Cable fault monitoring and early warning method and system

Abstract

The invention discloses a cable fault monitoring and early warning method and system, and relates to the technical field of fault monitoring. The method comprises the steps that cable operation state parameters and environment parameters are collected in real time through a distributed sensor array, and the cable operation state parameters comprise cable temperature, three-phase current, partial discharge signals, insulation resistance, grounding current and mechanical vibration amplitude; the environment parameters at least comprise temperature and humidity corresponding to a cable laying environment; preprocessing the cable operation state parameters and the environment parameters, and generating a standardized feature matrix based on preprocessed multi-source data; inputting the standardized feature matrix into a pre-trained fault prediction model for processing to obtain a fault type and a fault risk index corresponding to the current operation state of the cable; and triggering an early warning mechanism according to the fault type and the fault risk index, and outputting corresponding fault position information.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of fault monitoring, in particular to a cable fault monitoring and early warning method and system. BACKGROUND

[0002] With the acceleration of urbanization, the importance of cable comprehensive pipe gallery is gradually highlighted. However, due to the increasingly complex cable laying environment, the increasing power load and the long-term operation of the cable, the frequency of cable failure is gradually increasing, which not only affects the stability of power supply, but also may cause serious safety hazards. Therefore, how to accurately and effectively monitor the running state of the cable and timely warn potential faults has become a technical problem to be solved at present.

[0003] Most of the existing monitoring methods rely on manual inspection or simple sensor alarms, which have problems such as low efficiency, high false alarm rate and difficult positioning, and cannot meet the needs of modern power systems for high reliability and intelligence. Based on this, it is necessary to study a more scientific cable fault monitoring and early warning method and system to improve the intelligence, accuracy and reliability of cable fault monitoring. SUMMARY

[0004] To achieve the above-mentioned purpose, one aspect of an embodiment of the present application provides a cable fault monitoring and early warning method, which comprises: real-time acquisition of cable running state parameters and environmental parameters by a distributed sensor array, wherein the cable running state parameters include cable temperature, three-phase current, partial discharge signal, insulation resistance, grounding current and mechanical vibration amplitude, and the environmental parameters at least include temperature and humidity corresponding to the cable laying environment; preprocessing of the cable running state parameters and environmental parameters, and generating a standardized feature matrix based on the preprocessed multi-source data, wherein the preprocessing includes denoising and alignment; inputting the standardized feature matrix into a pre-trained fault prediction model for processing to obtain a fault type and a fault risk index corresponding to the current running state of the cable; wherein the fault prediction model comprises a deep feature extraction network and a fault mapping network, the deep feature extraction network is used for deep feature mining and pattern recognition on the standardized feature matrix to obtain a high-dimensional feature representation corresponding to the current running state of the cable; the fault mapping network is used for mapping the high-dimensional feature representation to real-time evaluate the current running state of the cable to obtain the fault type and the fault risk index; triggering an early warning mechanism according to the fault type and the fault risk index, and outputting corresponding fault location information.

[0005] In some embodiments, the preprocessing of the cable running state parameters and environmental parameters comprises: The wavelet threshold denoising and Kalman filtering algorithm are used to remove signal noise in the cable operation state parameters and environmental parameters; Then, the cable operation state parameters and environmental parameters are processed by time sequence alignment through an interpolation algorithm and a timestamp alignment technology, to obtain preprocessed multi-source data.

[0006] In some embodiments, the standardized feature matrix is generated based on the preprocessed multi-source data, including: The preprocessed multi-source data corresponding to the current moment and N moments before the current moment are obtained; The preprocessed multi-source data corresponding to each moment is subjected to feature extraction, to obtain a feature vector corresponding to each moment, wherein the feature vector includes dynamic change features and statistical features calculated based on the multiple cable operation state parameters and environmental parameters collected by the distributed sensor array; The feature vector is subjected to standardization processing, and the standardized feature vectors corresponding to the current moment and N moments before the current moment are combined into a two-dimensional matrix in time sequence, to obtain the standardized feature matrix.

[0007] In some embodiments, the preprocessed multi-source data corresponding to each moment is subjected to feature extraction, to obtain a feature vector corresponding to each moment, including: For the multiple cable operation state parameters and environmental parameters collected by the distributed sensor array at the target moment, the change rate of data collected by each collection point relative to adjacent collection points in each parameter is calculated, to obtain dynamic change features of the cable operation state parameters and environmental parameters at different positions; The mean, variance, peak value and valley value of each parameter are calculated, to obtain statistical features corresponding to the target moment; The dynamic change features and statistical features are integrated, to obtain a feature vector corresponding to the target moment.

[0008] In some embodiments, the dynamic change features and statistical features are integrated, to obtain a feature vector corresponding to the target moment, including: The fluctuation correlation degree between the dynamic change features corresponding to each parameter and the dynamic change features corresponding to other parameters is calculated; The dynamic change features, fluctuation correlation degrees and statistical features are spliced in a preset order, to obtain a feature vector corresponding to the target moment.

[0009] In some embodiments, the fault prediction model is obtained based on the following manner: Obtaining training sample data, wherein the training sample data includes cable operation state parameters, environmental parameters corresponding to a sample time period, and at least one fault type and a fault risk index corresponding thereto; Preprocessing the training sample data to generate a standardized feature matrix corresponding to the training sample data as input and the fault type and fault risk index corresponding to the training sample data as a label; Building an initial fault prediction model by using a deep learning framework, wherein the initial fault prediction model includes a deep feature extraction network and a fault mapping network; Iteratively training the initial fault prediction model by using the training sample data, and optimizing network parameters of the deep feature extraction network and the fault mapping network by using a back propagation algorithm until a prediction accuracy of the initial fault prediction model meets a preset condition, to obtain a trained fault prediction model. In some embodiments, triggering a warning mechanism according to the fault type and fault risk index includes: Mapping the fault type and fault risk index by using a preconfigured warning mechanism mapping table to obtain a warning mechanism corresponding to the current fault type and fault risk index.

[0010] In some embodiments, the fault location information is determined based on the following manner: Obtaining a preset spatial coordinate corresponding to each collection point in the distributed sensor array; Determining a corresponding key sensitive parameter according to the fault type; Calculating an abnormal deviation degree of the key sensitive parameter at each collection point, wherein the abnormal deviation degree is a ratio of an absolute value of a difference between a parameter value of the collection point and a preset normal threshold range and an upper limit or a lower limit of the threshold range; Obtaining the fault location information according to a spatial coordinate corresponding to a collection point with the maximum abnormal deviation degree.

