Fault analysis method, device, equipment and storage medium for power distribution equipment

Abstract

The present invention discloses a fault analysis method, apparatus, equipment and storage medium for power distribution equipment, comprising: obtaining historical alarm data and geographical environment data of the power distribution equipment, and determining the trend of environmental interference changes based on the historical alarm data and geographical environment data; wherein the environmental interference change trend is used to represent the dynamic influence trend of each geographical environment factor on the occurrence of power distribution equipment alarms; according to the environmental interference change trend, a convolutional neural network is used to extract features from the historical alarm data to obtain a comprehensive feature representation; the comprehensive feature representation is used to train the convolutional neural network to obtain a trained fault prediction model; real-time alarm data is obtained and input into the fault prediction model, and a preliminary fault cause is output; a long short-term memory network is used to perform time series analysis on the preliminary fault cause to obtain a fault diagnosis result; the present invention can improve the accuracy of fault cause identification and the intelligence level of alarm analysis, thereby ensuring the reliable operation of the power system.

Description

Technical Field

[0001] The present invention relates to the technical field of power systems, and in particular to a fault analysis method, apparatus, equipment and storage medium for power distribution equipment. Background Art

[0002] Alarm analysis in distribution communication networks can promptly identify equipment failures and reduce the risk of power outages, which is crucial to ensuring the stable operation of power systems. With the rapid development of smart grids, alarm analysis methods based on deep learning have gradually become a research hotspot, providing new possibilities for improving the accuracy and efficiency of fault diagnosis.

[0003] However, due to the influence of geographical environmental factors, the alarm characteristics of outdoor equipment show significant dynamics. Existing deep learning-based solutions tend to model static features, resulting in insufficient adaptability of the model in dynamic scenarios. It is difficult to accurately capture the complex relationship between fault causes and environmental factors. The analysis results often deviate from reality, and the identification of outdoor equipment fault causes faces the problem of reduced accuracy, which affects the intelligence level of distribution communication network alarm analysis and the reliable operation of the power system. Summary of the Invention

[0004] The embodiment of the present invention provides a fault analysis method for power distribution equipment, which can improve the accuracy of fault cause identification, enhance the intelligence level of alarm analysis in the power distribution communication network, and ensure the reliable operation of the power system.

[0005] In a first aspect, an embodiment of the present invention provides a fault analysis method for power distribution equipment, comprising:

[0006] Obtaining historical alarm data and geographical environment data of the power distribution equipment, and determining an environmental interference change trend based on the historical alarm data and the geographical environment data; wherein the environmental interference change trend is used to represent the dynamic impact trend of each geographical environment factor on the occurrence of power distribution equipment alarms;

[0007] According to the changing trend of the environmental interference, a convolutional neural network is used to extract features from the historical alarm data to obtain a comprehensive feature representation;

[0008] The convolutional neural network is trained using the comprehensive feature representation to obtain a trained fault prediction model;

[0009] Acquire real-time alarm data and input it into the fault prediction model to output preliminary fault causes;

[0010] A long short-term memory network is used to perform time series analysis on the preliminary fault cause to obtain a fault diagnosis result.

[0011] Furthermore, the acquiring of historical alarm data and geographical environment data of the power distribution equipment, and determining a trend of environmental interference changes based on the historical alarm data and the geographical environment data, includes:

[0012] Acquire historical alarm data of the power distribution equipment, and geographical environment data of the environment area where the power distribution equipment is located collected by various sensors; wherein the geographical environment data includes but is not limited to ambient temperature data, ambient humidity data, wind speed data, and vegetation coverage;

[0013] Calculate the equipment exposure distribution of the power distribution equipment in different seasons using a random forest algorithm based on the geographic environment data;

[0014] Calculating interference intensity of the device exposure of the power distribution equipment on the alarm frequency based on the device exposure distribution and the historical alarm data;

[0015] Using a clustering algorithm, the environmental area is divided into a high-impact area, a medium-impact area, and a low-impact area according to the geographical coordinates and the interference intensity;

[0016] Calculate the environmental interference change trend within a preset time period in each impact area; wherein, the environmental interference change trend is used to represent the dynamic impact trend of geographical environmental factors on the occurrence of distribution equipment alarms.

[0017] Furthermore, according to the changing trend of the environmental interference, a convolutional neural network is used to extract features from the historical alarm data to obtain a comprehensive feature representation, including:

[0018] Using a convolutional neural network, feature extraction is performed on the historical alarm data to obtain a feature map;

[0019] Based on the historical alarm data and the feature map, an attention mechanism is used to calculate the influence weights of different geographical environment factors on the alarm;

[0020] The feature maps are weightedly fused according to the influence weights to obtain a comprehensive feature representation after weighted fusion.

[0021] Furthermore, the convolutional neural network is used to extract features from the historical alarm data to obtain a feature map, including:

[0022] Using a convolutional neural network, the historical alarm data is organized into a feature expression matrix according to time and space dimensions; wherein the matrix element value of the feature expression matrix is ​​the number of alarms at a specific geographical location within a specific time period;

[0023] Analyzing the correlation between the alarm frequency and the geographical environment according to the feature expression matrix and the geographical environment data;

[0024] The convolution kernel size of the convolutional neural network is adjusted according to the correlation, and the feature expression matrix is ​​processed using the adjusted convolutional neural network to obtain an adjusted feature map.

[0025] Furthermore, the convolutional neural network is trained using the comprehensive feature representation to obtain a trained fault prediction model, including:

[0026] Inputting the feature representation into a convolutional neural network for training to obtain a first probability distribution of historical alarms; wherein the first probability distribution is the failure probability of each fault type in the historical alarms;

[0027] When the failure probability of a certain fault type is greater than a preset failure probability threshold, the model parameters of the convolutional neural network are updated based on the environmental interference change trend to obtain a fault prediction model.

