An ant colony optimization energy focused cable insulation damage positioning method and device

Abstract

The application discloses an ant colony optimization energy focusing cable insulation damage positioning method and device, and relates to the technical field of cable detection engineering, wherein the method comprises the following steps: generating a Gaussian envelope linear frequency modulation chirp detection signal and injecting the detection signal into a cable to be detected, collecting an incident signal and a return signal containing a reflection component; constructing a parameterized time-frequency analysis kernel function family under a Cohen time-frequency distribution framework, searching and optimizing the kernel parameters by using an ant colony optimization algorithm; calculating the time-frequency distribution and energy-time curve of the signal based on the optimized energy focusing time-frequency analysis kernel function, extracting the arrival time corresponding to the energy peak value, and realizing cable insulation damage positioning according to the reflection time delay and the propagation wave velocity. The application enhances the energy peak value focusing and peak position stability while suppressing the cross terms, significantly improves the defect positioning accuracy and robustness, and is suitable for engineering online detection and automatic implementation of cable insulation damage.

Description

Technical Field

[0001] This invention relates to the field of cable testing engineering technology, and in particular to an ant colony-optimized energy focusing method and device for locating cable insulation damage. Background Technology

[0002] Power cables and various transmission lines are critical infrastructure for energy and information transmission. During long-term operation, they are highly susceptible to localized defects or impedance discontinuities due to factors such as mechanical damage, chemical corrosion, moisture absorption, and insulation aging. Failure to detect and locate these defects promptly and accurately will affect power transmission efficiency and, in severe cases, may lead to line faults or even safety accidents. Therefore, researching high-precision cable defect location technology is of great significance for ensuring the safe and stable operation of the system.

[0003] Among various detection techniques, the time-frequency domain reflection method has significant advantages in combating noise interference and compensating for transmission line frequency dispersion effects because it can simultaneously provide joint time-domain and frequency-domain information of the signal. This method typically injects the detection signal into the cable under test and extracts the time delay characteristics characterizing the defect location by calculating and analyzing the time-frequency distribution of the echo signal. However, due to the significant attenuation of electromagnetic wave signals during transmission in the cable, the reflection method requires an accurate time-frequency energy distribution algorithm to distinguish the reflected wave peak value from electromagnetic interference.

[0004] In the calculation of time-frequency distributions, Cohen-type bilinear time-frequency distributions are the core tool for constructing signal energy characterizations. While the Wigner-Vell distribution (WVD) theoretically possesses high time-frequency clustering, its quadratic transform characteristics mean that when processing multi-component signals containing incident and reflected waves, it generates high-amplitude cross-term interference between the true signal terms. These spurious cross-terms often have high energy, easily confusing the true signal characteristics and leading to incorrect reflection peak identification.

[0005] To address the cross-term interference problem, existing techniques typically introduce different types of kernel functions to smooth and filter WVD. These kernel functions filter out cross-term components far from the origin by weighting the signal within the fuzzy domain. However, in practical applications, the design and parameter selection of existing kernel functions present an irreconcilable technical contradiction between "suppressing cross-terms" and "preserving self-term energy resolution."

[0006] On the one hand, if a kernel function with strong smoothing effect is selected, although it can effectively suppress cross-term interference, it will inevitably destroy the time-frequency structure of the signal term, causing the energy of the real signal to diffuse and leak in the time-frequency plane. This decrease in energy resolution is directly manifested as a widening of the time-frequency ridge and a blunting of the energy peak, reducing the extraction accuracy of the incident and reflection times.

[0007] On the other hand, if the smoothing effect of the kernel function is insufficient, it cannot completely eliminate cross terms, and residual spurious peaks may mask weak far-end reflection signals, leading to missed detections or misjudgments. In addition, traditional kernel functions usually use fixed mathematical forms and parameters, which cannot be adaptively adjusted according to different cable conditions and signal-to-noise ratio environments, making it difficult to achieve optimal focusing of incident and reflected signal energy. Summary of the Invention

[0008] The main objective of this invention is to provide an ant colony-optimized energy focusing cable insulation damage localization method. By optimizing the ant colony to obtain a higher-performance energy focusing kernel function, the reflected signal energy in time-frequency domain analysis is improved, thereby achieving more accurate cable insulation defect localization.

[0009] Another objective of this invention is to provide an ant colony-optimized energy-focusing cable insulation damage location device.

[0010] The third objective of this invention is to provide an electronic device.

[0011] To achieve the above objectives, a first aspect of the present invention proposes an ant colony-optimized method for locating energy-focusing cable insulation damage, comprising:

[0012] A Gaussian envelope linear frequency modulated chirp detection signal is generated and injected into the cable under test. The incident signal and the echo signal containing the reflected component are collected at the receiving end.

[0013] Within the Cohen-class time-frequency distribution framework, time-frequency transformations are performed on the incident signal and the echo signal respectively. The Born-Jordan kernel is selected as the basic kernel, and a scale factor is introduced to construct a family of parameterized time-frequency analysis kernel functions.

[0014] The search range of the kernel parameters of the parameterized time-frequency analysis kernel function family is set and discretized to construct candidate parameter combinations. The number of ants, the initial value of pheromone, the volatility coefficient and the weight of heuristic factor are initialized for the ant colony optimization algorithm.

[0015] For each set of candidate kernel parameters, the time-frequency distributions corresponding to the incident signal and the echo signal are calculated respectively. Peak detection is performed based on the time-frequency distributions, and feature indicators including peak energy, peak width, sidelobe energy and inter-peak separation are extracted.

[0016] Based on the aforementioned feature indicators, an objective function is constructed that includes a peak enhancement term, an energy leakage penalty term, and an inter-peak separability term, and this objective function is used as the fitness function of the ant colony optimization algorithm.

[0017] The candidate kernel parameters are iteratively searched using an ant colony optimization algorithm. The pheromone is updated according to the fitness function and volatilization is performed to obtain the optimal kernel parameters that maximize the objective function, thereby obtaining the time-frequency analysis kernel function for energy focusing.

[0018] The time-frequency distribution and energy-time curve of the signal are recalculated using the time-frequency analysis kernel function of the energy focusing, the arrival time of the energy peak corresponding to the incident wave and the reflected wave are determined, and the location of cable insulation damage is calculated based on the reflection delay and the wave velocity in the cable.

[0019] Optionally, a Gaussian envelope linear frequency modulated chirp detection signal is generated and injected into the cable under test. The incident signal and the echo signal containing the reflected component are acquired at the receiving end, including:

[0020] Based on the preset time-domain width parameter, time center parameter, frequency center parameter and frequency change rate, a linear frequency-modulated chirp detection signal with a Gaussian envelope is constructed.

