Method and device for identifying deterioration trend of hydraulic station of wind turbine generator

Abstract

The invention discloses a wind turbine generator hydraulic station degradation trend identification method and device, and the method comprises the steps: obtaining monitoring data of a plurality of influence factors in a preset time period, and forming time series data; determining an abnormal frequency of each influence factor serving as an evaluation index based on the time sequence data; normalizing each evaluation index to obtain an initial score of each evaluation index; determining a comprehensive score according to the initial score and the fusion weight; taking the time sequence data and the comprehensive score as input and expected output respectively, and training a comprehensive score model; and monitoring the hydraulic station in real time by using the comprehensive scoring model so as to identify a degradation trend and perform early warning according to a preset threshold value. According to the technical scheme, the health state of the hydraulic station of the wind turbine generator can be efficiently reflected at low cost, the degradation trend is recognized, and early warning is conducted.

Description

Technical Field

[0001] The present application relates to the field of neural network technology, and in particular to a method and device for identifying degradation trends of a hydraulic station of a wind turbine generator set. Background Art

[0002] The normal operation of a wind turbine depends on the health of its various subsystems, among which the hydraulic station is a key component to ensure the normal operation of the yaw and braking systems. Failure of the hydraulic station may cause the wind turbine to be unable to accurately align with the wind direction, or even cause a safety accident.

[0003] Traditional preventive maintenance relies on regular inspections and empirical judgments, which is not only costly but also fails to reflect the health status of equipment in real time. Summary of the invention

[0004] The embodiment of the present application solves the technical problem in the prior art that preventive maintenance is not only costly but also fails to reflect the health status of the equipment in real time by providing a method and device for identifying the degradation trend of a hydraulic station of a wind turbine generator set.

[0005] In a first aspect, an embodiment of the present application provides a method for identifying degradation trends of a hydraulic station of a wind turbine, the method comprising: acquiring monitoring data of a preset time period of multiple influencing factors to form time series data; based on the time series data, determining the abnormal frequency of each of the influencing factors as an evaluation indicator; normalizing each of the evaluation indicators to obtain an initial score for each of the evaluation indicators; determining a comprehensive score based on the initial score and a fusion weight; using the time series data and the comprehensive score as input and expected output, respectively, to train the comprehensive scoring model; and using the comprehensive scoring model to monitor the hydraulic station in real time to identify degradation trends and issue early warnings based on preset thresholds.

[0006] In combination with the first aspect, in a possible implementation method, determining the abnormal frequency of each of the influencing factors as an evaluation indicator based on the time series data includes: using a sliding window of a preset step size and size in the time series data to determine the abnormal frequency of each of the influencing factors in each sliding window.

[0007] In combination with the first aspect, in a possible implementation method, the multiple influencing factors include a single state factor and multiple state factors; determining the abnormal frequency of each of the influencing factors in each of the sliding windows includes: counting the number of abnormal occurrences of the monitoring data of the single state factor in each of the sliding windows, and determining the abnormal frequency of the single state factor; clustering the monitoring data with the multiple state factors, determining the abnormal boundary of each cluster, and determining the abnormal frequency of the multiple state factors based on the sum of the number of abnormal occurrences of each cluster in each of the sliding windows.

[0008] In combination with the first aspect, in a possible implementation method, the fusion weight is determined by the following steps: using the hierarchical analysis method and the entropy method to respectively calculate the first weight and the second weight of each of the evaluation indicators; calculating the fusion weight according to the preset weight balancing factor, the first weight and the second weight.

[0009] In combination with the first aspect, in a possible implementation manner, after forming the time series data, the method further includes: filling in missing values ​​and deleting outliers.

[0010] In combination with the first aspect, in a possible implementation manner, the multiple influencing factors include main system hydraulic pressure, yaw pressure, hydraulic oil temperature, wind speed, wind direction, power and hydraulic pump feedback time.

[0011] In combination with the first aspect, in a possible implementation, the comprehensive scoring model is a recurrent neural network.

[0012] In a second aspect, an embodiment of the present application provides a device for identifying a degradation trend of a hydraulic station of a wind turbine, the device comprising: a data acquisition module for acquiring monitoring data of a preset time period of multiple influencing factors to form time series data; an abnormality statistics module for determining the abnormal frequency of each of the influencing factors as an evaluation indicator based on the time series data; a normalization module for normalizing each of the evaluation indicators to obtain an initial score for each of the evaluation indicators; a comprehensive scoring module for determining a comprehensive score based on the initial score and a fusion weight; a model training module for training the comprehensive scoring model using the time series data and the comprehensive score as input and expected output, respectively; a degradation identification module for using the comprehensive scoring model to monitor the hydraulic station in real time to identify degradation trends and issue warnings based on preset thresholds.

