CN117907735A - Online intelligent monitoring system and method for power transmission line - Google Patents

Abstract

An on-line intelligent monitoring system and method for electric transmission line is disclosed. Firstly, acquiring current signals of all power transmission line branches in a preset time period to obtain a sequence of current signals, then uploading the sequence of current signals to an online intelligent power transmission line monitoring platform, then extracting current waveform correlation characteristics of the sequence of current signals to obtain a sequence of context current waveform characteristic vectors, and finally, identifying power transmission line branch fault characteristics in the sequence of context current waveform characteristic vectors to determine the fault branches. In this way, the current signal can be processed and diagnosed using advanced data processing and analysis algorithms without requiring time consuming inspection by hand.

Description

Online intelligent monitoring system and method for power transmission line

Technical Field

The application relates to the field of power transmission line monitoring, in particular to an online intelligent power transmission line monitoring system and method.

Background

With the development of economy, high-voltage transmission lines in various places are more and more, however, due to the long distance of the high-voltage transmission lines, the arrangement of geographical situation is complex and the limitation of test equipment, no effective, timely, simple and reliable daily maintenance method exists for the high-voltage transmission lines, especially for the lines which are located in remote areas such as original forests, oceans and the like or are not easy to enter the areas, only fault points are searched for maintenance after faults are generated, the high-voltage transmission lines cannot be prevented from being unburned, are very dangerous, the reaction repair time after the problems are generated is long, and the loss caused by power failure and the like is also large. For this reason, transmission line fault monitoring is becoming more and more necessary.

At present, three methods for monitoring faults of a power transmission line are mainly adopted, namely a short-circuit ranging method, a fault indicator method and a handheld monitor line inspection method. The short-circuit ranging method is to estimate the line length from the short-circuit point to the measuring point by measuring the line impedance from the short-circuit point to the measuring point so as to determine the position of the short-circuit point, and the method requires high equipment cost and cannot reflect fault branches. The fault indicator method is to install fault indicators along the line of the transmission line, when the line fails, fault characteristic current flows through the line of the fault section, such as: short circuit current, rush current, injection current, etc. By detecting the fault characteristic current, the fault indicator on the fault section line has an indication of mechanically turning over or lighting a lamp, and the fault indicator of the fault section line is not indicated, so that the line between the indication and the non-indication is the position of the fault point. The main problem of the fault indicator method is that the detection personnel must go along the fault line to inspect the indicator, which is long in time and large in workload. Similarly, the line patrol method for the handheld monitor is a method for detecting the characteristic current signal along the fault line by holding the monitor by a detector. This method also requires a long time and a large amount of effort.

Therefore, an online intelligent monitoring system and an online intelligent monitoring method for the power transmission line are expected.

Disclosure of Invention

The present application has been made to solve the above-mentioned technical problems. The embodiment of the application provides an online intelligent monitoring system and an online intelligent monitoring method for a power transmission line, which can collect and monitor current signals of branches of each power transmission line in real time, and automatically identify fault branches by analyzing current waveform characteristics without manual inspection.

According to one aspect of the application, there is provided an on-line intelligent monitoring method for a power transmission line, comprising:

Acquiring current signals of each power transmission line branch in a preset time period to obtain a sequence of the current signals;

uploading the sequence of the current signals to an online intelligent monitoring platform of the power transmission line; and

In the online intelligent monitoring platform of the power transmission line, processing and analyzing the sequence of the current signals to determine fault branches;

In the online intelligent monitoring platform of the power transmission line, the processing and analyzing the sequence of the current signals to determine fault branches comprises the following steps:

Extracting current waveform associated features of the sequence of current signals to obtain a sequence of context current waveform feature vectors; and

And identifying transmission line branch fault characteristics in the sequence of the context current waveform characteristic vectors to determine the fault branch.

According to another aspect of the present application, there is provided an on-line intelligent monitoring system for a power transmission line, comprising:

The current signal acquisition module is used for acquiring current signals of all power transmission line branches in a preset time period to obtain a sequence of the current signals;

the uploading module is used for uploading the sequence of the current signals to an online intelligent monitoring platform of the power transmission line; and

The analysis processing module is used for processing and analyzing the sequence of the current signals in the online intelligent monitoring platform of the power transmission line so as to determine fault branches;

wherein, the analysis processing module includes:

a current waveform associated feature extraction unit, configured to extract current waveform associated features of the sequence of current signals to obtain a sequence of context current waveform feature vectors; and

And the identification unit is used for identifying the fault characteristics of the power transmission line branches in the sequence of the context current waveform characteristic vectors so as to determine the fault branches.

