An ac-dc capacitor function test system

CN122944757APending Publication Date: 2026-10-20GUANGZHOU SAIRUI TESTING EQUIP CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611045675.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-07-14
Publication Date
2026-10-20

AI Technical Summary

Technical Problem

[0003]然而,基于现有技术进行交直流电容器测试时,在功率开关高速动作和强电磁干扰条件下,采样数据容易出现噪声伪峰和失真,影响电容量、等效串联电阻及损耗角正切等参数的判定准确性;在长时间纹波加载过程中,现有方案难以有效区分正常热漂移与真实性能退化,且故障保护响应速度不足,容易降低测试结果的可靠性并增加测试台架受冲击损坏的风险

Benefits of technology

[0034]1.本发明通过测试激励接入模块输出直流偏压及叠加其上的交流纹波激励,并由实时采集执行模块同步采集端电压、纹波电流、表面温度和共模噪声检测值,结合离散傅里叶变换提取基波相量并估算电容量、等效串联电阻和损耗角正切,实现了对交直流电容器在复合应力条件下的多参量联动测试与实时状态判定,提高了测试过程中的参数获取完整性和功能评估准确性;

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure FT_1
    Figure FT_1
  • Figure SMS_1
    Figure SMS_1
  • Figure SMS_15
    Figure SMS_15
Patent Text Reader

Abstract

This invention relates to the field of power electronics testing and capacitor detection technology, specifically an AC / DC capacitor functional testing system, including a test excitation access module, a real-time acquisition and execution module, an analysis and update module, and a hardware protection circuit. The test excitation access module outputs a DC bias voltage and an AC ripple excitation superimposed on it. The real-time acquisition and execution module synchronously acquires terminal voltage, ripple current, surface temperature, and common-mode noise detection values, and estimates capacitance, equivalent series resistance, and loss tangent. The sampling shielding unit sets a sampling suppression interval based on the power switching time, suspends effective sampling when noise exceeds the limit, and calls 2 to 5 effective samples before and after the suppression interval for weighted interpolation recovery. The analysis and update module generates and writes back a threshold lookup table, state machine configuration parameters, and control compensation parameters. When a fault is triggered, the hardware protection circuit executes protection according to the sequence of first blocking and then discharging. This invention can avoid misjudging switching noise as partial discharge of the capacitor.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of power electronics testing and capacitor testing technology, specifically to an AC / DC capacitor functional testing system. Background Technology

[0002] AC / DC capacitor functional testing refers to the detection and evaluation of the voltage, current, temperature, and electrical parameters of a capacitor under combined stress conditions of DC bias and AC ripple. Current AC / DC capacitor testing methods typically include constant operating condition loading, continuous sampling monitoring, and software-based protection methods with fixed thresholds.

[0003] However, when testing AC / DC capacitors using existing technologies, sampling data is prone to noise spurious peaks and distortion under conditions of high-speed operation of power switches and strong electromagnetic interference, affecting the accuracy of determining parameters such as capacitance, equivalent series resistance, and loss tangent. During long-term ripple loading, existing solutions struggle to effectively distinguish between normal thermal drift and actual performance degradation, and the fault protection response speed is insufficient, which can easily reduce the reliability of test results and increase the risk of impact damage to the test bench. Summary of the Invention

[0004] The purpose of this invention is to provide an AC / DC capacitor functional testing system that avoids data distortion under strong electromagnetic interference conditions and the difficulty in distinguishing between normal thermal drift and actual performance degradation during long-term testing. It also makes it easier to achieve accurate determination of electrical parameters and nanosecond-level fault hardware protection.

[0005] The objective of this invention can be achieved through the following technical solutions:

[0006] A functional testing system for AC / DC capacitors includes a test excitation access module, a real-time acquisition and execution module, an analysis and update module, and a hardware protection circuit. The test excitation access module is used to output a DC bias voltage and an AC ripple excitation superimposed on the DC bias voltage, and is connected to the AC / DC capacitor under test via a test bus. The test excitation access module includes a ripple injection power switch unit.

[0007] The real-time acquisition and execution module is used to acquire the terminal voltage, ripple current, surface temperature and common-mode noise detection values ​​of the AC / DC capacitor under test, perform noise reduction processing, parameter estimation and state determination on the acquired signals, and adjust at least one of the DC bias output, AC ripple amplitude, frequency and duty cycle of the test excitation input module according to the determination results.

[0008] The analysis and update module is used to receive data output from the real-time acquisition and execution module, perform thermal drift analysis and operating condition matching analysis, generate and write back threshold comparison table, state machine configuration parameters and control compensation parameters, and record test process data, degradation records and protection trigger logs.

[0009] The hardware protection circuit is connected to the real-time acquisition and execution module to receive drive lockout signals and discharge trigger signals when a fault is triggered, and to perform protection in the order of lockout followed by discharge.

[0010] Preferably, the real-time acquisition and execution module includes a sensor group acquisition unit, a high-speed sampling unit, a sampling shielding unit, a parameter estimation unit, a state machine protection unit, and a register loading unit;

[0011] The sensor group acquisition unit is connected to a high-voltage differential voltage sensor, a broadband current sensor, a temperature sensor, and a common-mode noise detection branch, respectively, to synchronously acquire capacitor terminal voltage, ripple current, surface temperature, and common-mode noise detection values, and to calculate the current change rate.

