Super capacitor soh estimation and life prediction method and system based on power adaptive segmentation

Abstract

This invention discloses a method and system for estimating the state of harm (SOH) and predicting the lifespan of supercapacitors based on power adaptive segmentation. The method includes: real-time acquisition of the supercapacitor's terminal voltage, charging and discharging current, and internal temperature; calculation of instantaneous power and power change rate; and dynamic adjustment of the slice length based on power fluctuation. At the end of each slice, the accumulated lifespan loss is calculated based on a three-dimensional nonlinear aging acceleration factor model of voltage, temperature, and time to obtain the current SOH value. During the no-load window period, the actual ESR and capacity are measured using test pulses to obtain the baseline measured SOH value and dynamically correct the model calibration coefficients. The remaining lifespan is predicted based on the difference between the current SOH and the scrap threshold, as well as the average aging rate. Based on a three-dimensional nonlinear aging model combined with an online self-calibration mechanism, the method solves the problem of model parameter drift during long-term operation. The power adaptive segmentation strategy ensures accuracy while achieving optimal allocation of computing resources, enabling online estimation and prediction of the entire lifespan without shutdown or disassembly.

Description

Technical Field

[0001] This invention belongs to the field of supercapacitor condition monitoring and health management technology, and relates to a method and system for supercapacitor SOH estimation and lifetime prediction based on power adaptive segmentation. Background Technology

[0002] Supercapacitors, with their high power density, long cycle life, and wide temperature range, have been widely used in key fields such as auxiliary power supplies for new energy vehicles, wind power pitch control systems, renewable energy recovery in rail transit, thermal power and energy storage frequency regulation in power systems, and aerospace. In actual operation, the State of Health (SOH) of a supercapacitor decays nonlinearly due to the combined effects of multiple factors, including voltage stress, ambient temperature, operating time, and power conditions. Accurately estimating SOH and predicting remaining life is of significant engineering importance for ensuring safe equipment operation, developing predictive maintenance strategies, and reducing operation and maintenance costs.

[0003] Existing technical solutions all have their own obvious shortcomings. The offline testing method (CN209296847U) requires shutdown and disassembly, making it impossible to achieve online real-time estimation; the fixed-time segmented accumulation method (CN110850345A) calculates aging loss by slicing at fixed time intervals, resulting in wasted computing power when power is stable, and decreased accuracy due to insufficient sampling granularity when power fluctuates drastically, making it unable to adapt to dynamic changes in operating conditions; the single-factor aging model (CN113253126A) relies only on temperature or voltage for linear fitting, ignoring the coupling effect of multiple factors, and the estimation error often reaches more than 15%; the online method without self-calibration (CN114579218A) relies only on the initial model for cumulative calculation, and after long-term operation, model parameter drift causes the error to accumulate to more than 20%, rendering it useless; in addition, most solutions are designed only for single scenarios such as automobiles or energy storage (CN217820826U), lacking versatility and difficult to migrate to other application fields.

[0004] In summary, there is currently a lack of a supercapacitor SOH estimation and remaining lifetime prediction method that can simultaneously achieve online real-time performance, adaptive segmentation of operating conditions, multi-factor nonlinear modeling, dynamic self-calibration, and versatility across multiple scenarios. Summary of the Invention

[0005] The purpose of this invention is to solve the problems in the prior art and provide a method and system for estimating the state of harm (SOH) and predicting the lifetime of a supercapacitor based on power adaptive segmentation.

[0006] To achieve the above objectives, the present invention employs the following technical solution:

[0007] A method for estimating the state of harm (SOH) and predicting the lifetime of supercapacitors based on power adaptive piecewise segmentation includes the following steps:

[0008] S1 collects the terminal voltage, charging and discharging current and internal temperature of the supercapacitor module in real time, and calculates the instantaneous power and power change rate at fixed time intervals. Based on the ratio of the power change rate to the rated power of the system, the operating conditions are divided into low fluctuation range, medium fluctuation range and high fluctuation range, and the slice length corresponding to each fluctuation range is dynamically determined. Among them, the low fluctuation range corresponds to the long slice, the medium fluctuation range corresponds to the medium slice, and the high fluctuation range corresponds to the short slice.

[0009] S2, at the end of each slice, record the average voltage, average temperature and duration within the slice. Based on the aging mechanism of supercapacitor electrolyte decomposition and electrode active material carbonization, establish a three-dimensional nonlinear aging acceleration factor model that simultaneously includes voltage stress, thermal stress and time stress, calculate the percentage of life loss within the slice, and accumulate the life loss of all slices to obtain the current health state SOH value.