Compared with the prior art, the system and the method for online intelligent monitoring of the power transmission line provided by the application have the advantages that firstly, current signals of all power transmission line branches in a preset time period are obtained to obtain a sequence of the current signals, then, the sequence of the current signals is uploaded to an online intelligent monitoring platform of the power transmission line, then, current waveform association characteristics of the sequence of the current signals are extracted to obtain a sequence of context current waveform characteristic vectors, and finally, power transmission line branch fault characteristics in the sequence of the context current waveform characteristic vectors are identified to determine the fault branches. In this way, the current signal can be processed and diagnosed using advanced data processing and analysis algorithms without requiring time consuming inspection by hand.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for the description of the embodiments will be briefly introduced below, the following drawings not being drawn to scale with respect to actual dimensions, emphasis instead being placed upon illustrating the gist of the present application.

Fig. 1 is a flowchart of an online intelligent monitoring method for a power transmission line according to an embodiment of the present application.

Fig. 2 is a flowchart of substep S130 of the power transmission line online intelligent monitoring method according to an embodiment of the present application.

Fig. 3 is a schematic diagram of the architecture of substep S130 of the power transmission line online intelligent monitoring method according to an embodiment of the present application.

Fig. 4 is a flowchart of substep S131 of the power transmission line online intelligent monitoring method according to an embodiment of the present application.

Fig. 5 is a flowchart of substep S132 of the power transmission line online intelligent monitoring method according to an embodiment of the present application.

Fig. 6 is a flowchart of sub-step S1321 of the power transmission line online intelligent monitoring method according to an embodiment of the present application.

Fig. 7 is a block diagram of an on-line intelligent monitoring system for a power transmission line according to an embodiment of the present application.

Fig. 8 is an application scenario diagram of an online intelligent monitoring method for a power transmission line according to an embodiment of the present application.

Detailed Description

The following description of the embodiments of the present application will be made clearly and fully with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some, but not all embodiments of the application. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are also within the scope of the application.

As used in the specification and in the claims, the terms "a," "an," "the," and/or "the" are not specific to a singular, but may include a plurality, unless the context clearly dictates otherwise. In general, the terms "comprises" and "comprising" merely indicate that the steps and elements are explicitly identified, and they do not constitute an exclusive list, as other steps or elements may be included in a method or apparatus.

Although the present application makes various references to certain modules in a system according to embodiments of the present application, any number of different modules may be used and run on a user terminal and/or server. The modules are merely illustrative, and different aspects of the systems and methods may use different modules.

A flowchart is used in the present application to describe the operations performed by a system according to embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed in order precisely. Rather, the various steps may be processed in reverse order or simultaneously, as desired. Also, other operations may be added to or removed from these processes.

Hereinafter, exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. It should be apparent that the described embodiments are only some embodiments of the present application and not all embodiments of the present application, and it should be understood that the present application is not limited by the example embodiments described herein.

Aiming at the technical problems, the technical conception of the application is to collect and monitor the current signals of all the power transmission line branches in real time and automatically identify the fault branches by analyzing the current waveform characteristics without manual inspection.

Based on this, fig. 1 is a flowchart of an online intelligent monitoring method for a power transmission line according to an embodiment of the present application. As shown in fig. 1, the method for online intelligent monitoring of a power transmission line according to an embodiment of the application includes the steps of: s110, acquiring current signals of each power transmission line branch in a preset time period to obtain a sequence of the current signals; s120, uploading the sequence of the current signals to an online intelligent monitoring platform of the power transmission line; and S130, processing and analyzing the sequence of the current signals in the online intelligent monitoring platform of the power transmission line to determine fault branches.

It should be noted that, in step S110, the acquisition of the current signal of the power transmission line may be performed in several ways: 1. sensor measurement: and installing a current sensor at a key position of the power transmission line, and acquiring current data of the power transmission line by measuring a current signal output by the current sensor. The sensor may be non-contact, such as a Current Transformer (CT) or Hall effect sensor, or contact, such as a current clamp (clamp-on ammeter). 2. A current sampler: a current sampler (current sampler) is used to sample the transmission line, and the current signal is converted into a digital signal for processing. Current samplers typically use analog-to-digital converters (ADCs) to convert continuous current signals to discrete digital signals that are then stored or transmitted to a monitoring system for analysis. 3. Communication interface: some power lines have installed thereon communication devices, such as terminal node units (RTUs) or intelligent power electronics devices (e.g., intelligent circuit breakers or intelligent current transformers), which can transmit current signals to a monitoring platform through a communication interface. Whichever method is used, the step S110 of acquiring current signals is to acquire and form a current signal sequence for each transmission line branch over a predetermined period of time.