[0013] In a third aspect, an embodiment of the present application provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method for identifying degradation trends of a hydraulic station of a wind turbine set as described in the first aspect or any possible implementation method of the first aspect is implemented.

[0014] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, which stores computer-readable instructions. When the computer-readable instructions are executed by a processor, the method for identifying degradation trends of a hydraulic station of a wind turbine generator system as described in the first aspect or any possible implementation method of the first aspect is implemented.

[0015] The technical solution provided in the embodiments of the present application has at least the following technical effects: The embodiment of the present application provides a method for identifying the deterioration trend of a hydraulic station of a wind turbine. The method uses the abnormal frequency of each influencing factor as an evaluation index, obtains the initial score of each evaluation index after normalizing each evaluation index, and determines the comprehensive score according to the initial score and the fusion weight, and uses the comprehensive scoring model trained by the time series data and the comprehensive score to monitor the hydraulic station to identify the deterioration trend and issue an early warning according to a preset threshold. The method for identifying the deterioration trend of a hydraulic station of a wind turbine solidifies the mapping relationship between the monitoring data and the comprehensive score through a comprehensive scoring model. The comprehensive scoring model can obtain the health status of the hydraulic station in real time by monitoring the hydraulic station, and then can monitor the deterioration trend of the yaw hydraulic station, and issue an early warning when the comprehensive score exceeds the preset threshold. After using this method to monitor the hydraulic station of a wind turbine, it is no longer necessary for operation and maintenance personnel to regularly arrive at the hydraulic station for inspection, saving labor costs, and monitoring the health status through a comprehensive scoring model avoids the subjectivity of empirical judgment, can more accurately judge the health status of the hydraulic station, and can identify deterioration trends and early warnings. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments of the present application will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0017] Figure 1 A flow chart of a method for identifying a deterioration trend of a hydraulic station of a wind turbine provided in an embodiment of the present application; Figure 2 A flowchart for determining the abnormal frequency of each influencing factor provided in an embodiment of the present application; Figure 3 A flowchart for determining fusion weights provided in an embodiment of the present application; Figure 4 A schematic diagram of a wind turbine hydraulic station degradation trend identification device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0018] The normal operation of a wind turbine depends on the health of its various subsystems, among which the hydraulic station is a key component to ensure the normal operation of the yaw and braking systems. Failure of the hydraulic station may cause the wind turbine to be unable to accurately align with the wind direction, or even cause a safety accident. Traditional preventive maintenance relies on regular inspections and empirical judgments, which is not only costly, but also unable to reflect the health status of the equipment in real time.

[0019] The embodiments of the present application provide a method and device for identifying the degradation trend of a hydraulic station of a wind turbine generator set, which can reflect the health status of the hydraulic station of a wind turbine generator set in a low-cost and efficient manner, identify the degradation trend and issue an early warning.

[0020] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0021] Below, some terms in the embodiments of the present application are explained to facilitate understanding by those skilled in the art.

[0022] Recurrent Neural Network (RNN) is a neural network model that can process sequence data. Recurrent Neural Network belongs to deep learning neural network. Unlike traditional feedforward neural network, RNN retains a hidden state when processing each input, and the hidden state will be passed to the next time step so that the model can remember the previous information.

[0023] Sliding Window is a technology used to manage and process data streams. It defines a window on the data stream, slides the window by a step size each time, and processes while covering different data fragments, thereby achieving efficient data processing, transmission control and resource management.

[0024] The analytic hierarchy process (AHP) is a decision-making method that decomposes elements related to decision-making into levels such as goals, criteria, and plans, and conducts qualitative and quantitative analysis on this basis. According to the nature of the problem and the overall goal to be achieved, the AHP decomposes the problem into different components, and clusters and combines the factors at different levels according to the mutual influence and affiliation between the factors to form a multi-level analysis structure model, so that the problem is ultimately reduced to the determination of the relative importance of the lowest level (plans and measures for decision-making) relative to the highest level (overall goal) or the ranking of the relative order of superiority and inferiority.

