Electrolytic capacitor life calculation method based on ripple analysis

Abstract

This invention relates to the field of power electronics technology, and in particular to a method for calculating the lifetime of electrolytic capacitors based on ripple analysis, comprising the following steps: S1, constructing a three-field coupled partial differential equation system; S2, real-time ripple data acquisition and preprocessing; S3, online identification of key electrical parameters; S4, calculation of dynamic heat loss and internal hot spot temperature; S5, prediction of remaining lifetime and output of confidence interval; S6, health management and early warning output. This invention, through the constructed three-field coupled partial differential equation system of electric, thermal, and concentration fields, deeply couples the evolution process of the electric field, thermal field, and concentration field, accurately capturing the nonlinear coupled positive feedback effect of electrolytic capacitor aging from a physical mechanism perspective. This makes the model's characterization of the electrolytic capacitor aging process more realistic, improving the accuracy of lifetime prediction from a physical mechanism perspective.

Description

Technical Field

[0001] This invention relates to the field of power electronics technology, and in particular to a method for calculating the lifetime of electrolytic capacitors based on ripple analysis. Background Technology

[0002] With the development of power electronics technology, DC-DC converters play a crucial role in renewable energy power generation, smart grids, aerospace, and marine applications. Due to the switching devices in DC-DC converters, their output voltage contains switching frequency harmonics and their integer multiples. Therefore, filter circuits, such as LC filters and LCL filters, are required on the output side of DC-DC converters. Capacitors are indispensable in filter circuits, and electrolytic capacitors, with their advantages of large capacitance, high energy density, and low cost, are widely used in the design of output-side filter circuits for DC-DC converters. However, due to limitations in their structure and material properties, electrolytic capacitors are among the components with the highest failure probability in power electronic equipment. Their aging failure is one of the main causes of unplanned equipment downtime. Therefore, accurately predicting the remaining lifespan of electrolytic capacitors is of great significance for the condition monitoring and intelligent operation and maintenance of power electronic equipment.

[0003] Currently, the main method for predicting the lifespan of electrolytic capacitors is the empirical model method. This method only considers temperature as the core influencing factor, ignoring the self-heating effect of the capacitor caused by ripple current and the impact of dynamic changes in capacitor parameters during aging on the lifespan, resulting in a large error in lifespan estimation. Therefore, it is necessary to design a method for calculating the lifespan of electrolytic capacitors based on ripple analysis. Summary of the Invention

[0004] In order to overcome the shortcomings of the prior art, the purpose of this invention is to provide a method for calculating the life of electrolytic capacitors based on ripple analysis.

[0005] The technical solution adopted in this invention is: a method for calculating the lifetime of electrolytic capacitors based on ripple analysis, comprising the following steps: S1. Construct a three-field coupled partial differential equation system: Establish a three-field coupled partial differential equation system of electro-thermal-concentration for electrolytic capacitors to characterize the dynamic evolution of ESR and C with internal hot spot temperature, electrolyte concentration, and ripple stress; complete the calibration of the physical parameters of the model in the equation system by combining accelerated aging experiments with nonlinear least squares method, and obtain the effective calibration parameters after model verification. S2. Real-time ripple data acquisition and preprocessing: By acquiring the ripple current, ripple voltage, shell temperature and ambient temperature data of the electrolytic capacitor in real time, the acquired data is sequentially processed to remove DC components, filter and denoise, and synchronize and align. S3. Online identification of key electrical parameters: For ESR, the frequency domain method is used and for C value, the ripple voltage charge-discharge slope method is used to perform targeted analysis on the preprocessed ripple data, so as to realize the real-time online identification of ESR and C value. S4. Calculation of dynamic heat loss and internal hot spot temperature: The real-time heat loss of the capacitor is calculated based on the identified ESR and RMS value of ripple current. The internal hot spot temperature of the capacitor is calculated by combining the thermal field equation in the three-field coupled partial differential equation system and the measured shell temperature. S5. Remaining Life Prediction and Confidence Interval Output: Substitute the currently identified electrical parameters, internal hotspot temperatures, and the model physical parameters calibrated by accelerated aging experiments obtained in S1 into a three-field coupled partial differential equation system. Use numerical solution methods and simulate the evolution of ESR and C values. Combine this with a general failure threshold to obtain the remaining life prediction value. The Monte Carlo simulation method was used to introduce random perturbations into the key parameters of the model, and multiple sets of remaining lifetime sample values ​​were obtained. The lower limit of the remaining lifetime was determined through statistical analysis. and upper limit value Simultaneously calculate the lifespan attrition rate, cumulative lifespan attrition, and remaining lifespan percentage; S6. Health Management and Early Warning Output: Calculate the corresponding health index based on the initial values ​​of ESR and C values ​​and the general failure threshold, and take the minimum of the two as the final health index. ,according to The system implements tiered early warnings and outputs these warnings.

