Wind power low-frequency power transmission line fault discrimination method and system

Abstract

The invention provides a wind power low-frequency power transmission line fault judgment method and system, and belongs to the technical field of power system protection. According to the method, sensors are installed at key nodes of a low-frequency power transmission line of a wind power plant, and multi-dimensional data and traveling wave signals are collected in real time. The collected data is preprocessed, and multi-dimensional fault features and traveling wave features are extracted. The features are fused to form a comprehensive feature vector, then historical fault data and the comprehensive feature vector are utilized, a support vector machine is used for training a fault classification model, and a fault positioning model is trained based on a regression model of traveling wave features. The models are optimized through an improved sparrow algorithm and a particle swarm algorithm, so that the accuracy of fault judgment and positioning is improved. And finally, the optimized model is utilized to analyze comprehensive feature vectors and traveling wave features which are acquired in real time, so that rapid judgment of fault types and accurate positioning of fault positions are realized. Line faults can be accurately judged, and the stability of a power system is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power system protection, and in particular to a method and system for determining faults of a wind power low-frequency transmission line. Background Art

[0002] In the field of wind power generation, low-frequency transmission lines are an important bridge connecting wind turbines and power grids. Their stability and reliability are crucial to the operation of the entire power system. However, since wind farms are usually located in remote areas with complex and changeable environments, transmission lines are susceptible to various factors and may fail. The existing technology has some shortcomings in fault identification of wind power low-frequency transmission lines, which are specifically manifested in the following aspects: (1) Insufficient accuracy in fault identification: Existing technologies often rely on a single monitoring parameter or simple threshold judgment to identify faults. This method is easily affected by noise interference and complex working conditions, resulting in low identification accuracy. There is a lack of comprehensive fault identification methods that comprehensively consider multi-dimensional data and traveling wave signals, which cannot accurately reflect the actual fault status of the transmission line.

[0003] (2) Fault location accuracy is limited: Existing fault location technologies are usually based on simple ranging principles, such as impedance ranging, but this method is easily affected by error accumulation under complex network structures and fault conditions, and the positioning accuracy is limited. Although traveling wave ranging technology has high positioning accuracy, its application and optimization in existing technologies are insufficient, making it difficult to achieve fast and accurate fault location in complex environments.

[0004] (3) Low degree of intelligence: The existing technology lacks advanced algorithm and model support, and cannot realize intelligent and automatic fault identification. When processing large amounts of data and complex working conditions, manual intervention is frequent, inefficient, and easily affected by human factors. Summary of the invention

[0005] The purpose of the present invention is to provide a method and system for determining wind power low-frequency transmission line faults, which can accurately determine line faults and improve the stability of the power system.

[0006] To achieve the above object, the present invention provides a method and system for determining a fault in a wind power low-frequency transmission line, comprising the following steps: S1. Install sensors at key nodes of low-frequency transmission lines in wind farms to collect multi-dimensional data and traveling wave signals in real time; S2. Preprocess the collected data to remove noise interference and extract multi-dimensional fault features and traveling wave features; S3, fusing the extracted multi-dimensional fault features and traveling wave features to form a comprehensive feature vector; S4. Using historical fault data and comprehensive feature vectors, a support vector machine is used to train a fault classification model; a regression model based on traveling wave features is used to train a fault location model; S5. Optimize the fault classification model using the improved sparrow algorithm; optimize the fault location model using the improved particle swarm algorithm; S6, using the fault classification model optimized in step S5, inputting the comprehensive feature vector collected in real time into the model to distinguish the fault type; S7. Using the fault location model optimized in step S5, input the real-time collected traveling wave characteristics into the model, and combine with the geographic information system to locate the fault position.

[0007] Preferably, in step S1, the multi-dimensional data includes current, voltage, temperature and vibration data.

[0008] Preferably, in step S2, multi-dimensional fault features are extracted by Fourier transform, and traveling wave features are extracted by wavelet transform.