[0028] Furthermore, the real-time alarm data is obtained and input into the fault prediction model to output preliminary fault causes, including:

[0029] Acquire real-time alarm data and geographic environment data of the power distribution equipment, input them into the fault prediction model, and output a second probability distribution corresponding to the alarm; wherein the second probability distribution is the fault probability of each fault type of the real-time alarm predicted by the fault prediction model;

[0030] Determining a maintenance difficulty index and a fault driving factor corresponding to each fault type in the second probability distribution; wherein the fault driving factor is a main driving factor causing the power distribution equipment to generate an alarm;

[0031] Calculating a contribution of the fault driving factor to the fault type, and adjusting the second probability distribution according to the contribution;

[0032] A preliminary fault cause is obtained according to the adjusted second probability distribution.

[0033] Furthermore, the long short-term memory network is used to perform a time series analysis on the preliminary fault cause to obtain a fault diagnosis result, including:

[0034] Acquire geographic environment data within a preset time period before the real-time alarm occurs, analyze the geographic environment data using a long short-term memory network, and output environmental time series characteristics;

[0035] A fault diagnosis result is obtained based on the environmental time sequence characteristics and the preliminary fault cause.

[0036] In a second aspect, an embodiment of the present invention provides a fault analysis device for power distribution equipment, comprising:

[0037] A data acquisition module, configured to acquire historical alarm data and geographic environment data of the power distribution equipment, and determine a trend of environmental interference changes based on the historical alarm data and the geographic environment data; wherein the trend of environmental interference changes is used to represent the dynamic impact trend of each geographic environment factor on the occurrence of power distribution equipment alarms;

[0038] A feature extraction module is used to extract features from the historical alarm data using a convolutional neural network according to the change trend of the environmental interference to obtain a comprehensive feature representation;

[0039] A model training module, configured to train the convolutional neural network using the comprehensive feature representation to obtain a trained fault prediction model;

[0040] A fault prediction module is used to obtain real-time alarm data and input it into the fault prediction model to output preliminary fault causes;

[0041] The fault diagnosis module is used to perform time series analysis on the preliminary fault cause using a long short-term memory network to obtain a fault diagnosis result.

[0042] In a third aspect, an embodiment of the present invention provides an electronic device, including:

[0043] memory for storing computer programs;

[0044] a processor for executing the computer program;

[0045] Wherein, when the processor executes the computer program, the fault analysis method for power distribution equipment described in any one of the first aspects above is implemented.

[0046] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed, the fault analysis method for the power distribution equipment described in any one of the first aspects above is implemented.

[0047] Compared with the prior art, the fault analysis method for distribution equipment provided by an embodiment of the present invention has the following beneficial effects: obtaining historical alarm data and geographical environment data of the distribution equipment, and determining the environmental interference change trend based on the historical alarm data and the geographical environment data; wherein the environmental interference change trend is used to represent the dynamic influence trend of each geographical environment factor on the occurrence of distribution equipment alarms; according to the environmental interference change trend, a convolutional neural network is used to extract features of the historical alarm data to obtain a comprehensive feature representation; the convolutional neural network is trained using the comprehensive feature representation to obtain a trained fault prediction model; real-time alarm data is obtained and input into the fault prediction model, and a preliminary fault cause is output; a long short-term memory network is used to perform time series analysis on the preliminary fault cause to obtain a fault diagnosis result; the present invention can improve the accuracy of fault cause identification, improve the intelligence level of alarm analysis of the distribution communication network, and ensure the reliable operation of the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical features of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments of the present invention. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0049] Figure 1 This is a flow chart of an embodiment of a method for analyzing a fault of a power distribution device provided by the present invention;

[0050] Figure 2 This is a structural diagram of an embodiment of a fault analysis device for power distribution equipment provided by the present invention;

[0051] Figure 3 It is a structural diagram of an embodiment of an electronic device provided by the present invention. DETAILED DESCRIPTION

[0052] 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. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0053] It should be noted that although the device schematics illustrate functional module divisions and the flowcharts illustrate logical sequences, in certain circumstances, the steps shown or described may be performed in a sequence that differs from the module divisions in the device or the sequence in the flowcharts. The terms "first," "second," and so on, in the specification, claims, and drawings, are used to distinguish similar items and are not necessarily used to describe a specific sequence or precedence.

[0054] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which the present invention pertains. The terms used herein are for the purpose of describing embodiments of the present invention only and are not intended to limit the present invention.

[0055] In a first aspect, an embodiment of the present invention provides a method for analyzing a fault of a power distribution device. Figure 1 , which is a flow chart of an embodiment of a fault analysis method for power distribution equipment provided by the present invention.

[0056] like Figure 1 As shown, the method includes the following steps:

[0057] S1: Obtain historical alarm data and geographical environment data of the power distribution equipment, and determine an environmental interference change trend based on the historical alarm data and the geographical environment data; wherein the environmental interference change trend is used to represent the dynamic impact trend of each geographical environment factor on the occurrence of power distribution equipment alarms;

[0058] S2: Based on the changing trend of the environmental interference, a convolutional neural network is used to extract features from the historical alarm data to obtain a comprehensive feature representation;

[0059] S3: Using the comprehensive feature representation to train the convolutional neural network to obtain a trained fault prediction model;

[0060] S4: Acquire real-time alarm data and input it into the fault prediction model to output preliminary fault causes;

[0061] S5: Use a long short-term memory network to perform a time series analysis on the preliminary fault cause to obtain a fault diagnosis result.