Fig. 2 is a flowchart of substep S130 of the power transmission line online intelligent monitoring method according to an embodiment of the present application. Fig. 3 is a schematic diagram of the architecture of substep S130 of the power transmission line online intelligent monitoring method according to an embodiment of the present application. As shown in fig. 2 and fig. 3, according to an embodiment of the present application, in the power transmission line online intelligent monitoring platform, a sequence of the current signal is processed and analyzed to determine a fault branch, including: s131, extracting current waveform association features of the sequence of the current signals to obtain a sequence of context current waveform feature vectors; and S132, identifying transmission line branch fault characteristics in the sequence of the context current waveform characteristic vectors to determine the fault branches.

Specifically, in the technical scheme of the application, firstly, current signals of branches of each power transmission line in a preset time period are obtained to obtain a sequence of the current signals; and uploading the sequence of the current signals to an online intelligent monitoring platform of the power transmission line. Here, by acquiring the current signals of the branches of each power transmission line, the current transmission and the running state of the power transmission line can be known in real time. The current signal contains information such as the current magnitude, waveform and frequency of the circuit, and can reflect the load condition and fault information of the circuit. In addition, by uploading the current signal to the online intelligent monitoring platform, the current signal can be processed and diagnosed by utilizing an advanced data processing and analysis algorithm, and time-consuming inspection is not required manually.

In the online intelligent monitoring platform of the power transmission line, the sequence of the current signals is firstly passed through a current waveform feature extractor based on a convolutional neural network model to obtain a sequence of current waveform feature vectors. Here, the current waveform feature extractor based on the convolutional neural network model may extract a representative feature distribution from the current signal, i.e., capture important features in the current waveform, such as transient features representing the occurrence of a fault, etc.

In an actual application scenario, considering that the power transmission line where the fault branch is located affects current fluctuation of surrounding power transmission lines, in the technical scheme of the application, the sequence of the current waveform feature vectors is further used for representing the current correlation fluctuation feature among the power transmission lines by capturing context correlation information among the current waveform feature vectors based on a context correlation encoder of the converter module, so that the sequence of the context current waveform feature vectors is obtained.

Then, calculating semantic information metric values of each context current waveform feature vector in the sequence of context current waveform feature vectors relative to the feature distribution entirety of the sequence of context current waveform feature vectors to obtain a plurality of semantic information metric values. Here, by calculating the semantic information metric value of each context current waveform feature vector with respect to the feature distribution entirety of the sequence of the context current waveform feature vectors, the contribution degree of each context current waveform feature vector with respect to the feature distribution entirety with respect to the fault feature can be quantified.

Accordingly, in step S131, as shown in fig. 4, extracting the current waveform correlation feature of the sequence of current signals to obtain a sequence of context current waveform feature vectors includes: s1311, extracting local current waveform characteristics of the sequence of current signals to obtain a sequence of current waveform characteristic vectors; and S1312, performing associated feature extraction on the sequence of current waveform feature vectors by using a deep learning network model to obtain the sequence of context current waveform feature vectors. It should be understood that it is possible to provide,

Wherein, in step S1311, extracting the local current waveform feature of the sequence of current signals to obtain a sequence of current waveform feature vectors, including: and passing the sequence of the current signals through a current waveform characteristic extractor based on a convolutional neural network model to obtain the sequence of the current waveform characteristic vectors.

