An intelligent low-voltage cable insulation monitoring method and system

Abstract

The present invention discloses an intelligent low-voltage cable insulation monitoring method and system, which includes data acquisition, preprocessing, establishing a low-voltage cable insulation evaluation model and low-voltage cable insulation monitoring. The present invention belongs to the field of cable insulation monitoring, and specifically refers to an intelligent low-voltage cable insulation monitoring method and system. This scheme enhances local discharge characteristics based on pulse injection branches; avoids excessive bias through a dynamic soft weight allocation mechanism; focuses resource allocation through a fault detection network; strengthens the contribution of minority cable samples in the overall gradient based on the construction of an insulation stability loss function; ignores early windows with insufficient information by introducing a window weight factor, focuses on critical sections, and constructs a window weight likelihood loss to enhance fault critical section identification; introduces a fault severity regression branch to promote the convergence of fault critical sections; and constructs a fault response adjustment based on a compensation function to reduce false positive and missed detection rates, thereby improving the cable insulation monitoring effect.

Description

Technical Field

[0001] The present invention relates to the technical field of cable insulation monitoring, and in particular to an intelligent low-voltage cable insulation monitoring method and system. Background Art

[0002] Low-voltage cable insulation monitoring methods rely on deploying sensors for partial discharge, leakage current, temperature, humidity, and other indicators to collect key diagnostic data in real time or periodically during cable operation or power outage maintenance, and employ data analysis techniques to assess insulation health. However, typical low-voltage cable insulation monitoring methods suffer from the inability to detect early-stage weak discharges, resulting in delayed early warnings of potential insulation failures and a scarcity of samples indicating insulation damage or degradation, leading to low monitoring reliability. Furthermore, these methods treat all monitoring windows equally, so early weak discharge and leakage current signals are often overwhelmed by noise, resulting in poor monitoring effectiveness. Summary of the Invention

[0003] In view of the above situation, in order to overcome the defects of the prior art, the present invention provides an intelligent low-voltage cable insulation monitoring method and system. In view of the problems that the general low-voltage cable insulation monitoring method is unable to detect early weak discharges, delays the early warning of potential insulation faults, and has a scarcity of samples of insulation damage or degradation, which leads to low monitoring reliability, this solution improves the recognition ability of rare fault modes by constructing four enhanced branches: disturbance injection, time distortion, sample synthesis and pulse injection; enhances local discharge characteristics based on the pulse injection branch; avoids excessive bias through a dynamic soft weight allocation mechanism; focuses resource allocation through a fault detection network; and constructs an insulation stability loss function based on the four enhanced branches. Strengthen the contribution of minority cable samples in the overall gradient; thereby improving the reliability of subsequent insulation monitoring; in view of the problem that general low-voltage cable insulation monitoring methods treat all monitoring windows equally, early weak discharge and leakage current signals are often submerged by noise, resulting in poor monitoring effect, this scheme introduces a window weight factor, ignores early windows with insufficient information, focuses on critical sections, and constructs a window weight likelihood loss to enhance the identification of critical fault sections; introduces a fault severity regression branch, incorporates continuous degradation signals into the penalty, and promotes the convergence of critical fault sections; and constructs a fault response adjustment based on the compensation function to reduce false positive and missed detection rates; thereby improving the cable insulation monitoring effect.

[0004] The technical solution adopted by the present invention is as follows: The present invention provides an intelligent low-voltage cable insulation monitoring method, which includes the following steps:

[0005] Step S1: data collection;

[0006] Step S2: preprocessing;

[0007] Step S3: establishing a low-voltage cable insulation evaluation model;

[0008] Step S4: Low voltage cable insulation monitoring.

[0009] Furthermore, in step S1, the data collection is to collect historical low-voltage cable monitoring data; mark the cable status as a data label; standardize the historical low-voltage cable monitoring data, and use sliding window segmentation to construct an initial low-voltage cable monitoring data set.

