Insulation polarization current information extraction method, system, device and medium

Abstract

The application discloses an insulation polarization current information extraction method, system, device and medium, and relates to the technical field of power equipment insulation state monitoring. The prior art only relies on single time point current data, ignores wide time domain dynamic change information, and causes the problem of incomplete evaluation. The application comprises the following steps: placing an insulation material sample in a three-electrode system to measure polarization current data in a wide time domain window; fitting the current data with a multi-exponential relaxation function to eliminate the interference of local current overshoot and extract the decay trend; extracting a transient characteristic quantity representing a transient process and a quasi-steady state characteristic quantity representing a quasi-steady state process from the fitted data; normalizing the two characteristic quantities to eliminate dimensional differences; substituting the normalized characteristic quantities into an S-shaped logic function to calculate a comprehensive state index; and outputting the comprehensive state index as characteristic information for evaluating the state of the insulation material. The technical scheme can comprehensively and accurately extract insulation polarization current information and evaluate the insulation state.

Description

Technical Field

[0001] This invention relates to the field of power systems, and more particularly to methods, systems, equipment, and media for extracting insulation polarization current information. Background Technology

[0002] In the condition assessment of high-voltage insulation materials, conductivity testing is one of the important means to evaluate their insulation performance. This method measures the polarization current and, based on the quasi-steady-state current value at the end of the polarization time, calculates the conductivity and threshold electric field at the corresponding moment, using these as key characteristic quantities for determining the insulation condition.

[0003] However, existing methods for evaluating insulation performance based on polarization current still have the following shortcomings: (1) Current data at a single moment is usually selected for evaluation, ignoring the insulation state information carried by the dynamic changes of current in a wide time domain during polarization, making it difficult to be fully representative; (2) The evaluation dimensions are relatively limited, and existing methods mostly use quasi-steady-state conductivity or threshold electric field as the basis for judgment, ignoring the characteristic quantities of the transient response of the material under the action of the electric field.

[0004] Taking the patent "A Method and System for Evaluating the Insulation Status of High-Voltage DC Cables" (Publication No.: CN202410637793.1) as an example, this method involves preparing samples by slicing the cable insulation or pressing the insulation material into sheets, measuring the polarization current of the samples under different electric field strengths and temperatures, and then substituting the obtained data into a theoretical model of a coaxial cable for status evaluation. The limitation of this method is that it relies solely on the current measurement value at a single time point as the evaluation basis, resulting in a lack of comprehensive representativeness. Over longer polarization times, the quasi-steady-state conductivity of intact samples and slightly aged samples may become similar, leading to a lack of comprehensive representativeness in the evaluation results based on a single moment's measurement parameters, which is insufficient to support accurate judgment of the insulation status. Furthermore, since the polarization current varies significantly at different time points, its evolution over time contains important insulation status information, but this method ignores current changes over a wide time domain.

[0005] To address the aforementioned issues, this patent proposes a wide-time-domain window polarization current information extraction method, which fully explores the polarization current's full-time-domain variation information and establishes an insulation performance evaluation model by combining transient and quasi-steady-state characteristic quantities. Summary of the Invention

[0006] The technical problem to be solved and the technical task proposed by this invention is to improve and refine existing technical solutions, and to provide a method, system, device, and medium for extracting insulation polarization current information, so as to achieve the purpose of comprehensively and accurately extracting insulation polarization current information and assessing insulation status. To this end, this invention adopts the following technical solution.

[0007] The first technical solution of the present invention is: a method for extracting insulation polarization current information based on a wide time-domain window, which includes the following steps:

[0008] S1: Place the solid insulating material sample to be evaluated between the high voltage electrode and the protected electrode of the three electrodes, and set a ring-shaped protective electrode around the protected electrode to eliminate surface leakage current interference.

[0009] S2: Measure the polarization current data of the insulating material within a wide time-domain window, and fit the current data using a multi-exponential relaxation function to eliminate the interference of local current overshoot and extract the current decay trend; wherein, the multi-exponential relaxation function is a linear superposition of multiple exponential decay terms with different time constants and a constant term of a DC current component.

[0010] S3: Based on the fitted current data, extract transient characteristic quantities that characterize the transient process and quasi-steady-state characteristic quantities that characterize the quasi-steady-state process;

[0011] S4: Normalize the transient and quasi-steady-state characteristic quantities respectively to eliminate dimensional differences;

[0012] S5: Substitute the normalized transient and quasi-steady-state characteristics into an S-type logic function to calculate a comprehensive state index between 0 and 1, which is used to evaluate the insulation state of the insulating material. The larger the comprehensive state index value, the worse the insulation performance.

