A method, device and equipment for early warning of a variable pitch system failure and a storage medium

Abstract

The application discloses a kind of variable pitch system fault early warning method, device, equipment and storage medium, applied to wind power equipment fault detection field, the method comprises: according to the equivalent equivalent torque of each blade in each time interval in motor torque data in mechanical operation data calculation;According to the correlation parameter and significance parameter between adjacent blades of the equivalent equivalent torque of each blade;According to the equivalent equivalent torque and motor current data, motor temperature data, pitch angle data and speed data in mechanical operation data calculation Mahalanobis distance;When judging mechanical failure based on Mahalanobis distance, correlation and significance level, output fault early warning.The method of the present application constructs correlation model and monitoring model by mechanical operating state data to carry out early warning of wind power equipment mechanical component failure, to avoid prior art only after mechanical component damage to carry out fault notification and cause unit failure.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of wind power equipment fault detection, and in particular to a variable pitch system fault early warning method, device, equipment and computer readable storage medium. BACKGROUND

[0002] With the advent of the era of parity in the wind power industry, the trend of large-scale wind turbine is further accelerated, and more and more large-capacity megawatt wind turbines are installed and operated. Wind turbine is operated in complex working environment for a long time, and bearings, bolts and other vulnerable parts are prone to failure, and mechanical and electrical system failures occur from time to time; the unit is scattered, and the maintenance is difficult, and the regular maintenance and after-sales service seriously affect the power generation benefit; the maintenance and replacement cost of large parts is high, and the proportion of the cost of the whole machine and the power generation cost is high, and the maintenance expenditure is large; the construction of wind power plant deviates from the city, and the inspection cost is high. Therefore, early fault warning and early intervention of the unit is an important means to improve the reliability of the unit and reduce the operation and maintenance cost.

[0003] At present, the fault warning of the mechanical parts of the variable pitch system in the wind power equipment is mainly through the threshold setting method, according to the material, process and fatigue calculation of the mechanical parts, the mechanical parts have the maximum value and the rated value, and different fault levels are set by different threshold values. The traditional mechanical part fault monitoring method is too simple and passive, and the mechanical part can only be monitored when it is obviously damaged, at which time the reliability and power generation performance of the unit have been seriously affected. SUMMARY

[0004] The purpose of the present application is to provide a variable pitch system fault early warning method, device, equipment and storage medium, which is applied to the field of wind power equipment fault detection, and the method constructs a correlation model and a monitoring model by mechanical operation state data to early warn the fault of the mechanical parts of the wind power equipment, so as to avoid the problem that the existing technology can only notify the fault after the mechanical parts are damaged.

[0005] To achieve the above purpose, the embodiment of the present application provides a variable pitch system fault early warning method, comprising:

[0006] According to the motor torque data in the mechanical operation data, the equivalent equivalent torque of each blade at each time interval is calculated;

[0007] According to the equivalent equivalent torque of each blade, the correlation parameter and the significance parameter between adjacent blades are calculated;

[0008] According to the equivalent equivalent torque, the motor current data, the motor temperature data, the pitch angle data and the speed data in the mechanical operation data, the Mahalanobis distance is calculated;

[0009] When judging mechanical failure based on the Mahalanobis distance, the correlation parameter and the significance parameter, outputting a failure warning.

[0010] Optionally, the calculating the equivalent equivalent torque of each blade in each time interval according to the motor torque data in the mechanical operation data comprises:

[0011] inputting the motor torque data of each blade into a first model, and calculating the equivalent equivalent torque of each blade in each time interval according to the first model; wherein the expression of the first model is:

[0012]

[0013] wherein, T j-eq-l is the equivalent equivalent torque of the jth blade in the lth time interval, p is a mechanical fatigue parameter, T j-l-i is the motor torque data of the jth blade at the ith time point in the lth time interval, t i is the time interval between the ith time point and the (i-1)th time point.

