Cable oscillation ultra-low frequency withstand voltage and partial discharge integrated test method and cable oscillation ultra-low frequency withstand voltage and partial discharge integrated test system

Abstract

The invention discloses a cable oscillation ultralow-frequency withstand voltage and partial discharge integrated test method and system, and relates to the technical field of distribution network cable detection, and the method comprises the steps: applying an oscillation ultralow-frequency voltage signal to a tested cable; the oscillation ultra-low frequency voltage signal is optimized and adjusted through a manifold learning algorithm and is used as a common excitation source for a withstand voltage test, dielectric loss detection and partial discharge detection; acquiring a voltage signal and a current signal of the tested cable under the excitation of the oscillation ultralow-frequency voltage signal, analyzing the deviation degree of the voltage signal and the current signal from the pre-established characteristic manifold, and extracting a dielectric loss characteristic signal and a partial discharge pulse signal; calculating a dielectric loss parameter based on the dielectric loss characteristic signal, determining a partial discharge position based on the partial discharge pulse signal, and generating a cable insulation state diagnosis report according to the dielectric loss parameter and the partial discharge position. According to the invention, the cable testing efficiency is improved, the equipment burden and the testing time are reduced, comprehensive evaluation and accurate fault positioning of the insulation state are realized, and a scientific basis is provided for preventive maintenance of a power system.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution network cable detection, and in particular to a cable oscillation ultra-low frequency withstand voltage and partial discharge integrated testing method and system. Background Art

[0002] The evaluation of cable insulation status in power distribution network systems has always been a key link in the reliable operation of power systems. With the continuous advancement of urbanization, underground distribution networks have become increasingly complex, and cable insulation failures have become one of the main types of distribution system failures. Traditional cable detection technologies mainly include three independent methods: dielectric loss test, withstand voltage test, and partial discharge test. Dielectric loss test usually uses power frequency or very low frequency signals to evaluate the overall insulation degradation state; withstand voltage test uses high voltage DC or 0.1Hz ultra-low frequency signal to verify the electrical strength of insulation; partial discharge detection requires high-frequency sampling technology to capture weak discharge pulse signals. These three detection methods require different excitation sources and measurement equipment for different insulation characteristics, resulting in complex on-site detection process, low efficiency, redundant equipment, and difficulty in comprehensive correlation analysis of detection results.

[0003] Existing technologies have many deficiencies in assessing the insulation status of cables. First, traditional detection technologies require multiple tests on the same cable, which not only prolongs the power outage time and increases the operation and maintenance costs, but may also accelerate insulation degradation due to repeated application of high voltage. Secondly, the test results are independent of each other, lack an effective correlation analysis mechanism, and cannot fully reflect the complexity of the insulation status. For example, dielectric loss testing may show that the insulation is aged overall, but it cannot locate the specific defect location; while partial discharge detection can locate defects, it is difficult to assess the overall degree of insulation degradation. In addition, traditional detection methods often use test signals with fixed parameters and lack the ability to adapt to different cable characteristics, resulting in limited test sensitivity and accuracy. Finally, existing technologies usually use empirical rules or simple threshold judgments in the signal processing and diagnosis process, and do not make full use of advanced signal processing and artificial intelligence technologies, making it difficult to cope with the diverse fault modes in complex cable systems. Summary of the invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the present invention provides a cable oscillation ultra-low frequency withstand voltage and partial discharge integrated testing method and system, which can solve the problems mentioned in the background technology.

[0006] To solve the above technical problems, the present invention provides the following technical solutions: a cable oscillation ultra-low frequency withstand voltage and partial discharge integrated test method, comprising: applying an oscillating ultra-low frequency voltage signal to the cable under test; the oscillating ultra-low frequency voltage signal is optimized and adjusted through a manifold learning algorithm, and serves as a common excitation source for withstand voltage test, dielectric loss detection and partial discharge detection; obtaining the voltage signal and current signal of the cable under test under the excitation of the oscillating ultra-low frequency voltage signal, analyzing the degree of deviation of the voltage signal and the current signal from a pre-built characteristic manifold, and extracting the dielectric loss characteristic signal and the partial discharge pulse signal; calculating the dielectric loss parameter based on the dielectric loss characteristic signal, determining the partial discharge position based on the partial discharge pulse signal, and generating a cable insulation status diagnosis report according to the dielectric loss parameter and the partial discharge position.

[0007] As a preferred scheme of the integrated cable oscillation ultra-low frequency withstand voltage and partial discharge test method described in the present invention, the optimization and adjustment of the manifold learning algorithm includes: establishing a parameter space according to the type of cable under test; generating a test parameter matrix in the parameter space; collecting test response characteristics to construct a characteristic manifold; and optimizing the frequency, amplitude and waveform parameters of the oscillating ultra-low frequency voltage signal based on the characteristic manifold.

[0008] As a preferred scheme of the integrated cable oscillation ultra-low frequency withstand voltage and partial discharge test method described in the present invention, wherein: obtaining the voltage signal and current signal of the tested cable under the excitation of the oscillating ultra-low frequency voltage signal, including: collecting voltage sampling data and current sampling data; performing signal preprocessing on the voltage sampling data and the current sampling data; converting the preprocessed sampling data into the voltage signal and the current signal, and mapping them to a preset feature space.

[0009] As a preferred scheme of the integrated test method for cable oscillation ultra-low frequency withstand voltage and partial discharge described in the present invention, wherein: the extraction of dielectric loss characteristic signal and partial discharge pulse signal includes: performing orthogonal decomposition on the voltage signal and the current signal to obtain orthogonal components; performing time-frequency analysis on the orthogonal components to obtain energy distribution characteristics; based on the energy distribution characteristics, the orthogonal components are distinguished into dielectric loss related components and partial discharge related components; partial discharge characteristics are extracted from the partial discharge related components to form the partial discharge pulse signal; phase difference and amplitude ratio parameters are extracted from the dielectric loss related components to form the dielectric loss characteristic signal.