[0013] S6: Output the comprehensive state index value as characteristic information extracted from the polarization current for evaluating the state of the insulating material.

[0014] This technical solution employs a three-electrode system and a protective electrode measurement architecture, effectively eliminating interference from surface leakage current and ensuring that the measured current signal truly reflects the polarization process within the insulating material. This provides an accurate and pure data foundation for all subsequent advanced information extraction. Using a multi-exponential relaxation function for fitting enables high-precision matching of complex current decay curves composed of multiple polarization mechanisms (fast, medium, and slow relaxation). It effectively suppresses local current overshoot noise during measurement, extracting a smooth and realistic current decay trend, fundamentally avoiding feature extraction errors caused by data fluctuations. The function structure, consisting of multiple exponential decay terms plus a constant term, directly corresponds to the physical polarization process of the insulating material, making the fitting results not only mathematically accurate but also possessing clear physical interpretability. Feature quantities are extracted from both transient and quasi-steady-state dimensions. Transient features reflect the speed of polarization establishment, while quasi-steady-state features reflect the final insulation strength of the material. This dual-feature fusion method provides richer state information, enabling a more comprehensive reflection of the insulating material's health status. By using a sigmoid logic function, the normalized dual features are fused into a comprehensive state index between 0 and 1, making the evaluation results very intuitive, easy to understand and compare.

[0015] As a preferred technical means: the exponential relaxation function in step S1 includes three exponential decay terms and one constant term. The three exponential decay terms are the first coefficient multiplied by the negative time of the natural constant divided by the first time constant power, plus the second coefficient multiplied by the negative time of the natural constant divided by the second time constant power, plus the third coefficient multiplied by the negative time of the natural constant divided by the third time constant power.

[0016] The function structure employed in this technical solution provides extremely high mathematical flexibility, enabling it to approximate complex, non-single-exponential current decay curves in actual measurements with exceptionally high accuracy. This is because the three exponential terms each have independent coefficients (…). ) and time constant ( These six adjustable parameters enable the function to characterize the fast, medium, and slow relaxation processes during current decay. Compared to single-exponential or double-exponential models, the triple-exponential model captures more subtle decay details, thus significantly reducing fitting errors and obtaining trend curves that more closely resemble the actual physical process.

[0017] This function is not a purely mathematical abstraction; each term corresponds to a specific physical polarization mechanism in the insulating material. Each exponentially decaying term can be understood as an equivalent capacitance of different sizes in the dielectric, accurately characterizing the dielectric polarization process. The constant term characterizes the DC component of the steady-state conduction current. Therefore, using this function for fitting is essentially a mathematical "decoupling" of multiple overlapping physical processes, allowing for a deeper analysis of the impact of various mechanisms on the overall polarization.

[0018] This technical solution effectively smooths and suppresses random fluctuations in measured data, extracting a stable and accurate decay trend. This is because by approximating discrete, noisy data points with a smooth, physically driven function curve, random interference in the measurement can be filtered out. This avoids the significant errors caused by directly differentiating or calculating the rate of change of the drastically fluctuating raw data, laying a reliable foundation for the subsequent accurate extraction of characteristic parameters (such as decay rate).

[0019] For most solid insulating materials, the polarization process mainly involves three dominant mechanisms. The three-exponential model adopted in this technical solution includes enough parameters to accurately describe this physical reality, while avoiding overfitting caused by introducing too many parameters (i.e., the model fits not only the trend but also the noise). This allows the model to maintain good robustness and practicality while ensuring accuracy.

[0020] As a preferred technical means: the transient characteristic quantity is the average value of the rate of change of current within a fixed time step in the initial stage of the polarization process; the quasi-steady-state characteristic quantity is the threshold electric field calculated by the steady-state current density under different electric fields.

