Annular piezoresistor aging performance test method based on data analysis

Abstract

The invention relates to the technical field of electronic component testing, in particular to an annular piezoresistor aging performance testing method based on data analysis. The method comprises the following steps: in a constant-temperature environment, applying an excitation signal consisting of a high-voltage pulse and a high-frequency sinusoidal signal to the annular piezoresistor; synchronously acquiring voltage and current signals, and extracting time-domain thermal characteristics and frequency-domain structural characteristics; calculating the ratio of the time-domain thermal characteristic change rate to the frequency-domain structural characteristic change rate to generate a differentiation index, and executing a hierarchical modulation strategy on the waveform parameter of the high-voltage pulse of the next period based on a numerical interval where the differentiation index is located; and calculating time derivatives of the two types of features in real time, and when absolute values of the two types of features are lower than a preset convergence threshold, terminating the aging test and outputting qualification judgment. According to the invention, by decoupling the thermal effect and the structural damage, the self-adaptive aging screening of the annular piezoresistor is realized, and the accuracy and the safety of the test are improved.

Description

Technical Field

[0001] This application relates to the field of electronic component testing technology. More specifically, this application relates to a data analysis-based method for testing the aging performance of a toroidal varistor. Background Technology

[0002] As a crucial overvoltage protection component in power systems and electronic equipment, the aging performance of toroidal varistors directly affects the safe and reliable operation of these devices. In practical applications, varistors are subjected to the combined effects of voltage and thermal stress over extended periods, gradually leading to performance degradation, manifested as increased leakage current and varistor voltage drift. Therefore, conducting aging performance tests on varistors before shipment to screen for potentially defective products is a critical step in ensuring product quality.

[0003] Existing varistor aging testing techniques typically employ a single excitation method of constant voltage stress or constant temperature stress, assessing the degree of aging by monitoring changes in leakage current or varistor voltage. This method has significant limitations: firstly, a single excitation method struggles to effectively distinguish between reversible performance drift caused by thermal effects and irreversible aging caused by microstructural damage, easily leading to misjudgments; secondly, fixed test parameters cannot be adaptively adjusted according to the actual state of the sample under test, resulting in excessively long test times or insufficient test intensity.

[0004] Furthermore, traditional methods primarily rely on monitoring time-domain parameters, neglecting the characterization of varistor nonlinear characteristics in the frequency domain. When the internal grain boundary structure of a varistor degrades, its nonlinear current-voltage characteristics change, manifesting as variations in harmonic components in the frequency domain. There is an urgent need for a method that can integrate time-domain and frequency-domain characteristics, decouple thermal effects from structural aging, and adaptively adjust test parameters according to actual conditions to address the problems of low testing efficiency and poor discrimination accuracy in existing technologies. Summary of the Invention

[0005] The purpose of this application is to propose a data analysis-based method for testing the aging performance of ring varistors, in order to solve the problems of low efficiency caused by distinguishing between thermal effects and structural aging and fixed test parameters in the prior art.

[0006] This application provides a data analysis-based aging performance testing method for a ring varistor, comprising: applying an excitation signal to the ring varistor in a constant temperature environment, the excitation signal consisting of periodic high-voltage pulses and a high-frequency sinusoidal signal superimposed between the high-voltage pulses; synchronously acquiring the voltage and current signals of the ring varistor, and extracting time-domain thermal features and frequency-domain structural features; wherein, the time-domain thermal feature is the charge relaxation constant, and the frequency-domain structural feature is the third harmonic distortion calculated based on the response data of the high-frequency sinusoidal signal; calculating the ratio of the rate of change of the time-domain thermal feature to the rate of change of the frequency-domain structural feature, generating a differentiation index, and performing a graded modulation strategy on the waveform parameters of the next period of high-voltage pulse based on the numerical range of the differentiation index; during the execution of the above steps, calculating the time derivatives of the time-domain thermal feature and the frequency-domain structural feature in real time, and terminating the aging test and outputting a pass / fail judgment when the absolute values ​​of both are lower than a preset convergence threshold.

