Intelligent fault positioning method and system for power transmission line of offshore wind plant

Abstract

The invention provides an intelligent fault positioning method and system for an offshore wind plant power transmission line, and the method comprises the steps: obtaining the multi-source monitoring data of a target power transmission line, and extracting the spatial-temporal characteristics in the multi-source monitoring data; calling the trained deep learning network to carry out abnormal mode recognition processing on the multi-source monitoring data, and generating a fault feature matrix; based on the fault feature matrix, calling a multi-modal data fusion model to perform joint analysis on the spatio-temporal features, and generating a comprehensive feature vector including electrical anomaly features, environmental interference features and equipment degradation features; and inputting the comprehensive feature vector into a fault positioning algorithm to carry out dynamic path optimization processing, and outputting a fault positioning result which comprises an optimal fault node sequence generated based on abnormal weight distribution and an error correction parameter. According to the method, a dynamic fault analysis framework of multi-source data fusion is constructed, so that the problem that a traditional method is insufficient in positioning precision under complex working conditions is solved.

Description

Technical Field

[0001] The present invention relates to the field of offshore wind farms, and in particular to an intelligent fault location method and system for offshore wind farm transmission lines. Background Art

[0002] With the rapid development of offshore wind farms, the safety and reliability of transmission lines have become crucial factors in ensuring the stable operation of wind farms. Currently, offshore wind farm transmission line fault location technology still has significant deficiencies in accuracy, real-time performance, and multi-source data fusion capabilities under complex operating conditions, impacting fault diagnosis efficiency and system recovery speed. Therefore, there is an urgent need for a fault location method and system that can integrate multi-source data, such as electrical signals, environmental parameters, and equipment status, to improve the accuracy, real-time performance, and robustness of fault location, thereby meeting the needs of modern offshore wind farms for efficient and intelligent fault diagnosis. Summary of the Invention

[0003] The present invention provides an intelligent fault location method and system for offshore wind farm transmission lines to improve the above-mentioned problems.

[0004] The present invention provides an intelligent fault location method for an offshore wind farm transmission line, comprising: acquiring multi-source monitoring data of a target transmission line, and extracting spatiotemporal features from the multi-source monitoring data, wherein the multi-source monitoring data includes electrical signal time series data, environmental parameter distribution data, and equipment status trend data; Calling the trained deep learning network to perform abnormal pattern recognition processing on the multi-source monitoring data to generate a fault feature matrix, wherein the fault feature matrix includes a mapping relationship between timestamps and spatial positions and an abnormal weight distribution; Based on the fault feature matrix, a multimodal data fusion model is called to jointly analyze the spatiotemporal features to generate a comprehensive feature vector including electrical anomaly features, environmental interference features, and equipment degradation features; The comprehensive feature vector is input into a fault location algorithm for dynamic path optimization processing, and a fault location result is output. The fault location result includes an optimal fault node sequence and error correction parameters generated based on abnormal weight distribution.

[0005] Preferably, the acquiring of multi-source monitoring data of the target transmission line includes: Collect electrical signal timing data through current transformers and voltage transformers; Environmental parameter distribution data is collected through temperature and humidity sensors, anemometers, and salt spray concentration detectors; equipment status trend data is collected through vibration sensors and temperature sensors.

[0006] Preferably, the calling of the trained deep learning network to perform abnormal pattern recognition processing on the multi-source monitoring data to generate a fault feature matrix includes: Extracting spatial features of the multi-source monitoring data using a convolutional neural network; Using a long short-term memory network to capture the time series variation patterns of the multi-source monitoring data; A fault feature matrix including a mapping relationship between timestamps and spatial positions is generated according to the spatial features and time series variation rules.

[0007] Preferably, the calling of a multimodal data fusion model to jointly analyze the spatiotemporal features to generate a comprehensive feature vector includes: Assign dynamic weights to electrical signal timing data, environmental parameter distribution data, and equipment status trend data respectively; The spatiotemporal features are weightedly summed based on the dynamic weights to generate a comprehensive feature vector including electrical anomaly features, environmental interference features, and equipment degradation features.