[0021] The dynamic decay of current during the initial polarization phase is highly sensitive to changes in the microstructure of insulating materials. Extracting the rate of change within this time window as a feature quantity allows for earlier and more sensitive capture of material aging information, outperforming methods that rely solely on the final steady-state value. Transient feature quantities are not isolated current points but quantify the rate of current change, directly reflecting the speed of polarization establishment and thus providing dynamic information about the material's dielectric relaxation characteristics—information unavailable from static steady-state values. By calculating the average value over a fixed time step, rather than a single instantaneous rate of change, short-term random fluctuations in the data are effectively smoothed, making the extracted feature quantities more stable and reliable. The threshold electric field is a critical parameter that marks the transition of a material from ohmic conductivity at low electric fields to nonlinear conductivity at high electric fields (such as space charge confinement current). It reflects the material's ability to withstand electric fields, i.e., its intrinsic insulating strength, more directly than a single steady-state current value. Compared to directly using the current value under a specific field strength, the threshold electric field calculated through a series of measurements is a derived characteristic parameter, less affected by fluctuations in a single measurement, and can more stably characterize the long-term insulation performance of the material. Quasi-steady-state characteristic quantities have a clear physical picture and are closely related to the conductivity mechanism of insulating materials, making the evaluation results not just numbers, but also possessing profound physical interpretation. This technical solution extracts features from both dynamic (transient) and static (quasi-steady-state) dimensions. Transient characteristic quantities capture the "speed" of the process, while quasi-steady-state characteristic quantities reflect the final "strength." The combination of the two provides a more comprehensive and three-dimensional assessment of the insulation state, overcoming the limitations of single-dimensional information. It establishes a multi-dimensional evaluation index system for insulation state diagnosis, more comprehensively reflecting the complex aging conditions of materials, and improving the accuracy and reliability of diagnosis.

[0022] As a preferred technical means: the normalization process of the transient feature quantity in step S4 is to subtract the minimum absolute value of the feature quantity among all samples from the absolute value of the feature quantity, and then divide by the difference between the maximum and minimum absolute values ​​of the feature quantity among all samples.

[0023] The normalization process for quasi-steady-state features involves subtracting the minimum value of the feature from the value of the feature across all samples, and then dividing by the difference between the maximum and minimum values ​​of the feature across all samples.

[0024] This technical solution not only eliminates the dimensions, but also unifies the evaluation direction through processing such as taking absolute values. More importantly, it truly preserves the relative gradient of the aging state, making the output results both scientific and intuitive, and very convenient for subsequent comprehensive evaluation and engineering applications.

[0025] As a preferred technical means: the S-type logic function in step S5 is: the comprehensive state index S equals 1 plus the first weight in the negative brackets of the natural constant multiplied by the normalized transient characteristic, plus the second weight multiplied by the normalized quasi-steady-state characteristic, plus the reciprocal of the power of a bias term.

[0026] The sigmoid logic function monotonically maps the calculation result of a linear combination to a value between 0 and 1. This allows the output comprehensive state index S to be intuitively understood as the probability or confidence level of a "poor" insulation state. The closer the S value is to 1, the higher the probability that the system is "confident" that the material's insulation state is poor; the closer it is to 0, the higher the probability that the state is good. The sigmoid logic function is a nonlinear function, highly sensitive to input changes within the intermediate range, but insensitive to input changes at extreme values. For samples with states near the critical point, even a small deterioration can cause a significant change in the S value, thus issuing a sensitive warning. For samples with extremely good or poor states, even with some parameter fluctuations, the S value will stabilize near 0 or 1, avoiding excessive oscillations in the evaluation results and enhancing the robustness of the evaluation. The function input is a linear weighted sum of normalized features, including weight coefficients and bias terms. By adjusting the weights and biases, the relative contributions of transient and quasi-steady-state features to the final evaluation result can be precisely controlled. Regardless of how the input feature values ​​change, the function's output is strictly limited to the (0,1) interval, ensuring a stable output range and facilitating system integration and comparison.

[0027] As a preferred technical means: the start time of the wide time domain window is 5 seconds after the system switching action, and the end time is 600 seconds; the average current change rate within a fixed time step of 5-100s during the initial polarization is calculated, and the average current change rate is used as a transient characteristic quantity; the fitted current value at 600s is selected as the quasi-steady-state current value, and the corresponding steady-state current density is calculated in combination with the measured electric field, thereby determining the threshold electric field, and the threshold electric field is used as a quasi-steady-state characteristic quantity.

[0028] This technical solution considers both transient and steady-state conditions, capturing the complete relaxation process. The start time is set at 5 seconds, effectively avoiding electromagnetic interference generated during high-voltage switch operation and the charging process caused by coupling capacitors in the equivalent topology of the measurement system. This prevents the acquisition of data that cannot effectively represent the dielectric polarization process, ensuring the stability and reliability of the initial data and providing a "clean" starting point for subsequent accurate fitting. The end time is set at 600 seconds, allowing the polarization process of most insulating materials (such as cross-linked polyethylene) to approach a quasi-steady state, thus enabling accurate extraction of steady-state current information. At the same time, the time is not too long, ensuring the engineering practicality and efficiency of the test, making it suitable for widespread application in laboratories or field settings.