Specifically, the current waveform feature extractor based on the convolutional neural network model comprises an input layer, a convolutional layer, an activation layer, a pooling layer and an output layer. It is noted that convolutional neural networks (Convolutional Neural Network, CNN) are a deep learning model that performs well when processing data (e.g., images) having a grid structure, and can also be used to process sequence data (e.g., time series). The convolutional neural network is mainly characterized in that local features of input data are extracted through convolutional operation, and abstraction and combination of features are carried out in a layer-by-layer stacking mode. It generally comprises the following core components: 1. input Layer (Input Layer): input data, such as image or sequence data, is received. 2. Convolution layer (Convolutional Layer): local features of the input data are extracted by a convolution operation. The convolution layer consists of a plurality of convolution kernels (filters), each of which is responsible for extracting a feature. The convolution kernel slides on the input data, and a feature map (feature map) is obtained by performing a convolution operation on the local area. 3 Activation Layer (Activation Layer): the output of the convolution layer is subjected to nonlinear transformation, and nonlinear characteristics are introduced. Common activation functions include ReLU (RECTIFIED LINEAR Unit), sigmoid, tanh, and the like. 4. Pooling layer (Pooling Layer): and downsampling the feature map to reduce the spatial dimension of the features. Common pooling operations include maximum pooling (Max Pooling) and average pooling (Average Pooling) for extracting spatial invariance of features and reducing the number of parameters. 5. Output Layer (Output Layer): and classifying, regressing or outputting other tasks on the characteristics of the last layer. The use of convolutional neural network models in current waveform feature extraction may help capture local patterns and features in the current signal. Through convolution operation, the model can automatically learn current waveform characteristics with different scales and frequencies, so that the characteristics of a current signal are better represented. This helps to improve the accuracy of fault analysis and diagnosis. In general, a current waveform feature extractor based on a convolutional neural network model extracts local features of a current signal sequence by utilizing a convolutional operation, and performs feature abstraction and combination through a hierarchical structure so as to obtain a sequence of current waveform feature vectors, and the feature extractor can effectively capture important features in the current signal and provide useful information for subsequent fault analysis and diagnosis.

Wherein, in step S1312, the deep learning network model is a context-dependent encoder based on a converter module; the method for extracting the associated features of the sequence of the current waveform feature vectors by using a deep learning network model to obtain the sequence of the context current waveform feature vectors comprises the following steps: passing the sequence of current waveform feature vectors through the context-dependent encoder based on a converter module to obtain the sequence of context current waveform feature vectors.

It should be noted that the converter module (transducer) is a deep learning model for processing sequence data, and adopts Self-Attention mechanism (Self-Attention) to capture the dependency relationship between different positions in the sequence, so that the modeling capability and the parallel computing performance are better. The converter module is mainly composed of the following two key components: 1. Self-Attention mechanism (Self-Attention): the self-attention mechanism is able to calculate the attention weight between each location and other locations in the sequence, capturing the contextual relationship in the sequence. It obtains a Query (Query), key (Key) and Value (Value) vector by linearly transforming the input sequence, and then calculates an attention weight and multiplies it with the Value vector to obtain a context representation. The self-attention mechanism can consider all positions in the sequence at the same time, thereby capturing global dependencies better. Feed forward neural network (Feed-Forward Neural Network): a feed-forward neural network is another important component in the converter module for non-linear transformation of the output of the self-attention mechanism. It consists of two fully connected layers and an activation function, enabling the introduction of more complex feature transformations and modeling capabilities. The converter module builds a deep network structure by stacking multiple self-attention layers and feedforward neural network layers. In processing the sequence data, the converter module can simultaneously consider global dependencies in the sequence and capture correlation features between different positions through a self-attention mechanism. In the context-dependent encoder, a sequence of current waveform feature vectors is subject to a dependent feature extraction using a network architecture based on a converter module. Through a self-attention mechanism, the converter module is able to learn the context and dependencies in the sequence of current waveform feature vectors, thereby generating a sequence of contextual current waveform feature vectors. The feature extraction mode can more comprehensively capture global features and associated features in the current signals, and is beneficial to improving the accuracy of fault analysis and diagnosis.

Specifically, passing the sequence of current waveform feature vectors through the context-dependent encoder of the converter-based module to obtain the sequence of context current waveform feature vectors includes: one-dimensional arrangement is carried out on the sequence of the current waveform feature vectors so as to obtain a current waveform global feature vector; calculating the product between the current waveform global eigenvector and the transpose vector of each current waveform eigenvector in the sequence of current waveform eigenvectors to obtain a plurality of self-attention correlation matrices; respectively carrying out standardization processing on each self-attention correlation matrix in the plurality of self-attention correlation matrices to obtain a plurality of standardized self-attention correlation matrices; obtaining a plurality of probability values by using a Softmax classification function through each normalized self-attention correlation matrix in the normalized self-attention correlation matrices; and weighting each current waveform feature vector in the sequence of current waveform feature vectors by taking each probability value in the plurality of probability values as a weight to obtain the sequence of context current waveform feature vectors.