[0011] In some embodiments, the method further includes: Regarding the spatial coordinate corresponding to the collection point with the maximum abnormal deviation degree as an initial fault location, and correcting the initial fault location by combining an abnormal deviation degree distribution of adjacent collection points to obtain the fault location information; Wherein, the correction process includes: Taking the initial fault location as a center, and then performing a weighted average calculation according to a spatial distance between adjacent collection points and the initial fault location and an abnormal deviation degree corresponding to the adjacent collection points; Determining one equivalent fault point on each of left and right sides of the initial fault location based on a calculation result of the weighted average calculation; The fault position information is obtained based on the spatial coordinates corresponding to the equivalent fault point.

[0012] Another aspect of the embodiments of the present specification also provides a cable fault monitoring and early warning system, which comprises: A monitoring data acquisition module is configured to collect cable operation state parameters and environmental parameters in real time through a distributed sensor array, wherein the cable operation state parameters include cable temperature, three-phase current, partial discharge signal, insulation resistance, grounding current, and mechanical vibration amplitude, and the environmental parameters at least include temperature and humidity corresponding to a cable laying environment. A preprocessing module is configured to preprocess the cable operation state parameters and environmental parameters, and generate a standardized feature matrix based on the preprocessed multi-source data, wherein the preprocessing includes denoising and alignment. A fault prediction module is configured to input the standardized feature matrix into a pre-trained fault prediction model for processing, so as to obtain a fault type and a fault risk index corresponding to a current operation state of the cable; wherein the fault prediction model comprises a deep feature extraction network and a fault mapping network, the deep feature extraction network is configured to perform deep feature mining and pattern recognition on the standardized feature matrix, so as to obtain a high-dimensional feature representation corresponding to the current operation state of the cable; and the fault mapping network is configured to map the high-dimensional feature representation, so as to perform real-time evaluation on the current operation state of the cable, and obtain the fault type and the fault risk index. An early warning module is configured to trigger an early warning mechanism according to the fault type and the fault risk index, and output corresponding fault position information.

[0013] The cable fault monitoring and early warning method and system provided by the embodiments of the present specification can bring at least the following beneficial effects: by preprocessing cable operation state parameters and environmental parameters collected in real time through a distributed sensor array to generate a standardized feature matrix, and then inputting the standardized feature matrix into a pre-trained fault prediction model for processing to obtain a corresponding fault type and fault risk index, the deep feature extraction network and the fault mapping network contained in the fault prediction model can be used to accurately and reliably perform real-time intelligent evaluation on the current operation state of the cable, so as to improve the intelligence, accuracy and reliability of cable fault monitoring to a certain extent.

[0014] Additional features will be set forth in the description that follows, and in part will be apparent from the description, or can be learned by practice of the presented embodiments of the disclosure. The features will be described in the following detailed description, which will set forth novel methods, tools, and combinations in connection with the described methods and tools. BRIEF DESCRIPTION OF DRAWINGS

[0015] The present specification will be further described in the manner of exemplary embodiments, which will be described in detail with reference to the accompanying drawings. The embodiments are not restrictive, and in the embodiments, the same reference numbers denote the same structures, in which: Figure 1 is an exemplary flow chart of a cable fault monitoring and early warning method according to some embodiments of the present specification; Figure 2 is an exemplary structural schematic diagram of a fault prediction model according to some embodiments of the present specification; Figure 3 is an exemplary structural schematic diagram of a deep feature extraction network according to some embodiments of the present specification; Figure 4 is an exemplary module diagram of a cable fault monitoring and early warning system according to some embodiments of the present specification. DETAILED DESCRIPTION

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present specification, the drawings needed to be used in the embodiment description will be briefly introduced below. Obviously, the drawings in the following description are only some examples or embodiments of the present specification, and for those skilled in the art, the present specification can be applied to other similar scenarios without creative labor on the basis of these drawings. Unless it is clear from the language environment or otherwise stated, the same reference numbers in the drawings represent the same structure or operation.

[0017] It should be understood that the "system", "device", "unit" and / or "module" used in the present specification is a method for distinguishing different components, elements, parts, portions or assemblies at different levels. However, if other words can achieve the same purpose, the words can be replaced by other expressions.

[0018] As shown in the present specification and claims, unless the context clearly indicates otherwise, "one", "a", "an", and / or "the" do not refer to the singular, but can also include the plural. Generally speaking, the terms "comprise" and "include" only indicate the inclusion of the steps and elements explicitly identified, and these steps and elements do not constitute an exclusive list, and the method or device can also include other steps or elements.

[0019] Flowcharts are used in the present specification to illustrate the operations performed by the system according to the embodiments of the present specification. It should be understood that the preceding or subsequent operations are not necessarily performed in sequence. On the contrary, each step can be processed in reverse order or simultaneously. At the same time, other operations can be added to these processes, or one or more steps of the operation can be removed from these processes.

[0020] The cable fault monitoring and early warning method and system provided by the embodiments of the present specification will be described in detail below with reference to the accompanying drawings.

[0021] Figure 1 is an exemplary flowchart of the cable fault monitoring and early warning method according to some embodiments of the present specification. Referring to Figure 1 , the cable fault monitoring and early warning method provided by the embodiments of the present application can include the following steps S110-S140: Step S110, real-time acquisition of cable operation state parameters and environmental parameters by a distributed sensor array. In some embodiments, step S110 can be performed by the monitoring data acquisition module 210 mentioned below.

[0022] Cable faults are related to a variety of factors, including cable factors and environmental factors, and can also be the result of long-term coupling of a variety of factors. For example, from the cable factors, high cable temperature can cause insulation material aging; three-phase current imbalance can cause local overheating of the cable, thereby accelerating the occurrence of faults; partial discharge signals can reflect internal defects (such as insulation damage, internal structure damage, etc.) or potential problems inside the cable; insulation resistance can reflect whether the insulation performance of the cable has deteriorated; ground current can reflect the state of the external grounding system of the cable; mechanical vibration amplitude can reflect changes in the environment around the cable or physical damage to the cable itself. For example, from the environmental factors, the temperature and humidity of the cable laying environment can also have some impact on the operation state of the cable (a high temperature and humidity environment can accelerate the aging process of the cable insulation material, and can also increase the probability of fault occurrence), and can therefore interfere with cable fault analysis and prediction due to their coupling effect.

[0023] Based on the above reasons, in some embodiments of the present application, real-time acquisition of cable operation state parameters and environmental parameters by a distributed sensor array can be performed, so that cable fault monitoring and early warning can be performed based on the cable operation state parameters and environmental parameters.

[0024] Specifically, in some embodiments of the present application, the cable operation state parameters can include cable temperature, three-phase current, partial discharge signal, insulation resistance, ground current, and mechanical vibration amplitude, and the environmental parameters can at least include temperature and humidity corresponding to the cable laying environment. In the embodiments of the present application, the environmental parameters can be used to assist in analyzing the reasons for changes in the cable operation state parameters, to avoid misjudgment due to a single factor, thereby improving the accuracy of fault monitoring and early warning.