[0010] S3: When the system enters the no-load standby state and the no-load current and no-load duration both meet the preset conditions, the online self-calibration process is started. During the no-load window period, a test pulse is applied and the voltage response is measured. The true ESR value and the true capacity value are calculated to obtain the reference SOH measured value. The current SOH estimated value accumulated in S2 is compared with the reference SOH measured value. If the deviation exceeds the preset threshold, the calibration coefficient in the three-dimensional nonlinear aging acceleration factor model is automatically corrected.

[0011] S4. Based on the difference between the current SOH value and the preset scrap threshold, combined with the average aging rate within the preset time period, calculate the remaining service life (RUL) and output the current SOH estimate and RUL prediction results.

[0012] The dynamic determination of the slice length corresponding to each fluctuation interval in S1 is specifically as follows:

[0013] When the ratio of the power change rate to the system rated power is less than the first preset threshold, it is determined to be a low fluctuation range, and the slice length is set to 10 to 60 seconds.

[0014] When the ratio is greater than or equal to the first preset threshold and less than the second preset threshold, it is determined to be in the medium fluctuation range, and the slice length is set to 2 to 10 seconds.

[0015] When the ratio is greater than or equal to the second preset threshold, it is determined to be a high fluctuation range, and the slice length is set to 0.5 seconds to 2 seconds;

[0016] The first preset threshold is less than the second preset threshold, and the slice length is adaptively shortened as power fluctuations intensify, in order to balance the consumption of computing resources and the accuracy of capturing transient aging loss.

[0017] The instantaneous power calculation cycle in S2 is once every 100 milliseconds; at the end of each slice, the average voltage, average temperature and duration are stored in the ring buffer of the main control unit for subsequent aging loss calculation module to call.

[0018] The three-dimensional nonlinear aging acceleration factor model calculates the aging acceleration factor using the following formula:

[0019]

[0020] in, These are the system calibration coefficients, configured according to the electrode material system of the supercapacitor at the time of manufacture; The aging activation energy is determined based on the electrode material system; R is the gas constant; T is the average absolute temperature within the current slice. The reference temperature is U; U is the average voltage within the current slice. α is the reference voltage; α is the voltage acceleration factor.

[0021] The percentage of lifetime loss for a single slice is calculated using the following formula:

[0022]

[0023] Among them, t slice The duration of the current slice is dynamically determined by the power adaptive segmentation logic; t rated Rated life under reference conditions; AF(T) avg U avg ) represents the aging acceleration factor calculated based on the current average temperature and average voltage of the slice; T avg U represents the average absolute temperature within the current slice. avg The average voltage within the current slice; k is the dynamic calibration coefficient, initially set to 1.0, corrected by the no-load online self-calibration process; the current health status (SOH) value is calculated cumulatively using the following formula:

[0024]

[0025] Where n is the total number of slices since it was put into operation.

[0026] The preset condition for the no-load current in S3 is: the current is less than or equal to a threshold value in the range of 0.02C to 0.1C, where C is the rated capacity of the supercapacitor; the preset condition for the no-load duration is: the duration is greater than or equal to a threshold value in the range of 15 minutes to 60 minutes.

[0027] The online self-calibration process in S3 specifically includes the following steps:

[0028] S3.1 After the no-load condition is met, a constant current pulse with an amplitude of 0.1C to 0.5C is applied to the supercapacitor, the voltage difference before and after the pulse is applied is measured, and the true ESR value is calculated according to Ohm's law.

[0029] S3.2, charge and discharge the supercapacitor with a small current for a preset time, and calculate the actual capacity value by combining the relationship curve between open circuit voltage and state of charge OCV-SOC.

[0030] S3.3 Calculate the measured baseline SOH value based on the actual ESR value and the actual capacity value. The baseline SOH is calculated as a weighted sum of the capacity SOH and the internal resistance SOH, where the weights can be configured according to the application scenario.

[0031] S3.4 If the relative error between the current SOH estimate and the baseline measured SOH exceeds 5%, the dynamic calibration coefficient in the three-dimensional nonlinear aging acceleration factor model is corrected so that the corrected SOH estimate approaches the baseline measured SOH.

[0032] The remaining useful life (RUL) is calculated using the following formula:

[0033]

[0034] Among them, SOH current This is an estimated value of the current health status, output by the SOH cumulative calculation module; SOH threshold The preset scrapping threshold can be configured within the range of 50% to 70% according to different application scenarios; aging The average aging rate over the past 30 days is expressed as a percentage per day, reflecting the impact of recent actual operating conditions on the aging rate.

[0035] The method also includes a multi-level early warning mechanism: a maintenance reminder signal is issued when the RUL is less than or equal to 180 days, and a replacement reminder signal is issued when the RUL is less than or equal to 90 days.

[0036] The algorithm parameters of the method are configurable, including: slice length range, power fluctuation interval division threshold, system calibration coefficient in the three-dimensional nonlinear aging acceleration factor model, aging activation energy and voltage acceleration coefficient, no-load trigger condition threshold, benchmark SOH calculation weight, and scrapping threshold. All of these can be configured according to different application scenarios, and can be adapted to any supercapacitor application scenario in new energy vehicles, rail transit, renewable energy storage, power system frequency regulation, port machinery, and aerospace without modifying the algorithm framework.