[0010] Furthermore, in step S2, the preprocessing is to construct four enhancement branches for the initial low-voltage cable monitoring dataset, each branch corresponding to an enhancement strategy, and automatically assign learning weights using a dynamic weighting and insulation stability loss mechanism, focusing on strengthening the enhancement effect on a minority of cable fault samples; thereby constructing a low-voltage cable monitoring dataset; specifically, including:

[0011] Step S21: Enhanced branch construction unit; the enhanced branch includes disturbance injection branch, time warping branch, sample synthesis branch and pulse injection branch; each enhanced branch is forward trained once to calculate the loss value of the enhanced cable sample on the main classification network , do softmax on the loss and get the learning weight , expressed as: ; Set up another fault detection network output , the training label is a reverse exponential mapping, expressed as: ;in, is the loss value of the enhanced cable sample on the main classification network after one forward training of the f-th enhanced branch; n and f are both enhanced branch indices; is the fault detection network training label; the main classification loss , enhanced branch weighted loss and fault detection network loss fusion, initial enhanced loss Expressed as: ; ;in, 、 and Is to enhance the loss weight; Use cross entropy loss; is the loss of the cable sample after the enhanced branch reinforcement; It is the main classification network map; is the cable sample enhanced by the enhancement branch; y is the true label of the cable sample;

[0012] Step S22: Sample balance compensation; design insulation stability loss function, let represents the classification error of the i-th reinforced cable sample, then the insulation stability loss is defined as ; ultimately enhance the total loss Expressed as: ;in, is the class probability value of the enhanced cable sample; and is the sensitivity coefficient; and is the offset coefficient; is a regulating factor; is the total classification error of the enhanced cable samples; is the control coefficient.

[0013] Furthermore, in step S3, the establishment of the low-voltage cable insulation evaluation model specifically includes the following steps:

[0014] Step S31: Model architecture design; based on the low-voltage cable monitoring dataset, Multi-ScaleCNN+BiLSTM+attention mechanism is used as the low-voltage cable insulation assessment model architecture; local features are extracted through multi-scale convolution ; Capture long-term and short-term dependencies through bidirectional LSTM; Aggregate feature information based on adaptive attention; Fully connected classification layer mapping, output probability distribution of each monitoring category ; Update parameters using gradient descent algorithm; Verify model performance based on accuracy and recall;

[0015] Step S32: loss function design; specifically including:

[0016] Step S321: Define the window weight factor, expressed as: ;in, is the window weight factor, x is the function variable; and is the window completeness threshold;

[0017] Step S322: Constructing window weight likelihood loss , the penalty is dynamically adjusted according to the ratio of the window length to the total length of the fault, which can be expressed as: ;in, is the length of the time interval of the i-th cable sample in the j-th window; is the length of the entire fault process of the i-th cable sample; and They are respectively the predicted cable status categories and Gaussian likelihood of ;

[0018] Step S323: Construct fault response adjustment; introduce fault severity regression branch, fault response adjustment Expressed as: ; ; ;in, It is the score of the window by the low voltage cable insulation evaluation model; is the compensation function; is the time interval length of the i-th cable sample in the j-1-th window; q is a constant coefficient used to control the magnitude of the overall effect; Q is a temperature parameter used to adjust the slope and change rate of the function; tanh(·) is the tanh function; and is the predicted severity score of the fault severity regression branch;

[0019] Step S324: Total loss Expressed as: ;in, is the total number of cable samples; is the monitoring loss coefficient.

[0020] Furthermore, in step S4, the low-voltage cable insulation monitoring is to collect low-voltage cable monitoring data in real time and input it into the low-voltage cable insulation evaluation model, and perform low-voltage cable insulation monitoring based on the cable status category output by the low-voltage cable insulation evaluation model; if the cable status category is one of moderate fault, serious fault and environmental abnormality, early warning processing is performed.

[0021] The present invention provides an intelligent low-voltage cable insulation monitoring system, comprising a data acquisition module, a preprocessing module, a low-voltage cable insulation evaluation model establishment module and a low-voltage cable insulation monitoring module;

[0022] The data acquisition module collects historical low-voltage cable monitoring data to construct an initial low-voltage cable monitoring data set;

[0023] The preprocessing module constructs four enhancement branches for the initial data set: disturbance injection, time warping, sample synthesis, and pulse injection, and constructs a low-voltage cable monitoring data set based on the insulation stability loss mechanism;

[0024] The low-voltage cable insulation assessment model establishment module establishes a low-voltage cable insulation assessment model based on the low-voltage cable monitoring data set combined with the window weight likelihood loss and the fault response adjustment of the fault severity regression branch;

[0025] The low-voltage cable insulation monitoring module monitors real-time low-voltage cable monitoring data based on a low-voltage cable insulation evaluation model.

