GIS vibration and discharge dual-mode source visualization positioning method

Abstract

The application relates to the technical field of GIS intelligent diagnosis, and discloses a GIS vibration and discharge bimodal source tracing visualization positioning method, which comprises the following steps: S1, a multidimensional bimodal collaborative acquisition platform is built, and dynamic working conditions are simulated; S2, discharge signal propagation time delay characteristics and vibration signal time-frequency domain characteristics are extracted; S3, a GIS device space-time topology constraint model is constructed, and bimodal signal correlation is calculated; S4, effective signals are screened, and a defect position range is preliminarily determined; and S5, defect coordinates are calculated, and GIS local discharge fault source tracing positioning three-dimensional visualization results are output. The technical scheme points are as follows: by fusing vibration-discharge bimodal signals and space-time topology constraints, the limitation of traditional single signal positioning is broken through, accurate source tracing positioning of GIS local discharge faults is realized, state monitoring scenes of GIS devices of different structural types are adapted, and new technical support is provided for fault positioning and operation and maintenance of power system GIS devices.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of GIS intelligent diagnosis, and particularly relates to a GIS vibration and discharge dual-mode source tracing visual positioning method. BACKGROUND

[0002] As a key primary equipment of the power system, the insulation state of GIS is related to the safe and stable operation of the power grid, and partial discharge is the core signal of insulation defect evolution. Therefore, accurate positioning of the partial discharge source is crucial for the operation and maintenance of GIS equipment. Current GIS partial discharge fault source tracing and positioning faces multiple technical difficulties: traditional positioning methods mostly rely on a single signal (such as ultra-high frequency or ultrasonic signal), which is easily disturbed by the closed structure of GIS and complex electromagnetic environment, and the positioning accuracy is difficult to guarantee; there are shortcomings in the signal fusion link, and the spatio-temporal correlation characteristics of multi-source signals such as discharge and vibration are not fully explored and fused, and key positioning information is easily overwhelmed by noise or irrelevant signals; the positioning algorithm is not well adapted to the GIS equipment scene, and the internal structure of GIS equipment is complex, with various signal propagation paths, and the conventional positioning algorithm is difficult to accurately calculate the defect position without considering the time-space topology characteristics of the equipment.

[0003] In the prior art, the research on multi-source signal positioning method mostly focuses on simple superposition of signals or improvement of single algorithm, and does not form a complete technical closed loop of "deep fusion of multi-source signals - accurate constraint of time-space topology - optimization of positioning algorithm adaptation" according to the structure of GIS equipment and signal propagation characteristics. For example, traditional multi-source positioning algorithms do not fully utilize the three-dimensional structure parameters of GIS equipment and the physical laws of signal propagation, resulting in large deviation of positioning results; some intelligent algorithms are used for positioning, but lack deep cooperation with the time-space topology model of GIS equipment, and the adaptation is insufficient, so that the source tracing and positioning technology is difficult to meet the demand of power grid for lean operation and maintenance of GIS equipment in actual engineering application. Therefore, it is urgent to develop a GIS vibration-discharge dual-mode source tracing visual positioning method with time-space topology constraint, which realizes accurate source tracing and positioning of GIS partial discharge fault by fusing dual-mode signals, constructing accurate time-space topology constraint model and optimizing multi-source positioning algorithm, and provides strong technical support for equipment operation and maintenance. SUMMARY

[0004] The purpose of the present application is to overcome the problems in the prior art that GIS partial discharge fault positioning relies on a single signal, time-space constraint is not accurate, and multi-source signal fusion is low, and to provide a GIS vibration and discharge dual-mode source tracing visual positioning method, which can be applied to the source tracing and positioning of partial discharge faults, accurate positioning of defects and intelligent operation and maintenance analysis of gas insulated switchgear of various voltage levels.

[0005] To achieve the above purpose, the technical scheme of the present application is as follows:

[0006] A GIS vibration and discharge dual-mode tracing visualization positioning method, comprising the following steps:

[0007] S1: Build a multi-dimensional dual-mode collaborative acquisition platform and simulate dynamic working conditions, and synchronously acquire GIS device partial discharge signals and vibration signals;

[0008] S2: Adopt a VMD-HHT-KPCA multi-algorithm fusion strategy to extract discharge signal propagation time delay features and vibration signal time-frequency domain features; wherein VMD is a variational mode decomposition, which is a signal self-adaptive decomposition algorithm, used to realize partial discharge signal noise reduction and modal separation; HHT is a Hilbert-Huang transform, used to extract vibration signal time-frequency domain features; KPCA is a kernel principal component analysis, used to reduce and optimize multi-dimensional features and eliminate redundant information;

[0009] S3: Combine three-dimensional electromagnetic simulation and improved IPSO algorithm to build a GIS device space-time topology constraint model and calculate the correlation degree of dual-mode signals;

[0010] S4: Screen effective signals and preliminarily determine the defect position range through correlation degree threshold screening and IPSO constraint fusion;

[0011] S5: Use improved WLS and WOA collaborative algorithm to iteratively calculate defect coordinates and output GIS partial discharge fault tracing positioning three-dimensional visualization results.

[0012] Further, the step S1 comprises the following sub-steps:

[0013] S11: Build a multi-dimensional dual-mode collaborative acquisition test platform, select a GIS test sample machine with gradient composite insulation defects, configure a UHF sensor to acquire partial discharge UHF electromagnetic signals, deploy a fiber bragg grating vibration sensing array to acquire vibration signals, cooperate with a pulse generator, a high-speed oscilloscope, a time scale synchronization device and a multi-channel data acquisition device to build a partial discharge-vibration dual-mode signal acquisition system, and realize high-sensitivity capture of UHF discharge electromagnetic signals and vibration displacement signals;

[0014] S12: Debug the GIS test sample machine operating parameters, adopt a gradient boosting combined with partial discharge pulse triggering voltage regulation mode, gradually boost from the initial voltage, when the UHF sensor detects a partial discharge UHF pulse, regulate the voltage with small fluctuations based on the voltage as a reference, simulate partial discharge and vibration coupling working conditions under actual power grid voltage fluctuations, and create dual-mode signal acquisition working conditions close to actual operation scenarios;

[0015] S13: Start the data acquisition device, realize the time synchronization of the partial discharge ultra-high frequency signal and the optical fiber vibration signal by means of the time tag synchronization device; for each gradient composite insulation defect, continuously and synchronously collect the partial discharge original ultra-high frequency signal and the vibration original optical fiber grating signal according to the adaptive acquisition period, collect multiple samples for each defect, and form a double-mode signal sample library containing the gradient composite defect type label, the working voltage information and the signal space-time correlation characteristics.