[0021] The detection signal is injected into the cable under test through capacitive coupling, inductive coupling, or direct coupling.

[0022] Signals are acquired at the receiving end of the cable under test to obtain an echo signal containing the incident wave component and the reflected wave component caused by the cable impedance discontinuity. The echo signal is then synchronously sampled to form the input signal for subsequent time-frequency analysis.

[0023] Optionally, within the Cohen-type time-frequency distribution framework, time-frequency transformations are performed on the incident signal and the echo signal respectively. The Born-Jordan kernel is selected as the base kernel, and a scaling factor is introduced to construct a family of parameterized time-frequency analysis kernel functions, including:

[0024] Within the framework of Cohen-class bilinear time-frequency distribution, a time-frequency transformation expression for the signal is established, where the kernel function is used to adjust the energy concentration characteristics of the time-frequency distribution;

[0025] The Born-Jordan kernel is selected as the basic kernel function, and the basic kernel is extended by introducing at least one scaling factor or weighting function to form an adjustable family of time-frequency analysis kernel functions.

[0026] Optionally, the search range of the kernel parameters of the parameterized time-frequency analysis kernel function family is set and discretized to construct candidate parameter combinations, and the ant colony optimization algorithm's ant number, initial pheromone value, volatile coefficient, and heuristic factor weights are initialized, including:

[0027] The scaling factor and weighting function parameters are set to have a range of values ​​and are discretized to generate multiple sets of candidate kernel parameter combinations;

[0028] Initialize the number of ants, initial pheromone value, pheromone volatility coefficient, and heuristic factor weights in the ant colony optimization algorithm to guide the search process for kernel parameter combinations.

[0029] Optionally, for each set of candidate kernel parameters, the time-frequency distributions corresponding to the incident signal and the echo signal are calculated respectively. Peak detection is performed based on the time-frequency distributions, and feature indicators including peak energy, peak width, sidelobe energy, and inter-peak separation are extracted, including:

[0030] Based on the corresponding candidate kernel parameters, the time-frequency distributions of the incident signal and the echo signal are calculated respectively.

[0031] Peak detection is performed on the time-frequency distribution, and the peak energy, peak width, sidelobe energy, and inter-peak time interval or separation degree corresponding to each major peak are extracted as characteristic indicators to measure the performance of time-frequency energy aggregation and separation.

[0032] Optionally, an objective function comprising a peak enhancement term, an energy leakage penalty term, and an inter-peak separability term is constructed based on the aforementioned feature indicators, and the objective function is used as the fitness function of the ant colony optimization algorithm, including:

[0033] A peak enhancement index is constructed based on the peak energy and peak width;

[0034] An energy leakage penalty index is constructed based on the sidelobe energy and the energy distribution outside the peak window;

[0035] Construct an inter-peak separability index based on the time interval or separability between adjacent peaks;

[0036] The peak enhancement index, energy leakage penalty index, and inter-peak separability index are weighted to construct an objective function for evaluating the merits of kernel parameter combinations, and the objective function is used as the fitness function of the ant colony optimization algorithm.

[0037] Optionally, the candidate kernel parameters are iteratively searched using an ant colony optimization algorithm, the pheromone is updated according to the fitness function and volatilization is performed to obtain the optimal kernel parameters that maximize the objective function, thereby obtaining the time-frequency analysis kernel function for energy focusing, including:

[0038] Based on the current pheromone distribution and heuristic factors, determine the state transition probability of the ant in the kernel parameter combination space;

[0039] Based on the fitness function, the selected kernel parameter combinations are incrementally updated with pheromones, and the unselected kernel parameter combinations are pheromone evaporation is performed.

[0040] Repeat the state transition and pheromone update process until the preset termination condition is met, and output the kernel parameter combination that makes the objective function reach the optimal value to obtain the time-frequency analysis kernel function for energy focusing.

[0041] Optionally, the time-frequency distribution and energy-time curve of the signal are recalculated using the time-frequency analysis kernel function of the energy focusing, the arrival times of the energy peaks corresponding to the incident and reflected waves are determined, and the location of cable insulation damage is calculated based on the reflection time delay and wave velocity in the cable, including:

[0042] The time-frequency distribution of the signal is recalculated using the energy-focused time-frequency analysis kernel function, and the corresponding energy-time curve is obtained.

[0043] Determine the arrival time of the energy peak corresponding to the incident wave and at least one reflected wave based on the energy-time curve, and calculate the reflection time delay;

[0044] Based on the reflection delay and the propagation wave velocity in the cable, calculate the location of cable insulation damage corresponding to each reflection peak.

[0045] To achieve the above objectives, a second aspect of the present invention provides an ant colony-optimized energy-focusing cable insulation damage location device, comprising:

[0046] The first module is used to generate a Gaussian envelope linear frequency modulated chirp detection signal and inject the detection signal into the cable under test, and collect the incident signal and the echo signal containing the reflected component at the receiving end.

[0047] The second module is used to perform time-frequency transformation on the incident signal and the echo signal respectively under the Cohen-type time-frequency distribution framework, select the Born-Jordan kernel as the basic kernel, and introduce a scale factor to construct a family of parameterized time-frequency analysis kernel functions;

[0048] The third module is used to set the search range and discretize the kernel parameters of the parameterized time-frequency analysis kernel function family, construct candidate parameter combinations, and initialize the number of ants, initial value of pheromone, volatility coefficient and heuristic factor weight of the ant colony optimization algorithm.

[0049] The fourth module is used to calculate the time-frequency distribution of the incident signal and the echo signal for each set of candidate kernel parameters, perform peak detection based on the time-frequency distribution, and extract feature indicators including peak energy, peak width, sidelobe energy and inter-peak separation.

[0050] The fifth module is used to construct an objective function containing a peak enhancement term, an energy leakage penalty term, and an inter-peak separability term based on the feature index, and to use the objective function as the fitness function of the ant colony optimization algorithm.

[0051] The sixth module is used to iteratively search the candidate kernel parameters using an ant colony optimization algorithm, update the pheromone according to the fitness function and perform volatilization, to obtain the optimal kernel parameters that maximize the objective function, thereby obtaining the time-frequency analysis kernel function for energy focusing;

[0052] The seventh module is used to recalculate the time-frequency distribution and energy-time curve of the signal using the time-frequency analysis kernel function of the energy focusing, determine the arrival time of the energy peak corresponding to the incident wave and the reflected wave, and calculate the location of cable insulation damage based on the reflection time delay and the wave velocity in the cable.