[0009] Preferably, in step S3, the extracted multi-dimensional fault features and traveling wave features are fused to form a comprehensive feature vector, and the specific operations are as follows: ; in, represents the comprehensive feature vector; Indicates the current characteristics; Indicates voltage characteristics; Indicates temperature characteristics; Indicates vibration characteristics; Represents the characteristics of traveling waves.

[0010] Preferably, in step S4, a support vector machine is used to train a fault classification model, and the specific operations are as follows: Find the optimal and : ; Subject to: ; in, represents the normal vector of the hyperplane; represents the bias term; represents the slack variable; Indicates Slack variables for samples; represents the regularization parameter; represents the number of samples; Indicates The category labels of samples; Indicates The comprehensive feature vector corresponding to the samples; The fault location model is trained using a regression model based on traveling wave features. The specific operations are as follows: Find the optimal model parameters by solving the following optimization problem : ; in, Indicates The actual fault location of each sample; The model is based on the characteristics of traveling waves Predicted failure location.

[0011] Preferably, in step S5, the fault classification model is optimized using an improved sparrow classification algorithm, and the specific operations are as follows: S511, initializing the sparrow population and sparrow speed; Sparrow population location ,in Indicates The position of each sparrow corresponds to the parameters of the fault classification model; The speed of sparrows ,in Indicates The speed of a sparrow; S512. Calculate the fitness of each sparrow, that is, the classification accuracy: ; in, , , , Respectively represent The number of true positives, true negatives, false positives, and false negatives corresponding to each sparrow; S513, Update sparrow speed: ; in, Indicates A new speed for each sparrow; Indicates The old speed of a sparrow; Indicates the disturbance amplitude, which controls the amplitude of the sine wave; Represents the angular frequency, which controls the period of the sine wave; Indicates the number of iterations; Update Sparrow position: ; in, Indicates the new location of the sparrow; Indicates The old position of a sparrow; The mean is 0 and the variance is Normally distributed random numbers; S514, recalculate the fitness, and when the fitness of the new position is higher than that of the old position, accept the new position; S515. Repeat steps S513-S514 until the maximum number of iterations is reached.

[0012] Preferably, in step S5, the improved particle swarm algorithm is used to optimize the fault location model, and the specific operations are as follows: S521, initializing particle group position and particle velocity; Particle swarm position ,in Indicates The positions of the particles correspond to the parameters of the fault location model; Particle Speed ,in Indicates The speed of a particle; S522. Calculate the fitness of each particle, that is, the positioning error: ; in, Indicates The particle prediction The fault location of each sample; S523, Dynamically adjust inertia weight: ; in, Indicates the current inertia weight; represents the maximum inertia weight; represents the minimum inertia weight; Indicates the maximum number of iterations; Indicates the number of iterations; S524. Update particle speed: ; in, Represents the updated particle velocity; , Both represent random numbers between [0,1]; Indicates The historical optimal position of a particle; represents the global optimal position; , represents the learning factor; S525, recalculate the fitness, and when the fitness of the new position is better than the individual historical optimal, update the individual optimal position; when the fitness of the new position is better than the global optimal, update the global optimal position; S526. Repeat steps S523-S525 until the maximum number of iterations is reached.

[0013] Preferably, in step S6, the fault classification model optimized in step S5 is used to input the comprehensive feature vector collected in real time into the model to distinguish the fault type. The specific operation is as follows: ; in, represents the decision function of the fault classification model; Represents the hyperplane normal vector of the optimized fault classification model; represents the bias term of the optimized fault classification model; Indicates that the comprehensive feature vector Functions that map to higher-dimensional spaces.

[0014] Preferably, in step S7, the fault location model optimized in step S5 is used to input the real-time collected traveling wave characteristics into the model, and combined with the geographic information system, the fault location is accurately located. The specific operations are as follows: S71. Input the traveling wave characteristics into the optimized fault location model to obtain the predicted fault location. ; S72, predicting the fault location Combined with geospatial data in a geographic information system that contains detailed geographic information about transmission lines; S73, using the spatial interpolation method in the geographic information system, based on the known control points and the predicted fault location, estimate the precise location of the fault point; S74, converting the predicted fault location into specific geographic coordinates, and performing coordinate system conversion to ensure consistency with data in the geographic information system; S75. Highlight the location of the fault point on the geographic information system map.