[0026] The beneficial effects achieved by the present invention using the above scheme are as follows:

[0027] (1) In view of the problems that general low-voltage cable insulation monitoring methods are unable to detect early weak discharges, delay the early warning of potential insulation faults, and have a scarcity of samples of insulation damage or degradation, which leads to low monitoring reliability, this scheme improves the recognition ability of rare fault modes by constructing four enhancement branches: disturbance injection, time distortion, sample synthesis, and pulse injection; enhances the local discharge characteristics based on the pulse injection branch; avoids excessive bias through a dynamic soft weight allocation mechanism; focuses resource allocation through a fault detection network; strengthens the contribution of minority cable samples in the overall gradient based on the construction of an insulation stability loss function; and thus improves the reliability of subsequent insulation monitoring.

[0028] (2) In view of the problem that the general low-voltage cable insulation monitoring method treats all monitoring windows equally, the early weak discharge and leakage current signals are often drowned by noise, which leads to poor monitoring effect. This scheme introduces a window weight factor to ignore the early windows with insufficient information, focus on the critical section, and construct a window weight likelihood loss to enhance the identification of the critical section of the fault; introduces a fault severity regression branch to include continuous degradation signals in the penalty to promote the convergence of the critical section of the fault; and constructs a fault response adjustment based on the compensation function to reduce the false positive and missed detection rates, thereby improving the cable insulation monitoring effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 A schematic flow chart of an intelligent low-voltage cable insulation monitoring method provided by the present invention;

[0030] Figure 2 This is a schematic diagram of an intelligent low-voltage cable insulation monitoring system provided by the present invention.

[0031] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention. DETAILED DESCRIPTION

[0032] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0033] In the description of the present invention, it should be understood that terms such as "up", "down", "front", "back", "left", "right", "top", "bottom", "inside" and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the accompanying drawings. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the system or element referred to must have a specific direction, be constructed and operated in a specific direction. Therefore, they should not be understood as limiting the present invention.

[0034] Example 1, see Figure 1 The present invention provides an intelligent low-voltage cable insulation monitoring method, which includes the following steps:

[0035] Step S1: Data collection: Collect historical low-voltage cable monitoring data to construct an initial low-voltage cable monitoring data set;

[0036] Step S2: Preprocessing: construct four enhancement branches for the initial data set: disturbance injection, time warping, sample synthesis, and pulse injection, and construct a low-voltage cable monitoring data set based on the insulation stability loss mechanism;

[0037] Step S3: Establishing a low-voltage cable insulation assessment model; Based on the low-voltage cable monitoring data set combined with the window weighted likelihood loss and the fault response adjustment of the fault severity regression branch, the low-voltage cable insulation assessment model is established;

[0038] Step S4: low-voltage cable insulation monitoring: monitoring the real-time low-voltage cable monitoring data based on the low-voltage cable insulation evaluation model.

[0039] Example 2, see Figure 1 This embodiment is based on the above embodiment. In step S1, data collection is to collect historical low-voltage cable monitoring data; the historical low-voltage cable monitoring data includes partial discharge count, average discharge amplitude, leakage current and ambient temperature and humidity; the cable status is marked as a data label; the cable status includes health, early insulation degradation, moderate fault, severe fault and environmental abnormality; the historical low-voltage cable monitoring data is standardized, and the sliding window segmentation is used to construct the initial low-voltage cable monitoring data set.

[0040] Example 3, see Figure 1 This embodiment is based on the above embodiment. In step S2, preprocessing is to construct four enhancement branches for the initial low-voltage cable monitoring data set. Each branch corresponds to an enhancement strategy, and adopts dynamic weighting and insulation stability loss mechanism to automatically assign learning weights, focusing on strengthening the enhancement effect of minority cable fault samples. This is of great significance for accurately identifying low-voltage cable insulation faults and ensuring the stable operation of the power system. Then, a low-voltage cable monitoring data set is constructed. Specifically, it includes:

[0041] Step S21: Enhancement branch construction unit; the enhancement branch includes a disturbance injection branch, a time warp branch, a sample synthesis branch, and a pulse injection branch. The disturbance injection branch simulates random interference from acquisition equipment noise and electromagnetic interference, and adds Gaussian white noise to the original cable sample data; the time warp branch simulates the fault development speed or sampling rate fluctuation, introduces nonlinear deformation on the time axis, and uses piecewise linear interpolation to resample the original signal to a new time axis; the sample synthesis branch uses interpolation between a few types of fault samples to expand the data; partial discharge is the most typical physical phenomenon of low-voltage cable insulation degradation, but its amplitude is low and its duration is short in the original monitoring data, making it easily ignored. Therefore, controllable PD pulse injection is introduced, and the pulse injection branch is expressed as: ;in, and are the cable sample data after pulse injection branch enhancement and the original cable sample data respectively; t is the cable sample sampling time; U is the number of partial discharge pulses, and u is the pulse index; is the amplitude of the u-th pulse; is the absolute time position of the u-th pulse; is the equivalent half-width of the pulse; It is a gating function that outputs 1 only in the pulse valid interval and 0 in the rest. Since different fault types have different performances in the monitoring data, some fault characteristics are relatively weak and more difficult to enhance. Therefore, the branch with the more difficult enhancement effect will obtain a higher learning weight. This can prevent the simple enhancement strategy from dominating the training and ensure that the model can fully learn various fault characteristics. Each enhanced branch is trained once in the forward direction to calculate the loss value of the enhanced cable sample on the main classification network. , do softmax on these losses and get the learning weights , expressed as: ; Set up another fault detection network output , the training label is a reverse exponential mapping, expressed as: ;in, is the loss value of the enhanced cable sample on the main classification network after one forward training of the f-th enhanced branch; n and f are both enhanced branch indices; is the fault detection network training label; through the fault detection network, ensure that the two groups of high loss → large label and high loss → large weight are aligned, so that the enhanced branch can allocate resources more reasonably during the training process; the main classification loss , enhanced branch weighted loss and fault detection network loss fusion, initial enhanced loss Expressed as: ; ;in, 、 and Is to enhance the loss weight; Use cross entropy loss; is the loss of the cable sample after the enhanced branch reinforcement; It is the main classification network map; is the cable sample after the enhancement branch; y is the true label of the cable sample; the contribution of multiple enhancement strategies is automatically balanced, and more attention is paid to the minority fault cable samples that are difficult to enhance, thereby improving the ability to identify rare fault modes and more accurately discovering potential insulation fault hazards in low-voltage cable insulation monitoring;

[0042] Step S22: Sample balance compensation; In the low-voltage cable insulation monitoring data, the number of healthy cable samples is usually much larger than the number of faulty cable samples; By controlling the overall loss contribution of the majority class in the enhanced cable samples, the model is prevented from overfitting a large number of healthy cable samples; and during the operation of low-voltage cables, insulation degradation and faults occur relatively rarely, and the characteristics contained in these minority class faulty cable samples are crucial for accurate fault monitoring; Therefore, the contribution of the minority class enhanced cable samples to the gradient is strengthened to enhance the learning strength of the fault mode; Design the insulation stability loss function, let represents the classification error of the i-th reinforced cable sample, then the insulation stability loss is defined as ; ultimately enhance the total loss Expressed as: ;in, is the class probability value of the enhanced cable sample; and Is the sensitivity coefficient, which is used to control the weight and influence of different parts in the enhanced loss function; and is the offset coefficient, which is used to adjust the offset of the enhanced loss function; is a regulating factor; is the total classification error of the enhanced cable samples; is the control coefficient; focus on rare but critical insulation degradation features; the features of the early stages of insulation degradation are often not obvious and account for a small proportion in the data set, but these features are the key to early fault detection; by increasing attention to these features, the subsequent fault detection rate is improved, potential insulation faults are discovered in a timely manner, and the safe operation of low-voltage cables is guaranteed.

[0043] By performing the above operations, in order to address the problems of general low-voltage cable insulation monitoring methods, such as the inability to detect early weak discharges, delayed early warning of potential insulation faults, and scarcity of samples of insulation damage or degradation, which in turn leads to low monitoring reliability, this scheme improves the ability to identify rare fault modes by constructing four enhancement branches: disturbance injection, time warping, sample synthesis, and pulse injection; enhances local discharge characteristics based on the pulse injection branch; avoids excessive bias through a dynamic soft weight allocation mechanism; focuses resource allocation through a fault detection network; and strengthens the contribution of minority cable samples in the overall gradient by constructing an insulation stability loss function, thereby improving the reliability of subsequent insulation monitoring.