[0053] To achieve the above objectives, a third aspect of this application provides an electronic device, including a processor and a memory; wherein the processor runs a program corresponding to the executable program code stored in the memory to implement the method described in the first aspect.

[0054] The embodiments of the present invention have the following beneficial effects: while suppressing cross terms, they enhance energy peak focusing and peak position stability, significantly improve defect location accuracy and robustness, and are suitable for engineering online detection and automation of cable insulation damage. Attached Figure Description

[0055] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:

[0056] Figure 1 A simplified text flowchart illustrating an ant colony-optimized method for locating energy-focusing cable insulation damage, provided as an embodiment of the present invention;

[0057] Figure 2 This is a deconstructed text flowchart of an ant colony-optimized method for locating energy-focusing cable insulation damage according to an embodiment of the present invention;

[0058] Figure 3 This is a time-frequency domain schematic diagram of an ant colony optimization energy focusing kernel function according to an embodiment of the present invention;

[0059] Figure 4 This is a schematic diagram illustrating the positioning capability of an ant colony-optimized energy-focusing cable insulation damage location method according to an embodiment of the present invention. Detailed Implementation

[0060] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0061] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0062] The following description, with reference to the accompanying drawings, describes an ant colony-optimized method and apparatus for locating energy-focusing cable insulation damage according to an embodiment of the present invention.

[0063] Example 1

[0064] This embodiment provides an ant colony-optimized method for locating energy-focusing cable insulation damage. For example... Figure 1 and Figure 2 As shown, the method includes the following steps:

[0065] S1, Generate a Gaussian envelope linear frequency modulated chirp detection signal and inject the detection signal into the cable under test, and collect the incident signal and the echo signal containing the reflection component at the receiving end.

[0066] In this embodiment, the detection signal is a Gaussian envelope linear frequency modulated chirped signal. Specifically, based on preset time-domain width parameters, time center parameters, frequency center parameters, and frequency change rate, a chirped signal with a Gaussian distribution in the time domain and linear frequency modulation characteristics in the frequency domain is constructed. By reasonably configuring the above parameters, sufficient frequency coverage can be ensured while maintaining signal energy concentration, thereby improving the excitation and response capability to local impedance discontinuities such as cable insulation damage. Compared with single-frequency or narrowband pulse signals, this type of detection signal has advantages such as strong noise immunity, controllable time-frequency characteristics, and suitability for subsequent time-frequency analysis and processing.

[0067] After generating the Gaussian envelope linear frequency modulated chirped detection signal, the detection signal is injected into the cable under test. In this embodiment, the injection method of the detection signal can be flexibly selected according to the actual application scenario. For example, capacitive coupling can be used to couple the high-frequency detection signal to the cable conductor without contact, or inductive coupling can be used to achieve signal injection through magnetic field induction, or, under safe conditions, direct coupling can be used to directly load the detection signal onto the cable port. All of the above injection methods can achieve effective propagation of the detection signal in the cable, and this application does not limit this method.

[0068] When the detection signal propagates along the cable under test, if there are changes in impedance due to insulation aging, local damage, or structural defects inside the cable, the detection signal will generate a reflected wave component at the corresponding location. In this embodiment, the propagating signal is acquired at the receiving end of the cable. The acquired signal simultaneously includes the incident wave component that propagates normally along the cable and one or more reflected wave components caused by impedance discontinuities, thereby forming an echo signal containing reflection information.

[0069] Furthermore, to ensure the accuracy of subsequent time-frequency analysis and energy focusing processing, in this embodiment, the echo signal is synchronously sampled to provide a unified sampling reference on the time axis, thereby forming input signal data for subsequent time-frequency transformation, peak detection, and damage localization calculation. Through the above steps, this application lays a reliable signal foundation for subsequent energy focusing analysis based on time-frequency distribution and ant colony optimization.

[0070] S2, within the Cohen-class time-frequency distribution framework, time-frequency transformations are performed on the incident signal and the echo signal respectively. The Born-Jordan kernel is selected as the basic kernel, and a scale factor is introduced to construct a family of parameterized time-frequency analysis kernel functions.

[0071] In one specific embodiment of this application, step S2 is performed, in which the incident signal and the echo signal are transformed in time and frequency under the Cohen-type time-frequency distribution framework, and an adjustable family of parameterized time-frequency analysis kernel functions is constructed for subsequent energy focusing and feature extraction.

[0072] In this embodiment, the Cohen-type time-frequency distribution belongs to a class of bilinear time-frequency analysis methods. By jointly representing the signal in both time and frequency dimensions, it can simultaneously characterize the time-domain and frequency-domain characteristics of the signal. Compared with traditional single time-domain or frequency-domain analysis methods, the Cohen-type time-frequency distribution has higher resolution in analyzing non-stationary signals, and is particularly suitable for analyzing echo signals with multiple reflections and time-varying frequencies during cable inspection.

[0073] Specifically, in the embodiments of this application, for a continuous-time signal x(t), its corresponding Cohen class time-frequency distribution can be expressed as:

[0074]

[0075] Where t represents the time variable, ω represents the angular frequency variable, and u, θ, and τ are integration variables. For kernel function, superscript Indicates conjugate. It is the imaginary unit. Kernel functions play a crucial role in Cohen-type time-frequency distributions. They are mainly used to suppress cross terms, adjust the smoothness of the time-frequency distribution, and control the energy concentration characteristics in the time-frequency plane. Different forms of kernel functions will directly affect the resolution and anti-interference ability of the time-frequency analysis results.

[0076] In this embodiment, to balance energy concentration and cross-term suppression performance, the Born-Jordan kernel is selected as the basic kernel function. The Born-Jordan kernel has a sound theoretical basis in time-frequency analysis, and it can suppress cross-terms caused by multi-component signals to a certain extent while maintaining good time-frequency resolution. In this embodiment, the basic kernel function can be expressed as follows: Where 'a' is a scale-dependent parameter used to adjust the kernel function. The range of support on the plane.

[0077] Furthermore, to enhance the adaptability of the time-frequency distribution to different signal characteristics, in this embodiment, instead of directly using a single Born-Jordan kernel, the basic kernel function is parametrically extended by introducing a scaling factor and an additional weighting function. Specifically, a scaling factor 'a' and a weighting function are introduced on top of the Born-Jordan kernel. Construct parameterized kernel function form Different scale factors and weighting function parameters correspond to different kernel function forms, thus forming an adjustable family of time-frequency analysis kernel functions.