[0062] In the specific implementation, the historical alarm data and geographical environment data of the distribution equipment are first obtained. The historical alarm data may include information such as the time, type, and severity of the equipment failure. The geographical environment data may include factors such as temperature, humidity, air pressure, and altitude. By analyzing the historical alarm data and geographical environment data, the dynamic impact trend of each geographical environment factor on the occurrence of distribution equipment alarms can be understood, thereby determining the trend of environmental interference changes. For example, an increase in temperature may increase the probability of failure of certain equipment, and changes in humidity may affect the insulation performance of the equipment.

[0063] According to the changing trend of environmental interference, convolutional neural networks are used to extract features from historical alarm data to obtain more representative comprehensive feature representations. These comprehensive feature representations can reflect the potential patterns and laws in historical alarm data, providing strong support for subsequent fault prediction and diagnosis. These feature representations are used to train the convolutional neural network to obtain a trained fault prediction model. During the training process, historical alarm data can be used as input and the corresponding fault type probability as output. The parameters of the convolutional neural network are continuously adjusted through the optimization algorithm, so that the model can accurately predict the fault type probability.

[0064] After obtaining a trained fault prediction model, real-time alarm data is obtained and input into the fault prediction model to obtain a preliminary prediction of the fault cause. For example, when the real-time alarm data shows that the equipment has an overcurrent alarm, the fault prediction model can predict possible fault causes, such as equipment overload, short circuit, etc., based on the characteristics of historical alarm data and environmental data. The long-short-term memory network is used to perform time series analysis on the preliminary fault cause. The preliminary fault cause is used as input and processed by the long-short-term memory network to obtain a more accurate fault diagnosis result.

[0065] To sum up, the present invention obtains historical alarm data and geographical environment data of distribution equipment, and determines the trend of environmental interference changes based on the historical alarm data and the geographical environment data; wherein, the environmental interference change trend is used to represent the dynamic influence trend of each geographical environment factor on the occurrence of distribution equipment alarms; according to the environmental interference change trend, a convolutional neural network is used to extract features of historical alarm data to obtain a comprehensive feature representation; the convolutional neural network is trained using the feature representation to obtain a trained fault prediction model; real-time alarm data is obtained and input into the fault prediction model, and a preliminary fault cause is output; a long short-term memory network is used to perform time series analysis on the preliminary fault cause to obtain a fault diagnosis result; the present invention takes into account the real-time impact of outdoor environmental factors, can improve the accuracy of fault cause identification, has high accuracy and reliability, can improve the intelligence level of distribution communication network alarm analysis, and provide strong support for the operation and maintenance of smart grids, thereby ensuring the reliable operation of the power system.

[0066] In an optional embodiment, the acquiring of historical alarm data and geographic environment data of the power distribution equipment, and determining the trend of environmental interference changes based on the real-time alarm data and environment data, includes:

[0067] Acquire historical alarm data of the power distribution equipment, and geographical environment data of the environment area where the power distribution equipment is located collected by various sensors; wherein the geographical environment data includes but is not limited to ambient temperature data, ambient humidity data, wind speed data, and vegetation coverage;

[0068] Calculate the equipment exposure distribution of the power distribution equipment in different seasons using a random forest algorithm based on the geographic environment data;

[0069] Calculating interference intensity of the device exposure of the power distribution equipment on the alarm frequency based on the device exposure distribution and the historical alarm data;

[0070] Using a clustering algorithm, the environmental area is divided into a high-impact area, a medium-impact area, and a low-impact area according to the geographical coordinates and the interference intensity;

[0071] Calculate the environmental interference change trend within a preset time period in each impact area; wherein, the environmental interference change trend is used to represent the dynamic impact trend of geographical environmental factors on the occurrence of distribution equipment alarms.

[0072] Specifically, historical alarm data of outdoor equipment in the power distribution and communication network within a preset time period is obtained, such as alarm data of voltage anomalies or communication interruptions, as well as geographical environment data of the environmental area where the power distribution equipment is located collected by various sensors. Sensors can be deployed on distribution poles or communication base stations to monitor environmental parameters such as temperature, humidity, and wind speed in real time. By combining remote sensing images with ground sensor data, the vegetation coverage rate around the equipment can be determined. For example, in summer, the vegetation is lush, with a coverage rate of 70%, which may block the ventilation of the equipment and lead to poor heat dissipation. In winter, the coverage rate drops to 30%.

[0073] The random forest algorithm is used to calculate the equipment exposure distribution of distribution equipment in different seasons. Equipment exposure can be understood as the degree to which distribution equipment is affected by various factors in a specific geographical environment. For example, in an environment with high vegetation coverage, equipment exposure may be low, while in an environment with low vegetation coverage, equipment exposure may be high. Since geographical environmental factors change in different seasons, the equipment exposure distribution of distribution equipment will also be different. Random forest is an integrated learning algorithm that improves the accuracy and stability of predictions by constructing multiple decision trees and combining their prediction results. For example, the algorithm votes through a large number of decision trees and concludes that equipment exposure decreases by 20% in summer and increases by 15% in winter.

[0074] Based on the equipment exposure distribution and historical alarm data, the interference intensity of the equipment exposure on the alarm frequency is determined. The interference intensity reflects the degree of influence of the equipment exposure on the alarm frequency of the distribution equipment. If the interference intensity is high, it means that the equipment exposure has a greater impact on the alarm frequency, and corresponding measures need to be taken to reduce the equipment exposure to reduce the occurrence of alarms. If the interference intensity is low, it means that the equipment exposure has a smaller impact on the alarm frequency, and the control of the equipment exposure can be appropriately relaxed. For example, when it is found that the exposure is higher than 0.6, the alarm frequency increases by 30%, indicating that the high exposure area is more affected by wind, sand or low temperature. It can be understood that when the interference intensity exceeds the preset interference intensity threshold, the support vector machine algorithm is used to judge the dynamic change trend of the interference intensity. The support vector machine uses time series data as input to divide the trend characteristics of increasing or decreasing interference intensity. For example, the support vector machine algorithm can predict that the interference intensity in a certain area will increase by 10% in the next three months. The advantage of this method is that it can quickly classify complex data and is suitable for dynamic environment analysis.

