Resistance Detection System and Method for Live Working on Distribution Networks

Abstract

The present application provides a resistance detection system and method for live working on distribution networks, which relates to the field of intelligent resistance detection. It uses data analysis technology based on artificial intelligence to respectively perform temporal feature correlation on the time queue of temperature values and the time queue of humidity values during the insulation detection process, and performs significant fusion interaction based on the query matching of principal components. Based on this, an insulation resistance measurement compensation factor is intelligently generated based on the multi-dimensional correlation representation features among the temperature temporal correlation features, the humidity temporal correlation features, and the significant interaction of temperature and humidity principal components, and the compensated insulation resistance value is calculated based on the compensation factor. In this way, through the automated measurement process, the accuracy of the measurement data can be ensured, and then a more accurate insulation resistance value can be obtained. Moreover, by dynamically compensating the insulation resistance value based on the real-time collected environmental data, the performance of the insulation material in the real working environment can be more accurately reflected.

Description

Technical Field

[0001] This application relates to the field of intelligent resistance detection, and more specifically, to a resistance detection system and method for live working on distribution networks. Background Art

[0002] Live working on distribution networks refers to the repair, maintenance, and transformation of power lines without power interruption. This working method can reduce power outage time, improve power supply reliability, and reduce economic losses. Therefore, high requirements are imposed on insulating materials for live working, and any decline in insulation performance may lead to safety accidents. Insulation resistance is one of the important indicators to measure the insulation performance of electrical equipment. In live working on distribution networks, accurately measuring the insulation resistance value is crucial for ensuring the safety of operators and the normal operation of equipment.

[0003] Chinese Patent CN113960368A proposes an insulation resistance detection tool for live working on distribution networks and its usage method. Specifically, the operator first wears insulating gloves, starts the insulation resistance tester, and adjusts the knob to fix the anti-slip gear. Then, hold the insulating handle, adjust the length and rotate the test contact to set an appropriate pole pitch, and finally lock the position of the contact to prepare for insulation resistance measurement.

[0004] However, although the above patent solves the traditional problem of relying on two workers for manual detection of insulation resistance values, it still requires a single operator to perform manual operations, and the process is cumbersome and prone to errors, thus affecting the accuracy and reliability of measurement. In addition, the limitation of this insulation resistance measurement method is that it does not consider the actual working environment of resistance measurement. That is to say, during the resistance measurement process, factors such as ambient temperature and humidity will have a significant impact on the measurement results. Specifically, an increase in temperature may lead to an increase in the molecular activity of insulating materials, thereby reducing the resistance value; while an increase in humidity may promote current leakage, further reducing the insulation performance. However, the solution in the above patent does not take into account the influence of environmental factors, so it cannot truly reflect the state of insulating materials under actual working conditions, resulting in an underestimation of the true performance of insulating materials, and further affecting the reliability and safety of insulation resistance measurement.

[0005] Therefore, an optimized resistance detection solution for live working on distribution networks is desired. Summary of the Invention

[0006] To solve the above technical problems, the present application is proposed. Embodiments of the present application provide a resistance detection system and method for live working on distribution networks, which use data analysis technology based on artificial intelligence to respectively perform temporal feature correlation on the time queue of temperature values and the time queue of humidity values during the insulation detection process, and perform significant fusion interaction based on the query matching of the principal components, so as to intelligently generate an insulation resistance measurement compensation factor based on the multi-dimensional correlation representation features between the temperature temporal correlation features, the humidity temporal correlation features, and the significant interaction between the temperature and humidity principal components, and calculate the compensated insulation resistance value based on the compensation factor. In this way, through an automated measurement process, the accuracy of the measurement data can be ensured, and thus a more accurate insulation resistance value can be obtained. And based on the real-time collected environmental data to dynamically compensate the insulation resistance value, the performance of the insulation material in the actual working environment can be more accurately reflected.

[0007] According to one aspect of the present application, there is provided a resistance detection system for live working on distribution networks, which includes: an insulation resistance detection tool for live working on distribution networks, a temperature sensor, a humidity sensor, and a resistance measurement compensation corrector;

[0008] The insulation resistance detection tool for live working on distribution networks is used to collect the initial insulation resistance value of the object to be measured;

[0009] The temperature sensor and the humidity sensor are respectively used to collect the time queue of temperature values and the time queue of humidity values during the insulation detection process;

[0010] The resistance measurement compensation corrector is used to compensate and correct the initial insulation resistance value based on the time queue of the temperature values and the time queue of the humidity values to obtain a compensated insulation resistance value;

[0011] Among them, the resistance measurement compensation corrector includes:

[0012] A temperature and humidity temporal correlation module for inputting the time queue of the temperature values and the time queue of the humidity values into a sequence encoder to obtain a temperature bidirectional temporal correlation feature vector and a humidity bidirectional temporal correlation feature vector;

[0013] A temperature-humidity principal component temporal significant matching module for inputting the temperature bidirectional temporal correlation feature vector and the humidity bidirectional temporal correlation feature vector into a feature principal component significant fusion network to obtain a temperature-humidity principal component temporal significant collaborative interaction representation vector;

[0014] An insulation measurement factor multi-dimensional correlation module for inputting the temperature-humidity principal component temporal significant collaborative interaction representation vector, the temperature bidirectional temporal correlation feature vector, and the humidity bidirectional temporal correlation feature vector into an insulation measurement factor multi-dimensional correlation analyzer to obtain an insulation measurement factor multi-dimensional correlation representation vector;

[0015] An insulation resistance measurement compensation factor generation module, configured to obtain an insulation resistance measurement compensation factor based on the multi-dimensional associated representation vector of the insulation measurement factor;

[0016] An insulation resistance value compensation module, configured to calculate the product between the insulation resistance measurement compensation factor and the initial insulation resistance value to obtain the compensated insulation resistance value.

[0017] In the above-mentioned resistance detection system for live working on distribution networks, the temperature and humidity time series association module is configured to: input the time queue of the temperature value and the time queue of the humidity value into a sequence encoder based on a bidirectional gated recurrent unit to obtain the temperature bidirectional time series association feature vector and the humidity bidirectional time series association feature vector.

[0018] In the above resistance detection system for live working on distribution networks, the temperature-humidity principal component time series significant matching module includes: a temperature-humidity two-way time series correlation feature standardization unit for standardizing the temperature two-way time series correlation feature vector and the humidity two-way time series correlation feature vector to obtain a standardized temperature two-way time series correlation feature vector and a standardized humidity two-way time series correlation feature vector; a temperature-humidity two-way time series correlation sample covariance matrix calculation unit for calculating the sample covariance matrix of the standardized temperature two-way time series correlation feature vector and the standardized humidity two-way time series correlation feature vector to obtain a temperature two-way time series correlation sample covariance matrix and a humidity two-way time series correlation sample covariance matrix; a temperature-humidity two-way time series correlation principal component feature extraction unit for extracting eigenvectors based on matrix decomposition from the temperature two-way time series correlation sample covariance matrix and the humidity two-way time series correlation sample covariance matrix to obtain a set of temperature two-way time series correlation principal component feature vectors and a set of humidity two-way time series correlation principal component feature vectors; a temperature-humidity two-way time series correlation feature query matching unit for inputting the set of temperature two-way time series correlation principal component feature vectors and the set of humidity two-way time series correlation principal component feature vectors into a maximum approximate query matching network to obtain a set of best matching pairs of temperature two-way time series correlation principal component feature vectors and humidity two-way time series correlation principal component feature vectors; a temperature-humidity two-way time series correlation principal component fusion unit for inputting each best matching pair of temperature two-way time series correlation principal component feature vectors and humidity two-way time series correlation principal component feature vectors in the set of best matching pairs of temperature two-way time series correlation principal component feature vectors and humidity two-way time series correlation principal component feature vectors into a semantic fine-grained gating joint module to obtain a set of temperature-humidity two-way time series correlation principal component fusion feature vectors; a temperature-humidity principal component time series significant collaborative interaction representation generation unit for cascading the set of temperature-humidity two-way time series correlation principal component fusion feature vectors to obtain the temperature-humidity principal component time series significant collaborative interaction representation vector.