[0078] In this embodiment, the parameterized time-frequency analysis kernel function family constructed in the above manner allows the energy concentration, time resolution, and frequency resolution of the time-frequency distribution to be adjusted according to the actual signal characteristics, providing a foundation for subsequent adaptive search of kernel parameters using ant colony optimization algorithms. Simultaneously, by performing time-frequency transformations on the incident and echo signals separately within the same time-frequency analysis framework, it can be ensured that the analysis results of both have a consistent reference scale, which is beneficial for subsequent accurate calculations of peak alignment, energy comparison, and reflection delay.

[0079] S3, set the search range for the kernel parameters of the parameterized time-frequency analysis kernel function family and discretize them, construct candidate parameter combinations, and initialize the number of ants, initial value of pheromone, volatility coefficient and heuristic factor weight of the ant colony optimization algorithm.

[0080] In one specific embodiment of this application, step S3 is performed to transform the parameterized time-frequency analysis kernel function family constructed in step S2 into a parameter space that can be searched by intelligent optimization algorithms, and to complete the initialization settings of the ant colony optimization algorithm, thereby providing an implementable computational basis for the subsequent adaptive optimization of kernel parameters.

[0081] In the embodiments of this application, the parameterized time-frequency analysis kernel function family The kernel parameter is determined by at least one scaling factor and one or more weighting function parameters. Since different kernel parameters will cause differences in the energy concentration, cross-term suppression ability, and peak separability of the time-frequency distribution, this application regards the kernel parameter as the variable to be optimized. First, a reasonable search range is set for it, and then it is discretized to form a finite candidate set so that the ant colony optimization algorithm can perform path selection and iterative convergence in the discrete parameter combination space.

[0082] Specifically, in this embodiment, the range of values ​​for the scaling factor 'a' is first set. The scaling factor is used to adjust the Born-Jordan basic kernel in... The support strength and smoothness on the plane are important considerations. If the values ​​are too small, the kernel function may become overly smoothed, weakening peak resolution. If the values ​​are too large, strong cross terms may be introduced, or energy leakage may increase. Therefore, in this embodiment, the upper and lower limits of 'a' can be set according to the length of the cable under test, the sampling rate, the frequency band range of the detection signal, and the expected defect reflection characteristics. This continuous interval is then discretized into several candidate value points with a preset step size.

[0083] Meanwhile, in the embodiments of this application, the weighting function is... The parameters are set with value ranges and discretized. The weighting function can be used to limit the effective support region of the kernel function or enhance the weight of certain regions to further improve the time-frequency energy focusing effect. The parameters of the weighting function may include, but are not limited to, the width, shape parameter, attenuation coefficient, or support boundary parameter of the window function. In this embodiment, the above parameters are also set with value ranges and discretized to obtain several sets of discrete candidate parameters. By combining the scale factor a with the weighting function parameters, multiple sets of candidate kernel parameter combinations can be formed as search nodes or candidate solution sets for subsequent ant colony optimization algorithms.

[0084] After constructing candidate parameter combinations, this embodiment further initializes the relevant hyperparameters of the ant colony optimization algorithm to guide the search strategy and convergence behavior. Specifically, the number of ants is initialized to determine the size of individuals participating in the search in each iteration; the initial pheromone values ​​corresponding to each candidate parameter combination are initialized to ensure that the algorithm maintains a relatively balanced exploration probability for each candidate combination in the initial stage; the pheromone evaporation coefficient is set to control the rate of forgetting historical search experience and avoid prematurely falling into local optima; and the heuristic factor weight is set to reflect the guiding role of prior tendencies for certain parameter combinations, such as a tendency to select parameter regions with tighter support or easier peak focusing. Through the above initialization settings, this embodiment enables the ant colony optimization algorithm to perform effective search within the discretized kernel parameter combination space and lays the foundation for subsequent pheromone updates and optimal parameter output based on the fitness of the objective function.

[0085] S4. For each set of candidate kernel parameters, calculate the time-frequency distribution corresponding to the incident signal and the echo signal respectively, perform peak detection based on the time-frequency distribution, and extract feature indicators including peak energy, peak width, sidelobe energy and inter-peak separation.

[0086] In one specific embodiment of this application, step S4 is performed to conduct time-frequency analysis on each set of candidate kernel parameters within the discretized candidate kernel parameter combination space and extract feature indicators that can be used to evaluate the "energy focusing effect". The output of this step provides basic data for subsequent construction of the objective function and optimization of the ant colony fitness.

[0087] In this embodiment, for each set of candidate kernel parameter combinations, they are first substituted into the parameterized time-frequency analysis kernel function family constructed in step S2 to obtain the corresponding kernel function form. Subsequently, Cohen-type time-frequency transforms are performed on the incident signal and the echo signal respectively, thereby obtaining the time-frequency distribution results under that set of kernel parameters. Since the incident signal and the echo signal differ in their propagation paths and reflection mechanisms, they may exhibit different energy accumulation patterns in the time-frequency plane. Therefore, this embodiment calculates the time-frequency distribution for both separately to ensure that subsequent peak time positioning and peak alignment have a consistent and reliable time-frequency basis.

[0088] After obtaining the time-frequency distribution, embodiments of this application perform peak detection based on the time-frequency distribution to identify the main energy peaks of the incident and reflected waves on the energy-time axis. Peak detection can be performed directly on the time-frequency plane, or it can be performed after obtaining the energy-time curve by energy aggregation of the time-frequency distribution. Embodiments of this application do not limit the specific implementation method of peak detection. For example, a method combining threshold decision and local maximum search can be used, or methods such as sliding window peak search and non-maximum suppression can be used to extract stable and reliable main peak positions.

[0089] To facilitate the quantification of the impact of candidate kernel parameters on energy focusing and inter-peak separability, embodiments of this application extract multiple feature indicators based on peak detection, including at least peak energy, peak width, sidelobe energy, and inter-peak time interval or inter-peak separation. Peak energy reflects the degree of energy concentration of the main peak; peak width reflects the extent of the main peak's expansion in the time or time-frequency dimension; a smaller peak width generally indicates more concentrated energy and a sharper peak; sidelobe energy characterizes the energy leakage level near the main peak and in non-main peak regions; and inter-peak time interval or inter-peak separation characterizes whether different peaks are easily distinguishable, especially in cases with multiple reflections or defects, where this indicator directly affects the discriminability and robustness of the localization.