[0010] As a preferred scheme of the integrated cable oscillation ultra-low frequency withstand voltage and partial discharge test method described in the present invention, wherein: analyzing the degree of deviation of the voltage signal and the current signal from the pre-built characteristic manifold, including: calculating the distance parameter between the characteristic parameters of the voltage signal and the current signal and the standard characteristic manifold; determining the deviation state according to the distance parameter; if the deviation state exceeds the reference range, adjusting the frequency and amplitude of the oscillating ultra-low frequency voltage signal to return the deviation state to the reference range; otherwise, maintaining the current oscillating ultra-low frequency voltage signal parameters and continuing the test process.

[0011] As a preferred solution of the integrated test method for cable oscillation ultra-low frequency withstand voltage and partial discharge described in the present invention, the partial discharge position is determined based on the partial discharge pulse signal, including: extracting the time characteristics of the partial discharge pulse signal; calculating the pulse propagation time difference; calculating the partial discharge source distance based on the propagation time difference and the cable propagation speed; and calibrating the physical position of the cable corresponding to the partial discharge source distance.

[0012] As a preferred solution of the integrated test method for cable oscillation ultra-low frequency withstand voltage and partial discharge described in the present invention, the partial discharge position is determined based on the partial discharge pulse signal, including: extracting the time characteristics of the partial discharge pulse signal; calculating the pulse propagation time difference; calculating the partial discharge source distance based on the propagation time difference and the cable propagation speed; and calibrating the physical position of the cable corresponding to the partial discharge source distance.

[0013] To further solve the above technical problems, the present invention provides the following technical solutions: a cable oscillation ultra-low frequency withstand voltage and partial discharge integrated test system, comprising: a signal excitation module, used to apply an oscillating ultra-low frequency voltage signal to the cable under test, the oscillating ultra-low frequency voltage signal is optimized and adjusted through a manifold learning algorithm, and serves as a common excitation source for withstand voltage test, dielectric loss detection and partial discharge detection; a signal processing module, used to obtain the voltage signal and current signal of the cable under test under the excitation of the oscillating ultra-low frequency voltage signal, analyze the degree of deviation of the voltage signal and the current signal from a pre-built characteristic manifold, and extract the dielectric loss characteristic signal and the partial discharge pulse signal; a state diagnosis module, used to calculate the dielectric loss parameter based on the dielectric loss characteristic signal, determine the partial discharge position based on the partial discharge pulse signal, and generate a cable insulation state diagnosis report according to the dielectric loss parameter and the partial discharge position.

[0014] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and is characterized in that when the processor executes the computer program, the steps of the above-mentioned integrated test method for cable oscillation ultra-low frequency withstand voltage and partial discharge are implemented.

[0015] A computer-readable storage medium having a computer program stored thereon, characterized in that when the computer program is executed by a processor, the steps of the cable oscillation ultra-low frequency withstand voltage and partial discharge integrated test method as described above are implemented.

[0016] Beneficial effects of the present invention: The present invention optimizes the oscillating ultra-low frequency voltage signal by applying the manifold learning algorithm, and successfully integrates the three tests of dielectric loss test, withstand voltage test and partial discharge detection that need to be carried out separately into a one-time joint test, which effectively solves the problems of extended power outage time, equipment redundancy and accelerated insulation degradation caused by multiple tests in the prior art. The modified sinusoidal oscillation ultra-low frequency voltage signal used in the signal excitation module adopts a special waveform design, which significantly enhances the dielectric loss characteristics and the sensitivity of partial discharge pulses while maintaining the requirements of the cable withstand voltage test, thereby improving the measurement accuracy and detection sensitivity. The signal processing module adopts a multi-resolution analysis method combined with a collective empirical mode decomposition algorithm to overcome the mode aliasing problem of the traditional EMD method, realize the efficient separation of dielectric loss characteristic signals and partial discharge pulse signals, and effectively improve the signal-to-noise ratio of the signal. The state diagnosis module establishes an accurate identification mechanism for complex faults, insulation degradation and local defects by comprehensively analyzing the dielectric loss parameters and the distribution characteristics of the partial discharge position, improves the positioning accuracy of the partial discharge source, and breaks through the technical bottleneck that the traditional method is difficult to diagnose comprehensively. In general, the present invention realizes a comprehensive assessment of the insulation status of the cable and accurate fault location, provides a scientific basis for preventive maintenance of the power system, and effectively improves the reliability and safety of the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0018] Figure 1 This is a schematic diagram of the overall process of the cable oscillation ultra-low frequency withstand voltage and partial discharge integrated test method proposed by the present invention; Figure 2 This is a diagram of the computer equipment in the integrated test method for cable oscillation ultra-low frequency withstand voltage and partial discharge proposed by the present invention. DETAILED DESCRIPTION

[0019] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0020] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0021] Example 1, reference Figure 1 , which is an embodiment of the present invention, provides an integrated test method for cable oscillation ultra-low frequency withstand voltage and partial discharge.

[0022] Figure 1 The overall process diagram of a cable oscillation ultra-low frequency withstand voltage and partial discharge integrated test method is shown, including the following steps: S1: Apply an oscillating ultra-low frequency voltage signal to the cable under test.

[0023] The oscillating ultra-low frequency voltage signal is optimized and adjusted by a manifold learning algorithm and serves as a common excitation source for withstand voltage test, dielectric loss test and partial discharge test.

[0024] Specifically, the optimization and adjustment of the manifold learning algorithm includes: establishing a parameter space according to the type of cable under test; generating a test parameter matrix in the parameter space; collecting test response characteristics to construct a characteristic manifold; and optimizing the frequency, amplitude and waveform parameters of the oscillating ultra-low frequency voltage signal based on the characteristic manifold.