[0025] It should be noted that in the present specification, the distributed sensor array refers to a network structure composed of multiple sensor nodes. In the embodiments of the present application, the multiple sensor nodes included in the distributed sensor array can be arranged at different cable positions (or different pipe gallery positions) to collect cable operating state parameters and environmental parameters corresponding to different cable positions (or different pipe gallery positions). In the embodiments of the present application, each sensor node can independently collect a specific type of data and transmit the data to a processing device (such as a cloud computing platform) for analysis and processing through a wireless or wired manner.

[0026] In some embodiments of the present application, the cable temperature can be collected by a temperature sensor arranged on the surface of the cable, and the three-phase current can be obtained by a current transformer installed at the cable joint. The partial discharge signal can be obtained by a high-frequency signal sensor (which can capture the tiny discharge phenomenon inside the cable insulation layer). The insulation resistance can be measured by a special insulation tester arranged for the cable, and the grounding current can be monitored by a grounding ring or a current clamp meter. The mechanical vibration amplitude can be collected by an acceleration sensor to detect abnormal vibration conditions (such as vibration caused by external force damage or state change of the cable) that can occur during the operation of the cable. For the temperature and humidity corresponding to the cable laying environment, a temperature and humidity sensor can be arranged in the pipe gallery to collect the temperature and humidity changes of the cable laying environment in real time.

[0027] It should be noted that the above-mentioned collection methods of the cable operating state parameters and the environmental parameters are only exemplary. In some other embodiments, any suitable alternative method can be used to collect the above-mentioned cable operating state parameters and environmental parameters. For example, in some embodiments, the cable temperature can also be measured non-contact by infrared thermal imaging technology, so as to avoid the interference caused by direct contact. In addition, the detection of the partial discharge signal can also be realized by an ultrasonic sensor, specifically, the weak acoustic signal generated by the discharge outside the cable insulation layer can be captured. For the collection of the three-phase current, in addition to the traditional current transformer, a Hall effect sensor can also be used to realize more accurate current monitoring. Such alternative methods are within the scope of protection of the present application.

[0028] In actual applications, different types of sensors can be flexibly selected or combined according to specific monitoring requirements, environmental conditions, cost budget and other factors, and the sensors can be arranged at different distribution densities to achieve the optimal monitoring effect.

[0029] In the embodiments of the present application, the sensors for collecting different parameters can be arranged at the same or different distribution positions and / or distribution densities (generally, the positions of different types of sensors are consistent at the head and tail ends). For example, in some embodiments, the distribution density of sensors can be increased at high-risk positions (such as cable joints, bending positions, or areas with large changes in environmental temperature) to improve the monitoring accuracy of the cable operation state in the area. In other low-risk or relatively stable environmental conditions, the distribution density of sensors can be appropriately reduced to effectively control the cost while ensuring the monitoring effect. In some embodiments, different types of sensors can have different distribution densities.

[0030] In step S120, the cable operation state parameters and the environmental parameters are preprocessed, and a standardized feature matrix is generated based on the preprocessed multi-source data. In some embodiments, step S120 can be performed by the preprocessing module 220 mentioned below.

[0031] In the embodiments of the present application, the preprocessing can include denoising and alignment. The denoising process refers to removing noise interference in the original data collected by each sensor node to improve the accuracy and reliability of the data. For example, in some embodiments, wavelet threshold denoising and Kalman filtering algorithm can be combined to perform denoising processing on the data collected by each sensor node, thereby eliminating random signal noise introduced by environmental factors or device limitations.

[0032] The alignment process refers to synchronizing the data collected by different sensor nodes in the time dimension to ensure that the multi-source data can reflect the cable operation state and environmental conditions at the same time. Specifically, in some embodiments, different types of sensors can collect data at different frequencies, which can cause misalignment of data on the time axis. To solve this problem, in some embodiments of the present application, interpolation algorithms and timestamp alignment techniques can be used to unify the data collected at different frequencies to the same time reference. For example, in some embodiments, for high-frequency sensor data, the data can be consistent with low-frequency data by downsampling; and for low-frequency data, interpolation methods can be used to supplement missing points, thereby achieving time alignment of multi-source data.

[0033] For example, in some embodiments, suppose one sensor collects data once every 10 seconds, while another sensor collects data once every 1 second. To achieve temporal alignment, the data collected by the high-frequency sensor can be downsampled to match the frequency of the data collected by the low-frequency sensor. Simultaneously, for the data collected by the low-frequency sensor, missing time points can be filled with estimated values ​​using linear interpolation or other interpolation methods (e.g., by averaging two consecutive time points), thus ensuring that all sensor data are analyzed on the same time reference.

[0034] It should be noted that, in the embodiments of this application, by aligning the above-mentioned cable operating status parameters and environmental parameters, not only can the consistency of timing be improved, but also an accurate and reliable data foundation can be provided for further feature extraction and cable fault analysis in the subsequent process.

[0035] Further technical details regarding the above preprocessing can be considered prior art and will not be discussed in detail in this specification.

[0036] Furthermore, after preprocessing the cable operating status parameters and environmental parameters as described above, a standardized feature matrix can be generated based on the preprocessed multi-source data. In this embodiment, the standardized feature matrix can be used to integrate information from all relevant parameters and transform it into a form suitable for subsequent analysis.

[0037] Specifically, in some embodiments of this application, preprocessed multi-source data corresponding to the current time and N times prior to the current time can be obtained; then, feature extraction is performed on the preprocessed multi-source data corresponding to each time to obtain the feature vector corresponding to each time; further, the feature vector can be standardized, and the standardized feature vectors corresponding to the current time and N times prior to the current time can be combined into a two-dimensional matrix in chronological order to obtain the standardized feature matrix.

[0038] For example, in some embodiments, the data collected at the current time and at N times prior to the current time can be represented as follows: T0: (a 1,0 a 2,0 a 3,0 , ...; b 1,0 b 2,0 , ...; ...; h 1,0 h 2,0 h 3,0 , ... T1: (a 1,1 a 2,1 a 3,1, …; b 1,1 , b 2,1 , …; …; null, null, null, …) T2: (a 1,2 , a 2,2 , a 3,2 , …; b 1,2 , b 2,2 , …; …; h 1,2 , h 2,2 , h 3,2 , …) … T N : (a 1,N , a 2,N , a 3,N , …; b 1,N , b 2,N , …; …; h 1,N , h 2,N , h 3,N , …) Wherein, T0 represents the current time, T1~T N represent the first~N time before the current time; a 1,0 , a 2,0 , a 3,0 , … represent the data collected by the first type of sensor at multiple sensor nodes at T0; similarly, b 1,1 , b 2,1 , … represent the data collected by the second type of sensor at multiple sensor nodes at T1; and so on. Null represents that the data collected by the sensor at a certain time (for example, T1) is null, and for this case, the mean of the collected data corresponding to the two nearest times (for example, T0 and T2) can be calculated to fill it (for example, the mean of h 1,0 and h 1,2 is calculated as the filling value of h 1,1 , the mean of h 2,0 and h 2,2 is calculated as the filling value of h 2,1 ……), so as to ensure that all sensor data is under the same time reference, facilitating subsequent analysis.