[0037] A supercapacitor SOH estimation and lifetime prediction system based on power adaptive segmentation includes:

[0038] The voltage sampling unit is connected to the positive and negative terminals of the supercapacitor module and is used to acquire the terminal voltage signal in real time.

[0039] The current sampling unit is connected in series in the charging and discharging circuit of the supercapacitor module to collect charging and discharging current signals in real time.

[0040] Temperature sampling unit, located inside or on the surface of the supercapacitor module, is used to collect temperature signals in real time;

[0041] The main control unit, an industrial-grade MCU or processor, is connected to the voltage sampling unit, current sampling unit, and temperature sampling unit, respectively; the main control unit integrates the following:

[0042] The power calculation and adaptive segmentation module is used to calculate instantaneous power and power change rate based on real-time collected voltage and current, and dynamically adjust the slice length according to the degree of power fluctuation.

[0043] The three-dimensional nonlinear aging acceleration factor calculation module is used to calculate the three-dimensional nonlinear aging acceleration factor and single-slice lifetime loss based on the average voltage, average temperature and duration in each slice.

[0044] The SOH cumulative calculation module is used to accumulate the lifetime loss of all slices to obtain the current SOH estimate.

[0045] The no-load online self-calibration module is used to measure the actual ESR and actual capacity during the system's no-load window period, obtain the measured value of the reference SOH, and dynamically correct the model calibration coefficients.

[0046] The Remaining Service Life (RUL) prediction module is used to calculate RUL based on the current SOH value, the preset scrap threshold, and the average aging rate, and output multi-level early warning signals.

[0047] The output unit is used to output the current SOH estimation result, RUL prediction result, and early warning signal.

[0048] Compared with the prior art, the present invention has the following beneficial effects:

[0049] This invention establishes a three-dimensional nonlinear acceleration factor model integrating voltage stress, thermal stress, and time stress based on the real aging mechanism of supercapacitor electrolyte decomposition and electrode active material carbonization. Combined with an online self-calibration mechanism during the no-load window period, the dynamic calibration coefficients are continuously corrected, ensuring stable control of SOH estimation error throughout the entire lifecycle. This solves the problem of estimation distortion caused by model parameter drift during long-term operation. This invention also proposes a power adaptive segmentation strategy for the first time, dynamically adjusting the slice length according to the real-time power fluctuation level. Longer slices indicate more stable power, significantly reducing redundant calculations; shorter slices indicate more severe power fluctuations, accurately capturing transient aging characteristics. This achieves an optimal balance between computing resources and estimation accuracy while ensuring precision. Furthermore, all calculations and calibrations are completed during continuous equipment operation, without the need for shutdown or disassembly.

[0050] This invention not only outputs the current State of Health (SOH) value but also predicts the remaining service life based on the average aging rate over the past 30 days. It also features a two-tiered early warning mechanism: a maintenance reminder for RUL ≤ 180 days and a replacement reminder for RUL ≤ 90 days. This allows maintenance personnel to plan spare parts procurement and replacement in advance, transforming unexpected downtime into planned maintenance and significantly reducing maintenance costs. The invention employs a configurable parameter architecture, allowing all key parameters to be flexibly set according to different application scenarios. It can be adapted to any supercapacitor application scenario, including new energy vehicles, rail transit, renewable energy storage, power system frequency regulation, port machinery, and aerospace, without modifying the algorithm framework. Furthermore, it fully reuses existing voltage, current, and temperature sampling hardware, and all innovative functions are implemented through software algorithm modules, resulting in zero hardware cost increase and demonstrating high engineering feasibility and industrialization potential. Attached Figure Description

[0051] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0052] Figure 1 This is a flowchart of the supercapacitor SOH estimation and lifetime prediction method based on power adaptive segmentation according to the present invention;

[0053] Figure 2 This is a structural diagram of the supercapacitor SOH estimation and lifetime prediction system based on power adaptive segmentation according to the present invention. Detailed Implementation

[0054] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0055] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0056] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0057] The present invention will now be described in further detail with reference to the accompanying drawings:

[0058] See Figure 1 The flowchart below shows the supercapacitor SOH estimation and lifetime prediction method based on power adaptive segmentation in this invention, which includes the following steps:

[0059] S1 dynamically divides operating conditions into intervals based on real-time power fluctuations and adaptively determines slice lengths. It collects real-time data on the supercapacitor module's terminal voltage, charging / discharging current, and internal temperature, and calculates instantaneous power and power change rate at fixed time intervals. Based on the ratio of the power change rate to the system's rated power, the operating conditions are divided into low-fluctuation, medium-fluctuation, and high-fluctuation intervals, and the slice length corresponding to each fluctuation interval is dynamically determined. Specifically, low-fluctuation intervals correspond to long slices, medium-fluctuation intervals to medium slices, and high-fluctuation intervals to short slices.