[0025] In the embodiment of the present application, the specific implementation process of step S1 may be: First, an oscillating ultra-low frequency voltage signal is applied to the cable under test. The oscillating ultra-low frequency voltage signal uses a frequency range of 0.01Hz-0.1Hz, and the specific frequency value is initially selected based on the cable type and length. In actual implementation, the oscillating ultra-low frequency voltage signal is generated by a half-bridge inverter circuit and applied to the cable under test through a high-voltage step-up transformer. The initial peak voltage is set to 1.5-1.7 times the rated voltage of the cable to ensure that the withstand voltage test requirements can be met at the same time.

[0026] In an optional embodiment, the oscillating ultra-low frequency signal can be generated by a full-bridge inverter circuit. Compared with a half-bridge inverter circuit, the full-bridge structure can provide higher voltage output efficiency and lower harmonic distortion, and is particularly suitable for testing scenarios of high-voltage cables of 35kV and above.

[0027] Among them, the oscillating ultra-low frequency voltage signal is optimized and adjusted through the manifold learning algorithm to achieve a common excitation source for the withstand voltage test, dielectric loss test and partial discharge test. The specific adjustment steps are as follows: (1) Establish parameter space according to the type of cable under test: The system first constructs a three-dimensional parameter space containing frequency range, amplitude range and waveform type based on the insulation type (such as XLPE, EPR, etc.), rated voltage level (such as 10kV, 35kV, etc.), cable cross-sectional area and length of the cable under test. For example, for 10kV, 300mm² XLPE cable, the frequency range is set to 0.01Hz-0.05Hz, and the amplitude range is set to 1.5U 0 -1.7U 0 (U 0 is the rated phase voltage), the waveform types include sine wave, modified square wave and trapezoidal wave.

[0028] In an optional embodiment, the parameter space can be expanded into a four-dimensional space, and the voltage rise rate is added as the fourth-dimensional parameter, so that the present invention can more accurately control the cable insulation compression process, which is particularly suitable for testing cables with severe aging and can effectively reduce the risk of breakdown during the test process.

[0029] (2) Generate a test parameter matrix in the parameter space: discretize the parameter space into multiple test points to form a test parameter matrix. Specifically, the frequency dimension is set at an interval of 0.005 Hz, and the amplitude dimension is set at an interval of 0.05 U. 0 For interval value selection, three typical waveforms are selected in the waveform type dimension to construct a parameter matrix containing about 100-200 test points. Each test point represents a set of optional oscillating ultra-low frequency voltage signal parameter combinations.

[0030] In an optional embodiment, an adaptive grid division method can be used to generate a parameter matrix, using a smaller grid spacing (such as a frequency spacing of 0.002 Hz) in the cable response sensitive area and a larger grid spacing (such as 0.01 Hz) in the response flat area. This non-uniform sampling method can improve the detection accuracy of the key parameter area while keeping the calculation amount basically unchanged.

[0031] (3) Collect test response characteristics to construct characteristic manifolds: Perform low voltage pre-tests (usually 30% of the rated voltage) on each parameter combination point in the parameter matrix to collect characteristic parameters of the cable response under test, including voltage-current phase difference, harmonic content, current density distribution, and transient response characteristics. These characteristic parameters are constructed into high-dimensional feature vectors, and the high-dimensional feature vectors are mapped to three-dimensional space through dimensionality reduction methods such as principal component analysis (PCA) and local linear embedding (LLE) to form a visual characteristic manifold. This characteristic manifold reflects the response characteristic distribution of the cable under test under different excitation parameters.

[0032] In an optional embodiment, a more advanced t-SNE (t-distributed StochasticNeighbor Embedding) or UMAP (Uniform Manifold Approximation and Projection) algorithm can be used for feature dimensionality reduction. These algorithms can better preserve the global topological relationship while maintaining the local structure, and are particularly suitable for feature analysis of complex cable systems. In addition, the pre-test voltage can be dynamically adjusted according to the cable condition. For newly laid cables, 50% of the rated voltage can be used for pre-testing to obtain richer feature information.

[0033] (4) Optimize the parameters of the oscillating ultra-low frequency voltage signal based on the characteristic manifold: According to the pre-established standard cable characteristic manifold library, analyze the topological structure of the characteristic manifold of the current cable under test and identify the optimal excitation area. Specifically, by calculating the manifold curvature and sensitivity distribution, determine the parameter area with the highest signal-noise distinction, and select the optimal parameter combination in this area as the final oscillating ultra-low frequency voltage signal parameters. For example, for a 10kV XLPE cable, the optimal parameters finally determined can be: frequency 0.035Hz, amplitude 1.6U 0 , the waveform type is modified sine wave, and the correction coefficient is 0.8.

[0034] It should be noted that when determining the optimal parameter region, the present invention uses Fisher's discriminant ratio to evaluate the signal-noise distinction: ; in, represents the Fisher discriminant ratio under parameter p, and are the means of the signal and noise, respectively. and are the variances of the signal and noise, respectively. It should be noted that parameter p refers to the parameter combination of the oscillating ultra-low frequency voltage signal, specifically including three key parameters: frequency, amplitude and waveform type (including waveform correction coefficient). The system needs to find the optimal parameter combination in the parameter space, so that under this parameter combination, the cable dielectric loss characteristics and partial discharge signals are most distinguishable from the background noise, that is, the Fisher discriminant ratio is the largest. The discrimination ratio is not only used for conventional signal differentiation in the present invention, but is also applied to the optimization process of the oscillating ultra-low frequency signal parameters. By maximizing the discrimination ratio, a technical breakthrough is achieved in that the three tests of withstand voltage test, dielectric loss detection and partial discharge detection share a common excitation source, which is impossible in the prior art. The traditional method requires different test signals to complete these three tests separately, while the present invention finds the optimal signal parameters that can simultaneously meet the requirements of the three tests through parameter optimization guided by the Fisher discriminant ratio.