[0039] In some embodiments of the present application, interpolation can also be performed in the quantity dimension for different types of sensors with different distribution densities. For example, when the number of a certain type of sensor is small, virtual sensor data points can be generated in the quantity dimension through an interpolation algorithm (e.g., linear interpolation). These virtual data points can reflect the spatial distribution characteristics of the type of sensor, thereby making up for the data sparsity problem caused by insufficient number of sensors, so as to facilitate the calculation of the fluctuation law between different types of data in the subsequent process. It should be noted that in the embodiments of the present application, when interpolation is performed in the quantity dimension, in order to ensure the accuracy of the interpolation, the interpolation can be performed in combination with the sensor positions corresponding to the existing data. For example, when interpolation is performed, the spatial distribution relationship can be determined by using the geographic coordinate information of the known sensor nodes, and then the values of the virtual data points can be calculated by linear sampling in combination with the data collected by the known sensor nodes, so as to more accurately estimate the values of the virtual data points and their reasonable positions in space.

[0040] In some embodiments of the present application, the data collected at the current time and the N times before the current time obtained through the above preprocessing can be respectively represented as (wherein m represents the total number of each type of data at the same time after alignment processing): T0: (a 1,0 , a 2,0 , …, a m,0 ; b 1,0 , b 2,0 , … b m,0 ; …; h 1,0 , h 2,0 , …, h m,0 ) T1: (a 1,1 , a 2,1 , …, a m,1 ; b 1,1 , b 2,1 , … b m,1 ; …; h 1,1 , h 2,1 , …, h m,1 ) T2: (a 1,2 , a 2,2 , …, a m,2 ; b 1,2 , b 2,2 , … b m,2 ; …; h 1,2 , h 2,2 , …, h m,2 ) … T N : (a 1,N , a2,N , ..., a m,N b 1,N b 2,N , ...b m,N ;…;h 1,N h 2,N , ..., h m,N ) In this embodiment, feature extraction can be performed on the preprocessed multi-source data corresponding to each time moment to obtain a feature vector corresponding to each time moment. The feature vector includes dynamic change features and statistical features calculated based on multiple cable operating status parameters and environmental parameters collected by the distributed sensor array.

[0041] Specifically, in some embodiments of this application, the target time (which may refer to T0 to T) can be targeted at a target time. N At any given moment, based on multiple cable operating status parameters and environmental parameters collected by the distributed sensor array, the rate of change of the data collected at each collection point for each parameter relative to adjacent collection points is calculated to obtain the dynamic change characteristics of the cable operating status parameters and environmental parameters at different locations. At the same time, the mean, variance, peak value, and valley value of each parameter can also be calculated to obtain the statistical characteristics corresponding to the target moment. Finally, the dynamic change characteristics and statistical characteristics can be integrated to obtain the feature vector corresponding to the target moment.

[0042] For example, for time T0, its corresponding dynamic change characteristics can be expressed as: Furthermore, in this embodiment, the mean, variance, peak value, and trough value of each parameter can be calculated to obtain the statistical characteristics corresponding to the target time. For example, for time T0, the corresponding statistical characteristics can be expressed as: in, Indicates the first Type of parameters, Indicates the first The mean of the parameters at time T0, Indicates the first Methods for setting parameters at time T0, Indicates the first The peak value of each parameter at time T0. Indicates the first The valley value of the parameter at time T0.

[0043] Further, the dynamic change features and the statistical features can be integrated to obtain the feature vector corresponding to the target moment. For example, for the moment T0, the corresponding feature vector can be represented as: It should be noted that the cable operation state parameters and the environmental parameters collected by the above-mentioned distributed sensor array may have certain normal differences at different positions due to the particularity of the environment (for example, there are electromagnetic interference, sharp temperature change or abnormal humidity, etc.). In the embodiments of the present application, the change rate of the data collected by each collection point in each parameter relative to the adjacent collection point is calculated as the corresponding dynamic change feature, and the dynamic change feature is integrated with the statistical features such as the mean value, variance, peak value and valley value of each parameter to obtain the feature vector corresponding to the target moment. The dynamic change feature and the statistical feature can better capture the dynamic change of the cable operation state parameters and the environmental parameters at different positions, so as to more accurately express the differences (including normal differences and abnormal differences) of the cable operation state parameters and the environmental parameters at different positions, so as to more accurately identify the cable fault in the subsequent process. Specifically, in the embodiments of the present application, the presentation rule of the differences of the above-mentioned cable operation state parameters and environmental parameters at different positions can be learned by the fault prediction model to be described later, and then the accurate identification of the cable fault (including existing fault and potential fault) can be realized, and the false positive rate of the cable fault monitoring can be reduced to a certain extent.

[0044] In some embodiments of the present application, the fluctuation correlation degree between the dynamic change features corresponding to each parameter and the dynamic change features corresponding to other parameters can also be calculated; then the dynamic change features, the fluctuation correlation degrees and the statistical features are spliced in a predetermined order to obtain the feature vector corresponding to the target moment.

[0045] ​​​​​​​​Specifically, in some embodiments of the present application, considering that there can be a certain coupling effect between the cable operation state parameters and the environmental parameters, and the cable fault can be jointly caused by the coupling effect of multiple different parameters (for example, the coupling effect can be embodied in that there is a certain correlation between the abnormal fluctuation of some parameters and the change of other parameters). Based on this, in some embodiments of the present application, in order to more sensitively identify the abnormality of the cable operation state parameters and the environmental parameters and predict the potential cable fault, the fluctuation correlation between the dynamic change feature corresponding to each parameter and the dynamic change feature corresponding to other parameters can be calculated, and then the dynamic change feature, the fluctuation correlation and the statistical feature are spliced in a preset order to obtain the feature vector corresponding to the target moment.

[0046] Exemplarily, in some embodiments of the present application, the fluctuation correlation can be calculated by the following formula: wherein, Corr(i, j) represents the fluctuation correlation between the i-th parameter and the j-th parameter, i∈(1, 2, 3, …, h), j∈(1, 2, 3, …, h), i≠j; represents a normalization operation; represents the k-th data in the dynamic change feature corresponding to the i-th parameter, represents the k-th data in the dynamic change feature corresponding to the j-th parameter.