[0090] In this embodiment, the peak energy can be measured using time-domain aggregation, frequency-domain aggregation, or time-frequency-domain aggregation, depending on implementation requirements. As a preferred implementation, the time-frequency distribution can be integrated and aggregated along the frequency dimension to obtain a time-dependent instantaneous energy curve, thereby enabling peak detection and energy quantization on the time axis. The frequency-domain aggregated signal energy can be expressed as:

[0091]

[0092] in, The time-frequency distribution obtained using candidate kernel functions, This corresponds to the energy-time curve. From this energy-time curve, indicators such as peak amplitude, peak energy, and peak width can be further extracted.

[0093] In this embodiment, peak energy and peak width together constitute an index system for measuring the "peak enhancement effect." Generally, provided the peak is detectable, higher peak energy and narrower peak width indicate better energy focusing capability in the time-frequency analysis results corresponding to the candidate kernel parameters. Sidelobe energy is used to measure the degree of energy diffusion and leakage; smaller sidelobe energy usually indicates weaker cross terms and energy diffusion, which is beneficial for improving the significance and positioning accuracy of the reflection peak. Inter-peak separation or inter-peak time interval further reflects whether different reflection peaks can be stably distinguished on the time axis, thus providing a guarantee for subsequent multi-peak positioning. By extracting and quantifying the above characteristic indicators, this embodiment provides a calculable and comparable evaluation basis for constructing the objective function and performing ant colony optimization search in subsequent steps.

[0094] S5. Construct an objective function based on the characteristic indicators, which includes a peak enhancement term, an energy leakage penalty term, and an inter-peak separability term, and use the objective function as the fitness function of the ant colony optimization algorithm.

[0095] In one specific embodiment of this application, step S5 is performed to transform the peak feature index extracted in step S4 into an evaluation criterion that can be directly used by the optimization algorithm, thereby constructing an objective function, which is then used as the fitness function of the ant colony optimization algorithm. Through this step, this application can establish a unified and comparable scale among candidate kernel parameter combinations, enabling the optimization process to search with the guidance of "more focused energy, lower sidelobes, and easier separation between peaks".

[0096] In this embodiment, the objective function includes at least three mutually constraining evaluation terms: a peak enhancement term, an energy leakage penalty term, and a peak-to-peak separability term. The peak enhancement term encourages candidate kernel parameters to produce sharper, more concentrated main peak energy representations. The energy leakage penalty term suppresses sidelobes and energy diffusion outside the window, preventing energy dispersion in non-target regions that could lead to indistinct peaks or an increase in false peaks. The peak-to-peak separability term maintains the distinguishability of incident peaks and reflection peaks, or multiple reflection peaks, on the time axis, thereby improving the localization capability and stability under multi-defect conditions. These three evaluation terms together constitute a comprehensive evaluation of the time-frequency distribution quality. Their essential goal is to maximize energy concentration while minimizing diffusion and cross-interference, and ensuring sufficient separation between peaks.

[0097] Specifically, in this embodiment, the peak enhancement index can be constructed from peak energy and peak width. Generally, a larger peak energy and a smaller peak width indicate that the energy is more concentrated near the peak. This embodiment can use peak energy and peak width to construct the peak enhancement index through multiplication, division, normalization ratios, or weighted combinations to characterize the sharpness and concentration of the main peak. For multi-peak scenarios, the amplitude or energy of multiple main peaks can be statistically analyzed to obtain the average peak amplitude or average peak energy, thus avoiding excessive influence of single-peak anomalies on the evaluation.

[0098] The energy leakage penalty index can be constructed based on the sidelobe energy and the energy distribution outside the peak window. In this embodiment, the sidelobe energy can be obtained by defining a sidelobe interval near the main peak and integrating or accumulating the energy in that interval; meanwhile, the leakage outside the window is used to measure the total energy in the region outside the main peak window. If this value is large, it indicates that the energy diffusion is significant or the cross term is strong, which can easily cause the reflection peak to be submerged or interference peaks to be generated, thereby reducing the reliability of positioning. Therefore, this embodiment introduces the average sidelobe energy and the leakage outside the window together as penalty terms into the objective function to drive the optimization process to select a kernel parameter combination with smaller leakage and better noise suppression.

[0099] The inter-peak separability index is used to measure the time interval or separation between adjacent peaks. In the embodiments of this application, the time index difference between adjacent peaks can be directly used as the inter-peak spacing index, or a normalized separation index combined with peak width can be used to reflect whether two peaks can still be distinguished under certain expansion conditions. The introduction of this item makes the optimization process not only pursue a sharper individual peak, but also avoids multiple peaks sticking together due to excessive smoothing or excessive compression, ensuring the feasibility of subsequent reflection delay extraction and multi-peak localization.

[0100] In this embodiment of the application, in order to achieve the above objectives, the peak enhancement index, leakage penalty index and inter-peak separability index are weighted and constructed to form an optimization objective function for evaluating the merits of kernel parameter combinations, and this objective function is used as the fitness function of the ant colony optimization algorithm.

[0101] As an exemplary implementation, the objective function of this application can be of the following form:

[0102]

[0103] Where k represents a set of candidate kernel parameter combinations or a set of parameter selections obtained by ant construction, and α, β, η are weight parameters used to balance the importance of peak enhancement, inter-peak separability and energy leakage penalty. The average amplitude or average energy of the multi-peak values ​​is used to characterize the overall intensity of the main peak. The average energy of the multi-peak sidelobes is used to characterize the level of sidelobe leakage. This is a parameter for non-zeroing the denominator, used to avoid instability caused by a zero or excessively small denominator in numerical calculations; For a time index of a certain peak, The time index for adjacent peaks. Used to characterize peak spacing or separability; This is an energy leakage term outside the multi-peak window, used to penalize energy diffusion in non-peak regions.

[0104] In this embodiment of the application, the energy leakage item outside the window This can be further expressed as:

[0105]

[0106] Where E(t) represents the total energy distribution of the energy-time curve over the entire time range. Let K represent the energy distribution within the peak window corresponding to the i-th main peak, and K be the number of detected main peaks. By subtracting the sum of the energies within each main peak window from the global energy, the total energy outside the main peak window can be obtained, reflecting the level of energy diffusion and leakage. This definition allows for more concentrated energy within the main peak window. The smaller the value, the larger the objective function becomes under the influence of the penalty term.

[0107] By defining the objective function as described above, this embodiment maps the time-frequency analysis results of "strong peak value, narrow peak width, low sidelobe, good inter-peak isolation, and small window leakage" to a higher fitness value. Subsequently, in the ant colony optimization process, by maximizing J(k) as the optimization direction, the energy concentration of the multi-peaks can be maximized and the energy diffusion minimized while maintaining the peak spacing and separability, thereby obtaining the energy focusing time-frequency analysis kernel function for subsequent positioning calculations.