[0012] The parameter estimation unit is used to estimate capacitance, equivalent series resistance, and loss tangent.

[0013] The register loading unit is used to receive the threshold lookup table, state machine configuration parameters and control compensation parameters issued by the analysis and update module, and write them into the comparison register and the state machine configuration register area.

[0014] Preferably, the sampling shielding unit sets the sampling suppression interval according to the switching time of the ripple injection power switching unit, and determines the noise threshold according to the background noise;

[0015] When the sampling time is within the sampling suppression interval and the common-mode noise detection value is higher than the noise threshold, effective sampling writing is paused.

[0016] During the pause of valid sampling writing, several valid samples are taken before and after the sampling suppression interval for weighted interpolation recovery, and the recovered sampled values ​​are output to the parameter estimation unit.

[0017] Preferably, the analysis and update module extracts operating parameters according to a continuous time window. The operating parameters include at least the average terminal voltage, the root mean square value of the ripple current, the temperature rise rate, and the ripple frequency.

[0018] The analysis and update module matches the current operating parameters with the weighted distance between the reference operating points in the thermoelectric parameter library, selects at least one neighboring operating point, and extracts the capacitance range, equivalent series resistance range and loss tangent range under the current operating condition based on the neighboring operating points, and discretizes the range into a two-dimensional lookup table indexed by temperature and voltage.

[0019] The reference operating point parameters include at least one of the following: capacitance, equivalent series resistance, loss tangent, and temperature rise rate threshold.

[0020] Preferably, the real-time acquisition and execution module calls up the upper and lower limits of capacitance, equivalent series resistance, and loss tangent in the two-dimensional lookup table according to the current surface temperature and capacitor terminal voltage, and compares the real-time estimated capacitance, equivalent series resistance, and loss tangent with the corresponding upper and lower limits respectively;

[0021] When the parameters estimated in real time are between the corresponding upper and lower limits or equal to the corresponding boundary values, output a thermal drift indicator;

[0022] When the parameters estimated in real time exceed the corresponding upper and lower limits, a degradation alarm is output and degradation information is recorded.

[0023] Preferably, the analysis and update module is also used to calculate the equivalent impedance amplitude within the time window based on the processed waveform, and generate control compensation parameters based on the change of the equivalent impedance amplitude relative to the equivalent impedance amplitude under the reference operating condition, and then write the compensated control parameters into the real-time acquisition and execution module.

[0024] The real-time acquisition and execution module adjusts the AC ripple injection process based on the compensated control parameters.

[0025] Preferably, the real-time acquisition and execution module constructs transient criteria based on the terminal voltage change rate, ripple current change rate, and common-mode noise detection value, and determines that a serious fault has occurred when the ripple current change rate is greater than the preset ripple current change rate threshold and the transient criteria are greater than the judgment threshold.

[0026] Among them, the ripple current change rate threshold is generated based on the preset normal test extreme value and failure test extreme value calibration, and the judgment threshold is generated based on the statistical mean and variance calibration of the transient criterion under steady-state conditions.

[0027] Preferably, the hardware protection circuit includes a power drive interlocking circuit, a crowbar discharge circuit, an energy dissipation resistor, and a protection isolation drive unit;

[0028] After the fault determination is established, the state machine protection unit first outputs a drive lockout signal within 50ns, and then outputs a crowbar conduction signal within 200ns after the lockout power drive, so that the surge energy is introduced into the dissipation branch.

[0029] Simultaneously, the preset sampling window data before and after the fault is triggered is frozen and sent to the analysis and update module.

[0030] Preferably, the analysis and update module includes a data recording unit and a database update unit;

[0031] The data logging unit is used to record parameter trajectories, protection trigger logs, and degradation information during the testing process;

[0032] The database update unit is used to correct the baseline operating point parameters in the thermoelectric parameter library based on the thermal drift trajectory, performance degradation records and sampling windows before and after each round of testing, and uses the corrected results as the basis for generating the threshold lookup table and state machine configuration parameters for the next round of testing.

[0033] The beneficial effects of this invention are:

[0034] 1. This invention outputs a DC bias voltage and an AC ripple excitation superimposed on it through a test excitation access module, and the real-time acquisition and execution module synchronously acquires the terminal voltage, ripple current, surface temperature and common-mode noise detection values. Combined with discrete Fourier transform to extract the fundamental phasor and estimate the capacitance, equivalent series resistance and loss tangent, it realizes multi-parameter linkage test and real-time status determination of AC and DC capacitors under composite stress conditions, and improves the completeness of parameter acquisition and the accuracy of functional evaluation during the test process.