[0016] Further, the step S11 includes the combined defects of conductor spikes and metal particles and the combined defects of insulator surface particles and dynamic suspended discharge.

[0017] Further, the step S2 includes the following sub-steps:

[0018] S21: For the synchronously collected partial discharge signal and the vibration signal, the spatial position information of the sensors in the GIS device space-time topology model is combined to calculate the medium difference time delay difference of the discharge signal propagation to different sensors; at the same time, the amplitude distribution of the discharge pulse is counted with the power frequency period as the time axis to generate an enhanced PRPD (Phase Resolved Partial Discharge) spectrum containing phase-amplitude-frequency information, wherein the PRPD spectrum is a phase-resolved partial discharge spectrum taking the phase, amplitude and frequency of the partial discharge signal as the core parameters, which is used to identify the defect type and severity of the partial discharge; for the vibration signal, the time-domain amplitude sequence is constructed, and the structure-related parameter of the sensor installation position is combined to generate a structure-related time-domain spectrum of the vibration signal.

[0019] S22: Extract the discharge signal characteristics of the discharge pulse from the enhanced PRPD spectrum, the discharge signal characteristics are the maximum amplitude, the average discharge amount and the positive and negative half-cycle pulse asymmetry, the variational mode decomposition algorithm is used for noise reduction processing to retain the integrity of the pulse characteristics of the discharge signal; from the structure-related time-domain spectrum of the vibration signal, the time-frequency domain characteristics are extracted by Hilbert-Huang transform, the time-frequency domain characteristics include the instantaneous amplitude peak value and the vibration energy root mean square value, the frequency domain characteristics are corrected by using the device structure modal parameters to enhance the correlation between the vibration characteristics and the device structure.

[0020] S23: The improved kernel principal component analysis algorithm is used for dimension reduction processing of the double-mode characteristics, the high-dimensional characteristics are mapped to the kernel space by the kernel function, the principal components meeting the requirement of cumulative contribution rate are selected, the low-dimensional characteristic matrix of the fusion discharge-vibration space-time characteristics is constructed, the differential characteristics of different defect types are retained and the characteristic redundancy is eliminated, and efficient input is provided for subsequent correlation degree calculation.

[0021] Further, the step S3 includes the following sub-steps:

[0022] S31: Based on the physical structure parameters of GIS equipment, a three-dimensional electromagnetic model containing metal shell, SF6 gas medium, internal conductor and insulator is established, the propagation path, attenuation law and boundary reflection characteristics of partial discharge electromagnetic signal in the equipment are simulated by finite element method, and the electromagnetic propagation theory basis for space-time topology constraint is provided; at the same time, the parameter optimization framework based on improved particle swarm algorithm is constructed, and the signal propagation characteristics output by the three-dimensional electromagnetic model and the sensor coordinates are used to initialize the particle population, and each particle corresponds to the optimized parameter vector of the bimodal signal correlation degree calculation model , particle parameters , and so on; at the same time, the basic parameters are set, including population size, maximum iteration number, learning factor;

[0023] S32: Define the bimodal signal correlation degree as the fitness function , and the core formula is:

[0024] (2);

[0025] Wherein is the space-time weight coefficient, which is determined by the particle parameters , is the time domain similarity of the discharge signal feature vector and the vibration signal feature vector , which is calculated by dynamic time warping algorithm, is the weight used to balance the frequency domain matching degree in the correlation degree calculation, is the frequency domain matching degree of the two, which is solved based on the cross power spectral density function, is the feature vector of the discharge signal used to calculate the frequency domain matching degree, is the feature vector of the vibration signal used to calculate the frequency domain matching degree;

[0026] The position updating formula of the improved particle swarm algorithm is:

[0027] (5);

[0028] Wherein is the inertia weight, is the velocity vector of particle i at the kth iteration, is the velocity vector of particle i at the k+1th iteration, which is used to describe the position updating trend of the particle in the parameter optimization process; is the learning factor, is a random number in the range of [0, 1], which increases the randomness of the algorithm and enhances the diversity of the search, is the individual optimal position, is the position vector of particle i at the kth iteration, is the global optimal position, the optimization weight coefficient and the feature matching threshold value are optimized to improve the accuracy of the correlation degree calculation, is the position vector of particle i at the k+1th iteration, and is the new position obtained after updating the speed;

[0029] S33: When the number of iterations reaches the set value or the change rate of the fitness function value of the continuous 5 iterations is less than the threshold value, terminate the iteration and output the optimal parameter combination , complete the parameter configuration of the correlation degree calculation model, and provide quantitative basis for the spatio-temporal correlation analysis of the dual-mode signal.

[0030] Further, the physical structure parameters in step S31 include cavity size, conductor layout, and insulating medium distribution.

[0031] Further, step S4 includes the following sub-steps:

[0032] S41: According to the correlation degree calculation results of the dual-mode signal, the signal effectiveness judgment threshold value is set by combining the spatio-temporal topology model of the GIS device, the effective discharge signal and vibration signal closely related to the defect source are screened out, and the invalid signals disturbed or irrelevant are removed;

[0033] S42: The spatial constraint conditions of the discharge signal source are constructed by using the propagation time delay characteristics of the effective discharge signal and the spatial structure parameters of the GIS device; at the same time, the spatial constraint conditions of the vibration signal source are constructed based on the propagation characteristics of the vibration signal and the device structure mode;

[0034] S43: The spatial constraint conditions of the discharge and vibration signals are fused to obtain the region where the defect may exist that meets the spatio-temporal correlation of the dual-mode signal, and the range of the defect position is preliminarily determined to provide a basis for subsequent accurate calculation of the defect coordinates.

[0035] Further, step S5 includes the following sub-steps:

[0036] S51: Taking the defect position range preliminarily determined in step S4 as the constraint boundary, the time difference of the discharge pulse arriving at different sensors and the path difference characteristics of the vibration signal propagation in multiple groups of effective dual-mode signals are extracted, and a positioning basic data set containing the three-dimensional coordinates of the sensors is established;

[0037] S52: An improved weighted least squares algorithm is used to construct a positioning model, taking the defect coordinates as the variable to be solved, combining the propagation speed parameters of the signal in different media, and establishing an objective function:

[0038] (9).

