Dual-layer perception-based distribution network cable environment perception method and system

Abstract

The present invention provides a dual-layer perception-based distribution network cable environment perception method, comprising: acquiring external environment data and internal operation data of a cable in a region to be measured; fusing the external environment data and the internal operation data, to obtain fused data; performing feature extraction on the fused data, to obtain temporary feature data; acquiring image data of the cable in the region to be measured; and inputting the external environment data, the temporary feature data and the image data into a preset environment evaluation model, to obtain an environment perception evaluation result of the cable in the region to be measured, the preset environment evaluation model being obtained on the basis of an improved CNN module and a Transformer module. The present invention improves the accuracy and effectiveness of cable environment perception, helps to find a fault in advance and take corresponding measures, and ensures the safe and stable operation of a network distribution system.

Description

A method and system for environmental sensing of distribution network cables based on dual-layer sensing Technical Field

[0001] This invention relates to the field of power system technology, and more specifically, to a method and system for sensing the environment of distribution network cables based on two-layer sensing. Background Technology

[0002] Cables, as the main channels for power transmission and distribution, are crucial to power distribution networks. However, if cable faults occur and cannot be monitored and repaired in a timely manner, it can lead to serious consequences such as power outages, equipment damage, and even fires. Therefore, timely monitoring of cable status and environmental sensing is particularly important. Cable environmental sensing, in particular, helps to achieve comprehensive monitoring of the environment surrounding the cable, promptly grasp the operating status of the power grid, reduce the probability of cable faults, and improve the safety and stability of the power distribution network system.

[0003] Currently, cable environmental sensing mainly relies on manual inspections and periodic checks, which suffers from drawbacks such as low efficiency, limited coverage, and difficulty in achieving comprehensive monitoring. Moreover, cable environmental sensing methods are often limited to a single data source, resulting in low accuracy in assessing cable condition. Technical issues

[0004] In view of the above-mentioned shortcomings of the prior art, the present invention addresses the problem of how to improve the accuracy of cable environment perception, thereby improving the reliability and safety of power systems. Solution

[0005] To address the aforementioned problems, this invention provides a method and device for sensing the environment of distribution network cables based on dual-layer sensing.

[0006] In a first aspect, the present invention provides a method for sensing the environment of distribution network cables based on dual-layer sensing, comprising:

[0007] Acquire external environmental data and internal operating data of the cable in the area to be tested;

[0008] The external environment data and the internal operation data are fused to obtain fused data, and features are extracted from the fused data to obtain temporary feature data;

[0009] Acquire image data of the cable in the area to be tested;

[0010] The external environment data, the temporary feature data, and the image data are input into a preset environmental assessment model to obtain the environmental perception assessment result of the cable in the area to be tested. The preset environmental assessment model is based on an improved CNN module and a Transformer module.

[0011] Optionally, the construction process of the improved CNN module includes:

[0012] The improved CNN module is obtained by adding a feature enhancement module to the main branch of the basic CNN model.

[0013] The feature enhancement module is based on a one-dimensional convolutional layer or a long short-term memory network.

[0014] Optionally, the step of inputting the external environment data, the temporary feature data, and the image data into a preset environmental assessment model to obtain the environmental perception assessment result of the cable in the area to be tested includes:

[0015] The improved CNN module is used to extract features from the image data to obtain first feature data;

[0016] Feature extraction is performed on the temporary feature data and the external environment data respectively to obtain the second feature data and the third feature data;

[0017] The first feature data, the second feature data, and the third feature data are input into the Transformer module to obtain the environmental perception assessment results of the cable in the area to be tested.

[0018] Optionally, the preset environmental assessment model further includes a self-attention module. The step of inputting the first feature data, the second feature data, and the third feature data into the Transformer module to obtain the environmental perception assessment result of the cable in the area to be tested includes:

[0019] The first feature data and the third feature data are input into the self-attention module to obtain fused feature data;

[0020] The fused feature data and the second feature data are input into the Transformer module to obtain the environmental perception assessment result of the cable in the area to be tested.

[0021] Optionally, inputting the first feature data and the third feature data into the self-attention module to obtain fused feature data includes:

[0022] Perform linear transformations on the first feature data and the third feature data respectively to obtain the corresponding first vector data and second vector data;

[0023] The attention weights are obtained based on the first vector data and the second vector data;

[0024] Based on the attention weights, the first vector data and the second vector data are fused to obtain fused feature data.

[0025] Optionally, the step of fusing the external environment data and the internal operating data to obtain fused data, and extracting features from the fused data to obtain temporary feature data, includes:

[0026] The external environment data and the internal operation data are preprocessed separately, and the processed data are then spliced ​​together to obtain the fused data;

[0027] Based on statistical algorithms, features are extracted from the fused data to obtain the temporary feature data.