[0075] A clustering algorithm is used to divide the environmental area according to geographic coordinates and interference intensity, and divide it into high-impact area, medium-impact area and low-impact area. The interference intensity in the high-impact area is higher and may be concentrated on the top of the mountain with sparse vegetation. The distribution equipment in these areas is more susceptible to geographical environmental factors, and the alarm frequency is relatively high. The interference intensity in the medium-impact area is moderate, and the alarm frequency of the distribution equipment in these areas is also at a medium level. The interference intensity in the low-impact area is lower, and the distribution equipment in these areas is less affected by geographical environmental factors, and the alarm frequency is also relatively low.

[0076] Calculate the environmental interference change trend within the preset time period for each impact area. For example, if the recent alarms in a high-impact area suddenly increase by 20%, it is concluded that the environmental interference change trend is worsening. Therefore, high-interference areas can be warned in advance based on the environmental interference change trend, reducing equipment failure rates and optimizing maintenance resource allocation.

[0077] In an optional embodiment, the feature extraction of the historical alarm data using a convolutional neural network based on the environmental interference change trend to obtain a comprehensive feature representation includes:

[0078] Using a convolutional neural network, feature extraction is performed on the historical alarm data to obtain a feature map;

[0079] Based on the historical alarm data and the feature map, an attention mechanism is used to calculate the influence weights of different geographical environment factors on the alarm;

[0080] The feature maps are weightedly fused according to the influence weights to obtain a comprehensive feature representation after weighted fusion.

[0081] Specifically, a convolutional neural network is used to extract features from historical alarm data to obtain a feature map. The historical alarm data and the feature map are combined to perform attention mechanism processing to analyze the influence weights of seasonal vegetation coverage, equipment exposure fluctuations and environmental interference distribution on the alarm characteristic change pattern. For example, assuming that the alarm frequency decreases by 15% when the vegetation coverage rate is high in summer, and the alarm frequency increases by 20% when the equipment exposure rate is high, the attention mechanism will assign a weight of 0.6 to the exposure and a weight of 0.3 to the vegetation. This weight distribution intuitively reflects the influence intensity of each factor. If the influence weight exceeds the preset influence weight threshold, the weight distribution is adjusted through the attention mechanism to obtain an updated weight distribution.

[0082] According to the updated weight distribution, the feature maps of seasonal vegetation, equipment exposure and environmental interference are weighted and fused to obtain a comprehensive feature representation. For example, the weights of seasonal vegetation, equipment exposure and environmental interference are 0.25, 0.65 and 0.1 respectively. The vegetation feature map can be multiplied by 0.25, the exposure feature map can be multiplied by 0.65, and the interference feature map can be multiplied by 0.1 to obtain a comprehensive feature representation. This weighting method highlights the main influencing factors and facilitates the subsequent analysis of the change pattern of the alarm characteristics.

[0083] In an optional embodiment, the use of a convolutional neural network to extract features from the historical alarm data to obtain a feature map includes:

[0084] Using a convolutional neural network, the historical alarm data is organized into a feature expression matrix according to time and space dimensions; wherein the matrix element value of the feature expression matrix is ​​the number of alarms at a specific geographical location within a specific time period;

[0085] Analyzing the correlation between the alarm frequency and the geographical environment according to the feature expression matrix and the geographical environment data;

[0086] The convolution kernel size of the convolutional neural network is adjusted according to the correlation, and the feature expression matrix is ​​processed using the adjusted convolutional neural network to obtain an adjusted feature map.

[0087] Specifically, historical alarm data may contain time series. Convolutional neural networks analyze these sequences through sliding windows to extract periodic features of abnormal occurrences. For example, if a device records 10 alarms every day in summer, the network may identify characteristic patterns during high temperature periods, such as concentrated alarms from 12 noon to 3 pm. This method can effectively mine hidden change patterns. Combining the change patterns in alarm data with geographical environmental factors can provide a more comprehensive understanding of the relationship between equipment alarms and the environment. For example, in a certain area, the temperature reaches 35 degrees and the humidity is 80% in summer. The alarm pattern shows frequent overheating problems, while the alarms decrease when the temperature is 5 degrees in winter. After fusion, it can be concluded that high temperature and high humidity have a greater impact on equipment. By taking environmental factors such as temperature, vegetation coverage, and wind speed as input, the environmental data and alarm data are integrated into a multidimensional vector to generate feature expressions. By combining feature expressions with dynamic changes, a feature expression is generated. The expression matrix can intuitively reflect dynamic changes such as seasonal fluctuations in equipment status. The support vector machine uses the feature matrix as input to classify intensity levels, including high intensity and low intensity. For example, when the exposure is 0.6, the alarm frequency increases by 30%, which is classified as high intensity. When the alarm frequency exceeds the preset threshold, such as the alarm frequency exceeds 50 times / month, the support vector machine can quickly quantify the impact of the change. By combining the intensity distribution and the geographical environment, areas with greater environmental impact can be accurately located. For example, the mountaintop area with high exposure and high wind speed in hilly areas is classified as a high dynamic change area with frequent alarms. This division can provide specific target areas for optimizing maintenance strategies and improve maintenance efficiency. When the alarms in the high dynamic change area increase sharply in the near future, the corresponding eigenvalue in the matrix is ​​adjusted upward. This update mechanism can reflect the changes in equipment status in real time and ensure the accuracy of the analysis. By continuously updating the expression matrix, changes in equipment status can be grasped in a timely manner.