[0035] 2. By setting a sampling suppression interval based on the switching time of the ripple injection power switch unit through the sampling shielding unit, and pausing effective sampling when the common-mode noise detection value is higher than the noise threshold, 2 to 5 effective samples are taken before and after the suppression interval for weighted interpolation recovery. At the same time, when samples are missing, the corresponding phase sampling value of the previous effective cycle is read back as a substitute sample. This effectively suppresses noise spurious peaks and sampling distortion caused by high-speed operation of the power switch and strong electromagnetic interference, avoids misjudging switching noise as abnormal capacitor discharge, and significantly improves the reliability of parameter estimation and judgment results.

[0036] 3. By analyzing and updating the module, operating parameters such as the average terminal voltage, root mean square value of ripple current, temperature rise rate, and ripple frequency are extracted according to continuous time windows. Combined with the thermoelectric parameter library, normalized weighted distance matching is performed. Nearby reference operating points are selected to generate capacitance range, equivalent series resistance range, and loss tangent range. These ranges are then discretized into a two-dimensional lookup table indexed by temperature and voltage. This allows the real-time acquisition and execution module to look up and compare the data based on the current temperature and voltage, thereby effectively distinguishing between normal thermal drift and actual performance degradation, and reducing the risk of false alarms and missed alarms caused by the fixed threshold judgment method.

[0037] 4. The analysis and update module calculates the equivalent impedance amplitude within the time window based on the processed waveform, and generates control compensation parameters by combining the impedance change under the reference operating condition. These parameters are then written back to the real-time acquisition and execution module to adjust the AC ripple injection process. This allows the ripple injection to be compensated and controlled according to the change in the hot impedance of the capacitor, reducing current fluctuations in long-term high-power ripple testing and improving the test condition maintenance capability and result consistency. Attached Figure Description

[0038] Other features, objects, and advantages of the invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0039] Figure 1 This is a schematic diagram of the AC / DC capacitor functional test system in an embodiment of this application. Detailed Implementation

[0040] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0041] Please see Figure 1 A functional testing system for AC / DC capacitors includes: a test excitation access module for outputting a DC bias voltage and an AC ripple excitation superimposed on the DC bias voltage, and connecting to the AC / DC capacitor under test via a test bus; the test excitation access module includes a ripple injection power switch unit.

[0042] The real-time acquisition and execution module is used to acquire the terminal voltage, ripple current, surface temperature and common-mode noise detection values ​​of the AC / DC capacitor under test, perform noise reduction processing, parameter estimation and state determination on the acquired signals, and adjust at least one of the DC bias output, AC ripple amplitude, frequency and duty cycle of the test excitation input module according to the determination results.

[0043] The analysis and update module is used to receive data output from the real-time acquisition and execution module, perform thermal drift analysis and working condition matching analysis, generate and write back threshold comparison table, state machine configuration parameters and control compensation parameters, and record test process data, degradation records and protection trigger logs.

[0044] The hardware protection circuit, connected to the real-time acquisition and execution module, is used to receive drive lockout signals and discharge trigger signals when a fault is triggered, and to perform protection in the order of lockout followed by discharge.

[0045] The real-time acquisition and execution module includes a sensor group acquisition unit, a high-speed sampling unit, a sampling shielding unit, a parameter estimation unit, a state machine protection unit, and a register loading unit;

[0046] The sensor group acquisition unit is connected to a high-voltage differential voltage sensor, a broadband current sensor, a temperature sensor, and a common-mode noise detection branch, respectively, to synchronously acquire capacitor terminal voltage, ripple current, surface temperature, and common-mode noise detection values, and to calculate the current change rate.

[0047] The parameter estimation unit is used to estimate capacitance, equivalent series resistance, and loss tangent; the register loading unit is used to receive the threshold lookup table, state machine configuration parameters, and control compensation parameters issued by the analysis update module, and write them into the comparison register and state machine configuration register area.

[0048] The sampling shielding unit sets the sampling suppression interval based on the switching time of the ripple injection power switch unit, and determines the noise threshold based on the background noise.

[0049] When the sampling time is within the sampling suppression interval and the common-mode noise detection value is higher than the noise threshold, effective sampling writing is paused.

[0050] During the pause of valid sampling writing, 2 to 5 valid samples are taken before and after the sampling suppression interval for weighted interpolation recovery, and the recovered sampled values ​​are output to the parameter estimation unit.

[0051] This embodiment provides an overall architecture for an AC / DC capacitor functional testing system and a sampling and reconstruction mechanism for high-frequency electromagnetic interference; this system decouples nonlinear analysis from the underlying high-speed protection control in a layered manner;

[0052] During system operation, the test excitation access module outputs a DC bias voltage and superimposes an AC ripple excitation to the two ends of the AC / DC capacitor under test; the sensor group acquisition unit in the real-time acquisition execution module synchronously acquires the terminal voltage, ripple current, surface temperature and common-mode noise detection values, and calculates the current change rate;

[0053] The acquired raw signal is denoised using a moving average filter, the expression of which is:

[0054]

[0055] in, This is the original sampling sequence. This is the filtered sequence. The length of the sliding window. This is the discrete index of the current sampling time. The sample index number is within the sliding window. The processed signal is sent to the parameter estimation unit to estimate the capacitance, equivalent series resistance, and loss tangent. The specific estimation logic is as follows: The parameter estimation unit uses discrete Fourier transform to extract the fundamental voltage phasor and fundamental current phasor of the terminal voltage and ripple current under the AC ripple excitation frequency. The complex impedance of the capacitor under the frequency is calculated by dividing the fundamental voltage phasor by the fundamental current phasor. The real part of the complex impedance is extracted as the estimated value of the equivalent series resistance.