[0028] Optionally, the preprocessing of the external environment data and the internal operating data, and the concatenation of the processed data to obtain the fused data, includes:

[0029] The external environment data and the internal operation data are preprocessed respectively to obtain the first temporary vector data and the second temporary vector data;

[0030] Obtain preset weight data, wherein the preset weight data includes first weight data and second weight data;

[0031] The weighted average value corresponding to the external environment data is calculated based on the first temporary vector data and the first weight data, and the weighted average value corresponding to the internal operation data is calculated based on the second temporary vector data and the second weight data.

[0032] The weighted average value corresponding to the external environment data and the weighted average value corresponding to the internal operation data are integrated to obtain the fused data.

[0033] Optionally, the distribution network cable environment sensing method based on dual-layer sensing further includes:

[0034] Based on the warning threshold, the warning status is determined according to the environmental perception assessment results.

[0035] Based on the aforementioned warning status, a warning message is issued.

[0036] Secondly, the present invention provides a distribution network cable environment sensing system based on dual-layer sensing, comprising:

[0037] The acquisition unit is used to acquire external environmental data and internal operating data of the cable in the area under test;

[0038] The processing unit is used to fuse the external environment data and the internal operating data to obtain fused data, and to extract features from the fused data to obtain temporary feature data.

[0039] The acquisition unit is also used to acquire image data of the cable in the area to be tested;

[0040] The processing unit is also used to input the external environment data, the temporary feature data and the image data into a preset environment assessment model to obtain the environmental perception assessment result of the cable in the area to be tested. The preset environment assessment model is obtained based on the CNN module and the Transformer module.

[0041] Thirdly, the present invention provides an electronic device, including a memory and a processor;

[0042] The memory is used to store computer programs;

[0043] The processor is configured to, when executing the computer program, implement the distribution network cable environment sensing method based on two-layer sensing as described in the first aspect.

[0044] Fourthly, the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the distribution network cable environment sensing method based on dual-layer sensing as described in the first aspect. Beneficial effects

[0045] The beneficial effects of the dual-layer sensing method and device for distribution network cables of the present invention are as follows: By acquiring external environmental data and internal operating data of the cable, wherein the external environmental data may include data such as temperature, humidity, vibration data, and electromagnetic field, and the internal operating data may include current data, the external environmental data and internal operating data are fused, that is, the influencing factors of distribution network cable environmental sensing are fused, providing a data foundation for subsequent cable environmental sensing. Furthermore, by combining image data and comprehensive analysis, the fused data obtained is more comprehensive and can provide more accurate environmental sensing assessment results. Moreover, through feature extraction and comprehensive analysis, temporary feature data can be obtained, which helps to improve the accuracy of the assessment. Finally, a pre-defined environmental assessment model was constructed using CNN and Transformer modules. These modules have good performance in image recognition and sequence modeling, which can improve the accuracy and efficiency of environmental perception assessment. Therefore, based on the dual-layer perception design, that is, by sensing and collecting data on the working status of the physical layer cable (such as external environmental data and internal operating data), combined with the intelligent algorithm analysis and processing of the decision layer, a comprehensive perception and monitoring of the cable environment and working status can be achieved. This enables the system to obtain more information and better integrate various data sources, thereby improving the level of cable environmental perception, helping to reduce the probability of fault occurrence and improve system safety.

[0046] Therefore, this invention utilizes external and internal data sources, as well as image data. Information from these different sources can provide a more comprehensive and multi-dimensional environmental perception. Combined with the preset environmental assessment model built by the CNN module and the Transformer module, it can gain a more comprehensive understanding of the environmental conditions of the cable, thereby improving the level of environmental perception. That is, it improves the accuracy and effectiveness of cable environmental perception, which helps to detect faults in advance and take corresponding measures to ensure the safe and stable operation of the distribution network system. Attached Figure Description

[0047] Figure 1 is a schematic flowchart of one of the distribution network cable environment sensing methods based on dual-layer sensing according to an embodiment of the present invention;

[0048] Figure 2 is a second schematic flowchart of a distribution network cable environment sensing method based on dual-layer sensing according to an embodiment of the present invention;

[0049] Figure 3 is a schematic diagram of the structure of a distribution network cable environment sensing system based on dual-layer sensing according to an embodiment of the present invention;

[0050] Figure 4 is a schematic diagram of the structure of an electronic device according to an embodiment of the present invention. Embodiments of the present invention

[0051] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0052] The term "comprising" and its variations as used herein are open-ended, meaning "including but not limited to"; the term "based on" means "at least partially based on"; the term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments"; and the term "optionally" means "optional embodiments". Definitions of other terms will be given in the following description. It should be noted that the concepts of "first," "second," etc., mentioned in this invention are used only to distinguish different devices, modules, or units, and are not intended to limit the order of functions performed by these devices, modules, or units or their interdependencies.

[0053] It should be noted that the terms "a" and "a plurality of" used in this invention are illustrative rather than restrictive. Those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0054] This embodiment provides a method, device, and electronic device for environmental sensing of distribution network cables based on dual-layer sensing.