[0006] As a further description of the above technical solution: The three-field coupled partial differential equation set in step S1 includes the electric field dynamic equation, the thermal field dynamic equation, and the concentration field diffusion equation. The dynamic equation of the electric field is: in, These are the material-related constants for ESR evolution; The activation energy for ESR evolution; It is the gas constant; This refers to the temperature of the hot spot inside the capacitor. Ripple current density; This represents the initial concentration of the electrolyte. Real-time electrolyte concentration; ESR is an empirical index of ESR evolution. The C-value is a material-related constant for evolution. The activation energy for C-value evolution; This represents the real-time RMS value of the ripple current. This is the rated ripple current of the capacitor; The empirical index for the evolution of C-value; The dynamic equation for the thermal field is: in, The dielectric density of the capacitor; The specific heat capacity of the capacitor dielectric at constant pressure; The rate of change of temperature over time; For Hamiltonian operators; The thermal conductivity of the capacitor dielectric; For heat source terms per unit volume; The concentration field diffusion equation is: in, The rate of change of electrolyte concentration over time; The diffusion coefficient is temperature-dependent. is the Laplace operator, representing the spatial second-order rate of change of concentration; Let $\frac{ ... The activation energy for electrolyte volatilization; The minimum effective electrolyte concentration; It is a nonnegative operator, characterized only if If the difference is true, take the difference; otherwise, take 0.

[0007] As a further description of the above technical solution: In step S2, the ripple current is acquired through a shunt; the ripple voltage is acquired through a voltage divider circuit and is sampled synchronously with the ripple current; the casing temperature is acquired through a surface-mount NTC thermistor; and the ambient temperature is acquired through an ambient temperature sensor. The data preprocessing includes removing DC components, filtering and noise reduction, and synchronous alignment.

[0008] As a further description of the above technical solution: The specific process of identifying ESR using the frequency domain method in step S3 is as follows: Perform a Fast Fourier Transform (FFT) on the ripple current and ripple voltage to calculate the switching frequency and its harmonics. The impedance calculation formula is: ,in, For frequency The impedance value below; The frequency domain amplitude of the ripple voltage; The frequency domain amplitude of the ripple current is used. After separating the effects of capacitive and inductive reactance, the ESR value at each frequency point is extracted. The arithmetic mean method is used to fuse the results of multiple frequency points to obtain the final ESR value.

[0009] As a further description of the above technical solution: The specific process of identifying the C value using the ripple voltage charge / discharge slope method in step S3 is as follows: Based on the preprocessed ripple voltage time-domain waveform, the charging and discharging phases within the ripple period are identified. The time period of voltage rise during the charging phase is selected, and the slope of voltage change with time is solved by linear fitting. Simultaneously calculate the average ripple current during this time period. Substitute the values ​​into the formula to calculate the value of C. The formula is: Where C is the capacitance value; This represents the average value of the ripple current during the charging and discharging phases. This represents the slope of the ripple voltage during the charging and discharging phase.

[0010] As a further description of the above technical solution: The formula for calculating real-time heat loss in step S4 is as follows: ,in, This refers to the real-time heat loss of the capacitor. RMS value of ripple current; ESR 实时 For real-time equivalent series resistance; The calculation model for the internal hotspot temperature includes a steady-state simplified model and a transient accurate model. The steady-state simplified model is as follows: ,in, This refers to the measured temperature of the capacitor casing. The thermal resistance from the internal hot spot of the capacitor to the outer casing; Transient exact model is ,in, for Internal hotspot temperature at any given moment; for The outer casing temperature at any given time; for Heat loss over time; for The transient thermal impedance function at time t.

[0011] As a further description of the above technical solution: In step S5, the numerical solution method is the finite difference method. The time and spatial domains of the three-field coupled partial differential equations are discretized. The time step is adaptively adjusted, ranging from 0.1h to 1h, and the spatial step is set according to the capacitor structure. This transforms the partial differential equations into a linear system of equations. The evolution curves of ESR and C values ​​over time are obtained through iterative solutions. The general failure threshold is: under the electrolyte volatilization-dominated mode, the failure threshold is... , The initial ESR value; under the dominant anodic foil corrosion mode, the failure threshold is... , The initial value of C is used; the electrolytic capacitor is determined to be in failure if any failure threshold is met; the remaining lifetime prediction value is the iteration time from the current moment to when the ESR or C value reaches the failure threshold.