[0056] The capacitance is estimated by taking the reciprocal of the product of the absolute value of the imaginary part of the complex impedance and the angular frequency; at the same time, the ratio of the real part of the complex impedance to the absolute value of the imaginary part is used as the estimated value of the loss tangent.

[0057] The estimation results are used for real-time status determination and sent to the analysis and update module. The analysis and update module receives the above data stream, performs thermal drift analysis and operating condition matching, and calculates and generates threshold lookup tables, state machine configuration parameters and control compensation parameters.

[0058] The register loading unit receives the above parameters and writes them into the comparison register and state machine configuration register area of ​​the real-time acquisition and execution module to complete the update of the judgment rules;

[0059] When the system detects a fault trigger condition, the hardware protection circuit receives the drive lockout signal and discharge trigger signal from the real-time acquisition and execution module, and performs physical-level disconnection and discharge.

[0060] If only the conventional continuous sampling method is used, common-mode interference can easily cause spurious peaks in the sampled data when the power switching device operates at high speed and causes high-frequency switching. Therefore, the sampling shielding unit establishes a sampling suppression interval based on the switching time of the ripple injected into the power switching unit and determines the noise threshold based on the background noise.

[0061] Within each sampling period, the system compares whether the current time is within the sampling suppression interval and determines whether the common-mode noise detection value is higher than the noise threshold; when both conditions are met—being within the interval and exceeding the noise threshold—the system suspends effective sampling and writing.

[0062] During the pause, the system retrieves 2 to 5 valid samples extracted before and after the suppression interval, and reconstructs the sampling waveform within that time period using a weighted interpolation algorithm. The specific weighting rule is to set the time span of the suppression interval edge based on the time distance of the valid samples, and allocate weights according to the inverse ratio of the time distance. The specific calculation logic is as follows:

[0063] The absolute value of the time difference between the sampling time of each extracted valid sample and the time point of the adjacent suppression interval edge is calculated. To avoid overflow caused by the sample times coinciding with the interval edge, a preset time constant is superimposed on the absolute value of the time difference. The value of the preset time constant is set to be between 0.1% and 1% of the current sampling period. The reciprocal of this shifted time difference is calculated, and then the reciprocal of the time difference of each sample is divided by the sum of the reciprocals of the time differences of all extracted valid samples to obtain the normalized weight, so that the sum of the total weights is strictly equal to the sum of the time differences of all samples. ;

[0064] That is, the closer the effective sample is to the edge of the suppression interval, the higher the weight. Based on this normalized weight distribution, the samples are linearly superimposed to reconstruct the disturbed waveform, and the recovered sampled values ​​are output to the parameter estimation unit.

[0065] The system establishes a first-in-first-out data buffer with a depth greater than the time span of the suppression interval in the underlying field-programmable gate array, applies an equivalent time pipeline delay to the real-time sampled data in the non-suppressed interval, and aligns the timestamps while satisfying the timing correspondence.

[0066] When extracting effective samples at the edge of the set suppression interval, if sensor communication abnormalities or sample value loss are detected, the system will automatically backtrack, read the corresponding phase sample value in the previous effective power frequency or ripple cycle as the substitute sample for the current cycle, and mark it with a low confidence estimation mark to prevent the waveform interpolation recovery process from failing due to the lack of local samples.

[0067] In the test bench for metallized film capacitors used in ultra-high voltage converter valves, after the test system is started, the test excitation access module injects DC high voltage and kilohertz-level AC ripple according to the set operating conditions; the real-time acquisition and execution module acquires the voltage, current and temperature status of the capacitor in real time; when the power semiconductor switch action causes strong transient fluctuations in ground potential, the system freezes the original sample writing within a nanosecond-level time window through the sampling shielding unit, and uses the undisturbed valid samples before and after the interval to smoothly reconstruct the disturbed waveform;

[0068] The reconstructed high-fidelity data is used for real-time parameter estimation, and the results are continuously transmitted to the analysis and update module to ensure that the judgment criteria of the test bench in complex electromagnetic environments are objective and accurate, and to avoid misjudging switching noise as partial discharge of capacitors.

[0069] In a preferred embodiment of the present invention, the analysis and update module extracts operating parameters according to a continuous time window. The operating parameters include at least the average terminal voltage, the root mean square value of ripple current, the temperature rise rate, and the ripple frequency.

[0070] The analysis and update module matches the current operating parameters with the weighted distance between the reference operating points in the thermoelectric parameter library, selects at least one neighboring operating point, and extracts the capacitance range, equivalent series resistance range and loss tangent range under the current operating condition based on the neighboring operating points, and discretizes the range into a two-dimensional lookup table indexed by temperature and voltage.

[0071] The reference operating point parameters include at least one of the following: capacitance, equivalent series resistance, loss tangent, and temperature rise rate threshold.