[0055] As shown in Figure 1, an embodiment of the present invention provides a distribution network cable environment sensing method based on dual-layer sensing, comprising:

[0056] S100 acquires external environmental data and internal operating data of the cable in the area under test.

[0057] Specifically, external environmental data may include temperature, humidity, vibration data, and electromagnetic fields, which can be acquired in real time using sensors and other devices. Internal operating data refers to the operating status and parameters of the cable itself, such as current data and voltage data. The acquired external environmental data and internal operating data provide the basic data for the subsequent environmental sensing process.

[0058] S200, the external environment data and the internal operation data are fused to obtain fused data; and features are extracted from the fused data to obtain temporary feature data.

[0059] Specifically, fusing external environmental data and internal operational data and extracting features yields more comprehensive and complete temporary feature data, which facilitates more accurate analysis and evaluation of the cable's operational status. When fusing external environmental data and internal operational data, temporal and spatial correspondences must be established between the two data sets to ensure data consistency. Furthermore, the data needs to be formatted and standardized to facilitate subsequent processing. Data fusion algorithms can be used to integrate external environmental data and internal operational data, forming a fused dataset.

[0060] Relevant features are extracted from the fused data. These features can be statistical quantities, frequency domain features, time domain features, etc., ensuring that the extracted features reflect important information about the cable's operating status. In other words, the temporary feature data can contain multiple features, reflecting the comprehensive characteristics of the cable's environment and operating status. This feature data can be used in subsequent applications such as model building, data analysis, and decision support.

[0061] By integrating external environmental data and internal operational data, more comprehensive and integrated data can be provided, which helps to conduct in-depth analysis of the cable's environmental perception. Furthermore, through feature extraction, the obtained temporary feature data can more accurately reflect the cable's operating status, enabling more precise environmental perception and assessment. Moreover, temporary feature data can provide important references for the system's automated operation and decision-making, improving system operating efficiency and stability.

[0062] S300, acquire image data of the cable in the area to be tested.

[0063] Specifically, a camera device can be installed at an appropriate location near the area to be tested to ensure that the surrounding environment of the cable can be captured. The camera device can be used independently or in conjunction with other sensor devices to obtain more comprehensive environmental data.

[0064] By capturing real-time images of the area around the cable, changes in the surrounding environment can be obtained in a timely manner. The image data can also be used to monitor the condition of the cable surface, such as damage, corrosion, and abnormal hot spots, providing important reference for cable maintenance.

[0065] Acquiring image data of cables in the area under test helps to enable real-time monitoring of the environment around the cables and quantitative assessment of the cable surface conditions, providing more comprehensive data support for environmental perception and decision-making.

[0066] S400, the external environment data, the temporary feature data and the image data are input into the preset environment assessment model to obtain the environmental perception assessment result of the cable in the area to be tested. The preset environment assessment model is obtained based on the CNN module and the Transformer module.

[0067] Specifically, external environment data, temporary feature data, and image data need to be preprocessed to ensure consistency in data format and dimensions. For example, external environment data and temporary feature data can be processed through feature extraction and other methods to be combined with image data and input into the model.

[0068] Image data is first processed through a CNN (Convolutional Neural Network) module for feature extraction and image recognition, extracting features from the image that are relevant to the cable's surrounding environment. The CNN module can identify damage to the cable surface, changes in the surrounding environment, and other conditions, providing important information for subsequent environmental perception and assessment.

[0069] Feature extraction is performed on external environmental data and temporary feature data separately. Finally, the output data of each feature extraction module is integrated through the Transformer module to obtain the environmental perception assessment result of the cable in the test area. That is, the model evaluates the environmental perception of the cable in the test area by learning the correlations and patterns between data. The final environmental perception assessment result is a comprehensive information on the environmental state of the cable in the test area. This comprehensive information summarizes the analysis results of external environmental data (such as temperature, humidity, vibration data, electromagnetic field), internal operating data (current information), temporary feature data, and image data. It should be noted that the environmental perception assessment result can be state information or comprehensive score data, and the specific assessment result can be set according to the actual situation.

[0070] Therefore, combining external environmental data, temporary feature data, and image data into the input model can improve the accuracy and comprehensiveness of environmental perception by utilizing multi-source information. Furthermore, the CNN and Transformer modules process different types of data respectively, and model fusion can comprehensively utilize information from different data types, improving the accuracy and robustness of environmental perception assessment. In other words, inputting external environmental data, temporary feature data, and image data into the CNN and Transformer modules of a pre-defined environmental assessment model enables the comprehensive utilization of multi-source information, improving the accuracy and effectiveness of environmental perception assessment.

