Power distribution network cable fault early warning method and system based on multi-terminal synchronous traveling wave detection

Abstract

The application discloses a kind of based on the cable fault early warning method and system of distribution network of multi-terminal synchronous traveling wave detection, and it is related to cable fault detection technical field.The present application comprises the following steps: at least three synchronous detection ends are deployed in each branch point of cable, and sampling rate is synchronously triggered;Establish master-slave type synchronous network, and carry out spectrum correction to the data collected;Based on distributed temperature sensor data, construct real-time correction model;Extract early fault feature, and construct three-dimensional feature tensor;Edge node preliminary diagnosis is carried out, and network is optimized using transfer learning to carry out feature classification;Combined with cable load rate and environmental temperature and humidity, establish early warning threshold adaptive model based on fuzzy PID.The present application is through multi-terminal collaborative perception and intelligent analysis, and the scheme changes fault processing mode from "passive repair" to "active defense", improves distribution network cable artificial inspection efficiency, and early detects early fault hidden danger.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of cable fault detection, and particularly relates to a distribution network cable fault early warning method and system based on multi-terminal synchronous traveling wave detection. Through multi-terminal collaborative sensing and intelligent analysis, the scheme changes the fault handling mode from "passive repair" to "active defense", significantly improving the reliability of distribution network power supply. BACKGROUND

[0002] With the continuous expansion of urban power grid scale, high-voltage power cables have become the core carrier of urban power transmission and distribution systems. However, the insulation deterioration, partial discharge and other fault hidden dangers caused by factors such as load fluctuation, environmental temperature and humidity changes, and mechanical stress during long-term operation of the cable are increasingly prominent, seriously threatening the reliability of power supply. In the prior art, cable state monitoring and fault early warning face the following key bottlenecks:

[0003] Threshold rigidity problem

[0004] Traditional monitoring systems mostly use fixed threshold criteria (such as partial discharge > 50pC, temperature rise rate > 3℃ / min), but the actual operation state of the cable is strongly coupled with dynamic load and environmental factors. For example, in a high-humidity environment in summer, the insulation performance of the cable will decrease with the increase of humidity, at which time the fixed threshold value may cause a false negative; while in the winter low-temperature high-load working condition, it is easy to cause false alarm due to the too conservative threshold value.

[0005] Feature extraction limitation

[0006] Existing fault feature extraction methods (such as wavelet transform, frequency spectrum analysis) are easily affected by power frequency harmonics and electromagnetic interference in a strong noise environment. Typical cases show that when the cable load rate > 80%, the traditional wavelet packet decomposition algorithm has a capture rate of early partial discharge features of 61.3%, and there is a feature aliasing phenomenon caused by spectrum leakage.

[0007] Insufficient synchronization accuracy

[0008] Multi-terminal traveling wave positioning technology relies on high-precision time synchronization, but existing systems mostly use single GPS clock synchronization (accuracy ±200ns), which is difficult to meet the sub-microsecond level synchronization requirement in complex electromagnetic environment. Analysis of a 500kV cable fault case shows that the positioning error caused by synchronization deviation is up to ±300 meters, which seriously delays the fault repair time.

[0009] Environmental adaptability defects

[0010] Traditional monitoring terminals are prone to baseline drift and sampling error in extreme temperature and humidity conditions (such as -40℃ to +70℃, humidity > 95%RH). Cable network monitoring data in a certain plateau area shows that the signal-to-noise ratio (SNR) of the vibration sensor signal decreases by about 15dB in a low-temperature environment in winter, resulting in an increase in the mechanical failure detection rate to 28%.

[0011] To solve the above problems, the present application uses deep fusion of dynamic parameter perception, adaptive feature extraction and high-precision synchronous acquisition technology to build a multi-dimensional collaborative monitoring system for cable status, aiming to break through the technical bottleneck and realize accurate early warning of cable hazards. SUMMARY

[0012] The purpose of the present application is to provide a distribution network cable fault early warning method and system based on multi-terminal synchronous traveling wave detection, which changes the fault handling mode from "passive repair" to "active defense" through multi-terminal collaborative perception and intelligent analysis, solving the problems of low efficiency of existing distribution network cable manual inspection, high concealment of early fault hazards and difficulty in detection.

[0013] To solve the above technical problems, the present application is realized by the following technical scheme:

[0014] The present application is a distribution network cable fault early warning method based on multi-terminal synchronous traveling wave detection, comprising the following steps:

[0015] Step S1: At least three synchronous detection terminals are deployed at each branch end of the cable, and the sampling rate is synchronously triggered;

[0016] Step S2: Establish a master-slave synchronous network, use FPGA to generate a sampling clock to realize multi-terminal synchronous start, and perform frequency spectrum correction on the collected data;

[0017] Step S3: Establish a cable network traveling wave propagation matrix model, and calculate the wave arrival direction by the cross-correlation time difference method;

[0018] Step S4: Based on the distributed temperature sensor data, a real-time correction model is established;

[0019] Step S5: Extract early fault features and construct a three-dimensional feature tensor;

[0020] Step S6: The edge node performs preliminary diagnosis and uses transfer learning to optimize the network for feature classification;

[0021] Step S7: Combine the cable load rate and environmental temperature and humidity to establish a fuzzy PID-based early warning threshold adaptive model.

[0022] As a preferred technical solution, in the step S1, the synchronous detection end is arranged at the first and last ends and the intermediate joints of the cable main line and the T branch, forming a multi-end monitoring array; the synchronous detection end comprises a high-frequency current transformer and a piezoelectric vibration sensor; the high-frequency current transformer is used to collect the core wire current and the shielding layer electromagnetic field, and the high-frequency current transformer can capture the high-frequency current signal generated when the partial discharge occurs in the cable by detecting the discharge pulse current signal flowing through the cable grounding wire, the neutral point wire and the cable body. These signals propagate along the outer shielding layer or the grounding wire of the cable, and the high-frequency current transformer as a "signal capturer" can accurately capture these weak signals and convert them into measurable electrical signals, providing a data basis for subsequent analysis and processing; the piezoelectric vibration sensor is used to collect mechanical vibration signals, mainly including vibration acceleration, vibration frequency and vibration amplitude, and can generate electric charges. By measuring these electric charges or the voltage signals converted therefrom, the size of the vibration acceleration can be determined. In addition, the sensor can also measure the frequency and amplitude of the vibration and other parameters.