[0019] In the above-mentioned resistance detection system for live working on distribution networks, the temperature and humidity two-way time-series correlation feature standardization unit is used to: calculate the mean and standard deviation of the temperature two-way time-series correlation feature vector respectively to obtain the temperature two-way time-series correlation feature mean and the temperature two-way time-series correlation feature standard deviation; after subtracting the temperature two-way time-series correlation feature vector from the temperature two-way time-series correlation feature mean by position, calculate the position-by-position division of the obtained temperature two-way time-series offset vector by the temperature two-way time-series correlation feature standard deviation to obtain the standardized temperature two-way time-series correlation feature vector; calculate the mean and standard deviation of the humidity two-way time-series correlation feature vector respectively to obtain the humidity two-way time-series correlation feature mean and the humidity two-way time-series correlation feature standard deviation; after subtracting the humidity two-way time-series correlation feature vector from the humidity two-way time-series correlation feature mean by position, calculate the position-by-position division of the obtained humidity two-way time-series offset vector by the humidity two-way time-series correlation feature standard deviation to obtain the standardized humidity two-way time-series correlation feature vector.

[0020] In the above-mentioned resistance detection system for live working on distribution networks, the temperature and humidity two-way time-series correlation sample covariance matrix calculation unit includes: a temperature two-way time-series correlation sample covariance matrix generation sub-unit, which is used to multiply the transposed vector of the standardized temperature two-way time-series correlation feature vector by the standardized temperature two-way time-series correlation feature vector, and then divide the obtained standardized temperature two-way time-series correlation matrix by the value obtained by subtracting 1 from the length of the standardized temperature two-way time-series correlation feature vector to obtain the temperature two-way time-series correlation sample covariance matrix; a humidity two-way time-series correlation sample covariance matrix generation sub-unit, which is used to multiply the transposed vector of the standardized humidity two-way time-series correlation feature vector by the standardized humidity two-way time-series correlation feature vector, and then divide the obtained standardized humidity two-way time-series correlation matrix by the value obtained by subtracting 1 from the length of the standardized humidity two-way time-series correlation feature vector to obtain the humidity two-way time-series correlation sample covariance matrix.

[0021] In the above resistance detection system for live working on distribution networks, the temperature-humidity two-way time-series correlation feature query and matching unit is configured to: extract a predetermined temperature two-way time-series correlation principal component feature vector from the set of temperature two-way time-series correlation principal component feature vectors; calculate the cosine similarity between the predetermined temperature two-way time-series correlation principal component feature vector and each humidity two-way time-series correlation principal component feature vector in the set of humidity two-way time-series correlation principal component feature vectors to obtain a set of matching query similarities; and use the humidity two-way time-series correlation principal component feature vector corresponding to the maximum matching query similarity in the set of matching query similarities and the predetermined temperature two-way time-series correlation principal component feature vector as the best matching pair of the predetermined temperature two-way time-series correlation principal component feature vector and the humidity two-way time-series correlation principal component feature vector.

[0022] In the above resistance detection system for live working on distribution networks, the temperature-humidity two-way time-series correlation principal component fusion unit is configured to: calculate the position difference, position dot product, and position addition between the best matching pairs of the temperature two-way time-series correlation principal component feature vector and the humidity two-way time-series correlation principal component feature vector respectively to obtain a temperature-humidity time-series correlation principal component difference vector, a temperature-humidity time-series correlation principal component dot product vector, and a temperature-humidity time-series correlation principal component sum vector; cascade the temperature-humidity time-series correlation principal component difference vector, the temperature-humidity time-series correlation principal component dot product vector, and the temperature-humidity time-series correlation principal component sum vector and perform one-dimensional convolutional coding to obtain a temperature-humidity time-series correlation principal component multi-dimensional fusion vector; and perform max pooling processing based on a local window on the temperature-humidity time-series correlation principal component multi-dimensional fusion vector to obtain the temperature-humidity two-way time-series correlation principal component fusion feature vector.

[0023] In the above resistance detection system for live working on distribution networks, the multi-dimensional correlation module of insulation measurement factors is configured to: input the temperature-humidity principal component time-series significantly collaborative interaction representation vector, the temperature two-way time-series correlation feature vector, and the humidity two-way time-series correlation feature vector into an insulation measurement factor multi-dimensional correlation analyzer based on an embedding layer to obtain the insulation measurement factor multi-dimensional correlation representation vector.

[0024] In the above resistance detection system for live working on distribution networks, the insulation resistance measurement compensation factor generation module is configured to: input the insulation measurement factor multi-dimensional correlation representation vector into a compensation factor generator based on a decoder to obtain the insulation resistance measurement compensation factor.

[0025] According to another aspect of the present application, there is provided a method for detecting resistance in live working on distribution networks, which includes:

[0026] Collecting the initial insulation resistance value of the object to be measured;

[0027] Collect the time queues of temperature values and humidity values during the insulation detection process;

[0028] Based on the time queue of the temperature values and the time queue of the humidity values, compensate and correct the initial insulation resistance value to obtain a compensated insulation resistance value;

[0029] Wherein, based on the time queue of the temperature values and the time queue of the humidity values, compensating and correcting the initial insulation resistance value to obtain a compensated insulation resistance value includes:

[0030] Input the time queue of the temperature values and the time queue of the humidity values into a sequence encoder to obtain a temperature bidirectional time series correlation feature vector and a humidity bidirectional time series correlation feature vector;

[0031] Input the temperature bidirectional time series correlation feature vector and the humidity bidirectional time series correlation feature vector into a feature principal component significant fusion network to obtain a temperature-humidity principal component time series significant collaborative interaction representation vector;

[0032] Input the temperature-humidity principal component time series significant collaborative interaction representation vector, the temperature bidirectional time series correlation feature vector and the humidity bidirectional time series correlation feature vector into an insulation measurement factor multi-dimensional correlation analyzer to obtain an insulation measurement factor multi-dimensional correlation representation vector;

[0033] Based on the insulation measurement factor multi-dimensional correlation representation vector, obtain an insulation resistance measurement compensation factor;

[0034] Calculate the product between the insulation resistance measurement compensation factor and the initial insulation resistance value to obtain the compensated insulation resistance value.