[0039] wherein is a weight coefficient based on the correlation degree of the dual-mode signal, is a three-dimensional spatial coordinate component of the GIS device partial discharge fault defect to be solved; is the X-axis coordinate value of the i th sensor in the three-dimensional coordinate system; is the Y-axis coordinate value of the i th sensor in the three-dimensional coordinate system; is the Z-axis coordinate value of the i th sensor in the three-dimensional coordinate system; is the signal propagation speed, is the time when the signal reaches the i th sensor;

[0040] S53: Introduce the whale optimization algorithm for iterative solution, take the position range determined in step S4 as the initial search space, update the candidate defect coordinates through the surrounding predation, bubble net attack and random search mechanism simulating the hunting behavior of humpback whales, and prune invalid solutions in combination with the spatial topological constraints of the GIS device in the iteration process, until the target function value is less than the set threshold or the maximum iteration number is reached, and finally output the optimal defect coordinates , realize three-dimensional visualization output of GIS partial discharge fault tracing and positioning results.

[0041] Compared with the prior art, the beneficial effects of the present application are:

[0042] 1. Multi-modal deep fusion, through vibration-discharge dual-mode signal cooperation, combined with three-dimensional electromagnetic simulation modeling, breaking through the limitations of traditional single signal positioning, improving the accuracy of GIS partial discharge fault tracing and positioning;

[0043] 2. Spatiotemporal constraint precision, constructing a spatiotemporal topological constraint model containing three-dimensional electromagnetic propagation characteristics, deeply integrating geometric space and electromagnetic physical laws, enhancing the practicality and precision of positioning constraints;

[0044] 3. Algorithm efficient adaptation, improving the cooperation of weighted least squares and whale optimization algorithm, with the help of the preliminary positioning range constraint, improving the algorithm convergence speed while ensuring the positioning accuracy, adapting to the positioning scene under the complex structure of GIS equipment;

[0045] 4. Strengthening operation and maintenance support, realizing accurate tracing and positioning of GIS partial discharge faults, providing accurate basis for equipment operation and maintenance, helping to quickly evaluate the insulation state of GIS, and ensuring the safe and stable operation of power grids. BRIEF DESCRIPTION OF DRAWINGS

[0046] 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 embodiment description.

[0047] Figure 1 is a total flow chart of the present application;

[0048] Figure 2 is a bimodal signal multi-source acquisition and transmission architecture flow chart of the present application;

[0049] Figure 3 is a three-dimensional electromagnetic simulation model construction and optimization flow chart of the present application;

[0050] Figure 4 is a bimodal driven intelligent defect positioning and decision flow chart of the present application;

[0051] Figure 5 is a GIS space-time topology constraint local discharge signal angle-power density three-dimensional distribution schematic diagram of the present application. The diagram takes "azimuth angle" and "elevation angle" as spatial angle dimensions and "power density" as signal energy dimension, and directly presents the spatial propagation law and energy concentration interval of the local discharge signal under the GIS device structure constraint.

[0052] Figure 6 is a fault three-dimensional positioning imaging flow chart based on the improved optimization positioning algorithm and GIS geometric model of the present application;

[0053] Figure 7 is a GIS device local discharge fault three-dimensional positioning visualization schematic diagram of the present application. DETAILED DESCRIPTION

[0054] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0055] In the description of the present application, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer" and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the device or element referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation on the present application. In addition, the terms "first", "second", "third" are only for the purpose of description, and cannot be understood as indicating or implying relative importance.

[0056] In the description of the present application, it should be noted that unless otherwise explicitly specified and limited, the terms "mounting", "connection", "connecting" should be understood broadly, for example, it can be fixed connection, or detachable connection, or integrally connected, it can be mechanical connection, or electrical connection, it can be directly connected, or indirectly connected through intermediate medium, or the internal communication of two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances. In addition, the technical features involved in the different embodiments of the present application described below can be combined with each other as long as they do not conflict with each other.

[0057] As shown in Figure 1 , it is a flow chart of the method of the present application. First, the GIS device partial discharge signal and the vibration signal are synchronously collected, then the propagation time delay and other characteristics of the discharge signal and the time domain and frequency domain characteristics of the vibration signal are extracted respectively, then the GIS device space-time topology constraint model containing the three-dimensional electromagnetic simulation model is constructed, the correlation degree is calculated combining the double modal signal characteristics, and then the effective signal is screened based on the space-time topology constraint and the correlation degree, and the defect position range is preliminarily determined, finally the multi-source positioning algorithm is used to iteratively calculate the defect coordinates, and the tracing positioning result of the GIS partial discharge fault is output.

[0058] As shown in Figure 2 , it is a flow chart of the double modal signal multi-source collection and transmission architecture of the present application.

[0059] A GIS vibration and discharge double modal tracing visual positioning method, comprising the following steps:

[0060] S1: Build a multi-dimensional double modal collaborative collection platform and simulate dynamic working conditions, and synchronously collect GIS device partial discharge signals and vibration signals;

[0061] Traditional GIS partial discharge positioning relies on a single signal, such as ultra-high frequency or ultrasonic signal, and a conventional collection mode, which has the problems of single signal dimension, only reflecting partial discharge or vibration in one aspect, and weak anti-interference ability, that is, the effective signal is easily submerged due to the influence of complex GIS electromagnetic environment and mechanical noise, resulting in insufficient reliability of positioning data basis.

[0062] This step effectively solves the problems of missing double modal signal correlation characteristics and insufficient effective information caused by single signal and poor time synchronization in the traditional collection method by introducing multi-source heterogeneous sensor networking, time synchronization technology and dynamic gain control. Among them, the time synchronization technology ensures the accurate alignment of the discharge and vibration signals in the time dimension, and the dynamic gain control intelligently adjusts the sensor gain according to the signal strength, thereby ensuring the effective collection of weak signals. These measures collectively provide high-quality double modal raw data for subsequent positioning tasks.

[0063] The step S1 comprises the following sub-steps:

[0064] S11: Build a multi-dimensional dual-mode collaborative acquisition test platform.

[0065] Select a GIS test prototype with gradient composite insulation defects, including the combination of conductor spikes and metal particles, where the tip curvature radius of the conductor spikes is less than or equal to 0.5 mm, and the particle size of the metal particles ranges from 50 to 200 microns. It also includes the combination of insulator surface micro-gap and dynamic suspended discharge, with a micro-gap width of 20 to 100 microns and a suspended potential difference ranging from 0 to 10 kV. These defects all have significant size differences and evolution characteristics.