[0007] This invention simultaneously acquires time-domain thermal characteristics and frequency-domain structural characteristics through a composite excitation signal, and utilizes a differentiation index to decouple and distinguish between thermal effects and structural aging. The graded modulation strategy based on the differentiation index can adaptively adjust test parameters according to the actual state of the sample under test. When thermal drift is dominant, it reduces stress intensity to avoid misjudgment; when structural aging is dominant, it increases stress intensity to accelerate defect exposure, thereby significantly improving test efficiency while ensuring test accuracy.

[0008] Optionally, the step of generating the excitation signal includes: preset a basic clock cycle, outputting a rectangular high-voltage pulse during the duty cycle conduction period of the basic clock cycle, wherein the amplitude of the high-voltage pulse is set to the nominal varistor voltage of the ring varistor; during the duty cycle off period of the basic clock cycle, firstly, the control circuit enters a high-impedance state to acquire the voltage attenuation waveform, and after the acquisition is completed, outputting a zero-volt bias voltage and superimposing the high-frequency sine signal onto the zero-bias voltage.

[0009] Optionally, the calculation process of the time-domain thermal characteristics includes: extracting the voltage decay waveform data after the falling edge of the high-voltage pulse; constructing a mathematical model based on the exponential decay law, and mapping the voltage decay waveform data to the mathematical model using the least squares method; extracting the parameter term in the mathematical model that characterizes the time required for the voltage to decay to a specific proportion of the initial value, and determining it as the charge relaxation constant.

[0010] Optionally, the calculation process of the frequency domain structural features includes: extracting the current response data sequence under the excitation of the high-frequency sinusoidal signal; performing a fast Fourier transform on the current response data sequence to separate the amplitude of the fundamental component and the amplitude of the third harmonic component; calculating the ratio of the amplitude of the third harmonic component to the amplitude of the fundamental component, and determining the ratio as the third harmonic distortion.

[0011] By extracting the third harmonic distortion through frequency domain analysis, the changes in the nonlinear characteristics of the varistor's grain boundary structure can be sensitively reflected. When grain boundary degradation occurs, the nonlinear coefficient changes, leading to changes in the proportion of harmonic components, thereby enabling early warning of structural aging.

[0012] Optionally, the calculation process of the differentiation index includes: obtaining the absolute value of the relative change of the charge relaxation constant of the current period with respect to the initial period, as the numerator; obtaining the absolute value of the change of the third harmonic distortion of the current period with respect to the initial period, and superimposing a non-zero regularization factor as the denominator; calculating the ratio of the numerator to the denominator to obtain the differentiation index; such that the differentiation index is positively correlated with the change amplitude of the charge relaxation constant and negatively correlated with the fluctuation amplitude of the third harmonic distortion.

[0013] Optionally, the graded modulation strategy includes: if the differentiation index is greater than the first preset threshold, determining that the current state is a thermal drift-dominated state; increasing the high-voltage pulse interval time of the next cycle while keeping the amplitude of the high-voltage pulse unchanged until the differentiation index falls back below the first preset threshold. If the differentiation index is less than the second preset threshold, determining that the current state is a structural aging-dominated state; controlling the signal generator to execute stress screening logic, increasing the amplitude of the high-voltage pulse or increasing the pulse width of the high-voltage pulse to accelerate the evolution of potential defect regions; the second preset threshold is less than the first preset threshold.

[0014] By identifying the dominant state of structural aging and automatically increasing stress intensity, it is possible to accelerate the exposure and evolution of potential defects, shorten the testing cycle, and improve screening efficiency.

[0015] Optionally, the graded modulation strategy further includes: if the differentiation index is not less than the second preset threshold and not greater than the first preset threshold, determining that the current state is a stable evolution state; keeping the high voltage pulse waveform parameters of the current cycle unchanged, and continuing to perform the aging test of the next cycle.