[0012] As a further description of the above technical solution: The Monte Carlo simulation method obtained in step S5 and The specific process is as follows: (1) Select , , , , , For the key perturbation parameters of the model, a random perturbation of ±5% is introduced into each parameter within the range of 95%-105% of its calibration value; (2) Generate 100 sets of mutually independent perturbation parameter combinations by random sampling; (3) Substitute each set of parameters into the three-field coupled partial differential equation system in turn, and use the finite difference method to solve independently to obtain the corresponding remaining lifetime prediction value, forming 100 remaining lifetime sample values. (4) Sort the 100 remaining lifetime sample values ​​in ascending order, and take the 2.5th percentile value as the lower limit of remaining lifetime. The 97.5th percentile value is taken as the upper limit of the remaining lifespan. ; The formula for the lifespan attrition rate is: ,in, For discrete, small time steps; This refers to the single-step long lifespan consumption rate; The formula for cumulative lifespan consumption is: ,in, This refers to the cumulative lifespan attrition rate. The formula for the percentage of remaining lifespan is: ,in, This represents the percentage of remaining lifespan.

[0013] As a further description of the above technical solution: The health index in step S6 Based on the aging mode, it was determined that under the electrolyte volatilization-dominated mode... ; Under the dominant mode of anodic foil corrosion, ; in, As a health index; The ESR failure threshold; For real-time equivalent series resistance; This is the initial value for ESR; The C-value is the failure threshold; This is the real-time C value; The initial value for C; The graded early warning is as follows: when 0.1≤HI≤0.3, a Level 1 early warning is issued, recommending that the electrolytic capacitor be replaced in the near future; when HI<0.1, a Level 2 early warning is issued, requiring the electrolytic capacitor to be replaced immediately; the early warning output method includes one or more of HMI display, CAN communication upload, and relay contact alarm.

[0014] The present invention has the following beneficial effects: 1. This invention uses a set of coupled partial differential equations for the electric, thermal, and concentration fields to deeply couple the evolution of these fields. The heat source term is determined by the ripple current and the ESR, which changes dynamically with temperature and concentration. The evolution of the ESR is related to the electrolyte concentration and the temperature of the internal hot spots. The change in electrolyte concentration is also affected by temperature. This invention accurately captures the nonlinear coupled positive feedback effect of electrolytic capacitor aging from a physical mechanism perspective, making the model's characterization of the electrolytic capacitor aging process more realistic and improving the accuracy of lifetime prediction from a physical mechanism perspective.

[0015] 2. This invention constructs a health index based on aging modes and electrical parameters, strictly follows the hierarchical early warning rules of the technical solution, and realizes intelligent management of the entire process from condition monitoring to life prediction to early warning output. Early warning information is output in multiple ways, providing clear and explicit decision-making basis for operation and maintenance personnel, effectively avoiding unplanned shutdowns of power electronic equipment caused by electrolytic capacitor failure, and improving the stability and reliability of equipment operation. Attached Figure Description

[0016] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0017] Reference Figure 1 This invention provides a method for calculating the lifetime of electrolytic capacitors based on ripple analysis, comprising the following steps: S1. Construct a system of three coupled partial differential equations: By introducing an electrolyte concentration field, a system of partial differential equations deeply coupled with the electric field, thermal field, and concentration field is constructed to accurately characterize the dynamic process and positive feedback effect of electrolytic capacitor aging. The implementation process is divided into three stages: model construction, parameter calibration, and model verification, as detailed below: Model Construction: A three-field coupled partial differential equation system was constructed strictly according to the formulas of the electric field dynamic equation, thermal field dynamic equation, and concentration field diffusion equation in the technical solution. The physical meaning of each equation was clarified: the electric field equation describes the evolution of ESR and C values ​​with temperature, electrolyte concentration, and ripple stress; the thermal field equation reflects the internal temperature distribution of the capacitor and the heat loss transmission process; the concentration field equation describes the spatiotemporal changes in concentration caused by electrolyte evaporation, and is expressed through the heat source term. The correlation between ESR and temperature and concentration achieves three-field coupling.

[0018] The dynamic equation of the electric field is: in, These are the material-related constants for ESR evolution; The activation energy for ESR evolution; It is the gas constant; This refers to the temperature of the hot spot inside the capacitor. Ripple current density; This represents the initial concentration of the electrolyte. Real-time electrolyte concentration; ESR is an empirical index of ESR evolution. The C-value is a material-related constant for evolution. The activation energy for C-value evolution; This represents the real-time RMS value of the ripple current. This is the rated ripple current of the capacitor; The empirical index for the evolution of C-value; The dynamic equation of the thermal field is: in, The dielectric density of the capacitor; The specific heat capacity of the capacitor dielectric at constant pressure; The rate of change of temperature over time; For Hamiltonian operators; The thermal conductivity of the capacitor dielectric; For heat source terms per unit volume; The concentration field diffusion equation is: in, The rate of change of electrolyte concentration over time; The diffusion coefficient is temperature-dependent. is the Laplace operator, representing the spatial second-order rate of change of concentration; Let $\frac{ ... The activation energy for electrolyte volatilization; The minimum effective electrolyte concentration; It is a nonnegative operator, characterized only if If the difference is true, take the difference; otherwise, take 0.