[0060] The slice length corresponding to each fluctuation range is dynamically determined, specifically as follows:

[0061] When the ratio of the power change rate to the system rated power is less than the first preset threshold, it is determined to be a low fluctuation range, and the slice length is set to 10 to 60 seconds.

[0062] When the ratio is greater than or equal to the first preset threshold and less than the second preset threshold, it is determined to be in the medium fluctuation range, and the slice length is set to 2 to 10 seconds.

[0063] When the ratio is greater than or equal to the second preset threshold, it is determined to be a high fluctuation range, and the slice length is set to 0.5 seconds to 2 seconds;

[0064] The first preset threshold is less than the second preset threshold, and the slice length is adaptively shortened as power fluctuations intensify, in order to balance the consumption of computing resources and the accuracy of capturing transient aging loss.

[0065] S2, based on a three-dimensional nonlinear aging model of voltage, temperature, and time, calculates the lifetime loss of each slice and accumulates it to obtain the current SOH value. At the end of each slice, the average voltage, average temperature, and duration within the slice are recorded. Based on the aging mechanism of supercapacitor electrolyte decomposition and electrode active material carbonization, a three-dimensional nonlinear aging acceleration factor model that simultaneously includes voltage stress, thermal stress, and time stress is established to calculate the percentage of lifetime loss within the slice. The lifetime loss of all slices is then accumulated to obtain the current health state SOH value.

[0066] The instantaneous power is calculated every 100 milliseconds. At the end of each slice, the average voltage, average temperature and duration are stored in the ring buffer of the main control unit for subsequent aging loss calculation module to call.

[0067] The three-dimensional nonlinear aging acceleration factor model calculates the aging acceleration factor using the following formula:

[0068]

[0069] in, These are the system calibration coefficients, configured according to the electrode material system of the supercapacitor at the time of manufacture; The aging activation energy is determined based on the electrode material system; R is the gas constant; and T is the average absolute temperature within the current slice. The reference temperature is U; U is the average voltage within the current slice. α is the reference voltage; α is the voltage acceleration factor.

[0070] The percentage of lifetime loss for a single slice is calculated using the following formula:

[0071]

[0072] Among them, t slice The duration of the current slice is dynamically determined by the power adaptive segmentation logic; t rated Rated life under reference conditions; AF(T) avg U avg ) represents the aging acceleration factor calculated based on the current average temperature and average voltage of the slice; T avg U represents the average absolute temperature within the current slice. avg The average voltage within the current slice; k is the dynamic calibration coefficient, initially set to 1.0, corrected by the no-load online self-calibration process; the current health status (SOH) value is calculated cumulatively using the following formula:

[0073]

[0074] Where n is the total number of slices since it was put into operation.

[0075] S3: During the no-load window period, the actual ESR and capacitance are measured online using test pulses to obtain the reference SOH, and the model calibration coefficients are dynamically corrected. When the system enters the no-load standby state and both the no-load current and no-load duration meet the preset conditions, the online self-calibration process is initiated. During the no-load window period, test pulses are applied and the voltage response is measured to calculate the actual ESR value and the actual capacitance value, thereby obtaining the measured reference SOH value. The current SOH estimate accumulated in S2 is compared with the measured reference SOH value. If the deviation exceeds the preset threshold, the calibration coefficients in the three-dimensional nonlinear aging acceleration factor model are automatically corrected. The preset condition for the no-load current is: the current is less than or equal to a threshold in the range of 0.02C to 0.1C, where C is the rated capacitance of the supercapacitor. The preset condition for the no-load duration is: the duration is greater than or equal to a threshold in the range of 15 minutes to 60 minutes.

[0076] The online self-calibration process specifically includes the following steps:

[0077] S3.1 After the no-load condition is met, a constant current pulse with an amplitude of 0.1C to 0.5C is applied to the supercapacitor, the voltage difference before and after the pulse is applied is measured, and the true ESR value is calculated according to Ohm's law.

[0078] S3.2, charge and discharge the supercapacitor with a small current for a preset time, and calculate the actual capacity value by combining the relationship curve between open circuit voltage and state of charge OCV-SOC.

[0079] S3.3 Calculate the measured baseline SOH value based on the actual ESR value and the actual capacity value. The baseline SOH is calculated as a weighted sum of the capacity SOH and the internal resistance SOH, where the weights can be configured according to the application scenario.

[0080] S3.4 If the relative error between the current SOH estimate and the baseline measured SOH exceeds 5%, the dynamic calibration coefficient in the three-dimensional nonlinear aging acceleration factor model is corrected so that the corrected SOH estimate approaches the baseline measured SOH.