[0088] It can be understood that the feature expression matrix can reflect the temporal and spatial distribution of alarms. By extracting geographical environment related data from the feature expression matrix and determining the change trend, the impact of the geographical environment on the alarm can be captured. Specifically, column data related to rainfall and terrain height can be extracted from the matrix. For example, when the rainfall in a certain area increases by 20 mm, the alarm frequency increases by 30%. By comparing the data of different time periods, the change trend of the alarm frequency is obtained. The convolution kernel size is adjusted according to the change trend to obtain the depth adjustment parameter. For example, if the change trend shows that the alarm is concentrated in a local area, the convolution kernel size can be adjusted from 5×5 to 3×3 to focus on the detail features. If the trend is relatively gentle, a larger size such as 7×7 is maintained. The depth adjustment parameter is set to the pixel range covered by the convolution kernel accordingly. The depth adjustment parameter is used to update the hierarchical structure to obtain an adaptive The feature extraction model can improve the adaptability of the model to dynamic data. For example, small convolution kernels are used in shallow networks to extract local features, and large convolution kernels are integrated in deep networks to capture global trends, forming a model with three layers of convolution. The parameters of each layer are dynamically updated according to the adjustment results. The matrix structure is processed by the adaptive feature extraction model to obtain a feature map with adjusted width, which can further expand the expressive ability of the features. If the change trend exceeds the preset threshold, the feature map is environmentally modulated to obtain the final feature map. For example, when the alarm frequency growth rate exceeds 50%, the feature map can be modulated by weighted rainfall data. For example, the characteristic value of the area with larger rainfall is increased by 20%, thereby highlighting the environmental impact. By adjusting the convolution kernel size and environmental modulation and other steps, the model can adapt more flexibly to alarm changes in different geographical environments.

[0089] In an optional embodiment, the using the comprehensive feature representation to train the convolutional neural network to obtain a trained fault prediction model includes:

[0090] Inputting the feature representation into a convolutional neural network for training to obtain a first probability distribution of historical alarms; wherein the first probability distribution is the failure probability of each fault type in the historical alarms;

[0091] When the failure probability of a certain fault type is greater than a preset failure probability threshold, the model parameters of the convolutional neural network are updated based on the environmental interference change trend to obtain a fault prediction model.

[0092] Specifically, the feature representation is input into a convolutional neural network for training. The convolutional neural network can automatically learn the complex relationship between different features, so as to better extract the probability distribution of historical faults and obtain the first probability distribution of historical alarms, that is, the failure probability of each fault type in the historical alarm. For example, the historical alarm data contains 1,000 equipment failure records. After training, the network outputs the probability of each fault type, such as the probability of circuit short circuit is 0.6 and the probability of mechanical wear is 0.3. This probability distribution reflects the potential law of failure and helps to identify risk points in advance.

[0093] Furthermore, assume that the preset circuit short circuit probability threshold is 0.4, and the actual probability reaches 0.6, indicating an abnormal situation. In this case, introducing geographic data, such as rainfall or altitude in the area where the equipment is located, can further reveal the impact of environmental changes on equipment failures. For example, an increase in rainfall from 50 mm to 200 mm per month may lead to abnormal humidity, which in turn affects the failure probability. Time series analysis can be used to determine whether the failure probability increases with the increase in geographical environmental factors such as rainfall. For example, data for three consecutive months shows that the probability increases from 0.4 to 0.7, and the trend is clearly upward. Model parameters such as the convolution kernel size or learning rate can be dynamically adjusted according to the environmental change trend. When the trend increases, the learning rate is adjusted from 0.01 to 0.005 to adapt to the new data more finely. This adjustment can improve the model's responsiveness to dynamic environments and enable it to better adapt to the ever-changing actual situation. The convolutional neural network after the model parameters are updated is the fault prediction model, which can be used for fault prediction.

[0094] In an optional embodiment, acquiring real-time alarm data and inputting it into the fault prediction model to output a preliminary fault cause includes:

[0095] Acquire real-time alarm data and geographic environment data of the power distribution equipment, input them into the fault prediction model, and output a second probability distribution corresponding to the alarm; wherein the second probability distribution is the fault probability of each fault type of the real-time alarm predicted by the fault prediction model;

[0096] Determining a maintenance difficulty index and a fault driving factor corresponding to each fault type in the second probability distribution; wherein the fault driving factor is a main driving factor causing the power distribution equipment to generate an alarm;

[0097] Calculating a contribution of the fault driving factor to the fault type, and adjusting the second probability distribution according to the contribution;

[0098] A preliminary fault cause is obtained according to the adjusted second probability distribution.

[0099] Specifically, real-time alarm data is processed through a fault prediction model. When a certain power equipment is operating, the real-time alarm data may include information such as voltage anomalies and current fluctuations. The fault prediction model analyzes this data, extracts the offset characteristics of equipment exposure and environmental interference, and generates a second probability distribution, that is, the failure probability of each fault type in the real-time alarm. For example, assuming that a certain transformer operates in a high-temperature environment, its real-time alarm data is input into the fault prediction model. The model identifies that the probability of overheating failure is 70%, while the probability of short-circuit failure is 20%.

[0100] It should be noted that after obtaining the second probability distribution, it is necessary to determine the maintenance difficulty index and fault driving factors corresponding to each fault type. The cluster analysis method can be used to group the second probability distribution, and the second probability distribution of fault types can be divided into high-risk and low-risk categories. The trend of changes in the maintenance difficulty index can be obtained from the grouping results. For example, an overheating fault caused by temperature increase may increase the maintenance difficulty index from 3 to 5 due to the deterioration of the equipment's heat dissipation conditions. The fault driving factor is the main driving factor that causes the distribution equipment to alarm. For example, temperature increase may be the main driving factor that causes overheating faults.