[0016] Optionally, the method further includes, during the execution of the graded modulation strategy: real-time monitoring of the leakage current value of the ring varistor; if the leakage current value exceeds a preset safety cutoff threshold, or the change in the third harmonic distortion within a single test cycle exceeds a preset abrupt change threshold; generating a defect rejection instruction, cutting off the excitation signal, terminating the test, and marking the currently tested ring varistor as a defective product.

[0017] By monitoring the sudden changes in leakage current and harmonic distortion in real time, seriously defective samples can be detected in a timely manner and the test can be terminated immediately, which not only protects the safety of the test equipment, but also improves the rejection efficiency of defective products.

[0018] Optionally, the determination of terminating the aging test includes: constructing a sliding time window containing the most recent N test cycles; calculating the slope of the change curve of the time-domain thermal characteristics and the slope of the change curve of the frequency-domain structural characteristics within the sliding time window; determining that the aging state has reached a steady state and triggering the pass determination only when the absolute value of the slope of the change curve of the time-domain thermal characteristics is less than a first convergence factor and the absolute value of the slope of the change curve of the frequency-domain structural characteristics is less than a second convergence factor.

[0019] Optionally, the first preset threshold and the second preset threshold include: establishing a pure thermal excitation experimental group, applying only temperature field changes to the sample without applying high-voltage pulses, statistically analyzing the probability density distribution of the differential index calculated under this state, and setting the lower limit of the confidence interval of the probability density distribution as the first preset threshold; establishing a defect destruction experimental group, applying destructive high-voltage pulses to the sample until breakdown, retrospectively analyzing the average value of the differential index within a preset time period before breakdown, and setting the average value as the second preset threshold.

[0020] The threshold parameters are determined by experimental calibration, which gives the threshold setting clear physical meaning and statistical basis, thus improving the scientificity and reliability of the hierarchical modulation strategy.

[0021] The beneficial effects of this application are as follows: This invention simultaneously acquires time-domain thermal characteristics and frequency-domain structural characteristics through a composite excitation signal, and utilizes a differentiation index to achieve decoupled discrimination between thermal effects and structural aging. This mechanism abandons the traditional single-parameter monitoring approach and uses dual-domain feature fusion analysis to achieve accurate identification of the aging mechanism of varistors. The graded modulation strategy based on the differentiation index can adaptively adjust the test parameters according to the actual state of the sample under test. When thermal drift is dominant, the stress intensity is reduced to avoid misjudgment, and when structural aging is dominant, the stress intensity is increased to accelerate defect exposure. Thus, while ensuring test accuracy, test efficiency is significantly improved, and test time and energy consumption are reduced. Attached Figure Description

[0022] Figure 1This is a flowchart of a data analysis-based aging performance testing method for a toroidal varistor according to an embodiment of this application.

[0023] Figure 2 This is a characteristic change diagram of a data analysis-based aging performance testing method for a toroidal varistor according to an embodiment of this application.

[0024] Figure 3 This is a comparison chart showing the effects of a data analysis-based aging performance testing method for a ring varistor according to an embodiment of this application. Detailed Implementation

[0025] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Figure 1 The diagram shown is a flowchart of a data analysis-based aging performance testing method for a toroidal varistor according to an embodiment of this application.

[0026] S1: Apply an excitation signal and acquire voltage and current response data.

[0027] In a constant temperature environment, an excitation signal is applied to the ring-shaped varistor. The excitation signal consists of periodic high-voltage pulses and high-frequency sinusoidal signals superimposed between the high-voltage pulses.

[0028] The testing system includes a constant temperature chamber, a signal generator, a high-voltage amplifier, a data acquisition unit, and a controller. The constant temperature chamber provides a stable temperature environment for the tested toroidal varistor, with a preferred temperature control accuracy of ±0.5°C. The signal generator produces the basic excitation waveform. The high-voltage amplifier amplifies the waveform output from the signal generator to the required current level for the test. The data acquisition unit synchronously acquires the voltage and current signals of the tested sample, with a preferred sampling rate of 1MHz or higher. The controller is a microprocessor control unit responsible for executing the control algorithm, data processing, and test procedure management.