[0081] S4 calculates the remaining useful life (RUL) based on the current SOH, the preset scrap threshold, and the recent average aging rate, and outputs the prediction result. Based on the difference between the current SOH value and the preset scrap threshold, combined with the average aging rate within a preset time period, it calculates the remaining useful life (RUL) and outputs the current SOH estimate and the RUL prediction result.

[0082] The remaining useful life (RUL) is calculated using the following formula:

[0083]

[0084] Among them, SOH current This is an estimated value of the current health status, output by the SOH cumulative calculation module; SOH threshold The preset scrapping threshold can be configured within the range of 50% to 70% according to different application scenarios; aging The average aging rate over the past 30 days is expressed as a percentage per day, reflecting the impact of recent actual operating conditions on the aging rate.

[0085] The method also includes a multi-level early warning mechanism: a maintenance reminder signal is issued when the RUL is less than or equal to 180 days, and a replacement reminder signal is issued when the RUL is less than or equal to 90 days.

[0086] The algorithm parameters are configurable, including: slice length range, power fluctuation interval division threshold, system calibration coefficient in the three-dimensional nonlinear aging acceleration factor model, aging activation energy and voltage acceleration coefficient, no-load trigger condition threshold, benchmark SOH calculation weight, and scrapping threshold. All of these can be configured according to different application scenarios, and can be adapted to any supercapacitor application scenario in new energy vehicles, rail transit, renewable energy storage, power system frequency regulation, port machinery, and aerospace without modifying the algorithm framework.

[0087] like Figure 2 The diagram shown is a structural diagram of the supercapacitor SOH estimation and lifetime prediction system based on power adaptive segmentation in this invention, including:

[0088] The voltage sampling unit is connected to the positive and negative terminals of the supercapacitor module and is used to acquire the terminal voltage signal in real time.

[0089] The current sampling unit is connected in series in the charging and discharging circuit of the supercapacitor module to collect charging and discharging current signals in real time.

[0090] Temperature sampling unit, located inside or on the surface of the supercapacitor module, is used to collect temperature signals in real time;

[0091] The main control unit, an industrial-grade MCU or processor, is connected to the voltage sampling unit, current sampling unit, and temperature sampling unit, respectively; the main control unit integrates the following:

[0092] The power calculation and adaptive segmentation module is used to calculate instantaneous power and power change rate based on real-time collected voltage and current, and dynamically adjust the slice length according to the degree of power fluctuation.

[0093] The three-dimensional nonlinear aging acceleration factor calculation module is used to calculate the three-dimensional nonlinear aging acceleration factor and single-slice lifetime loss based on the average voltage, average temperature and duration in each slice.

[0094] The SOH cumulative calculation module is used to accumulate the lifetime loss of all slices to obtain the current SOH estimate.

[0095] The no-load online self-calibration module is used to measure the actual ESR and actual capacity during the system's no-load window period, obtain the measured value of the reference SOH, and dynamically correct the model calibration coefficients.

[0096] The Remaining Service Life (RUL) prediction module is used to calculate RUL based on the current SOH value, the preset scrap threshold, and the average aging rate, and output multi-level early warning signals.

[0097] The output unit is used to output the current SOH estimation result, RUL prediction result, and early warning signal.

[0098] Example

[0099] This embodiment provides a method and system for estimating the state of harm (SOH) and predicting the lifetime of a supercapacitor based on power adaptive segmentation.

[0100] The overall system framework is as follows: Supercapacitor energy storage modules (arbitrary series-parallel structure) are connected to voltage sampling units, current sampling units, and temperature sampling units, respectively. These sampling units reuse existing system hardware. The output of each sampling unit is connected to the main control unit, which can be any industrial-grade MCU or processor. The main control unit integrates a power calculation and adaptive segmentation module, a three-dimensional nonlinear aging acceleration factor calculation module, a SOH accumulation calculation module, an online self-calibration module under no-load conditions, and a remaining useful life (RUL) prediction module.

[0101] The specific steps of the method described in this embodiment are as follows:

[0102] Step 1: Power-adaptive segmented slicing. Fixed-time slicing is no longer used; instead, the slice length is dynamically adjusted based on the degree of power fluctuation, balancing accuracy and computing power.

[0103] (1) Parameter calculation.

[0104] Instantaneous power calculation, performed every 100ms:

[0105]

[0106] in, Let be the terminal voltage at time t. Let t be the charging and discharging current at time t.

[0107] Calculation of power change rate:

[0108]

[0109] in, This is the system's rated power.

[0110] (2) Dynamic slicing rules.