[0014] Optionally, the calculating the correlation parameter and the significance parameter between adjacent blades according to the equivalent equivalent torque of each blade comprises:

[0015] inputting the equivalent equivalent torque of each blade into a second model, and calculating the correlation parameter between adjacent blades according to the second model;

[0016] determining the significance parameter according to the correlation parameter;

[0017] wherein the expression of the second model is:

[0018]

[0019] wherein, r j is the correlation parameter between the jth blade and the (j+1)th blade, n is the number of equivalent equivalent torques, T j-eq-l is the equivalent equivalent torque of the jth blade in the lth time interval, T j+1-eq-l is the equivalent equivalent torque of the (j+1)th blade in the lth time interval, is the average value of the equivalent equivalent torque of the jth blade, is the average value of the equivalent equivalent torque of the (j+1)th blade.

[0020] Optionally, the determining the significance parameter according to the correlation parameter comprises:

[0021] inputting the correlation parameter into a third model, and calculating an auxiliary parameter according to the third model;

[0022] determining the saliency parameter according to an auxiliary parameter distribution table;

[0023] wherein an expression of the third model is:

[0024]

[0025] wherein h j is the auxiliary parameter between the jth paddle and the j+1th paddle, r j is the correlation parameter between the jth paddle and the j+1th paddle, and n is the number of equivalent torques.

[0026] Optionally, the calculation of the Mahalanobis distance according to the equivalent torque and the motor current data, the motor temperature data, the pitch angle data and the speed data in the mechanical operation data comprises:

[0027] constructing a data matrix according to the equivalent torque and the motor current data, the motor temperature data, the pitch angle data and the speed data in the mechanical operation data;

[0028] reconstructing the data matrix to obtain a reconstruction matrix by dimension reduction;

[0029] calculating the Mahalanobis distance according to the reconstruction matrix.

[0030] Optionally, the reconstruction of the data matrix to obtain a reconstruction matrix by dimension reduction comprises:

[0031] standardizing the data matrix to obtain a standardized matrix;

[0032] calculating a covariance matrix of the standardized matrix, and obtaining an eigenvalue matrix and an eigenvector matrix of the covariance matrix;

[0033] performing PCA dimension reduction on the data matrix according to the eigenvalue matrix and the eigenvector matrix to obtain the reconstruction matrix.

[0034] Optionally, the calculation of the Mahalanobis distance according to the reconstruction matrix comprises:

[0035] inputting the reconstruction matrix into a fourth model, and calculating the Mahalanobis distance according to the fourth model, wherein an expression of the fourth model is:

[0036]

[0037] wherein, is the Mahalanobis distance, For the reconstruction matrix, T is the transpose of the matrix, S is the covariance matrix of the reconstruction matrix, and μ' is the mean value of the reconstruction matrix.

[0038] To achieve the above object, the application further provides a variable pitch system fault early warning device, comprising:

[0039] A first calculation module is configured to calculate the equivalent equivalent torque of each blade at each time interval according to the motor torque data in the mechanical operation data.

[0040] A second calculation module is configured to calculate the correlation parameter and the significance parameter between adjacent blades according to the equivalent equivalent torque of each blade.

[0041] A third calculation module is configured to calculate the Mahalanobis distance according to the equivalent equivalent torque and the motor current data, motor temperature data, pitch angle data and speed data in the mechanical operation data.

[0042] A judgment module is configured to output a fault warning when a mechanical fault is judged based on the Mahalanobis distance, the correlation parameter and the significance parameter.

[0043] To achieve the above object, the application further provides a variable pitch system fault early warning device, comprising:

[0044] A memory is configured to store a computer program.

[0045] A processor is configured to execute the computer program to realize the variable pitch system fault early warning method according to any one of the above.

[0046] To achieve the above object, the application further provides a computer readable storage medium, wherein the computer readable storage medium stores computer executable instructions, and the computer executable instructions are executed by a processor to realize the variable pitch system fault early warning method according to any one of the above.

[0047] It can be seen that the method calculates the equivalent equivalent torque of each blade at each time interval according to the motor torque data in the mechanical operation data, calculates the correlation parameter and the significance parameter between adjacent blades according to the equivalent equivalent torque of each blade, calculates the Mahalanobis distance according to the equivalent equivalent torque and the motor current data, motor temperature data, pitch angle data and speed data in the mechanical operation data, and outputs a fault warning when a mechanical fault is judged based on the Mahalanobis distance, the correlation parameter and the significance parameter. The problem that the prior art can only perform fault notification after the mechanical components are damaged and the power generation function of the unit is faulty is avoided. BRIEF DESCRIPTION OF DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description only constitute a part of the embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of the provided drawings.