[0035] In an optional embodiment, the optimal parameters can be selected in combination with a deep reinforcement learning method. By establishing a comprehensive reward function for dielectric loss detection accuracy, partial discharge sensitivity, and test safety, the intelligent agent is trained to explore the optimal strategy in the parameter space. This method has better adaptability to complex insulation structures (such as multi-branch cable systems). Another optional solution is to use a genetic algorithm to optimize the parameter selection process. By simulating the biological evolution process, the optimal parameter combination is iteratively searched, and the search efficiency for large parameter spaces is high.

[0036] After completing the parameter optimization, the power inverter unit is controlled by the digital signal processor DSP to generate an oscillating ultra-low frequency voltage signal that meets the optimization parameter requirements, and is applied to the cable under test through a high-voltage transformer, thereby realizing a common excitation source for the withstand voltage test, dielectric loss test and partial discharge test, and providing high-quality basic signal excitation for the subsequent comprehensive cable insulation status evaluation.

[0037] Based on the above optimization process, the present invention proposes a modified sinusoidal oscillation ultra-low frequency voltage signal with a special structure: ; in, is the voltage value at time t, A is the optimized amplitude, f is the optimized frequency, is the waveform correction factor (usually a value between 0.1 and 0.3). This formula synthesizes the basic sine wave and its third harmonic to produce a special waveform. While maintaining the cable withstand voltage test requirements, by adjusting The parameters significantly enhance the sensitivity to dielectric loss characteristics and partial discharge pulses. Experiments show that when When the value is 0.2, compared with the pure sine wave signal, the corrected waveform can improve the dielectric loss measurement accuracy by 30%, and at the same time increase the partial discharge detection sensitivity by about 25pC.

[0038] In an optional embodiment, the present invention may have the ability to fine-tune parameters online, dynamically adjusting the frequency and amplitude of the oscillating ultra-low frequency signal according to the real-time response characteristics of the cable during the test, making the test process safer and more efficient. For example, when a sudden change in dielectric loss characteristics is detected, the voltage amplitude may be appropriately reduced or the frequency may be adjusted to prevent potential breakdown risks.

[0039] The present invention improves the test efficiency, integrating the three tests (withstand voltage, dielectric loss and partial discharge) that traditionally need to be performed separately into a one-time joint test, shortening the test time from the original 4-6 hours to 1-2 hours. In addition, due to the use of optimized ultra-low frequency signals, the size and weight of the test equipment are greatly reduced, making it easier to carry and operate on site.

[0040] S2: obtaining the voltage signal and current signal of the tested cable under the excitation of the oscillating ultra-low frequency voltage signal, analyzing the degree of deviation of the voltage signal and the current signal from the pre-built characteristic manifold, and extracting the dielectric loss characteristic signal and the partial discharge pulse signal.

[0041] Specifically, the method of obtaining the voltage signal and current signal of the cable under test under the excitation of the oscillating ultra-low frequency voltage signal includes: collecting voltage sampling data and current sampling data; performing signal preprocessing on the voltage sampling data and the current sampling data; converting the preprocessed sampling data into the voltage signal and the current signal, and mapping them to a preset feature space.

[0042] Specifically, the extraction of dielectric loss characteristic signal and partial discharge pulse signal includes: performing orthogonal decomposition on the voltage signal and the current signal to obtain orthogonal components; performing time-frequency analysis on the orthogonal components to obtain energy distribution characteristics; based on the energy distribution characteristics, the orthogonal components are divided into dielectric loss related components and partial discharge related components; extracting partial discharge characteristics from the partial discharge related components to form the partial discharge pulse signal; extracting phase difference and amplitude ratio parameters from the dielectric loss related components to form the dielectric loss characteristic signal.

[0043] Specifically, the analysis of the degree of deviation of the voltage signal and the current signal from the pre-built characteristic manifold includes: calculating the distance parameters between the characteristic parameters of the voltage signal and the current signal and the standard characteristic manifold; determining the deviation state according to the distance parameters; if the deviation state exceeds the reference range, adjusting the frequency and amplitude of the oscillating ultra-low frequency voltage signal to return the deviation state to the reference range; otherwise, maintaining the current oscillating ultra-low frequency voltage signal parameters and continuing the test process.

[0044] In the embodiment of the present application, the specific implementation process of step S2 may be: First, the voltage and current signals of the tested cable are obtained under the excitation of the modified sinusoidal oscillation ultra-low frequency voltage signal. A high-precision sensor array is used for acquisition, in which the voltage acquisition adopts a dual-channel solution combining a resistor divider and a high-voltage probe, and the current acquisition adopts a dual-channel solution combining a current transformer and a Rogowski coil. The voltage sampling frequency is set to 10MHz and the current sampling frequency is set to 20MHz to ensure that the low-frequency dielectric loss characteristics and high-frequency partial discharge pulses can be captured simultaneously.

[0045] In an optional embodiment, a photoelectric hybrid sensor array can also be used for collection, converting voltage and current signals into optical signals for transmission, effectively avoiding interference of strong electromagnetic fields on weak signals, which is particularly suitable for ultra-high voltage cable testing environments and can improve the signal-to-noise ratio by 5-10dB.

[0046] Among them, the acquired voltage signal and current signal need to go through key steps such as signal preprocessing, feature space mapping and orthogonal decomposition, which are specifically implemented as follows: First, signal preprocessing is performed on the voltage sampling data and the current sampling data: Preprocessing operations are performed by adaptive wavelet denoising, baseline drift correction, outlier processing and phase synchronization correction. The denoising process adopts the same wavelet basis function selection strategy as step S1, using symlet wavelet for low-frequency signals and Daubechies wavelet for high-frequency pulse signals. The baseline correction window length is set to 3-5 oscillation cycles, and the phase correction accuracy is better than 0.01°, ensuring the accuracy of dielectric loss measurement.