[0008] Preferably, the step of inputting the comprehensive feature vector into a fault location algorithm for dynamic path optimization processing and outputting a fault location result includes: Calculate the distance metric between the candidate fault node and the actual fault signature; Combined with the overall smoothness constraint of the path, the optimal fault node sequence is generated; Error correction parameters are generated according to the optimal fault node sequence.

[0009] Preferably, the dynamic path optimization process adopts the following formula:

[0010] in, represents the optimal fault node sequence, L i represents the i-th candidate fault node, F i Indicates that L i The corresponding actual fault characteristics, w i represents the abnormal weight of the i-th node, d(L i , F i ) represents the distance measure between the candidate node and the actual fault feature, Δ(L) represents the overall smoothness constraint of the path, and λ represents the smoothness weight coefficient.

[0011] An embodiment of the present invention further provides an intelligent fault location system for an offshore wind farm transmission line, comprising: a data acquisition unit, configured to acquire multi-source monitoring data of a target transmission line and extract spatiotemporal features from the multi-source monitoring data, wherein the multi-source monitoring data includes electrical signal time series data, environmental parameter distribution data, and equipment status trend data; an identification unit, configured to call a trained deep learning network to perform abnormal pattern recognition processing on the multi-source monitoring data and generate a fault feature matrix, wherein the fault feature matrix includes a mapping relationship between timestamps and spatial positions and an abnormal weight distribution; A joint analysis unit is used to call a multimodal data fusion model to jointly analyze the spatiotemporal features based on the fault feature matrix, and generate a comprehensive feature vector including electrical anomaly features, environmental interference features, and equipment degradation features; a path optimization unit is used to input the comprehensive feature vector into a fault location algorithm for dynamic path optimization processing, and output a fault location result, which includes an optimal fault node sequence and error correction parameters generated based on the abnormal weight distribution.

[0012] Preferably, the path optimization unit is specifically used to: Calculate the distance metric between the candidate fault node and the actual fault signature; Combined with the overall smoothness constraint of the path, the optimal fault node sequence is generated; Error correction parameters are generated according to the optimal fault node sequence.

[0013] Preferably, the dynamic path optimization process adopts the following formula:

[0014] in, represents the optimal fault node sequence, L i represents the i-th candidate fault node, F i Indicates that L i The corresponding actual fault characteristics, w i represents the abnormal weight of the i-th node, d(L i , F i ) represents the distance measure between the candidate node and the actual fault feature, Δ(L) represents the overall smoothness constraint of the path, and λ represents the smoothness weight coefficient.

[0015] In summary, the present invention solves the technical bottleneck of insufficient positioning accuracy of traditional methods under complex working conditions by constructing a dynamic fault analysis framework of multi-source data fusion. By introducing a spatiotemporal feature extraction mechanism, the adaptive matching capability of fault features across regions and time periods is achieved; by combining a multimodal data fusion model with comprehensive feature vector analysis, a breakthrough is achieved in the joint reasoning of electrical anomalies and environmental interference; by integrating anomaly weight distribution and a dynamic path optimization algorithm with error correction parameters, the generated fault location results not only have an adaptive anomaly fault tolerance mechanism, but can also autonomously derive the optimal fault node sequence based on multi-source data features, providing an innovative technical implementation path for efficient and intelligent fault diagnosis of modern offshore wind farms. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 1 is a flow chart of an intelligent fault location method for an offshore wind farm transmission line provided by a first embodiment of the present invention; Figure 2 3 is a schematic structural diagram of an intelligent fault location system for offshore wind farm transmission lines provided by a second embodiment of the present invention. DETAILED DESCRIPTION

[0017] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the 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.

[0018] In order to better understand the technical solution of the present invention, the embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0019] It should be understood that the embodiments described are only a portion of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without creative work are within the scope of protection of the present invention.

[0020] The terms used in the embodiments of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The singular forms "a", "an", "the" and "the" used in the embodiments of the present invention and the appended claims are also intended to include plural forms unless the context clearly indicates otherwise.