[0015] The present invention also provides a wind power low-frequency transmission line fault identification system, comprising: Sensor units are used to collect multi-dimensional data and traveling wave signals in real time at key nodes of low-frequency transmission lines in wind farms; Data preprocessing unit, used to remove noise interference and extract multi-dimensional fault features and traveling wave features; A feature fusion unit is used to fuse the extracted multi-dimensional fault features and traveling wave features to form a comprehensive feature vector; A model training unit, used to train a fault classification model and a fault location model; A model optimization unit, used to optimize the fault classification model using an improved sparrow algorithm and to optimize the fault location model using an improved particle swarm algorithm; A fault identification unit is used to input the comprehensive feature vector collected in real time into the optimized fault classification model to identify the fault type; The fault location unit is used to input the real-time collected traveling wave characteristics into the optimized fault location model and realize the location of the fault by combining with the geographic information system.

[0016] Therefore, the present invention adopts the above-mentioned method and system for determining faults in wind power low-frequency transmission lines, and the beneficial technical effects are as follows: (1) Improve the accuracy of fault identification: By installing sensors at key nodes of low-frequency transmission lines in wind farms, multi-dimensional data including current, voltage, temperature and vibration, as well as traveling wave signals, can be collected in real time to more comprehensively reflect the working status and fault characteristics of the transmission lines.

[0017] The fault classification model is trained using support vector machine and optimized with improved sparrow algorithm, which improves the accuracy of the model in distinguishing fault types.

[0018] (2) Improve fault location accuracy: The fault location model is trained using a regression model based on traveling wave characteristics and optimized using an improved particle swarm algorithm to achieve accurate and rapid location of the fault.

[0019] In conjunction with the Geographic Information System (GIS), the predicted fault location is combined with the detailed geographic information of the transmission line, which further improves the positioning accuracy and facilitates the rapid search and repair of the fault point.

[0020] (3) Enhanced intelligence: The method and system realize intelligent and automatic fault identification by introducing advanced algorithms and models, such as support vector machine, improved sparrow algorithm and particle swarm algorithm.

[0021] When processing large amounts of data and complex working conditions, human intervention is reduced, work efficiency is improved, and the impact of human factors on fault identification and positioning results is reduced.

[0022] (4) Improving power system stability: By accurately determining the type and location of line faults, appropriate protective measures can be taken in a timely manner to prevent the fault from expanding and causing more serious power system failures. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1This is a flow chart of a method for determining a fault in a low-frequency wind power transmission line according to the present invention; Figure 2 This is a structural diagram of a wind power low-frequency transmission line fault identification system of the present invention. DETAILED DESCRIPTION

[0024] The technical solution of the present invention is further described below through the accompanying drawings and embodiments.

[0025] Unless otherwise defined, technical or scientific terms used in the present invention shall have the common meanings understood by one having ordinary skills in the field to which the present invention belongs.

[0026] Embodiment 1 like Figure 1 FIG. 1 is a flow chart of a method for determining a fault of a wind power low-frequency transmission line according to the present invention, which specifically includes the following steps: S1. Install sensors at key nodes of low-frequency transmission lines in wind farms to collect multi-dimensional data and traveling wave signals in real time. The multi-dimensional data includes current, voltage, temperature and vibration data.

[0027] Current and voltage monitoring: Current transformers (CTs) and voltage transformers (PTs) are installed at key nodes of transmission lines to monitor current and voltage in real time. These sensors are installed on the high-voltage side near transformers or circuit breakers to ensure that the working status of the line can be accurately measured.

[0028] Temperature monitoring: Temperature sensors, such as thermocouples or infrared sensors, are installed at key locations such as conductor joints and switch contacts of transmission lines to monitor temperature in real time.

[0029] Vibration monitoring: Vibration acceleration sensors are installed at key points of the wind turbine's transmission chain to monitor vibration.

