Wind power plant fault early warning system and method based on big data

Abstract

The invention discloses a wind power plant fault early warning system and method based on big data. The wind power plant fault early warning system comprises a data acquisition module, a data processing module and a fault early warning module, relates to the technical field of wind power management, and solves the technical problem of low operation maintenance efficiency and safety of a wind power plant caused by difficulty in accurately identifying potential faults of a fan by performing fault judgment through a fixed threshold in an existing scheme. According to the method, the position distribution and the historical operation data of the wind driven generators are acquired, and the regional division and the environmental data analysis are combined, so that the operation state of the wind power plant is finely monitored. The method comprises the following steps: firstly, dividing a wind power plant into a plurality of monitoring areas according to geographic positions, collecting historical environment data of each area and extracting environment characteristics, further analyzing the operation deviation degree of each wind driven generator by combining the characteristics, and realizing fault early warning according to the operation deviation degree. Therefore, the operation and maintenance efficiency and safety of the wind power plant can be improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of wind power management, and specifically relates to a wind farm fault early warning system and method based on big data. Background Art

[0002] With the growing global demand for clean energy, wind power, as a key renewable energy source, has seen rapid development and utilization. As the primary facility for converting wind energy into electricity, the operational efficiency and stability of wind farms directly impact the safety and reliability of power supply. However, in actual operation, critical equipment such as wind turbines in wind farms are prone to various failures due to factors such as equipment aging, environmental factors, and improper maintenance. These failures not only lead to reduced power generation but, in severe cases, can also cause safety incidents.

[0003] Existing technologies monitor wind turbine operating data in real time within a wind farm and compare it with preset warning thresholds. Based on the threshold comparison results, these systems can provide basic fault warnings for wind turbines. However, in practice, these fixed thresholds are prone to false or missed faults, resulting in low accuracy in wind farm fault warnings and inefficient wind farm operation and maintenance, as well as low safety.

[0004] The present invention proposes a wind farm fault early warning system and method based on big data to solve the above technical problems. Summary of the Invention

[0005] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes a wind farm fault early warning system and method based on big data, which is used to solve the technical problem that the existing solution is difficult to accurately identify potential faults of wind turbines through fault judgment through fixed thresholds, resulting in low operation and maintenance efficiency and safety of wind farms.

[0006] To achieve the above objectives, a first aspect of the present invention provides a wind farm fault early warning method based on big data, comprising:

[0007] Obtain location distribution information and historical operating data of multiple wind turbines in the target wind farm;

[0008] Based on the location distribution information of wind turbines, the wind farm is divided into several areas to obtain monitoring areas;

[0009] Collect environmental data of each monitoring area in several consecutive periods to obtain historical environmental data of each monitoring area;

[0010] Analyze the historical environmental data of each monitoring area separately to obtain the environmental characteristic data of the corresponding monitoring area;

[0011] Based on the environmental characteristic data, the historical operating data of each wind turbine is analyzed to obtain the deviation degree score of each wind turbine;

[0012] Fault warning is provided to the target wind farm based on the deviation degree score.

[0013] Preferably, the dividing the wind farm into regions based on the location distribution information of the wind turbines includes:

[0014] Extract the location distribution information of multiple wind turbines in the target wind farm; construct a plane coordinate system with any point in the target wind farm as the origin, the east direction as the horizontal axis positive direction, and the north direction as the vertical axis positive direction;

[0015] Marking coordinate points corresponding to the wind turbines in a plane coordinate system according to the location distribution information of the wind turbines;

[0016] The area where several wind turbines with adjacent coordinate points and the same wind turbine axis are located is divided into a monitoring area, and the monitoring area is numbered i; where i = 1, 2, ..., n, n is the total number of monitoring areas in the target wind farm.

[0017] Preferably, the analysis of the historical environmental data of each monitoring area includes:

[0018] Extract historical environmental data from each monitoring area; the historical environmental data includes wind speed, wind direction and rainfall;

[0019] After eliminating the maximum and minimum values ​​of wind speed and rainfall, the corresponding average values ​​are calculated to obtain the average wind speed and average rainfall; the wind direction with the highest frequency is marked as the characteristic wind direction; the average wind speed, average rainfall and characteristic wind direction are integrated into environmental characteristic data.