[0071] In this embodiment, external environmental data and internal operational data of the cable are acquired. External environmental data may include temperature, humidity, vibration, and electromagnetic field data, while internal operational data may include current data. These two data are then fused, integrating the influencing factors of the distribution network cable environment perception. This provides a data foundation for subsequent cable environment perception. Combined with image data, the fused data obtained through comprehensive analysis is more comprehensive and provides more accurate environmental perception assessment results. Furthermore, through feature extraction and comprehensive analysis, temporary feature data can be obtained, which helps improve the accuracy of the assessment. Finally, a pre-defined environmental assessment model constructed using CNN and Transformer modules is employed. These modules have good performance in image recognition and sequence modeling, improving the accuracy and efficiency of environmental perception assessment. Therefore, based on a two-layer perception design—that is, by perceiving and collecting data on the physical layer cable operating status (such as external environmental data and internal operational data), combined with intelligent algorithm analysis and processing at the decision layer—comprehensive perception and monitoring of the cable environment and operating status is achieved. This allows the system to obtain more information and better integrate various data sources, thereby improving the level of cable environmental perception, helping to reduce the probability of faults and improve system safety.

[0072] Therefore, this embodiment utilizes external and internal data sources, as well as image data. Information from these different sources can provide a more comprehensive and multi-dimensional environmental perception. Combined with the preset environmental assessment model built by the CNN module and the Transformer module, it can gain a more comprehensive understanding of the environmental conditions of the cable, thereby improving the level of environmental perception. That is, it improves the accuracy and effectiveness of cable environmental perception, which helps to detect faults in advance and take corresponding measures to ensure the safe and stable operation of the distribution network system.

[0073] Optionally, the CNN module includes a basic CNN module and an improved CNN module, and the construction process of the improved CNN module includes:

[0074] The improved CNN module is obtained by adding a feature enhancement module to the main branch of the basic CNN model.

[0075] The feature enhancement module is based on a one-dimensional convolutional layer or a long short-term memory network.

[0076] Specifically, when building an improved CNN module, the first step is to utilize the backbone of the basic CNN model and add a feature enhancement module to it, thereby obtaining the improved CNN module. The basic CNN model typically includes components such as convolutional layers, pooling layers, and fully connected layers to extract features from the input image.

[0077] In the backbone of a basic CNN model, features often gradually decrease, potentially resulting in the loss of some information. Therefore, a feature enhancement module is needed to further extract and enhance features. This module can be built based on one-dimensional convolutional layers or LSTMs. One-dimensional convolutional layers are suitable for processing sequential data, capturing local correlations within the data; while LSTMs are suitable for processing time-series data, learning long-term dependencies within the sequence. By using one-dimensional convolutional layers or LSTMs as part of the feature enhancement module, more high-level features are extracted from the backbone of the basic CNN model, thereby increasing the model's understanding and representation capabilities of the input data.

[0078] By adding a feature enhancement module, the improved CNN module can better capture abstract features and local relationships in the input data, improving the feature representation ability. The improved CNN module makes the basic CNN model more generalizable and robust, thereby improving the model's performance on different tasks. Furthermore, the introduction of the feature enhancement module helps the model better handle complex data structures and relationships, improving the model's understanding of the input data.

[0079] Optionally, the step of inputting the external environment data, the temporary feature data, and the image data into a preset environmental assessment model to obtain the environmental perception assessment result of the cable in the area to be tested includes:

[0080] The improved CNN module is used to extract features from the image data to obtain first feature data;

[0081] Feature extraction is performed on the temporary feature data and the external environment data respectively to obtain the second feature data and the third feature data;

[0082] The first feature data, the second feature data, and the third feature data are input into the Transformer module to obtain the environmental perception assessment results of the cable in the area to be tested.

[0083] Specifically, using an improved CNN module to extract features from image data can better capture abstract features and local correlations in the input data, thereby improving the feature representation capability.

[0084] Feature extraction is performed on temporary feature data and external environment data to obtain second feature data and third feature data, which respectively contain important feature information about temporary feature data and external environment data.

[0085] The first, second, and third feature data are input into the Transformer module. The Transformer module helps the model better understand and integrate feature data from these different sources, thereby improving the accuracy of environmental perception assessment. This module can learn the correlations between features, further optimizing the assessment results of the cable environment in the tested area.

[0086] Assuming the image data displays visual information about the cable's surroundings, and the temporary feature data includes the cable's temperature change rate, while external environmental data includes temperature, humidity, and other information, the model can be used to obtain the cable's operating status under different temperature, humidity, and other environmental conditions. This helps in early maintenance and prevention of faults.

[0087] By integrating multiple data sources and combining improved CNN and Transformer modules, a more comprehensive understanding of the environment surrounding the cable in the test area can be achieved, improving environmental awareness, reducing the likelihood of unexpected failures, lowering maintenance costs, and enhancing equipment reliability and safety.

[0088] Optionally, the preset environmental assessment model further includes a self-attention module. The step of inputting the first feature data, the second feature data, and the third feature data into the Transformer module to obtain the environmental perception assessment result of the cable in the area to be tested includes:

[0089] The first feature data and the third feature data are input into the self-attention module to obtain fused feature data;

[0090] The fused feature data and the second feature data are input into the Transformer module to obtain the environmental perception assessment result of the cable in the area to be tested.