[0066] The platform is configured with a very high frequency sensor covering a bandwidth of 300 MHz to 1.5 GHz, and local partial discharge very high frequency electromagnetic signals are collected with a layout spacing of less than or equal to 10 cm. At the same time, a fiber grating vibration sensing array with a sampling frequency not less than 200 kHz is deployed to collect vibration signals. The system also integrates a pulse generator with a pulse peak of 0 to 30 kV, a high-speed oscilloscope with a bandwidth not less than 1 GHz, a nanosecond-level time synchronization device, and a multi-channel data acquisition device with a channel number not less than 16, together forming a complete partial discharge-vibration dual-mode signal acquisition system. The system ultimately realizes high-sensitivity capture of very high frequency discharge electromagnetic signals and vibration displacement signals, with an amplitude resolution of the electromagnetic signal reaching below 1 millivolt and a measurement accuracy of the vibration displacement signal reaching below 1 micron.

[0067] S12: Debug the operating parameters of the GIS test prototype. Adopt a voltage regulation method combining gradient voltage rise and local discharge pulse triggering, starting from an initial voltage of 0 kV and gradually rising at a rate of 0.5 kV / s. When the very high frequency sensor detects a local discharge very high frequency pulse with an amplitude not less than 5 mV, the voltage is taken as the reference to perform small amplitude fluctuation regulation within ±5% amplitude, to simulate the local discharge and vibration coupling conditions under actual power grid voltage fluctuations and create a dual-mode signal acquisition condition close to actual operation scenarios.

[0068] S13: Start the data acquisition device. With the aid of a time synchronization device with a time synchronization accuracy of within 10 nanoseconds, accurate time synchronization of the local discharge very high frequency signal and the fiber vibration signal is realized. For each type of gradient composite insulation defect, the local discharge original very high frequency signal and the vibration original fiber grating signal are continuously and synchronously acquired according to an adaptive acquisition period, which can be intelligently adjusted within the range of 100 hertz to 10 kilohertz according to the discharge pulse frequency. Not less than 20 groups of samples are acquired for each defect, each group containing more than 1000 pulse periods of data, ultimately forming a dual-mode signal sample library containing gradient composite defect type labels, working voltage information, and signal space-time correlation characteristics.

[0069] S2: A VMD-HHT-KPCA multi-algorithm fusion strategy is adopted to extract the propagation delay characteristics of the discharge signal and the time-frequency domain characteristics of the vibration signal;

[0070] Traditional partial discharge or vibration signal feature extraction relies on fixed analysis rules for single signals, resulting in insufficient mining of signal correlation features and a lack of multi-dimensional feature fusion, leading to fuzzy correlations between dual-modal signals. This step, based on the original synchronous signals of partial discharge and vibration, uses a cross-correlation algorithm combined with signal propagation path analysis to extract features such as propagation delay and amplitude attenuation of discharge signals from different sensors for the discharge signal. For the vibration signal, variational mode decomposition (VMD) is applied to decompose the signal, obtaining each modal component, and then extracting features such as peak value and root mean square in the time domain, and center frequency and spectral entropy in the frequency domain. This addresses the problems of traditional methods, which suffer from single-signal analysis and limited feature dimensions, resulting in missing correlation features between dual-modal signals and weak distinguishing power of localization input features, providing accurate dual-modal feature data for subsequent correlation calculation and localization.

[0071] Step S2 includes the following sub-steps:

[0072] S21: For the synchronously acquired partial discharge and vibration signals, combined with the spatial location information of the sensors in the spatiotemporal topology model of the GIS equipment, the spatial distance accuracy between sensors is ≤5mm, according to the formula... ,in For the discharge signal to propagate from sensor i to sensor j The time delay, Let i be the spatial distance between sensors i and j. This represents the propagation speed of the signal within the GIS internal medium, with a value approximately 2 × 10⁻⁶. 8 m / s, calculate the time delay difference of the discharge signal propagating to different sensors due to differences in the medium.

[0073] Simultaneously, using a power frequency cycle of 50Hz and period T=20ms as the time axis, the phase interval is subdivided into 360 equal parts. The amplitude distribution of the discharge pulse within each phase interval is statistically analyzed to generate an enhanced PRPD spectrum containing phase-amplitude-frequency information, where the phase resolution is 1°, the amplitude quantization accuracy is 1mV, and the frequency statistical accuracy is 1Hz. For the vibration signal, a sampling frequency is constructed. The time-domain amplitude sequence, combined with the structural modal correlation parameters of the sensor installation location, including the cavity natural frequency, is used. Modal testing was conducted to determine an error ≤1%, and a structural correlation time-domain spectrum of the vibration signal was generated with a time resolution of 5. μ s.

[0074] S22: Extract the maximum amplitude of the discharge pulse from the enhanced PRPD spectrum. , average discharge amount (N is the total number of pulses, is the discharge amount of the kth pulse), positive and negative half-cycle pulse asymmetry , , respectively, the number of positive and negative half-cycle pulses) and other discharge signal characteristics, noise reduction processing is performed using the variational mode decomposition (VMD) algorithm, and the constraint variational model of VMD is:

[0075] (1);

[0076] wherein is the kth decomposition modal component of the vibration signal, which is used to split different frequency components of the vibration signal; is the kth vibration component corresponding to the dominant angular frequency; is a time-frequency domain transformation term, j is an imaginary unit, t is a time variable, and is used to convert the vibration component from the time domain to the frequency domain, reflecting its frequency characteristics; is a time function of the discharge pulse, which is used to represent the instantaneous time of the occurrence of partial discharge in the GIS device; is a fusion feature function of the discharge-vibration signal; is the original bimodal mixed signal; the pulse feature integrity of the discharge signal is preserved through this algorithm. From the structure-related time domain map of the vibration signal, the instantaneous amplitude peak value , vibration energy root mean square value ( M is the number of sampling points, is the amplitude of the mth sampling point) and other time-frequency domain features are extracted by Hilbert-Huang transform (HHT), and the frequency domain features are corrected using the device structure modal parameters. The correction formula is ( is the corrected frequency domain feature, is the measured frequency domain feature, is the nominal frequency of the signal), which enhances the correlation between the vibration features and the device structure.