[0020] Preferably, the analyzing the historical operating data of each wind turbine based on the environmental characteristic data includes:

[0021] Extracting historical operating data of each wind turbine and environmental characteristic data of the monitoring area where each wind turbine is located; the historical operating data includes power generation, main shaft speed, and bearing temperature;

[0022] Performing linear fitting on the historical operating data of each wind turbine to obtain an operating change curve; wherein the operating change curve includes a power generation change curve, a main shaft speed change curve, and a bearing temperature change curve;

[0023] Inputting the environmental characteristic data of the monitoring area where each wind turbine is located into the wind turbine operation prediction model to obtain the standard operating data of the corresponding wind turbine under the corresponding environmental characteristic data; wherein the wind turbine operation prediction model is constructed based on an artificial intelligence model; the standard operating data includes standard values ​​of power generation, main shaft speed, and bearing temperature;

[0024] Performing linear fitting on the standard operating data of each wind turbine in several consecutive cycles to obtain corresponding standard operating curves; wherein the standard operating curves include a power generation standard curve, a main shaft speed standard curve, and a bearing temperature standard curve;

[0025] Calculate the deviation score between the operation variation curve of each wind turbine and the corresponding standard operation curve; wherein the deviation score includes the power generation deviation score, the main shaft speed deviation score and the bearing temperature deviation score.

[0026] Preferably, performing linear fitting on the historical operating data of each wind turbine respectively includes:

[0027] The historical operating data of each wind turbine was extracted; with time as the independent variable and power generation as the dependent variable, a power generation change curve was generated through linear fitting; with time as the independent variable and main shaft speed as the dependent variable, a main shaft speed change curve was generated through linear fitting; with time as the independent variable and bearing temperature as the dependent variable, a bearing temperature change curve was generated through linear fitting.

[0028] Preferably, the wind turbine operation prediction model is constructed based on an artificial intelligence model, including:

[0029] Extract environmental characteristic data of several continuous periods and corresponding wind turbine operation data; integrate each set of environmental characteristic data and corresponding operation data to obtain several sets of standard training data; use the several sets of standard training data to train the artificial intelligence model, and finally obtain a wind turbine operation prediction model with environmental characteristic data as input and wind turbine operation data as output; wherein the artificial intelligence model includes a BP neural network model or an RBF neural network model.

[0030] Preferably, the step of calculating the deviation score between the operation variation curve of each wind turbine and the corresponding standard operation curve includes:

[0031] Extracting the operation variation curve of each wind turbine and the corresponding standard operation curve;

[0032] By formula Calculate the deviation score PLFj of wind turbine j;

[0033] Among them, FDLj(t) is the power generation change curve of wind turbine j, and Aj(t) is the power generation standard curve of wind turbine j; ZZSj(t) is the main shaft speed change curve of wind turbine j, and Bj(t) is the main shaft speed standard curve of wind turbine j; ZWDj(t) is the bearing temperature change curve of wind turbine j, and Cj(t) is the bearing temperature standard curve of wind turbine j;

[0034] k1, k2, and k3 are all proportionality coefficients greater than 0; t is time, t1 is the start time of several consecutive periods, t2 is the end time of several consecutive periods, and t1 < t < t2; j = 1, 2,..., m, where m is the total number of wind turbines in the target wind farm.

[0035] Preferably, the fault warning for the target wind farm based on the deviation degree score includes:

[0036] Extract the deviation degree scores of each wind turbine;

[0037] Judge whether the deviation degree score is greater than the preset score threshold; if yes, mark the operating state of the corresponding wind turbine as a fault and add the number of the corresponding wind turbine to the first warning sequence; if not, mark the operating state of the corresponding wind turbine as normal; where the operating state includes normal and fault;

[0038] Sort the warning sequence in descending order of the deviation degree score to obtain the second warning sequence; generate the fault warning information for the corresponding wind turbines in sequence based on the second warning sequence.