[0091] Specifically, firstly, the first and third feature data are input into the self-attention module. Self-attention is a mechanism used to dynamically calculate weights for each position in the input sequence to better capture the dependencies between different positions in the sequence. Through the self-attention module, the model learns and focuses on the most important parts of the first and third feature data, generating fused feature data that contains key information from both data sources.

[0092] The fused feature data and the second feature data are input into the Transformer module. The Transformer module can further integrate and process these feature data, thereby improving the accuracy and robustness of the evaluation results. The obtained environmental perception evaluation results of the cable in the test area comprehensively consider image features, temporary features, and external environmental data, providing a more comprehensive understanding of the environment surrounding the cable.

[0093] By leveraging information from diverse data sources and employing self-attention and Transformer mechanisms, the correlations between features can be more effectively integrated and learned, improving the accuracy and robustness of environmental perception assessment. This comprehensive approach helps improve equipment operating efficiency, extend cable life, and enhance equipment reliability and safety.

[0094] Optionally, inputting the first feature data and the third feature data into the self-attention module to obtain fused feature data includes:

[0095] Perform linear transformations on the first feature data and the third feature data respectively to obtain the corresponding first vector data and second vector data;

[0096] The attention weights are obtained based on the first vector data and the second vector data;

[0097] Based on the attention weights, the first vector data and the second vector data are fused to obtain fused feature data.

[0098] Specifically, linear transformations are performed on the first and third feature data to obtain the corresponding first and second vector data. The linear transformation maps the original feature data to a new vector space through matrix multiplication and the addition of bias terms, and learns the relationships and feature representations between the data.

[0099] Attention weights are derived from the first and second vector data. These weights measure the importance of each feature data point, allowing for greater focus on key information during fusion. Attention weights are obtained by calculating the similarity or correlation between the first and second vector data, and then normalizing it using softmax to obtain the weight values.

[0100] For example, a self-attention module actually consists of three key parts: query, key, and value. First, for the first and third feature data, a linear transformation (usually involving a fully connected layer) is applied to obtain the query vector, key vector, and value vector, respectively. The purpose of this linear transformation is to map the original feature data to a higher-dimensional space, allowing the model to better learn the relationships between the data.

[0101] For each query vector, its similarity to all key vectors needs to be calculated. This can be obtained by taking the dot product of the query vector and the key vectors. Then, these similarities are subjected to a softmax operation to obtain attention weights that have a normalizing effect.

[0102] For each query vector, based on the calculated attention weights, the attention weights are used as the corresponding numerical vectors for weighting and then summed in a weighted manner to obtain a weighted numerical representation based on the attention weights. This step is essentially the model learning how to fuse information between different elements.

[0103] For the first and third feature data, weighted numerical representations based on the self-attention mechanism are obtained through the above steps. Then, these two fused numerical representations are added or concatenated to obtain the final fused feature data.

[0104] In summary, the process of inputting the first and third feature data into the attention module mainly involves calculating the similarity between the query vector and the key vector to obtain attention weights, then multiplying the attention weights by the numerical vectors to obtain a weighted numerical representation, and finally fusing the obtained weighted numerical representations to obtain the final fused feature data. This method enables the Transformer model to capture the complex relationships between the input data and generate more representative and informative feature representations.

[0105] Based on the obtained attention weights, the first and second vector data are weighted and fused to obtain fused feature data. The fusion process involves weighted summation of the two vector data, ensuring that important features are retained after fusion, while the influence of unimportant features is reduced.

[0106] Through the above processing, information from different feature data can be utilized more fully, and key information from the data can be effectively fused together through the operation of the self-attention mechanism. This method of fusing feature data can help improve the sensitivity and accuracy of the evaluation model to environmental changes, and contribute to improving the efficiency of cable maintenance and the reliability of equipment.

[0107] Optionally, the step of fusing the external environment data and the internal operating data to obtain fused data, and extracting features from the fused data to obtain temporary feature data, includes:

[0108] The external environment data and the internal operation data are preprocessed separately, and the processed data are then spliced ​​together to obtain the fused data;

[0109] Based on statistical algorithms, features are extracted from the fused data to obtain the temporary feature data.

[0110] Specifically, the first step involves preprocessing both external environmental data and internal operational data, including data cleaning, missing value handling, and feature normalization. The processed data may include environmental data such as temperature, humidity, voltage, and current, as well as internal operational data such as equipment operating status and fault records. The preprocessed data is then concatenated to obtain fused data.

[0111] Statistical algorithms, such as mean, standard deviation, maximum, and minimum, are used to extract features from the fused data. These statistical features can reflect the overall distribution, average level, and degree of variation of the data, helping to reveal the patterns and characteristics of the data.

[0112] After feature extraction using statistical algorithms, provisional feature data is obtained. This feature data may include various statistical indicators, such as average temperature, maximum humidity, and standard deviation of current. This provisional feature data helps to understand the overall characteristics of the data and provides a foundation for subsequent modeling and analysis.