[0029] The steps for generating the excitation signal include: presetting a basic clock cycle; outputting a rectangular high-voltage pulse during the duty cycle conduction period of the basic clock cycle, wherein the amplitude of the high-voltage pulse is set to the nominal varistor voltage of the ring varistor; and outputting a zero-volt voltage during the duty cycle off period of the basic clock cycle, and superimposing the high-frequency sine wave signal onto the zero bias voltage.

[0030] Specifically, the base clock cycle can be set in two ways: a fixed cycle mode and an adaptive cycle mode. In the fixed cycle mode, the base clock cycle is set to a fixed value, such as 1 second or 2 seconds. This mode is suitable for batch testing scenarios. In the adaptive cycle mode, the system dynamically adjusts the base clock cycle according to the differentiation index to achieve intelligent testing.

[0031] The amplitude of the high-voltage pulse is set to the nominal varistor voltage of the tested ring varistor, i.e., the voltage value measured under a 1mA reference current. The pulse width of the high-voltage pulse is preferably 10ms to 100ms, and the duty cycle is preferably 10% to 30%. The pulse width needs to be long enough to allow the varistor to conduct fully and generate a thermal effect, but it cannot be too long to avoid excessive damage.

[0032] The frequency of the high-frequency sinusoidal signal is preferably between 1 kHz and 10 kHz, and the amplitude is preferably between 5% and 20% of the nominal varistor voltage. The high-frequency sinusoidal signal is superimposed on the zero-bias voltage of the high-voltage pulse gap to detect the small-signal nonlinear characteristics of the varistor. Frequency selection needs to consider the frequency response characteristics of the varistor; excessively high frequencies may cause parasitic capacitance effects to interfere with the measurement results.

[0033] S2: Extract time-domain thermal features and frequency-domain structural features.

[0034] The voltage and current signals of the ring varistor are acquired synchronously, and the time-domain thermal characteristics and frequency-domain structural characteristics are extracted; wherein, the time-domain thermal characteristics are the charge relaxation constants, and the frequency-domain structural characteristics are the third harmonic distortion calculated based on the response data of the high-frequency sinusoidal signal.

[0035] The calculation process of the time-domain thermal characteristics includes: extracting the voltage decay waveform data after the falling edge of the high-voltage pulse; constructing a mathematical model based on the exponential decay law, and mapping the voltage decay waveform data to the mathematical model using the least squares method; extracting the parameter term in the mathematical model that characterizes the time required for the voltage to decay to a specific proportion of the initial value, and determining it as the charge relaxation constant.

[0036] Specifically, the voltage decay waveform data is extracted starting from the falling edge of the high-voltage pulse, and the duration is preferably 2 to 5 times the pulse width. The mathematical model based on the exponential decay law is expressed as: ; in, for The voltage value at time , in volts. The initial voltage amplitude, in volts. Let be the charge relaxation constant, with dimensions in seconds. This is the steady-state voltage value, measured in volts, which is close to zero under ideal conditions.

[0037] The collected voltage decay data were fitted using the least squares method, and the parameter $\tau$ was extracted as the charge relaxation constant. The charge relaxation constant reflects the time characteristics of charge redistribution within the varistor and is closely related to the temperature state of the material. When the temperature of the varistor increases due to thermal effects, the carrier mobility increases, and the charge relaxation constant decreases; when the temperature recovers, the charge relaxation constant rises again.

[0038] The calculation process of the frequency domain structural features includes: extracting the current response data sequence under the excitation of the high-frequency sinusoidal signal; performing a fast Fourier transform on the voltage response data sequence to separate the amplitude of the fundamental component and the amplitude of the third harmonic component; calculating the ratio of the amplitude of the third harmonic component to the amplitude of the fundamental component, and determining the ratio as the third harmonic distortion.