[0049] Figure 1 A flow chart of a variable pitch system fault early warning method provided for the embodiments of the present application;

[0050] Figure 2 A specific embodiment diagram of a variable pitch system fault early warning method provided for the embodiments of the present application;

[0051] Figure 3 A structural block diagram of a variable pitch system fault device provided for the embodiments of the present application. DETAILED DESCRIPTION

[0052] The technical solutions in the embodiments of the present application will be described clearly and completely in the following with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments only constitute a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0053] The variable pitch system undertakes important tasks such as absorbing wind energy by the unit, controlling power and speed balance, aerodynamic brake, facing complex wind conditions such as extreme turbulence and high wind shear, and protecting the unit in the process of operation. The mechanical connection in the variable pitch system mainly consists of a hub, a variable pitch motor, a variable pitch reducer, a variable pitch bearing and a blade. The reliability of the mechanical components of the variable pitch system is greatly challenged under the variable wind conditions and high-frequency load changes. When the mechanical components in the variable pitch system have serious faults, the unit is prone to major safety accidents, which is not conducive to safe operation. For the mechanical components in the variable pitch system, except that the variable pitch motor torque can be directly measured, there is no monitoring sensor for other components, and there is a lack of effective and comprehensive monitoring means. Once the mechanical components in the system have hidden troubles due to excessive load, it is difficult for the operation and maintenance personnel to discover and eliminate the hidden troubles in time. The present application constructs a correlation model and a monitoring model by collecting multiple mechanical operating state data to perform fault early warning on the variable pitch system.

[0054] The following will be described in combination with Figure 1 , Figure 1 A flow chart of a variable pitch system fault early warning method provided for the embodiments of the present application, which can include:

[0055] S101: Calculate the equivalent equivalent torque of each blade in each time interval according to the motor torque data in the mechanical operation data.

[0056] The embodiment does not limit the acquisition method of the mechanical operation data used, which can generally be data collected from a wind turbine sensor. The embodiment also does not limit the time interval between each data sample. In general, in order to make the prediction result more accurate without wasting computing resources, the sensor can collect mechanical operation data at a level of seconds. The embodiment also does not limit the specific content of the collected mechanical operation data, which can at least include motor torque data, motor current data, motor temperature data, pitch angle data, and speed data. Further, after obtaining the mechanical operation data, the mechanical operation data can be preprocessed to eliminate invalid data to make the result more accurate, such as eliminating power-off data, i.e., data with motor current of 0; eliminating communication error data, i.e., data with motor torque, motor current, and motor temperature remaining unchanged. The data collection and data calculation in the embodiment can be completed by an edge collection device and an edge server.

[0057] The equivalent equivalent torque data in the embodiment can be calculated from the motor torque data in each time interval. The embodiment does not limit the size of the time interval, which can be calculated using motor torque data for each day or for each seven days. Further, the embodiment does not limit the number of equivalent equivalent torque data required for each correlation parameter and significance parameter calculation, which is generally not less than 30. The embodiment does not limit the number of blades, which can generally be 3. The time point in the embodiment can refer to the time point of collecting each data in each time interval. The time point corresponding to the first motor torque data collected in each time interval can be the first time point, and the time point corresponding to the second motor torque data collected in each time interval can be the second time point. For example, if the time interval is one day and the motor torque data is collected every second in each day, the first second in each day is the first time point, and the second second is the second time point.

[0058] In the embodiment, the motor torque data of each blade can be input into the first model to calculate the equivalent equivalent torque of each blade in each time interval. The expression of the first model can be:

[0059]

[0060] In the formula, T j-eq-l is the equivalent equivalent torque of the jth blade in the lth time interval, p is the mechanical fatigue parameter, T j-l-i is the motor torque data of the jth blade at the ith time point in the lth time interval, ti is the time interval between the i-th time point and the i-1-th time point.

[0061] S102: Calculate the correlation parameter and the significance parameter between adjacent blades according to the equivalent equivalent torque of each blade.