[0113] Integrating external environmental data and internal operational data, and extracting statistical features, helps to comprehensively consider the impact of environmental factors and equipment operating status on equipment performance. By acquiring temporary feature data, we can better understand the characteristics and patterns of the data, providing a more comprehensive and accurate feature representation for subsequent analysis, evaluation, and modeling, thereby improving the efficiency of equipment maintenance and the reliability of equipment operation.

[0114] Optionally, the preprocessing of the external environment data and the internal operating data, and the concatenation of the processed data to obtain the fused data, includes:

[0115] The external environment data and the internal operation data are preprocessed respectively to obtain the first temporary vector data and the second temporary vector data;

[0116] Obtain preset weight data, wherein the preset weight data includes first weight data and second weight data;

[0117] The weighted average value corresponding to the external environment data is calculated based on the first temporary vector data and the first weight data, and the weighted average value corresponding to the internal operation data is calculated based on the second temporary vector data and the second weight data.

[0118] The weighted average value corresponding to the external environment data and the weighted average value corresponding to the internal operation data are integrated to obtain the fused data.

[0119] Specifically, preprocessing of external environmental data involves data cleaning and standardization to ensure consistency and comparability. Preprocessing of internal operational data involves handling missing values, detecting outliers, and transforming the data to meet processing requirements. Furthermore, data consistency must also be ensured during the processing of internal operational data.

[0120] For example: the processed external environment data, i.e., the first temporary vector data VE = [E1, E2, ..., EN]; the processed internal operation data, i.e., the first temporary vector data VI = [I1, I2, ..., IM];

[0121] Obtaining preset weight data means determining their weights in the fusion process, i.e., their respective importance ratios. These weights can be set in advance based on experience, or determined through data analysis or model training.

[0122] If the weights of the external environment data are set as w1, w2, ..., wN, and the weights of the internal operating data are set as v1, v2, ..., vM, then the weights of the second weighted data are set as w1, v2, ..., vM.

[0123] Calculate the weighted average of the external environment data and the internal operating data separately. The weighted average of the external environment data is: WE = w1 * E1 + w2 * E2 + ... + wN * EN; the weighted average of the internal operating data is: WI = v1 * I1 + v2 * I2 + ... + vM * IM.

[0124] Finally, the weighted average of the external environmental data and the internal operational data is integrated to obtain fused data: WE + WI. This fused data takes into account preprocessed data, weighted data, and the calculation of the weighted average to comprehensively reflect the combined situation of external environmental data and internal operational data. This integration method can provide more comprehensive and accurate information for environmental perception assessment.

[0125] Optionally, the distribution network cable environment sensing method based on dual-layer sensing further includes:

[0126] Based on the warning threshold, the warning status is determined according to the environmental perception assessment results.

[0127] Based on the aforementioned warning status, a warning message is issued.

[0128] Specifically, firstly, based on the environmental perception assessment results and the set warning thresholds, it is determined whether the current cable environment requires a warning. The warning thresholds are indicators set based on historical data and expert experience. For example, if the final environmental perception assessment result is a comprehensive score, the warning thresholds are set as follows: 50-60 is normal, less than 50 is an abnormal alarm, and more than 60 is an emergency alarm; that is, a comprehensive score of different environmental perception indicators can more comprehensively reflect the cable environment status. Through simple scoring, operators can quickly understand the overall cable environment status, facilitating rapid decision-making and handling.

[0129] By setting warning thresholds of 50-60 as normal, less than 50 as an abnormal alarm, and more than 60 as an emergency alarm, the system takes into account the environmental conditions corresponding to different warning levels. When the environmental perception assessment result reaches the emergency alarm level, it indicates a very serious problem, requiring immediate action to prevent malfunctions and safety incidents.

[0130] Setting early warning thresholds not only ensures the safety and stability of cable system operation, but also allows operators to have a clear understanding of the environmental conditions, thus helping to improve operation and maintenance efficiency.

[0131] Once the warning status is confirmed, the procedure for issuing warning information will be triggered immediately.

[0132] Warning information can be disseminated through various channels, such as SMS, email, system alerts, and app push notifications. Warning information can be simultaneously sent to the monitoring center, maintenance personnel, and management personnel so that they can take timely action.

[0133] Warning information typically needs to include the warning level, warning content, possible impact, and recommended handling measures so that the recipient can respond quickly and accurately.

[0134] Real-time early warnings help managers and maintenance personnel quickly understand the cable environment status, identify potential problems early, and prevent failures. Timely dissemination of warning information improves the efficiency and accuracy of fault handling, helping to reduce the impact of faults on the system and users. It also helps strengthen the monitoring and management of cable environment safety, improving equipment reliability and safety.

[0135] In some embodiments, as shown in Figure 3, the process of the distribution network cable environmental sensing method based on dual-layer sensing is as follows: First, acquire the external environmental data and internal operating data of the cable in the area to be measured. The external environmental data includes temperature, humidity, vibration data and electromagnetic field, and the internal operating data includes current data, which includes information such as current magnitude, waveform and frequency. It should be noted that for each data in the external environmental data and internal operating data, the mean, maximum value and peak value can be selected, and the specific selection can be set according to the actual situation.