[0039] Preferably, the generating the fault warning information for the corresponding wind turbines in sequence based on the second warning sequence includes:

[0040] Extract the numbers of the corresponding wind turbines from the second warning sequence in sequence, and query the basic information of the corresponding wind turbines from the database according to the numbers; generate the corresponding fault warning information according to the basic information of the wind turbines; where the basic information includes location information and model information.

[0041] The second aspect of the present invention provides a wind farm fault warning system based on big data, including: a data processing module, and a data acquisition module and a fault warning module connected thereto;

[0042] The data acquisition module: is used to obtain the position distribution information and historical operation data of multiple wind turbines in the target wind farm; divide the wind farm into several monitoring areas based on the position distribution information of the wind turbines; collect the environmental data of each monitoring area in several consecutive periods to obtain the historical environmental data of each monitoring area;

[0043] The data processing module is used to analyze the historical environmental data of each monitoring area to obtain the environmental characteristic data of the corresponding monitoring area; based on the environmental characteristic data, the historical operating data of each wind turbine is analyzed to obtain the deviation degree score of each wind turbine;

[0044] The fault warning module is used to provide fault warning for the target wind farm based on the deviation degree score.

[0045] Compared with the prior art, the present invention has the following beneficial effects:

[0046] 1. This invention achieves refined monitoring of wind farm operating status by acquiring the location distribution and historical operating data of wind turbines, combined with regional division and environmental data analysis. This method first divides the wind farm into multiple monitoring areas based on geographic location, collects historical environmental data from each area, and extracts environmental characteristics. These characteristics are then combined to analyze the degree of operational deviation of each wind turbine, ultimately providing fault warnings. This regionalized and characterized analysis approach improves the accuracy of anomaly identification and reduces false alarm and missed alarm rates, thereby improving the operational and maintenance efficiency and safety of wind farms. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0048] Figure 1 This is an overall flow chart of the wind farm fault early warning method based on big data of the present invention;

[0049] Figure 2 This is a schematic diagram of the principle of the wind farm fault early warning system based on big data of the present invention;

[0050] Figure 3 This is a flow chart of the present invention for performing fault early warning on a target wind farm based on the deviation degree score. DETAILED DESCRIPTION

[0051] The technical solutions of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0052] See also Figure 1-Figure 3The first embodiment of the present invention provides a wind farm fault early warning method based on big data, comprising:

[0053] S1: Obtain the location distribution information and historical operation data of multiple wind turbines in the target wind farm;

[0054] S2: Divide the wind farm into several monitoring areas based on the location distribution information of the wind turbines;

[0055] S3: Collect environmental data of each monitoring area in several consecutive periods to obtain historical environmental data of each monitoring area;

[0056] S4: Analyze the historical environmental data of each monitoring area separately to obtain the environmental characteristic data of the corresponding monitoring area;

[0057] S5: Analyze the historical operating data of each wind turbine based on the environmental characteristic data to obtain a deviation score for each wind turbine;

[0058] S6: Provide fault warning for the target wind farm based on the deviation degree score.

[0059] In this embodiment, the wind farm is divided into regions based on the location distribution information of the wind turbines, including:

[0060] Extract the location distribution information of multiple wind turbines in the target wind farm; construct a plane coordinate system with any point in the target wind farm as the origin, the east direction as the horizontal axis positive direction, and the north direction as the vertical axis positive direction;

[0061] Marking coordinate points corresponding to the wind turbines in a plane coordinate system according to the location distribution information of the wind turbines;

[0062] The area where several wind turbines with adjacent coordinate points and the same wind turbine axis are located is divided into a monitoring area, and the monitoring area is numbered i; where i = 1, 2, ..., n, n is the total number of monitoring areas in the target wind farm.