[0047] After preprocessing, the continuous voltage signal and current signal are reconstructed by the B-spline interpolation algorithm and mapped to the preset feature space. When the preprocessed signal is mapped to the preset feature space, the manifold learning technology based on the kernel method is adopted. Specifically, the Gaussian radial basis function is used as the kernel function, and the bandwidth parameter is adaptively set to 0.5 times the average distance between samples. The mapping process includes two steps: feature embedding and manifold reconstruction. The embedding dimension is determined according to the cumulative variance contribution rate. Generally, the number of dimensions with a cumulative contribution rate of 95% is selected. The feature space includes the phase-amplitude plane, the harmonic component space, the time-frequency distribution plane and the statistical feature space, which is consistent with the feature manifold framework established in step S1. Through dimensionality reduction techniques such as principal component analysis, the most discriminative feature subspace is extracted to lay the foundation for subsequent analysis.

[0048] Secondly, the voltage signal and current signal are orthogonally decomposed to obtain mutually orthogonal sub-components. Specifically, the ensemble empirical mode decomposition algorithm is used to overcome the mode aliasing problem of the traditional EMD method by introducing noise assistance and extreme point optimization strategies. Time-frequency analysis is performed on each sub-component obtained by decomposition, and a hybrid algorithm of S transform and synchronous squeezing transform is used to achieve high-resolution time-frequency representation and reveal the distribution characteristics of signal energy on the time-frequency plane.

[0049] When performing time-frequency analysis on orthogonal components, the instantaneous energy density at each time-frequency point is calculated, and a three-dimensional energy distribution spectrum is constructed. The spectrum uses a logarithmic scale on the time axis to better show the energy distribution at different time scales. The energy distribution characteristics are obtained by calculating the statistical moments on the time-frequency plane, including the first-order moment (energy center), second-order moment (energy diffusion) and higher-order moment (energy skewness and kurtosis), forming a 12-dimensional feature vector.

[0050] Based on the time-frequency characteristics, the fuzzy C-means clustering algorithm is applied to divide all orthogonal components into dielectric loss related components and partial discharge related components. The clustering process adopts the same weight adjustment strategy as the Fisher discriminant analysis in step S1 to ensure classification accuracy. For the components in the fuzzy area, the system further applies wavelet packet energy entropy analysis for fine classification.

[0051] From the PD-related components, a dynamic threshold detection algorithm is used to identify PD pulse events, with a sensitivity coefficient set to 3.5-4.5. For each detected pulse event, the system extracts its waveform features, time features, and statistical features, and classifies the discharge type into internal discharge, surface discharge, corona discharge, or suspended discharge through a pattern recognition algorithm.

[0052] When extracting partial discharge features from partial discharge related components to form partial discharge pulse signals, six key parameters are first extracted for each detected partial discharge event: peak value, rise time, duration, phase angle, energy and waveform factor. These parameters are then integrated into a standardized partial discharge pulse signal through a weighted synthesis method, and the weight coefficient is determined by the Fisher discriminant ratio. The final partial discharge pulse signal retains the timing information and amplitude characteristics of the original partial discharge event, while removing random noise and interference components, and the signal-to-noise ratio is improved by 8-12dB.

[0053] From the dielectric loss related components, the fundamental frequency component is first extracted, the center frequency is set to the fundamental frequency (0.1Hz) of the oscillating ultra-low frequency voltage signal, and the bandwidth is 2% of the fundamental frequency. The phase difference between the voltage and current fundamental components is accurately calculated by combining zero crossing detection and Hilbert transform, and the dielectric loss tangent value is calculated based on this. At the same time, the system analyzes the change characteristics of the dielectric loss value with frequency to obtain the frequency response characteristics that reflect the dielectric polarization loss mechanism.

[0054] Finally, the degree of deviation of the acquired voltage and current signals from the pre-built characteristic manifold is analyzed. First, the key characteristic parameters are extracted to form a characteristic vector of the same dimension as in step S1, and the geodesic distance between it and the standard characteristic manifold is calculated. The degree of deviation is evaluated at multiple scales, and the overall deviation index is calculated comprehensively.

[0055] According to the deviation index value, the current test status is divided into three categories: normal status (deviation index <1.0), slight deviation (1.0≤deviation index <1.5) and severe deviation (deviation index ≥1.5). For severe deviation status, the parameter adjustment process is started; otherwise, the current parameter settings are maintained and the test process continues.

[0056] When parameters need to be adjusted, the optimal frequency and amplitude adjustment is calculated based on manifold learning and optimal control theory, with the goal of returning the test features to the standard manifold while minimizing the amplitude of parameter changes. Parameter adjustment is smooth to avoid sudden changes. After adjustment, the system automatically re-evaluates the deviation status and confirms the adjustment effect.

[0057] In an optional embodiment, a model predictive control method can be used to optimize the parameter adjustment process, taking into account the cumulative effect of multi-step adjustments and avoiding the shock that may be caused by simple adjustments. In addition, the system can also integrate a reinforcement learning module to continuously optimize the adjustment strategy and achieve adaptive control by recording the parameter adjustment history and effects.

[0058] Through the implementation of the above S2 step, the present invention can stably obtain the response signal of the tested cable under the excitation of the optimized oscillating ultra-low frequency voltage signal, accurately extract the dielectric loss characteristics and partial discharge pulse signal, and dynamically monitor and adjust the test state during the test to ensure data quality. This provides high-quality basic data support for subsequent insulation status evaluation.