[0021] It should be understood that the term "and / or" as used herein is merely a description of the relationship between associated objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A exists alone, A and B exist simultaneously, or B exists alone. Furthermore, the character " / " in this document generally indicates that the associated objects are in an "or" relationship.

[0022] The word "if," as used herein, may be interpreted as "at the time of" or "when" or "in response to determining" or "in response to detecting," depending on the context. Similarly, the phrases "if it is determined" or "if (stated condition or event) is detected" may be interpreted as "when it is determined" or "in response to the determination" or "when detecting (stated condition or event)" or "in response to detecting (stated condition or event)," depending on the context.

[0023] The "first" and "second" mentioned in the embodiments are merely used to distinguish similar objects and do not represent a specific ordering of the objects. It is understood that the specific order or precedence of "first" and "second" can be interchanged where appropriate. It should be understood that the objects distinguished by "first" and "second" can be interchanged where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein.

[0024] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0025] Multiple embodiments of the present invention provide an intelligent fault location method and system for offshore wind farm transmission lines, which builds a dynamic fault analysis framework by combining electrical signal time series data, environmental parameter distribution data, and equipment status trend data, significantly improving the accuracy, real-time, and robustness of fault location under complex working conditions. Figure 1 The specific implementation methods of this application are described in detail.

[0026] First, as attached Figure 1 As shown, the method provided by the embodiment of the present application includes multiple key steps, from acquiring multi-source monitoring data to finally outputting the optimal fault node sequence. Specifically, it includes: S101, acquiring multi-source monitoring data of a target transmission line and extracting spatiotemporal features from the multi-source monitoring data, wherein the multi-source monitoring data includes electrical signal time series data, environmental parameter distribution data, and equipment status trend data.

[0027] Among them, in actual applications, offshore wind farm transmission lines are usually in a complex operating environment, such as harsh conditions such as high humidity, strong winds and waves, and salt spray corrosion. These factors may cause the traditional single data source analysis method to be unable to accurately identify the fault location. Therefore, the embodiment of the present application first collects multi-source monitoring data of the target transmission line through a variety of sensors. These data mainly include electrical signal timing data, environmental parameter distribution data, and equipment status trend data. Among them, the electrical signal timing data is collected by current transformers, voltage transformers and other equipment to reflect the real-time operating status of the transmission line; the environmental parameter distribution data is collected by temperature and humidity sensors, anemometers, salt spray concentration detectors and other equipment to describe the impact of the external environment on the transmission line; the equipment status trend data is collected by vibration sensors, temperature sensors and other equipment to monitor the health status of transmission line related equipment. The acquisition frequency of the above-mentioned multi-source monitoring data needs to be set according to actual needs to ensure that the time resolution of the data can meet the requirements of subsequent analysis.

[0028] Next, after acquiring multi-source monitoring data, spatiotemporal feature extraction is required. The key to this process lies in utilizing deep learning networks to effectively capture the temporal and spatial information in the data. Specifically, the deep learning network employs an architecture that combines a convolutional neural network with a long short-term memory (LSTM) network. The convolutional neural network extracts spatial features, while the LSTM network captures patterns of time series variation. This architecture enables the extraction of spatiotemporal features from multi-source monitoring data, encompassing mappings between timestamps and spatial locations. For example, for electrical signal time series data, the network identifies abnormal fluctuations within a specific time period and correlates them with the spatial location of the transmission line. For environmental parameter distribution data, the network analyzes trends in environmental parameters across different regions and maps them to specific nodes on the transmission line. For equipment status trend data, the network captures patterns of equipment status over time and correlates these patterns with spatial location. The key to this process is ensuring that the extracted spatiotemporal features fully reflect the coupling relationship between electrical signals, environmental parameters, and equipment status, thereby laying the foundation for subsequent anomaly pattern recognition.

[0029] S102: Calling the trained deep learning network to perform abnormal pattern recognition processing on the multi-source monitoring data to generate a fault feature matrix, where the fault feature matrix includes a mapping relationship between timestamps and spatial positions and an abnormal weight distribution.