[0062] In this embodiment, after the equivalent equivalent torque of each blade in each time interval is calculated, a correlation model can be constructed according to the equivalent equivalent torque to obtain the correlation parameter and the significance parameter between adjacent blades.

[0063] This embodiment does not limit the calculation method of the correlation parameter and the significance parameter between adjacent blades, and generally the equivalent equivalent torque of each blade can be input into a second model to calculate the correlation parameter between adjacent blades according to the second model; wherein the expression of the second model is:

[0064]

[0065] In the formula, r j is the correlation parameter between the j-th blade and the j+1-th blade, n is the number of equivalent equivalent torques, T j-eq-l is the equivalent equivalent torque of the j-th blade in the l-th time interval, T j+1-eq-l is the equivalent equivalent torque of the j+1-th blade in the l-th time interval, is the average value of the equivalent equivalent torque of the j-th blade, is the average value of the equivalent equivalent torque of the j+1-th blade.

[0066] After the correlation parameter r j is calculated, the significance parameter p j between the corresponding adjacent two blades can be calculated according to r j -value, and the specific calculation process can be to input the correlation parameter into a third model to calculate the auxiliary parameter according to the third model; wherein the expression of the third model is:

[0067]

[0068] In the formula, h j is the auxiliary parameter between the j-th blade and the j+1-th blade, r j is the correlation parameter between the j-th blade and the j+1-th blade, and n is the number of equivalent equivalent torques.

[0069] This embodiment can determine the significance parameter through the auxiliary parameter, and this embodiment does not limit the specific process of determining the significance parameter through the auxiliary parameter, and generally the auxiliary parameter h jThen, the corresponding auxiliary parameter and saliency parameter table can be consulted to determine the saliency parameter p between every two blades j -value.

[0070] S103: Calculate Mahalanobis distance according to equivalent equivalent torque and motor current data, motor temperature data, pitch angle data and speed data in mechanical operation data.

[0071] In this embodiment, a monitoring model can be constructed according to equivalent equivalent torque and motor current data, motor temperature data, pitch angle data and speed data in mechanical operation data to calculate the Mahalanobis distance of the blade system. This embodiment does not limit the parameters used to calculate the Mahalanobis distance. When other mechanical operation parameters can also calculate the Mahalanobis distance, other parameters can be introduced into the embodiment of the application.

[0072] This embodiment does not limit the specific calculation method of the Mahalanobis distance. Generally, the equivalent equivalent torque and the motor current data, the motor temperature data, the pitch angle data and the speed data in the mechanical operation data can be constructed into a data matrix:

[0073]

[0074] In the formula, is the constructed data matrix, a represents the number of blades, T a-eq , I a , T a-emp , θ a , v a respectively represent the equivalent equivalent torque data group, the motor current data group, the motor temperature data group, the pitch angle data group and the speed data group of the a-th blade.

[0075] The motor current data group, the motor temperature data group, the pitch angle data group and the speed data group in this embodiment are composed of data at each time point in all time intervals, that is, each data group contains n*i data. Since each time interval corresponds to only one equivalent equivalent torque, in order to make the number of equivalent equivalent torque data group correspond to that of other data groups, the value of the n-th group of all i equivalent equivalent torques in the equivalent equivalent torque data group is assigned to the n-th equivalent equivalent torque data.

[0076] The above data group can be a column vector, that is, the data matrix is a matrix of n*i rows and 5*a columns. In this embodiment, the data matrix is constructed as follows: The standardization process is not limited in the embodiment, and a Z-Score standardization method can be generally used. The Z-Score is a name of a standardization method, each data in each data group is subtracted by the average value of the corresponding data group and then divided by the standard deviation of the corresponding data group, that is:

[0077]

[0078] In the formula, is a data matrix constructed in the embodiment, is a standardized matrix after the standardization process, μ represents the average value of each column data group in the data matrix , and σ represents the standard deviation of each column data group in the data matrix .

[0079] In the embodiment, after the standardized matrix is obtained, a covariance matrix of the standardized matrix can be obtained.