[0111] The power fluctuation range, slice length range, value selection principle, and applicable operating conditions are as follows:

[0112] when When the power is in a low-fluctuation range, the slice length is 10s to 60s. The principle behind this is that when the power is stable, voltage / temperature fluctuations are much smaller than the resolution of the aging model, so long slices do not affect accuracy and significantly reduce computational load. Suitable operating conditions include steady-state charging and discharging of wind, solar, and energy storage systems, and UPS standby.

[0113] when When the fluctuation range is moderate, the slice length ranges from 2s to 10s. The principle behind this value selection is to balance accuracy and computing power during moderate fluctuations. It is suitable for AGVs operating at constant speed and vehicles idling.

[0114] when When the power fluctuation range is high, the slice length ranges from 0.5s to 2s. The principle behind this is that when power fluctuates drastically, a short slice captures transient aging losses and avoids errors. Suitable operating conditions include vehicle starting, regenerative braking, and UPS transient discharge.

[0115] (3) Slice data recording.

[0116] After each slice is completed, the average voltage of that slice is recorded. Average temperature Duration Stored in the MCU's circular buffer.

[0117] Step 2: Three-dimensional nonlinear aging loss calculation. Based on the real aging mechanism of supercapacitors, namely "electrolyte decomposition and electrode active material carbonization", a general three-dimensional aging acceleration factor model is established to accurately quantify the lifespan loss of each slice.

[0118] (1) Formula for aging acceleration factor:

[0119]

[0120] Parameter description: These are the system calibration coefficients, configured according to the electrode material system of the supercapacitor at the time of manufacture; The aging activation energy is determined based on the electrode material system; R is the gas constant; T is the average absolute temperature within the current slice. The reference temperature is U; U is the average voltage within the current slice. α is the reference voltage; α is the voltage acceleration factor.

[0121] (2) Calculation of single-slice lifetime loss:

[0122]

[0123] Among them, t slice The duration of the current slice is dynamically determined by the power adaptive segmentation logic; t rated Rated life under reference conditions; AF(T) avg U avg ) represents the aging acceleration factor calculated based on the current average temperature and average voltage of the slice; T avg U represents the average absolute temperature within the current slice. avg is the average voltage within the current slice; k is the dynamic calibration coefficient, initially set to 1.0, and corrected by the unloaded online self-calibration process.

[0124] (3) Current cumulative SOH calculation:

[0125]

[0126] Where n is the total number of slices since the system was put into operation. The SOH lower limit is the industry-standard scrapping threshold (configurable from 50% to 70%).

[0127] Step 3: Online self-calibration during the idle window. This addresses model drift issues after long-term operation, ensuring a total lifecycle error of ≤3%.

[0128] (1) Triggering conditions.

[0129] The no-load current threshold ranges from 0.02C to 0.1C (C is the rated capacity). The principle for this value is that when the current is ≤0.05C, the polarization voltage is ≤1mV, which has no impact on OCV measurement. The no-load duration ranges from 15 minutes to 60 minutes. The principle for this value is that the polarization recovery time constant of the supercapacitor is 1 to 10 minutes, with 30 minutes being a general redundancy value to ensure complete polarization recovery.

[0130] (2) Calibration steps.

[0131] Step 1: After the system meets the no-load conditions, output a constant current pulse of 0.1C~0.5C, measure the voltage difference before and after the pulse, and calculate the actual ESR according to formula (6):

[0132]

[0133] Step 2: Charge and discharge with a small current for 10 minutes, and calculate the actual capacity using the OCV-SOC curve. .

[0134] Step 3: Calculate the baseline SOH:

[0135]

[0136] The weights can be configured according to the scenario.

[0137] Step 4: If the error between the estimated value and the true value is >5%, automatically correct the calibration coefficient. To correct the error.

[0138] Step 4: Remaining useful life (RUL) prediction.

[0139] RUL calculation:

[0140]

[0141] Among them, SOH current This is an estimated value of the current health status, output by the SOH cumulative calculation module; SOH threshold The preset scrapping threshold can be configured within the range of 50% to 70% according to different application scenarios; aging The average aging rate over the past 30 days is expressed as a percentage per day, reflecting the impact of recent actual operating conditions on the aging rate.

[0142] This embodiment supports multi-level warnings: maintenance reminder when RUL ≤ 180 days, and replacement reminder when RUL ≤ 90 days.

[0143] The complete workflow is as follows:

[0144] The system collects voltage, current, and temperature data in real time, calculates instantaneous power and power change rate, dynamically adjusts slice length, calculates three-dimensional aging acceleration factor and loss for each slice, accumulates loss to obtain current SOH, and determines whether the no-load condition is met. If it is met, online self-calibration and model parameter correction are performed; otherwise, the current SOH and remaining lifetime are directly output.