[0063] This method constructs a plane coordinate system based on the location distribution of wind turbines and divides the wind farm into zones based on the wind turbine axis, thereby achieving a rational division of the monitoring area. This method divides adjacent wind turbines with the same axial orientation into the same monitoring area, fully accounting for the similarities in the spatial distribution and operating characteristics of the units, and improving the pertinence and accuracy of subsequent environmental data collection and analysis. Compared to traditional extensive division methods, this solution can more realistically reflect the environmental differences and unit operating status within each zone, providing a foundation for subsequent accurate fault warnings for wind farms.

[0064] In this embodiment, the historical environmental data of each monitoring area is analyzed separately, including:

[0065] Extract historical environmental data from each monitoring area; the historical environmental data includes wind speed, wind direction and rainfall;

[0066] After eliminating the maximum and minimum values ​​of wind speed and rainfall, the corresponding average values ​​are calculated to obtain the average wind speed and average rainfall; the wind direction with the highest frequency is marked as the characteristic wind direction; the average wind speed, average rainfall and characteristic wind direction are integrated into environmental characteristic data.

[0067] In this embodiment, the historical operating data of each wind turbine is analyzed based on the environmental characteristic data, including:

[0068] Extracting historical operating data of each wind turbine and environmental characteristic data of the monitoring area where each wind turbine is located; the historical operating data includes power generation, main shaft speed, and bearing temperature;

[0069] Performing linear fitting on the historical operating data of each wind turbine to obtain an operating change curve; wherein the operating change curve includes a power generation change curve, a main shaft speed change curve, and a bearing temperature change curve;

[0070] Inputting the environmental characteristic data of the monitoring area where each wind turbine is located into the wind turbine operation prediction model to obtain the standard operating data of the corresponding wind turbine under the corresponding environmental characteristic data; wherein the wind turbine operation prediction model is constructed based on an artificial intelligence model; the standard operating data includes standard values ​​of power generation, main shaft speed, and bearing temperature;

[0071] Performing linear fitting on the standard operating data of each wind turbine in several consecutive cycles to obtain corresponding standard operating curves; wherein the standard operating curves include a power generation standard curve, a main shaft speed standard curve, and a bearing temperature standard curve;

[0072] Calculate the deviation score between the operation variation curve of each wind turbine and the corresponding standard operation curve; wherein the deviation score includes the power generation deviation score, the main shaft speed deviation score and the bearing temperature deviation score.

[0073] The present invention constructs a dynamic and quantifiable fault warning and evaluation system by combining historical operating data with environmental characteristic data. First, key operating parameters such as power generation, main shaft speed and bearing temperature are extracted, and linear fitting is performed to obtain their changing trends, thereby reflecting the historical operating status of the equipment; secondly, the environmental characteristic data of the area are input into the wind turbine operation prediction model built based on artificial intelligence to obtain standard operating data that matches the current environment, and a standard operating curve is formed based on this fitting to establish a dynamic benchmark that takes environmental factors into account; on this basis, by calculating the degree of deviation between the actual operating curve and the standard curve, a multi-dimensional quantitative evaluation of the abnormal operating conditions of the wind turbine is performed. This method effectively overcomes the problems of poor adaptability and high false alarm rate of the traditional fixed threshold warning method, which is conducive to improving the accuracy of fault warning for wind farms.

[0074] In this embodiment, linear fitting is performed on the historical operating data of each wind turbine, including:

[0075] The historical operating data of each wind turbine was extracted; with time as the independent variable and power generation as the dependent variable, a power generation change curve was generated through linear fitting; with time as the independent variable and main shaft speed as the dependent variable, a main shaft speed change curve was generated through linear fitting; with time as the independent variable and bearing temperature as the dependent variable, a bearing temperature change curve was generated through linear fitting.

[0076] In this embodiment, the wind turbine operation prediction model is constructed based on an artificial intelligence model, including:

[0077] Extract environmental characteristic data of several continuous cycles and corresponding operating data of wind turbines; integrate each group of environmental characteristic data and corresponding operating data to obtain several groups of standard training data; use 80% of the several groups of standard training data as training data and 20% as test data; use the training data to train the artificial intelligence model; use the test data to test the trained artificial intelligence model, and adjust the artificial intelligence model according to the test results; finally obtain a wind turbine operation prediction model with environmental characteristic data as input and wind turbine operation data as output; wherein the artificial intelligence model includes a BP neural network model or an RBF neural network model.