[0035] Compared with the prior art, the resistance detection system and method for live working on distribution networks provided by the present application adopt data analysis technology based on artificial intelligence to respectively perform time series feature correlation on the time queues of temperature values and humidity values during the insulation detection process and perform significant fusion interaction of query matching based on principal components. Based on this, an insulation resistance measurement compensation factor is intelligently generated based on the multi-dimensional correlation representation features among the temperature time series correlation features, the humidity time series correlation features, and the significant interaction of temperature and humidity principal components, and the compensated insulation resistance value is calculated based on the compensation factor. In this way, through an automated measurement process, the accuracy of measurement data can be ensured, and thus a more accurate insulation resistance value can be obtained. And dynamically compensating the insulation resistance value based on the real-time collected environmental data can more accurately reflect the performance of the insulation material in the actual working environment. Description of the Drawings

[0036] To more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings. In the drawings:

[0037] Figure 1 It is a block diagram of a resistance detection system for live working on distribution networks according to an embodiment of the present application.

[0038] Figure 2 It is a block diagram of a resistance measurement compensation corrector in a resistance detection system for live working on distribution networks according to an embodiment of the present application.

[0039] Figure 3 It is a schematic diagram of data flow of a resistance measurement compensation corrector in a resistance detection system for live working on distribution networks according to an embodiment of the present application.

[0040] Figure 4 It is a block diagram of a temperature-humidity principal component time series significant matching module in a resistance detection system for live working on distribution networks according to an embodiment of the present application.

[0041] Figure 5 It is a block diagram of a temperature-humidity two-way time series correlation sample covariance matrix calculation unit in a resistance detection system for live working on distribution networks according to an embodiment of the present application.

[0042] Figure 6 It is a flowchart of a resistance detection method for live working on distribution networks according to an embodiment of the present application. Detailed implementation manners

[0043] Next, example embodiments according to the present application will be described in detail with reference to the drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. It should be understood that the present application is not limited by the example embodiments described here.

[0044] Live working on distribution networks aims to maintain and transform power lines without interrupting power supply to improve power supply stability and reduce economic losses. During this process, accurately measuring the insulation resistance is the key to ensuring operation safety and normal operation of equipment. Chinese Patent CN113960368A introduces an insulation resistance detection tool and its usage method. This insulation resistance detection tool measures the insulation resistance through manual operation. Although it improves the convenience of measurement, there are still problems such as complex operation and easy error. In addition, this method fails to fully consider the influence of environmental factors such as temperature and humidity on the measurement results, which may lead to misjudgment of the performance of insulating materials, thus affecting the measurement accuracy and operation safety.

[0045] Based on this, the present application proposes an optimized resistance detection system for live working on distribution networks. Figure 1 FIG. is a system block diagram of a resistance detection system for live working on distribution networks according to an embodiment of the present application. As Figure 1 shown, in the resistance detection system 100 for live working on distribution networks, it includes: an insulation resistance detection tool 110 for live working on distribution networks, a temperature sensor 120, a humidity sensor 130, and a resistance measurement compensation corrector 140; the insulation resistance detection tool 110 for live working on distribution networks is used to collect the initial insulation resistance value of the object to be measured; the temperature sensor 120 and the humidity sensor 130 are respectively used to collect the time queue of temperature values and the time queue of humidity values during the insulation detection process; the resistance measurement compensation corrector 140 is used to compensate and correct the initial insulation resistance value based on the time queue of temperature values and the time queue of humidity values to obtain a compensated insulation resistance value.

[0046] It should be understood that in order to obtain the reference data of the insulating material without considering the influence of environmental factors, it is first necessary to collect the initial insulation resistance value of the object to be measured through the insulation resistance detection tool for live working on distribution networks. Insulation resistance is one of the key indicators to measure the performance of insulating materials. A high insulation resistance value usually means good insulation performance. The initial insulation resistance value collected by the detection tool is the basic value for subsequent environmental factor compensation. Then, considering that temperature and humidity are important environmental factors affecting the performance of insulating materials, different temperature and humidity conditions will cause changes in the insulation resistance value. In order to monitor these parameters in real time to help evaluate the authenticity of the insulation performance, in the technical solution of the present application, a temperature sensor and a humidity sensor are respectively used to obtain the time queue data of temperature values and the time queue data of humidity values during the insulation detection process. That is, when calculating the compensation of the insulation resistance, the changes in these two environmental factors of temperature and humidity are taken into account. Finally, in the resistance measurement compensation corrector, the obtained temperature and humidity time queue data are analyzed, and then the initial insulation resistance value is compensated and corrected to obtain a compensated insulation resistance value, so as to improve the accuracy and reliability of the insulation resistance measurement.

[0047] Accordingly, the resistance measurement compensation corrector performs temporal feature correlation on the time queue of the temperature value and the time queue of the humidity value respectively by adopting artificial intelligence-based data analysis and processing technologies, and performs significant fusion interaction based on the query matching of the principal components, so as to intelligently generate an insulation resistance measurement compensation factor based on the multi-dimensional correlation representation features among the temperature temporal correlation features, the humidity temporal correlation features, and the significant interaction between the temperature and humidity principal components, and calculate the compensated insulation resistance value based on the compensation factor. In this way, environmental data can be collected in real time, and the insulation resistance value can be dynamically compensated according to these data to reflect the performance in the real working environment. At the same time, through the automated measurement process, human operation errors are reduced, the accuracy of the measurement data is ensured, and thus a more accurate insulation resistance value is obtained.

[0048] Figure 2 It is a block diagram of a resistance measurement compensation corrector in a resistance detection system for live working on a distribution network according to an embodiment of the present application. Figure 3 It is a schematic diagram of data flow of a resistance measurement compensation corrector in a resistance detection system for live working on a distribution network according to an embodiment of the present application. As Figure 2 and Figure 3 shown, in the resistance measurement compensation corrector 140, it includes: a temperature and humidity temporal correlation module 141, configured to input the time queue of the temperature value and the time queue of the humidity value into a sequence encoder to obtain a temperature bidirectional temporal correlation feature vector and a humidity bidirectional temporal correlation feature vector; a temperature-humidity principal component temporal significant matching module 142, configured to input the temperature bidirectional temporal correlation feature vector and the humidity bidirectional temporal correlation feature vector into a feature principal component significant fusion network to obtain a temperature-humidity principal component temporal significant collaborative interaction representation vector; an insulation measurement factor multi-dimensional correlation module 143, configured to input the temperature-humidity principal component temporal significant collaborative interaction representation vector, the temperature bidirectional temporal correlation feature vector, and the humidity bidirectional temporal correlation feature vector into an insulation measurement factor multi-dimensional correlation analyzer to obtain an insulation measurement factor multi-dimensional correlation representation vector; an insulation resistance measurement compensation factor generation module 144, configured to obtain an insulation resistance measurement compensation factor based on the insulation measurement factor multi-dimensional correlation representation vector; and an insulation resistance value compensation module 145, configured to calculate the product between the insulation resistance measurement compensation factor and the initial insulation resistance value to obtain the compensated insulation resistance value.