[0023] As a preferred technical solution, in the step S2, a master-slave synchronous network is established based on Beidou / GPS; wherein, the master clock node periodically sends 1PPS second pulse, and the slave node transmits PTPv2 protocol message through the cable sheath to realize that the time deviation of all nodes is less than 50ns; the specific process of the spectrum correction is as follows:

[0024] Step S21: adopt parameterized Vincent window to perform signal truncation, and the specific formula is as follows:

[0025] ;

[0026] In the formula, is a dynamic adjustment factor, and the value range is [0.08, 0.16], and self-adaptive adjustment is realized by real-time monitoring of background noise spectral density; is the weighting coefficient value of the window function at the discrete time point , is the discrete time index, is the total length of the window function, is the main lobe shaping core, is the side lobe suppression adjustment;

[0027] Step S22: perform 4096-point FFT on the windowed new signal to obtain the initial spectrum , and calculate the maximum value point of the amplitude in the main lobe range;

[0028] Step S23: construct a cubic spline interpolation function in the interval to solve the real frequency deviation ; ​

[0029] Step S24: Establishing amplitude-phase correction model , performing iterative promotion of frequency resolution, wherein, is the corrected signal amplitude, is the original measured amplitude, is the frequency deviation factor, is the spectrum leakage compensation term;

[0030] Step S25: Real-time monitoring of power frequency deviation, establishing dynamic compensation matrix, and eliminating the influence of frequency fluctuation;

[0031] Step S26: Constructing multiple filtering architecture for anti-aliasing processing; the front stage of the anti-aliasing processing adopts an analog anti-aliasing filter (cut-off frequency 500 kHz, transition band attenuation 120 dB / oct), and the rear stage digital filter uses an elliptical filter (passband ripple <0.01 dB, stopband attenuation >90 dB); after spectrum correction, three polynomial frequency domain interpolation is performed to compensate for the amplitude attenuation caused by the window function, and the specific formula is as follows: ; wherein, is the final amplitude after frequency domain interpolation compensation, is the current analyzed frequency point, is the actual frequency of the signal or the center frequency of the main lobe of the spectrum, is the sampling frequency.

[0032] As a preferred technical solution, in the step S3, the specific process of establishing the cable network wave propagation matrix model is as follows:

[0033] Step S31: Abstracting the power distribution network cable system as a weighted graph ;

[0034] wherein, the vertex set represents the physical node, the edge set represents the parameters of the cable section, and the weight matrix stores the propagation delay and attenuation coefficient between nodes;

[0035] Step S32: Establishing time-varying propagation matrix , obtaining the transmission function in the wave propagation process:

[0036] ;

[0037] wherein, is the multipath reflection coefficient, is the path delay, is the cable frequency-dependent loss transfer function;

[0038] Step S33: Establishing joint differential equation of wave propagation;

[0039] Step S34: Calculate the cross-correlation of the two signals after continuous wavelet transform:

[0040] ;

[0041] wherein, is the wavelet coefficient matrix, is the weight function, focusing on the main frequency band, is the corresponding traveling wave characteristic scale;

[0042] Step S35: Construct an interpolation function within ±3 sampling points of the cross-correlation peak value, and the specific formula is:

[0043] ;

[0044] wherein, is the cubic spline basis function, is the true time difference estimate value, is the value of finding the maximum value of the objective function, is the weighted sum of discrete sampling points, is the cross-correlation function value of the signal at the time delay , is the cubic spline basis function; Step S36: Construct a cost function to distinguish direct waves and reflected waves, and solve the wave direction of arrival;

[0045] Through cubic spline interpolation, the discrete cross-correlation function is continuously reconstructed, breaking through the traditional sampling interval limit, and finally the accurate time delay that makes the reconstructed cross-correlation function reach the peak value is solved

[0046] .

[0047] As a preferred technical solution, in the step S36, the wave direction of arrival solving process is as follows:

[0048] Step S361: Calculate the time difference between each two terminals, forming a time difference matrix ;

[0049] Step S362: Use the propagation matrix to perform path inversion:

[0050] ;

[0051] wherein, is the parameter vector or matrix to be optimized, is the actual observed time difference matrix, is the propagation model matrix,​ is the square of the Frobenius norm, is the regularization coefficient, is the L1 norm;

[0052] The purpose of the DOA solution is to find the best match between the time difference model and the target. Minimum), and can ensure the sparsity of the solution ( smallest) In cable fault detection, the solution corresponds to the most likely fault location and intensity distribution;

[0053] Step S363: Construct array flow matrix , solve the spatial spectrum peak, the calculation formula is:

[0054] ;

[0055] Where, is the optimal direction of arrival, that is, the incident angle of the signal source. By maximizing the spatial spectrum function, the true direction of the signal source is determined. To find the maximum value of an expression value, find the value that minimizes the denominator within the angle search range (Because the smaller the denominator, the larger the overall score), is the conjugate transpose of the steering vector, is the noise subspace matrix, which is obtained by eigendecomposing the covariance matrix; is the noise subspace projection matrix, which represents the projection of the signal on the noise subspace. When it is orthogonal to the noise subspace, the projection energy of the matrix is ​​minimum and the corresponding denominator value is minimum;

[0056] The specific function is to search for the smallest denominator , find the direction of the signal source,

[0057] when and noise subspace When orthogonal, the denominator approaches zero and the fraction is maximum, corresponding to the true direction of the signal source; the orthogonal characteristics of the noise subspace and the signal subspace are utilized to suppress noise interference and improve the direction estimation accuracy.

[0058] As a preferred technical solution, in step S4, the material characteristics are analyzed:

[0059] Establishing the dielectric constant of cable insulation materials Regression equation with temperature:

[0060] ;

[0061] In the formula, the coefficient T is temperature, measured by thermal aging test;

[0062] Derivation of the conductor expansion coefficient model:

[0063] ;

[0064] In the formula, is the expansion coefficient of the conductor, is the length of the object at temperature T, is the initial length at the reference temperature By fusing distributed temperature sensing and cable material temperature characteristics, the wave velocity calculation error is reduced, and the traveling wave positioning accuracy under complex working conditions is effectively improved.