[0078] In this embodiment, calculating the deviation score between the operation variation curve of each wind turbine and the corresponding standard operation curve includes:

[0079] Extracting the operation variation curve of each wind turbine and the corresponding standard operation curve;

[0080] By formula Calculate the deviation score PLFj of wind turbine j;

[0081] Among them, FDLj(t) is the power generation change curve of wind turbine j, and Aj(t) is the standard power generation curve of wind turbine j; ZZSj(t) is the main shaft speed change curve of wind turbine j, and Bj(t) is the standard main shaft speed curve of wind turbine j; ZWDj(t) is the bearing temperature change curve of wind turbine j, and Cj(t) is the standard bearing temperature curve of wind turbine j;

[0082] k1, k2, and k3 are all proportionality coefficients greater than 0; t is time, t1 is the start time of several consecutive periods, t2 is the end time of several consecutive periods, and t1 < t < t2; j = 1, 2,..., m, where m is the total number of wind turbines in the target wind farm.

[0083] It should be noted that the values of the proportionality coefficients k1, k2, and k3 are related to the operating years of the wind turbine; when the operating years are longer, the actual values of the proportionality coefficients k1, k2, and k3 are set larger.

[0084] The present invention calculates the deviation degree scores of key operating parameters such as power generation, main shaft speed, and bearing temperature from the corresponding standard operating data within several consecutive periods; comprehensively considers the influence of the deviation degrees of power generation, main shaft speed, and bearing temperature on the failure risk of the wind turbine, provides an accurate basis for subsequent failure warning of the wind farm according to the deviation degree scores, thereby facilitating the improvement of the accuracy of failure warning of the wind farm, and further facilitating the improvement of the operation and maintenance efficiency and safety of the wind farm.

[0085] In this embodiment, based on the deviation degree scores, failure warning for the target wind farm includes:

[0086] Extract the deviation degree scores of each wind turbine;

[0087] Judge whether the deviation degree score is greater than a preset score threshold; if so, mark the operating state of the corresponding wind turbine as a failure, and add the number of the corresponding wind turbine to the first warning sequence; if not, mark the operating state of the corresponding wind turbine as normal; where the operating state includes normal and failure;

[0088] Sort the warning sequence in descending order of the deviation degree scores to obtain the second warning sequence; generate failure warning information for the corresponding wind turbines in sequence based on the second warning sequence.

[0089] In this embodiment, generating failure warning information for the corresponding wind turbines in sequence based on the second warning sequence includes:

[0090] The numbers of the corresponding wind turbines are extracted from the second warning sequence in sequence, and the basic information of the corresponding wind turbines is queried from the database according to the numbers; the corresponding fault warning information is generated according to the basic information of the wind turbines; wherein the basic information includes location information and model information.

[0091] A second embodiment of the present invention provides a wind farm fault warning system based on big data, comprising: a data processing module, and a data acquisition module and a fault warning module connected thereto;

[0092] Data acquisition module: used to obtain the location distribution information and historical operating data of multiple wind turbines in the target wind farm; divide the wind farm into several monitoring areas based on the location distribution information of the wind turbines; collect environmental data of each monitoring area over several consecutive periods to obtain historical environmental data of each monitoring area;

[0093] Data processing module: used to analyze the historical environmental data of each monitoring area separately to obtain the environmental characteristic data of the corresponding monitoring area; based on the environmental characteristic data, the historical operating data of each wind turbine is analyzed separately to obtain the deviation degree score of each wind turbine;

[0094] Fault warning module: used to provide fault warning for target wind farms based on deviation degree scores.