It should be understood that in the application scenario of the present application, fault branches in each power transmission line need to be identified, and fault feature distributions with different forms exist in current waveform feature distributions expressed by current signals of each power transmission line, and correlation about fault features in the current waveform feature distributions expressed by current signals of each power transmission line is measured by calculating semantic information metric values, so as to distinguish other power transmission lines running normally. Specifically, when the semantic information measurement value is larger, the transmission line corresponding to the semantic information measurement value is obviously distributed with fault characteristics; otherwise, when the semantic information measurement value is smaller, the current waveform characteristic distribution of the power transmission line corresponding to the semantic information measurement value is normal. In this way the location of a possible faulty branch is identified.

And then comparing the maximum one of the plurality of semantic information metric values with a predetermined threshold value, and determining that the power transmission line branch corresponding to the maximum one is a fault branch in response to the maximum one of the plurality of semantic information metric values being greater than the predetermined threshold value.

Accordingly, in step S132, as shown in fig. 5, identifying transmission line branch fault characteristics in the sequence of contextual current waveform feature vectors to determine the fault branch includes: s1321, calculating semantic information metric values of the feature distribution of each context current waveform feature vector in the sequence of context current waveform feature vectors relative to the sequence of context current waveform feature vectors to obtain a plurality of semantic information metric values; and S1322, comparing the maximum one of the plurality of semantic information metric values with a preset threshold value, and determining that the power transmission line branch corresponding to the maximum one is the fault branch in response to the maximum one of the plurality of semantic information metric values being greater than the preset threshold value.

It should be appreciated that two steps S1321 and S1322 are involved in identifying transmission line branch fault signatures in the sequence of contextual current waveform signature vectors to determine a fault branch. In step S1321, by analyzing each feature vector in the sequence of the feature vectors of the contextual current waveform, their feature distribution with respect to the overall sequence can be calculated. The semantic information metric value of the feature distribution can be used for measuring the importance of each feature vector in the sequence and the association degree of each feature vector with the fault branch, and a group of values for representing the importance of each feature vector in the sequence can be obtained by calculating the semantic information metric values of a plurality of feature vectors. In step S1322, the largest value among the semantic information metric values of the plurality of feature vectors is found by comparing them. The maximum value is then compared with a predetermined threshold. And if the maximum value is greater than the preset threshold value, determining that the power transmission line branch corresponding to the maximum value is a fault branch. This is because the largest semantic information metric value indicates the highest importance of the corresponding feature vector in the sequence and the highest degree of association with the faulty branch. Through the two steps, the sequence of the characteristic vectors of the context current waveform can be analyzed and judged, the characteristic vector related to the fault branch is found, and the fault branch is determined. The method utilizes the semantic information metric value to evaluate the importance of the feature vector and the association degree with the fault branch, thereby realizing the identification and the positioning of the fault branch.

In step S1321, as shown in fig. 6, calculating semantic information metric values of each context current waveform feature vector in the sequence of context current waveform feature vectors with respect to a feature distribution of the sequence of context current waveform feature vectors to obtain a plurality of semantic information metric values, including: s13211, cascading the sequence of the context current waveform feature vectors into cascading feature vectors; s13212, performing feature distribution optimization on the cascade feature vector to obtain an optimized cascade feature vector; and S13213, calculating semantic information metric values of each context current waveform feature vector in the sequence of context current waveform feature vectors relative to the optimized cascade feature vector to obtain the plurality of semantic information metric values.

It should be understood that in calculating the semantic information metric values of the feature vectors in the context current waveform feature vector sequence with respect to the feature distribution ensemble of the ensemble sequence to obtain a plurality of semantic information metric values, three steps S13211, S13212 and S13213 are involved. In step S13211, the sequences of the context current waveform feature vectors are connected according to a certain order to form a cascade feature vector, and the cascade feature vector integrates the information of the whole sequence, so that the features in the sequence can be more comprehensively represented. In step S13212, the feature distribution optimization process is performed on the cascade feature vector, where the optimization process may include some mathematical transformations, filtering or other processing methods, so that the cascade feature vector better conforms to the feature distribution of the fault branch, and the importance and distribution situation of the features in the sequence can be better reflected by the optimized cascade feature vector. In step S13213, by calculating a semantic information metric value of each feature vector in the context current waveform feature vector sequence relative to the optimized cascade feature vector, the metric value may measure similarity and importance between each feature vector and the optimized cascade feature vector, thereby obtaining a plurality of semantic information metric values. Through the three steps, the feature distribution optimization and the semantic information metric value calculation can be carried out on the context current waveform feature vector sequence, so that a plurality of semantic information metric values are obtained. These metric values can be used to evaluate the importance of each feature vector in the sequence and the degree of association with the optimized concatenated feature vector for further use in the identification and localization of faulty branches.