[0049] In an embodiment of the present application, the temperature-humidity time-sequence correlation module 141 is configured to input the time queue of the temperature values and the time queue of the humidity values into a sequence encoder to obtain a temperature bidirectional time-sequence correlation feature vector and a humidity bidirectional time-sequence correlation feature vector. Specifically, in an embodiment of the present application, the temperature-humidity time-sequence correlation module is configured to: input the time queue of the temperature values and the time queue of the humidity values into a sequence encoder based on a bidirectional gated recurrent unit to obtain the temperature bidirectional time-sequence correlation feature vector and the humidity bidirectional time-sequence correlation feature vector. It should be understood that in resistance detection, temperature and humidity change continuously over time, and the current temperature and humidity values may be related to the values before and after. These changes in the time dimension may have a significant impact on the insulation resistance value. Those of ordinary skill in the art should know that the gated recurrent unit (GRU) is a variant of a recurrent neural network that effectively captures long-distance dependencies in the input sequence by introducing a gating mechanism (update gate and reset gate) to control the flow of information. The bidirectional gated recurrent unit is an extension of the gated recurrent unit that can process information both forward and backward from the input sequence data. This means that for the input sequence data, the model not only considers the previous information (past context) but also the subsequent information (future context). This bidirectional feature enables the model to more comprehensively understand the context of the input sequence data and thus effectively represent the time-sequence features of the input sequence data. Based on this, in the technical solution of the present application, the time queue of the temperature values and the time queue of the humidity values are input into a sequence encoder based on a bidirectional gated recurrent unit to capture and mine the variation rules and feature information of temperature and humidity over time, obtaining a temperature bidirectional time-sequence correlation feature vector and a humidity bidirectional time-sequence correlation feature vector.

[0050] In an embodiment of the present application, the temperature-humidity principal component time-sequence significant matching module 142 is configured to input the temperature bidirectional time-sequence correlation feature vector and the humidity bidirectional time-sequence correlation feature vector into a feature principal component significant fusion network to obtain a temperature-humidity principal component time-sequence significant collaborative interaction representation vector. Correspondingly, considering that both the temperature bidirectional time-sequence correlation feature vector and the humidity bidirectional time-sequence correlation feature vector contain important time-sequence data features and there is a key interaction between them in the time dimension. In order to perform significant fusion based on the key feature information in their time sequences to provide a more reliable basis for the generation of subsequent compensation factors, in the technical solution of the present application, the temperature bidirectional time-sequence correlation feature vector and the humidity bidirectional time-sequence correlation feature vector are input into a feature principal component significant fusion network to obtain a temperature-humidity principal component time-sequence significant collaborative interaction representation vector.

[0051] It should be understood that the feature principal component significant fusion network is a technology integrating feature dimensionality reduction, principal component analysis, and key feature collaborative interaction, aiming to create a prominent and concise feature association mapping for extracting the core interactions between feature vectors. Specifically, first, the temperature bidirectional time-series correlation feature vector and the humidity bidirectional time-series correlation feature vector are standardized to eliminate the influence caused by different dimensions of different feature vectors, making the features comparable, thereby obtaining the standardized temperature bidirectional time-series correlation feature vector and the standardized humidity bidirectional time-series correlation feature vector. Then, based on the standardized features, the respective sample covariance matrices are calculated, providing data support for subsequent principal component feature extraction. Next, the sample covariance matrices are used to extract eigenvectors based on matrices to obtain the set of temperature bidirectional time-series correlation principal component feature vectors and the set of humidity bidirectional time-series correlation principal component feature vectors. In particular, the eigenvector extraction here uses the principal component analysis algorithm to extract the key features of the features. That is, principal component analysis (PCA) is a commonly used data dimensionality reduction technique that can extract the main components in temperature and humidity data, namely those features that can best explain the data changes. Subsequently, the set of temperature bidirectional time-series correlation principal component feature vectors and the set of humidity bidirectional time-series correlation principal component feature vectors are subjected to maximum approximation matching to obtain the best matching pairs. Specifically, the cosine similarity between each temperature bidirectional time-series correlation principal component feature vector and all the humidity bidirectional time-series correlation principal component feature vectors in the set of humidity bidirectional time-series correlation principal component feature vectors is calculated, and the temperature bidirectional time-series correlation principal component feature vector and the humidity bidirectional time-series correlation principal component feature vector corresponding to the maximum cosine value are used as the best matching pair. After that, each pair of best matching pairs obtained is input into the semantic fine-grained gating joint module for fine-grained analysis and processing to more deeply understand the semantic interaction content between temperature and humidity, improve the grasp of the data meaning and the ability to capture key significant features, and obtain the set of temperature-humidity bidirectional time-series correlation principal component fusion feature vectors. Finally, the set of principal component fusion feature vectors is cascaded to integrate the global interaction information in the entire time domain, obtaining the temperature-humidity principal component time-series significant collaborative interaction representation vector.

[0052] Specifically, Figure 4 is a block diagram of the temperature-humidity principal component time-series significant matching module in the live working resistance detection system for distribution networks according to the embodiment of the present application. As Figure 4As shown, the temperature-humidity principal component time series significant matching module 142 includes: a temperature-humidity two-way time series correlation feature normalization unit 1421, configured to perform normalization processing on the temperature two-way time series correlation feature vector and the humidity two-way time series correlation feature vector to obtain a normalized temperature two-way time series correlation feature vector and a normalized humidity two-way time series correlation feature vector; a temperature-humidity two-way time series correlation sample covariance matrix calculation unit 1422, configured to calculate the sample covariance matrix of the normalized temperature two-way time series correlation feature vector and the normalized humidity two-way time series correlation feature vector to obtain a temperature two-way time series correlation sample covariance matrix and a humidity two-way time series correlation sample covariance matrix; a temperature-humidity two-way time series correlation principal component feature extraction unit 1423, configured to perform eigenvector extraction based on matrix decomposition on the temperature two-way time series correlation sample covariance matrix and the humidity two-way time series correlation sample covariance matrix to obtain a set of temperature two-way time series correlation principal component feature vectors and a set of humidity two-way time series correlation principal component feature vectors; a temperature-humidity two-way time series correlation feature query matching unit 1424, configured to input the set of temperature two-way time series correlation principal component feature vectors and the set of humidity two-way time series correlation principal component feature vectors into a maximum approximate query matching network to obtain a set of best matching pairs of temperature two-way time series correlation principal component feature vectors and humidity two-way time series correlation principal component feature vectors; a temperature-humidity two-way time series correlation principal component fusion unit 1425, configured to input each best matching pair of the temperature two-way time series correlation principal component feature vector and the humidity two-way time series correlation principal component feature vector in the set of best matching pairs of the temperature two-way time series correlation principal component feature vector and the humidity two-way time series correlation principal component feature vector into a semantic fine-grained gating joint module to obtain a set of temperature-humidity two-way time series correlation principal component fusion feature vectors; a temperature-humidity principal component time series significant collaborative interaction representation generation unit 1426, configured to cascade the set of temperature-humidity two-way time series correlation principal component fusion feature vectors to obtain the temperature-humidity principal component time series significant collaborative interaction representation vector.