[0065] As a preferred technical solution, in step S5, the early fault feature is extracted, and the specific process of constructing the three-dimensional feature tensor is as follows:

[0066] Step S51: dynamically select the optimal wavelet basis function according to the cable fault feature frequency band, and construct the parameterized Morlet wavelet family: ; wherein, according to According to the signal instantaneous frequency dynamic adjustment;

[0067] Step S52: introduce local-global weight factor, dynamic energy entropy ratio calculation, the calculation formula is as follows:

[0068] ; In the formula, is the wavelet energy entropy ratio of the kth node, is the wavelet transform coefficient, which represents the time-frequency feature of the signal at scale and translation parameter

[0069] Step S53: 6-layer wavelet packet decomposition, covering 0-30MHz frequency band, calculating the normalized WEE value at each node, forming a 64-dimensional feature vector, and retaining the top 8 most discriminant features through KL divergence;

[0070] Step S54: calculate the time synchronization difference degree based on the multi-terminal traveling wave arrival time difference:

[0071] ;

[0072] In the formula, is the wave head arrival time marked by each terminal, and the time resolution is improved through cross-correlation algorithm;

[0073] ​Step S55: fuse electromagnetic / vibration dual-mode data and construct a hybrid frequency spectrum matrix; wherein, the row vector is 1MHz frequency spectrum data (0-500kHz resolution 0.5Hz) collected by each terminal; and the column vector is a time sequence from 10 seconds before the fault to 200ms after the fault;

[0074] Step S56: perform truncated SVD decomposition , and retain the first three singular values , to form a three-dimensional feature tensor formula ; , respectively, are the left singular vector matrix, the singular value matrix, and the right singular vector matrix; and the three-dimensional feature tensor is the time synchronization difference degree, the modal energy distribution entropy, and the frequency spectrum singular value decomposition coefficient.

[0075] As a preferred technical solution, in the step S6, the specific process of the preliminary diagnosis of the edge node is as follows:

[0076] Step S61: adopt a sliding window mechanism to dynamically segment the real-time traveling wave signal, and the window length is adaptively adjusted according to the cable length;

[0077] Step S62: extract the third-order feature vector of the segmented waveform: time domain feature, frequency domain feature, and morphological feature; wherein, the time domain feature includes the wave crest coefficient and the waveform factor; the frequency domain feature includes the main frequency band energy proportion; and the morphological feature includes the rising slope;

[0078] Step S63: introduce a dynamic time adjustment optimization algorithm to solve the problem of waveform time axis shift, and the specific formula is as follows:

[0079] ;

[0080] In the formula, the distance metric The Manhattan distance and the cosine similarity are combined by weighting; an adaptive noise suppression layer is added, and when the signal-to-noise ratio (SNR) is less than 15dB, Gaussian filtering is automatically activated to improve the comparison stability;

[0081] Step S64: establish a fault waveform feature library, and perform three-level similarity determination; the fault waveform feature library contains reference waveforms of 16 typical fault modes (grounding / short circuit / disconnection, etc.); the three-level similarity determination includes a coarse screening stage: calculating the waveform envelope area similarity (threshold>85%); a fine screening stage: Hausdorff distance calculation (threshold<0.15); and a verification stage: adjacent terminal waveform polarity consistency verification.

[0082] As a preferred technical solution, in the step S6, the network structure includes: constructing a 12-layer ResNet-1D structure, which includes:

[0083] Input layer: accept 128-dimensional feature vector (64-dimensional in time domain + 64-dimensional in frequency domain);

[0084] Residual block: expand the receptive field with dilated convolution (dilation = 2);

[0085] Attention mechanism: dynamically adjust channel weights by embedding SE module;

[0086] When optimizing the network through transfer learning, the multi-terminal data received by the cloud is spatio-temporally aligned, improved SVD noise reduction processing is performed, and the first three principal components are retained; multi-scale feature fusion is performed, which specifically includes:

[0087] Low-level features: first-order difference signals of the original waveform;

[0088] Intermediate features: node energies of wavelet packet decomposition;

[0089] High-level features: singular value coefficients of time-frequency joint distribution;

[0090] Adopting a triplet loss function , wherein, is a boundary parameter, are the anchor point, positive sample and negative sample respectively; when dynamically adjusting the class weight, the class weight is automatically calculated based on the fault occurrence frequency (the weight of rare faults is increased by 3-5 times), and a fuzzy membership function is introduced to process boundary fuzzy samples.

[0091] As a preferred technical solution, in the step S7, the specific steps of establishing the early warning threshold adaptive model based on fuzzy PID are as follows:

[0092] Step S71: measure the cable core current in real time through the Rogowski coil and calculate the load rate; ;

[0093] In the formula, is the load rate of the cable at time t, is the current effective value at time t, is the rated current of the cable;

[0094] Step S72: deploy distributed temperature and humidity sensors, install one set every 200 meters of cable, and establish an environmental factor matrix , wherein, respectively represent the cable outer surface temperature, the cable environment humidity and the cable core conductor operating temperature, and the conductor temperature is measured by non-contact infrared thermal imager;

[0095] Step S73: define three fuzzy input quantities, including load rate deviation, temperature rise rate and four-degree influence factor, establish a fuzzy rule base, and there are 81 control rules, and the rule weight is dynamically adjusted according to the aging characteristics of the cable insulation material;

[0096] Step S74: PID parameter self-integration, output quantity deblurring formula: ; wherein, is the activation degree, is the corresponding rule output value, and the proportional-integral-derivative coefficient is updated in real time to ensure that the adjustment period is less than 500 ms;

[0097] Step S75: Establishing an initial threshold matrix according to the cable model ; wherein, are respectively the partial discharge threshold, the harmonic distortion rate and the traveling wave amplitude threshold, and the adaptive adjustment equation is constructed, and the specific formula is as follows: ; wherein, is the reference threshold, is the proportional coefficient, is the load rate deviation, is the integral coefficient, is the temperature rise rate integral term, is the differential coefficient, is the temperature influence factor conversion rate, and the correction amplitude is limited within the range of ±40% to prevent over-regulation;

[0098] Step S76: Introducing a Sigmoid function for nonlinear amplitude limiting to ensure that the threshold is always in the safe interval and complete the dynamic calculation of the early warning threshold.

[0099] The application is a kind of power distribution network cable fault early warning system based on multi-terminal synchronous traveling wave detection, including a distributed traveling wave sensor array, a quantum time synchronization network, an intelligent analysis hub and an anti-interference transmission system.

