution curve corresponding to the second fan parameter within the first fan parameter interval by using a kernel density estimation method according to the fan characteristic data corresponding to the first fan parameter interval; a third determination module, configured to determine a main peak interval corresponding to the first fan parameter interval according to the fan probability density distribution curve and a preset main peak threshold, and determine a cleaning result corresponding to the first fan parameter interval according to the fan characteristic data corresponding to the first fan parameter interval and the main peak interval; the main peak interval is a second fan parameter interval; a merging module, configured to merge the cleaning results corresponding to the first fan parameter intervals to obtain a cleaning result of the fan characteristic data; Wherein, the third determination module is specifically used to: determine the peak-valley interval corresponding to the first wind turbine parameter interval according to the wind turbine probability density distribution curve and the preset main peak threshold; determine the main peak interval corresponding to the first wind turbine parameter interval according to the wind turbine probability density distribution curve, the peak-valley interval and the preset boundary threshold; The third determination module is specifically used to: determine the boundary interval corresponding to the first wind turbine parameter interval according to the wind turbine probability density distribution curve, the peak-valley interval and the preset boundary threshold; determine the main peak interval corresponding to the first wind turbine parameter interval according to the boundary interval and the preset extension coefficient; Among them, the third determination module is specifically used to: determine the first set of first fan parameters corresponding to the maximum value and the second set of first fan parameters corresponding to the minimum value of the fan probability density distribution curve according to the wind turbine probability density distribution curve; determine the peak second fan parameter and the trough second fan parameter according to the first set, the second set and the preset main peak threshold; determine the peak-valley interval corresponding to the first fan parameter interval according to the peak second fan parameter and the trough second fan parameter.

6. The device according to claim 5, characterized in that The first determining module is specifically configured to: According to the number of preset first fan parameter intervals, the fan characteristic data is divided into fan characteristic data corresponding to each first fan parameter interval; For each first fan parameter interval, calculating a mean and a standard deviation of a second fan parameter in the first fan parameter interval; An assumed distribution interval is determined according to the mean value and the standard deviation, and the fan characteristic data within the assumed distribution interval is used as the fan characteristic data corresponding to the first fan parameter interval.

7. The device according to claim 5, characterized in that The device also includes: A fourth determination module, configured to determine design parameters of the target device according to the target device corresponding to the wind turbine characteristic data; a fifth determination module, configured to determine a value interval of the fan characteristic data according to the design parameters, and use the fan characteristic data within the value interval as the fan characteristic data; The sixth determination module is used to determine the fault point according to the fault record of the target device, and use the fan characteristic data without each of the fault points as the fan characteristic data.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 4 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.

Citation Information

Patent Citations

  • A data cleaning method based on two-dimensional probability density estimation and a quartile method

    CN109918364A

  • Wind turbine generator abnormal data cleaning method

    CN110134919A

  • Wind turbine generator set power characteristic evaluation method and system

    CN111242411A