[0108] S6, the candidate kernel parameters are iteratively searched using the ant colony optimization algorithm, the pheromone is updated according to the fitness function and volatilization is performed, and the optimal kernel parameters that maximize the objective function are obtained, thereby obtaining the time-frequency analysis kernel function for energy focusing.

[0109] In one specific embodiment of this application, step S6 is performed, in which the ant colony optimization algorithm is used to iteratively search the candidate kernel parameter combination constructed in step S3, and the objective function constructed in step S5 is used as the fitness function to drive pheromone updates and volatilization, thereby obtaining the optimal kernel parameter combination that maximizes the objective function, and finally obtaining the time-frequency analysis kernel function of energy focusing.

[0110] In this embodiment, candidate kernel parameter combinations can be viewed as discrete search space nodes. The ant colony optimization algorithm uses multiple ants to perform state transitions and path construction within this discrete space, gradually strengthening better-performing parameter combinations and suppressing poor-performing ones, thus achieving an adaptive optimization process from extensive exploration to convergence. Compared to exhaustive search, ant colony optimization can significantly reduce computational cost when there are many parameter dimensions and a large number of combinations, while also possessing a certain global search capability, which helps avoid getting trapped in local optima for kernel parameters.

[0111] In this embodiment, the state transition probability of the ant in the kernel parameter combination space is first determined based on the current pheromone distribution and the heuristic factor. Specifically, in the d-th iteration, the ant selects the next candidate parameter combination starting from its current state. The probability can be defined as:

[0112]

[0113] in, This indicates the combination with candidate parameters in the d-th iteration. The corresponding pheromone variable reflects the cumulative advantages and disadvantages of this combination in historical searches; The heuristic factor can be used to incorporate prior preferences, such as favoring parameter regions with tighter support, more suitable smoothness, or a higher likelihood of improving energy focusing. ρ and σ are weighting coefficients used to adjust the relative contributions of pheromones and the heuristic factor in state transition decisions. When ρ is large, the algorithm tends to utilize historical experience; when σ is large, the algorithm tends to rely more on heuristic guidance, thus achieving a balance between exploration and convergence.

[0114] In each iteration, this embodiment constructs a set of kernel parameter combinations k for each ant. Then, based on the time-frequency distribution calculation and peak feature extraction in step S4, and the objective function definition in step S5, the fitness value J(k) of the solution obtained by the ant is calculated. The higher the fitness value, the better the overall performance of the kernel parameter combination in terms of peak enhancement, sidelobe suppression, window leakage reduction, and inter-peak separability, and the higher the probability of selection in subsequent iterations.

[0115] In this embodiment, the pheromone increment update is based on fitness. If ant i selects a candidate parameter combination in the d-th iteration... Then, perform an incremental update on the pheromone variable for the candidate combination:

[0116]

[0117] Where q is a scaling factor used to adjust the amplification of pheromone increment by fitness, and J(i) represents the objective function value or fitness value corresponding to the parameter combination constructed by ant i. If ant i does not select a candidate parameter combination... Then the pheromone increment is 0, that is:

[0118]

[0119] Through this mechanism, the better-performing parameter combinations will accumulate more pheromones in the same round of iteration, thus having a greater probability of being selected in the next round of iteration.

[0120] To prevent the algorithm from prematurely falling into local optima due to the infinite accumulation of pheromones, and to ensure that the search process can continuously explore new parameter combinations, this embodiment introduces a volatility coefficient ψ to perform pheromone volatilization and enhancement. The pheromone update can be expressed as:

[0121]

[0122] in, This indicates the residual pheromones after evaporation. The value of ψ is used to control the rate of forgetting; the larger the value of ψ, the faster the evaporation and the stronger the exploratory nature. This represents the cumulative pheromone enhancement term from K ants, used to strengthen the candidate parameter combination that performs better in this round.

[0123] In this embodiment, the iterative process of "state transition probability calculation—parameter combination construction—fitness evaluation—pheromone incremental update—pheromone evaporation" is repeatedly executed until a preset termination condition is met, at which point the optimal result is output. The termination condition may include, but is not limited to: reaching a preset upper limit for the number of iterations, the optimal fitness improvement being less than a threshold for several consecutive iterations, or the optimal kernel parameter combination remaining unchanged for several iterations. After the termination condition is met, the kernel parameter combination that maximizes the objective function is output as the optimal kernel parameter, and the time-frequency analysis kernel function for energy focusing is determined accordingly. Through the above-mentioned iterative pheromone update mechanism, this embodiment enables the search process to gradually converge towards a parameter combination with a larger objective function, ultimately obtaining an ant colony-optimized energy focusing kernel function, providing crucial support for subsequent accurate location of time-frequency energy peaks and calculation of cable insulation damage locations.

[0124] In one possible embodiment, the energy focusing kernel function optimized by the ant colony is as follows: Figure 3 As shown.

[0125] S7. Using the time-frequency analysis kernel function of the energy focusing, the time-frequency distribution and energy-time curve of the signal are recalculated to determine the arrival time of the energy peak corresponding to the incident wave and the reflected wave, and the location of the cable insulation damage is calculated based on the reflection delay and the wave velocity in the cable.

[0126] In one specific embodiment of this application, step S7 is performed. After the optimization and determination of the energy focusing time-frequency analysis kernel function is completed, the energy focusing time-frequency analysis kernel function is used to perform the final time-frequency analysis and location calculation on the detection signal, thereby realizing the determination of the location of cable insulation damage.

[0127] In this embodiment, the energy focusing time-frequency analysis kernel function output in step S6 is first used to re-perform the Cohen-type time-frequency transform on the acquired incident and echo signals. Since this kernel function has already achieved a superior state in terms of peak enhancement, energy leakage suppression, and inter-peak separability through ant colony optimization, the recalculated time-frequency distribution exhibits a more concentrated energy distribution and a clearer multi-peak structure compared to the result under the initial kernel function conditions. Based on this, energy aggregation is performed on the time-frequency distribution in either the frequency or time-frequency dimension to obtain the corresponding energy-time curve, which is used for subsequent peak detection and time delay calculation.

[0128] In this embodiment, the main energy peaks in the signal are detected based on the energy-time curve. The detected energy peaks typically include an incident wave peak corresponding to the initial injection of the detection signal, and one or more reflected wave peaks generated by discontinuities in the cable's internal impedance. By performing a peak search on the energy-time curve, the arrival time of each energy peak on the time axis can be determined, thereby obtaining the time difference between the incident wave and each reflected wave, i.e., the reflection delay.