[0030] Wind speed and direction monitoring: A wind speed sensor and a wind direction sensor are installed on the top of the nacelle of the wind turbine. They can be mechanical wind speed sensors or ultrasonic wind speed sensors to monitor wind speed and wind direction in real time.

[0031] Traveling wave signal monitoring: Traveling wave sensors are installed along the transmission lines to collect traveling wave signals.

[0032] The types and installation of sensors are well known to those skilled in the art and will not be described in detail here.

[0033] S2. Preprocess the collected data to remove noise interference, extract multi-dimensional fault features through Fourier transform, and extract traveling wave features through wavelet transform.

[0034] S3. Fuse the extracted multi-dimensional fault features and traveling wave features to form a comprehensive feature vector. The specific operations are as follows: ; in, represents the comprehensive feature vector; Indicates the current characteristics; Indicates voltage characteristics; Indicates temperature characteristics; Indicates vibration characteristics; Represents the characteristics of traveling waves.

[0035] S4. Use historical fault data and comprehensive feature vectors to train the fault classification model using a support vector machine; use a regression model based on traveling wave features to train the fault location model.

[0036] Use support vector machine to train the fault classification model. The specific operations are as follows: Find the optimal and : ; Subject to: ; in, represents the normal vector of the hyperplane; represents the bias term; represents the slack variable; Indicates Slack variables for samples; represents the regularization parameter; represents the number of samples; Indicates The category labels of samples; Indicates The comprehensive feature vector corresponding to the samples; The fault location model is trained using a regression model based on traveling wave features. The specific operations are as follows: Find the optimal model parameters by solving the following optimization problem : ; in, Indicates The actual fault location of each sample; The model is based on the characteristics of traveling waves Predicted failure location.

[0037] S5. Use the improved sparrow algorithm to optimize the fault classification model. The specific operations are as follows: S511, initializing the sparrow population and sparrow speed; Sparrow population location ,in Indicates The position of each sparrow corresponds to the parameters of the fault classification model; The speed of sparrows ,in Indicates The speed of a sparrow; S512. Calculate the fitness of each sparrow, that is, the classification accuracy: ; in, , , , They are The number of true positives, true negatives, false positives, and false negatives corresponding to each sparrow; S513, Update sparrow speed: ; in, Indicates A new speed for each sparrow; Indicates The old speed of a sparrow; Indicates the disturbance amplitude, which controls the amplitude of the sine wave; Represents the angular frequency, which controls the period of the sine wave; Indicates the number of iterations; Update Sparrow position: ; in, Indicates a new location for the sparrow; Indicates The old position of a sparrow; The mean is 0 and the variance is Normally distributed random numbers; S514, recalculate the fitness, and when the fitness of the new position is higher than that of the old position, accept the new position; S515. Repeat steps S513-S514 until the maximum number of iterations is reached.

[0038] The improved particle swarm algorithm is used to optimize the fault location model. The specific operations are as follows: S521, initializing particle group position and particle velocity; Particle swarm position ,in Indicates The positions of the particles correspond to the parameters of the fault location model; Particle Speed ,in Indicates The speed of a particle; S522. Calculate the fitness of each particle, that is, the positioning error: ; in, Indicates The particle prediction The fault location of each sample; S523, Dynamically adjust inertia weight: ; in, Indicates the current inertia weight; represents the maximum inertia weight; represents the minimum inertia weight; Indicates the maximum number of iterations; t Indicates the number of iterations; S524. Update particle speed: ; in, Represents the updated particle velocity; , Represents a random number between [0,1]; Indicates The historical optimal position of a particle; represents the global optimal position; , represents the learning factor; S525, recalculate the fitness, and when the fitness of the new position is better than the individual historical optimal, update the individual optimal position; when the fitness of the new position is better than the global optimal, update the global optimal position; S526. Repeat steps S523-S525 until the maximum number of iterations is reached.

[0039] S6. Using the fault classification model optimized in step S5, the comprehensive feature vector collected in real time is input into the model to distinguish the fault type. The specific operation is as follows: ; in, represents the decision function of the fault classification model; Represents the hyperplane normal vector of the optimized fault classification model; represents the bias term of the optimized fault classification model; Indicates that the comprehensive feature vector Functions that map to higher-dimensional spaces.