[0025] The entropy method is an objective method for determining the weight of indicators. It is based on the concept of entropy in information theory and is used to measure the uncertainty or discreteness of indicator information. In a multi-indicator comprehensive evaluation system, the entropy method can assign weights according to the degree of variation of each indicator. Indicators with a large degree of variation have a greater impact on the comprehensive evaluation.

[0026] The embodiment of the present application provides a method for identifying a deterioration trend of a hydraulic station of a wind turbine generator set, the method comprising: Figure 1Steps 101 to 106 are shown.

[0027] Step 101: Acquire monitoring data of a preset time period of multiple influencing factors to form time series data.

[0028] The hydraulic station of the wind turbine is equipped with multiple types of sensors such as temperature sensors, hydraulic sensors, wind speed sensors, etc. These sensors are used to monitor the hydraulic station of the wind turbine, collect monitoring data of multiple influencing factors affecting the hydraulic station of the wind turbine in real time, and form time series data stored in chronological order.

[0029] Exemplarily, the multiple influencing factors include main system hydraulic pressure, yaw pressure, hydraulic oil temperature, wind speed, wind direction, power and hydraulic pump feedback time; when executing step 101, the monitoring data of the main system hydraulic pressure, the monitoring data of the yaw pressure, the monitoring data of the hydraulic oil temperature, the monitoring data of the wind speed, the monitoring data of the wind direction, the monitoring data of the power and the hydraulic pump feedback time are obtained to form time series data including the main system hydraulic pressure, yaw pressure, hydraulic oil temperature, wind speed, wind direction, power and hydraulic pump feedback time.

[0030] Step 102: Based on the time series data, determine the abnormal frequency of each influencing factor as an evaluation indicator.

[0031] The abnormal frequencies of these influencing factors are used as evaluation indicators to reflect the health status of the wind turbine hydraulic station. The greater the abnormal frequencies of these influencing factors, the worse the health status of the wind turbine hydraulic station.

[0032] Exemplarily, when executing step 102, the abnormal frequency of the main system hydraulic pressure, the abnormal frequency of the yaw pressure, the abnormal frequency of the hydraulic oil temperature, the abnormal frequency of the wind speed, the abnormal frequency of the wind direction, the abnormal frequency of the power and the abnormal frequency of the hydraulic pump feedback time in the time series data are determined.

[0033] Step 103: normalize each evaluation indicator to obtain an initial score for each evaluation indicator.

[0034] Different evaluation indicators may have different dimensions. By normalizing each evaluation indicator in step 103 , the dimension effects between different evaluation indicators can be eliminated, and the degradation trends of the evaluation indicators can be unified.

[0035] Negative normalization is used to process data sets where smaller values ​​indicate better results. Exemplarily, when executing step 103, each evaluation indicator may be negatively normalized to obtain an initial score for each evaluation indicator.

[0036] Step 104: Determine a comprehensive score based on the initial score and the fusion weight.

[0037] Specifically, when executing step 104, the product of the initial score of each evaluation indicator and the corresponding fusion weight is added to obtain a comprehensive score, as shown in the following formula: in, For the comprehensive rating, is the matrix of initial scores of the evaluation indicators, is the matrix of fusion weights.

[0038] The comprehensive score can reflect the overall health status of the hydraulic station of the wind turbine. The fusion weight is set before executing step 104, and can be set directly based on experience, or can be set using methods such as principal component analysis and hierarchical analysis.

[0039] Step 105: Use the time series data and the comprehensive score as input and expected output respectively to train the comprehensive score model.

[0040] After executing step 105, a trained comprehensive scoring module is obtained. The comprehensive scoring model can output a comprehensive score reflecting the health status of the hydraulic station based on the monitoring data obtained from various types of sensors. The comprehensive scoring model solidifies the mapping relationship between the monitoring data and the comprehensive score, and can effectively identify abnormal periodicity and trends in the monitoring data, thereby being able to monitor the degradation trend of the yaw hydraulic station.

[0041] For example, the comprehensive scoring model is a recurrent neural network, which enables it to learn the pattern of the evolution of the health status of the hydraulic station over time; the recurrent neural network uses its memory units to capture the long-term dependencies in the time series data, which can more comprehensively evaluate the overall health status of the hydraulic station, thereby improving the prediction accuracy. Of course, the comprehensive scoring model can also be other types of models such as long short-term memory artificial neural network (English: Long Short-Term Memory, referred to as: LSTM).