[0039] Specifically, the extraction window for the voltage response data sequence is the off-time of the high-voltage pulse gap, ensuring that the response under high-frequency sinusoidal signal excitation is not affected by the residual effect of the high-voltage pulse. The Fast Fourier Transform (FFT) uses the standard FFT algorithm, and the number of transformation points is preferably an integer power of 2, such as 1024 or 2048 points.

[0040] The formula for calculating third harmonic distortion is: ; in, The third harmonic distortion is... The amplitude of the fundamental component. This represents the amplitude of the third harmonic component.

[0041] The third harmonic distortion reflects the nonlinear characteristics of a varistor. The volt-ampere characteristic of a varistor is highly nonlinear; when the grain boundary structure degrades, the nonlinear coefficient changes, leading to a change in the proportion of harmonic components. The reason for choosing the third harmonic over the second harmonic is that the volt-ampere characteristic of a varistor is symmetrical; even-order harmonic components are smaller, while odd-order harmonic components better reflect changes in nonlinear characteristics.

[0042] like Figure 2 The diagram shows the characteristic changes of a data analysis-based aging performance testing method for a toroidal varistor according to an embodiment of this application. The red area represents the light red shaded region where the curves are superimposed, indicating the confidence interval for thermal noise fluctuations. The initially visible red area is actually within the normal fluctuation range of thermal characteristics under ambient temperature disturbances. The blue curve shows an exponential increase in structural damage as aging progresses.

[0043] S3: Calculate the differentiation index and execute the hierarchical modulation strategy.

[0044] The ratio of the rate of change of the time-domain thermal characteristics to the rate of change of the frequency-domain structural characteristics is calculated to generate a differentiation index. Based on the numerical range of the differentiation index, a graded modulation strategy is executed on the waveform parameters of the high-voltage pulse in the next cycle.

[0045] The calculation process of the differentiation index includes: obtaining the relative change of the charge relaxation constant of the current period with respect to the initial period as the numerator; obtaining the absolute value of the change of the third harmonic distortion of the current period with respect to the initial period, and superimposing a non-zero regularization factor as the denominator; and calculating the ratio of the numerator to the denominator to obtain the differentiation index.

[0046] The formula for calculating the differentiation index is: ; Wherein, DI is the differentiation index, which is dimensionless; This represents the change in the charge relaxation constant of the current period relative to the initial period. Let be the charge relaxation constant for the initial period. This represents the change in the third harmonic distortion of the current period relative to the initial period. It is a non-zero regularization factor, preferably with a value between 0.01 and 0.1.

[0047] The physical meaning of the differentiation index is as follows: when thermal effects dominate, the charge relaxation constant changes significantly while the third harmonic distortion changes relatively little, resulting in a larger differentiation index; when structural aging dominates, the third harmonic distortion changes significantly while the charge relaxation constant changes relatively little, resulting in a smaller differentiation index. The differentiation index can effectively distinguish between two different aging mechanisms.

[0048] The hierarchical modulation strategy includes handling the thermal drift-dominated state and the structural aging-dominated state.

[0049] For the thermal drift-dominant state: the differentiation index is compared with a first preset threshold; if the differentiation index is greater than the first preset threshold, the current state is determined to be the thermal drift-dominant state; the control signal generator executes cooling logic, increases the high-voltage pulse interval time of the next cycle, and keeps the amplitude of the high-voltage pulse unchanged until the differentiation index falls back below the first preset threshold.

[0050] Specifically, the cooling logic can be implemented using linear incremental and exponential incremental methods. The linear incremental method increases the pulse interval by a fixed increment, such as 100ms each time. The exponential incremental method multiplies the pulse interval by a fixed coefficient, such as 1.5 each time. The upper limit of the pulse interval is preferably 5 times the initial interval to avoid excessively prolonged testing time.

[0051] For the structural aging-dominated state: the differentiation index is compared with a second preset threshold, and the second preset threshold is less than the first preset threshold; if the differentiation index is less than the second preset threshold, the current state is determined to be the structural aging-dominated state; the control signal generator executes stress screening logic to increase the amplitude of the high-voltage pulse or increase the pulse width of the high-voltage pulse to accelerate the evolution of potential defect areas.