[0100] The distributed traveling wave sensor array includes a multi-modal sensor cluster and an edge computing node; the multi-modal sensor cluster deploys electromagnetic-ultrasonic composite sensors at cable joints, integrates high-frequency current transformers and piezoelectric vibration sensors, and is used for synchronous acquisition of multiple physical quantities of traveling waves; the edge computing node is provided with an FPGA preprocessing chip in each terminal, which is used for wavelet packet threshold denoising and traveling wave arrival time marking;

[0101] The quantum time synchronization network includes a Beidou / GPS-rubidium atomic clock dual-mode time service module and a PTPv2 protocol optimization channel; the Beidou / GPS-rubidium atomic clock dual-mode time service module is used for time synchronization error of each terminal; the PTPv2 protocol optimization channel is used to establish a hybrid communication link based on the cable metal sheath to ensure that the time synchronization signal can still be transmitted when the communication is interrupted;

[0102] The intelligent analysis hub comprises a feature fusion engine, a deep prediction model and a dynamic risk assessment module; the feature fusion engine adopts a multi-source data fusion algorithm of an attention mechanism to establish a time domain-frequency domain-energy domain joint feature matrix; the deep prediction model is based on a spatiotemporal sequence prediction network of TCN-GRU, and embedded with cable topology prior knowledge, and is used for early fault feature extraction; and the dynamic risk assessment module is used for combining an improved D-S fusion algorithm of evidence theory to construct a five-level early warning index system.

[0103] The anti-interference transmission system is a power line carrier and NB-IoT dual-channel redundant transmission.

[0104] The present application has the following advantages:

[0105] (1) The present application deploys at least three synchronous detection ends at each branch end of the cable to establish a master-slave synchronous network, corrects the collected data to establish a cable network traveling wave propagation matrix model, establishes a fuzzy PID-based early warning threshold adaptive model, realizes early warning of faults, effectively identifies early weak faults, and realizes the technical leap from passive positioning to active early warning through the innovative combination of multi-physical quantity fusion and deep learning technology.

[0106] (2) The present application integrates the cable network topology, material parameters and environmental temperature and humidity into the time-varying propagation matrix to reduce model errors, simultaneously considers the propagation differences of electromagnetic traveling waves and mechanical stress waves, realizes composite wave direction of arrival (DOA) solving, introduces quantum bit coding into the particle swarm algorithm to make the DOA estimation resolution reach 0.5°, and effectively suppresses false alarms caused by reflected waves and reduces the false alarm rate through propagation matrix inversion and cost function constraints.

[0107] (3) The fuzzy PID threshold adjustment model of the present application reduces the false alarm rate and the missed alarm rate under complex working conditions such as ±40% load fluctuation, sudden change of temperature and humidity (ΔT>10℃ / h, ΔH>30%RH / h), etc.; based on the local-global weight factor and adaptive threshold denoising technology, the effective signal separation degree is improved by 10% compared with the traditional wavelet packet method.

[0108] (4) The three-dimensional feature tensor of time synchronization difference, modal energy distribution entropy and spectral singular value of the present application can simultaneously represent the multi-physical field coupling effect of electrical, mechanical and thermal fields, and improve the fault type recognition accuracy.

[0109] Of course, any product implementing the present application does not necessarily need to achieve all the advantages described above at the same time. BRIEF DESCRIPTION OF DRAWINGS

[0110] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without any creative effort on the basis of these drawings.

[0111] Figure 1 A flow chart of a cable fault early warning method based on multi-terminal synchronous traveling wave detection of a distribution network of the present application;

[0112] Figure 2 A structural schematic diagram of a cable fault early warning system based on multi-terminal synchronous traveling wave detection of the present application. DETAILED DESCRIPTION

[0113] The technical solutions in the embodiments of the present application will be described clearly and completely in the following with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without any creative effort belong to the protection scope of the present application.

[0114] In addition, the technical features involved in each of the embodiments of the present application described below can be combined with each other as long as they do not conflict with each other.

[0115] In order to make the purpose, technical solutions and advantages of the present application more clear, the following will combine the drawings in the embodiments of the present application. Figure 2 The embodiments of the present application will be described in further detail.

[0116] Before introducing the embodiments of the present application, first, the application of traveling wave on the cable of the distribution network will be described.

[0117] The traveling wave fault location technology is based on the traveling wave theory. When a fault occurs in a power transmission line, a sudden change of voltage and current will be generated at the fault point, forming a traveling wave. The traveling wave propagates along the power transmission line to both ends, and its propagation characteristics are related to the characteristic impedance, wave impedance and distribution parameters of the line. In the propagation process of the traveling wave, reflection and transmission will occur when the traveling wave encounters a line impedance discontinuity point (such as the end of the line, a branch point or a fault point).

[0118] The traveling wave propagates along the power transmission line at a speed close to the speed of light, and its propagation speed is related to the electrical parameters of the line, usually between 290,000 and 310,000 km / s. When the traveling wave reaches the measurement points at both ends of the line, reflected waves and transmitted waves will be generated. By accurately measuring the time difference of the arrival of the traveling wave, combined with the wave speed of the line, the distance of the fault point from the measurement point can be calculated.

[0119] For the purpose, technical solutions and advantages of the present application to be more clearly understood, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only intended to explain the present application and not to limit the present application. Figures 1-2 The present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only intended to explain the present application and not to limit the present application.

[0120] Embodiment one, please refer to Figure 1 The present application is a kind of based on multi-terminal synchronous traveling wave detection distribution network cable fault early warning method, comprising the following steps:

[0121] Step S1: at least three synchronous detection ends are deployed at each branch point of the cable, and the sampling rate is synchronously triggered;

[0122] Step S2: establish a master-slave synchronous network, use FPGA to generate a sampling clock to realize multi-terminal synchronous start, and correct the frequency spectrum of the collected data;

[0123] Step S3: establish a cable network traveling wave propagation matrix model, and calculate the wave direction by the cross-correlation time difference method;

[0124] Step S4: based on the distributed temperature sensor data, build a real-time correction model;

[0125] Step S5: extract early fault features and build a three-dimensional feature tensor;

[0126] Step S6: the edge node preliminarily diagnoses and uses transfer learning to optimize the network for feature classification;

[0127] Step S7: combine the cable load rate and environmental temperature and humidity to establish a fuzzy PID-based early warning threshold adaptive model.