[0030] After completing spatiotemporal feature extraction, the next step is to use the trained deep learning network to identify anomaly patterns in the multi-source monitoring data and generate a fault signature matrix. The network is trained on a historical fault dataset, which includes multi-source monitoring data under normal operating conditions and multi-source monitoring data under known fault conditions. During network training, a supervised learning approach is used, taking both normal and faulty data as input and outputting the corresponding fault type and location. Through multiple iterative optimizations, the network learns the characteristic patterns of different fault types and generates a fault signature matrix containing mappings between timestamps and spatial locations, as well as anomaly weight distributions. For example, when a short circuit occurs on a transmission line, the network generates a fault signature matrix based on current surges in the electrical signal time series data, local temperature rises in the environmental parameter distribution data, and vibration anomalies in the equipment status trend data. Each element in the matrix represents an anomaly weight at a specific time point and spatial location. These anomaly weights reflect the probability of a fault at that location and provide a crucial basis for subsequent multimodal data fusion.

[0031] S103 , based on the fault feature matrix, calling a multimodal data fusion model to jointly analyze the spatiotemporal features, and generating a comprehensive feature vector including electrical anomaly features, environmental interference features, and equipment degradation features.

[0032] After generating the fault feature matrix, this embodiment further utilizes a multimodal data fusion model to jointly analyze the spatiotemporal features and generate a comprehensive feature vector. The core concept of the multimodal data fusion model is to overcome the information fragmentation problem of single-source analysis in complex scenarios by integrating information from different data sources. Specifically, the model uses a weighted fusion approach, assigning different weights to electrical signal time series data, environmental parameter distribution data, and equipment status trend data, and then performing a weighted summation. The weight assignment is dynamically adjusted based on the importance of each data source to fault location. For example, in high-humidity environments, the weight of environmental parameter distribution data may be increased to better reflect the impact of excessive humidity on transmission line insulation performance; whereas in areas with severe equipment aging, the weight of equipment status trend data may be increased to more accurately capture the impact of equipment degradation on the fault. In this way, the multimodal data fusion model can generate a comprehensive feature vector that incorporates electrical anomaly characteristics, environmental interference characteristics, and equipment degradation characteristics, thereby achieving a comprehensive characterization of the fault.

[0033] S104: Input the comprehensive feature vector into a fault location algorithm for dynamic path optimization processing, and output a fault location result. The fault location result includes an optimal fault node sequence and error correction parameters generated based on abnormal weight distribution.

[0034] After generating the comprehensive feature vector, the next step is to input it into the fault location algorithm for dynamic path optimization processing, outputting the optimal fault node sequence and error correction parameters. The dynamic path optimization algorithm formula used in the embodiment of the present application is as follows:

[0035] in, represents the optimal fault node sequence, L i represents the i-th candidate fault node, F i Indicates that L i The corresponding actual fault characteristics, w i represents the abnormal weight of the i-th node, d(L i , F i ) represents the distance measure between the candidate node and the actual fault feature, Δ(L) represents the overall smoothness constraint of the path, and λ represents the smoothness weight coefficient.

[0036] The first half of the formula ensures the accuracy of the fault location result through weighted distance measurement, and the second half enhances the continuity and robustness of the path through smoothness constraints. Specifically, in the calculation process, the priority of each candidate fault node is first determined according to the abnormal weight distribution in the comprehensive feature vector; then the distance measurement function is used to determine the priority of each candidate fault node. d(L i , F i ) Calculate the matching degree between candidate nodes and actual fault characteristics; finally, combine the smoothness constraint Δ(L) The overall continuity of the path is optimized to generate the optimal fault node sequence. In addition, the introduction of error correction parameters further enhances the robustness of the system, enabling it to adaptively adjust the positioning results under complex working conditions.

[0037] In order to verify the effectiveness of this embodiment, a specific application scenario is described below.