[0052] Specifically, the stress screening logic is implemented using two methods: amplitude enhancement and pulse width enhancement. Amplitude enhancement increases the high-voltage pulse amplitude by a fixed percentage, for example, 5% each time, with an upper limit of 120% of the nominal varistor voltage. Pulse width enhancement increases the high-voltage pulse width by a fixed percentage, for example, 20% each time, with an upper limit of three times the initial pulse width.

[0053] If the differentiation index is between the second preset threshold and the first preset threshold, the current state is determined to be a stable evolution state. The high-voltage pulse waveform parameters of the current cycle remain unchanged, and the next cycle test is continued.

[0054] The calibration methods for the first preset threshold and the second preset threshold include: establishing a pure thermal excitation experimental group, applying only temperature field changes to the sample without applying a high-voltage pulse, statistically analyzing the probability density distribution of the differential index calculated under this state, and setting the lower limit of the confidence interval of the probability density distribution as the first preset threshold; establishing a defect destruction experimental group, applying a destructive high-voltage pulse to the sample until breakdown, retrospectively analyzing the average value of the differential index within a preset time period before breakdown, and setting the average value as the second preset threshold.

[0055] Specifically, the implementation steps of the pure thermal excitation experimental group are as follows: select several qualified samples and place them in a constant temperature chamber; generate thermal excitation by changing the temperature of the constant temperature chamber (e.g., cyclically changing it within the range of 25°C to 85°C) without applying a high voltage pulse; collect the charge relaxation constant and third harmonic distortion at each temperature point and calculate the differentiation index; statistically analyze the probability density distribution of the differentiation index of all data points and take the lower limit of the 95% confidence interval as the first preset threshold.

[0056] The implementation steps of the defect failure test group are as follows: select several samples, apply gradually increasing high voltage pulses until the samples break down and fail; retrospectively analyze the differential index data within the 10 test cycles before breakdown; calculate the average value of these data as the second preset threshold.

[0057] The implementation of the graded modulation strategy also includes a safety protection mechanism: real-time monitoring of the leakage current value of the ring varistor; if the leakage current value exceeds the safety cutoff threshold, or the change in the third harmonic distortion within a single test cycle exceeds the sudden change threshold; a defect rejection instruction is generated, the excitation signal is forcibly cut off, the test is terminated, and the currently tested ring varistor is marked as a defective product.

[0058] The safe cut-off threshold is preferably 10 times the nominal leakage current. The mutation threshold is preferably 50% of the initial third harmonic distortion. When the safety protection mechanism is triggered, the system immediately cuts off the high-voltage output, records the fault information, and marks the current sample as a defective product.

[0059] S4: Determine the convergence condition and output the test results.

[0060] During the execution of the above steps, the time derivatives of the time-domain thermal characteristics and the frequency-domain structural characteristics are calculated in real time. When the absolute values ​​of both are lower than the preset convergence threshold, the aging test is terminated and a pass / fail judgment is output.

[0061] The determination of terminating the aging test includes: constructing a sliding time window containing the most recent N test cycles; calculating the slope of the change curve of the time-domain thermal characteristics and the slope of the change curve of the frequency-domain structural characteristics within the sliding time window; and determining that the aging state has reached a steady state and triggering the qualification determination only when the slope of the change curve of the time-domain thermal characteristics is less than a first convergence factor and the slope of the change curve of the frequency-domain structural characteristics is less than a second convergence factor.

[0062] Specifically, the size N of the sliding time window is preferably 10 to 20 test cycles. The window size needs to be large enough to eliminate the influence of random fluctuations, but not too large to avoid response lag.

[0063] The slope of the curve was calculated using a least-squares linear fitting method. For the time-domain thermal characteristics, the charge relaxation constant values ​​for each period within the sliding window were used to construct a data sequence, and the slope was obtained by fitting the data. For frequency domain structural characteristics, the third harmonic distortion values ​​of each period within the sliding window are used to construct a data sequence, and the slope is obtained by fitting. .