[0047] In the embodiments of the present application, the fluctuation correlation of each parameter with respect to other h-1 parameters can be integrated with the aforementioned dynamic change feature and statistical feature to obtain the feature vector corresponding to the target moment.

[0048] In the embodiments of the present application, after obtaining the feature vector corresponding to each moment by the above-mentioned manner, the feature vector can be standardized, and the standardized feature vectors corresponding to the current moment and the N moments before the current moment are combined in time sequence into a two-dimensional matrix, so as to obtain the standardized feature matrix.

[0049] In the embodiments of the present application, the aforementioned standardization processing can be understood as adjusting the order position of each parameter in the aforementioned feature vector according to the standard format, so as to obtain the standardized feature vector corresponding to each moment. Further, the standardized feature vectors corresponding to the current moment (for example, the T0 moment) and the N moments before the current moment can be combined in time sequence into a two-dimensional matrix to obtain the standardized feature matrix. In some embodiments, each row in the two-dimensional matrix can be used to represent a standardized feature vector corresponding to a moment.

[0050] In step S130, the standardized feature matrix is input into the pre-trained fault prediction model for processing to obtain the fault type and the fault risk index corresponding to the current operation state of the cable. In some embodiments, step S130 can be performed by the fault prediction module 230 mentioned below.

[0051] Figure 2 is an exemplary structural schematic diagram of a fault prediction model according to some embodiments of the present specification. Refer to Figure 2 In the embodiments of the present application, the fault prediction model can include a deep feature extraction network and a fault mapping network, wherein the deep feature extraction network is configured to perform deep feature mining and pattern recognition on the standardized feature matrix to obtain a high-dimensional feature representation corresponding to the current operation state of the cable; and the fault mapping network is configured to map the high-dimensional feature representation to perform real-time evaluation on the current operation state of the cable to obtain the fault type and the fault risk index. For example only, in the embodiments of the present application, the fault type can include, but is not limited to, short-circuit fault, open-circuit fault, insulation aging fault, overheat fault, damage fault, and ground fault.

[0052] In some embodiments of the present application, the fault prediction model can be obtained based on the following manner: First, training sample data is obtained. In the embodiments of the present application, the training sample data includes cable operation state parameters, environmental parameters corresponding to a sample time period, and at least one fault type (including known fault type and potential fault type) and fault risk index corresponding thereto. In the embodiments of the present application, the fault risk index can be between 0 and 1, and the closer the fault risk index is to 1, the higher the possibility of cable failure. Then, the training sample data can be preprocessed according to the preprocessing method described above, and the standardized feature matrix corresponding to the training sample data is generated as input, and at least one fault type and fault risk index corresponding to the training sample data are taken as labels. Further, an initial fault prediction model can be constructed using a deep learning framework. In the embodiments of the present application, the initial fault prediction model includes a deep feature extraction network and a fault mapping network. More details about the deep feature extraction network and the fault mapping network can be referred to later. Finally, the initial fault prediction model can be iteratively trained by the training sample data, and the network parameters of the deep feature extraction network and the fault mapping network can be optimized by a backpropagation algorithm, until the prediction accuracy of the initial fault prediction model meets a preset condition, to obtain a trained fault prediction model.

[0053] Figure 3is an exemplary structural schematic diagram of a deep feature extraction network according to some embodiments of the present specification. Referring to Figure 3 In some embodiments of the present application, the deep feature extraction network can include a plurality of feature extraction layers (which can be a plurality of convolutional layers in nature), wherein each feature extraction layer can perform feature extraction at different levels on the standardized feature matrix, thereby obtaining feature representations at different depths (i.e. Figure 3 w1, w2, w3, …, wx) shown.

[0054] Continuing to refer to Figure 3 In some embodiments of the present application, the deep feature extraction network can further include a feature fusion layer, which can be used to fuse the plurality of feature representations at different depths obtained by the plurality of feature extraction layers, thereby obtaining a high-dimensional feature representation. It should be noted that in embodiments of the present application, the feature fusion layer can learn the weights corresponding to the feature representations obtained by each feature extraction layer during the training process. It should also be noted that in embodiments of the present application, by performing feature extraction at multiple depths on the aforementioned standardized feature matrix and then fusing the feature representations at multiple depths, the high-dimensional feature representation is obtained, which can mine more abstract feature representations, thereby capturing more complex feature patterns and feature semantic information, which helps to more accurately convert the information reflected by the high-dimensional feature representation into the corresponding fault type and fault risk index through the feature mapping network in the subsequent process.

[0055] In embodiments of the present application, the aforementioned feature mapping network can be a multi-layer perceptron, which can perform nonlinear mapping on the high-dimensional feature representation output by the aforementioned deep feature extraction network to convert the high-dimensional feature representation into a specific fault type and fault risk index. Similarly, during the training process, the feature mapping network can also continuously adjust its internal parameters to ensure that the error between the output result and the actual label is minimized.

[0056] In some embodiments of the present application, an attention mechanism can also be introduced into the fault mapping network, so that the feature mapping network pays more attention to the feature dimensions that play a key role in cable fault prediction in the aforementioned high-dimensional feature representation during processing.

[0057] In embodiments of the present application, a large number of training sample data obtained in the aforementioned manner can be used to iteratively train the initial fault prediction model, and the network parameters of the deep feature extraction network and the fault mapping network can be optimized through a backpropagation algorithm until the prediction accuracy of the initial fault prediction model meets a preset condition (e.g., the prediction result accuracy is greater than 99%), thereby obtaining a trained fault prediction model.

[0058] More details about training the above failure prediction model and optimizing the network parameters of the deep feature extraction network and the failure mapping network can be considered as prior art, which will not be discussed in detail in this specification.

[0059] At step S140, an early warning mechanism is triggered according to the failure type and the failure risk index, and the corresponding failure location information is output. In some embodiments, step S140 can be performed by early warning module 240 mentioned below.

[0060] In some embodiments of the present application, the failure type and the failure risk index can be mapped by a pre-configured early warning mechanism mapping table to obtain the early warning mechanism corresponding to the current failure type and failure risk index.

[0061] Specifically, in some embodiments of the present application, the early warning mechanism mapping table can include a plurality of failure types, their corresponding failure risk index ranges and corresponding early warning measures. For example, for a high-risk failure type, a higher level of early warning measure can be configured, such as triggering an emergency alarm or notifying maintenance personnel to handle immediately. For a low-risk failure type, a lower level of early warning measure can be taken, such as recording failure information or sending a notification to remind relevant personnel to observe later. In some embodiments, for a high-risk failure type, but the corresponding failure risk index is low (e.g. lower than 50%), a lower level of early warning measure can be taken.