[0080]

[0081] In the formula, is the covariance matrix, n is the number of each data group, is the standardized matrix, T represents the transpose of the matrix. According to the covariance matrix , an eigenvalue matrix and an eigenvector matrix are obtained. m is the rank of the matrix, and k is a parameter to be obtained.

[0082]

[0083] The PCA (Principal Component Analysis) dimension reduction process in the embodiment can be that the first k eigenvectors with an eigenvalue accumulation greater than a preset proportion are selected. The preset proportion is not limited in the embodiment, and can be generally 90%. The process of obtaining k can be as follows:

[0084]

[0085] In the formula, λ i is an eigenvalue in the eigenvalue matrix, m is the rank of the matrix, and k is a parameter to be obtained.

[0086] After the value of k is obtained, the first k data in the eigenvector matrix are selected to form a dimension reduction matrix . According to the matrix , a reconstruction matrix The calculation of the reconstruction matrix can be performed according to the following formula:

[0087]

[0088] In the formula, is a reconstruction matrix, is a normalization matrix, is a dimension reduction matrix, and T is the transpose of the matrix.

[0089] After the reconstruction matrix is calculated, the reconstruction matrix can be input into a fourth model to calculate the Mahalanobis matrix, where the fourth model can be:

[0090]

[0091] In the formula, is a Mahalanobis distance, is a reconstruction matrix, T is the transpose of the matrix, S is the covariance matrix of the reconstruction matrix, and μ' is the mean value of the reconstruction matrix.

[0092] S104: When the mechanical failure is judged based on the Mahalanobis distance, the correlation parameter, and the significance parameter, output a failure warning.

[0093] In this embodiment, after the correlation parameter is obtained by constructing the correlation model and the Mahalanobis distance is obtained by constructing the monitoring model, whether the machine fails can be judged according to the Mahalanobis distance, the correlation parameter, and the significance parameter.

[0094] In this embodiment, when the significance parameter is less than the first threshold value and the correlation parameter is greater than the second threshold value, it can be judged that the running state between the two blades is normal. This embodiment does not limit the setting method of the first threshold value and the second threshold value, nor does it limit the specific numerical value. Generally, the first threshold value can be 0.05, and the second threshold value can be 0.75. Further, this embodiment does not limit the way of judging the system running failure through the running state between the two blades. It can be that when any two blades run abnormally, the entire system is judged to run abnormally, or when all adjacent blades run abnormally, the entire system is judged to run abnormally. For example, when the system has three blades, the correlation parameters between the first blade and the third blade are r1 and p1-value; the correlation parameters between the second blade and the first blade are r2 and p2-value; and the correlation parameters between the third blade and the second blade are r3 and p3-value. In this embodiment, it can be that when the first blade and the third blade, the second blade and the first blade, and the third blade and the second blade run abnormally at the same time, the system is judged to run abnormally, or it can be that when any one group runs abnormally, the system is judged to run abnormally.

[0095] In the embodiment, when the Mahalanobis distance obtained exceeds a third threshold value, it is determined that the entire system is in an abnormal state. The embodiment does not limit the setting manner and specific value of the third threshold value, which can be obtained according to a predetermined confidence probability 1-z combined with a chi-square distribution quantile table.

[0096] The embodiment does not limit the output manner of the fault warning. The fault warning can be output when the system running fault is determined through the correlation parameter and the significance parameter and the entire system running fault is determined when the system running fault is determined through the Mahalanobis distance. The fault warning can also be output when the entire system running fault is determined through the correlation parameter and the significance parameter or the Mahalanobis distance.

[0097] The embodiment constructs a correlation model to obtain the correlation parameter and the significance parameter through the mechanical running state data, constructs a monitoring model to obtain the Mahalanobis distance, analyzes the fault condition of the variable pitch system according to the correlation parameter, the significance parameter and the Mahalanobis distance output by the model, and pre-warns the fault, thereby avoiding the problem that the power generation function of the unit is faulty because the fault notification is performed only after the mechanical component is damaged in the prior art.

[0098] The following will be described in detail Figure 2 , Figure 2 The embodiment provides a specific embodiment of a variable pitch system fault warning method, which can include the following steps.

[0099] 1. Calculate the equivalent equivalent torque of each blade at each time interval according to the motor torque data in the mechanical running data.