[0019] Parameter calibration: (1) Sample selection: Select electrolytic capacitors of the same batch and specifications as the electrolytic capacitors to be monitored as calibration samples to avoid parameter calibration errors caused by differences in production process and material batches; (2) Platform construction: A dedicated accelerated aging test platform is built, equipped with a constant temperature control module, a ripple current output module, and a data acquisition module, which can realize precise control of temperature and ripple current and synchronous acquisition of multi-source data. (3) Operating Condition Settings: Multiple temperature gradients of 65℃, 85℃, 95℃, and 105℃ are set, along with a 0.5... 1.0 1.2 Multiple sets of ripple current stress gradients cover the typical operating conditions of power electronic equipment; (4) Data acquisition: The calibration sample was placed in the above multi-gradient experimental conditions, and the full-dimensional evolution data of ESR, C value, internal hot spot temperature and real-time electrolyte concentration during the aging process of the sample were continuously collected to provide a real and complete experimental basis for parameter fitting. (5) Fitting and Solving: Based on the three-field coupled partial differential equation system, the nonlinear least squares method is adopted, with the mean square error as the objective function, to minimize the deviation between the calculated model value and the experimental measured value, and the solution is obtained iteratively. , , , , , The calibration value.

[0020] Model Validation: The fitted model physical parameter calibration values ​​are substituted into the three-field coupled partial differential equation system, and the equation system is numerically solved using the finite difference method to obtain the evolution curves of ESR and C values ​​over time. The simulated evolution curves are compared with the measured evolution curves from accelerated aging experiments to verify the goodness of fit of the model and the model's ability to capture the nonlinear coupling positive feedback effect of electrolytic capacitor aging. Only calibration parameters whose simulated curves closely match the measured curves and whose coupling effects are effectively reflected are stored in the parameter library as valid calibration parameters to provide a basis for numerical solutions in subsequent steps.

[0021] S2. Real-time ripple data acquisition and preprocessing: Real-time acquisition and standardized preprocessing of multi-source data are achieved to provide a high-quality data source for subsequent electrical parameter identification. The implementation process is divided into two stages: real-time data acquisition and data preprocessing, which is completely consistent with the technical solution requirements, as detailed below: Real-time data acquisition 1.1 Ripple current: A shunt is used to collect the data, and the sampling rate is strictly set to ≥20 times the switching frequency of the power electronic equipment (e.g., when the switching frequency is 5kHz, the sampling rate is ≥100kHz) to ensure complete capture of the waveform characteristics of the ripple current. 1.2 Ripple Voltage: High-precision resistor voltage divider circuit is used for sampling. The voltage division ratio matches the rated voltage of the capacitor. The bandwidth of the sampling circuit is consistent with that of the current sampling and is strictly synchronized with the ripple current sampling. The sampling clock error is ≤1μs. 1.3 Case Temperature: A surface-mount NTC thermistor (accuracy ±0.1℃) is mounted at the center of the capacitor case, with a sampling rate ≥1Hz; 1.4 Ambient Temperature: A digital temperature sensor (accuracy ±0.5℃) is used to collect the ambient temperature inside the equipment cabinet, with a sampling rate ≥0.1Hz; 1.5 All acquired data is converted into digital signals by a high-speed AD converter and transmitted to the FPGA embedded processing unit to prepare for subsequent preprocessing and analysis.

[0022] Data preprocessing: Perform the preprocessing requirements of DC component removal, filtering and noise reduction, and synchronization alignment as outlined in the technical solution. The specific operations are as follows: 1.1 DC Component Removal: A digital high-pass filter (cutoff frequency 0.1Hz) is used to remove the DC component from the ripple voltage and ripple current, retaining only the AC ripple portion. The engineering calculation formula is as follows: ,in, These are the original collected values. This is the ripple value after DC removal; 1.2 Filtering and Denoising: Wavelet thresholding is used to denoise the device. The db4 wavelet is selected as the base wavelet and the decomposition layer is 5. The wavelet coefficients are processed by the soft thresholding method to eliminate random noise such as sensor noise and circuit electromagnetic interference. 1.3 Synchronization Alignment: Based on the sampling clock of the ripple current, the ripple voltage, casing temperature and ambient temperature data are synchronized on the time axis to ensure that the multi-source data under the same timestamp correspond one-to-one, and the synchronization error is ≤1ms.