[0128] In step S1, the synchronous detection end is deployed at the first and last ends and the intermediate joints of the cable trunk line and the T-branch, forming a multi-terminal monitoring array; the synchronous detection end includes a high-frequency current transformer and a piezoelectric vibration sensor; the high-frequency current transformer is used to collect the core wire current and the electromagnetic field of the shielding layer, and the high-frequency current transformer can capture the high-frequency current signal generated by the local discharge in the cable by detecting the discharge pulse current signal in the cable grounding wire, the neutral point connection and the cable body. These signals propagate along the outer shielding layer or grounding wire of the cable, and the high-frequency current transformer as a "signal capturer" can accurately capture these weak signals and convert them into measurable electrical signals, providing a data basis for subsequent analysis and processing; the piezoelectric vibration sensor is used to collect mechanical vibration signals, mainly including vibration acceleration, vibration frequency and vibration amplitude, and can generate electric charge. By measuring these electric charges or the voltage signals converted therefrom, the size of the vibration acceleration can be determined. In addition, the sensor can also measure the frequency and amplitude of the vibration and other parameters.

[0129] In step S2, a master-slave synchronization network is established based on BeiDou / GPS. The master clock node periodically sends 1PPS pulses per second, and the slave nodes transmit PTPv2 protocol messages through the cable sheath, achieving a time deviation of less than 50ns for all nodes. The specific process of spectrum correction is as follows:

[0130] Step S21: Use the parameterized Vincent window to perform signal truncation. The specific formula is as follows:

[0131] ;

[0132] Where, is a dynamic adjustment factor with a value range of [0.08, 0.16], which is adaptively adjusted by real-time monitoring of the background noise spectral density; is the window function at discrete time points The weighted coefficient value at is the discrete time index, is the total length of the window function, As the core of the main valve shaping, Adjust for sidelobe suppression;

[0133] Step S22: Perform 4096-point FFT on the windowed new signal to obtain the initial spectrum , calculate the maximum amplitude point within the main lobe range ;

[0134] Step S23: Construct a cubic spline interpolation function within the interval to solve the true frequency deviation ;

[0135] Step S24: Establishing an amplitude-phase correction model , iteratively improve the frequency resolution, where, is the signal amplitude after correction, is the original measured amplitude, is the frequency deviation factor, is the spectrum leakage compensation term;

[0136] Step S25: Real-time monitoring of the power frequency offset, establishing a dynamic compensation matrix, and eliminating the impact of frequency fluctuations;

[0137] Step S26: Construct a multi-filter architecture for anti-aliasing processing. The front-end of the anti-aliasing process uses an analog anti-aliasing filter (cutoff frequency 500kHz, transition band attenuation 120dB / oct), and the back-end digital filter uses an elliptical filter (passband ripple <0.01dB, stopband attenuation >90dB). After spectrum correction, perform cubic polynomial frequency domain interpolation to compensate for the amplitude attenuation caused by the window function. The specific formula is as follows: Where, is the final amplitude after frequency domain interpolation compensation, is the frequency point currently analyzed, is the actual frequency of the signal or the center frequency of the main lobe of the spectrum, is the sampling frequency.

[0138] In step S3, the specific process of establishing the cable network traveling wave propagation matrix model is as follows:

[0139] Step S31: Abstract the distribution network cable system into a weighted graph ;

[0140] In the formula, the vertex set Represents physical nodes, edge sets Represents the parameters of the cable segment, the weight matrix Store the propagation delay and attenuation coefficient between each node;

[0141] Step S32: Establish a time-varying propagation matrix , obtain the transfer function during traveling wave transmission:

[0142] ;

[0143] Where, is the multipath reflection coefficient, is the path delay, is the cable frequency-dependent loss transfer function;

[0144] Step S33: establishing a joint differential equation for traveling wave propagation;

[0145] Step S34: Two-way signal Calculate the cross-correlation after performing a continuous wavelet transform:

[0146] ;

[0147] Where, is the wavelet coefficient matrix, is the weight function, focusing on the main frequency band, is the characteristic scale of the corresponding traveling wave;

[0148] Step S35: construct an interpolation function within ±3 sampling points of the cross-correlation peak. The specific formula is:

[0149] ;

[0150] Where, is the cubic spline basis function, is the true time difference estimate, To find the maximum value of the objective function value, To perform weighted summation of discrete sampling points, For signal In the delay The cross-correlation function value at is the cubic spline basis function;

[0151] Step S36: constructing a cost function to distinguish direct waves from reflected waves and calculating the direction of arrival;

[0152] The discrete cross-correlation function is continuously reconstructed through cubic spline interpolation, breaking through the traditional sampling interval limitation and ultimately solving the precise time delay that makes the reconstructed cross-correlation function reach its peak. .

[0153] In step S36, the direction of arrival solution process is as follows:

[0154] Step S361: Calculate the time difference between each of the N terminals , forming a time difference matrix ;

[0155] Step S362: Using the propagation matrix Perform path inversion:

[0156] ;

[0157] Where, is the parameter vector or matrix to be optimized, is the actual observed time difference matrix, is the propagation model matrix, is the square of the Frobenius norm, is the regularization coefficient, is the L1 norm;

[0158] The purpose of the DOA solution is to find the best match between the time difference model and the target. Minimum), and can ensure the sparsity of the solution ( smallest) In cable fault detection, the solution corresponds to the most likely fault location and intensity distribution;

[0159] Step S363: Construct array flow matrix , solve the spatial spectrum peak, the calculation formula is:

[0160] ;

[0161] Where, is the optimal direction of arrival, that is, the incident angle of the signal source, To find the maximum value of an expression value, is the conjugate transpose of the steering vector, is the noise subspace matrix, is the noise subspace projection matrix, representing the projection of the signal on the noise subspace; the specific role is to find the signal source direction by searching for the minimum time division mother .

[0162] In step S4, the material characteristics are analyzed:

[0163] The regression equation of the dielectric constant of the cable insulation material and the temperature is established:

[0164] ;

[0165] In the formula, the coefficient is measured by a thermal aging test, and T is the temperature;

[0166] The conductor expansion coefficient model is derived:

[0167] ;

[0168] In the formula, is the expansion coefficient of the conductor, is the length of the object at temperature T, is the initial length at the reference temperature By fusing distributed temperature sensing and cable material temperature characteristics, the error in wave velocity calculation is reduced, and the positioning accuracy under complex working conditions is effectively improved.