[0145] In this embodiment, the modules within the main control unit operate collaboratively according to the above-described process. The power calculation and adaptive segmentation module executes step 1; the three-dimensional nonlinear aging acceleration factor calculation module executes the aging acceleration factor calculation and single-slice lifetime loss calculation in step 2; the SOH accumulation calculation module executes the SOH accumulation calculation in step 2; the no-load online self-calibration module executes step 3; and the remaining lifetime prediction module executes step 4. All of the above modules are software algorithm modules running within the main control unit, reusing the existing voltage, current, and temperature sampling hardware of the system, eliminating the need for additional hardware.

[0146] This embodiment enables online real-time SOH estimation without downtime or disassembly; the error over the entire lifecycle is ≤3%; it is universally applicable across all scenarios, with one algorithm adapting to all overcapacity systems; it features lifelong self-calibration without drift; it can predict remaining lifespan and enable preventative maintenance; it is implemented purely in software with zero hardware cost; and it is adaptive in terms of computing power without consuming resources.

[0147] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for estimating the state of harm (SOH) and predicting the lifetime of a supercapacitor based on power adaptive segmentation, characterized in that, Includes the following steps: S1 collects the terminal voltage, charging and discharging current and internal temperature of the supercapacitor module in real time, and calculates the instantaneous power and power change rate at fixed time intervals. Based on the ratio of the power change rate to the rated power of the system, the operating conditions are divided into low fluctuation range, medium fluctuation range and high fluctuation range, and the slice length corresponding to each fluctuation range is dynamically determined. Among them, the low fluctuation range corresponds to the long slice, the medium fluctuation range corresponds to the medium slice, and the high fluctuation range corresponds to the short slice. S2, at the end of each slice, record the average voltage, average temperature and duration within the slice. Based on the aging mechanism of supercapacitor electrolyte decomposition and electrode active material carbonization, establish a three-dimensional nonlinear aging acceleration factor model that simultaneously includes voltage stress, thermal stress and time stress, calculate the percentage of life loss within the slice, and accumulate the life loss of all slices to obtain the current health state SOH value. S3: When the system enters the no-load standby state and the no-load current and no-load duration both meet the preset conditions, the online self-calibration process is started. During the no-load window period, a test pulse is applied and the voltage response is measured. The true ESR value and the true capacity value are calculated to obtain the reference SOH measured value. The current SOH estimated value accumulated in S2 is compared with the reference SOH measured value. If the deviation exceeds the preset threshold, the calibration coefficient in the three-dimensional nonlinear aging acceleration factor model is automatically corrected. S4. Based on the difference between the current SOH value and the preset scrap threshold, combined with the average aging rate within the preset time period, calculate the remaining service life (RUL) and output the current SOH estimate and RUL prediction results.

2. The method for estimating the state of harm (SOH) and predicting the lifetime of a supercapacitor based on power adaptive segmentation as described in claim 1, characterized in that, The dynamic determination of the slice length corresponding to each fluctuation interval in S1 is specifically as follows: When the ratio of the power change rate to the system rated power is less than the first preset threshold, it is determined to be a low fluctuation range, and the slice length is set to 10 to 60 seconds. When the ratio is greater than or equal to the first preset threshold and less than the second preset threshold, it is determined to be in the medium fluctuation range, and the slice length is set to 2 to 10 seconds. When the ratio is greater than or equal to the second preset threshold, it is determined to be a high fluctuation range, and the slice length is set to 0.5 seconds to 2 seconds; The first preset threshold is less than the second preset threshold, and the slice length is adaptively shortened as power fluctuations intensify, in order to balance the consumption of computing resources and the accuracy of capturing transient aging loss.

3. The method for estimating the state of harm (SOH) and predicting the lifetime of a supercapacitor based on power adaptive segmentation as described in claim 1, characterized in that, The instantaneous power calculation cycle in S2 is once every 100 milliseconds; at the end of each slice, the average voltage, average temperature and duration are stored in the ring buffer of the main control unit for subsequent aging loss calculation module to call.

4. The method for estimating the state of harm (SOH) and predicting the lifetime of a supercapacitor based on power adaptive segmentation as described in claim 1, characterized in that, The three-dimensional nonlinear aging acceleration factor model calculates the aging acceleration factor using the following formula: ； in, These are the system calibration coefficients, configured according to the electrode material system of the supercapacitor at the time of manufacture; The aging activation energy is determined based on the electrode material system; R is the gas constant; and T is the average absolute temperature within the current slice. The reference temperature is U; U is the average voltage within the current slice. α is the reference voltage; α is the voltage acceleration factor.

5. The method for estimating the state of harm (SOH) and predicting the lifetime of a supercapacitor based on power adaptive segmentation as described in claim 4, characterized in that, The percentage of lifetime loss for a single slice is calculated using the following formula: ； Among them, t slice The duration of the current slice is dynamically determined by the power adaptive segmentation logic; t rated Rated life under reference conditions; AF(T) avg U avg ) represents the aging acceleration factor calculated based on the current average temperature and average voltage of the slice; T avg U represents the average absolute temperature within the current slice. avg The average voltage within the current slice; k is the dynamic calibration coefficient, initially set to 1.0, corrected by the no-load online self-calibration process; the current health status (SOH) value is calculated cumulatively using the following formula: ； Where n is the total number of slices since it was put into operation.