[0042] Step 106: Use the comprehensive scoring model to monitor the hydraulic station in real time to identify degradation trends and issue warnings based on preset thresholds.

[0043] The comprehensive scoring model can output a comprehensive score based on the monitoring data obtained by various types of sensors. The comprehensive score reflects the health status of the hydraulic station. The comprehensive scoring model can identify the degradation trend of the hydraulic station through changes in the comprehensive score and issue an early warning when the comprehensive score exceeds the preset threshold.

[0044] The method for identifying the deterioration trend of the hydraulic station of a wind turbine solidifies the mapping relationship between monitoring data and comprehensive scores through a comprehensive scoring model. The comprehensive scoring model can obtain the health status of the hydraulic station in real time by monitoring the hydraulic station, and then monitor the deterioration trend of the yaw hydraulic station, and issue an early warning when the comprehensive score exceeds the preset threshold. After using this method to monitor the hydraulic station of a wind turbine, it is no longer necessary for operation and maintenance personnel to regularly visit the hydraulic station for inspection, saving labor costs. In addition, by monitoring the health status through a comprehensive scoring model, the subjectivity of empirical judgment is avoided, and the health status of the hydraulic station can be judged more accurately, and degradation trends and early warnings can be identified.

[0045] In some embodiments of the present application, step 102 specifically includes: using a sliding window of a preset step size and size in the time series data, and determining the abnormal frequency of each influencing factor in each sliding window.

[0046] The sliding window can slide from the starting point to the end point of the time series data, and each time it slides, the abnormal frequency of each influencing factor in the sliding window is calculated, thereby obtaining a time series of abnormal frequencies.

[0047] Specifically, the multiple influencing factors include single state factors and multiple state factors. Exemplarily, the single state factors include hydraulic oil temperature, wind speed, wind direction, yaw pressure, and hydraulic pump feedback time; the multiple state factors include main system hydraulic pressure and power. Determining the abnormal frequency of each influencing factor in each sliding window specifically includes step 201 and step 202.

[0048] Step 201: Count the number of abnormal occurrences of monitoring data of a single state factor in each sliding window to determine the abnormal frequency of the single state factor.

[0049] A single state factor has a certain normal value range. When the monitoring data of a single state factor exceeds the normal value range, it is determined to be abnormal. By counting the number of abnormalities in the sliding window, the frequency of the abnormality can be determined.

[0050] Exemplarily, the normal range of the hydraulic oil temperature is 30°C-70°C; when the hydraulic oil temperature is less than 30°C, the hydraulic equipment is difficult to start; when the hydraulic oil temperature is greater than 70°C, the service life of the hydraulic equipment will be reduced; when executing step 201 to count the abnormal frequency of the hydraulic oil temperature, the sliding window is slid from the starting point to the end point of the time series data, and the number of abnormal hydraulic oil temperatures in the window is counted each time it slides, thereby determining the abnormal frequency of the hydraulic oil temperature.

[0051] Step 202: cluster the monitoring data with multiple state factors, determine the abnormal boundary of each cluster, and determine the abnormal frequency with multiple state factors according to the sum of the number of abnormalities of each cluster appearing in each sliding window.

[0052] There is no definite normal value range for multiple state factors, and it is necessary to determine whether they are normal according to the different states they are in. After clustering the monitoring data of multiple state factors in step 202, each cluster represents a state of the multiple state factors, and the abnormal boundary of each cluster can be determined according to the preset clustering radius, and the monitoring data outside the abnormal boundary is the abnormal data of the multiple state factors.

[0053] Exemplarily, in step 202 , a K-nearest neighbor clustering algorithm may be used to perform clustering.

[0054] In some embodiments of the present application, the method determines the fusion weight through step 301 and step 302.

[0055] Step 301: Calculate the first weight and the second weight of each evaluation index respectively using the analytic hierarchy process and the entropy method.

[0056] When the first weight of each evaluation index is calculated using the hierarchical analysis method in step 301, the relationship between the evaluation indexes is first analyzed to establish a hierarchical structure of the evaluation indexes.

[0057] Then, when determining the weights between the evaluation indicators at each level, all evaluation indicators are compared with each other, and relative scales are used to minimize the difficulty of comparing evaluation indicators of different natures to improve accuracy. For the importance of each evaluation indicator at the same level with respect to a criterion at the previous level, a pairwise judgment matrix is ​​constructed: .