[0136] External environmental data and internal operational data are fused to obtain fused data. This can be achieved through data normalization or standardization: mapping values ​​from different ranges to a unified standard range to eliminate the influence of different dimensions. Feature extraction is then performed on the fused data to obtain temporary feature data.

[0137] The acquired image data, external environment data, and temporary feature data of the cable in the test area are input into a preset environmental assessment model to obtain the environmental perception assessment result of the cable in the test area. The processing procedure of the preset environmental assessment model is as follows: an improved CNN module extracts features from the image data to obtain first feature data; different Transformer modules extract features from the temporary feature data and external environment data to obtain second and third feature data; the first and third feature data are input into a self-attention module to obtain fused feature data; finally, the fused feature data and the second feature data are input into the Transformer module to obtain the environmental perception assessment result of the cable in the test area.

[0138] It should be noted that the self-attention module can choose the same network structure as the Transformer module, and the specific choice depends on the actual situation.

[0139] As shown in Figure 3, an embodiment of the present invention provides a distribution network cable environment sensing system 300 based on dual-layer sensing, comprising:

[0140] Acquisition unit 310 is used to acquire external environmental data and internal operating data of the cable in the area under test;

[0141] Processing unit 320 is used to fuse the external environment data and the internal operating data to obtain fused data; and to extract features from the fused data to obtain temporary feature data;

[0142] The acquisition unit 310 is also used to acquire image data of the cable in the area to be tested;

[0143] The processing unit 320 is further configured to input the external environment data, the temporary feature data and the image data into a preset environment assessment model to obtain the environmental perception assessment result of the cable in the area to be tested. The preset environment assessment model is obtained based on the CNN module and the Transformer module.

[0144] The distribution network cable environment sensing system based on dual-layer sensing in this embodiment is used to implement the distribution network cable environment sensing method based on dual-layer sensing as described above. Its advantages over the prior art are the same as the advantages of the distribution network cable environment sensing method based on dual-layer sensing over the prior art, and will not be repeated here.

[0145] Optionally, the processing unit 320 is specifically used to: add a feature enhancement module to the backbone branch of the basic CNN model to obtain the improved CNN module; wherein the feature enhancement module is obtained based on a one-dimensional convolutional layer or a long short-term memory network.

[0146] Optionally, the processing unit 320 is further configured to: preprocess the external environment data and the internal operating data respectively, and concatenate the processed data to obtain the fused data; and extract features from the fused data based on statistical algorithms to obtain the temporary feature data.

[0147] As shown in Figure 4, an electronic device 400 provided in this embodiment of the invention includes a memory 410 and a processor 420; the memory 410 is used to store a computer program; the processor 420 is used to implement the above-described method for sensing the environment of distribution network cables based on dual-layer sensing when executing the computer program.

[0148] Alternatively, an electronic device 400 includes a memory 410 and a processor 420 coupled to the memory 410; the memory 410 is configured to store a computer program; and the processor 420 is configured to perform the following operations when the computer program is executed:

[0149] Acquire external environmental data and internal operating data of the cable in the area to be tested;

[0150] The external environment data and internal operation data are fused to obtain fused data; and features are extracted from the fused data to obtain temporary feature data.

[0151] Acquire image data of the cable in the area to be tested;

[0152] The external environment data, the temporary feature data, and the image data are input into a preset environmental assessment model to obtain the environmental perception assessment result of the cable in the area to be tested. The preset environmental assessment model is based on an improved CNN module and a Transformer module.

[0153] This invention provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the distribution network cable environment sensing method based on two-layer sensing as described above.

[0154] Alternatively, a non-volatile computer-readable storage medium storing a computer program that, when executed by a processor, causes the processor to perform the following operations:

[0155] Acquire external environmental data and internal operating data of the cable in the area to be tested;

[0156] The external environment data and internal operation data are fused to obtain fused data; and features are extracted from the fused data to obtain temporary feature data.

[0157] Acquire image data of the cable in the area to be tested;

[0158] The external environment data, the temporary feature data, and the image data are input into a preset environmental assessment model to obtain the environmental perception assessment result of the cable in the area to be tested. The preset environmental assessment model is based on an improved CNN module and a Transformer module.

[0159] The present invention will now be described an electronic device 400 that can serve as a server or client of the present invention, which is an example of a hardware device that can be applied to various aspects of the present invention. Electronic device 400 is intended to represent various forms of digital electronic computer devices, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. Electronic device 400 can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0160] Electronic device 400 includes a computing unit that can perform various appropriate actions and processes based on a computer program stored in read-only memory (ROM) or a computer program loaded from a storage unit into random access memory (RAM). The RAM may also store various programs and data required for device operation. The computing unit, ROM, and RAM are interconnected via a bus. Input / output (I / O) interfaces are also connected to the bus.

[0161] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc. In this application, the units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of the embodiments of the present invention according to actual needs. Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated units can be implemented in hardware or as software functional units.