[0029] Selecting the initial polarization phase (5-100s) allows for the most dramatic decay of the polarization current in the insulating material. This phase contains the richest dynamic information and is most sensitive to the material's aging state. Extracting features from this range enables the early detection of subtle signs of insulation performance degradation. Calculating the average rate of change within this range, compared to selecting the current value at a single time point, provides a more stable characteristic, stronger resistance to random noise interference, and a more reliable characterization of the decay rate trend. Selecting the fitted value at 600s as the quasi-steady-state current avoids potential fluctuation errors from directly reading the raw 600s data. By extrapolating the fitted function at 600s, a smoother and more accurate steady-state current estimate is obtained. In a 10-minute test, both the high dynamic information in the initial phase (before 100s) and the quasi-steady-state information in the final phase (600s) are fully utilized, maximizing test time utilization.

[0030] As a preferred technical approach: when the solid insulation material to be evaluated is cross-linked polyethylene, the first and second weights in the S-type logic function are both 0.5, and the bias term is -1, in order to adapt to the polarization characteristics and aging behavior of cross-linked polyethylene materials.

[0031] This technical solution balances the contribution of different feature information to the evaluation results and simplifies the application process of the model. It ensures that when performing state evaluation on XLPE, a specific material, it can output highly reliable, consistent, and easy-to-interpret results.

[0032] The second technical solution of the present invention is: an insulation polarization current information extraction system based on a wide time-domain window, employing the aforementioned insulation polarization current information extraction method based on a wide time-domain window, wherein the insulation polarization current information extraction system includes:

[0033] The data acquisition module is used to acquire polarization current data of insulating materials measured within a wide time-domain window;

[0034] The data processing module, which is communicatively connected to the data acquisition module, is used to fit the polarization current data using a multi-exponential relaxation function to suppress data fluctuations and obtain a smooth current decay trend curve.

[0035] The feature extraction module is communicatively connected to the data processing module and is used to extract a transient feature quantity to characterize the rate of transient current decay in the early stage of polarization, and a quasi-steady-state feature quantity to characterize the insulation strength of the material, based on the fitted current decay trend curve.

[0036] The normalization module is communicatively connected to the feature extraction module and is used to normalize the transient feature quantities and quasi-steady-state feature quantities to obtain dimensionless first normalization parameters and second normalization parameters, respectively.

[0037] The state assessment module is communicatively connected to the normalization module. It is used to linearly combine the first normalization parameter and the second normalization parameter, and input the combination result into an S-type logic function to map a comprehensive state index with a value range between 0 and 1. The magnitude of the comprehensive state index value is positively correlated with the degree of deterioration of the insulation material.

[0038] The output module, which is communicatively connected to the state assessment module, is used to output the comprehensive state index value as characteristic information extracted from the polarization current for assessing the state of the insulating material.

[0039] The data processing module uses a multi-exponential relaxation function to fit the polarization current data, effectively suppressing local fluctuations in the original data while preserving the complete decay trend of the current from transient to quasi-steady state within a wide time-domain window. This balances data smoothness and trend integrity, providing stable and reliable foundational data for subsequent feature extraction. The feature extraction module extracts "transient features characterizing the rate of decay of the transient current in the initial stage of polarization" and "quasi-steady-state features characterizing the insulation strength of the material," corresponding to the dynamic changes and static strength characteristics of the insulation material during polarization, respectively. This achieves multi-dimensional, full-cycle information capture of the insulation state, avoiding the limitations of single feature representation. The normalization module transforms the transient and quasi-steady-state features into dimensionless normalized parameters, resolving the dimensional differences caused by the different physical meanings of the two types of features. This provides a unified basis for comparison and calculation of feature information that was previously impossible to combine directly, making linear combination feasible for the subsequent state assessment module. The condition assessment module maps the combined normalized parameters to a "comprehensive condition index between 0 and 1" using an S-shaped logic function. The index magnitude is positively correlated with the degree of insulation degradation, simplifying and integrating complex feature information. The quantification results are intuitive and easy to understand, facilitating rapid judgment and comparison of insulation condition in engineering practice. From data acquisition to outputting the comprehensive condition index, the system forms a complete closed loop of "raw data acquisition → processing → feature extraction → normalization → assessment → result output," directly producing feature information usable for insulation condition assessment, thus enhancing the system's practicality and engineering application value.