[0129] Furthermore, in this embodiment, X effective peak values ​​are detected in the energy-time curve, where the peak value corresponding to the incident wave is used as the reference peak, and its time index is 0. The time index of the x-th reflected peak value relative to the incident peak value is denoted as... By using the above method, different reflection peaks can be uniformly mapped to the same time reference coordinate system, which facilitates distance calculation and multi-peak location analysis.

[0130] After obtaining the reflection delay, this embodiment calculates the location of cable insulation damage by combining the propagation wave velocity in the cable. Specifically, since the detection signal propagates in the cable to the damage location, is reflected, and then returns to the receiving end, its propagation path includes both round trip directions; therefore, there is a proportional relationship between the corresponding spatial distance and the time delay. In this embodiment, the cable insulation damage location corresponding to the x-th reflection peak... It can be represented as:

[0131]

[0132] Where v represents the propagation speed of the detection signal in the cable. This represents the time delay of the reflected peak relative to the incident peak. By introducing a 1 / 2 coefficient, the total round-trip propagation time delay can be converted into a one-way propagation distance, thus obtaining the position of the damage point relative to the signal injection end or reference port.

[0133] In this application embodiment, the propagation wave velocity v in the cable can be obtained in various ways, such as by calibrating and measuring a cable of known length using the time-domain reflectometry method, or by estimating the propagation characteristic parameters of the cable through vector impedance analysis, frequency domain measurement, etc. This application does not limit the method of obtaining the wave velocity; as long as an effective propagation wave velocity parameter matching the cable under test can be obtained, it can be used for the above-mentioned positioning calculation.

[0134] Through the above steps, the embodiments of this application can utilize the energy focusing time-frequency analysis kernel function obtained by ant colony optimization to achieve stable separation and accurate time delay extraction of incident waves and multiple reflected waves in complex echo environments. This enables high-precision positioning of cable insulation damage locations in the presence of single or multiple insulation damage points, thereby improving the reliability and engineering applicability of the detection results.

[0135] In addition, refer to Figure 4 Experiments have shown that by using the energy focusing kernel function proposed in this invention to detect and locate two artificial defects at 10m and 20m of a 30m long cable, the peak energy of the reflected waves at the two locations can be identified to 88% and 65% of the injected signal, respectively. This is 10.17 to 20.56% higher than that of common kernel functions.

[0136] Example 2

[0137] This invention also provides an ant colony-optimized energy-focusing cable insulation damage location device, the device comprising:

[0138] The first module is used to generate a Gaussian envelope linear frequency modulated chirp detection signal and inject the detection signal into the cable under test, and collect the incident signal and the echo signal containing the reflected component at the receiving end.

[0139] The second module is used to perform time-frequency transformation on the incident signal and the echo signal respectively under the Cohen-type time-frequency distribution framework, select the Born-Jordan kernel as the basic kernel, and introduce a scale factor to construct a family of parameterized time-frequency analysis kernel functions;

[0140] The third module is used to set the search range and discretize the kernel parameters of the parameterized time-frequency analysis kernel function family, construct candidate parameter combinations, and initialize the number of ants, initial value of pheromone, volatility coefficient and heuristic factor weight of the ant colony optimization algorithm.

[0141] The fourth module is used to calculate the time-frequency distribution of the incident signal and the echo signal for each set of candidate kernel parameters, perform peak detection based on the time-frequency distribution, and extract feature indicators including peak energy, peak width, sidelobe energy and inter-peak separation.

[0142] The fifth module is used to construct an objective function containing a peak enhancement term, an energy leakage penalty term, and an inter-peak separability term based on the feature index, and to use the objective function as the fitness function of the ant colony optimization algorithm.

[0143] The sixth module is used to iteratively search the candidate kernel parameters using an ant colony optimization algorithm, update the pheromone according to the fitness function and perform volatilization, to obtain the optimal kernel parameters that maximize the objective function, thereby obtaining the time-frequency analysis kernel function for energy focusing;

[0144] The seventh module is used to recalculate the time-frequency distribution and energy-time curve of the signal using the time-frequency analysis kernel function of the energy focusing, determine the arrival time of the energy peak corresponding to the incident wave and the reflected wave, and calculate the location of cable insulation damage based on the reflection time delay and the wave velocity in the cable.

[0145] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0146] Example 3

[0147] To implement the methods of the above embodiments, the present invention also provides an electronic device, which includes a memory and a processor; wherein the processor reads executable program code stored in the memory to run a program corresponding to the executable program code, so as to implement the various steps of the methods described above.

[0148] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0149] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0150] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

Claims

1. A method for locating cable insulation damage using energy focusing optimized by ant colony control, characterized in that, include: A Gaussian envelope linear frequency modulated chirp detection signal is generated and injected into the cable under test. The incident signal and the echo signal containing the reflected component are collected at the receiving end. Within the Cohen-class time-frequency distribution framework, time-frequency transformations are performed on the incident signal and the echo signal respectively. The Born-Jordan kernel is selected as the basic kernel, and a scale factor is introduced to construct a family of parameterized time-frequency analysis kernel functions. The search range of the kernel parameters of the parameterized time-frequency analysis kernel function family is set and discretized to construct candidate parameter combinations. The number of ants, the initial value of pheromone, the volatility coefficient and the weight of heuristic factor are initialized for the ant colony optimization algorithm. For each set of candidate kernel parameters, the time-frequency distributions corresponding to the incident signal and the echo signal are calculated respectively. Peak detection is performed based on the time-frequency distributions, and feature indicators including peak energy, peak width, sidelobe energy and inter-peak separation are extracted. Based on the aforementioned feature indicators, an objective function is constructed that includes a peak enhancement term, an energy leakage penalty term, and an inter-peak separability term, and this objective function is used as the fitness function of the ant colony optimization algorithm. The candidate kernel parameters are iteratively searched using an ant colony optimization algorithm. The pheromone is updated according to the fitness function and volatilization is performed to obtain the optimal kernel parameters that maximize the objective function, thereby obtaining the time-frequency analysis kernel function for energy focusing. The time-frequency distribution and energy-time curve of the signal are recalculated using the time-frequency analysis kernel function of the energy focusing, the arrival time of the energy peak corresponding to the incident wave and the reflected wave are determined, and the location of cable insulation damage is calculated based on the reflection delay and the wave velocity in the cable.