[0072] The real-time acquisition and execution module calls up the upper and lower limits of capacitance, equivalent series resistance, and loss tangent in the two-dimensional lookup table based on the current surface temperature and capacitor terminal voltage, and compares the real-time estimated capacitance, equivalent series resistance, and loss tangent with the corresponding upper and lower limits respectively.

[0073] When the real-time estimated parameter is between or equal to the corresponding boundary value, a thermal drift flag is output; when the real-time estimated parameter exceeds the corresponding upper or lower limit, a degradation alarm is output and degradation information is recorded.

[0074] The analysis and update module is also used to calculate the equivalent impedance amplitude within the time window based on the processed waveform, and generate control compensation parameters based on the change of the equivalent impedance amplitude relative to the equivalent impedance amplitude under the reference operating condition. The compensated control parameters are then written into the real-time acquisition and execution module. The real-time acquisition and execution module adjusts the AC ripple injection process based on the compensated control parameters.

[0075] The analysis and update module includes a data recording unit and a database update unit; the data recording unit is used to record parameter trajectories, protection trigger logs, and degradation information during the test process;

[0076] The database update unit is used to correct the baseline operating point parameters in the thermoelectric parameter library based on the thermal drift trajectory, performance degradation records and sampling windows before and after each round of testing, and uses the corrected results as the basis for generating the threshold lookup table and state machine configuration parameters for the next round of testing.

[0077] This embodiment includes a thermoelectric coupling drift compensation and control compensation mechanism for the system under long-term high-power ripple test conditions. During the long-term loading of AC ripple, Joule heat accumulates inside the capacitor, causing the temperature to rise and its electrical parameters to drift nonlinearly. If a fixed judgment threshold is used, it is often impossible to distinguish between normal physical thermal drift and actual dielectric performance degradation.

[0078] Therefore, the analysis and update module extracts operating parameters including the mean terminal voltage, the root mean square value of ripple current, the temperature rise rate, and the ripple frequency according to a continuous time window.

[0079] After obtaining the current operating parameters, since the dimensions and numerical ranges of parameters such as terminal voltage, ripple current, and temperature rise rate differ by orders of magnitude, the system uses historical boundary extreme values ​​from the parameter library to normalize each operating parameter. The normalization calculation logic is as follows:

[0080] Subtract the minimum extreme value of the corresponding parameter in the historical parameter database from each currently collected operating condition parameter, then divide by the difference between the maximum and minimum extreme values ​​plus a system-preset bias constant. The value of this preset bias constant is set to the sensor range of the corresponding parameter. to This is to prevent calculation overflow caused by dividing by zero when the extreme value difference is zero due to the parameter not fluctuating within the time window, thus robustly mapping all operating parameters with different dimensions to... dimensional interval;

[0081] After normalization, the system assigns fixed distance measurement weights to the average terminal voltage, ripple frequency, root mean square value of ripple current, and temperature rise rate. , , and Based on this, the weighted Euclidean distance between the current operating parameters and each reference operating point in the preset thermoelectric parameter library is calculated:

[0082]

[0083] in, This is the index number of the operating condition parameter. For the first Distance measurement weights for each working condition parameter For the currently collected number Normalized operating parameters The first reference operating point in the pre-set thermoelectric parameter library One normalized parameter;

[0084] The system selects the K nearest neighboring operating points with the smallest distance. Based on the parameters of these neighboring points, the system calculates a weighted average of the parameters corresponding to the K neighboring points and adds a margin calculated based on the variance of these K samples. The specific rules for obtaining this margin are as follows:

[0085] Calculate the statistical standard deviation of the parameters for K samples. And multiplied by a preset statistical fault tolerance factor. , Based on a device confidence level ranging from 2 to 3.5, let the weighted average of the corresponding parameters be... Then, the upper limit of the allowable parameter range under the current working condition is calculated using the weighted average as the central axis. The lower limit is This allows us to obtain the upper and lower limits of the capacitance range, the equivalent series resistance range, and the loss tangent range, and to discretize the above continuous ranges into a two-dimensional lookup table indexed by temperature and voltage.

[0086] After receiving the two-dimensional lookup table, the real-time acquisition and execution module locates the corresponding table entry based on the currently acquired surface temperature and terminal voltage, and calls the upper and lower limits of capacitance, equivalent series resistance, and loss tangent.

[0087] The system compares the real-time estimated parameter values ​​with the extracted upper and lower limits; if the estimated parameter is within the upper and lower limits, the system outputs a thermal drift indicator; if the estimated parameter exceeds the corresponding upper and lower limit range, the system outputs a degradation alarm command and records degradation information.

[0088] Meanwhile, since thermal drift causes changes in equivalent impedance, the analysis and update module calculates the equivalent impedance amplitude within the time window based on the processed waveform, compares the changes in equivalent impedance amplitude under the benchmark condition to generate control compensation parameters, and writes them back to the real-time acquisition and execution module to adjust the control parameters and stabilize the AC ripple injection process.