[0059] S3: Calculate dielectric loss parameters based on the dielectric loss characteristic signal, determine the partial discharge position based on the partial discharge pulse signal, and generate a cable insulation status diagnosis report according to the dielectric loss parameters and the partial discharge position.

[0060] Specifically, determining the partial discharge position based on the partial discharge pulse signal includes: extracting the time characteristics of the partial discharge pulse signal; calculating the pulse propagation time difference; calculating the partial discharge source distance according to the propagation time difference and the cable propagation speed; and calibrating the physical position of the cable corresponding to the partial discharge source distance.

[0061] Specifically, the generation of the cable insulation status diagnosis report includes: determining the state of the insulating medium based on the dielectric loss parameter; determining the defect characteristics based on the partial discharge position; if the dielectric loss parameter exceeds the standard range and the partial discharge position shows a regular distribution, it is determined to be a complex fault; if only the dielectric loss parameter exceeds the standard range, it is determined to be insulation degradation; if only the partial discharge position shows a regular distribution, it is determined to be a local defect; and the insulating medium state and the defect characteristics are combined to form a diagnosis result including fault location information and reliability evaluation.

[0062] In a preferred embodiment of the present invention, the specific implementation process of step S3 is as follows: Based on the dielectric loss characteristic signal and partial discharge pulse signal obtained in step S2, the insulation status diagnosis is performed. First, the dielectric loss parameters are calculated to extract multi-dimensional features to reflect the overall degradation of the insulation medium; secondly, the partial discharge position is determined to accurately locate local defects; finally, these two types of information are comprehensively analyzed to generate a comprehensive insulation status diagnosis report.

[0063] The dielectric loss parameter is calculated using the frequency segment vector analysis method. The system calculates the baseband dielectric loss tangent value based on the phase difference and amplitude ratio extracted in step S2. , through the following vector formula: ; in, Indicates the phase difference between voltage and current. The measurement accuracy is better than 0.0001, which meets the measurement requirements of ultra-low loss XLPE cables.

[0064] Analyze the relationship between dielectric loss and frequency and construct the dielectric loss spectrum characteristics. to Get in range -Frequency curve. Fitted by Cole-Cole model: ; in, Indicates the zero-frequency dielectric loss value, Indicates the high-frequency limit dielectric loss value, represents the characteristic frequency, represents the frequency dispersion index, The system extracts these model parameters as dielectric loss spectrum characteristics to reflect the loss characteristics under different polarization mechanisms.

[0065] In an optional embodiment, a three-dimensional dielectric loss characterization technology can be used to construct a three-dimensional characteristic surface of the dielectric loss value with frequency and electric field strength to more comprehensively reflect the insulation degradation characteristics. This method is particularly suitable for degradation diagnosis of complex insulation systems and can improve the accuracy of water tree aging identification by 15-20%.

[0066] The measured dielectric loss value is corrected for temperature to ensure the comparability of test results at different temperatures. The correction adopts an exponential model: ; in, Indicates the dielectric loss value at standard temperature. Indicates the dielectric loss value at the measurement temperature. represents the temperature correction factor, To measure temperature, For standard temperature (20°C). For XLPE insulation, The value is 0.025-0.035; for oil-paper insulation, The value range is 0.045-0.055.

[0067] Analyze the voltage nonlinear characteristics of dielectric loss, which is a sensitive indicator of water tree aging. By fitting the voltage-dielectric loss curve: ; in, The electric field strength is The dielectric loss value at Indicates the reference electric field strength The dielectric loss value under is a nonlinear index. Healthy insulation Close to 0, water tree aging insulation Obviously greater than 0.

[0068] Calculate the historical change rate of dielectric loss , using normalization method to eliminate the influence of test condition differences: ; in, Represents the time interval between two tests (years). Through the above calculations, the system forms a multi-dimensional dielectric loss characteristic vector including base frequency dielectric loss value, frequency characteristic parameters, temperature correction coefficient, voltage nonlinear index and historical change rate.

[0069] For the determination of the PD position, the time characteristics of the PD pulse signal are first extracted. A multi-resolution singular point detection algorithm based on wavelet transform is used to accurately locate the pulse wavefront. The algorithm uses adaptive threshold technology to process the maximum value of the wavelet coefficient, and the positioning accuracy is better than 1 / 10 of the sampling time interval, achieving sub-sampling accuracy. The system simultaneously identifies the pulse peak time and termination time, and establishes a complete time series feature model.

[0070] The double-ended synchronous measurement or single-ended reflection method is used to calculate the propagation time difference of the partial discharge pulse. The double-ended measurement uses high-precision time synchronization technology based on the PTP protocol, and the synchronization error is less than 50ns. The system calculates the time difference between the same discharge event arriving at both ends. : ; in, and are the time it takes for the pulse to reach the two ends of the cable. In the reflection method, the system measures the original pulse Reflection echo from its end The time interval between : ; The system calculates the location of the PD source based on the propagation time difference and the pulse propagation velocity. For the two-terminal measurement method, the distance from the PD source to the reference end is Calculated as: ; in, is the total length of the cable, is the propagation speed of the pulse in the cable. For the reflection method, the distance from the partial discharge source to the measurement end is Calculated as: ; The present invention uses adaptive calibration technology to determine the pulse propagation velocity. By injecting calibration pulses at known locations in the cable, the actual propagation time is measured and the propagation velocity is dynamically calculated. The present invention takes into account the dispersion characteristics of different frequency components and improves positioning accuracy through a dispersion compensation algorithm. For XLPE insulated cables, the typical propagation velocity is 0.5-0.6 times the speed of light; for oil-paper insulated cables, the propagation velocity is 0.3-0.4 times the speed of light.