[0169] In step S5, the early fault features are extracted, and the specific process of constructing the three-dimensional feature tensor is as follows:

[0170] Step S51: dynamically select the optimal wavelet basis function according to the cable fault feature frequency band, and construct the parameterized Morlet wavelet family: ; wherein, according to According to the signal instantaneous frequency, the dynamic adjustment is as follows:

[0171] Step S52: introduce local-global weight factors, and dynamically calculate the energy entropy ratio, the calculation formula is as follows:

[0172] ; in the formula, is the wavelet energy entropy ratio of the kth node, is the wavelet transform coefficient, representing the time-frequency feature of the signal at scale and translation parameter ;

[0173] ​Step S53: 6-layer wavelet packet decomposition, covering 0-30MHz frequency band, calculating the normalized WEE value at each node to form a 64-dimensional feature vector, and retaining the top 8 most discriminative features through KL divergence;

[0174] Step S54: Calculate the time synchronization difference based on multi-terminal traveling wave arrival time difference:

[0175] ;

[0176] In the formula, The wave front arrival time of each terminal is marked, and the time resolution is improved through cross-correlation algorithm;

[0177] Step S55: Fusion of electromagnetic / vibration dual-mode data, construction of hybrid frequency spectrum matrix; the row vector is 1MHz frequency spectrum data (0-500kHz resolution 0.5Hz) collected by each terminal; the column vector is the time sequence from 10 seconds before failure to 200ms after failure;

[0178] Step S56: Perform truncated SVD decomposition , retain the top 3 singular values , form a three-dimensional feature tensor formula ; , respectively, singular value matrix, right singular vector matrix; the three-dimensional feature tensor is time synchronization difference, modal energy distribution entropy, and spectral singular value decomposition coefficient.

[0179] In step S6, the specific process of preliminary diagnosis of the edge node is as follows:

[0180] Step S61: Adopt sliding window mechanism to dynamically segment real-time traveling wave signals, and the window length is adaptively adjusted according to the cable length;

[0181] Step S62: Extract the third-order feature vector of the segmented waveform: time domain feature, frequency domain feature and morphological feature; wherein the time domain feature includes wave peak coefficient and waveform factor; the frequency domain feature includes main frequency band energy proportion; the morphological feature includes rising slope;

[0182] Step S63: Introduce dynamic time adjustment optimization algorithm to solve the problem of waveform time axis shift, and the specific formula is as follows:

[0183] ;

[0184] In the formula, the distance measure The Manhattan distance and cosine similarity are combined by weighting; an adaptive noise suppression layer is added, and when the signal-to-noise ratio (SNR) is less than 15dB, Gaussian filtering is automatically activated to improve the stability of comparison;

[0185] Step S64: Establish a fault waveform feature library and perform a three-level similarity determination. The fault waveform feature library contains reference waveforms for 16 typical fault modes (such as grounding, short circuit, and line break). The three-level similarity determination includes a coarse screening stage: calculating the waveform envelope area similarity (threshold > 85%); a fine screening stage: calculating the Hausdorff distance (threshold < 0.15); and a verification stage: checking the polarity consistency of adjacent terminal waveforms.

[0186] In step S6, the network structure includes: constructing a 12-layer ResNet-1D structure, including:

[0187] Input layer: accepts 128-dimensional feature vectors (64 dimensions in time domain + 64 dimensions in frequency domain);

[0188] Residual block: uses dilated convolution (dilation=2) to expand the receptive field;

[0189] Attention mechanism: embedding SE module to dynamically adjust channel weights;

[0190] When performing transfer learning to optimize the network, the multi-terminal data received from the cloud is spatially and temporally aligned, and an improved SVD noise reduction process is performed to retain the first three principal components. Multi-scale feature fusion is also performed, specifically including:

[0191] Low-level features: first-order difference signal of the original waveform;

[0192] Intermediate features: node energy of wavelet packet decomposition;

[0193] Advanced features: singular value coefficients of the joint time-frequency distribution;

[0194] Using triplet loss function , where is the boundary parameter, They are anchor points, positive samples and negative samples respectively; when the dynamic category weight is adjusted, the category weight is automatically calculated based on the frequency of fault occurrence (the weight of rare faults is increased by 3 times to 5 times), and the fuzzy membership function is introduced to process the boundary fuzzy samples.

[0195] In step S7, the specific steps of establishing the early warning threshold adaptive model based on fuzzy PID are as follows:

[0196] Step S71: measuring the cable core current in real time through the Rogowski coil and calculating the load rate; ;

[0197] Where, is the load rate of the cable at time t, is the effective value of the current at time t, is the rated current of the cable;

[0198] Step S72: deploy distributed temperature and humidity sensors, install 1 set every 200 meters of cable section, and establish an environmental factor matrix wherein, respectively represent the cable outer surface temperature, the ambient humidity where the cable is located, and the operating temperature of the cable core conductor, and the conductor temperature is measured non-contact by an infrared thermal imager;

[0199] Step S73: define three fuzzy input quantities, including load rate deviation, temperature rise rate, and four-degree influence factor, establish a fuzzy rule base, and there are 81 control rules, and the rule weight is dynamically adjusted according to the aging characteristics of the cable insulation material;

[0200] Step S74: PID parameter self-tuning, and the output quantity is disengaged from the formula: ; wherein, is the activation degree, is the corresponding rule output value, the proportional-integral-derivative coefficient is updated in real time, and the adjustment period is ensured to be less than 500 ms;

[0201] Step S75: establish an initial threshold matrix according to the cable model ; wherein, respectively are the partial discharge threshold, the harmonic distortion rate, and the traveling wave amplitude threshold, an adaptive adjustment equation is constructed, and the specific formula is as follows: ; wherein, is the reference threshold, is the proportional coefficient, is the load rate deviation, is the integral coefficient, is the temperature rise rate integral term, is the differential coefficient, is the temperature influence factor transformation rate, the correction amplitude is limited within ±40%, and over-regulation is prevented;

[0202] Step S76: introduce a Sigmoid function for nonlinear amplitude limiting, ensure that the threshold is always in the safe range, and complete the dynamic calculation of the warning threshold.