[0101] By calculating the contribution of fault driving factors to the fault type through regression analysis, for example, calculating that the probability of overheating failure increases by 5% for every 1 degree increase in temperature, this can accurately reflect the impact of the fault driving factors on the fault type in the current scenario. Based on the calculated contribution, the second probability distribution is adjusted to make the probability distribution more accurately reflect the current scenario. The adjusted probability distribution can improve the targetedness of the prediction and provide a more reliable basis for subsequent fault cause analysis.

[0102] Based on the adjusted second probability distribution, the mapping relationship between equipment exposure fluctuations and cause distribution is analyzed to determine the preliminary cause of the fault. For example, if the exposure of a certain equipment increases from 60% to 80% due to windy and sandy weather, the main cause of the fault changes from overheating to insulation aging. This mapping relationship helps to trace the root cause of the fault.

[0103] In an optional embodiment, the use of a long short-term memory network to perform a time series analysis on the preliminary fault cause to obtain a fault diagnosis result includes:

[0104] Acquire geographic environment data within a preset time period before the real-time alarm occurs, analyze the geographic environment data using a long short-term memory network, and output environmental time series characteristics;

[0105] A fault diagnosis result is obtained based on the environmental time sequence characteristics and the preliminary fault cause.

[0106] Specifically, geographic environment data within a preset time period before the real-time alarm occurs is obtained, and the data is analyzed using a long short-term memory network to output environmental time series characteristics. The long short-term memory network can analyze the time series pattern of geographic environment data. For example, in a device operation scenario, assuming that a device records data for 30 consecutive days, the daily illumination angle offset ranges from 5 degrees to 15 degrees, and the humidity level fluctuates between 60% and 80%. The long short-term memory network can analyze the time series pattern of these data and output initial time series characteristics, such as the periodic peak of the illumination angle offset occurs at 10 am, and the humidity level reaches its maximum at night. This method can effectively identify the dynamic laws of the data and provide a basis for subsequent fault diagnosis.

[0107] Factors with long-term changing trends, such as dynamic offset of illumination angle and periodic changes in humidity levels, are extracted from the environmental time series features. For analysis of illumination angle offset, we can focus on the frequency and amplitude of the offset. For example, if the illumination angle offsets 10 degrees per day and exceeds 12 degrees on two days per week, this can be determined to be a dynamic offset caused by external occlusion. For analysis of periodic changes in humidity levels, we can understand the changing trend of humidity levels and their impact on equipment. For example, humidity may rise sharply from 70% to 85% during the rainy season, persist for three days, and then fall back.

[0108] It is understandable that vegetation density will change the lighting and humidity environment of the equipment with the seasons. For example, a 20% increase in vegetation density in summer may reduce the lighting angle offset by 5 degrees and increase the humidity by 10%. By analyzing the seasonal modulation effect of vegetation density changes, we can understand the impact of vegetation density changes on the lighting and humidity environment of the equipment, as well as the potential impact on equipment failure. If the modulation impact intensity exceeds the standard, the long short-term memory network will adjust the time series data processing flow, assigning more weight to short-term fluctuations in humidity changes, reducing the long-term dependence on light offset, and obtaining the corrected time series correlation.

[0109] Based on the corrected time series correlation and the distribution of fault causes, the fault diagnosis result is obtained. For example, if the humidity periodically exceeds 80% and lasts for 5 days, and the fault probability increases from 10% to 30%, the fault diagnosis result is a circuit short circuit caused by excessive humidity.

[0110] In a second aspect, an embodiment of the present invention provides a fault analysis device for a power distribution device, see Figure 2 , which is a structural diagram of an embodiment of a fault analysis device for power distribution equipment provided by the present invention.

[0111] like Figure 2 As shown, the device includes:

[0112] A data acquisition module 21 is configured to acquire historical alarm data and geographic environment data of the power distribution equipment, and determine a trend of environmental interference changes based on the historical alarm data and the environment data; wherein the trend of environmental interference changes is used to represent the dynamic impact trend of each geographic environment factor on the occurrence of power distribution equipment alarms;

[0113] A feature extraction module 22 is configured to extract features from the historical alarm data using a convolutional neural network based on the change trend of the environmental interference to obtain a comprehensive feature representation;

[0114] A model training module 23 is used to train the convolutional neural network using the comprehensive feature representation to obtain a trained fault prediction model;

[0115] The fault prediction module 24 is used to obtain real-time alarm data and input it into the fault prediction model to output preliminary fault causes;

[0116] The fault diagnosis module 25 is used to perform time series analysis on the preliminary fault cause using a long short-term memory network to obtain a fault diagnosis result.

[0117] In an optional implementation, the data acquisition module 21 is further configured to:

[0118] Acquire historical alarm data of the power distribution equipment, and geographical environment data of the environment area where the power distribution equipment is located collected by various sensors; wherein the geographical environment data includes but is not limited to ambient temperature data, ambient humidity data, wind speed data, and vegetation coverage;

[0119] Calculate the equipment exposure distribution of the power distribution equipment in different seasons using a random forest algorithm based on the geographic environment data;

[0120] Calculating interference intensity of the device exposure of the power distribution equipment on the alarm frequency based on the device exposure distribution and the historical alarm data;

[0121] Using a clustering algorithm, the environmental area is divided into a high-impact area, a medium-impact area, and a low-impact area according to the geographical coordinates and the interference intensity;

[0122] Calculate the environmental interference change trend within a preset time period in each impact area; wherein, the environmental interference change trend is used to represent the dynamic impact trend of geographical environmental factors on the occurrence of distribution equipment alarms.