[0023] S3. Online identification of key electrical parameters: An online identification process was designed separately for ESR and C values. The preprocessed ripple data was analyzed to achieve real-time online identification of ESR and C values, as detailed below: 3.1 ESR Online Identification: ESR is a core sensitive parameter for the aging of electrolytic capacitors, and its identification accuracy directly affects the heat loss and life prediction results. A frequency domain method is used for accurate correction, and the specific process is as follows: Precise correction using frequency domain method: 1.1 Perform Fast Fourier Transform (FFT) on the preprocessed ripple voltage and ripple current, with the number of transform points set to 1024 to ensure that the frequency domain resolution meets the requirements of switching frequency analysis; 1.2 Extracting Feature Frequency Points: Based on the switching frequency of power electronic equipment Select (baseband) (Second harmonic) (Third harmonic) Three characteristic frequency points to avoid single frequency point being affected by harmonic interference or waveform distortion; 1.3 Calculate the impedance at each frequency point: based on the frequency domain amplitude and Substitute into the impedance formula Calculate the impedance magnitude at the three frequency points respectively. ; 1.4 Separating capacitive and inductive reactance: Given the equivalent inductance of an electrolytic capacitor According to the formula Calculate the inductive reactance at each frequency; assume the current C value is the rated capacitance value. (Initial stage) or the identification value of the previous cycle (subsequent cycle), according to the formula Calculate the capacitive reactance at each frequency point; 1.5 Extracting the pure resistive component: Calculate the ESR value at each frequency point using the geometric relationship between the impedance magnitude and the reactance component. ESR(f0), ESR(2f0), and ESR(3f0) were obtained respectively. 1.6 Final ESR Value of Fusion Output: The identification results of three frequency points are fused using the arithmetic mean method, and the formula is as follows: If the impedance calculation at a certain frequency point is abnormal (such as the amplitude being too small, resulting in insufficient signal-to-noise ratio), the point will be automatically removed, and the average value of the remaining two points will be used to ensure the robustness of the identification results.

[0024] 1.7 Real-time performance guarantee: The ESR identification algorithm is executed in parallel in the FPGA, with a single frame data processing time of ≤10ms. The data update frequency is synchronized with the ripple sampling frequency (≥20 times the switching frequency), which meets the requirements of online real-time monitoring.

[0025] 3.2C value online identification: The change in the C value directly reflects the degree of corrosion of the anode foil. The charge-discharge slope method is used, and the specific process is as follows: Charge / discharge slope method: 1.1 Ripple Voltage Stage Division: Based on the pre-processed ripple voltage time-domain waveform The charging and discharging stages are divided using the slope threshold method, and a slope threshold is set. ,when When it is determined to be in the charging stage, When the discharge phase is determined, items with an absolute slope value less than [value missing] are excluded. The smooth section avoids interference from ripple fluctuations; 1.2 Effective Stage Selection: During the charging stage, the voltage is selected from... Rise to The time period is taken as the effective charging period. The voltage change in this interval is highly linear, which can effectively avoid the nonlinear effects of initial charging delay and saturation stage. 1.3 Slope Calculation: Linear fitting is performed on the voltage data of the effective charging segment. The fitting formula is as follows: , where the slope The goodness of fit must meet the requirements; otherwise, a new effective segment must be selected. 1.4 Average Current Calculation: Extract the ripple current data corresponding to the effective charging segment and calculate the average current within that time period. ,in The number of sampling points for the effective charging segment; 1.5C value calculation: Substitute into the charge / discharge slope method formula This gives us the current value of C.

[0026] To verify the accuracy of the modular identification method for ESR and C value, a simulation model was built in Matlab to simulate different working conditions (normal waveform, 5% harmonic distortion, 10% harmonic distortion), with 0.5%-2% white noise added. The identification results are compared with the true values, as shown in Table 1. The average relative error of ESR is ≤2.2%, and the average relative error of C value is ≤1.5%. Compared with the traditional single method (ESR error ≥3.5%, C value error ≥2.8%), the identification accuracy is significantly improved.