Specifically, in step S13213, calculating semantic information metric values of each context current waveform feature vector in the sequence of context current waveform feature vectors relative to the optimized cascade feature vector to obtain the plurality of semantic information metric values includes: calculating semantic information metric values of each context current waveform feature vector in the sequence of context current waveform feature vectors relative to the optimized cascade feature vector by using the following semantic information metric formula to obtain a plurality of semantic information metric values; wherein, the semantic information measurement formula is:

s_i＝σ(A*h_i+B*V′)

Where a is a vector of 1×n _w, N _w is the dimension of the context current waveform feature vector, h _i is the i-th context current waveform feature vector, V' is the optimized cascade feature vector, B is a vector of 1×n _h, N _h is the dimension of the optimized cascade feature vector, σ is a Sigmoid function, and s _i is the i-th semantic information metric value.

In the above technical solution, each of the sequence of context current waveform feature vectors expresses a context-associated feature representation of a waveform semantic feature of a current signal of a corresponding power transmission line branch, but after current waveform semantic feature extraction and context encoding, the sequence of context current waveform feature vectors as a whole may have inconsistency and instability of a timing feature distribution, that is, may have an abnormal local distribution that diverges with respect to the overall feature distribution, thereby affecting accuracy of calculating a plurality of semantic information metric values of semantic weights of the respective context current waveform feature vectors with respect to an overall of the sequence of the plurality of context current waveform feature vectors, taking into account source data distribution pattern differences of the current signal of the power transmission line branch in a time-series direction. Based on this, the applicant of the present application optimizes the sequence of the context current waveform feature vectors as a whole.

Accordingly, in one example, feature distribution optimization is performed on the cascade feature vector to obtain an optimized cascade feature vector, including: performing feature distribution optimization on the cascade feature vector by using the following optimization formula to obtain the optimized cascade feature vector; wherein, the optimization formula is:

Wherein V is a cascading feature vector obtained after cascading the sequence of the context current waveform feature vector, V _i is a feature value of an ith position of the cascading feature vector, i ₁ and i ₂ are a 1-norm and a 2-norm of the cascading feature vector V respectively, L is a length of the cascading feature vector V, α is a weight super-parameter related to V _i, exp (·) represents an exponential operation of a numerical value, the exponential operation of the numerical value represents a natural exponential function value calculated by the numerical value as a power, and V' _i is a feature value of the ith position of the optimized cascading feature vector.

Here, the global feature distribution of the cascade feature vector V has a certain repeatability to the local mode change through the structural consistency and stability representation of the overall feature distribution of the cascade feature vector V obtained after the sequence cascade of the context current waveform feature vector under the rigid structure of the absolute distance and the non-rigid structure of the spatial distance respectively, so that when the semantic representation weights of the overall feature distribution of the context current waveform feature vector relative to the cascade feature vector are calculated, the scale and rotation change of the weight representation are robust, and the accuracy of the semantic information metric values is improved.

In summary, the power transmission line online intelligent monitoring method based on the embodiment of the application is explained, and can process and diagnose the current signal by utilizing an advanced data processing and analysis algorithm without time-consuming inspection by manpower.

Fig. 7 is a block diagram of an on-line intelligent monitoring system 100 for a power transmission line according to an embodiment of the present application. As shown in fig. 7, an online intelligent monitoring system 100 for a power transmission line according to an embodiment of the present application includes: a current signal acquisition module 110, configured to acquire current signals of each power transmission line branch in a predetermined period of time to obtain a sequence of current signals; the uploading module 120 is configured to upload the sequence of the current signal to an online intelligent monitoring platform of the power transmission line; the analysis processing module 130 is used for processing and analyzing the sequence of the current signals in the online intelligent monitoring platform of the power transmission line to determine fault branches; wherein the analysis processing module 130 includes: a current waveform associated feature extraction unit, configured to extract current waveform associated features of the sequence of current signals to obtain a sequence of context current waveform feature vectors; and an identification unit for identifying transmission line branch fault characteristics in the sequence of the contextual current waveform feature vectors to determine the fault branch.

Here, it will be understood by those skilled in the art that the specific functions and operations of the respective modules in the above-described power transmission line online intelligent monitoring system 100 have been described in detail in the above description of the power transmission line online intelligent monitoring method with reference to fig. 1 to 6, and thus, repetitive descriptions thereof will be omitted.