[0064] First convergence factor The setting of the second convergence factor is related to the measurement accuracy of the charge relaxation constant, and is preferably 0.1% of the initial charge relaxation constant per cycle. The setting is related to the measurement accuracy of the third harmonic distortion, and is preferably 0.05% of the initial third harmonic distortion per cycle.

[0065] The logic for convergence determination is: when When all conditions are met simultaneously, the aging process is considered to have reached a steady state. Steady state means that the performance parameters of the tested sample no longer change significantly under the current stress conditions, the aging process tends to be in equilibrium, and the sample can be considered to have passed the aging screening test.

[0066] When the convergence condition is met, the system performs the following operations: stops applying the excitation signal; records the final charge relaxation constant, third harmonic distortion, total number of test cycles, and total test time; outputs the pass / fail result; and stores the test data in the database for subsequent analysis.

[0067] If the convergence condition is not met within the preset maximum number of test cycles, the system determines that the test has timed out. Test timeout may indicate a continuous degradation trend in the sample, requiring manual verification or an extension of the test period. The maximum number of test cycles is preferably 500 to 1000 cycles.

[0068] like Figure 3 The figure shown is a comparison of the effects of a data analysis-based aging performance testing method for a ring varistor according to an embodiment of this application. Compared with traditional methods, the method of this invention can effectively distinguish between reversible performance drift caused by thermal effects and irreversible degradation caused by structural aging, avoiding misjudgment; by adaptively adjusting test parameters through a graded modulation strategy, the test time is shortened while ensuring test accuracy; and the dual-domain feature fusion analysis improves the detection rate of early defects, thus enhancing the defect detection rate.

[0069] The above description is merely a preferred embodiment of this application, but the scope of protection of this application is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in this application, based on the technical solution and inventive concept of this application, should be within the scope of protection of this application.

Claims

1. A data analysis-based method for testing the aging performance of a ring varistor, characterized in that, The testing method includes: In a constant temperature environment, an excitation signal is applied to the ring-shaped varistor. The excitation signal consists of periodic high-voltage pulses and high-frequency sinusoidal signals superimposed between the high-voltage pulses. The voltage and current signals of the ring varistor are acquired synchronously, and the time-domain thermal characteristics and frequency-domain structural characteristics are extracted; wherein, the time-domain thermal characteristics are the charge relaxation constant, and the frequency-domain structural characteristics are the third harmonic distortion calculated based on the response data of the high-frequency sinusoidal signal; The ratio of the rate of change of the time-domain thermal characteristics to the rate of change of the frequency-domain structural characteristics is calculated to generate a differentiation index. Based on the numerical range of the differentiation index, a graded modulation strategy is executed on the waveform parameters of the high-voltage pulse in the next cycle. During the execution of the above steps, the time derivatives of the time-domain thermal characteristics and the frequency-domain structural characteristics are calculated in real time. When the absolute values ​​of both are lower than the preset convergence threshold, the aging test is terminated and a pass / fail judgment is output.

2. The method for testing the aging performance of a ring varistor based on data analysis according to claim 1, characterized in that, The steps for generating the excitation signal include: A basic clock cycle is preset, and a rectangular wave high voltage pulse is output during the duty cycle conduction period of the basic clock cycle. The amplitude of the high voltage pulse is set to the nominal varistor voltage of the ring varistor. During the duty cycle off period of the basic clock cycle, the control circuit first enters a high impedance state to acquire the voltage decay waveform. After the acquisition is completed, a zero-volt bias voltage is output, and the high-frequency sine wave signal is superimposed on the zero-bias voltage.

3. The method for testing the aging performance of a ring varistor based on data analysis according to claim 1, characterized in that, The calculation process of the time-domain thermal characteristics includes: Extract the voltage attenuation waveform data after the falling edge of the high-voltage pulse; A mathematical model based on the exponential decay law is constructed, and the voltage decay waveform data is mapped to the mathematical model using the least squares method; The time constant in the mathematical model is extracted and determined as the charge relaxation constant.