[0095] Some of the data in the above formula are calculated by removing the dimensions and taking their numerical values. The formula is a formula that is closest to the actual situation obtained by software simulation of a large amount of collected data; the preset parameters and preset thresholds in the formula are set by technical personnel in this field according to actual conditions or obtained through simulation of a large amount of data.

[0096] Working principle of the present invention:

[0097] The present invention obtains the location distribution information and historical operation data of multiple wind turbines in a target wind farm; divides the wind farm into regions based on the location distribution information of the wind turbines to obtain a number of monitoring areas; collects environmental data of each monitoring area in a number of continuous periods to obtain historical environmental data of each monitoring area; analyzes the historical environmental data of each monitoring area to obtain environmental characteristic data of the corresponding monitoring area; analyzes the historical operation data of each wind turbine based on the environmental characteristic data to obtain a deviation degree score of each wind turbine; and performs fault warning for the target wind farm based on the deviation degree score.

[0098] The above embodiments are only used to illustrate the technical method of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A wind farm fault early warning method based on big data, characterized in that: Obtain location distribution information and historical operating data of multiple wind turbines in the target wind farm; Based on the location distribution information of wind turbines, the wind farm is divided into several areas to obtain monitoring areas; Collect environmental data of each monitoring area in several consecutive periods to obtain historical environmental data of each monitoring area; Analyze the historical environmental data of each monitoring area separately to obtain the environmental characteristic data of the corresponding monitoring area; Based on the environmental characteristic data, the historical operating data of each wind turbine is analyzed to obtain the deviation degree score of each wind turbine; Fault warning is provided to the target wind farm based on the deviation degree score.

2. The wind farm fault early warning method based on big data according to claim 1 is characterized in that: The dividing the wind farm into regions based on the location distribution information of the wind turbines includes: Extract the location distribution information of multiple wind turbines in the target wind farm; construct a plane coordinate system with any point in the target wind farm as the origin, the east direction as the horizontal axis positive direction, and the north direction as the vertical axis positive direction; Marking coordinate points corresponding to the wind turbines in a plane coordinate system according to the location distribution information of the wind turbines; The area where several wind turbines with adjacent coordinate points and the same wind turbine axis are located is divided into a monitoring area, and the monitoring area is numbered i; where i = 1, 2, ..., n, n is the total number of monitoring areas in the target wind farm.

3. The wind farm fault early warning method based on big data according to claim 1 is characterized in that: The historical environmental data of each monitoring area are analyzed separately, including: Extract historical environmental data from each monitoring area; the historical environmental data includes wind speed, wind direction and rainfall; After eliminating the maximum and minimum values ​​of wind speed and rainfall, the corresponding average values ​​are calculated to obtain the average wind speed and average rainfall; the wind direction with the highest frequency is marked as the characteristic wind direction; the average wind speed, average rainfall and characteristic wind direction are integrated into environmental characteristic data.

4. The wind farm fault early warning method based on big data according to claim 1 is characterized in that: The analysis of the historical operating data of each wind turbine based on the environmental characteristic data includes: Extracting historical operating data of each wind turbine and environmental characteristic data of the monitoring area where each wind turbine is located; the historical operating data includes power generation, main shaft speed, and bearing temperature; Performing linear fitting on the historical operating data of each wind turbine to obtain an operating change curve; wherein the operating change curve includes a power generation change curve, a main shaft speed change curve, and a bearing temperature change curve; Inputting the environmental characteristic data of the monitoring area where each wind turbine is located into the wind turbine operation prediction model to obtain the standard operating data of the corresponding wind turbine under the corresponding environmental characteristic data; wherein the wind turbine operation prediction model is constructed based on an artificial intelligence model; the standard operating data includes standard values ​​of power generation, main shaft speed, and bearing temperature; Performing linear fitting on the standard operating data of each wind turbine in several consecutive cycles to obtain corresponding standard operating curves; wherein the standard operating curves include a power generation standard curve, a main shaft speed standard curve, and a bearing temperature standard curve; Calculate the deviation degree scores between the operation change curves of each wind turbine and the corresponding standard operation curves; among them, the deviation degree scores include power generation deviation scores, main shaft speed deviation scores, and bearing temperature deviation scores.