[0077] S23: The improved kernel principal component analysis (KPCA) algorithm is used to reduce the dimension of the bimodal features. KPCA maps high-dimensional features to kernel space through a kernel function ( is a mapping function), and , are the feature vectors of the i and j samples (corresponding to the extracted features of the bimodal signal), respectively. The eigenvalue problem of the covariance matrix is ( K is the kernel matrix, The principal components with a cumulative contribution rate of 90% or more are selected to construct a low-dimensional feature matrix fusing the space-time characteristics of discharge and vibration, and the dimension of the feature matrix is reduced from 20 dimensions or more to 5-8 dimensions, so as to retain the differential characteristics of different defect types and eliminate feature redundancy, thereby providing efficient input for subsequent correlation degree calculation.

[0078] As shown in Figure 3 the flow chart of the three-dimensional electromagnetic simulation model construction and optimization of the application.

[0079] S3: Combined with three-dimensional electromagnetic simulation and improved IPSO algorithm, a GIS device space-time topology constraint model is constructed and the correlation degree of double-mode signals is calculated.

[0080] The traditional double-mode signal correlation degree calculation relies on empirical formula or simple statistical method, and there are problems of not fully considering the space-time topology characteristics of GIS device and single correlation degree calculation dimension, which leads to inaccurate reflection of the real correlation relationship between double-mode signals. In this step, a high-precision three-dimensional electromagnetic simulation model containing GIS cavity, conductor, insulator and other structures is established by means of three-dimensional electromagnetic simulation software, the propagation path, attenuation law and space-time distribution characteristics of partial discharge signal and vibration signal in the device are simulated, and a GIS device space-time topology constraint model is constructed. Then, according to the features of discharge signal such as propagation time delay and amplitude attenuation, and the features of vibration signal such as time domain peak value and frequency domain center frequency, an improved mutual information and cosine similarity fusion algorithm is introduced, combined with the physical law of signal propagation in the space-time topology constraint model, the physical law includes the attenuation coefficient of medium to signal and the influence of spatial distance of different position sensors on propagation time, the correlation degree between double-mode signal features is calculated. The problems of inaccurate double-mode signal correlation degree calculation and low reliability of subsequent positioning and diagnosis basis caused by lack of device space-time topology constraint and single feature correlation dimension in traditional method are solved, and a reliable double-mode correlation basis is provided for accurate positioning and diagnosis.

[0081] The step S3 includes the following sub-steps:

[0082] S31: Based on the specific physical structure parameters of GIS device, including cavity size error not more than 2mm, conductor layout deviation not more than 1mm and insulating medium distribution uniformity error within 5%, a three-dimensional electromagnetic model containing metal shell, SF6 gas medium with purity not less than 99.9% and pressure in the range of 0.4 to 0.6 megapascal, internal conductor and insulator is established.

[0083] The model is simulated by a finite element method, the meshing precision of which is less than 1 mm, and the time step is not more than 1 ns, to simulate the propagation path, attenuation law and boundary reflection characteristics of the partial discharge electromagnetic signal inside the equipment, and output the signal propagation characteristic parameters with an error of no more than 2% compared with the theoretical value and a wave speed ratio deviation of no more than 3% compared with the theoretical value, thereby providing an electromagnetic propagation theoretical basis for the space-time topology constraint.

[0084] Meanwhile, a parameter optimization framework based on an improved particle swarm algorithm (IPSO) is constructed. The framework is based on the signal propagation characteristics output by the three-dimensional electromagnetic model and the sensor coordinates with a positioning accuracy of less than or equal to 1 mm, and initializes a particle population with a size of 30 to 50, each particle corresponding to a to-be-optimized parameter vector in the bimodal signal correlation degree calculation model , which contains time-space weight coefficients, feature matching thresholds and other parameters, and the dimension n is between 6 and 10. The algorithm sets the maximum number of iterations to 100 to 200, and sets the learning factors c1 and c2 to 2.0, and the inertia weight is set to a strategy of linearly decreasing from 0.9 to 0.4.

[0085] S32: Defining the bimodal signal correlation degree as the fitness function , the core formula of which is:

[0086] (2) ;

[0087] wherein is the time-space weight coefficient determined by the particle parameter , and the value range is [0, 1], is the time domain similarity of the discharge signal feature vector and the vibration signal feature vector .

[0088] The time domain similarity is calculated by the dynamic time warping algorithm, and the warping path constraint window width is ≤10% of the signal length, and the calculation formula is:

[0089] (3) ;

[0090] is the frequency domain matching degree of the two, and DTW is the abbreviation of dynamic time warping, which physically means an algorithm for measuring the similarity between two time series of different lengths: by elastically aligning the time series, the minimum cumulative distance between the two sequences is calculated to quantify the shape matching degree of the sequences. Based on the cross power spectral density function, the formula is as follows:

[0091] (4);

[0092] wherein, (1) is the correlation degree of the discharge signal and the vibration signal, is the cross power spectral density, , and (2) and (3) are the auto power spectral densities of the discharge signal and the vibration signal, respectively.

[0093] By improving the position update formula of the particle swarm algorithm:

[0094] (5);

[0095] wherein is the inertia weight, is the learning factor, is a random number in the interval [0, 1], is the velocity vector of the i-th particle at the k-th iteration, representing the current motion trend of the particle in the parameter optimization process, and corresponds to the parameter update step of the bimodal signal correlation degree model; is the velocity vector of the i-th particle at the k+1-th iteration; is the position vector of the i-th particle at the k-th iteration, corresponding to a set of parameter combinations to be optimized of the bimodal signal correlation degree model; is the position vector of the i-th particle at the k+1-th iteration; is the individual optimal position, is the global optimal position, the optimization weight coefficient and the feature matching threshold value, so that the average error of the correlation degree calculation is ≤5%, and the accuracy of the correlation degree calculation is improved.

[0096] S33: When the number of iterations reaches the preset upper limit of 150 times, or the change rate of the fitness function value of 5 consecutive iterations is less than or equal to the threshold value of 1x10 -4 , the algorithm terminates the iteration and outputs the optimal parameter combination , so as to complete the parameter configuration of the correlation degree calculation model. It has been verified that under this configuration, the coincidence degree of the correlation degree calculation result and the actual physical correlation is not less than 90%, which provides a reliable quantitative basis for the spatio-temporal correlation analysis of the bimodal signal. Through the analysis of the model, the spatial propagation law and energy distribution characteristics of the partial discharge signal under the GIS spatio-temporal topological constraint can be intuitively presented, as shown in Figure 5 , which clearly reflects the power density difference of the partial discharge signal under different azimuth angles and elevation angles, providing visual signal feature support for subsequent effective signal screening and defect position range determination.