As described above, the power transmission line online intelligent monitoring system 100 according to the embodiment of the present application may be implemented in various wireless terminals, for example, a server or the like having a power transmission line online intelligent monitoring algorithm. In one example, the transmission line online intelligent monitoring system 100 according to an embodiment of the present application can be integrated into a wireless terminal as a software module and/or hardware module. For example, the transmission line online intelligent monitoring system 100 may be a software module in the operating system of the wireless terminal, or may be an application developed for the wireless terminal; of course, the transmission line online intelligent monitoring system 100 can also be one of a plurality of hardware modules of the wireless terminal.

Alternatively, in another example, the transmission line online intelligent monitoring system 100 and the wireless terminal may be separate devices, and the transmission line online intelligent monitoring system 100 may be connected to the wireless terminal through a wired and/or wireless network and transmit interactive information in a agreed data format.

Fig. 8 is an application scenario diagram of an online intelligent monitoring method for a power transmission line according to an embodiment of the present application. As shown in fig. 8, in this application scenario, first, current signals (for example, D illustrated in fig. 8) of respective transmission line branches within a predetermined period of time are acquired to obtain a sequence of current signals, and then the sequence of current signals is input to a server (for example, S illustrated in fig. 8) in which a transmission line online intelligent monitoring algorithm is deployed, wherein the server can process the sequence of current signals using the transmission line online intelligent monitoring algorithm to determine a fault branch.

Furthermore, those skilled in the art will appreciate that the various aspects of the application are illustrated and described in the context of a number of patentable categories or circumstances, including any novel and useful procedures, machines, products, or materials, or any novel and useful modifications thereof. Accordingly, aspects of the application may be performed entirely by hardware, entirely by software (including firmware, resident software, micro-code, etc.) or by a combination of hardware and software. The above hardware or software may be referred to as a "data block," module, "" engine, "" unit, "" component, "or" system. Furthermore, aspects of the application may take the form of a computer product, comprising computer-readable program code, embodied in one or more computer-readable media.

Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. It will be further understood that terms, such as those defined in commonly used dictionaries, should be interpreted as having a meaning that is consistent with their meaning in the context of the relevant art and will not be interpreted in an idealized or overly formal sense unless expressly so defined herein.

The foregoing is illustrative of the present invention and is not to be construed as limiting thereof. Although a few exemplary embodiments of this invention have been described, those skilled in the art will readily appreciate that many modifications are possible in the exemplary embodiments without materially departing from the novel teachings and advantages of this invention. Accordingly, all such modifications are intended to be included within the scope of this invention as defined in the following claims. It is to be understood that the foregoing is illustrative of the present invention and is not to be construed as limited to the specific embodiments disclosed, and that modifications to the disclosed embodiments, as well as other embodiments, are intended to be included within the scope of the appended claims. The invention is defined by the claims and their equivalents.

Claims (10)

1. An online intelligent monitoring method for a power transmission line is characterized by comprising the following steps:

Acquiring current signals of each power transmission line branch in a preset time period to obtain a sequence of the current signals;

uploading the sequence of the current signals to an online intelligent monitoring platform of the power transmission line; and

In the online intelligent monitoring platform of the power transmission line, processing and analyzing the sequence of the current signals to determine fault branches;

In the online intelligent monitoring platform of the power transmission line, the processing and analyzing the sequence of the current signals to determine fault branches comprises the following steps:

Extracting current waveform associated features of the sequence of current signals to obtain a sequence of context current waveform feature vectors; and

And identifying transmission line branch fault characteristics in the sequence of the context current waveform characteristic vectors to determine the fault branch.

2. The method of on-line intelligent monitoring of a power transmission line according to claim 1, wherein extracting current waveform correlation features of the sequence of current signals to obtain a sequence of context current waveform feature vectors comprises:

Extracting local current waveform characteristics of the sequence of current signals to obtain a sequence of current waveform characteristic vectors; and

And performing associated feature extraction on the sequence of the current waveform feature vectors by using a deep learning network model to obtain the sequence of the context current waveform feature vectors.

3. The method for online intelligent monitoring of a power transmission line according to claim 2, wherein extracting the local current waveform characteristics of the sequence of current signals to obtain the sequence of current waveform characteristic vectors comprises:

And passing the sequence of the current signals through a current waveform characteristic extractor based on a convolutional neural network model to obtain the sequence of the current waveform characteristic vectors.

4. The method for online intelligent monitoring of a power transmission line according to claim 3, wherein the current waveform feature extractor based on the convolutional neural network model comprises an input layer, a convolutional layer, an activation layer, a pooling layer and an output layer.