[0053] More specifically, in the embodiments of the present application, the temperature and humidity bidirectional time series correlation feature standardization unit is configured to: calculate the mean and standard deviation of the temperature bidirectional time series correlation feature vector respectively to obtain the temperature bidirectional time series correlation feature mean and the temperature bidirectional time series correlation feature standard deviation; after subtracting the temperature bidirectional time series correlation feature vector from the temperature bidirectional time series correlation feature mean in a position-by-position manner, calculate the position-by-position division of the obtained temperature bidirectional time series offset vector by the temperature bidirectional time series correlation feature standard deviation to obtain the standardized temperature bidirectional time series correlation feature vector; calculate the mean and standard deviation of the humidity bidirectional time series correlation feature vector respectively to obtain the humidity bidirectional time series correlation feature mean and the humidity bidirectional time series correlation feature standard deviation; after subtracting the humidity bidirectional time series correlation feature vector from the humidity bidirectional time series correlation feature mean in a position-by-position manner, calculate the position-by-position division of the obtained humidity bidirectional time series offset vector by the humidity bidirectional time series correlation feature standard deviation to obtain the standardized humidity bidirectional time series correlation feature vector.

[0054] More specifically, Figure 5 FIG. is a block diagram of a temperature and humidity bidirectional time series correlation sample covariance matrix calculation unit in a live working resistance detection system for a distribution network according to an embodiment of the present application. As Figure 5 shown, the temperature and humidity bidirectional time series correlation sample covariance matrix calculation unit 1422 includes: a temperature bidirectional time series correlation sample covariance matrix generation subunit 14221, configured to multiply the transposed vector of the standardized temperature bidirectional time series correlation feature vector by the standardized temperature bidirectional time series correlation feature vector, and then perform position-by-position division of the obtained standardized temperature bidirectional time series correlation matrix by a value obtained by subtracting one from the length of the standardized temperature bidirectional time series correlation feature vector to obtain the temperature bidirectional time series correlation sample covariance matrix; a humidity bidirectional time series correlation sample covariance matrix generation subunit 14222, configured to multiply the transposed vector of the standardized humidity bidirectional time series correlation feature vector by the standardized humidity bidirectional time series correlation feature vector, and then perform position-by-position division of the obtained standardized humidity bidirectional time series correlation matrix by a value obtained by subtracting one from the length of the standardized humidity bidirectional time series correlation feature vector to obtain the humidity bidirectional time series correlation sample covariance matrix.

[0055] More specifically, in the embodiments of the present application, the temperature-humidity two-way time-series correlation feature query and matching unit is configured to: extract a predetermined temperature two-way time-series correlation principal component feature vector from the set of temperature two-way time-series correlation principal component feature vectors; calculate the cosine similarity between the predetermined temperature two-way time-series correlation principal component feature vector and each humidity two-way time-series correlation principal component feature vector in the set of humidity two-way time-series correlation principal component feature vectors to obtain a set of matching query similarities; and use the humidity two-way time-series correlation principal component feature vector corresponding to the maximum matching query similarity in the set of matching query similarities and the predetermined temperature two-way time-series correlation principal component feature vector as the best matching pair of the predetermined temperature two-way time-series correlation principal component feature vector and the humidity two-way time-series correlation principal component feature vector.

[0056] More specifically, in the embodiments of the present application, the temperature-humidity two-way time-series correlation principal component fusion unit is configured to: calculate the position difference, position dot product, and position addition between the best matching pairs of the temperature two-way time-series correlation principal component feature vectors and the humidity two-way time-series correlation principal component feature vectors respectively to obtain a temperature-humidity time-series correlation principal component difference vector, a temperature-humidity time-series correlation principal component dot product vector, and a temperature-humidity time-series correlation principal component sum vector; cascade the temperature-humidity time-series correlation principal component difference vector, the temperature-humidity time-series correlation principal component dot product vector, and the temperature-humidity time-series correlation principal component sum vector and perform one-dimensional convolutional coding to obtain a temperature-humidity time-series correlation principal component multi-dimensional fusion vector; and perform maximum pooling processing based on a local window on the temperature-humidity time-series correlation principal component multi-dimensional fusion vector to obtain the temperature-humidity two-way time-series correlation principal component fusion feature vector.

[0057] In the embodiments of the present application, specifically, the temperature-humidity principal component time-series significant matching module is configured to: input the temperature two-way time-series correlation feature vector and the humidity two-way time-series correlation feature vector into a feature principal component significant fusion network and process them according to the following significant fusion formula to obtain the temperature-humidity principal component time-series significant collaborative interaction representation vector; where the significant fusion formula is:

[0058]

[0059]

[0060]

[0061]

[0062]

[0063]

[0064]

[0065]

[0066]

[0067]

[0068]

[0069]

[0070] wherein, and represent the temperature two-way time series correlation feature vector and the humidity two-way time series correlation feature vector respectively, and are the means of the temperature two-way time series correlation feature vector and the humidity two-way time series correlation feature vector respectively, and are the standard deviations of the temperature two-way time series correlation feature vector and the humidity two-way time series correlation feature vector respectively, and are the standardized temperature two-way time series correlation feature vector and the standardized humidity two-way time series correlation feature vector respectively, and are respectively and transpose vectors, and are the lengths of the standardized temperature two-way time series correlation feature vector and the standardized humidity two-way time series correlation feature vector respectively, and are the temperature two-way time series correlation sample covariance matrix and the humidity two-way time series correlation sample covariance matrix respectively, and are the temperature two-way time series correlation principal component orthogonal matrix and the humidity two-way time series correlation principal component orthogonal matrix respectively, and are the temperature two-way time series correlation diagonal matrix and the humidity two-way time series correlation diagonal matrix respectively, is the temperature two-way time series correlation diagonal matrix with the elements on the diagonal being , are the weight values of the respective temperature two-way time series correlation principal component feature vectors, is the humidity two-way time series correlation diagonal matrix with the elements on the diagonal being , They are the weight values of the humidity two-way time series associated principal component feature vectors respectively. and are respectively and the transposed matrices of, is each temperature two-way time series associated principal component feature vector in the set of the temperature two-way time series associated principal component feature vectors, is each humidity two-way time series associated principal component feature vector in the set of the humidity two-way time series associated principal component feature vectors, is to calculate the and vector inner product between, is to calculate the one-norm of the vector, is to return the value corresponding to the maximum value, is the maximum approximate matching value, , and are respectively by position difference, by position dot product and by position addition, is the concatenation processing, is the one-dimensional convolutional coding operation, is the max pooling operation, is the th temperature-humidity two-way time series associated principal component fusion feature vector in the set of the temperature-humidity two-way time series associated principal component fusion feature vectors, is the number of feature vectors in the set of the temperature-humidity two-way time series associated principal component fusion feature vectors, is the temperature-humidity principal component time series significant collaborative interaction representation vector.

[0071] In an embodiment of the present application, the insulation measurement factor multi-dimensional correlation module 143 is configured to input the temperature-humidity principal component time-series significantly collaborative interaction representation vector, the temperature two-way time-series correlation feature vector, and the humidity two-way time-series correlation feature vector into an insulation measurement factor multi-dimensional correlation analyzer to obtain an insulation measurement factor multi-dimensional correlation representation vector. Specifically, in an embodiment of the present application, the insulation measurement factor multi-dimensional correlation module is configured to: input the temperature-humidity principal component time-series significantly collaborative interaction representation vector, the temperature two-way time-series correlation feature vector, and the humidity two-way time-series correlation feature vector into an insulation measurement factor multi-dimensional correlation analyzer based on an embedding layer to obtain the insulation measurement factor multi-dimensional correlation representation vector. It should be understood that in order to further integrate and analyze the information in temperature, humidity, and temperature-humidity interaction features to generate a multi-dimensional feature that can comprehensively reflect the impact of environmental conditions on insulation resistance measurement, in the technical solution of the present application, the temperature-humidity principal component time-series significantly collaborative interaction representation vector, the temperature two-way time-series correlation feature vector, and the humidity two-way time-series correlation feature vector are input into an insulation measurement factor multi-dimensional correlation analyzer based on an embedding layer to capture the complex time-series relationships between the various features, and an insulation measurement factor multi-dimensional correlation representation vector is obtained.