[0058] The relative weight of the compared evaluation index to the criterion is calculated by the judgment matrix, and a consistency test is performed. The first weight of each evaluation index can be solved by the following methods: arithmetic mean method, geometric mean method and eigenvalue method.

[0059] Exemplarily, the first weight of the evaluation index is calculated using the arithmetic mean method: in, is the first weight of the i-th evaluation index, n is the number of samples, is the comparison result between the i-th evaluation index and the j-th evaluation index.

[0060] When the entropy method is used in step 301 to calculate the second weight of each evaluation index, the entropy value of each evaluation index is first calculated by the following formula: ;in, is the entropy value of the jth evaluation index, and n is the number of samples; is the proportion of the i-th sample under the j-th evaluation index; Then, the entropy redundancy is calculated by the following formula: ;in, is the entropy redundancy of the jth evaluation index; entropy redundancy is the complement of the entropy value, which indicates the amount of information of the evaluation index; Finally, the second weight of each evaluation index is calculated according to the entropy redundancy. The calculation formula of the second weight is as follows: ;in, Represents the second weight of the i-th evaluation index.

[0061] Step 302: Calculate a fusion weight according to a preset weight balancing factor, a first weight, and a second weight.

[0062] Specifically, when executing step 302, the following formula is calculated: ;in, is the fusion weight of the i-th evaluation index, is the weight balancing factor, is the first weight of the i-th evaluation index, is the second weight of the i-th evaluation index.

[0063] The first weight obtained by the analytic hierarchy process is the subjective weight, and the second weight of each evaluation index calculated by the entropy method according to the objective characteristics of the data is the objective weight. The advantage of combining the two analytic hierarchy processes and the entropy method to obtain the fusion weight is that the analytic hierarchy process can make up for the deficiency of the entropy method in that it is strongly dependent on data, while the entropy method can correct the tendency of the analytic hierarchy process to be overly dependent on subjective judgment, thereby making the obtained fusion weight more comprehensive and scientific, and then according to the initial score of each evaluation index and the fusion weight and the comprehensive score obtained after that, the objectivity and reliability of the state assessment can be enhanced.

[0064] In some embodiments of the present application, after forming the time series data, the method further includes: filling in missing values ​​and deleting outliers.

[0065] Specifically, when filling missing values, you can use the mean, median or mode to fill missing values; use methods such as standard deviation and interquartile range to identify outliers, and then delete the outliers.

[0066] like Figure 2 As shown, an embodiment of the present application provides a wind turbine hydraulic station degradation trend identification device 400, which includes a data acquisition module 401, an abnormal statistics module 402, a normalization module 403, a comprehensive scoring module 404, a model training module 405 and a degradation identification module 406.

[0067] The data acquisition module 401 is used to obtain monitoring data of a preset time period of multiple influencing factors to form time series data. The abnormal statistics module 402 is used to determine the abnormal frequency of each influencing factor as an evaluation indicator based on the time series data. The normalization module 403 is used to normalize each evaluation indicator to obtain an initial score for each evaluation indicator. The comprehensive scoring module 404 is used to determine the comprehensive score based on the initial score and the fusion weight. The model training module 405 is used to train the comprehensive scoring model using the time series data and the comprehensive score as input and expected output, respectively. The degradation identification module 406 is used to use the comprehensive scoring model to monitor the hydraulic station in real time to identify degradation trends and issue early warnings based on preset thresholds.

[0068] The abnormality statistics module 402 is specifically used to: use a sliding window of a preset step length and size in the time series data to determine the abnormal frequency of each influencing factor in each sliding window.

[0069] The multiple influencing factors include single-state factors and multiple-state factors; the abnormal statistical module 202 is used to determine the abnormal frequency of each influencing factor in each sliding window, specifically for: counting the number of abnormal occurrences of the monitoring data of the single-state factor in each sliding window, and determining the abnormal frequency of the single-state factor; clustering the monitoring data with multiple state factors, determining the abnormal boundary of each cluster, and determining the abnormal frequency of the multiple state factors based on the sum of the number of abnormal occurrences of each cluster in each sliding window.

[0070] The wind turbine hydraulic station degradation trend identification device also includes a weight determination module, which is used to: use a hierarchical analysis method and an entropy method to respectively calculate the first weight and the second weight of each evaluation index; and calculate a fusion weight based on a preset weight balance factor, the first weight and the second weight.