5. The wind farm fault early warning method based on big data according to claim 4 is characterized in that: The linear fitting of the historical operation data of each wind turbine respectively includes: Extract the historical operation data of each wind turbine; with time as the independent variable and power generation as the dependent variable, generate a power generation change curve through linear fitting; with time as the independent variable and main shaft speed as the dependent variable, generate a main shaft speed change curve through linear fitting; with time as the independent variable and bearing temperature as the dependent variable, generate a bearing temperature change curve through linear fitting.

6. The wind farm fault early warning method based on big data according to claim 4 is characterized in that: The wind turbine operation prediction model is constructed based on an artificial intelligence model, including: Extract the environmental feature data of several consecutive periods and the operation data of the corresponding wind turbine; integrate each group of environmental feature data and the corresponding operation data respectively to obtain several groups of standard training data; use several groups of standard training data to train the artificial intelligence model, and finally obtain a wind turbine operation prediction model with environmental feature data as the input and the operation data of the wind turbine as the output; among them, the artificial intelligence model includes a BP neural network model or an RBF neural network model.

7. The wind farm fault early warning method based on big data according to claim 4 is characterized in that: The calculation of the deviation degree scores between the operation change curves of each wind turbine and the corresponding standard operation curves includes: Extract the operation change curves of each wind turbine and the corresponding standard operation curves; By formula Calculate the deviation score PLFj of wind turbine j; Among them, FDLj(t) is the power generation change curve of wind turbine j, Aj(t) is the power generation standard curve of wind turbine j; ZZSj(t) is the main shaft speed change curve of wind turbine j, Bj(t) is the main shaft speed standard curve of wind turbine j; ZWDj(t) is the bearing temperature change curve of wind turbine j, Cj(t) is the bearing temperature standard curve of wind turbine j; k1, k2, and k3 are all proportionality coefficients greater than 0; t is time, t1 is the start time of several consecutive periods, t2 is the end time of several consecutive periods, and t1 < t < t2; j = 1, 2,..., m, and m is the total number of wind turbines in the target wind farm.

8. The wind farm fault early warning method based on big data according to claim 1 is characterized in that: The fault warning for the target wind farm based on the deviation degree scores includes: Extract the deviation degree scores of each wind turbine; Judge whether the deviation degree score is greater than the preset score threshold; if so, mark the operation status of the corresponding wind turbine as a fault and add the number of the corresponding wind turbine to the first warning sequence; if not, mark the operation status of the corresponding wind turbine as normal; among them, the operation status includes normal and fault; Sort the warning sequence in descending order of the deviation degree scores to obtain the second warning sequence; generate the fault warning information of the corresponding wind turbine based on the second warning sequence in turn.

9. The wind farm fault early warning method based on big data according to claim 8, characterized in that: The generation of the fault warning information of the corresponding wind turbine based on the second warning sequence in turn includes: The numbers of the corresponding wind turbines are extracted from the second warning sequence in sequence, and the basic information of the corresponding wind turbines is queried from the database according to the numbers; the corresponding fault warning information is generated according to the basic information of the wind turbines; wherein the basic information includes location information and model information.

10. A wind farm fault early warning system based on big data, used to implement the wind farm fault early warning method based on big data according to any one of claims 1 to 9, characterized in that: include: Data processing module, and the data acquisition module and fault warning module connected thereto; The data acquisition module is used to obtain the location distribution information and historical operating data of multiple wind turbines in the target wind farm; divide the wind farm into several monitoring areas based on the location distribution information of the wind turbines; and collect environmental data of each monitoring area over several consecutive periods to obtain historical environmental data of each monitoring area. The data processing module is used to analyze the historical environmental data of each monitoring area to obtain the environmental characteristic data of the corresponding monitoring area; Based on the environmental characteristic data, the historical operating data of each wind turbine is analyzed to obtain the deviation degree score of each wind turbine; The fault warning module is used to provide fault warning for the target wind farm based on the deviation degree score.

Images