[0203] Embodiment two, as shown in Figure 2 , the present application is a kind of based on multi-terminal synchronous traveling wave detection of distribution network cable fault early warning system, can be used to complete the method content of embodiment 1 of the present application, including: distributed traveling wave sensor array, quantum time synchronization network, intelligent analysis center and anti-interference transmission system;

[0204] The distributed traveling wave sensor array comprises a multi-modal sensor cluster and an edge computing node; the multi-modal sensor cluster deploys an electromagnetic-ultrasonic composite sensor at a cable joint, integrates a high-frequency current transformer and a piezoelectric vibration sensor, and is used for synchronous acquisition of multiple physical quantities of traveling waves; and the edge computing node is provided with an FPGA pre-processing chip in each terminal, which is used for wavelet packet threshold denoising and traveling wave arrival time marking.

[0205] The quantum time synchronization network comprises a Beidou / GPS-rubidium atomic clock dual-mode time service module and a PTPv2 protocol optimized channel; the Beidou / GPS-rubidium atomic clock dual-mode time service module is used for time synchronization error of each terminal; and the PTPv2 protocol optimized channel is used for establishing a hybrid communication link based on a cable metal sheath, and ensuring that the time synchronization signal can still be transmitted when communication is interrupted.

[0206] The intelligent analysis hub comprises a feature fusion engine, a deep prediction model and a dynamic risk assessment module; the feature fusion engine adopts a multi-source data fusion algorithm of an attention mechanism, and establishes a time domain-frequency domain-energy domain joint feature matrix; the deep prediction model is a spatiotemporal sequence prediction network based on TCN-GRU, and embeds cable topology prior knowledge, and is used for early fault feature extraction; and the dynamic risk assessment module is used for constructing a five-level early warning index system in combination with an improved D-S fusion algorithm of evidence theory.

[0207] The anti-interference transmission system is a dual-channel redundant transmission of a power line carrier and an NB-IoT.

[0208] It is worth noting that, in the above system embodiments, each unit included is only divided according to functional logic, but is not limited to the above division, as long as the corresponding functions can be realized; in addition, the specific names of each functional unit are only for easy mutual differentiation, and do not limit the protection scope of the present application.

[0209] In addition, those skilled in the art can understand that all or part of the steps of the above-mentioned embodiments can be completed by programs instructing related hardware, and the corresponding programs can be stored in a computer readable storage medium.

[0210] The preferred embodiments of the application disclosed above are only used to help explain the application. The preferred embodiments do not describe all the details, nor limit the application to the specific embodiments described. Obviously, many modifications and variations can be made according to the content of the present application. The embodiments are selected and described in detail in order to better explain the principles and practical applications of the application, so that those skilled in the art can well understand and utilize the application. The application is limited only by the claims and their full scope and equivalents.

Claims

1. A distribution network cable fault early warning method based on multi-terminal synchronous traveling wave detection, characterized in that: The steps include: Step S1: deploy at least three synchronous detection terminals at each branch end of the cable, with sampling rates triggered synchronously; Step S2: Establish a master-slave synchronization network, use FPGA to generate sampling clock to achieve multi-terminal synchronous startup, and perform spectrum correction on the collected data; Step S3: Establish a cable network traveling wave propagation matrix model and calculate the direction of arrival by using the cross-correlation time difference method; Step S4: constructing a real-time correction model based on distributed temperature sensor data; Step S5: extracting early fault features and constructing a three-dimensional feature tensor; Step S6: preliminary diagnosis of edge nodes and feature classification using transfer learning optimized network; Step S7: combining the cable load rate and the ambient temperature and humidity, establishing a fuzzy PID-based early warning threshold adaptive model; In step S3, the specific process of establishing the cable network traveling wave propagation matrix model is as follows: Step S31: Abstract the distribution network cable system into a weighted graph ; In the formula, the vertex set Represents physical nodes, edge sets Represents the parameters of the cable segment, the weight matrix Store the propagation delay and attenuation coefficient between each node; Step S32: Establish a time-varying propagation matrix , obtain the transfer function during traveling wave transmission: ； Where, is the multipath reflection coefficient, is the path delay, is the cable frequency-dependent loss transfer function; Step S33: establishing a joint differential equation for traveling wave propagation; Step S34: Two-way signal Calculate the cross-correlation after performing a continuous wavelet transform: ； Where, is the wavelet coefficient matrix, is the weight function, focusing on the main frequency band, is the characteristic scale of the corresponding traveling wave; Step S35: construct an interpolation function within ±3 sampling points of the cross-correlation peak. The specific formula is: ； Where, is the cubic spline basis function, is the true time difference estimate, To find the maximum value of the objective function value, To perform weighted summation of discrete sampling points, For signal In the delay The cross-correlation function value at is the cubic spline basis function; Step S36: constructing a cost function to distinguish direct waves from reflected waves and calculating the direction of arrival; In step S5, the specific process of extracting early fault features and constructing a three-dimensional feature tensor is as follows: Step S51: Dynamically select the optimal wavelet basis function according to the cable fault characteristic frequency band and construct a parameterized Morlet wavelet family: ;in, Dynamic adjustment according to the instantaneous frequency of the signal; Step S52: Introduce the local-global weight factor and calculate the dynamic energy entropy ratio. The calculation formula is as follows: Where, is the wavelet energy entropy ratio of the kth node, is the wavelet transform coefficient, which indicates the signal at scale and translation parameters The time-frequency characteristics of Step S53: 6-layer wavelet packet decomposition, covering the 0-30 MHz frequency band, calculating the normalized WEE value for each node to form a 64-dimensional feature vector, and retaining the top 8 most discriminative features through KL divergence; Step S54: Calculate the time synchronization difference based on the arrival time difference of the traveling waves of multiple terminals: ； Where, The arrival time of the wave head marked for each terminal is improved with time resolution through cross-correlation algorithm; Step S55: Fuse the electromagnetic / vibration dual-modal data to construct a hybrid spectrum matrix, where the row vectors are the 1 MHz spectrum data collected by each terminal, and the column vectors are the time series from 10 seconds before the fault to 200 ms after the fault. Step S56: Perform truncated SVD decomposition , retain the first 3 singular values , forming a three-dimensional feature tensor formula: ; are the left singular vector matrix, singular value matrix, and right singular vector matrix respectively; the three-dimensional eigentensors are the time synchronization difference, modal energy distribution entropy, and spectrum singular value decomposition coefficient respectively; In step S7, the specific steps of establishing the early warning threshold adaptive model based on fuzzy PID are as follows: Step S71: measuring the cable core current in real time through the Rogowski coil and calculating the load rate; ; Where, is the load rate of the cable at time t, is the effective value of the current at time t, is the rated current of the cable; Step S72: Deploy distributed temperature and humidity sensors, install one set every 200 meters of cable segment, and establish an environmental factor matrix ,in, Respectively represent the cable outer surface temperature, the cable's ambient humidity and the operating temperature of the cable core conductor. The conductor temperature is measured non-contact by an infrared thermal imager. Step S73: Define three fuzzy input quantities, including load rate deviation, temperature rise rate, and four-degree influence factors, and establish a fuzzy rule base with a total of 81 control rules. The rule weights are dynamically adjusted according to the aging characteristics of the cable insulation material; Step S74: Customize PID parameters and output defuzzification formula: Where, is the activation degree, To output the corresponding rule value, the proportional-integral-differential coefficients are updated in real time to ensure that the adjustment cycle is less than 500ms; Step S75: Establish an initial threshold matrix based on the cable model Where, The adaptive adjustment equations are constructed for the partial discharge threshold, harmonic distortion rate and traveling wave amplitude threshold respectively. The specific formulas are as follows: Where, is the baseline threshold, is the proportionality coefficient, is the load factor deviation, is the integration coefficient, is the integral term of the temperature rise rate, is the differential coefficient, The temperature influence factor conversion rate is limited to ±40%; Step S76: Introduce the Sigmoid function to perform nonlinear limiting to ensure that the threshold is always within the safe range and complete the dynamic calculation of the warning threshold.