Claims

1. A method for sensing the environment of distribution network cables based on dual-layer sensing, characterized in that, include: Acquire external environmental data and internal operating data of the cable in the area to be tested; The external environment data and the internal operating data are fused to obtain fused data, including: The external environment data and the internal operating data are preprocessed separately, and the processed data are then concatenated to obtain the fused data, including: The external environment data and the internal operation data are preprocessed respectively to obtain the first temporary vector data and the second temporary vector data; Obtain preset weight data, wherein the preset weight data includes first weight data and second weight data; The weighted average value corresponding to the external environment data is calculated based on the first temporary vector data and the first weight data, and the weighted average value corresponding to the internal operation data is calculated based on the second temporary vector data and the second weight data. The weighted average value corresponding to the external environment data and the weighted average value corresponding to the internal operation data are integrated to obtain the fused data; Feature extraction is performed on the fused data to obtain temporary feature data; Acquire image data of the cable in the area to be tested; The external environment data, the temporary feature data, and the image data are input into a preset environmental assessment model to obtain the environmental perception assessment result of the cable in the area to be tested. The preset environmental assessment model is based on an improved CNN module and a Transformer module and includes: The improved CNN module is used to extract features from the image data to obtain first feature data; Feature extraction is performed on the temporary feature data and the external environment data respectively to obtain the second feature data and the third feature data; The preset environment assessment model also includes a self-attention module; The first feature data, the second feature data, and the third feature data are input into the Transformer module to obtain the environmental perception assessment results of the cable in the area under test, including: The first feature data and the third feature data are input into the self-attention module to obtain fused feature data, including: Perform linear transformations on the first feature data and the third feature data respectively to obtain the corresponding first vector data and second vector data; The attention weights are obtained based on the first vector data and the second vector data; Based on the attention weights, the first vector data and the second vector data are fused to obtain fused feature data; The fused feature data and the second feature data are input into the Transformer module to obtain the environmental perception assessment result of the cable in the area to be tested.

2. The distribution network cable environment sensing method based on dual-layer sensing according to claim 1, characterized in that, The construction process of the improved CNN module includes: The improved CNN module is obtained by adding a feature enhancement module to the main branch of the basic CNN model. The feature enhancement module is based on a one-dimensional convolutional layer or a long short-term memory network.

3. The distribution network cable environment sensing method based on dual-layer sensing according to claim 1, characterized in that, The process of extracting features from the fused data to obtain temporary feature data includes: extracting features from the fused data based on statistical algorithms to obtain the temporary feature data.

4. The distribution network cable environment sensing method based on dual-layer sensing according to claim 1, characterized in that, Also includes: Based on the warning threshold, the warning status is determined according to the environmental perception assessment results. Based on the aforementioned warning status, a warning message is issued.

5. A distribution network cable environment sensing system based on dual-layer sensing, characterized in that, include: The acquisition unit is used to acquire external environmental data and internal operating data of the cable in the area under test; The processing unit is configured to fuse the external environment data and the internal operating data to obtain fused data, including: preprocessing the external environment data and the internal operating data respectively, and concatenating the processed data to obtain the fused data, including: preprocessing the external environment data and the internal operating data respectively to obtain first temporary vector data and second temporary vector data; obtaining preset weight data, the preset weight data including first weight data and second weight data; and calculating the fused data based on the first temporary vector data and the first weight data. The weighted average value corresponding to the external environment data is calculated, and the weighted average value corresponding to the internal operating data is calculated based on the second temporary vector data and the second weight data; the weighted average value corresponding to the external environment data and the weighted average value corresponding to the internal operating data are integrated to obtain the fused data; and features are extracted from the fused data to obtain temporary feature data. The acquisition unit is also used to acquire image data of the cable in the area to be tested; The processing unit is further configured to input the external environment data, the temporary feature data, and the image data into a preset environmental assessment model to obtain the environmental perception assessment result of the cable in the test area. The preset environmental assessment model is based on an improved CNN module and a Transformer module, and includes: extracting features from the image data using the improved CNN module to obtain first feature data; extracting features from the temporary feature data and the external environment data respectively to obtain second feature data and third feature data; wherein, the preset environmental assessment model further includes a self-attention module; and inputting the first feature data, the second feature data, and the third feature data into the Transformer module. In the `ormer` module, obtaining the environmental perception assessment result of the cable in the area to be tested includes: inputting the first feature data and the third feature data into the self-attention module to obtain fused feature data, including: performing linear transformations on the first feature data and the third feature data respectively to obtain corresponding first vector data and second vector data; obtaining attention weights based on the first vector data and the second vector data; fusing the first vector data and the second vector data based on the attention weights to obtain fused feature data; and inputting the fused feature data and the second feature data into the `Transformer` module to obtain the environmental perception assessment result of the cable in the area to be tested.

6. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the distribution network cable environment sensing method based on dual-layer sensing as described in any one of claims 1 to 4.

Images