[0072] In an embodiment of the present application, the insulation resistance measurement compensation factor generation module 144 and the insulation resistance value compensation module 145 are respectively configured to obtain an insulation resistance measurement compensation factor based on the insulation measurement factor multi-dimensional correlation representation vector, and calculate the product of the insulation resistance measurement compensation factor and the initial insulation resistance value to obtain the compensated insulation resistance value. Specifically, in an embodiment of the present application, the insulation resistance measurement compensation factor generation module is configured to: input the insulation measurement factor multi-dimensional correlation representation vector into a compensation factor generator based on a decoder to obtain the insulation resistance measurement compensation factor. That is, the insulation measurement factor multi-dimensional correlation representation vector obtained by performing multi-dimensional correlation analysis using the temperature-humidity principal component time-series significantly collaborative interaction representation vector, the temperature two-way time-series correlation feature vector, and the humidity two-way time-series correlation feature vector is used for decoding processing, so as to intelligently generate an insulation resistance measurement compensation factor, and calculate the compensated insulation resistance value based on the compensation factor. In this way, environmental data can be collected in real time, and the insulation resistance value can be dynamically compensated according to these data to reflect the performance in the actual working environment. At the same time, through an automated measurement process, human operation errors are reduced, ensuring the accuracy of the measurement data, and thus a more accurate insulation resistance value is obtained.

[0073] Preferably, considering that the temperature-humidity principal component time-series significant co-interaction representation vector, the temperature two-way time-series correlation feature vector, and the humidity two-way time-series correlation feature vector respectively represent the temperature-humidity time-series co-feature, the temperature two-way local time-series correlation feature, and the humidity two-way local time-series correlation feature. Considering the possible feature redundancy in the feature set composed of the temperature-humidity principal component time-series significant co-interaction representation vector, the temperature two-way time-series correlation feature vector, and the humidity two-way time-series correlation feature vector, when aligning the multi-dimensional correlation analysis of the insulation measurement factor based on the embedding layer, there will be a multi-modal semantic fusion shift caused by the uneven distribution of attention weights under feature redundancy, so as to further improve the multi-modal semantic fusion regression comprehensibility of the multi-dimensional correlation representation vector of the insulation measurement factor, thereby improving the accuracy of the insulation resistance measurement compensation factor obtained by inputting it into the compensation factor generator based on the decoder.

[0074] Therefore, in a preferred embodiment, when inputting the multi-dimensional correlation representation vector of the insulation measurement factor into the compensation factor generator based on the decoder to obtain the insulation resistance measurement compensation factor, the multi-dimensional correlation representation vector of the insulation measurement factor is optimized, including the steps of:

[0075] Arrange the respective eigenvalues of the multi-dimensional correlation representation vector of the insulation measurement factor in ascending order to obtain the multi-dimensional correlation order coding vector of the insulation measurement factor;

[0076] In response to the covariance between the eigenvalue and the eigenvalue of the multi-dimensional correlation order coding vector of the insulation measurement factor being less than or equal to the distance difference hyperparameter , calculate the weighted sum of the eigenvalue and the eigenvalue as the optimized eigenvalue:

[0077]

[0078]

[0079] Wherein, and respectively represent the eigenvalue and the eigenvalue of the multi-dimensional correlation order coding vector of the insulation measurement factor, is the feature mean of the multi-dimensional correlation order coding vector of the insulation measurement factor, represents the distance difference hyperparameter, and and are weight hyperparameters, represents the optimized Eigenvalue;

[0080] Calculate the hyperbolic sine function value of the sum of the squares of all eigenvalues of the multi-dimensional correlation representation vector of the insulation measurement factor:

[0081]

[0082] where, represents the length of the multi-dimensional correlation representation vector of the insulation measurement factor, represents the hyperbolic sine function, represents the hyperbolic sine function value;

[0083] And calculate its exponential value with the natural constant as the base and then divide it by the square of the length of the multi-dimensional correlation representation vector of the insulation measurement factor to obtain the multi-dimensional correlation potential manifold value of the insulation measurement factor , where, represents the natural constant, represents the multi-dimensional correlation potential manifold value of the insulation measurement factor;

[0084] In response to the absolute value of the covariance between the eigenvalue and the eigenvalue of the multi-dimensional correlation order-encoded vector of the insulation measurement factor being greater than the distance difference hyperparameter , multiply the multi-dimensional correlation potential manifold value of the insulation measurement factor by the eigenvalue, and then calculate the weighted subtraction between the product and the eigenvalue to obtain the optimized eigenvalue:

[0085]

[0086] where, and respectively represent the eigenvalue and the eigenvalue of the multi-dimensional correlation order-encoded vector of the insulation measurement factor, represents the multi-dimensional correlation potential manifold value of the insulation measurement factor, and are weight hyperparameters, represents the optimized eigenvalue;

[0087] On the basis of keeping the first eigenvalue of the multi-dimensional correlation order-encoded vector of the insulation measurement factor unchanged, combine the optimized eigenvalue to obtain the optimized multi-dimensional correlation representation vector of the insulation measurement factor.

[0088] Therefore, for the semantic feature set of the multi-dimensional associated representation vector of the insulation measurement factor, under the condition of the trajectory distribution of the preset semantic space, the cross-domain semantic localization efficiency attenuation phenomenon mainly stems from the long-range action distance beyond the local association window. By adopting a non-linear manifold latent unit representation method based on covariance tensor integration, the multi-modal associated topological structure hidden in its parameter space is analyzed, and by constructing a non-linear manifold latent unit model with scale adaptability, an adaptive association representation mechanism between the parameters of the real-time interaction scene time-series association mapping matrix is re-established. Thus, through this reconstruction strategy of the physical state evolution trajectory, the continuous state transition modeling under long-range parameter association is effectively realized, the semantic field reconstruction efficiency is significantly improved, and at the same time, the original coding representation characteristics of the multi-dimensional associated representation vector of the insulation measurement factor are maintained, improving the accuracy of the insulation resistance measurement compensation factor obtained by inputting the multi-dimensional associated representation vector of the insulation measurement factor into the decoder-based compensation factor generator. In this way, environmental data can be collected in real time, and the insulation resistance value can be dynamically compensated according to these data to reflect the performance in the real working environment. At the same time, through the automated measurement process, human operation errors are reduced, ensuring the accuracy of the measurement data, and thus obtaining a more accurate insulation resistance value.