[0071] Furthermore, the calculated partial discharge source distance is mapped to the actual physical location of the cable. The present invention integrates the cable path geographic information system, which contains key node information such as the laying path, joint location and branch point. The distance value is converted into geographic coordinates or physical description position through piecewise linear mapping and node matching algorithm. The system takes into account the nonlinear factors of cable laying, such as the winding coefficient and the length change caused by temperature, to further improve the calibration accuracy. The final positioning accuracy is within 1% of the cable length.

[0072] Based on the calculated dielectric loss parameters, the system evaluates the insulation dielectric condition. The system uses a multi-threshold grading evaluation method to classify the insulation condition into five levels based on cable type-specific criteria: 1. Excellent status: , , ; 2. Good status means: , , ; 2. Note that the status is: , , ; 3. The deterioration state is: , , ; 4. Dangerous conditions are: , , .

[0073] in, Indicates the standard dielectric loss value of the cable type. It also analyzes the shape of the dielectric loss frequency characteristic curve to identify different types of insulation aging mechanisms, such as water tree aging, oxidation aging or thermal aging.

[0074] Furthermore, the location distribution characteristics of the partial discharge source are analyzed to determine the nature of the defect. The system identifies typical location distribution patterns: middle joint aggregation, terminal concentration, uniform distribution, segmented concentration, and single point protrusion. Define the partial discharge distribution variation coefficient As a quantitative indicator of the uniformity of the position distribution: ; in, is the location standard deviation, is the average position. represents uniform distribution, Indicates the aggregate distribution. The system combines the PD intensity, phase characteristics and waveform characteristics to further determine the defect type.

[0075] Furthermore, the cable fault type is determined by combining dielectric loss parameters and partial discharge location characteristics. The determination logic is as follows: 1. Compound failure: or ,and or 80% of the partial discharge is concentrated in a specific location; 2. Insulation degradation: only or , and the partial discharge activity is not obvious (discharge amplitude <50pC or discharge rate <10 times / minute); 3. Local defects: only or the partial discharges are regularly distributed, and and .

[0076] Based on the above analysis, the present invention forms a final diagnosis report. The report contains basic information, test results, status assessment, fault diagnosis and maintenance suggestions. The system adopts a predictive maintenance strategy to predict the optimal maintenance time window based on the degradation trend.

[0077] The present invention quantitatively evaluates the reliability of the diagnostic results. The evaluation indicators include data quality index , model fit Matching rate with historical cases The system uses a weighted calculation method to obtain the final confidence index CI: ; in, , and are the weight coefficients of each indicator. The system divides CI into three levels: high reliability (>90%), medium reliability (70%-90%), and low reliability (<70%) to help users correctly evaluate and use diagnostic results.

[0078] In an optional embodiment, a deep learning diagnostic model can be integrated, which can identify complex fault modes through a large number of historical case training. The model adopts a hybrid architecture of convolutional neural network and long short-term memory network. The input features include dielectric loss spectrum, partial discharge spectrum and location distribution map, and the output fault type probability distribution and location estimation. This method can improve the diagnostic accuracy by 15%-25% when the data volume is sufficient.

[0079] Through the implementation of the above S3 step, the present invention can accurately calculate the dielectric loss parameters and determine the partial discharge position based on the dielectric loss characteristic signal and the partial discharge pulse signal, and generate a comprehensive and accurate cable insulation status diagnosis report by integrating these two aspects of information. This diagnostic method combines global dielectric loss analysis with local defect location, realizes a comprehensive evaluation of the cable insulation status, provides a scientific basis for preventive maintenance and fault prediction, and effectively improves the reliability and safety of the power system.

[0080] In summary, the present invention optimizes the oscillating ultra-low frequency voltage signal by applying the manifold learning algorithm, and successfully integrates the three tests of dielectric loss test, withstand voltage test and partial discharge detection that need to be carried out separately into a one-time joint test, which effectively solves the problems of extended power outage time, equipment redundancy and accelerated insulation degradation caused by multiple tests in the prior art. The modified sinusoidal oscillation ultra-low frequency voltage signal used in the signal excitation module adopts a special waveform design, which significantly enhances the dielectric loss characteristics and the sensitivity of the partial discharge pulse while maintaining the requirements of the cable withstand voltage test, and improves the measurement accuracy and detection sensitivity. The signal processing module adopts a multi-resolution analysis method combined with a collective empirical mode decomposition algorithm to overcome the mode aliasing problem of the traditional EMD method, realize the efficient separation of the dielectric loss characteristic signal and the partial discharge pulse signal, and effectively improve the signal-to-noise ratio of the signal. The state diagnosis module establishes an accurate identification mechanism for complex faults, insulation degradation and local defects by comprehensively analyzing the dielectric loss parameters and the partial discharge position distribution characteristics, improves the localization accuracy of the partial discharge source, and breaks through the technical bottleneck that the traditional method is difficult to diagnose comprehensively. In general, the present invention realizes a comprehensive assessment of the insulation status of the cable and accurate fault location, provides a scientific basis for preventive maintenance of the power system, and effectively improves the reliability and safety of the power system.

[0081] Embodiment 2 is an embodiment of the present invention, which provides a cable oscillation ultra-low frequency withstand voltage and partial discharge integrated test system, including: The signal excitation module is used to apply an oscillating ultra-low frequency voltage signal to the cable under test. The oscillating ultra-low frequency voltage signal is optimized and adjusted through a manifold learning algorithm and serves as a common excitation source for withstand voltage test, dielectric loss test and partial discharge test. The signal processing module is used to obtain the voltage signal and current signal of the tested cable under the excitation of the oscillating ultra-low frequency voltage signal, analyze the deviation degree of the voltage signal and the current signal from the pre-built characteristic manifold, and extract the dielectric loss characteristic signal and the partial discharge pulse signal; The status diagnosis module is used to calculate the dielectric loss parameters based on the dielectric loss characteristic signal, determine the partial discharge position based on the partial discharge pulse signal, and generate a cable insulation status diagnosis report based on the dielectric loss parameters and the partial discharge position.