[0027] Table 1 shows the simulation error table for ESR and C value module identification. S4. Calculation of dynamic heat loss and internal hot spot temperature: Based on the ESR identified in step S3, real-time heat loss is calculated. Combining the thermal field equation from step S1 and the measured shell temperature, the internal hot spot temperature is calculated, as follows: Real-time heat loss calculation: A formula is directly used for real-time engineering calculations. The formula is: in, For the real-time heat loss of the capacitor, The RMS value is the ripple current, and ESR is the real-time equivalent series resistance identified in step S3. The heat loss value is updated once per frame of data. Temperature unit conversion: If the collected casing temperature is in Celsius. It needs to be converted to thermodynamic temperature. Substitute the values ​​into the formula, and the conversion formula is: The temperature units are consistent with those in step S1. Internal hotspot temperature calculation: Provides two calculation methods: a steady-state simplified model and a transient accurate model. Simplified steady-state model: Applicable to equipment steady-state operation (load fluctuation ≤5%, temperature change ≤1℃ / min), the formula is as follows: ,in, This refers to the measured temperature of the capacitor casing. The thermal resistance from the internal hot spot of the capacitor to the outer casing; Transient accurate model: Applicable to equipment operating under varying conditions (load fluctuation > 5%, temperature change > 1℃ / min), the formula is as follows: ,in, for Internal hotspot temperature at any given moment; for The outer casing temperature at any given time; for Heat loss at any time; for The transient thermal impedance function at any given time, calibrated from the capacitor datasheet or through offline experiments, can accurately capture the dynamic changes in hot spot temperature.

[0028] S5. Remaining lifetime prediction and confidence interval output: The remaining lifetime is predicted using a numerical solution method, and the confidence interval is output through Monte Carlo simulation, as detailed below: Numerical solution of the three-field coupled equations: The real-time ESR and C values ​​identified in step S3, the internal hotspot temperatures calculated in step S4, and the model physical parameters calibrated by accelerated aging experiments obtained in step S1 are substituted into the three-field coupled partial differential equations and solved numerically using the finite difference method. The time and spatial domains are discretized, with the time step adaptively adjusted (0.1h-1h) and the spatial step set according to the capacitor structure. The partial differential equations are transformed into a linear equation system, and the evolution curves of ESR and C values ​​over time are obtained through iterative solution.

[0029] Failure Threshold Determination and Remaining Life Calculation: The corresponding failure threshold is determined based on a general failure threshold. Under the electrolyte evaporation-dominated mode, the failure threshold is: , The initial ESR value; under the dominant anodic foil corrosion mode, the failure threshold is... , The initial value for C is used; failure is determined when any failure threshold is met. Starting from the current moment, the solution is iterated until either the ESR or the C value reaches the failure threshold; the iteration time at this point is the predicted remaining lifetime. (h).

[0030] Confidence interval calculation: The Monte Carlo simulation method is used to quantify the uncertainty of prediction. The specific operation is as follows: (1) Select , , , , , For the key perturbation parameters of the model, a random perturbation of ±5% is introduced into each parameter within the range of 95%-105% of its calibration value; (2) Generate 100 sets of mutually independent perturbation parameter combinations by random sampling; (3) Substitute each set of parameters into the three-field coupled partial differential equation system in turn, and use the finite difference method to solve independently to obtain the corresponding remaining lifetime prediction value, forming 100 remaining lifetime sample values. (4) Sort the 100 remaining lifetime sample values ​​in ascending order, and take the 2.5th percentile value as the lower limit of remaining lifetime. The 97.5th percentile value is taken as the upper limit of the remaining lifespan. This constitutes a 95% confidence interval. .

[0031] Cumulative lifespan consumption calculation: Discretize time into tiny time steps. (Taking 0.1h), the lifespan attrition rate, cumulative lifespan attrition, and remaining lifespan percentage are calculated in real time using the following formula: Remaining lifetime prediction simulation verification: Typical operating conditions in the technical solution were set up to verify the prediction accuracy of the physical model. As shown in Table 2, the average relative error of the remaining lifetime prediction was only 6.2%, and the actual coverage of the 95% confidence interval reached 93.8%, which meets the high-precision prediction requirements of the technical solution.

[0032] Table 2 shows the simulation results for remaining lifetime prediction. S6. Health Management and Early Warning Output: Based on the initial values ​​of ESR and C and the failure threshold, a health index is calculated, and tiered early warnings are implemented with multiple methods for outputting warning information, as detailed below: Health Index (HI) Calculation: The health index is calculated based on the general failure threshold, corresponding to the ESR and C values. The health index ranges from 0 to 1, with values ​​closer to 1 indicating better health. Electrolyte volatilization is the primary factor: ESR-related formulas are used. ,in (The failure threshold specified in the technical solution); Anodic foil corrosion is the dominant factor: using the C-value related formula, ,in (The failure threshold specified in the technical solution).

[0033] Tiered early warning implementation: Two levels of early warning are set based on health index and confidence interval width, and targeted operation and maintenance suggestions are provided: Level 1 warning: Triggered when 0.1≤HI≤0.3. The technical solution recommends replacing the electrolytic capacitor in the near future and increasing the monitoring frequency to twice the original frequency. Level 2 warning: Triggered when HI < 0.1. The technical solution stipulates that the electrolytic capacitor must be replaced immediately to prevent the capacitor failure from causing equipment failure.