[0062] In some embodiments of the present application, the corresponding failure location information can be output at the same time as the early warning, so as to help maintenance personnel quickly locate the failure point and improve the efficiency of failure handling. In some embodiments of the present application, the failure location information can be determined based on the following manner: First, the pre-set spatial coordinates corresponding to each collection point in the distributed sensor array are obtained; then, the corresponding key sensitive parameters are determined according to the failure type output by the failure prediction model; further, the abnormal deviation degrees of the key sensitive parameters at each collection point can be calculated; finally, the spatial coordinates corresponding to the collection point with the largest abnormal deviation degree can be obtained to obtain the failure location information.

[0063] In embodiments of the present application, the key sensitive parameter refers to a cable operating state parameter or an environmental parameter that has a significant influence or correlation with the failure type. For example, for an overheating failure, the corresponding key sensitive parameters can include cable temperature, etc.; for a damage failure, the corresponding key sensitive parameters can include partial discharge signal, etc.; for an insulation aging failure, the corresponding key sensitive parameters can include insulation resistance, etc.; for a short circuit failure, the corresponding key sensitive parameters can include three-phase current, etc.

[0064] In the embodiments of the present application, the abnormal deviation degree of the key sensitive parameter at each collection point can be calculated. For example, in some embodiments, the abnormal deviation degree can be the ratio of the absolute value of the difference between the parameter value of the collection point and the upper limit or lower limit of the preset normal threshold range. For example, in some embodiments, the absolute value of the difference between the parameter value of the collection point and the upper limit or lower limit of the preset normal threshold range that is relatively closer can be calculated, and then the ratio of the absolute value of the difference to the upper limit or lower limit of the preset normal threshold range that is relatively closer is taken as the abnormal deviation degree corresponding to the collection point.

[0065] Further, after the abnormal deviation degree corresponding to each collection point is calculated by the above-mentioned method for the pre-processed cable operation state parameters and environmental parameters, the fault location information can be obtained according to the spatial coordinates corresponding to the collection point with the maximum abnormal deviation degree.

[0066] It should be noted that in the embodiments of the present application, the fault type determined by the above-mentioned process can be taken as the guidance information, so that the fault location can be more accurately and conveniently located by determining the corresponding key sensitive parameter according to the fault type, and then calculating the abnormal deviation degree corresponding to each collection point according to the key sensitive parameter, and determining the fault location information according to the abnormal deviation degree.

[0067] It should also be noted that in some embodiments of the present application, the collection point with the maximum abnormal deviation degree mentioned above can be a virtual data point obtained by the above-mentioned interpolation processing, and the parameter value corresponding to the virtual data point can deviate from the actual value, that is, the fault location information obtained by the above-mentioned method can also have a certain error. In order to solve the problem, in some embodiments of the present application, the accuracy and reliability of the fault location information can be further improved by regarding the spatial coordinates corresponding to the collection point with the maximum abnormal deviation degree as the initial fault location, and then correcting the initial fault location in combination with the abnormal deviation degree distribution of the adjacent collection points to obtain the fault location information.

[0068] Specifically, in some embodiments of the present application, the correction process can include the following steps: First, the initial fault location is taken as the center; then the spatial distance between the adjacent collection points and the initial fault location and the abnormal deviation degree corresponding to the adjacent collection points are calculated by weighted average; further, based on the calculation result of the weighted average calculation, an equivalent fault point is determined on the left and right sides of the initial fault location respectively; finally, based on the spatial coordinates corresponding to the equivalent fault points, the fault location information is obtained.

[0069] For example, assuming that the initial fault location obtained through the above manner is O, and the adjacent collection points on the left and right sides of O are O1 and O2, wherein the abnormal deviation degree corresponding to the O1 point is y1, the spatial distance between the O1 point and the O point is d1, the abnormal deviation degree corresponding to the O2 point is y2, and the spatial distance between the O2 point and the O point is d2, the calculation result y can be obtained through weighted average calculation according to the spatial distances (d1 and d2) between the adjacent collection points and the initial fault location and the abnormal deviation degrees (y1 and y2) corresponding to the adjacent collection points. .

[0070] In the embodiments of the present application, y The calculation process can be represented as follows: Further, an equivalent fault point corresponding to an abnormal deviation degree of y may be determined on the left and right sides of the initial fault location O, respectively, and then a position range is determined according to the spatial coordinates corresponding to the two equivalent fault points, and the position range is taken as the fault location information. It should be noted that in the embodiments of the present application, the initial fault location is corrected through the above manner, and a more accurate and reliable fault location prediction range can be obtained.

[0071] Figure 4 is a module schematic diagram of a cable fault monitoring and early warning system according to some embodiments of the present application. In some embodiments, Figure 4 The cable fault monitoring and early warning system 200 shown in some embodiments of the present application can be implemented in the form of software and / or hardware, for example, can be configured in the form of software and / or hardware to a processing device and / or terminal device, to process the cable operation state parameters and environmental parameters collected in real time by the distributed sensor array, and to evaluate the current operation state of the cable in real time based on the cable operation state parameters and environmental parameters, to obtain the corresponding fault type and fault risk index, and then to trigger the early warning mechanism according to the fault type and fault risk index, and to output the corresponding fault location information.

[0072] Referring to Figure 4 , the cable fault monitoring and early warning system 200 provided by the embodiments of the present application can include a monitoring data acquisition module 210, a preprocessing module 220, a fault prediction module 230, and an early warning module 240. Among them: The monitoring data acquisition module 210 can be used to collect cable operation state parameters and environmental parameters in real time through a distributed sensor array.

[0073] The preprocessing module 220 can be used to preprocess the cable operation state parameters and environmental parameters, and generate a standardized feature matrix based on the preprocessed multi-source data.

[0074] The failure prediction module 230 can be configured to input the standardized feature matrix into a pre-trained failure prediction model to obtain a failure type and a failure risk index corresponding to the current operation state of the cable.

[0075] The early warning module 240 can be configured to trigger an early warning mechanism according to the failure type and the failure risk index, and output corresponding failure location information.

[0076] Further details regarding the above-mentioned modules can be found elsewhere in this specification (e.g. Figures 1-3 and related descriptions), which will not be repeated here.

[0077] It should be understood that, Figure 4 The cable failure monitoring and early warning system 200 and its modules shown can be implemented in various ways. For example, in some embodiments, the system and its modules can be implemented in hardware, software, or a combination of software and hardware. The hardware portion can be implemented with dedicated logic; the software portion can be stored in a memory and executed by a suitable instruction execution system, such as a microprocessor or a specially designed hardware. Those skilled in the art can understand that the above-mentioned methods and systems can be implemented using computer executable instructions and / or included in processor control code, such as provided on a carrier medium, such as a magnetic disk, CD or DVD-ROM, a programmable memory, such as a read-only memory (firmware), or a data carrier, such as an optical or electronic signal carrier. The system and its modules of this specification can not only be implemented in hardware circuits, such as very large scale integrated circuits or gate arrays, semiconductors, such as logic chips, transistors, or programmable hardware devices, such as field programmable gate arrays, programmable logic devices, etc., but also in software, for example, executed by various types of processors, and also by a combination of the above-mentioned hardware circuits and software (e.g., firmware).