[0123] In an optional embodiment, the feature extraction module 22 is further configured to:

[0124] Using a convolutional neural network, feature extraction is performed on the historical alarm data to obtain a feature map;

[0125] Based on the historical alarm data and the feature map, an attention mechanism is used to calculate the influence weights of different geographical environment factors on the alarm;

[0126] The feature maps are weightedly fused according to the influence weights to obtain a comprehensive feature representation after weighted fusion.

[0127] In an optional embodiment, the feature extraction module 22 is further configured to:

[0128] Using a convolutional neural network, the historical alarm data is organized into a feature expression matrix according to time and space dimensions; wherein the matrix element value of the feature expression matrix is ​​the number of alarms at a specific geographical location within a specific time period;

[0129] Analyzing the correlation between the alarm frequency and the geographical environment according to the feature expression matrix and the geographical environment data;

[0130] The convolution kernel size of the convolutional neural network is adjusted according to the correlation, and the feature expression matrix is ​​processed using the adjusted convolutional neural network to obtain an adjusted feature map.

[0131] In an optional embodiment, the model training module 23 is further configured to:

[0132] Inputting the feature representation into a convolutional neural network for training to obtain a first probability distribution of historical alarms; wherein the first probability distribution is the failure probability of each fault type in the historical alarms;

[0133] When the failure probability of a certain fault type is greater than a preset failure probability threshold, the model parameters of the convolutional neural network are updated based on the environmental interference change trend to obtain a fault prediction model.

[0134] In an optional implementation, the fault prediction module 24 is further configured to:

[0135] Acquire real-time alarm data and geographic environment data of the power distribution equipment, input them into the fault prediction model, and output a second probability distribution corresponding to the alarm; wherein the second probability distribution is the fault probability of each fault type of the real-time alarm predicted by the fault prediction model;

[0136] Determining a maintenance difficulty index and a fault driving factor corresponding to each fault type in the second probability distribution; wherein the fault driving factor is a main driving factor causing the power distribution equipment to generate an alarm;

[0137] Calculating a contribution of the fault driving factor to the fault type, and adjusting the second probability distribution according to the contribution;

[0138] A preliminary fault cause is obtained according to the adjusted second probability distribution.

[0139] In an optional implementation, the fault diagnosis module 25 is further configured to:

[0140] Acquire geographic environment data within a preset time period before the real-time alarm occurs, analyze the geographic environment data using a long short-term memory network, and output environmental time series characteristics;

[0141] A fault diagnosis result is obtained based on the environmental time sequence characteristics and the preliminary fault cause.

[0142] In a third aspect, an embodiment of the present invention provides an electronic device, see Figure 3 , which is a schematic structural diagram of an electronic device provided by an embodiment of the present invention.

[0143] like Figure 3 As shown, the device includes:

[0144] Memory 31, for storing computer programs;

[0145] a processor 32, configured to execute the computer program;

[0146] Wherein, when the processor 32 executes the computer program, the fault analysis method for power distribution equipment as described in any of the above embodiments is implemented.

[0147] Exemplarily, the computer program may be divided into one or more modules / units, which are stored in the memory 31 and executed by the processor 32 to implement the present invention. The one or more modules / units may be a series of computer program instruction segments capable of implementing specific functions, and the instruction segments are used to describe the execution process of the computer program in the electronic device.

[0148] The processor 32 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0149] The memory 31 can be used to store the computer programs and / or modules. The processor 32 implements the various functions of the electronic device by running or executing the computer programs and / or modules stored in the memory 31 and accessing the data stored in the memory 31. The memory 31 may primarily include a program storage area and a data storage area. The program storage area may store an operating system and at least one application required for a function (such as sound playback or image playback); the data storage area may store data generated based on the use of the mobile phone (such as audio data and a phone book). Furthermore, the memory 31 may include high-speed random access memory (RAM) and non-volatile memory, such as a hard disk, internal memory, a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) card, a flash card, at least one disk storage device, a flash memory device, or other volatile solid-state storage device.

[0150] It should be noted that the above electronic devices include, but are not limited to, processors and memories. Those skilled in the art will understand that Figure 3 The structural diagram is merely an example of the electronic device described above and does not limit the electronic device. The electronic device may include more components than shown in the figure, or may combine certain components, or may include different components.

[0151] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed, the fault analysis method for the power distribution equipment described in any of the above embodiments is implemented.

[0152] It should be understood that the present invention can implement all or part of the processes in the above-mentioned method for analyzing the fault of power distribution equipment by instructing related hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, it can implement the steps of the above-mentioned method for analyzing the fault of power distribution equipment. The computer program includes computer program code, which can be in source code form, object code form, executable file, or some intermediate form. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium.

[0153] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. It should be pointed out that for those skilled in the art, several equivalent obvious variations and / or equivalent replacements can be made without departing from the technical principles of the present invention. These obvious variations and / or equivalent replacements should also be regarded as the scope of protection of the present invention.

Claims

1. A method for analyzing a fault of a power distribution device, characterized in that: include: Obtaining historical alarm data and geographical environment data of the power distribution equipment, and determining an environmental interference change trend based on the historical alarm data and the geographical environment data; wherein the environmental interference change trend is used to represent the dynamic impact trend of each geographical environment factor on the occurrence of power distribution equipment alarms; According to the changing trend of the environmental interference, a convolutional neural network is used to extract features from the historical alarm data to obtain a comprehensive feature representation; The convolutional neural network is trained using the comprehensive feature representation to obtain a trained fault prediction model; Acquire real-time alarm data and input it into the fault prediction model to output preliminary fault causes; Using a long short-term memory network to perform time series analysis on the preliminary fault cause to obtain a fault diagnosis result; The acquiring of historical alarm data and geographical environment data of the power distribution equipment, and determining a trend of environmental interference changes based on the historical alarm data and the geographical environment data, includes: Acquire historical alarm data of the power distribution equipment, and geographical environment data of the environment area where the power distribution equipment is located collected by various sensors; wherein the geographical environment data includes but is not limited to ambient temperature data, ambient humidity data, wind speed data, and vegetation coverage; Calculate the equipment exposure distribution of the power distribution equipment in different seasons using a random forest algorithm based on the geographic environment data; Calculating interference intensity of the device exposure of the power distribution equipment on the alarm frequency based on the device exposure distribution and the historical alarm data; Using a clustering algorithm, the environmental area is divided into a high-impact area, a medium-impact area, and a low-impact area according to the geographical coordinates and the interference intensity; Calculate the environmental interference change trend within a preset time period in each impact area; wherein, the environmental interference change trend is used to represent the dynamic impact trend of geographical environmental factors on the occurrence of distribution equipment alarms.