[0089] In summary, the resistance detection system 100 for live working on distribution networks according to the embodiments of the present application is elucidated. It uses artificial intelligence-based data analysis technology to perform time-series feature association on the time queue of temperature values and the time queue of humidity values respectively during the insulation detection process, and significantly fuses and interacts based on the query matching of the principal components. Based on the multi-dimensional associated representation features between the temperature time-series association features, the humidity time-series association features, and the significant interaction of the temperature and humidity principal components, an insulation resistance measurement compensation factor is intelligently generated, and the compensated insulation resistance value is calculated based on the compensation factor. In this way, through the automated measurement process, the accuracy of the measurement data can be ensured, and thus a more accurate insulation resistance value can be obtained. And based on the real-time collected environmental data, the insulation resistance value is dynamically compensated, which can more accurately reflect the performance of the insulation material in the real working environment.

[0090] As described above, the resistance detection system 100 for live working on distribution networks according to the embodiments of the present application can be implemented in various terminal devices, such as a server for resistance detection in live working on distribution networks. In one example, the resistance detection system 100 for live working on distribution networks according to the embodiments of the present application can be integrated into the terminal device as a software module and / or a hardware module. For example, the resistance detection system 100 for live working on distribution networks can be a software module in the operating system of the terminal device, or can be an application program developed for the terminal device; of course, the resistance detection system 100 for live working on distribution networks can also be one of the many hardware modules of the terminal device.

[0091] Alternatively, in another example, the resistance detection system 100 for live working on distribution network and the terminal device may also be separate devices, and the resistance detection system 100 for live working on distribution network can be connected to the terminal device through a wired and / or wireless network and transmit interaction information in accordance with a predefined data format.

[0092] Figure 6 FIG. is a flowchart of a resistance detection method for live working on distribution network according to an embodiment of the present application. As Figure 6 shown, in the resistance detection method for live working on distribution network, it includes: S110, collecting an initial insulation resistance value of a measured object; S120, collecting a time queue of temperature values and a time queue of humidity values during the insulation detection process; S130, based on the time queue of the temperature values and the time queue of the humidity values, compensating and correcting the initial insulation resistance value to obtain a compensated insulation resistance value; wherein, in step S130, it includes: inputting the time queue of the temperature values and the time queue of the humidity values into a sequence encoder to obtain a temperature bidirectional time series correlation feature vector and a humidity bidirectional time series correlation feature vector; inputting the temperature bidirectional time series correlation feature vector and the humidity bidirectional time series correlation feature vector into a feature principal component significant fusion network to obtain a temperature-humidity principal component time series significant collaborative interaction representation vector; inputting the temperature-humidity principal component time series significant collaborative interaction representation vector, the temperature bidirectional time series correlation feature vector and the humidity bidirectional time series correlation feature vector into an insulation measurement factor multi-dimensional correlation analyzer to obtain an insulation measurement factor multi-dimensional correlation representation vector; based on the insulation measurement factor multi-dimensional correlation representation vector, obtaining an insulation resistance measurement compensation factor; calculating the product between the insulation resistance measurement compensation factor and the initial insulation resistance value to obtain the compensated insulation resistance value.

[0093] Here, those skilled in the art can understand that the specific operations of each step in the above resistance detection method for live working on distribution network have been described in detail in the description of the above Figures 1 to 5 resistance detection system for live working on distribution network, and therefore, the repeated description thereof will be omitted.

[0094] In summary, the resistance detection method for live working on distribution networks according to the embodiments of the present application is elucidated. It uses data analysis technology based on artificial intelligence to perform temporal feature correlation on the time queue of temperature values and the time queue of humidity values during the insulation detection process respectively, and performs significant fusion interaction based on query matching of principal components. Based on this, an insulation resistance measurement compensation factor is intelligently generated based on the multi-dimensional correlation representation features between the temperature temporal correlation features, the humidity temporal correlation features, and the significant interaction between temperature and humidity principal components, and the compensated insulation resistance value is calculated based on the compensation factor. In this way, through an automated measurement process, the accuracy of measurement data can be ensured, and thus a more accurate insulation resistance value can be obtained. Moreover, by dynamically compensating the insulation resistance value based on the real-time collected environmental data, the performance of the insulation material in the actual working environment can be more accurately reflected.

[0095] In several embodiments provided by the present application, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative. For example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation.

[0096] The modules described as separate components may or may not be physically separated. The components shown as modules may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0097] In addition, in each embodiment of the present application, the functional modules can be integrated in one processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above integrated unit can be implemented in the form of hardware, or in the form of hardware plus software functional modules.

[0098] For those skilled in the art, it is obvious that the present application is not limited to the details of the above exemplary embodiments, and can be implemented in other specific forms without departing from the spirit or basic characteristics of the present application.

Claims

1. A resistance detection system for live working in a distribution network, characterized in that: include: Insulation resistance detection tools, temperature sensors, humidity sensors and resistance measurement compensation correctors for live distribution network work; The insulation resistance detection tool for live distribution network operation is used to collect the initial insulation resistance value of the object under test; The temperature sensor and the humidity sensor are used to collect the time queue of the temperature value and the time queue of the humidity value during the insulation detection process respectively; The resistance measurement compensation corrector is used to compensate and correct the initial insulation resistance value based on the time queue of the temperature value and the time queue of the humidity value to obtain a compensated insulation resistance value; Wherein, the resistance measurement compensation corrector comprises: A temperature and humidity time series association module, used for inputting the time queue of the temperature value and the time queue of the humidity value into a sequence encoder to obtain a temperature bidirectional time series association feature vector and a humidity bidirectional time series association feature vector; A temperature-humidity principal component time series significant matching module is used to input the temperature bidirectional time series associated feature vector and the humidity bidirectional time series associated feature vector into a feature principal component significant fusion network to obtain a temperature-humidity principal component time series significant collaborative interaction representation vector; An insulation measurement factor multidimensional association module is used to input the temperature-humidity principal component time series significant collaborative interaction representation vector, the temperature bidirectional time series association feature vector, and the humidity bidirectional time series association feature vector into an insulation measurement factor multidimensional association analyzer to obtain an insulation measurement factor multidimensional association representation vector; An insulation resistance measurement compensation factor generation module, used to obtain an insulation resistance measurement compensation factor based on the insulation measurement factor multi-dimensional association representation vector; The insulation resistance value compensation module is used to calculate the product between the insulation resistance measurement compensation factor and the initial insulation resistance value to obtain the compensated insulation resistance value.

2. The resistance detection system for live working in distribution network according to claim 1, characterized in that: The temperature and humidity time series association module is used to: input the time queue of the temperature value and the time queue of the humidity value into a sequence encoder based on a bidirectional gated cyclic unit to obtain the temperature bidirectional time series association feature vector and the humidity bidirectional time series association feature vector.