[0078] It should be noted that the above description of the cable failure monitoring and early warning system 200 is provided for illustrative purposes only and is not intended to limit the scope of this specification. It can be understood that those skilled in the art can make any combination of the modules or connect the sub-systems to other modules according to the description of this specification without departing from the principles of this specification. For example, Figure 4 The monitoring data acquisition module 210, the preprocessing module 220, the failure prediction module 230 and the early warning module 240 described in the above can be different modules in a system, or one module can implement the functions of two or more modules described above. Variations such as this are within the scope of protection of this specification.

[0079] In summary, the beneficial effects that can be brought about by the embodiments of this specification include but are not limited to: (1) In the cable fault monitoring and early warning method and system provided by some embodiments of the present specification, the real-time collected cable operation state parameters and environmental parameters of the distributed sensor array are preprocessed to generate a standardized feature matrix, and then the standardized feature matrix is input into a pre-trained fault prediction model for processing to obtain the corresponding fault type and fault risk index. The current operation state of the cable can be accurately and reliably evaluated in real time based on the deep feature extraction network and fault mapping network included in the fault prediction model, thereby improving the intelligence, accuracy and reliability of cable fault monitoring to a certain extent.

[0080] (2) In the cable fault monitoring and early warning method and system provided by some embodiments of the present specification, the change rate of the data collected by each collection point in each parameter with respect to the adjacent collection point is calculated as the corresponding dynamic change feature, and the dynamic change feature is integrated with the statistical features such as mean value, variance, peak value and valley value of each parameter to obtain the feature vector corresponding to the target time. The dynamic change of the cable operation state parameters and environmental parameters at different positions can be better captured through the dynamic change feature and the statistical feature, thereby more detailedly expressing the differences of the cable operation state parameters and environmental parameters at different positions to more accurately identify the cable fault in the subsequent process.

[0081] (3) In the cable fault monitoring and early warning method and system provided by some embodiments of the present specification, the corresponding key sensitive parameters are determined according to the fault type, then the abnormal deviation degree of each collection point is calculated according to the key sensitive parameters, and the fault location information is determined according to the abnormal deviation degree. The fault type determined in the foregoing process can be used as guide information to more accurately and conveniently locate the fault position.

[0082] It should be noted that different embodiments may have different beneficial effects, and in different embodiments, the beneficial effects that can be produced can be any one or a combination of the above, or any other beneficial effects that can be obtained.

[0083] The above has described the basic concepts, and it is obvious that the above detailed disclosure is only used as an example and does not limit the present specification. Although it is not explicitly stated here, those skilled in the art can make various modifications, improvements and corrections to the present specification. Such modifications, improvements and corrections are suggested in the present specification, so such modifications, improvements and corrections still belong to the spirit and scope of the exemplary embodiments of the present specification.

[0084] Furthermore, this specification uses specific terms to describe embodiments thereof. For example, "an embodiment," "one embodiment," and / or "some embodiments" refer to a particular feature, structure, or characteristic associated with at least one embodiment of this specification. Therefore, it should be emphasized and noted that references to "an embodiment," "one embodiment," or "an alternative embodiment" in different locations throughout this specification do not necessarily refer to the same embodiment. Moreover, certain features, structures, or characteristics in one or more embodiments of this specification can be appropriately combined.

[0085] Furthermore, those skilled in the art will understand that various aspects of this specification can be described and illustrated in several patentable ways or situations, including any new and useful combination of processes, machines, products, or substances, or any new and useful improvements thereof. Accordingly, various aspects of this specification can be implemented entirely by hardware, entirely by software (including firmware, resident software, microcode, etc.), or by a combination of hardware and software. All of the above hardware or software may be referred to as a “data block,” “module,” “engine,” “unit,” “component,” or “system.” Furthermore, various aspects of this specification may be represented as a computer product located on one or more computer-readable media, including computer-readable program code.

[0086] Computer storage media may contain a propagated data signal containing computer program code, for example, on baseband or as part of a carrier wave. This propagated signal may take various forms, including electromagnetic, optical, and suitable combinations thereof. Computer storage media can be any computer-readable medium other than a computer-readable storage medium, which can be connected to an instruction execution system, apparatus, or device to enable communication, propagation, or transmission of a program for use. The program code located on the computer storage medium can be propagated through any suitable medium, including radio, cable, fiber optic cable, RF, or similar media, or any combination of the above media.

[0087] Computer program code for carrying out operations of the aspects described in this specification can be written in any combination of one or more programming languages, including an object-oriented programming language such as Java, Scala, Smalltalk, Eiffel, JADE, Emerald, C++, C#, VB.NET, Python, and others, conventional procedural programming languages, such as the C programming language, Visual Basic, Fortran 2003, Perl, COBOL 2002, PHP, ABAP, dynamic programming languages such as Python, Ruby and Groovy, or other programming languages. The program code can execute entirely on the user's computer, or it can be executed as a standalone software package, or it can execute partly on the user's computer and partly on a remote computer or server, or it can execute entirely on a remote computer or server. In the latter scenario, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection can be made to an external computer (for example, through the Internet using an Internet Service Provider). In some embodiments, electronic program code can be downloaded via software

[0088] In addition, the order of execution or performance of the operations of the aspects described in this specification illustrated and described herein is not essential, unless otherwise specified. That is, it is contemplated that the operations of the aspects can be performed in any order, unless otherwise specified, and that the examples of the aspects described in this specification are not limited to any particular ordering.

[0089] Similarly, it is contemplated that, where different features of the aspects described in this specification are described in the context of separate aspects, such separate aspects can also be combined, unless otherwise specified. Unless otherwise specified, it is also contemplated that the aspects described in this specification can be implemented in hardware, software, firmware, middleware, microcode, or any combination thereof.

[0090] Finally, it should be noted that the language used in the specification has been principally selected for readability and instructional purposes and can not have been selected to delineate or circumscribe the inventive subject matter. Accordingly, the disclosure of the aspects is intended to be illustrative, but not limiting, of the aspects described in the following claims.