[0044] Example 4, see Figure 1 This embodiment is based on the above embodiment. In step S3, establishing a low-voltage cable insulation evaluation model specifically includes the following steps:

[0045] Step S31: Model architecture design; based on the low-voltage cable monitoring dataset, Multi-ScaleCNN+BiLSTM+attention mechanism is used as the low-voltage cable insulation assessment model architecture; local features are extracted through multi-scale convolution , expressed as: ; Capture long-term and short-term dependencies through bidirectional LSTM, expressed as: ; Based on adaptive attention aggregation feature information, it is expressed as: ; Fully connected classification layer mapping, outputting the probability distribution of each monitoring category , expressed as: ;in, is the jth cable sample in the low-voltage cable monitoring dataset; and is the convolution kernel weight and bias; ReLU is the ReLU activation function; and are the forward and reverse hidden state variables of the current time step respectively; It is a bidirectional long short-term memory network; is the feature sequence of the previous time step; is the attention weight; and are the forward and reverse hidden state variables at k time steps respectively; is the weight vector of the attention mechanism; z is the feature vector after aggregation by the attention mechanism; softmax is the softmax function; and They are the weight matrix and bias vector of the fully connected layer respectively; the gradient descent algorithm is used to update the parameters; the model performance is verified based on the accuracy and recall rate;

[0046] Step S32: loss function design; specifically including:

[0047] Step S321: define a window weight factor; short windows with insufficient information do not contribute to the loss, while the weight increases for complete windows closer to the fault endpoint, expressed as: In low-voltage cable insulation monitoring, early event fragments may contain more noise and incomplete information. Through the window weight factor, those fragments with insufficient information are ignored, and the focus is on processing more complete fragments that are closer to the fault end. is the window weight factor, x is the function variable; and is the window completeness threshold;

[0048] Step S322: Constructing window weight likelihood loss , the penalty is dynamically adjusted according to the ratio of the window length to the total length of the fault, which can be expressed as: ;in, is the length of the time interval of the i-th cable sample in the j-th window; is the length of the entire fault process of the i-th cable sample; and They are respectively the predicted cable status categories and Gaussian likelihood; in low-voltage cable insulation monitoring, segments closer to the critical insulation failure section with high partial discharge intensity and increased leakage current are more likely to contain key information about the fault; by constructing a window-weighted likelihood loss, these segments are given greater loss weights, and the evaluation model will pay more attention to these segments during training, thereby improving the sensitivity of early warning;

[0049] Step S323: Construct fault response adjustment; for three common errors in the monitoring process, including: early alarm, the model judges the segment as a fault but has not yet reached the actual starting point; delayed alarm, the window crosses the actual end point but the score is insufficient; score reversal, the current window is closer to the fault than the previous window, but the score is lower; construct fault response adjustment; increase the penalty when the current window score is higher than the previous window and meets the abnormal conditions of temperature, humidity, and PD characteristics; and introduce a fault severity regression branch to quantify the continuous degree of insulation degradation, accelerate the convergence of the evaluation model to the fault stage, and adjust the fault response. Expressed as: ; ; ;in, It is the score of the window by the low voltage cable insulation evaluation model; is the compensation function; is the time interval length of the i-th cable sample in the j-1-th window; q is a constant coefficient used to control the magnitude of the overall effect; Q is a temperature parameter used to adjust the slope and change rate of the function; tanh(·) is the tanh function; and is the predicted severity score of the fault severity regression branch, and the predicted severity score ∈ [0,1], where 0 indicates healthy insulation and 1 indicates critical failure; is the severity weight coefficient. In low-voltage cable insulation monitoring, environmental disturbances such as sudden temperature rises can lead to false positive monitoring results. By constructing a fault response adjustment, we can suppress these erroneous monitoring results caused by environmental disturbances and ensure the stability of continuous monitoring. When early alarms or score reversals occur, we increase the penalty intensity, making the evaluation model more cautious in its judgment.

[0050] Step S324: Total loss Expressed as: ;in, is the total number of cable samples; is the monitoring loss coefficient.