[0038] Suppose that a short circuit fault occurs in the transmission line of an offshore wind farm within a certain period of time, and the fault location is located on the sea surface about 5 kilometers away from the substation. At this time, the system first collects the electrical signal time series data, environmental parameter distribution data and equipment status trend data within this period of time through multi-source sensors. Subsequently, the system uses a deep learning network to extract spatiotemporal features from these data to generate spatiotemporal features containing a mapping relationship between timestamps and spatial positions. Next, the system calls the trained deep learning network to perform abnormal pattern recognition processing on the multi-source monitoring data to generate a fault feature matrix. On this basis, the system further calls the multimodal data fusion model to jointly analyze the spatiotemporal features and generate a comprehensive feature vector. Finally, the system inputs the comprehensive feature vector into the dynamic path optimization algorithm to generate the optimal fault node sequence and error correction parameters. Experimental results show that the embodiment of the present application can accurately identify the fault location and maintain high positioning accuracy and robustness under complex working conditions.

[0039] In summary, this embodiment solves the technical bottleneck of insufficient positioning accuracy of traditional methods under complex working conditions by constructing a dynamic fault analysis framework that integrates multi-source data. By introducing a spatiotemporal feature extraction mechanism, it achieves the ability to adaptively match fault features across regions and time periods. By combining a multimodal data fusion model with comprehensive feature vector analysis, it achieves a breakthrough in the joint reasoning of electrical anomalies and environmental interference. By integrating anomaly weight distribution with a dynamic path optimization algorithm for error correction parameters, the generated fault location results not only have an adaptive anomaly fault tolerance mechanism, but can also autonomously derive the optimal fault node sequence based on multi-source data features, providing an innovative technical implementation path for efficient and intelligent fault diagnosis in modern offshore wind farms.

[0040] See also Figure 2The second embodiment of the present invention further provides an intelligent fault location system for an offshore wind farm transmission line, which includes: a data acquisition unit 210, used to acquire multi-source monitoring data of a target transmission line and extract spatiotemporal features from the multi-source monitoring data, wherein the multi-source monitoring data includes electrical signal time series data, environmental parameter distribution data, and equipment status trend data; an identification unit 220, used to call a trained deep learning network to perform abnormal pattern recognition processing on the multi-source monitoring data to generate a fault feature matrix, wherein the fault feature matrix includes a mapping relationship between timestamps and spatial positions and an abnormal weight distribution; a joint analysis unit 230, used to call a multimodal data fusion model based on the fault feature matrix to perform joint analysis on the spatiotemporal features to generate a comprehensive feature vector including electrical abnormality features, environmental interference features, and equipment degradation features; a path optimization unit 240, used to input the comprehensive feature vector into a fault location algorithm for dynamic path optimization processing, and output a fault location result, wherein the fault location result includes an optimal fault node sequence and error correction parameters generated based on the abnormal weight distribution.

[0041] Preferably, the path optimization unit 240 is specifically used to: calculate the distance metric between the candidate fault nodes and the actual fault characteristics; generate an optimal fault node sequence in combination with the overall smoothness constraint of the path; and generate error correction parameters based on the optimal fault node sequence.

[0042] Preferably, the dynamic path optimization process adopts the following formula:

[0043] in, represents the optimal fault node sequence, L i represents the i-th candidate fault node, F i Indicates that L i The corresponding actual fault characteristics, w i represents the abnormal weight of the i-th node, d(L i , F i ) represents the distance measure between the candidate node and the actual fault feature, Δ(L) represents the overall smoothness constraint of the path, and λ represents the smoothness weight coefficient.

[0044] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. An intelligent fault location method for offshore wind farm transmission lines, characterized in that: include: Acquire multi-source monitoring data of a target transmission line and extract spatiotemporal features from the multi-source monitoring data, wherein the multi-source monitoring data includes electrical signal time series data, environmental parameter distribution data, and equipment status trend data; Calling the trained deep learning network to perform abnormal pattern recognition processing on the multi-source monitoring data to generate a fault feature matrix, wherein the fault feature matrix includes a mapping relationship between timestamps and spatial positions and an abnormal weight distribution; Based on the fault feature matrix, a multimodal data fusion model is called to jointly analyze the spatiotemporal features to generate a comprehensive feature vector including electrical anomaly features, environmental interference features, and equipment degradation features; The comprehensive feature vector is input into a fault location algorithm for dynamic path optimization processing, and a fault location result is output. The fault location result includes an optimal fault node sequence and error correction parameters generated based on abnormal weight distribution.