Claims

1. A method for monitoring and early warning of cable faults, characterized in that, include: The cable operating status parameters and environmental parameters are collected in real time by a distributed sensor array. The cable operating status parameters include cable temperature, three-phase current, partial discharge signal, insulation resistance, grounding current and mechanical vibration amplitude. The environmental parameters include at least the temperature and humidity of the cable laying environment. The cable operating status parameters and environmental parameters are preprocessed, and a standardized feature matrix is ​​generated based on the preprocessed multi-source data. The preprocessing includes noise reduction and alignment. The standardized feature matrix is ​​input into a pre-trained fault prediction model for processing to obtain the fault type and fault risk index corresponding to the current operating state of the cable. The fault prediction model includes a deep feature extraction network and a fault mapping network. The deep feature extraction network performs deep feature mining and pattern recognition on the standardized feature matrix to obtain a high-dimensional feature representation corresponding to the current operating state of the cable. The fault mapping network maps the high-dimensional feature representation to perform real-time evaluation of the current operating state of the cable, obtaining the fault type and fault risk index. The warning mechanism is triggered based on the fault type and fault risk index, and the corresponding fault location information is output.

2. The cable fault monitoring and early warning method as described in claim 1, characterized in that, The preprocessing of the cable's operating status parameters and environmental parameters includes: Wavelet threshold denoising and Kalman filtering algorithms are used to remove signal noise from the cable's operating status parameters and environmental parameters; Then, the cable operating status parameters and environmental parameters are time-series aligned using interpolation algorithms and timestamp alignment techniques to obtain preprocessed multi-source data.

3. The cable fault monitoring and early warning method as described in claim 2, characterized in that, The generation of a standardized feature matrix based on preprocessed multi-source data includes: Obtain the preprocessed multi-source data corresponding to the current time and N times preceding the current time; Feature extraction is performed on the preprocessed multi-source data corresponding to each time moment to obtain the feature vector corresponding to each time moment. The feature vector includes dynamic change features and statistical features calculated based on multiple cable operation status parameters and environmental parameters collected by the distributed sensor array. The feature vectors are standardized, and the standardized feature vectors corresponding to the current time and the N times before the current time are combined into a two-dimensional matrix in chronological order to obtain the standardized feature matrix.

4. The cable fault monitoring and early warning method as described in claim 3, characterized in that, The step of extracting features from the preprocessed multi-source data at each time step to obtain the feature vector at each time step includes: For multiple cable operating status parameters and environmental parameters collected by the distributed sensor array at the target time, the rate of change of the data collected at each collection point for each parameter relative to the adjacent collection points is calculated to obtain the dynamic change characteristics of the cable operating status parameters and environmental parameters at different locations. Calculate the mean, variance, peak value, and trough value of each parameter to obtain the statistical characteristics corresponding to the target time. By integrating the dynamic change features and statistical features, the feature vector corresponding to the target time is obtained.

5. The cable fault monitoring and early warning method as described in claim 4, characterized in that, The process of integrating the dynamic change features and statistical features to obtain the feature vector corresponding to the target time includes: Calculate the fluctuation correlation between the dynamic change characteristics corresponding to each parameter and the dynamic change characteristics corresponding to other parameters; The dynamic change features, fluctuation correlation degree, and statistical features are concatenated in a preset order to obtain the feature vector corresponding to the target time.

6. The cable fault monitoring and early warning method as described in claim 1, characterized in that, The fault prediction model is obtained based on the following method: Acquire training sample data, wherein the training sample data includes cable operating status parameters, environmental parameters, and at least one fault type and fault risk index corresponding to the sample time period; The training sample data is preprocessed to generate a standardized feature matrix corresponding to the training sample data as input, and the fault type and fault risk index corresponding to the training sample data are used as labels. An initial fault prediction model is constructed using a deep learning framework, wherein the initial fault prediction model includes a deep feature extraction network and a fault mapping network; The initial fault prediction model is iteratively trained using the training sample data, and the network parameters of the deep feature extraction network and the fault mapping network are optimized using the backpropagation algorithm until the prediction accuracy of the initial fault prediction model meets the preset conditions, thus obtaining the trained fault prediction model.

7. The cable fault monitoring and early warning method as described in claim 1, characterized in that, The mechanism for triggering an early warning based on the fault type and fault risk index includes: The fault type and fault risk index are mapped by a pre-configured early warning mechanism mapping table to obtain the early warning mechanism corresponding to the current fault type and fault risk index.

8. The cable fault monitoring and early warning method as described in claim 1, characterized in that, The fault location information is determined based on the following method: Obtain the preset spatial coordinates corresponding to each acquisition point in the distributed sensor array; Determine the corresponding key sensitive parameters based on the fault type; Calculate the abnormal deviation of the key sensitive parameter at each collection point, wherein the abnormal deviation is the ratio of the absolute value of the difference between the parameter value at that collection point and the preset normal threshold range to the upper or lower limit of the threshold range; The fault location information is obtained based on the spatial coordinates corresponding to the acquisition point with the largest abnormal deviation.

9. The cable fault monitoring and early warning method as described in claim 8, characterized in that, The method further includes: The spatial coordinates corresponding to the acquisition point with the largest abnormal deviation are regarded as the initial fault location, and the initial fault location is corrected by combining the abnormal deviation distribution of adjacent acquisition points to obtain the fault location information. The correction process includes: Taking the initial fault location as the center, a weighted average is calculated based on the spatial distance between adjacent acquisition points and the initial fault location, as well as the abnormal deviation degree corresponding to the adjacent acquisition points. Based on the calculation results of the weighted average, an equivalent fault point is determined on the left and right sides of the initial fault location, respectively. The fault location information is obtained based on the spatial coordinates corresponding to the equivalent fault point.

10. A cable fault monitoring and early warning system, characterized in that, include: The monitoring data acquisition module is used to collect cable operating status parameters and environmental parameters in real time through a distributed sensor array. The cable operating status parameters include cable temperature, three-phase current, partial discharge signal, insulation resistance, grounding current and mechanical vibration amplitude. The environmental parameters include at least the temperature and humidity corresponding to the cable laying environment. The preprocessing module is used to preprocess the cable operating status parameters and environmental parameters, and generate a standardized feature matrix based on the preprocessed multi-source data. The preprocessing includes noise reduction and alignment. The fault prediction module is used to input the standardized feature matrix into a pre-trained fault prediction model for processing, so as to obtain the fault type and fault risk index corresponding to the current operating state of the cable. The fault prediction model includes a deep feature extraction network and a fault mapping network. The deep feature extraction network is used to perform deep feature mining and pattern recognition on the standardized feature matrix to obtain a high-dimensional feature representation corresponding to the current operating state of the cable. The fault mapping network is used to map the high-dimensional feature representation to perform real-time evaluation of the current operating state of the cable, thereby obtaining the fault type and fault risk index. The early warning module is used to trigger an early warning mechanism based on the fault type and fault risk index, and output the corresponding fault location information.

Images