[0051] By performing the above operations, in view of the problem that general low-voltage cable insulation monitoring methods treat all monitoring windows equally, early weak discharge and leakage current signals are often overwhelmed by noise, which leads to poor monitoring effect. This scheme introduces a window weight factor to ignore early windows with insufficient information, focus on critical sections, and construct a window weight likelihood loss to enhance the identification of critical fault sections. It also introduces a fault severity regression branch to incorporate continuous degradation signals into the penalty to promote the convergence of critical fault sections. It also constructs a fault response adjustment based on the compensation function to reduce false positive and missed detection rates, thereby improving the cable insulation monitoring effect.

[0052] Example 5, see Figure 1 This embodiment is based on the above embodiment. In step S4, the low-voltage cable insulation monitoring is to collect low-voltage cable monitoring data in real time and input it into the low-voltage cable insulation evaluation model, and perform low-voltage cable insulation monitoring based on the cable status category output by the low-voltage cable insulation evaluation model; if the cable status category is one of moderate fault, serious fault and environmental abnormality, early warning processing is performed.

[0053] Example 6, see Figure 2 , this embodiment is based on the above embodiment, and the present invention provides an intelligent low-voltage cable insulation monitoring system, including a data acquisition module, a preprocessing module, a low-voltage cable insulation evaluation model establishment module and a low-voltage cable insulation monitoring module;

[0054] The data acquisition module collects historical low-voltage cable monitoring data to construct an initial low-voltage cable monitoring data set;

[0055] The preprocessing module constructs four enhancement branches for the initial data set: disturbance injection, time warping, sample synthesis, and pulse injection, and constructs a low-voltage cable monitoring data set based on the insulation stability loss mechanism;

[0056] The low-voltage cable insulation assessment model establishment module establishes a low-voltage cable insulation assessment model based on the low-voltage cable monitoring data set combined with the window weight likelihood loss and the fault response adjustment of the fault severity regression branch;

[0057] The low-voltage cable insulation monitoring module monitors real-time low-voltage cable monitoring data based on a low-voltage cable insulation evaluation model.

[0058] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0059] While the embodiments of the present invention have been shown and described, it will be apparent to those skilled in the art that various changes, modifications, substitutions, and alterations can be made to the embodiments without departing from the principles and spirit of the invention.

[0060] The present invention and its embodiments are described above. This description is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if a person skilled in the art is inspired by this and, without departing from the purpose of the present invention, designs structures and embodiments similar to this technical solution without inventiveness, they shall fall within the scope of protection of the present invention.

Claims

1. An intelligent low-voltage cable insulation monitoring method, characterized by: The method comprises the following steps: Step S1: Data collection: Collect historical low-voltage cable monitoring data to construct an initial low-voltage cable monitoring data set; Step S2: Preprocessing: construct four enhancement branches for the initial data set: disturbance injection, time warping, sample synthesis, and pulse injection, and construct a low-voltage cable monitoring data set based on the insulation stability loss mechanism; Step S3: Establishing a low-voltage cable insulation assessment model; Based on the low-voltage cable monitoring data set combined with the window weighted likelihood loss and the fault response adjustment of the fault severity regression branch, the low-voltage cable insulation assessment model is established; Step S4: low-voltage cable insulation monitoring: monitoring the real-time low-voltage cable monitoring data based on the low-voltage cable insulation evaluation model; In step S3, the establishment of the low-voltage cable insulation evaluation model specifically includes the following steps: Step S31: Model architecture design; based on the low-voltage cable monitoring dataset, Multi-ScaleCNN+BiLSTM+attention mechanism is used as the low-voltage cable insulation assessment model architecture; local features are extracted through multi-scale convolution ; Capture long-term and short-term dependencies through bidirectional LSTM; Aggregate feature information based on adaptive attention; Fully connected classification layer mapping, output probability distribution of each monitoring category ; Update parameters using gradient descent algorithm; Verify model performance based on accuracy and recall; Step S32: loss function design; specifically including: Step S321: Define the window weight factor, expressed as: ;in, is the window weight factor, x is the function variable; and is the window completeness threshold; Step S322: Constructing window weight likelihood loss ; Step S323: Constructing fault response adjustment ; Step S324: Total loss Expressed as: ;in, is the total number of cable samples; is the monitoring loss coefficient; i is the sample index; j is the window index.