4. The method for testing the aging performance of a ring varistor based on data analysis according to claim 1, characterized in that, The calculation process of the frequency domain structure features includes: Extract the current response data sequence under the high-frequency sinusoidal signal excitation; Perform a Fast Fourier Transform on the current response data sequence to separate the amplitude of the fundamental component and the amplitude of the third harmonic component; Calculate the ratio of the amplitude of the third harmonic component to the amplitude of the fundamental component, and determine the ratio as the third harmonic distortion.

5. The method for testing the aging performance of a ring varistor based on data analysis according to claim 1, characterized in that, The calculation process of the differentiation index includes: Obtain the absolute value of the relative change of the charge relaxation constant in the current period with respect to the initial period, and use it as the numerator. Obtain the absolute value of the change in the third harmonic distortion of the current period relative to the initial period, and add a non-zero regularization factor as the denominator. The difference index is obtained by calculating the ratio of the numerator to the denominator; such that the difference index is positively correlated with the change in the charge relaxation constant and negatively correlated with the fluctuation of the third harmonic distortion.

6. The method for testing the aging performance of a ring varistor based on data analysis according to claim 1, characterized in that, The hierarchical modulation strategy includes: If the differentiation index is greater than the first preset threshold, the current state is determined to be a thermal drift-dominated state; the high-voltage pulse interval time of the next cycle is increased, while the amplitude of the high-voltage pulse remains unchanged, until the differentiation index falls back below the first preset threshold. If the differentiation index is less than the second preset threshold, the current state is determined to be a structural aging-dominated state; the control signal generator executes stress screening logic to increase the amplitude of the high-voltage pulse or increase the pulse width of the high-voltage pulse to accelerate the evolution of potential defect areas; the second preset threshold is less than the first preset threshold.

7. The method for testing the aging performance of a ring varistor based on data analysis according to claim 6, characterized in that, The hierarchical modulation strategy also includes: If the differentiation index is not less than the second preset threshold and not greater than the first preset threshold, the current state is determined to be a stable evolution state; the high voltage pulse waveform parameters of the current cycle remain unchanged, and the aging test of the next cycle continues.

8. The method for testing the aging performance of a ring varistor based on data analysis according to claim 7, characterized in that, In the process of implementing the hierarchical modulation strategy, the method further includes: The leakage current value of the ring varistor is monitored in real time; if the leakage current value exceeds the preset safety cutoff threshold, or the change in the third harmonic distortion in a single test cycle exceeds the preset sudden change threshold, a defect rejection instruction is generated, the excitation signal is cut off, the test is terminated, and the ring varistor currently being tested is marked as a defective product.

9. The method for testing the aging performance of a ring varistor based on data analysis according to claim 1, characterized in that, The determination of terminating the aging test includes: Construct a sliding time window containing the most recent N test cycles; Calculate the slope of the time-domain thermal characteristic change curve and the slope of the frequency-domain structural characteristic change curve within the sliding time window, respectively. The aging state is determined to have reached a steady state only when the absolute value of the slope of the time-domain thermal characteristic change curve is less than the first convergence factor and the absolute value of the slope of the frequency-domain structural characteristic change curve is less than the second convergence factor, thus triggering the qualification determination.

10. The method for testing the aging performance of a ring varistor based on data analysis according to claim 7, characterized in that, The first preset threshold and the second preset threshold include: A pure thermal excitation experimental group was established, in which only temperature field changes were applied to the sample without high-pressure pulses. The probability density distribution of the differential index calculated under this state was statistically analyzed, and the lower limit of the confidence interval of the probability density distribution was set as the first preset threshold. A defect destruction experimental group was established, and a destructive high-voltage pulse was applied to the sample until it broke down. The average value of the differential index within a preset time period before the breakdown was analyzed, and the average value was set as the second preset threshold.

Images