[0040] S7. Using the fault location model optimized in step S5, input the real-time collected traveling wave characteristics into the model, and combine with the geographic information system to achieve accurate location of the fault. The specific operations are as follows: S71. Input the traveling wave characteristics into the optimized fault location model to obtain the predicted fault location. ; S72, predicting the fault location Combined with geospatial data in a geographic information system that contains detailed geographic information about transmission lines; S73, using the spatial interpolation method in the geographic information system, based on the known control points and the predicted fault location, estimate the precise location of the fault point; S74, converting the predicted fault location into specific geographic coordinates, and performing coordinate system conversion to ensure consistency with data in the geographic information system; S75. Highlight the location of the fault point on the geographic information system map.

[0041] It is worth noting that the contents not elaborated in detail in the present invention are all prior art and are well known to those skilled in the art.

[0042] Therefore, the present invention adopts the above-mentioned wind power low-frequency transmission line fault judgment method and system, which can accurately judge line faults and improve the stability of the power system.

[0043] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims

1. A method for determining faults in a wind power low-frequency transmission line, characterized in that: The following steps are involved: S1. Install sensors at key nodes of low-frequency transmission lines in wind farms to collect multi-dimensional data and traveling wave signals in real time; S2. Preprocess the collected data to remove noise interference and extract multi-dimensional fault features and traveling wave features; S3, fusing the extracted multi-dimensional fault features and traveling wave features to form a comprehensive feature vector; S4. Using historical fault data and comprehensive feature vectors, a support vector machine is used to train a fault classification model; a regression model based on traveling wave features is used to train a fault location model; S5. Optimize the fault classification model using the improved sparrow algorithm; optimize the fault location model using the improved particle swarm algorithm; S6, using the fault classification model optimized in step S5, inputting the comprehensive feature vector collected in real time into the model to distinguish the fault type; S7. Using the fault location model optimized in step S5, input the real-time collected traveling wave characteristics into the model, and combine with the geographic information system to locate the fault position.

2. A method for determining a wind power low-frequency transmission line fault according to claim 1, characterized in that: In step S1 , the multi-dimensional data includes current, voltage, temperature and vibration data.

3. A method for determining a wind power low-frequency transmission line fault according to claim 2, characterized in that: In step S2, multi-dimensional fault features are extracted by Fourier transform, and traveling wave features are extracted by wavelet transform.

4. A method for determining a wind power low-frequency transmission line fault according to claim 3, characterized in that: In step S3, the extracted multi-dimensional fault features and traveling wave features are fused to form a comprehensive feature vector. The specific operation is as follows: ； in, represents the comprehensive feature vector; Indicates the current characteristics; Indicates voltage characteristics; Indicates temperature characteristics; Indicates vibration characteristics; Represents the characteristics of traveling waves.

5. A method for determining a wind power low-frequency transmission line fault according to claim 4, characterized in that: In step S4, a support vector machine is used to train a fault classification model. The specific operations are as follows: Find the optimal and : ； Subject to: ； in, represents the normal vector of the hyperplane; represents the bias term; represents the slack variable; Indicates Slack variables for samples; represents the regularization parameter; represents the number of samples; Indicates The category labels of samples; Indicates The comprehensive feature vector corresponding to the samples; The fault location model is trained using a regression model based on traveling wave features. The specific operations are as follows: Find the optimal model parameters by solving the following optimization problem : ； in, Indicates The actual fault location of each sample; The model is based on the characteristics of traveling waves Predicted failure location.