[0089] Generate control compensation parameters The specific logic is as follows: the analysis and update module will update the equivalent impedance amplitude within the current time window. Equivalent impedance amplitude under reference operating conditions The ratio is used as a basic proportional coefficient. When the impedance decreases due to heating, this ratio is less than 1. This basic proportional coefficient is then compared with the preset AC ripple voltage amplitude initial setting command. Multiplying these values ​​yields the static feedforward compensation target value, which enables the voltage excitation command to scale synchronously and proportionally with the impedance, thereby maintaining a constant injected ripple current.

[0090] Calculate the equivalent impedance magnitude within the current time window. Its equivalent impedance magnitude within the previous continuous time window The difference is extracted as the dynamic impedance drift, and the dynamic impedance drift is multiplied by the impedance correction gain configured in the system. This gain is equivalent to a reference current in physical dimensions to ensure that the result after multiplication is converted to voltage dimensions, thus obtaining a dynamic compensation term;

[0091] The static feedforward compensation target value is added to the dynamic compensation term to generate the final control compensation parameters, and the calculation formula is as follows:

[0092]

[0093] Initial instructions for reference impedance for The expected ripple current is As the test progresses, the current time window impedance... Due to fever decreasing to The impedance of the previous time window for Impedance correction gain Set as The target value for static feedforward compensation is then... ;

[0094] Dynamic impedance drift is The dynamic compensation item is Final control compensation parameters ;

[0095] After receiving the parameter, the real-time acquisition and execution module directly uses it as the pulse width modulation duty cycle control reference for the ripple injection power switch unit, thereby stabilizing the AC ripple injection process by adjusting the control parameters.

[0096] After a single round of testing, the data recording unit saves the parameter trajectory, protection trigger logs, and degradation information for the entire testing process. The database update unit uses an exponential smoothing algorithm to correct the baseline operating point parameters in the thermoelectric parameter library. The correction formula is as follows:

[0097]

[0098] in, For the updated baseline parameters, For historical baseline parameters, This represents the average value of the parameters recorded in this test. A smoothing coefficient ranging from 0.1 to 0.3 is used as the basis for generating a two-dimensional comparison table for the next round of testing;

[0099] During the process of matching and looking up operating parameters, if the current operating parameters deviate from the weighted distance of all reference points in the thermoelectric parameter library by more than the preset distance threshold due to extreme stress on the tested capacitor, i.e., the weighted distance exceeds the preset extreme value condition, the analysis and update module will pause the generation of a new two-dimensional reference table. The system will automatically maintain the use of the valid reference table in the previous time window for parameter judgment and write the operating condition limit violation record to the test log to prevent the judgment standard from fluctuating beyond the set threshold due to model extrapolation distortion.

[0100] During the composite stress aging test of the metallized film capacitor for the UHV converter valve, which lasted for hundreds of hours, the surface temperature of the capacitor gradually increased as ripple current was continuously injected, and its measured equivalent series resistance changed.

[0101] The analysis and update module acquires temperature and voltage status according to a fixed time window, matches the allowable normal resistance fluctuation range under the current stress from the historical thermoelectric parameter database, and generates a two-dimensional comparison table which is sent to the real-time acquisition and execution module.

[0102] The real-time acquisition and execution module compares the real-time temperature and voltage with a table to confirm that the current resistance change does not exceed the corresponding upper and lower limits, and determines it to be normal thermal drift.

[0103] Meanwhile, the analysis and update module calculates the decrease in the equivalent impedance of the capacitor, generates control compensation parameters and sends them to the real-time acquisition and execution module, which then fine-tunes the ripple output gain to suppress low-frequency fluctuations in the test current.

[0104] After a single aging test is completed, the system adds the actual temperature rise and parameter drift trajectory of this test to the database, so that the judgment threshold of subsequent test batches is more in line with the actual physical characteristics of the components in that batch.

[0105] In a preferred embodiment of the present invention, the real-time acquisition and execution module constructs a transient criterion based on the terminal voltage change rate, ripple current change rate, and common-mode noise detection value, and determines that a serious fault has occurred when the ripple current change rate is greater than a preset ripple current change rate threshold and the transient criterion is greater than a judgment threshold.

[0106] Among them, the ripple current change rate threshold is generated based on the preset normal test extreme value and failure test extreme value calibration, and the judgment threshold is generated based on the statistical mean and variance calibration of the transient criterion under steady-state conditions.

[0107] The hardware protection circuit includes a power drive interlock circuit, a crowbar discharge circuit, an energy dissipation resistor, and a protection isolation drive unit;

[0108] After a fault is determined, the state machine protection unit first outputs a drive lockout signal within 50ns, and then outputs a crowbar conduction signal within 200ns after the lockout power drive, so that the surge energy is introduced into the dissipation branch; at the same time, it freezes the preset sampling window data before and after the fault is triggered and sends it to the analysis and update module.

[0109] This embodiment includes a nanosecond-level hardware protection mechanism for the system under capacitor dielectric breakdown or transient energy avalanche.

[0110] In conventional test protection logic, shutdown actions are usually performed by relying on operating system software polling or communication interruption. This method has a millisecond-level response delay. When a capacitor experiences transient breakdown, surge energy flows backward and damages the test equipment before the software response.