2. The intelligent low-voltage cable insulation monitoring method according to claim 1, characterized in that: In step S2, the preprocessing is to construct four enhancement branches for the initial low-voltage cable monitoring data set, each branch corresponding to an enhancement strategy, and automatically assign learning weights using a dynamic weighting and insulation stability loss mechanism, focusing on strengthening the enhancement effect on minority cable fault samples; Then, a low-voltage cable monitoring data set is constructed; specifically, it includes: Step S21: Enhanced branch construction unit; the enhanced branch includes disturbance injection branch, time warping branch, sample synthesis branch and pulse injection branch; each enhanced branch is forward trained once to calculate the loss value of the enhanced cable sample on the main classification network , do softmax on the loss and get the learning weight , expressed as: ; Set up another fault detection network output , the training label is a reverse exponential mapping, expressed as: ;in, is the loss value of the enhanced cable sample on the main classification network after one forward training of the f-th enhanced branch; n and f are both enhanced branch indices; is the fault detection network training label; the main classification loss , enhanced branch weighted loss and fault detection network loss fusion, initial enhanced loss Expressed as: ; ;in, 、 and Is to enhance the loss weight; Use cross entropy loss; is the loss of the cable sample after the enhanced branch reinforcement; It is the main classification network map; is the cable sample enhanced by the enhancement branch; y is the true label of the cable sample; Step S22: Sample balance compensation.

3. The intelligent low-voltage cable insulation monitoring method according to claim 2, characterized in that: In step S2, the sample balance compensation is to design an insulation stability loss function, let represents the classification error of the i-th reinforced cable sample, then the insulation stability loss is defined as ; ultimately enhance the total loss Expressed as: ;in, is the class probability value of the enhanced cable sample; and is the sensitivity coefficient; and is the offset coefficient; is a regulating factor; is the total classification error of the enhanced cable samples; is the control coefficient.

4. The intelligent low-voltage cable insulation monitoring method according to claim 3, characterized in that: In step S3, the constructed window weighted likelihood loss The penalty is dynamically adjusted according to the ratio of the window length to the total length of the fault, which can be expressed as: ;in, is the length of the time interval of the i-th cable sample in the j-th window; is the length of the entire fault process of the i-th cable sample; and They are respectively the predicted cable status categories and Gaussian likelihood of .

5. The intelligent low-voltage cable insulation monitoring method according to claim 4, characterized in that: In step S3, the fault response adjustment is constructed Is to introduce fault severity regression branch, fault response adjustment Expressed as: ; ; ;in, It is the score of the window by the low voltage cable insulation evaluation model; is the compensation function; is the time interval length of the i-th cable sample in the j-1-th window; q is a constant coefficient used to control the magnitude of the overall effect; Q is a temperature parameter used to adjust the slope and change rate of the function; tanh(·) is the tanh function; and is the predicted severity score of the fault severity regression branch.

6. The intelligent low-voltage cable insulation monitoring method according to claim 5, characterized in that: In step S1, the data collection is to collect historical low-voltage cable monitoring data; mark the cable status as a data label; standardize the historical low-voltage cable monitoring data, and use sliding window segmentation to construct an initial low-voltage cable monitoring data set.

7. The intelligent low-voltage cable insulation monitoring method according to claim 6, characterized in that: In step S4, the low-voltage cable insulation monitoring is to collect low-voltage cable monitoring data in real time and input it into a low-voltage cable insulation evaluation model, and perform low-voltage cable insulation monitoring based on the cable status category output by the low-voltage cable insulation evaluation model; If the cable status category is one of moderate fault, severe fault and environmental abnormality, an early warning process is performed.

8. An intelligent low-voltage cable insulation monitoring system, for implementing an intelligent low-voltage cable insulation monitoring method according to any one of claims 1 to 7, characterized in that: It includes data acquisition module, pre-processing module, low-voltage cable insulation evaluation model building module and low-voltage cable insulation monitoring module; The data acquisition module collects historical low-voltage cable monitoring data to construct an initial low-voltage cable monitoring data set; The preprocessing module constructs four enhancement branches for the initial data set: disturbance injection, time warping, sample synthesis, and pulse injection, and constructs a low-voltage cable monitoring data set based on the insulation stability loss mechanism; The low-voltage cable insulation assessment model establishment module establishes a low-voltage cable insulation assessment model based on the low-voltage cable monitoring data set combined with the window weight likelihood loss and the fault response adjustment of the fault severity regression branch; The low-voltage cable insulation monitoring module monitors real-time low-voltage cable monitoring data based on a low-voltage cable insulation evaluation model.

Images