[0071] The wind turbine hydraulic station degradation trend identification device also includes a preprocessing module, which is used to fill in missing values ​​and delete abnormal values.

[0072] The devices or modules described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. For the convenience of description, the above devices are described in various modules according to their functions. When implementing the embodiments of the present application, the functions of each module can be implemented in the same or multiple software and / or hardware. Of course, the module that implements a certain function can also be implemented by combining multiple sub-modules or sub-units.

[0073] An embodiment of the present application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the method for identifying the degradation trend of a hydraulic station of a wind turbine generator set as described above is implemented.

[0074] An embodiment of the present application provides a computer-readable storage medium, which stores computer-readable instructions. When the computer-readable instructions are executed by a processor, the method for identifying the degradation trend of a hydraulic station of a wind turbine generator system as described above is implemented.

[0075] The above-mentioned storage media include but are not limited to random access memory (RAM), read-only memory (ROM), cache, hard disk drive (HDD) or memory card (Memory Card).

[0076] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other. Each embodiment focuses on the differences from other embodiments.

[0077] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit the present application. Although the present application has been described in detail with reference to the aforementioned embodiments, a person of ordinary skill in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some or all of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the present application.

Claims

1. A method for identifying the degradation trend of a hydraulic station of a wind turbine generator system, characterized in that: include: Obtain monitoring data of multiple influencing factors for a preset time period to form time series data; Based on the time series data, determining the abnormal frequency of each of the influencing factors as an evaluation indicator; Normalizing each of the evaluation indicators to obtain an initial score for each of the evaluation indicators; Determining a comprehensive score based on the initial score and the fusion weight; Taking the time series data and the comprehensive score as input and expected output respectively, training the comprehensive score model; The comprehensive scoring model is used to monitor the hydraulic station in real time to identify degradation trends and issue early warnings based on preset thresholds.

2. The method according to claim 1, characterized in that The determining, based on the time series data, the abnormal frequency of each of the influencing factors as an evaluation index comprises: A sliding window of preset step length and size is used in the time series data to determine the abnormal frequency of each of the influencing factors in each of the sliding windows.

3. The method according to claim 2, characterized in that The multiple influencing factors include a single state factor and multiple state factors; The determining the abnormal frequency of each of the influencing factors in each of the sliding windows includes: Counting the number of abnormalities of the monitoring data of the single state factor in each sliding window to determine the abnormal frequency of the single state factor; The monitoring data with the multiple state factors are clustered to determine the abnormal boundary of each cluster, and the abnormal frequency with the multiple state factors is determined according to the sum of the number of abnormalities of each cluster appearing in each sliding window.

4. The method according to claim 1, characterized in that The fusion weight is determined by the following steps: Calculate the first weight and the second weight of each evaluation index respectively by using the hierarchical analysis method and the entropy method; The fusion weight is calculated according to a preset weight balancing factor, the first weight and the second weight.

5. The method according to claim 1, characterized in that After forming the time series data, the method further includes: Fill in missing values ​​and remove outliers.

6. The method according to claim 1, characterized in that The multiple influencing factors include main system hydraulic pressure, yaw pressure, hydraulic oil temperature, wind speed, wind direction, power and hydraulic pump feedback time.

7. The method according to claim 1, characterized in that The comprehensive scoring model is a recurrent neural network.

8. A device for identifying degradation trend of hydraulic station of wind turbine generator set, characterized in that: include: A data acquisition module is used to acquire monitoring data of a preset time period of multiple influencing factors to form time series data; An abnormality statistics module, used to determine the abnormal frequency of each of the influencing factors as an evaluation index based on the time series data; A normalization module, used to normalize each of the evaluation indicators to obtain an initial score for each of the evaluation indicators; A comprehensive scoring module, used to determine a comprehensive score based on the initial score and the fusion weight; A model training module, used to take the time series data and the comprehensive score as input and expected output respectively, and train the comprehensive score model; A degradation identification module is used to monitor the hydraulic station in real time using the comprehensive scoring model to identify degradation trends and issue early warnings based on preset thresholds.

9. An electronic device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the method for identifying the degradation trend of a hydraulic station of a wind turbine generator set as claimed in any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-readable instructions, and when the computer-readable instructions are executed by a processor, the method for identifying degradation trends of a hydraulic station of a wind turbine generator set according to any one of claims 1 to 7 is implemented.

Images