[0111] To this end, the real-time acquisition and execution module directly constructs transient criteria at the underlying level based on the terminal voltage change rate, ripple current change rate, and common-mode noise detection value.

[0112] Based on the physical characteristics of ripple current overshoot during non-self-healing capacitor breakdown, and the accompanying voltage drop slope exceeding a preset extreme value, a transient criterion is constructed. At that time, the underlying logic normalizes the collected physical change rates by dividing them by the normal fluctuation extreme values ​​under the rated operating conditions of the system, and designs the transient criterion as: the rate of change of terminal voltage drop. Extreme values ​​of normal fluctuations in relative terminal voltage The absolute value of the per-unit value, and the rate of change of the ripple current rise. relative ripple current normal fluctuation extreme value The product of the per-unit values, plus the penalty coefficient of the common-mode noise mutation mapping. The calculation formula is as follows:

[0113]

[0114] Among them, the penalty coefficient of the common-mode noise mutation mapping The establishment rule is: extract the current common-mode noise detection value. The mean of background common-mode noise under steady-state conditions The difference, when the difference is lower than the calibrated interference tolerance When the amplitude exceeds the interference tolerance, the penalty coefficient is zero. When the amplitude exceeds the interference tolerance, the penalty coefficient is increased proportionally to the excess amplitude, combined with the penalty gain. Positive linear increase, its calculation formula is: when hour, ;when hour, ;

[0115] For example, in a certain test, , Detected The absolute value after standardization is 3; After standardization, it becomes 4; at this point for , for , for , ;

[0116] The difference is , Final transient criterion ;

[0117] The system continuously monitors the rate of change of ripple current. When it exceeds the preset threshold for the rate of change of ripple current and the transient criterion exceeds the judgment threshold, that is, when the current increases sharply and the voltage drops sharply at the same time, causing the product term to be significantly amplified, for example, when the transient criterion 17 exceeds the judgment threshold 10, the system immediately determines that a serious fault has occurred.

[0118] Once a fault is determined, in order to meet the nanosecond-level hardware protection response requirements, the state machine protection unit relies on the pure hardware logic implementation of the field programmable gate array. It directly performs threshold discrimination on the sampled data stream through the internal high-speed digital comparator, avoiding software interrupt response and bus transmission delay.

[0119] Its internal clock operates at a frequency of at least 200MHz, which keeps the single-step logic judgment time within 5ns; therefore, the state machine protection unit can bypass the software scheduling layer and directly send a drive lockout signal to the power drive lockout circuit within 50ns to cut off the power output of the test excitation access module.

[0120] Within 200ns after the locking action, a crowbar conduction signal is output to trigger the crowbar discharge circuit, which diverts the residual surge energy on the bus to the energy dissipation resistor. At the same time, the system freezes the preset sampling window data before and after the fault trigger point and sends it to the analysis and update module for subsequent fault source analysis.

[0121] In the execution logic of transient judgment, if the ripple current change rate exceeds the preset change rate limit due to external mechanical vibration or loose wiring of the broadband current sensor, the system will perform cross-validation by combining the common-mode noise weight and the terminal voltage change rate in the transient criterion.

[0122] If the current change rate is found to exceed the preset upper limit but the terminal voltage does not drop accordingly and the common mode noise is at the normal background level, the system will intercept this serious fault judgment, determine it as a sensor link abnormality and trigger the equipment maintenance instruction, thereby avoiding the test bench from being shut down due to poor single-point hardware contact.

[0123] During the ultimate failure boundary exploration test of the metallized film capacitor used in the UHV converter valve, when the applied DC bias voltage exceeded the tolerance limit of the dielectric film, non-self-healing breakdown occurred inside the capacitor; the real-time acquisition and execution module captured the sharp increase in the ripple current change rate and the accompanying abnormal common-mode noise within a nanosecond cycle, and the transient criterion was greater than the judgment threshold.

[0124] Without the need for analysis and update module intervention, the state machine protection unit immediately outputs a drive lockout signal to the power drive lockout circuit, cuts off the power output of the test excitation access module, and outputs a crowbar conduction signal to the crowbar discharge circuit, so as to quickly dissipate the surge energy generated by the breakdown on the energy dissipation resistor.

[0125] Meanwhile, the system fully preserves high-frequency waveform slices before and after breakdown and uploads them to the analysis and update module, protecting the test bench from backflow damage while providing the system with accurate data on the instantaneous physical failure.

Claims

1. A functional testing system for AC / DC capacitors, characterized in that, include: The test excitation access module is used to output a DC bias voltage and an AC ripple excitation superimposed on the DC bias voltage, and is connected to the AC / DC capacitor under test via the test bus. The test excitation access module includes a ripple injection power switch unit. The real-time acquisition and execution module is used to acquire the terminal voltage, ripple current, surface temperature and common-mode noise detection value of the AC / DC capacitor under test, perform noise reduction processing, parameter estimation and state determination on the acquired signal, and adjust at least one of the DC bias output, AC ripple amplitude, frequency and duty cycle of the test excitation access module according to the determination result. The analysis and update module is used to receive the data output by the real-time acquisition and execution module, perform thermal drift analysis and working condition matching analysis, generate and write back the threshold comparison table, state machine configuration parameters and control compensation parameters, and record test process data, degradation records and protection trigger logs. The hardware protection circuit, connected to the real-time acquisition and execution module, is used to receive the drive lockout signal and the discharge trigger signal when a fault is triggered, and to perform protection in the order of lockout followed by discharge.