[0097] S4: Screening of effective signals and preliminary determination of defect position range through correlation degree threshold screening and IPSO constraint fusion.

[0098] As shown in Figure 4As shown, it is a smart defect positioning and decision-making flowchart driven by the bimodal mode of the application. The existing defect positioning method depends on a single signal or a simple multi-signal fusion rule, and there are problems of inaccurate signal effectiveness judgment, insufficient combination of device space-time topology characteristics and bimodal signal correlation, and insufficient accuracy of defect position range determination, which is difficult to accurately narrow the position range of GIS partial discharge fault.

[0099] This step is based on the correlation calculation model and GIS device space-time topology constraint model constructed in S3 to effectively screen the collected discharge and vibration bimodal signals: calculate the correlation of each sensor collected signal and bimodal signal, set the correlation threshold to 0.7, and screen out effective signals with a correlation higher than the threshold; combined with the space-time topology constraints of GIS equipment in signal propagation, such as path restrictions and cavity boundary reflections of signal propagation physical characteristics, and using multi-source signal positioning methods based on signal arrival time difference and amplitude difference basic physical principles, the characteristics of effective signals and their correlation information can be fused, and the range of defect position can be determined, providing a reliable basis for subsequent accurate positioning.

[0100] Step S4 includes the following sub-steps:

[0101] S41: combined with the GIS device space-time topology model constructed by three-dimensional electromagnetic simulation, the model cavity size error is not more than 2mm, the conductor layout deviation is not more than 1mm, and according to the bimodal signal correlation calculation result with an error of less than 5%, the signal effectiveness judgment threshold is set to 0.75. A large number of experiments have shown that under this threshold, the recognition accuracy of effective signals can reach more than 90%. Based on this threshold, the system can screen out effective discharge signals and vibration signals closely related to the defect source space-time, and successfully eliminate invalid signals disturbed or irrelevant, so as to control the proportion of invalid signals to less than 10%.

[0102] S42: use the propagation time delay characteristics of effective discharge signals, which are according to the formula:

[0103] (6) ;

[0104] Wherein, is the time delay of discharge signal from sensor i to sensor j, is the spatial distance of sensors i and j, is the propagation speed of the signal in the GIS internal medium, which is about 2×10 8 m / s, error ≤2%.

[0105] At the same time, combined with the spatial structure parameters of GIS equipment, the spatial structure parameters include cavity length, conductor spacing, etc., with an accuracy of ≤1mm, the spatial coordinates The spatial constraint condition of the discharge signal source is represented.

[0106] The constraint condition is represented as:

[0107] (7);

[0108] wherein, is the theoretical time delay of the signal generated by the partial discharge defect with coordinates (x, y, z) to the i-th sensor; is the theoretical time delay of the signal of the same defect to the j-th sensor; is the measured time delay difference, is the time delay measurement error, ≤1 ns;

[0109] Meanwhile, based on the propagation characteristics of the vibration signal, the propagation characteristics are that the vibration wave propagates at a speed of about 5000 m / s in the GIS metal shell, with an error of ≤3%, and the inherent frequency measurement error of the equipment structure mode is ≤1%, the spatial constraint condition of the vibration signal source is constructed, which is also represented by the spatial coordinates , and the constraint condition is . is the measured time difference of the vibration signal to different sensors, is the time difference measurement error, is the theoretical time delay of the vibration signal generated by the defect with coordinates (x, y, z) to the i-th sensor, is the theoretical time delay of the vibration signal of the same defect to the j-th sensor, .

[0110] S43: Fuse the spatial constraint conditions of the discharge and vibration signals, and solve the possible existence area of the defect satisfying the space-time correlation of the double-mode signals by using an improved particle swarm optimization (IPSO) algorithm. In the particle swarm optimization algorithm, the particle position represents the spatial coordinates of the defect , and the fitness function is the degree of satisfying the spatial constraint conditions of the discharge and vibration signals, and the fitness function formula is:

[0111] (8);

[0112] wherein, is a weight coefficient, satisfying , and ; is the calculated discharge signal time delay difference, is the measured discharge signal time delay difference; is the calculated vibration signal time difference, is the measured vibration signal time difference.

[0113] By setting the population size as 40, the maximum number of iterations as 150, the learning factors c1 and c2 as 2.0, and adopting the iterative optimization strategy of linearly decreasing the inertia weight from 0.9 to 0.4, when the algorithm meets the convergence condition that the change rate of the fitness function value of 5 consecutive iterations is not more than 1x10 -4 The error of the preliminary positioning result in space can be controlled within 5mm, which provides a reliable foundation for the subsequent accurate calculation of defect coordinates.

[0114] S5: applying the improved WLS and WOA collaborative algorithm, iteratively calculating the defect coordinates and outputting the three-dimensional visualization result of GIS partial discharge fault tracing positioning.

[0115] The existing defect positioning methods mostly rely on a single algorithm or do not fully integrate multi-modal signal features, and have the problems of poor algorithm robustness, insufficient combination of spatial and temporal topological constraints and dual-modal signal correlation, resulting in insufficient accuracy of defect coordinate calculation, and it is difficult to accurately output the three-dimensional positioning result of GIS partial discharge fault.

[0116] As shown in Figure 6 , the three-dimensional fault positioning imaging flowchart based on the improved optimization positioning algorithm and GIS geometric model of the present application.

[0117] This step is based on the correlation calculation model and GIS device spatial and temporal topological constraint model constructed in S3. For the effective discharge and vibration signals after the preliminary determination range of S4, the improved weighted centroid multi-source positioning algorithm is used: the correlation of the dual-modal signals is used to assign weights to each sensor signal, and the principle is that the higher the correlation, the greater the weight, and the weight value range is 0.5 to 1.0. Then, the time difference and amplitude of the signals are used, and the GIS device spatial and temporal topological parameters (including cavity geometric size, sensor spatial position, etc.) with an accuracy of not less than 1mm are fused, and the three-dimensional coordinates of the defect are finally determined through iterative calculation; finally, based on the three-dimensional model of the GIS device, the calculated defect coordinates are three-dimensionally visualized and presented, and the position of the partial discharge fault is intuitively displayed, providing accurate spatial positioning basis for fault tracing and maintenance of the GIS device.