[0040] The third technical solution of the present invention is: a computer device, the device including one or more processors and one or more memories, wherein the one or more memories store at least one piece of program code, and when the program code is executed by the one or more processors, it implements the aforementioned method for extracting insulation polarization current information based on a wide time-domain window.

[0041] The fourth technical solution of the present invention is: a storage medium storing at least one piece of program code, characterized in that, when the program code is executed by a processor, it implements the steps of the aforementioned method for extracting insulation polarization current information based on a wide time-domain window.

[0042] Beneficial effects: This technical solution enriches the evaluation dimensions of insulation status by extracting transient and quasi-steady-state characteristic quantities from the polarization current of the insulation material, realizes quantitative assessment of insulation status, effectively improves the accuracy of insulation assessment, and is applicable to insulation status assessment of high-voltage cables and transformers. Attached Figure Description

[0043] Figure 1 This is a numerical diagram of the polarization current of a cross-linked polyethylene sample under a polarization field.

[0044] Figure 2 This is the comprehensive state index of the present invention. S Schematic diagram.

[0045] Figure 3 This is a graph showing the measured data of the cross-linked polyethylene sample of this invention. Detailed Implementation

[0046] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings.

[0047] Example 1

[0048] S1: Measurement Preparation and Three-Electrode System Setup

[0049] 1. Measurement Equipment Selection: A three-electrode system conforming to IEC 62631 is adopted to eliminate the interference of surface leakage current on polarization current measurement, ensuring that the data only reflects the polarization process within the insulating material. The solid insulating material sample to be evaluated is placed between the high-voltage electrode and the protected electrode of the three-electrode system. A ring-shaped protective electrode is set around the protected electrode to form a core current path of "high-voltage electrode-sample-protected electrode," while the protective electrode shields the leakage current on the sample surface.

[0050] 2. Volume resistivity calculation logic: The volume resistivity of solid insulating materials is a core parameter characterizing their insulation state. Based on measurement data from a three-electrode system, the volume resistivity is calculated using the following formula ( ρ ) and volume conductivity ( σ ):

[0051] (1) Formula for calculating volume resistivity: ,in The volume resistivity of the sample. U This is the external voltage applied between the high-voltage electrode and the protected electrode. I The polarization current flowing through the sample is measured.

[0052] (2) Formula for calculating the effective area of ​​the electrode: ,in The effective area of ​​the electrode. The diameter of the ring-shaped protective electrode. To protect the gap between the electrode and the protected electrode;

[0053] (3) Formula for calculating volume resistivity: ,in The average thickness of the sample to be evaluated;

[0054] (4) Formula for calculating volume conductivity: ,in σ The volume conductivity is negatively correlated with insulation performance (the lower the conductivity, the better the insulation performance).

[0055] Note: Other electrode size parameters involved in the formula are defined as follows: The inner diameter of the annular protective electrode. The outer diameter of the annular protective electrode. The diameter of the protected electrode (measuring electrode) is given. The above parameters need to be calibrated in advance according to the IEC62631 standard and sample size.

[0056] 3. Physical characteristics of polarization current: Insulating materials can be equivalently represented as a capacitor-resistor parallel (RC) circuit. Therefore, the measured value of polarization current shows a trend of "gradually decreasing with time". Its ideal decay law can be expressed as:

[0057]

[0058] in To initially apply voltage, Equivalent resistance This is the equivalent capacitance. For polarization time, The time constant (reflecting the rate of current decay) The smaller the value, the faster the current decays and the worse the insulation performance.

[0059] In actual measurements, the polarization current may fluctuate locally due to electromagnetic interference, system noise, etc. Taking a 200μm thick cross-linked polyethylene sample as an example (test polarization field strength 35kV / mm, test time 600s): there is a 5s delay in the system switching action, so the effective recording time range is 5-600s; the measured current is 175pA at 5s, drops to 14pA at 100s, and is about 10pA in the quasi-steady-state stage at 600s, but there is still significant fluctuation in the current in this stage, which needs to be eliminated by subsequent fitting processing.