[0100] 2. Calculate the correlation parameter and the significance parameter between adjacent blades according to the equivalent equivalent torque of each blade.

[0101] 3. Construct a data matrix according to the equivalent equivalent torque, the motor current data, the motor temperature data, the pitch angle data and the speed data.

[0102] 4. Obtain the eigenvalue matrix and the eigenvector matrix of the covariance matrix of the data matrix.

[0103] 5. Perform PCA dimension reduction on the data matrix according to the eigenvalue matrix and the eigenvector matrix to obtain a reconstruction matrix.

[0104] 6. Perform PCA dimension reduction according to the eigenvalue matrix and the eigenvector matrix to obtain a reconstruction matrix.

[0105] 7. When the system is determined to be abnormal according to the correlation parameter and the significance parameter and the system is determined to be abnormal through the Mahalanobis distance, output a fault warning.

[0106] The following description is made in connection with Figure 3 , Figure 3 A structural block diagram of a pre-warning device of a variable pitch system is provided for an embodiment of the present application, which can include:

[0107] A first calculation module 100 is configured to calculate equivalent equivalent torque of each blade at each time interval according to motor torque data in mechanical operation data;

[0108] A second calculation module 200 is configured to calculate correlation parameters and significance parameters between adjacent blades according to the equivalent equivalent torque of each blade;

[0109] A third calculation module 300 is configured to calculate Mahalanobis distance according to the equivalent equivalent torque and motor current data, motor temperature data, pitch angle data and speed data in the mechanical operation data;

[0110] A judgment module 400 is configured to output a fault pre-warning when judging mechanical failure based on the Mahalanobis distance, the correlation parameters and the significance parameters.

[0111] Based on the above embodiments, the present application constructs a correlation model to obtain correlation parameters and significance parameters by using mechanical operation state data, constructs a monitoring model to obtain Mahalanobis distance, analyzes the failure condition of the variable pitch system according to the correlation parameters, the significance parameters and the Mahalanobis distance output by the model, and pre-warns the failure, thereby avoiding the problem that the power generation function of the unit fails due to the fact that the prior art can only notify the failure after the mechanical components are damaged.

[0112] Based on the above embodiments, the first calculation module 100 can include:

[0113] A first model unit is configured to input motor torque data of each blade into a first model, and calculate equivalent equivalent torque of each blade at each time interval according to the first model; wherein the expression of the first model is:

[0114]

[0115] In the formula, T j-eq-l is the equivalent equivalent torque of the jth blade at the lth time interval, p is a mechanical fatigue parameter, T j-l-i is the motor torque data of the jth blade at the lth time interval and the ith time point, t i is the time interval between the ith time point and the (i-1)th time point.

[0116] Based on the above embodiments, the second calculation module 200 can include:

[0117] a second model unit configured to input the equivalent equivalent torques of each of the blades into a second model, and calculate the correlation parameters between adjacent blades according to the second model;

[0118] a significance unit configured to determine the significance parameters according to the correlation parameters;

[0119] wherein an expression of the second model is:

[0120]

[0121] wherein r j is the correlation parameter between the jth blade and the j+1th blade, n is a number of the equivalent equivalent torques, T j-eq-l is the equivalent equivalent torque of the jth blade in the lth time interval, T j+1-eq-l is the equivalent equivalent torque of the j+1th blade in the lth time interval, is an average of the equivalent equivalent torques of the jth blade, is an average of the equivalent equivalent torques of the j+1th blade.

[0122] According to the above embodiments, the significance unit can include:

[0123] a third model sub-unit configured to input the correlation parameters into a third model, and calculate auxiliary parameters according to the third model;

[0124] a determination sub-unit configured to determine the significance parameters according to an auxiliary parameter distribution table;

[0125] wherein an expression of the third model is:

[0126]

[0127] wherein h j is the auxiliary parameter between the jth blade and the j+1th blade, r j is the correlation parameter between the jth blade and the j+1th blade, n is a number of the equivalent equivalent torques.