5. The method for online intelligent monitoring of a power transmission line according to claim 4, wherein the deep learning network model is a context-dependent encoder based on a converter module;

the method for extracting the associated features of the sequence of the current waveform feature vectors by using a deep learning network model to obtain the sequence of the context current waveform feature vectors comprises the following steps:

Passing the sequence of current waveform feature vectors through the context-dependent encoder based on a converter module to obtain the sequence of context current waveform feature vectors.

6. The method of on-line intelligent monitoring of a power transmission line of claim 5, wherein passing the sequence of current waveform feature vectors through the context-dependent encoder based on a converter module to obtain the sequence of context current waveform feature vectors comprises:

One-dimensional arrangement is carried out on the sequence of the current waveform feature vectors so as to obtain a current waveform global feature vector;

Calculating the product between the current waveform global eigenvector and the transpose vector of each current waveform eigenvector in the sequence of current waveform eigenvectors to obtain a plurality of self-attention correlation matrices;

Respectively carrying out standardization processing on each self-attention correlation matrix in the plurality of self-attention correlation matrices to obtain a plurality of standardized self-attention correlation matrices;

obtaining a plurality of probability values by using a Softmax classification function through each normalized self-attention correlation matrix in the normalized self-attention correlation matrices; and

And weighting each current waveform characteristic vector in the sequence of current waveform characteristic vectors by taking each probability value in the plurality of probability values as a weight so as to obtain the sequence of the context current waveform characteristic vectors.

7. The method of on-line intelligent monitoring of power transmission lines of claim 6, wherein identifying power transmission line branch fault signatures in the sequence of contextual current waveform signature vectors to determine the fault branches comprises:

calculating semantic information metric values of the feature distribution of each context current waveform feature vector in the sequence of context current waveform feature vectors relative to the sequence of context current waveform feature vectors to obtain a plurality of semantic information metric values; and

Comparing the largest one of the plurality of semantic information metric values with a predetermined threshold, and determining that the power transmission line branch corresponding to the largest one is the faulty branch in response to the largest one of the plurality of semantic information metric values being greater than the predetermined threshold.

8. The method of on-line intelligent monitoring of a power transmission line according to claim 7, wherein calculating semantic information metric values of each of the sequence of contextual current waveform feature vectors relative to a feature distribution ensemble of the sequence of contextual current waveform feature vectors to obtain a plurality of semantic information metric values comprises:

cascading the sequence of the context current waveform feature vectors into cascading feature vectors;

Performing feature distribution optimization on the cascade feature vectors to obtain optimized cascade feature vectors; and

And calculating semantic information metric values of each context current waveform feature vector in the sequence of context current waveform feature vectors relative to the optimized cascade feature vector to obtain the plurality of semantic information metric values.

9. The method of on-line intelligent monitoring of a power transmission line according to claim 8, wherein calculating semantic information metric values of each context current waveform feature vector in the sequence of context current waveform feature vectors relative to the optimized cascade feature vector to obtain the plurality of semantic information metric values comprises:

calculating semantic information metric values of each context current waveform feature vector in the sequence of context current waveform feature vectors relative to the optimized cascade feature vector by using the following semantic information metric formula to obtain a plurality of semantic information metric values;

Wherein, the semantic information measurement formula is:

s_i＝σ(A*h_i+B*V′)

Where a is a vector of 1×n _w, N _w is the dimension of the context current waveform feature vector, h _i is the i-th context current waveform feature vector, V' is the optimized cascade feature vector, B is a vector of 1×n _n, N _h is the dimension of the optimized cascade feature vector, σ is a Sigmoid function, and s _i is the i-th semantic information metric value.

10. An on-line intelligent monitoring system for a power transmission line, comprising:

The current signal acquisition module is used for acquiring current signals of all power transmission line branches in a preset time period to obtain a sequence of the current signals;

the uploading module is used for uploading the sequence of the current signals to an online intelligent monitoring platform of the power transmission line; and

The analysis processing module is used for processing and analyzing the sequence of the current signals in the online intelligent monitoring platform of the power transmission line so as to determine fault branches;

wherein, the analysis processing module includes:

a current waveform associated feature extraction unit, configured to extract current waveform associated features of the sequence of current signals to obtain a sequence of context current waveform feature vectors; and

And the identification unit is used for identifying the fault characteristics of the power transmission line branches in the sequence of the context current waveform characteristic vectors so as to determine the fault branches.