[0082] Example 3, reference Figure 2, is an embodiment of the present invention, which is different from the previous embodiment in that: if the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0083] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0084] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0085] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0086] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A cable oscillation ultra-low frequency withstand voltage and partial discharge integrated test method, characterized in that: include: Apply an oscillating ultra-low frequency voltage signal to the cable under test; The oscillating ultra-low frequency voltage signal is optimized and adjusted by a manifold learning algorithm and serves as a common excitation source for withstand voltage test, dielectric loss test and partial discharge test; Acquire the voltage signal and current signal of the tested cable under the excitation of the oscillating ultra-low frequency voltage signal, analyze the degree of deviation of the voltage signal and the current signal from the pre-built characteristic manifold, and extract the dielectric loss characteristic signal and the partial discharge pulse signal; A dielectric loss parameter is calculated based on the dielectric loss characteristic signal, a partial discharge position is determined based on the partial discharge pulse signal, and a cable insulation status diagnosis report is generated according to the dielectric loss parameter and the partial discharge position.

2. The cable oscillation ultra-low frequency withstand voltage and partial discharge integrated test method according to claim 1, characterized in that: The optimization and adjustment of the manifold learning algorithm includes: establishing a parameter space according to the type of cable under test; generating a test parameter matrix in the parameter space; collecting test response characteristics to construct a characteristic manifold; and optimizing the frequency, amplitude and waveform parameters of the oscillating ultra-low frequency voltage signal based on the characteristic manifold.

3. The cable oscillation ultra-low frequency withstand voltage and partial discharge integrated test method according to claim 2, characterized in that: Obtaining the voltage signal and current signal of the tested cable under the excitation of the oscillating ultra-low frequency voltage signal includes: collecting voltage sampling data and current sampling data; performing signal preprocessing on the voltage sampling data and the current sampling data; converting the preprocessed sampling data into the voltage signal and the current signal, and mapping them to a preset feature space.

4. The cable oscillation ultra-low frequency withstand voltage and partial discharge integrated test method according to claim 3, characterized in that: The extraction of dielectric loss characteristic signal and partial discharge pulse signal comprises: performing orthogonal decomposition on the voltage signal and the current signal to obtain orthogonal components; performing time-frequency analysis on the orthogonal components to obtain energy distribution characteristics; distinguishing the orthogonal components into dielectric loss related components and partial discharge related components based on the energy distribution characteristics; extracting partial discharge characteristics from the partial discharge related components to form the partial discharge pulse signal; and extracting phase difference and amplitude ratio parameters from the dielectric loss related components to form the dielectric loss characteristic signal.

5. The cable oscillation ultra-low frequency withstand voltage and partial discharge integrated test method according to claim 4, characterized in that: Analyzing the degree of deviation of the voltage signal and the current signal from a pre-established characteristic manifold includes: calculating a distance parameter between the characteristic parameters of the voltage signal and the current signal and a standard characteristic manifold; determining a deviation state according to the distance parameter; if the deviation state exceeds a reference range, returning the deviation state to the reference range by adjusting the frequency and amplitude of the oscillating ultra-low frequency voltage signal; otherwise, maintaining the current oscillating ultra-low frequency voltage signal parameters and continuing the test process.

6. The cable oscillation ultra-low frequency withstand voltage and partial discharge integrated test method according to claim 5, characterized in that: Determining the partial discharge position based on the partial discharge pulse signal includes: extracting the time feature of the partial discharge pulse signal; calculating the pulse propagation time difference; calculating the partial discharge source distance according to the propagation time difference and the cable propagation speed; and calibrating the cable physical position corresponding to the partial discharge source distance.

7. The cable oscillation ultra-low frequency withstand voltage and partial discharge integrated test method according to claim 6, characterized in that: A cable insulation status diagnostic report is generated according to the dielectric loss parameter and the partial discharge position, including: determining the state of the insulating medium based on the dielectric loss parameter; determining the defect characteristics based on the partial discharge position; if the dielectric loss parameter exceeds the standard range and the partial discharge position presents a regular distribution, it is determined to be a composite fault; if only the dielectric loss parameter exceeds the standard range, it is determined to be insulation degradation; if only the partial discharge position presents a regular distribution, it is determined to be a local defect; and the insulating medium state and the defect characteristics are combined to form a diagnostic result including fault location information and reliability evaluation.

8. A cable oscillation ultra-low frequency withstand voltage and partial discharge integrated test system, based on the cable oscillation ultra-low frequency withstand voltage and partial discharge integrated test method according to any one of claims 1 to 7, characterized in that: include, A signal excitation module is used to apply an oscillating ultra-low frequency voltage signal to the cable under test. The oscillating ultra-low frequency voltage signal is optimized and adjusted by a manifold learning algorithm and serves as a common excitation source for withstand voltage test, dielectric loss test and partial discharge test; A signal processing module, used to obtain the voltage signal and current signal of the tested cable under the excitation of the oscillating ultra-low frequency voltage signal, analyze the degree of deviation of the voltage signal and the current signal from the pre-built characteristic manifold, and extract the dielectric loss characteristic signal and the partial discharge pulse signal; The status diagnosis module is used to calculate the dielectric loss parameter based on the dielectric loss characteristic signal, determine the partial discharge position based on the partial discharge pulse signal, and generate a cable insulation status diagnosis report according to the dielectric loss parameter and the partial discharge position.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the cable oscillation ultra-low frequency withstand voltage and partial discharge integrated testing method described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the cable oscillation ultra-low frequency withstand voltage and partial discharge integrated testing method according to any one of claims 1 to 7 are implemented.

Images