2. A distribution network cable fault early warning method based on multi-terminal synchronous traveling wave detection according to claim 1, characterized in that: In step S1, the synchronous detection end is deployed at the beginning and end and the middle joint of the cable trunk line and the T-connected branch to form a multi-terminal monitoring array; the synchronous detection end includes a high-frequency current transformer and a piezoelectric vibration sensor; the high-frequency current transformer is used to collect the core wire current; the piezoelectric vibration sensor is used to shield the electromagnetic field and mechanical vibration signal of the layer.

3. A distribution network cable fault early warning method based on multi-terminal synchronous traveling wave detection according to claim 1, characterized in that: In step S2, a master-slave synchronization network is established based on Beidou / GPS; wherein the master clock node periodically sends 1PPS second pulses, and the slave node transmits PTPv2 protocol messages through the cable sheath; the specific process of the spectrum correction is as follows: Step S21: using a parameterized Vincent window to perform signal truncation; Step S22: Perform a 4096-point FFT on the windowed new signal to obtain an initial spectrum, and calculate the maximum amplitude point within the main lobe range; Step S23: constructing a cubic spline interpolation function within the specified interval to solve the true frequency deviation; Step S24: establishing an amplitude-phase correction model for iteration; Step S25: Real-time monitoring of power frequency offset and establishment of a dynamic compensation matrix; Step S26: constructing a multi-filter architecture to perform anti-aliasing processing.

4. A distribution network cable fault early warning method based on multi-terminal synchronous traveling wave detection according to claim 1, characterized in that: In step S36, the direction of arrival solution process is as follows: Step S361: Calculate the time difference between each of the N terminals to form a time difference matrix; Step S362: performing path inversion using the propagation matrix; Step S363: construct an array flow matrix and calculate the spatial spectrum peak value.

5. The distribution network cable fault early warning method based on multi-terminal synchronous traveling wave detection according to claim 1 is characterized in that: In step S6, the specific process of the edge node preliminary diagnosis is as follows: Step S61: Dynamically segment the real-time traveling wave signal using a sliding window mechanism, with the window length adaptively adjusted according to the cable length; Step S62: extracting the third-order feature vectors of the segmented waveform: time domain features, frequency domain features, and morphological features; Step S63: introducing a dynamic time adjustment optimization algorithm and adding an adaptive noise suppression layer; Step S64: Establish a fault waveform feature library and perform three-level similarity determination.

6. A distribution network cable fault early warning method based on multi-terminal synchronous traveling wave detection according to claim 1, characterized in that: In step S6, when the transfer learning optimizes the network, the multi-terminal data received from the cloud are temporally and spatially aligned, and an improved SVD noise reduction process is performed to retain the first three principal components; multi-scale feature fusion is performed, and a triplet loss function is used to perform dynamic category weight adjustment.

7. A distribution network cable fault early warning system based on multi-terminal synchronous traveling wave detection, used to implement the distribution network cable fault early warning method based on multi-terminal synchronous traveling wave detection as described in any one of claims 1 to 6, comprising a distributed traveling wave sensor array, a quantum time synchronization network, an intelligent analysis hub, and an anti-interference transmission system, characterized in that: The distributed traveling wave sensor array includes a multimodal sensor cluster and an edge computing node. The multimodal sensor cluster deploys electromagnetic-ultrasonic composite sensors at cable joints, integrating high-frequency current transformers and piezoelectric vibration sensors for synchronous acquisition of multiple traveling wave physical quantities. The edge computing node has a built-in FPGA pre-processing chip for each terminal, which is used for wavelet packet threshold denoising and traveling wave arrival time marking. The quantum time synchronization network includes a BeiDou / GPS-rubidium atomic clock dual-mode timing module and a PTPv2 protocol optimization channel; the BeiDou / GPS-rubidium atomic clock dual-mode timing module is used to reduce the time synchronization error of each terminal; the PTPv2 protocol optimization channel is used to establish a hybrid communication link based on the cable metal sheath to ensure that the time synchronization signal can still be transmitted when the communication is interrupted; The intelligent analysis hub includes a feature fusion engine, a deep prediction model, and a dynamic risk assessment module; The feature fusion engine uses a multi-source data fusion algorithm with an attention mechanism to establish a joint feature matrix of the time domain, frequency domain, and energy domain. The deep prediction model is based on a TCN-GRU spatiotemporal sequence prediction network, embedding prior knowledge of cable topology for early fault feature extraction. The dynamic risk assessment module is used to combine the DS fusion algorithm with the improved evidence theory to build a five-level early warning indicator system; The anti-interference transmission system is a power line carrier and NB-IoT dual-channel redundant transmission.

Images