[0118] The step S5 includes the following sub-steps:

[0119] S51: Taking the defect location range preliminarily determined in step S4 as a constraint boundary, the range is a cuboid region with a spatial error ≤5 mm, and the specific dimensions of the length, width and height are determined according to the GIS device model and the previous analysis, such as a range of 200 mm in length, 150 mm in width and 100 mm in height in a certain model of GIS. Within this range, the time difference of the discharge pulse reaching different sensors in 8-12 groups of effective dual-mode signals is extracted, with a measurement accuracy ≤1 ns, and the path difference characteristics of the vibration signal propagation are extracted, with a calculation error not exceeding 2 mm. Finally, combined with the three-dimensional coordinates of the sensor itself with a positioning accuracy better than 1 mm, a reference data set for precise positioning is constructed.

[0120] S52: An improved weighted least squares (WLS) algorithm is used to construct a positioning model, taking the defect coordinates as the variables to be solved, combining the parameters of the discharge signal propagation speed in the SF6 gas inside the GIS (error ≤2%) and the vibration signal propagation speed in the metal shell (error ≤3%) to establish the objective function:

[0121] (9);

[0122] (10);

[0123] wherein, is the weight coefficient based on the correlation degree calculation of the dual-mode signals, is the three-dimensional spatial coordinate component of the GIS device local discharge fault defect to be solved; is the X-axis coordinate value of the i-th sensor in the three-dimensional coordinate system; is the Y-axis coordinate value of the i-th sensor in the three-dimensional coordinate system; is the Z-axis coordinate value of the i-th sensor in the three-dimensional coordinate system; is the signal propagation speed, is the time when the signal reaches the i-th sensor; the higher the correlation degree , the greater the weight, The value range of is [0.6, 1.0], is selected according to the type of signal, , is the time when the signal reaches the i-th sensor (measurement accuracy ≤1 ns).

[0124] S53: introduce whale optimization algorithm (WOA) for iterative solution, set the population size to 30-50, and the maximum iteration number to 100-150. The position range determined in step S4 is the initial search space, and the position update formula is obtained by simulating the surrounding hunting behavior of humpback whale hunting:

[0125] (11) ;

[0126] wherein decreases linearly from 2 to 0, is a random number in the interval [0, 1], is the current optimal solution: bubble net attack mechanism updates the candidate defect coordinates.

[0127] The position update formula of the bubble net attack mechanism is: (12) ;

[0128] In the iteration process, in combination with the GIS device spatial topological constraint with an accuracy of not less than 1mm, the spatial topological constraint includes cavity boundary restriction and wall coordinates, the invalid solution is pruned. As shown in Figure 7 When the target function value is less than or equal to the set threshold of 1x10 -6 , or the maximum iteration number has been reached, the iteration process is terminated, and the optimal defect coordinates with a positioning error of not more than 3mm are finally output. The GIS partial discharge fault tracing and positioning result realized accordingly can be output through a three-dimensional visual model, and the geometric error between the model and the actual GIS structure is controlled within 2mm.

[0129] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited thereto, any person skilled in the art can easily think of changes or replacements within the technical range disclosed by the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A GIS vibration and discharge dual-mode source visualization positioning method, characterized in that, The method comprises the following steps: S1: build a multi-dimensional dual-mode collaborative acquisition platform and simulate dynamic working conditions, and synchronously acquire GIS device partial discharge signals and vibration signals; S2: adopt a VMD-HHT-KPCA multi-algorithm fusion strategy to extract discharge signal propagation time delay features and vibration signal time-frequency domain features; S3: combine three-dimensional electromagnetic simulation and an improved IPSO algorithm to construct a GIS device space-time topology constraint model and calculate dual-mode signal correlation degrees; S4: screen effective signals and preliminarily determine a defect position range through correlation degree threshold screening and IPSO constraint fusion; S5: use an improved WLS and WOA collaborative algorithm to iteratively calculate defect coordinates and output GIS partial discharge fault tracing positioning three-dimensional visualization results.

2. The GIS vibration and discharge dual-mode traceable visualization positioning method according to claim 1, characterized in that, The step S1 comprises the following sub-steps: S11: build a multi-dimensional dual-mode collaborative acquisition test platform, select a GIS test sample machine carrying gradient composite insulation defects, configure a UHF sensor to acquire partial discharge UHF electromagnetic signals, deploy a fiber bragg grating vibration sensor array to acquire vibration signals, cooperate with a pulse generator, a high-speed oscilloscope, a time scale synchronization device and a multi-channel data acquisition device to construct a partial discharge-vibration dual-mode signal acquisition system, and realize high-sensitivity capture of UHF discharge electromagnetic signals and vibration displacement signals; S12: debug GIS test sample machine operating parameters, adopt a gradient voltage boosting combined with partial discharge pulse triggering voltage adjustment mode, gradually boost voltage from an initial voltage, when a UHF sensor detects a partial discharge UHF pulse, perform small amplitude fluctuation voltage adjustment based on the voltage, simulate partial discharge and vibration coupling working conditions under actual power grid voltage fluctuation, and create dual-mode signal acquisition working conditions close to actual operation scenarios; S13: start a data acquisition device, realize time synchronization of partial discharge UHF signals and fiber vibration signals by means of a time scale synchronization device, continuously synchronously acquire partial discharge original UHF signals and vibration original fiber bragg grating signals according to an adaptive acquisition period for each gradient composite insulation defect, acquire multiple samples for each defect, and form a dual-mode signal sample library containing gradient composite defect type labels, working condition voltage information and signal space-time correlation features.

3. The GIS vibration and discharge dual-mode traceable visualization positioning method according to claim 2, characterized in that, The composite insulation defects in the step S11 include combined defects of conductor spikes and metal particles and combined defects of insulator surface particles and dynamic suspended discharge.