2. The AC / DC capacitor functional testing system according to claim 1, characterized in that, The real-time acquisition and execution module includes a sensor group acquisition unit, a high-speed sampling unit, a sampling shielding unit, a parameter estimation unit, a state machine protection unit, and a register loading unit; The sensor group acquisition unit is connected to a high-voltage differential voltage sensor, a broadband current sensor, a temperature sensor, and a common-mode noise detection branch, respectively, to synchronously acquire capacitor terminal voltage, ripple current, surface temperature, and common-mode noise detection values, and to calculate the current change rate. The parameter estimation unit is used to estimate capacitance, equivalent series resistance, and loss tangent; the register loading unit is used to receive the threshold lookup table, state machine configuration parameters, and control compensation parameters issued by the analysis update module, and write them into the comparison register and state machine configuration register area.

3. The AC / DC capacitor functional testing system according to claim 2, characterized in that, The sampling shielding unit sets the sampling suppression interval according to the switching time of the ripple injection power switching unit, and determines the noise threshold according to the background noise. When the sampling time is within the sampling suppression interval and the common-mode noise detection value is higher than the noise threshold, effective sampling writing is paused. During the pause of valid sampling writing, 2 to 5 valid samples are taken before and after the sampling suppression interval for weighted interpolation recovery, and the recovered sampled values ​​are output to the parameter estimation unit.

4. The AC / DC capacitor functional testing system according to claim 1, characterized in that, The analysis and update module extracts operating parameters according to a continuous time window. The operating parameters include at least the average terminal voltage, the root mean square value of ripple current, the temperature rise rate, and the ripple frequency. The analysis and update module matches the current operating parameters with the weighted distance between the reference operating points in the thermoelectric parameter library, selects at least one neighboring operating point, and extracts the capacitance range, equivalent series resistance range and loss tangent range under the current operating condition based on the neighboring operating points, and discretizes the range into a two-dimensional lookup table indexed by temperature and voltage. The reference operating point parameters include at least one of the following: capacitance, equivalent series resistance, loss tangent, and temperature rise rate threshold.

5. The AC / DC capacitor functional testing system according to claim 4, characterized in that, The real-time acquisition and execution module calls up the upper and lower limits of capacitance, equivalent series resistance, and loss tangent in the two-dimensional lookup table according to the current surface temperature and capacitor terminal voltage, and compares the real-time estimated capacitance, equivalent series resistance, and loss tangent with the corresponding upper and lower limits respectively. When the parameter obtained by real-time estimation is between the corresponding upper and lower limits or equal to the corresponding boundary value, a thermal drift flag is output; when the parameter obtained by real-time estimation exceeds the corresponding upper and lower limits, a degradation alarm is output and degradation information is recorded.

6. The AC / DC capacitor functional testing system according to claim 1, characterized in that, The analysis and update module is also used to calculate the equivalent impedance amplitude within the time window based on the processed waveform, and generate control compensation parameters based on the change of the equivalent impedance amplitude relative to the equivalent impedance amplitude under the reference operating condition, and then write the compensated control parameters into the real-time acquisition and execution module. The real-time acquisition and execution module adjusts the AC ripple injection process based on the compensated control parameters.

7. The AC / DC capacitor functional testing system according to claim 2, characterized in that, The real-time acquisition and execution module constructs transient criteria based on the terminal voltage change rate, ripple current change rate, and common-mode noise detection value. When the ripple current change rate is greater than the preset ripple current change rate threshold and the transient criteria are greater than the judgment threshold, a serious fault is determined to have occurred. The ripple current change rate threshold is generated based on the preset normal test extreme value and failure test extreme value calibration, and the judgment threshold is generated based on the statistical mean and variance calibration of the transient criterion under steady-state conditions.

8. The AC / DC capacitor functional testing system according to claim 7, characterized in that, The hardware protection circuit includes a power drive interlocking circuit, a crowbar discharge circuit, an energy dissipation resistor, and a protection isolation drive unit. After the fault determination is established, the state machine protection unit first outputs a drive lockout signal within 50ns, and then outputs a crowbar conduction signal within 200ns after the lockout power drive, so that the surge energy is introduced into the dissipation branch. Simultaneously, the preset sampling window data before and after the fault is triggered is frozen and sent to the analysis and update module.

9. The AC / DC capacitor functional testing system according to claim 1, characterized in that, The analysis and update module includes a data recording unit and a database update unit; The data recording unit is used to record parameter trajectories, protection trigger logs, and degradation information during the testing process; The database update unit is used to correct the baseline operating point parameters in the thermoelectric parameter library based on the thermal drift trajectory, performance degradation records and sampling windows before and after each round of testing, and to use the corrected results as the basis for generating the threshold lookup table and state machine configuration parameters in the next round of testing.