[0034] Early warning information output: Early warning information and health status data (remaining life, health index, aging mode, hot spot temperature, etc.) are output using one or more of the following methods: HMI display, CAN communication upload, and relay contact alarm, to ensure that maintenance personnel can perceive the situation in a timely manner and provide a basis for decision-making for the intelligent operation and maintenance of power electronic equipment.

[0035] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for calculating the lifetime of an electrolytic capacitor based on ripple analysis, characterized in that, Includes the following steps: S1. Construct a three-field coupled partial differential equation system: Establish a three-field coupled partial differential equation system of electro-thermal-concentration for electrolytic capacitors to characterize the dynamic evolution of ESR and C with internal hot spot temperature, electrolyte concentration, and ripple stress; complete the calibration of the physical parameters of the model in the equation system by combining accelerated aging experiments with nonlinear least squares method, and obtain the effective calibration parameters after model verification. S2. Real-time ripple data acquisition and preprocessing: By acquiring the ripple current, ripple voltage, shell temperature and ambient temperature data of the electrolytic capacitor in real time, the acquired data is sequentially processed to remove DC components, filter and denoise, and synchronize and align. S3. Online identification of key electrical parameters: For ESR, the frequency domain method is used and for C value, the ripple voltage charge-discharge slope method is used to perform targeted analysis on the preprocessed ripple data, so as to realize the real-time online identification of ESR and C value. S4. Calculation of dynamic heat loss and internal hot spot temperature: The real-time heat loss of the capacitor is calculated based on the identified ESR and RMS value of ripple current. The internal hot spot temperature of the capacitor is calculated by combining the thermal field equation in the three-field coupled partial differential equation system and the measured shell temperature. S5. Remaining Life Prediction and Confidence Interval Output: Substitute the currently identified electrical parameters, internal hotspot temperatures, and the model physical parameters calibrated by accelerated aging experiments obtained in S1 into a three-field coupled partial differential equation system. Use numerical solution methods and simulate the evolution of ESR and C values. Combine this with a general failure threshold to obtain the remaining life prediction value. The Monte Carlo simulation method was used to introduce random perturbations into the key parameters of the model, and multiple sets of remaining lifetime sample values ​​were obtained. The lower limit of the remaining lifetime was determined through statistical analysis. and upper limit value Simultaneously calculate the lifespan attrition rate, cumulative lifespan attrition, and remaining lifespan percentage; S6. Health Management and Early Warning Output: Calculate the corresponding health index based on the initial values ​​of ESR and C values ​​and the general failure threshold, and take the minimum of the two as the final health index. ,according to The system implements tiered early warnings and outputs these warnings.

2. The method for calculating the lifetime of an electrolytic capacitor based on ripple analysis according to claim 1, characterized in that, The three-field coupled partial differential equation set in step S1 includes the electric field dynamic equation, the thermal field dynamic equation, and the concentration field diffusion equation. The dynamic equation of the electric field is: in, These are the material-related constants for ESR evolution; The activation energy for ESR evolution; It is the gas constant; This refers to the temperature of the hot spot inside the capacitor. Ripple current density; This represents the initial concentration of the electrolyte. Real-time electrolyte concentration; ESR is an empirical index of ESR evolution. The C-value is a material-related constant for evolution. The activation energy for C-value evolution; This represents the real-time RMS value of the ripple current. This is the rated ripple current of the capacitor; The empirical index for the evolution of C-value; The dynamic equation for the thermal field is: in, The dielectric density of the capacitor; The specific heat capacity of the capacitor dielectric at constant pressure; The rate of change of temperature over time; For Hamiltonian operators; The thermal conductivity of the capacitor dielectric; For heat source terms per unit volume; The concentration field diffusion equation is: in, The rate of change of electrolyte concentration over time; The diffusion coefficient is temperature-dependent. is the Laplace operator, representing the spatial second-order rate of change of concentration; Let $\frac{ ... The activation energy for electrolyte volatilization; The minimum effective electrolyte concentration; It is a nonnegative operator, characterized only if If the difference is true, take the difference; otherwise, take 0.

3. The method for calculating the lifetime of an electrolytic capacitor based on ripple analysis according to claim 1, characterized in that, In step S2, the ripple current is acquired through a shunt; the ripple voltage is acquired through a voltage divider circuit and is sampled synchronously with the ripple current; the casing temperature is acquired through a surface-mount NTC thermistor; and the ambient temperature is acquired through an ambient temperature sensor. The data preprocessing includes removing DC components, filtering and noise reduction, and synchronous alignment.