[0128] According to the above embodiments, the third calculation module 300 can include:

[0129] a construction unit configured to construct a data matrix according to the equivalent equivalent torques and the motor current data, the motor temperature data, the pitch angle data and the speed data in the mechanical operation data;

[0130] a reconstruction unit configured to perform dimensionality reduction reconstruction on the data matrix to obtain a reconstruction matrix;

[0131] a calculation unit configured to calculate the Mahalanobis distance according to the reconstruction matrix.

[0132] According to the above embodiments, the reconstruction unit can include:

[0133] a normalization sub-unit configured to normalize the data matrix to obtain a normalized matrix;

[0134] a covariance sub-unit configured to calculate a covariance matrix of the normalized matrix, and obtain an eigenvalue matrix and an eigenvector matrix of the covariance matrix;

[0135] a dimension reduction sub-unit configured to perform PCA dimension reduction on the data matrix according to the eigenvalue matrix and the eigenvector matrix to obtain the reconstruction matrix.

[0136] According to the above embodiments, the calculation unit can include:

[0137] a fourth model sub-unit configured to input the reconstruction matrix into a fourth model, and calculate the Mahalanobis distance according to the fourth model, wherein an expression of the fourth model is:

[0138]

[0139] wherein, is the Mahalanobis distance, is the reconstruction matrix, T is a transpose of a matrix, S is a covariance matrix of the reconstruction matrix, and μ' is a mean value of the reconstruction matrix.

[0140] According to the above embodiments, the present application further provides a variable pitch system fault early warning device, which can include a memory and a processor, wherein the memory has a computer program stored therein, and the processor can implement the steps provided in the above embodiments when calling the computer program in the memory. Of course, the device can also include various necessary network interfaces, power supplies and other components, etc.

[0141] The present application further provides a computer readable storage medium having a computer program stored thereon, which can implement the variable pitch system fault early warning method provided in the embodiments of the present application when executed by a terminal or a processor. The storage medium can include: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various media that can store program codes.

[0142] Various embodiments are described herein with reference to the following items, which are presented by way of example to provide context for various embodiments of the present application. 1. A method for wireless communication, comprising: receiving a first signal from a first wireless communication device; receiving a second signal from a second wireless communication device; and determining a location of the first wireless communication device based at least in part on the first signal and the second signal.

[0143] Finally, it should be noted that, in the specification, relational terms such as first and second, and the like, can be used solely to distinguish one entity or action from another entity or action without necessarily implying any actual relationship or order between such entities or actions. Moreover, the terms "comprises", "comprising", or any other variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can also include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element preceded by "comprises a" does not, without more constraints, foreclose the existence of additional identical elements in the process, method, article, or apparatus that comprises the recited element.

[0144] The above describes in detail a variable pitch system fault early warning method, device, equipment and storage medium provided by the present application. The principle and implementation mode of the present application are described by applying specific examples. The above embodiment description is only used to help understand the method of the present application and its core idea. Meanwhile, for those skilled in the art, according to the idea of the present application, the specific implementation mode and application range will be changed. In summary, the content of the specification should not be understood as a limitation of the present application.

Claims

1. A method of failure early warning for a variable pitch system, characterized in that, The method comprises the following steps: calculating equivalent equivalent torque of each blade at each time interval according to motor torque data in mechanical operation data; calculating correlation parameters and significance parameters between adjacent blades according to the equivalent equivalent torque of each blade; calculating Mahalanobis distance according to the equivalent equivalent torque and motor current data, motor temperature data, pitch angle data and speed data in the mechanical operation data; outputting a fault warning when judging mechanical failure based on the Mahalanobis distance, the correlation parameters and the significance parameters; The method for calculating equivalent equivalent torque of each blade at each time interval according to motor torque data in mechanical operation data comprises the following steps: inputting motor torque data of each blade into a first model to calculate equivalent equivalent torque of each blade at each time interval according to the first model; wherein the expression of the first model is: ； wherein, is the equivalent equivalent torque of the jth paddle in the lth time interval, p is a mechanical fatigue parameter, is the motor torque data of the jth paddle at the ith time point in the lth time interval, is the time interval between the ith time point and the i-1th time point.