6. A method for determining a wind power low-frequency transmission line fault according to claim 5, characterized in that: In step S5, the fault classification model is optimized using the improved sparrow classification algorithm, and the specific operations are as follows: S511, initializing the sparrow population and sparrow speed; Sparrow population location ,in Indicates The position of each sparrow corresponds to the parameters of the fault classification model; The speed of sparrows ,in Indicates The speed of a sparrow; S512. Calculate the fitness of each sparrow, that is, the classification accuracy: ； in, , , , Respectively represent The number of true positives, true negatives, false positives, and false negatives corresponding to each sparrow; S513, Update sparrow speed: ； in, Indicates A new speed for each sparrow; Indicates The old speed of a sparrow; Indicates the disturbance amplitude, which controls the amplitude of the sine wave; Represents the angular frequency, which controls the period of the sine wave; Indicates the number of iterations; Update Sparrow position: ； in, Indicates a new location for the sparrow; Indicates The old position of a sparrow; The mean is 0 and the variance is Normally distributed random numbers; S514, recalculate the fitness, and when the fitness of the new position is higher than that of the old position, accept the new position; S515. Repeat steps S513-S514 until the maximum number of iterations is reached.

7. A method for determining a wind power low-frequency transmission line fault according to claim 6, characterized in that: In step S5, the improved particle swarm algorithm is used to optimize the fault location model. The specific operations are as follows: S521, initializing particle group position and particle velocity; Particle swarm position ,in Indicates The positions of the particles correspond to the parameters of the fault location model; Particle Speed ,in Indicates The speed of a particle; S522. Calculate the fitness of each particle, that is, the positioning error: ； in, Indicates The particle prediction The fault location of each sample; S523, Dynamically adjust inertia weight: ； in, Indicates the current inertia weight; represents the maximum inertia weight; represents the minimum inertia weight; Indicates the maximum number of iterations; Indicates the number of iterations; S524. Update particle speed: ； in, Represents the updated particle velocity; , Both represent random numbers between [0,1]; Indicates The historical optimal position of a particle; represents the global optimal position; , represents the learning factor; S525, recalculate the fitness, and when the fitness of the new position is better than the individual historical optimal, update the individual optimal position; when the fitness of the new position is better than the global optimal, update the global optimal position; S526. Repeat steps S523-S525 until the maximum number of iterations is reached.

8. A method for determining a wind power low-frequency transmission line fault according to claim 7, characterized in that: In step S6, the fault classification model optimized in step S5 is used to input the comprehensive feature vector collected in real time into the model to distinguish the fault type. The specific operation is as follows: ； in, represents the decision function of the fault classification model; Represents the hyperplane normal vector of the optimized fault classification model; represents the bias term of the optimized fault classification model; Indicates that the comprehensive feature vector Functions that map to higher-dimensional spaces.

9. A method for determining a wind power low-frequency transmission line fault according to claim 8, characterized in that: In step S7, the fault location model optimized in step S5 is used to input the real-time collected traveling wave characteristics into the model, and combined with the geographic information system, the fault location is accurately located. The specific operations are as follows: S71. Input the traveling wave characteristics into the optimized fault location model to obtain the predicted fault location. ; S72, predicting the fault location Combined with geospatial data in a geographic information system that contains detailed geographic information about transmission lines; S73, using the spatial interpolation method in the geographic information system, based on the known control points and the predicted fault location, estimate the precise location of the fault point; S74, converting the predicted fault location into specific geographic coordinates, and performing coordinate system conversion; S75. Highlight the location of the fault point on the geographic information system map.

10. A wind power low-frequency transmission line fault identification system, characterized in that: include: Sensor units are used to collect multi-dimensional data and traveling wave signals in real time at key nodes of low-frequency transmission lines in wind farms; Data preprocessing unit, used to remove noise interference and extract multi-dimensional fault features and traveling wave features; A feature fusion unit is used to fuse the extracted multi-dimensional fault features and traveling wave features to form a comprehensive feature vector; A model training unit, used to train a fault classification model and a fault location model; A model optimization unit, used to optimize the fault classification model using an improved sparrow algorithm and to optimize the fault location model using an improved particle swarm algorithm; A fault identification unit is used to input the comprehensive feature vector collected in real time into the optimized fault classification model to identify the fault type; The fault location unit is used to input the real-time collected traveling wave characteristics into the optimized fault location model and realize the location of the fault by combining with the geographic information system.

Images