2. The fault analysis method for power distribution equipment according to claim 1, wherein: The method of extracting features from the historical alarm data using a convolutional neural network based on the changing trend of the environmental interference to obtain a comprehensive feature representation includes: Using a convolutional neural network, feature extraction is performed on the historical alarm data to obtain a feature map; Based on the historical alarm data and the feature map, an attention mechanism is used to calculate the influence weights of different geographical environment factors on the alarm; The feature maps are weightedly fused according to the influence weights to obtain a comprehensive feature representation after weighted fusion.

3. The fault analysis method for power distribution equipment according to claim 2, characterized in that: The convolutional neural network is used to extract features from the historical alarm data to obtain a feature map, including: Using a convolutional neural network, the historical alarm data is organized into a feature expression matrix according to time and space dimensions; wherein the matrix element value of the feature expression matrix is ​​the number of alarms at a specific geographical location within a specific time period; Analyzing the correlation between the alarm frequency and the geographical environment according to the feature expression matrix and the geographical environment data; The convolution kernel size of the convolutional neural network is adjusted according to the correlation, and the feature expression matrix is ​​processed using the adjusted convolutional neural network to obtain an adjusted feature map.

4. The fault analysis method for power distribution equipment according to claim 1, wherein: The method of training the convolutional neural network using the comprehensive feature representation to obtain a trained fault prediction model includes: Inputting the feature representation into a convolutional neural network for training to obtain a first probability distribution of historical alarms; wherein the first probability distribution is the failure probability of each fault type in the historical alarms; When the failure probability of a certain fault type is greater than a preset failure probability threshold, the model parameters of the convolutional neural network are updated based on the environmental interference change trend to obtain a fault prediction model.

5. The fault analysis method for power distribution equipment according to claim 1, wherein: The real-time alarm data is obtained and input into the fault prediction model to output preliminary fault causes, including: Acquire real-time alarm data and geographic environment data of the power distribution equipment, input them into the fault prediction model, and output a second probability distribution corresponding to the alarm; wherein the second probability distribution is the fault probability of each fault type of the real-time alarm predicted by the fault prediction model; Determining a maintenance difficulty index and a fault driving factor corresponding to each fault type in the second probability distribution; wherein the fault driving factor is a main driving factor causing the power distribution equipment to generate an alarm; Calculating a contribution of the fault driving factor to the fault type, and adjusting the second probability distribution according to the contribution; A preliminary fault cause is obtained according to the adjusted second probability distribution.

6. The fault analysis method for power distribution equipment according to claim 1, characterized in that: The use of the long short-term memory network to perform time series analysis on the preliminary fault cause to obtain a fault diagnosis result includes: Acquire geographic environment data within a preset time period before the real-time alarm occurs, analyze the geographic environment data using a long short-term memory network, and output environmental time series characteristics; A fault diagnosis result is obtained based on the environmental time sequence characteristics and the preliminary fault cause.

7. A fault analysis device for power distribution equipment, characterized in that: include: A data acquisition module, configured to acquire historical alarm data and geographic environment data of the power distribution equipment, and determine a trend of environmental interference changes based on the historical alarm data and the geographic environment data; wherein the trend of environmental interference changes is used to represent the dynamic impact trend of each geographic environment factor on the occurrence of power distribution equipment alarms; A feature extraction module is used to extract features from the historical alarm data using a convolutional neural network according to the change trend of the environmental interference to obtain a comprehensive feature representation; A model training module, configured to train the convolutional neural network using the comprehensive feature representation to obtain a trained fault prediction model; A fault prediction module is used to obtain real-time alarm data and input it into the fault prediction model to output preliminary fault causes; a fault diagnosis module, configured to perform a time series analysis on the preliminary fault cause using a long short-term memory network to obtain a fault diagnosis result; The data acquisition module is also used for: Acquire historical alarm data of the power distribution equipment, and geographical environment data of the environment area where the power distribution equipment is located collected by various sensors; wherein the geographical environment data includes but is not limited to ambient temperature data, ambient humidity data, wind speed data, and vegetation coverage; Calculate the equipment exposure distribution of the power distribution equipment in different seasons using a random forest algorithm based on the geographic environment data; Calculating interference intensity of the device exposure of the power distribution equipment on the alarm frequency based on the device exposure distribution and the historical alarm data; Using a clustering algorithm, the environmental area is divided into a high-impact area, a medium-impact area, and a low-impact area according to the geographical coordinates and the interference intensity; Calculate the environmental interference change trend within a preset time period in each impact area; wherein, the environmental interference change trend is used to represent the dynamic impact trend of geographical environmental factors on the occurrence of distribution equipment alarms.

8. An electronic device, characterized in that: include: memory for storing computer programs; a processor for executing the computer program; Wherein, when the processor executes the computer program, the fault analysis method for power distribution equipment according to any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed, the fault analysis method for power distribution equipment according to any one of claims 1 to 6 is implemented.

Images