6. The method for estimating the state of harm (SOH) and predicting the lifetime of a supercapacitor based on power adaptive segmentation as described in claim 1, characterized in that, The preset condition for the no-load current in S3 is: the current is less than or equal to a threshold value in the range of 0.02C to 0.1C, where C is the rated capacity of the supercapacitor; the preset condition for the no-load duration is: the duration is greater than or equal to a threshold value in the range of 15 minutes to 60 minutes.

7. The method for estimating the state of harm (SOH) and predicting the lifetime of a supercapacitor based on power adaptive segmentation as described in claim 1, characterized in that, The online self-calibration process in S3 specifically includes the following steps: S3.1 After the no-load condition is met, a constant current pulse with an amplitude of 0.1C to 0.5C is applied to the supercapacitor, the voltage difference before and after the pulse is applied is measured, and the true ESR value is calculated according to Ohm's law. S3.2, charge and discharge the supercapacitor with a small current for a preset time, and calculate the actual capacity value by combining the relationship curve between open circuit voltage and state of charge OCV-SOC. S3.3 Calculate the measured baseline SOH value based on the actual ESR value and the actual capacity value. The baseline SOH is calculated as a weighted sum of the capacity SOH and the internal resistance SOH, where the weights can be configured according to the application scenario. S3.4 If the relative error between the current SOH estimate and the baseline measured SOH exceeds 5%, the dynamic calibration coefficient in the three-dimensional nonlinear aging acceleration factor model is corrected so that the corrected SOH estimate approaches the baseline measured SOH.

8. The method for estimating the state of harm (SOH) and predicting the lifetime of a supercapacitor based on power adaptive segmentation as described in claim 1, characterized in that, The remaining useful life (RUL) is calculated using the following formula: ； Among them, SOH current This is an estimated value of the current health status, output by the SOH cumulative calculation module; SOH threshold The preset scrapping threshold can be configured within the range of 50% to 70% according to different application scenarios; aging The average aging rate over the past 30 days is expressed as a percentage per day, reflecting the impact of recent actual operating conditions on the aging rate. The method also includes a multi-level early warning mechanism: a maintenance reminder signal is issued when the RUL is less than or equal to 180 days, and a replacement reminder signal is issued when the RUL is less than or equal to 90 days.

9. The method for estimating the state of harm (SOH) and predicting the lifetime of a supercapacitor based on power adaptive segmentation as described in claim 1, characterized in that, The algorithm parameters of the method are configurable, including: slice length range, power fluctuation interval division threshold, system calibration coefficient in the three-dimensional nonlinear aging acceleration factor model, aging activation energy and voltage acceleration coefficient, no-load trigger condition threshold, benchmark SOH calculation weight, and scrapping threshold. All of these can be configured according to different application scenarios, and can be adapted to any supercapacitor application scenario in new energy vehicles, rail transit, renewable energy storage, power system frequency regulation, port machinery, and aerospace without modifying the algorithm framework.

10. A supercapacitor SOH estimation and lifetime prediction system based on power adaptive segmentation, characterized in that, include: The voltage sampling unit is connected to the positive and negative terminals of the supercapacitor module and is used to acquire the terminal voltage signal in real time. The current sampling unit is connected in series in the charging and discharging circuit of the supercapacitor module to collect charging and discharging current signals in real time. Temperature sampling unit, located inside or on the surface of the supercapacitor module, is used to collect temperature signals in real time; The main control unit, an industrial-grade MCU or processor, is connected to the voltage sampling unit, current sampling unit, and temperature sampling unit, respectively; the main control unit integrates the following: The power calculation and adaptive segmentation module is used to calculate instantaneous power and power change rate based on real-time collected voltage and current, and dynamically adjust the slice length according to the degree of power fluctuation. The three-dimensional nonlinear aging acceleration factor calculation module is used to calculate the three-dimensional nonlinear aging acceleration factor and single-slice lifetime loss based on the average voltage, average temperature and duration in each slice. The SOH cumulative calculation module is used to accumulate the lifetime loss of all slices to obtain the current SOH estimate. The no-load online self-calibration module is used to measure the actual ESR and actual capacity during the system's no-load window period, obtain the measured value of the reference SOH, and dynamically correct the model calibration coefficients. The Remaining Service Life (RUL) prediction module is used to calculate RUL based on the current SOH value, the preset scrap threshold, and the average aging rate, and output multi-level early warning signals. The output unit is used to output the current SOH estimation result, RUL prediction result, and early warning signal.

Images