[0060] S2: Polarization current data preprocessing

[0061] Traditional methods, which select a single-moment value or calculate the average value over the last few seconds as the current value under quasi-steady-state conditions, are easily affected by numerical fluctuations in accuracy. Therefore, this embodiment selects a multi-exponential relaxation function to fit the current data. The advantages of this process are: (1) it smooths current fluctuations within the entire time-domain window, avoiding calculation errors caused by local fluctuations when extracting current changes over a wide time domain; (2) it stably characterizes the current trend, avoiding the influence of the number of time points selected for the average value when calculating the threshold electric field or conductivity value under quasi-steady-state conditions. This embodiment preferably uses a tri-exponential relaxation function to fit the original polarization current data. The function expression is:

[0062]

[0063] in, I The fitted polarization current value is... t Polarization time;

[0064] , , These are the current amplitude coefficients for the three polarization stages of "fast relaxation", "medium relaxation" and "slow relaxation", which quantify the contribution of each stage to the total polarization current (the larger the coefficient, the higher the initial current amplitude of the corresponding stage).

[0065] , , These are the time constants for the three polarization stages, reflecting the rate of current decay in each stage. < < (corresponding to the early, middle, and late stages of polarization, respectively).

[0066] To fit the reference current, the basic value of the quasi-steady-state current after the polarization process has stabilized is given.

[0067] The fitting result must satisfy the coefficient of determination R. 2 A value greater than 0.994 ensures that the fitted curve faithfully reflects the decay trend of the original data while smoothing out local fluctuations.

[0068] S3: Extraction of transient and quasi-steady-state features

[0069] Based on the fitted data, the current value change within 5-100s accounts for 97.6% of the change within 5-600s. Therefore, the mean current change rate within a fixed time step (e.g., 10s) during the initial polarization phase was calculated. α , as a characteristic parameter x1. Directly select the current value at 600s as the steady-state current value, and further calculate the steady-state current density under different electric fields to obtain the threshold electric field. E th , as a characteristic parameter x 2.

[0070] S4: Normalization

[0071] Feature parameters x 1 and x The numerical range and units of 2 differ, making it impossible to use them as a standard for evaluation. Normalization is used here to convert them into a unified numerical range, ensuring their fair participation in subsequent model calculations.

[0072] As aging intensifies, the time constant... RC = τ There will be a decreasing trend, the current decays rapidly, and the characteristic parameters will decrease. x 1 is negative; for ease of calculation, the characteristic parameter is... x 1. Absolute value range normalization is adopted, and the normalization formula is:

[0073]

[0074] in α This represents the average current decay rate of the current sample. α min and α max The mean of the minimum and maximum decay rates observed in different aged samples. x 1′∈[0,1] is the result after normalization.

[0075] As aging intensifies, the threshold electric field tends to decrease, therefore the characteristic parameter x 2. Use the positive correlation normalization formula:

[0076]

[0077] in E th The threshold electric field of the current sample. E min and E max The minimum and maximum threshold electric fields observed for different aged samples. x 2′∈[0,1] represents the normalized result. Through the above processing, the dimensional standards of the two characteristic parameters are unified.

[0078] S5: Overall Status Index S Evaluate

[0079] After normalization, the result is substituted into the sigmoid logic function to obtain the overall state index. S :

[0080]

[0081] Comprehensive Status Index S The larger the size, the worse the insulation performance. The weights for the normalized quasi-steady-state characteristics (usually negative, due to the threshold electric field) E th It is positively correlated with insulation performance, but its value decreases with aging, so a negative weight is needed to make its contribution to the S value positively correlated with the degree of degradation. Weights for normalized transient characteristics (usually set to positive values ​​because of the mean current decay rate) It is negatively correlated with insulation performance, and its value increases with aging. Therefore, positive weighting makes its contribution to the S value positively correlated with the degree of degradation. b For the bias term (adjustment curve in) S =0.5 (position). Comprehensive state index S Schematic diagram as follows Figure 2 As shown, the horizontal axis is The vertical axis is S ∈(0,1), when z When =0, S =0.5 is the critical point for the aging process. The parameters are adjusted based on the type of material being evaluated and the aging state. , , For example, cross-linked polyethylene materials can... , All are set to 0.5. b Set it to -1.

[0082] Taking the measured data of cross-linked polyethylene samples aged for 24 hours and 72 hours as examples, the measured data are as follows: Figure 3 As shown. Calibration based on a large amount of experimental data. α min , α max , E min and E max The following results were then obtained from the calculation:

[0083]

[0084] The measurement data shows that the comprehensive state index SThe evaluation dimensions were expanded to include changes in the quasi-steady-state threshold electric field and current changes over a wide time domain, which can effectively characterize insulation state parameters. As the aging time increased from 24h to 72h, the comprehensive state index S increased from 0.553 to 0.558, conforming to the pattern of "accelerated aging → increased S", thus verifying the effectiveness of this method.