2. The method according to claim 1, characterized in that, A Gaussian envelope linear frequency modulated chirp detection signal is generated and injected into the cable under test. The incident signal and the echo signal containing the reflected component are acquired at the receiving end, including: Based on the preset time-domain width parameter, time center parameter, frequency center parameter and frequency change rate, a linear frequency-modulated chirp detection signal with a Gaussian envelope is constructed. The detection signal is injected into the cable under test through capacitive coupling, inductive coupling, or direct coupling. Signals are acquired at the receiving end of the cable under test to obtain an echo signal containing the incident wave component and the reflected wave component caused by the cable impedance discontinuity. The echo signal is then synchronously sampled to form the input signal for subsequent time-frequency analysis.

3. The method according to claim 1, characterized in that, Within the Cohen-class time-frequency distribution framework, time-frequency transformations are performed on the incident signal and the echo signal, respectively. The Born-Jordan kernel is selected as the basic kernel, and a family of parameterized time-frequency analysis kernel functions is constructed by introducing a scaling factor, including: Within the framework of Cohen-class bilinear time-frequency distribution, a time-frequency transformation expression for the signal is established, where the kernel function is used to adjust the energy concentration characteristics of the time-frequency distribution; The Born-Jordan kernel is selected as the basic kernel function, and the basic kernel is extended by introducing at least one scaling factor or weighting function to form an adjustable family of time-frequency analysis kernel functions.

4. The method according to claim 3, characterized in that, The kernel parameters of the parameterized time-frequency analysis kernel function family are set with a search range and discretized to construct candidate parameter combinations. The ant colony optimization algorithm is then initialized with the ant count, initial pheromone value, volatile coefficient, and heuristic factor weights, including: The scaling factor and weighting function parameters are set to have a range of values ​​and are discretized to generate multiple sets of candidate kernel parameter combinations; Initialize the number of ants, initial pheromone value, pheromone volatility coefficient, and heuristic factor weights in the ant colony optimization algorithm to guide the search process for kernel parameter combinations.

5. The method according to claim 4, characterized in that, For each set of candidate kernel parameters, the time-frequency distributions corresponding to the incident signal and the echo signal are calculated respectively. Peak detection is performed based on the time-frequency distributions, and feature indicators including peak energy, peak width, sidelobe energy, and inter-peak separation are extracted, including: Based on the corresponding candidate kernel parameters, the time-frequency distributions of the incident signal and the echo signal are calculated respectively. Peak detection is performed on the time-frequency distribution, and the peak energy, peak width, sidelobe energy, and inter-peak time interval or separation degree corresponding to each major peak are extracted as characteristic indicators to measure the performance of time-frequency energy aggregation and separation.

6. The method according to claim 5, characterized in that, Based on the aforementioned feature indicators, an objective function is constructed that includes a peak enhancement term, an energy leakage penalty term, and an inter-peak separability term. This objective function is then used as the fitness function for the ant colony optimization algorithm, including: A peak enhancement index is constructed based on the peak energy and peak width; An energy leakage penalty index is constructed based on the sidelobe energy and the energy distribution outside the peak window; Construct an inter-peak separability index based on the time interval or separability between adjacent peaks; The peak enhancement index, energy leakage penalty index, and inter-peak separability index are weighted to construct an objective function for evaluating the merits of kernel parameter combinations, and the objective function is used as the fitness function of the ant colony optimization algorithm.

7. The method according to claim 6, characterized in that, The candidate kernel parameters are iteratively searched using an ant colony optimization algorithm. The pheromone is updated according to the fitness function and volatilization is performed to obtain the optimal kernel parameters that maximize the objective function, thereby obtaining the time-frequency analysis kernel function for energy focusing, including: Based on the current pheromone distribution and heuristic factors, determine the state transition probability of the ant in the kernel parameter combination space; Based on the fitness function, the selected kernel parameter combinations are incrementally updated with pheromones, and the unselected kernel parameter combinations are pheromone evaporation is performed. Repeat the state transition and pheromone update process until the preset termination condition is met, and output the kernel parameter combination that makes the objective function reach the optimal value to obtain the time-frequency analysis kernel function for energy focusing.

8. The method according to claim 7, characterized in that, The time-frequency distribution and energy-time curve of the signal are recalculated using the energy-focusing time-frequency analysis kernel function to determine the arrival times of the energy peaks corresponding to the incident and reflected waves. The location of cable insulation damage is then calculated based on the reflection time delay and wave velocity in the cable, including: The time-frequency distribution of the signal is recalculated using the energy-focused time-frequency analysis kernel function, and the corresponding energy-time curve is obtained. Determine the arrival time of the energy peak corresponding to the incident wave and at least one reflected wave based on the energy-time curve, and calculate the reflection time delay; Based on the reflection delay and the propagation wave velocity in the cable, calculate the location of cable insulation damage corresponding to each reflection peak.

9. A colony-optimized energy-focusing cable insulation damage location device, characterized in that, include: The first module is used to generate a Gaussian envelope linear frequency modulated chirp detection signal and inject the detection signal into the cable under test, and collect the incident signal and the echo signal containing the reflected component at the receiving end. The second module is used to perform time-frequency transformation on the incident signal and the echo signal respectively under the Cohen-type time-frequency distribution framework, select the Born-Jordan kernel as the basic kernel, and introduce a scale factor to construct a family of parameterized time-frequency analysis kernel functions; The third module is used to set the search range and discretize the kernel parameters of the parameterized time-frequency analysis kernel function family, construct candidate parameter combinations, and initialize the number of ants, initial value of pheromone, volatility coefficient and heuristic factor weight of the ant colony optimization algorithm. The fourth module is used to calculate the time-frequency distribution of the incident signal and the echo signal for each set of candidate kernel parameters, perform peak detection based on the time-frequency distribution, and extract feature indicators including peak energy, peak width, sidelobe energy and inter-peak separation. The fifth module is used to construct an objective function containing a peak enhancement term, an energy leakage penalty term, and an inter-peak separability term based on the feature index, and to use the objective function as the fitness function of the ant colony optimization algorithm. The sixth module is used to iteratively search the candidate kernel parameters using an ant colony optimization algorithm, update the pheromone according to the fitness function and perform volatilization, to obtain the optimal kernel parameters that maximize the objective function, thereby obtaining the time-frequency analysis kernel function for energy focusing; The seventh module is used to recalculate the time-frequency distribution and energy-time curve of the signal using the time-frequency analysis kernel function of the energy focusing, determine the arrival time of the energy peak corresponding to the incident wave and the reflected wave, and calculate the location of cable insulation damage based on the reflection time delay and the wave velocity in the cable.

10. An electronic device, characterized in that, Including processor and memory; The processor reads executable program code stored in the memory to run a program corresponding to the executable program code, so as to implement the method as described in any one of claims 1-8.

Images