3. The resistance detection system for live working in distribution network according to claim 2, characterized in that: The temperature-humidity main component time series significant matching module includes: A temperature and humidity bidirectional time series correlation feature standardization unit, used for performing standardization processing on the temperature bidirectional time series correlation feature vector and the humidity bidirectional time series correlation feature vector to obtain a standardized temperature bidirectional time series correlation feature vector and a standardized humidity bidirectional time series correlation feature vector; A temperature and humidity bidirectional time series correlation sample covariance matrix calculation unit, used to calculate the sample covariance matrix of the standardized temperature bidirectional time series correlation feature vector and the standardized humidity bidirectional time series correlation feature vector to obtain a temperature bidirectional time series correlation sample covariance matrix and a humidity bidirectional time series correlation sample covariance matrix; A temperature and humidity bidirectional time series correlation principal component feature extraction unit, used for performing matrix decomposition-based feature vector extraction on the temperature bidirectional time series correlation sample covariance matrix and the humidity bidirectional time series correlation sample covariance matrix to obtain a set of temperature bidirectional time series correlation principal component feature vectors and a set of humidity bidirectional time series correlation principal component feature vectors; A temperature and humidity bidirectional time series correlation feature query matching unit, used to input the set of temperature bidirectional time series correlation principal component feature vectors and the set of humidity bidirectional time series correlation principal component feature vectors into a maximum approximate query matching network to obtain a set of best matching pairs of temperature bidirectional time series correlation principal component feature vectors and humidity bidirectional time series correlation principal component feature vectors; A temperature-humidity bidirectional time series correlation principal component fusion unit is used to input each best matching pair of the temperature bidirectional time series correlation principal component feature vector and the humidity bidirectional time series correlation principal component feature vector in the set of best matching pairs of the temperature bidirectional time series correlation principal component feature vector and the humidity bidirectional time series correlation principal component feature vector into a semantic fine-grained gating joint module to obtain a set of temperature-humidity bidirectional time series correlation principal component fusion feature vectors; The temperature-humidity principal component time series significant collaborative interaction representation generation unit is used to cascade the set of the temperature-humidity bidirectional time series associated principal component fusion feature vectors to obtain the temperature-humidity principal component time series significant collaborative interaction representation vector.

4. The resistance detection system for live working in distribution network according to claim 3, characterized in that: The temperature and humidity bidirectional time series correlation feature standardization unit is used to: Calculating the mean and standard deviation of the temperature bidirectional time series association feature vector respectively to obtain the temperature bidirectional time series association feature mean and the temperature bidirectional time series association feature standard deviation; After subtracting the temperature bidirectional time series association feature vector from the temperature bidirectional time series association feature mean by position, the calculated temperature bidirectional time series offset vector and the temperature bidirectional time series association feature standard deviation are divided by position to obtain the standardized temperature bidirectional time series association feature vector; Calculating the mean and standard deviation of the humidity bidirectional time series correlation feature vector respectively to obtain the humidity bidirectional time series correlation feature mean and the humidity bidirectional time series correlation feature standard deviation; After subtracting the humidity bidirectional time series association feature vector from the humidity bidirectional time series association feature mean by position, the calculated humidity bidirectional time series offset vector is divided by the humidity bidirectional time series association feature standard deviation by position to obtain the standardized humidity bidirectional time series association feature vector.

5. The resistance detection system for live working in distribution network according to claim 4, characterized in that: The temperature and humidity bidirectional time series correlation sample covariance matrix calculation unit includes: a temperature bidirectional time series association sample covariance matrix generating subunit, configured to multiply the transposed vector of the standardized temperature bidirectional time series association feature vector by the standardized temperature bidirectional time series association feature vector, and then divide the obtained standardized temperature bidirectional time series association matrix by a value obtained by subtracting one from the length of the standardized temperature bidirectional time series association feature vector to obtain the temperature bidirectional time series association sample covariance matrix; The humidity bidirectional time series correlation sample covariance matrix generating subunit is used to multiply the transposed vector of the standardized humidity bidirectional time series correlation feature vector by the standardized humidity bidirectional time series correlation feature vector, and then divide the obtained standardized humidity bidirectional time series correlation matrix by the value obtained by subtracting one from the length of the standardized humidity bidirectional time series correlation feature vector to obtain the humidity bidirectional time series correlation sample covariance matrix.

6. The resistance detection system for live working in distribution network according to claim 5, characterized in that: The temperature and humidity bidirectional time series correlation feature query and matching unit is used to: Extracting a predetermined temperature bidirectional time series correlation principal component feature vector from the set of temperature bidirectional time series correlation principal component feature vectors; Calculating the cosine similarity between the predetermined temperature bidirectional time series association principal component feature vector and each humidity bidirectional time series association principal component feature vector in the set of humidity bidirectional time series association principal component feature vectors to obtain a set of matching query similarities; The humidity bidirectional time series association principal component feature vector corresponding to the largest matching query similarity in the set of matching query similarities and the predetermined temperature bidirectional time series association principal component feature vector are taken as the best matching pair of the predetermined temperature bidirectional time series association principal component feature vector and the humidity bidirectional time series association principal component feature vector.

7. The resistance detection system for live working in distribution network according to claim 6, characterized in that: The temperature-humidity bidirectional time series correlation principal component fusion unit is used to: Respectively calculating the position difference, position dot product and position addition between the best matching pairs of the temperature bidirectional time series association principal component eigenvector and the humidity bidirectional time series association principal component eigenvector to obtain a temperature-humidity time series association principal component difference vector, a temperature-humidity time series association principal component dot product vector and a temperature-humidity time series association principal component sum vector; The temperature-humidity time series associated principal component difference vector, the temperature-humidity time series associated principal component dot product vector and the temperature-humidity time series associated principal component sum vector are cascaded and then one-dimensional convolutional coding is performed to obtain a temperature-humidity time series associated principal component multi-dimensional fusion vector; The temperature-humidity time series correlation principal component multi-dimensional fusion vector is subjected to a local window-based maximum pooling process to obtain the temperature-humidity bidirectional time series correlation principal component fusion feature vector.

8. The resistance detection system for live working in distribution network according to claim 7, characterized in that: The insulation measurement factor multidimensional association module is used to: input the temperature-humidity principal component time series significant collaborative interaction representation vector, the temperature bidirectional time series association feature vector and the humidity bidirectional time series association feature vector into the insulation measurement factor multidimensional association analyzer based on the embedding layer to obtain the insulation measurement factor multidimensional association representation vector.

9. The resistance detection system for live-line working in distribution network according to claim 8, characterized in that: The insulation resistance measurement compensation factor generating module is used to: input the insulation measurement factor multi-dimensional association representation vector into a compensation factor generator based on a decoder to obtain the insulation resistance measurement compensation factor.

10. A resistance detection method for live working in a distribution network, characterized in that: include: Collect the initial insulation resistance value of the object under test; Collect the time queue of temperature value and humidity value during insulation detection; Based on the time queue of the temperature value and the time queue of the humidity value, compensating and correcting the initial insulation resistance value to obtain a compensated insulation resistance value; The method of compensating and correcting the initial insulation resistance value based on the time queue of the temperature value and the time queue of the humidity value to obtain a compensated insulation resistance value includes: Inputting the time queue of the temperature value and the time queue of the humidity value into a sequence encoder to obtain a temperature bidirectional time series association feature vector and a humidity bidirectional time series association feature vector; Inputting the temperature bidirectional time series correlation feature vector and the humidity bidirectional time series correlation feature vector into a feature principal component significant fusion network to obtain a temperature-humidity principal component time series significant collaborative interaction representation vector; Input the temperature-humidity principal component time series significant synergistic interaction representation vector, the temperature bidirectional time series association feature vector, and the humidity bidirectional time series association feature vector into an insulation measurement factor multidimensional association analyzer to obtain an insulation measurement factor multidimensional association representation vector; Based on the multi-dimensional association representation vector of the insulation measurement factor, obtaining an insulation resistance measurement compensation factor; The product of the insulation resistance measurement compensation factor and the initial insulation resistance value is calculated to obtain the compensated insulation resistance value.

Images