[0085] S6: Output the comprehensive state index value as characteristic information extracted from the polarization current for evaluating the state of the insulating material. It can be displayed on a computer screen or transmitted to a database or a higher-level monitoring system.

[0086] At the same time, it can output the original fitting curve, feature values ​​and other intermediate results for users to refer to.

[0087] Example 2

[0088] An insulation polarization current information extraction system based on a wide time-domain window, employing the aforementioned insulation polarization current information extraction method based on a wide time-domain window, the insulation polarization current information extraction system comprising:

[0089] The data acquisition module is used to acquire polarization current data of insulating materials measured within a wide time-domain window;

[0090] The data processing module, which is communicatively connected to the data acquisition module, is used to fit the polarization current data using a multi-exponential relaxation function to suppress data fluctuations and obtain a smooth current decay trend curve.

[0091] The feature extraction module is communicatively connected to the data processing module and is used to extract a transient feature quantity to characterize the rate of transient current decay in the early stage of polarization, and a quasi-steady-state feature quantity to characterize the insulation strength of the material, based on the fitted current decay trend curve.

[0092] The normalization module is communicatively connected to the feature extraction module and is used to normalize the transient feature quantities and quasi-steady-state feature quantities to obtain dimensionless first normalization parameters and second normalization parameters, respectively.

[0093] The state assessment module is communicatively connected to the normalization module. It is used to linearly combine the first normalization parameter and the second normalization parameter, and input the combination result into the S-type logic function to map it into a comprehensive state index with a value range between 0 and 1. The magnitude of the comprehensive state index value is positively correlated with the degree of deterioration of the insulation material.

[0094] The output module, which is communicatively connected to the state assessment module, is used to output the comprehensive state index value as characteristic information extracted from the polarization current for assessing the state of the insulating material.

[0095] The data processing is highly targeted, effectively suppressing fluctuations and preserving the complete decay trend through multi-exponential relaxation function fitting, providing a reliable foundation for feature extraction; the feature extraction dimensions are comprehensive, simultaneously capturing transient decay characteristics and quasi-steady-state insulation strength information, avoiding the limitations of single feature representation; normalization processing eliminates dimensional differences, laying the foundation for comprehensive evaluation of feature quantities; the evaluation results are intuitively quantified, mapped to a comprehensive state index of 0-1 through an S-shaped logistic function, and positively correlated with the degree of insulation degradation, facilitating engineering applications; a complete closed loop of "data acquisition-processing-evaluation-output" is formed, realizing the integration of information extraction and state evaluation, making it highly practical.

[0096] It is understood that the detailed functional implementation of each of the above modules can be found in the description of the aforementioned method embodiments, and will not be elaborated further here.

[0097] Example 3

[0098] This embodiment provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements a method for extracting insulation polarization current information based on a wide time-domain window as described in any embodiment of the present invention.

[0099] Example 4

[0100] This embodiment provides a computer-readable storage medium storing a computer program thereon. When the computer program is executed by a processor, it implements a method for extracting insulation polarization current information based on a wide time-domain window as described in any embodiment of the present invention.

[0101] Those skilled in the art will recognize that the units and algorithm steps described in the embodiments disclosed in this invention can be implemented using electronic hardware, computer software, or a combination of electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0102] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0103] In the several embodiments provided in this application, any function, if implemented as a software functional unit and sold or used as an independent product, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0104] The insulation polarization current information extraction method, system, device and medium shown above are specific embodiments of the present invention, which have demonstrated the substantial features and progress of the present invention. Equivalent modifications can be made to them according to actual use needs, under the guidance of the present invention, and all such modifications are within the protection scope of this solution.

Claims

1. A method for extracting insulation polarization current information based on a wide time-domain window, characterized in that: Includes the following steps: S1: Place the solid insulating material sample to be evaluated between the high voltage electrode and the protected electrode of the three electrodes, and set a ring-shaped protective electrode around the protected electrode to eliminate surface leakage current interference. S2: Measure the polarization current data of the insulating material within a wide time-domain window, and fit the current data using a multi-exponential relaxation function to eliminate the interference of local current overshoot and extract the current decay trend; wherein, the multi-exponential relaxation function is a linear superposition of multiple exponential decay terms with different time constants and a constant term of a DC current component. S3: Based on the fitted current data, extract transient characteristic quantities that characterize the transient process and quasi-steady-state characteristic quantities that characterize the quasi-steady-state process; S4: Normalize the transient and quasi-steady-state characteristic quantities respectively to eliminate dimensional differences; S5: Substitute the normalized transient and quasi-steady-state characteristics into an S-type logic function to calculate a comprehensive state index between 0 and 1, which is used to evaluate the insulation state of the insulating material. The larger the comprehensive state index value, the worse the insulation performance. S6: Output the comprehensive state index value as characteristic information extracted from the polarization current for evaluating the state of the insulating material.