4. The method for calculating the lifetime of an electrolytic capacitor based on ripple analysis according to claim 1, characterized in that, The specific process of identifying ESR using the frequency domain method in step S3 is as follows: Perform a Fast Fourier Transform (FFT) on the ripple current and ripple voltage to calculate the switching frequency and its harmonics. The impedance calculation formula is: ,in, For frequency The impedance value below; The frequency domain amplitude of the ripple voltage; The frequency domain amplitude of the ripple current is used. After separating the effects of capacitive and inductive reactance, the ESR value at each frequency point is extracted. The arithmetic mean method is used to fuse the results of multiple frequency points to obtain the final ESR value.

5. The method for calculating the lifetime of an electrolytic capacitor based on ripple analysis according to claim 1, characterized in that, The specific process of identifying the C value using the ripple voltage charge / discharge slope method in step S3 is as follows: Based on the preprocessed ripple voltage time-domain waveform, the charging and discharging phases within the ripple period are identified. The time period of voltage rise during the charging phase is selected, and the slope of voltage change with time is solved by linear fitting. Simultaneously calculate the average ripple current during this time period. Substitute the values ​​into the formula to calculate the value of C. The formula is: Where C is the capacitance value; This represents the average value of the ripple current during the charging and discharging phases. This represents the slope of the ripple voltage during the charging and discharging phase.

6. The method for calculating the lifetime of an electrolytic capacitor based on ripple analysis according to claim 1, characterized in that, The formula for calculating real-time heat loss in step S4 is as follows: ,in, This refers to the real-time heat loss of the capacitor. RMS value of ripple current; ESR 实时 For real-time equivalent series resistance; The calculation model for the internal hotspot temperature includes a steady-state simplified model and a transient accurate model. The steady-state simplified model is as follows: ,in, This refers to the measured temperature of the capacitor casing. The thermal resistance from the internal hot spot of the capacitor to the outer casing; Transient exact model is ,in, for Internal hotspot temperature at any given moment; for The outer casing temperature at any given time; for Heat loss over time; for The transient thermal impedance function at time t.

7. The method for calculating the lifetime of an electrolytic capacitor based on ripple analysis according to claim 1, characterized in that, In step S5, the numerical solution method is the finite difference method. The time and spatial domains of the three-field coupled partial differential equations are discretized. The time step is adaptively adjusted, ranging from 0.1h to 1h, and the spatial step is set according to the capacitor structure. This transforms the partial differential equations into a linear system of equations. The evolution curves of ESR and C values ​​over time are obtained through iterative solutions. The general failure threshold is: under the electrolyte volatilization-dominated mode, the failure threshold is... , The initial ESR value; under the dominant anodic foil corrosion mode, the failure threshold is... , The initial value of C is used; the electrolytic capacitor is determined to be in failure if any failure threshold is met; the remaining lifetime prediction value is the iteration time from the current moment to when the ESR or C value reaches the failure threshold.

8. The method for calculating the lifetime of an electrolytic capacitor based on ripple analysis according to claim 1, characterized in that, The Monte Carlo simulation method obtained in step S5 and The specific process is as follows: (1) Select , , , , , For the key perturbation parameters of the model, a random perturbation of ±5% is introduced into each parameter within the range of 95%-105% of its calibration value; (2) Generate 100 sets of mutually independent perturbation parameter combinations by random sampling; (3) Substitute each set of parameters into the three-field coupled partial differential equation system in turn, and use the finite difference method to solve independently to obtain the corresponding remaining lifetime prediction value, forming 100 remaining lifetime sample values. (4) Sort the 100 remaining lifetime sample values ​​in ascending order, and take the 2.5th percentile value as the lower limit of remaining lifetime. The 97.5th percentile value is taken as the upper limit of the remaining lifespan. ; The formula for the lifespan attrition rate is: ,in, For discrete, small time steps; This refers to the single-step long lifespan consumption rate; The formula for cumulative lifespan consumption is: ,in, This refers to the cumulative lifespan attrition rate. The formula for the percentage of remaining lifespan is: ,in, This represents the percentage of remaining lifespan.

9. The method for calculating the lifetime of an electrolytic capacitor based on ripple analysis according to claim 1, characterized in that, The health index in step S6 Based on the aging mode, it was determined that under the electrolyte volatilization-dominated mode... ; Under the dominant mode of anodic foil corrosion, ; in, As a health index; The ESR failure threshold; For real-time equivalent series resistance; This is the initial value for ESR; The C-value is the failure threshold; This is the real-time C value; The initial value for C; The graded early warning is as follows: when 0.1≤HI≤0.3, a first-level early warning is issued, suggesting that the electrolytic capacitor be replaced in the near future; when HI<0.1, a second-level early warning is issued, requiring the electrolytic capacitor to be replaced immediately; the early warning output method includes one or more of HMI display, CAN communication upload, and relay contact alarm.

Images