4. The GIS vibration and discharge dual-mode traceable visualization positioning method according to claim 1, characterized in that, The step S2 comprises the following sub-steps: S21: for the synchronously acquired partial discharge signals and vibration signals, combine the spatial position information of sensors in a GIS device space-time topology model to calculate medium difference time delay differences of discharge signals propagating to different sensors; meanwhile, take a power frequency period as a time axis to count discharge pulse amplitude distribution, generate an enhanced PRPD spectrum containing phase-amplitude-frequency information, and for vibration signals, construct a time domain amplitude sequence, combine structure mode related parameters of sensor installation positions to generate a structure correlation time domain spectrum of vibration signals; S22: Extract the discharge signal features of the discharge pulse from the enhanced PRPD pattern, the discharge signal features being maximum amplitude, average discharge amount, and positive and negative half-cycle pulse asymmetry, and perform noise reduction processing on the discharge signal features by using a variational mode decomposition algorithm to retain the integrity of the pulse features of the discharge signal; From the structural correlation time-domain pattern of the vibration signal, extract time-frequency domain features by using a Hilbert-Huang transform, the time-frequency domain features including instantaneous amplitude peak value and vibration energy root mean square value, and correct the frequency domain features by using device structural modal parameters to enhance the correlation between the vibration features and the device structure; S23: Perform dimension reduction processing on the dual-modal features by using an improved kernel principal component analysis algorithm, map the high-dimensional features to a kernel space by using a kernel function, select principal components whose cumulative contribution rate meets the requirements, construct a low-dimensional feature matrix that fuses discharge-vibration time-space features, retain the differential features of different defect types and eliminate feature redundancy, and provide efficient input for subsequent correlation degree calculation.

5. The GIS vibration and discharge dual-mode traceable visualization positioning method according to claim 1, characterized in that, The step S3 includes the following sub-steps: S31: Based on the physical structure parameters of the GIS device, a three-dimensional electromagnetic model containing the metal shell, SF6 gas medium, internal conductor and insulator is established, the propagation path, attenuation law and boundary reflection characteristics of the partial discharge electromagnetic signal in the device interior are simulated by the finite element method, and the electromagnetic propagation theoretical basis for the space-time topology constraint is provided; at the same time, a parameter optimization framework based on an improved particle swarm algorithm is constructed, and the signal propagation characteristics output by the three-dimensional electromagnetic model and the sensor coordinates are used as the basis to initialize the particle population, and each particle corresponds to the optimized parameter vector of the bimodal signal correlation degree calculation model , point particle parameters , and so on; at the same time, the basic parameters including the population size, the maximum number of iterations, and the learning factor are set. S32: defining the bimodal signal correlation degree as the fitness function The core formula is: ； wherein is the spatio-temporal weight coefficient determined by the particle parameters , is the discharge signal feature vector and the vibration signal feature vector , is the time-domain similarity between the two, calculated using the dynamic time warping algorithm, is the frequency-domain matching degree between the two, solved based on the cross-power spectral density function, is the discharge signal feature vector used to calculate the frequency-domain matching degree, is the vibration signal feature vector used to calculate the frequency-domain matching degree. The position updating formula of the improved particle swarm algorithm is as follows: ； wherein is an inertia weight, is the velocity vector of particle i at the kth iteration, is the velocity vector of particle i at the k+1th iteration, used to describe the position update trend of the particle in the parameter optimization process; is a learning factor, is a random number in the range of [0, 1], which increases the randomness of the algorithm and enhances the diversity of the search, is the individual optimal position, is the position vector of particle i at the kth iteration, is the global optimal position, which optimizes the weight coefficient and the feature matching threshold, and improves the accuracy of the correlation degree calculation, is the position vector of particle i at the k+1th iteration, which is the new position obtained after the velocity is updated; S33: When the number of iterations reaches a set value or the rate of change of the fitness function value of the continuous 5 iterations is less than the threshold value, terminate the iteration, and output the optimal parameter combination , complete the parameter configuration of the correlation degree calculation model, and provide a quantitative basis for the spatio-temporal correlation analysis of the dual-mode signal.

6. The GIS vibration and discharge dual-mode traceable visualization positioning method according to claim 5, characterized in that, The physical structure parameters in the step S31 include cavity size, conductor layout, and insulating medium distribution.

7. The GIS vibration and discharge dual-mode traceable visualization positioning method according to claim 1, characterized in that, The step S4 includes the following sub-steps: S41: In combination with the time-space topology model of the GIS device, set a signal validity determination threshold according to the correlation degree calculation result of the dual-modal signals, select effective discharge signals and vibration signals that are closely associated with the time-space of the defect source, and eliminate invalid signals that are disturbed or irrelevant; S42: Construct the spatial constraint condition of the discharge signal source by using the propagation time delay features of the effective discharge signal and the spatial structure parameters of the GIS device; at the same time, construct the spatial constraint condition of the vibration signal source based on the propagation characteristics of the vibration signal and the device structure modal; S43: Fuse the spatial constraint conditions of the discharge and vibration signals, solve to obtain a region where the defect is likely to exist and meet the time-space correlation of the dual-modal signals, and preliminarily determine the range of the defect position, thereby providing a basis for subsequent accurate calculation of the defect coordinates.

8. The GIS vibration and discharge dual-mode traceable visualization positioning method according to claim 1, characterized in that, The step S5 includes the following sub-steps: S51: With the defect position range preliminarily determined in step S4 as a constraint boundary, extract the time difference of discharge pulses reaching different sensors and the path difference characteristics of vibration signal propagation in multiple sets of effective bimodal signals, and establish a positioning base data set containing the three-dimensional coordinates of the sensors ; S52: Construct the positioning model by using the improved weighted least square algorithm, and take the defect coordinates as the variables to be solved, combine the propagation speed parameters of the signals in different media, and establish a target function: ； wherein is a weight coefficient based on the correlation degree of the dual-mode signals, is a three-dimensional spatial coordinate component of the GIS device partial discharge fault defect to be solved; is the X-axis coordinate value of the i th sensor in the three-dimensional coordinate system; is the Y-axis coordinate value of the i th sensor in the three-dimensional coordinate system; is the Z-axis coordinate value of the i th sensor in the three-dimensional coordinate system; is the signal propagation speed, is the time when the signal reaches the i th sensor; S53: Introduce whale optimization algorithm for iterative solution, take the position range determined in step S4 as the initial search space, update the candidate defect coordinates through the surrounding predation, bubble net attack and random search mechanism simulating the hunting behavior of humpback whales, combine the spatial topological constraints of GIS equipment to prune invalid solutions in the iteration process, until the target function value is less than the set threshold or the maximum iteration number is reached, and finally output the optimal defect coordinates , and realize three-dimensional visualization output of GIS partial discharge fault tracing and positioning results.

Images