2. The method for extracting insulation polarization current information based on a wide time-domain window according to claim 1, characterized in that, The exponential relaxation function described in step S2 includes three exponentially decaying terms and one constant term. The three exponentially decaying terms are: ;in, , , These are the current amplitude coefficients for the three polarization stages; , , These are the time constants for the three polarization stages. t For polarization time.

3. The method for extracting insulation polarization current information based on a wide time-domain window according to claim 1, characterized in that, The transient characteristic quantity is the average value of the rate of change of current within a fixed time step in the initial stage of the polarization process; the quasi-steady-state characteristic quantity is the threshold electric field calculated by the steady-state current density under different electric fields.

4. The method for extracting insulation polarization current information based on a wide time-domain window according to claim 3, characterized in that, In step S4, the normalization process for the transient feature is to subtract the minimum absolute value of the feature from the absolute value of the feature among all samples, and then divide by the difference between the maximum and minimum absolute values ​​of the feature among all samples. The normalization process for quasi-steady-state features involves subtracting the minimum value of the feature from the value of the feature across all samples, and then dividing by the difference between the maximum and minimum values ​​of the feature across all samples.

5. The method for extracting insulation polarization current information based on a wide time-domain window according to claim 1, characterized in that, The S-type logic function mentioned in step S5 is: in, The weights for normalized transient features, The weights for normalized quasi-steady-state characteristic quantities, These are the normalized transient characteristics; This refers to the normalized quasi-steady-state characteristic quantity; b This is a bias term.

6. The method for extracting insulation polarization current information based on a wide time-domain window according to claim 1, characterized in that, The start time of the wide time domain window is 5 seconds after the system switching action, and the end time is 600 seconds; the average current change rate within a fixed time step of 5-100s during the initial polarization is calculated, and the average current change rate is used as a transient characteristic quantity. The fitted current value at 600s is selected as the quasi-steady-state current value. The corresponding steady-state current density is calculated by combining the measured electric field, and then the threshold electric field is determined. The threshold electric field is used as the quasi-steady-state characteristic quantity.

7. The method for extracting insulation polarization current information based on a wide time-domain window according to claim 5, characterized in that, When the solid insulation material to be evaluated is cross-linked polyethylene, in the S-type logic function and All are 0.5, with a bias term of -1, to adapt to the polarization characteristics and aging behavior of cross-linked polyethylene materials.

8. A system for extracting insulation polarization current information based on a wide time-domain window, characterized in that: The insulation polarization current information extraction method based on a wide time-domain window, as described in any one of claims 1-6, comprises: The data acquisition module is used to acquire polarization current data of insulating materials measured within a wide time-domain window; The data processing module, which is communicatively connected to the data acquisition module, is used to fit the polarization current data using a multi-exponential relaxation function to suppress data fluctuations and obtain a smooth current decay trend curve. The feature extraction module is communicatively connected to the data processing module and is used to extract a transient feature quantity to characterize the rate of transient current decay in the early stage of polarization, and a quasi-steady-state feature quantity to characterize the insulation strength of the material, based on the fitted current decay trend curve. The normalization module is communicatively connected to the feature extraction module and is used to normalize the transient feature quantities and quasi-steady-state feature quantities to obtain dimensionless first normalization parameters and second normalization parameters, respectively. The state assessment module is communicatively connected to the normalization module. It is used to linearly combine the first normalization parameter and the second normalization parameter, and input the combination result into an S-type logic function to map a comprehensive state index with a value range between 0 and 1. The magnitude of the comprehensive state index value is positively correlated with the degree of deterioration of the insulation material. The output module, which is communicatively connected to the state assessment module, is used to output the comprehensive state index value as characteristic information extracted from the polarization current for assessing the state of the insulating material.

9. A computer device, characterized in that: The device includes one or more processors and one or more memories, wherein at least one piece of program code is stored in the one or more memories, and when the program code is executed by the one or more processors, it implements the insulation polarization current information extraction method based on a wide time-domain window as described in any one of claims 1-7.

10. A storage medium storing at least one line of program code, characterized in that, When the program code is executed by the processor, it implements the steps of the insulation polarization current information extraction method based on a wide time-domain window as described in any one of claims 1-7.

Images