2. The pre-failure warning method for a pitch system according to claim 1, characterized in that, The method for calculating correlation parameters and significance parameters between adjacent blades according to the equivalent equivalent torque of each blade comprises the following steps: inputting the equivalent equivalent torque of each blade into a second model to calculate the correlation parameters between adjacent blades according to the second model; determining the significance parameters according to the correlation parameters; wherein the expression of the second model is: ； wherein is the correlation parameter between the jth blade and the j+1th blade, n is the number of the equivalent equivalent torques, is the equivalent equivalent torque of the jth blade in the lth time interval, is the equivalent equivalent torque of the j+1th blade in the lth time interval, is the average of the equivalent equivalent torques of the jth blade, is the average of the equivalent equivalent torques of the j+1th blade.

3. The pre-failure warning method for a pitch system according to claim 2, characterized in that, The method for determining the significance parameters according to the correlation parameters comprises the following steps: inputting the correlation parameters into a third model to calculate auxiliary parameters according to the third model; determining the significance parameters according to an auxiliary parameter distribution table; wherein the expression of the third model is: ； wherein is the auxiliary parameter between the jth blade and the j+1th blade, is the correlation parameter between the jth blade and the j+1th blade, n is the number of the equivalent equivalent torques.

4. The pre-failure warning method for a pitch system according to claim 1, characterized in that, The method for calculating Mahalanobis distance according to the equivalent equivalent torque and motor current data, motor temperature data, pitch angle data and speed data in the mechanical operation data comprises the following steps: constructing a data matrix according to the equivalent equivalent torque and the motor current data, the motor temperature data, the pitch angle data and the speed data in the mechanical operation data; reconstructing the data matrix to obtain a reconstruction matrix by dimension reduction; calculating the Mahalanobis distance according to the reconstruction matrix.

5. The pre-failure warning method for a variable pitch system according to claim 4, characterized in that, The method for reconstructing the data matrix to obtain a reconstruction matrix by dimension reduction comprises the following steps: standardizing the data matrix to obtain a standardized matrix; calculating the covariance matrix of the standardized matrix, and obtaining the eigenvalue matrix and eigenvector matrix of the covariance matrix; performing PCA dimension reduction on the data matrix according to the eigenvalue matrix and the eigenvector matrix to obtain the reconstruction matrix.

6. The pre-failure warning method for a variable pitch system according to claim 4, characterized in that, The method for calculating the Mahalanobis distance according to the reconstruction matrix comprises the following steps: inputting the reconstruction matrix into a fourth model to calculate the Mahalanobis distance according to the fourth model, wherein the expression of the fourth model is: ； wherein is the Mahalanobis distance, is the reconstruction matrix, T is the transpose of a matrix, S is the covariance matrix of the reconstruction matrix, is the mean of the reconstruction matrix.

7. A pre-failure warning device for a variable pitch system, characterized in that The method comprises the following steps: The first calculation module is configured to calculate equivalent equivalent torque of each blade at each time interval according to motor torque data in mechanical operation data; The second calculation module is configured to calculate correlation parameters and significance parameters between adjacent blades according to the equivalent equivalent torque of each blade; a third calculation module, configured to calculate a Mahalanobis distance according to the equivalent equivalent torque and motor current data, motor temperature data, pitch angle data and speed data in the mechanical operation data; a judgment module, configured to output a fault warning when judging a mechanical fault based on the Mahalanobis distance, the correlation parameter and the significance parameter; the equivalent equivalent torque of each of the blades in each time interval is calculated according to the motor torque data in the mechanical operation data, comprising: the motor torque data of each of the blades is input into a first model, and the equivalent equivalent torque of each of the blades in each time interval is calculated according to the first model; wherein the expression of the first model is: ； wherein is the equivalent equivalent torque of the jth paddle in the lth time interval, p is a mechanical fatigue parameter, is the motor torque data of the jth paddle at the ith time point in the lth time interval, is the time interval between the ith time point and the i-1th time point.

8. A pre-failure warning device for a variable pitch system, characterized in that comprising: a memory, configured to store a computer program; a processor, configured to implement the variable pitch system fault warning method according to any one of claims 1 to 6 when executing the computer program.

9. A computer-readable storage medium, characterized in that, The computer readable storage medium stores computer executable instructions, and the computer executable instructions are executed by the processor to implement the variable pitch system fault warning method according to any one of claims 1 to 6.

Images