2. The method according to claim 1, wherein The acquiring of multi-source monitoring data of the target transmission line comprises: Collect electrical signal timing data through current transformers and voltage transformers; Environmental parameter distribution data is collected through temperature and humidity sensors, anemometers, and salt spray concentration detectors; equipment status trend data is collected through vibration sensors and temperature sensors.

3. The method according to claim 1, wherein The calling of the trained deep learning network to perform abnormal pattern recognition processing on the multi-source monitoring data to generate a fault feature matrix includes: Extracting spatial features of the multi-source monitoring data using a convolutional neural network; Using a long short-term memory network to capture the time series variation patterns of the multi-source monitoring data; A fault feature matrix including a mapping relationship between timestamps and spatial positions is generated according to the spatial features and time series variation rules.

4. The method according to claim 1, wherein The calling of the multimodal data fusion model to jointly analyze the spatiotemporal features to generate a comprehensive feature vector includes: Assign dynamic weights to electrical signal timing data, environmental parameter distribution data, and equipment status trend data respectively; The spatiotemporal features are weightedly summed based on the dynamic weights to generate a comprehensive feature vector including electrical anomaly features, environmental interference features, and equipment degradation features.

5. The method according to claim 1, wherein The step of inputting the comprehensive feature vector into a fault location algorithm for dynamic path optimization processing and outputting a fault location result includes: Calculate the distance metric between the candidate fault node and the actual fault signature; Combined with the overall smoothness constraint of the path, the optimal fault node sequence is generated; Error correction parameters are generated according to the optimal fault node sequence.

6. The method according to claim 5, wherein The dynamic path optimization process uses the following formula: in, represents the optimal fault node sequence, L i represents the i-th candidate fault node, F i Indicates that L i The corresponding actual fault characteristics, w i represents the abnormal weight of the i-th node, d(L i , F i ) represents the distance measure between the candidate node and the actual fault feature, Δ(L) represents the overall smoothness constraint of the path, and λ represents the smoothness weight coefficient.

7. An intelligent fault location system for offshore wind farm transmission lines, characterized in that: include: a data acquisition unit, configured to acquire multi-source monitoring data of a target transmission line and extract spatiotemporal features from the multi-source monitoring data, wherein the multi-source monitoring data includes electrical signal time series data, environmental parameter distribution data, and equipment status trend data; an identification unit, configured to call a trained deep learning network to perform abnormal pattern recognition processing on the multi-source monitoring data and generate a fault feature matrix, wherein the fault feature matrix includes a mapping relationship between timestamps and spatial positions and an abnormal weight distribution; A joint analysis unit is used to call a multimodal data fusion model to jointly analyze the spatiotemporal features based on the fault feature matrix, and generate a comprehensive feature vector including electrical anomaly features, environmental interference features, and equipment degradation features; a path optimization unit is used to input the comprehensive feature vector into a fault location algorithm for dynamic path optimization processing, and output a fault location result, which includes an optimal fault node sequence and error correction parameters generated based on the abnormal weight distribution.

8. The system according to claim 7, wherein: The path optimization unit is specifically used for: Calculate the distance metric between the candidate fault node and the actual fault signature; Combined with the overall smoothness constraint of the path, the optimal fault node sequence is generated; Error correction parameters are generated according to the optimal fault node sequence.

9. The system according to claim 8, wherein The dynamic path optimization process uses the following formula: in, represents the optimal fault node sequence, L i represents the i-th candidate fault node, F i Indicates that L i The corresponding actual fault characteristics, w i represents the abnormal weight of the i-th node, d(L i , F i ) represents the distance measure between the candidate node and the actual fault feature, Δ(L) represents the overall smoothness constraint of the path, and λ represents the smoothness weight coefficient.

Images