A method and system for estimating the state of health of a lithium-sulfur battery

Abstract

The application discloses a kind of battery health estimation method and system of lithium-sulfur battery, it is related to battery management technical field, the early degradation signal of lithium-sulfur battery in the cycle aging process can be comprehensively captured from electricity, mechanics and relaxation dynamics three angles, the sensitivity and reliability of health recession early warning are significantly improved;Adopt the judgment mechanism that difference value and continuous three effective determination cycle same sign test of adjacent cycle are combined, effectively eliminate battery individual difference, temperature drift and sensor zero drift etc. Common mode interference, simultaneously through three to two voting mechanism, the false alarm risk caused by single parameter fluctuation is greatly reduced.

Description

Technical Field

[0001] This invention relates to the field of battery management technology, and in particular to a method and system for estimating the health of lithium-sulfur batteries. Background Technology

[0002] Lithium-sulfur batteries, due to their extremely high theoretical energy density and low sulfur resource cost, are considered an important development direction for next-generation high-energy-density energy storage systems. Accurate estimation of the state of health (SQH) during the cyclic aging process of lithium-sulfur batteries is crucial for the safety control and lifespan prediction of battery management systems. Currently, SQH estimation methods for lithium-sulfur batteries can be mainly divided into two categories: one is model-based parameter identification methods, which establish electrochemical models or equivalent circuit models and use algorithms such as extended Kalman filtering and recursive least squares to identify parameters such as battery internal resistance and polarization capacitance online, thereby inferring the SQH; the other is data-driven methods, which extract feature quantities from the charge-discharge curves (such as time integrals during constant current charging and changes in discharge voltage plateaus) and combine them with machine learning models (such as support vector regression and Gaussian process regression) to establish a SQH mapping relationship. In addition, some studies have estimated the SQH by measuring the ohmic or DC internal resistance of the battery under specific operating conditions and using empirical formulas for internal resistance growth and capacity decay.

[0003] However, existing methods have significant limitations when applied to lithium-sulfur batteries. For example, CN118068208A discloses a method and apparatus for estimating battery health status. This method obtains the relationship curve between the capacity increment and voltage of the battery to be estimated, divides it into multiple voltage ranges, converts the capacity increment measured at different temperatures to the capacity increment at a preset temperature, and then sums the capacity increments of each voltage range to obtain the battery charging capacity, ultimately determining the battery health status. While this method improves the estimation accuracy under different operating conditions through temperature compensation, its core relies on a complete capacity increment curve. Typically, the battery needs to be charged at a very low rate (generally below 0.05C) with constant current to obtain a high-resolution IC curve, which is difficult to achieve in practical applications (such as electric vehicles and energy storage systems). Furthermore, this method is based on the solid-state diffusion and phase transition mechanism of lithium-ion batteries. The charging and discharging process of lithium-sulfur batteries involves multiple redox reactions and the shuttle effect of lithium polysulfides. The shape of its capacity increment curve differs significantly from that of lithium-ion batteries, and its voltage plateau characteristics are not obvious. Directly applying this method results in significant errors.

[0004] Another common technical solution, such as CN121232062B, involves testing the battery under set experimental conditions to construct a data table and estimation model relating temperature, current, depth of discharge, and rate of change of state of health. This is then combined with actual operating condition electrical data to obtain initial values ​​of the state of health, and a final estimated value is obtained through weighted fusion. While this method has strong adaptability to operating conditions, it is essentially a lookup-fitting method based on a large amount of prior data. It requires systematic orthogonal experiments for different battery models to construct the data table and model, resulting in high calibration costs and long cycles. More importantly, this method primarily utilizes electrical parameters (voltage, current, temperature, depth of discharge) during the charging and discharging process, completely ignoring the significant electro-mechanical-thermal multi-physics coupling characteristics of lithium-sulfur batteries during cycling. During the charging and discharging of lithium-sulfur batteries, the deposition / dissolution of lithium metal at the negative electrode leads to significant changes in electrode thickness, which in turn alters the pressure on the back of the current collector. This mechanical deformation information is highly correlated with the battery's health status. Simultaneously, during the open-circuit voltage relaxation process after discharge, due to the redistribution of lithium polysulfides in the electrolyte and side reactions, multiple relaxation troughs appear on the voltage rate curve. The time intervals and widths of these troughs contain rich information about kinetic degradation. These mechanical and relaxation kinetic characteristics have not been effectively utilized in existing technologies. Summary of the Invention

[0005] In view of the problems existing in the above-mentioned background technology, the present invention is proposed.

[0006] Therefore, the problem to be solved by the present invention is how to achieve early and accurate determination of the health degradation of lithium-sulfur batteries by monitoring the trend of the same sign change of the three characteristic differences between adjacent cycles.

[0007] To solve the above-mentioned technical problems, the present invention provides the following technical solution: In a first aspect, the present invention provides a method for estimating the battery health of a lithium-sulfur battery, comprising: In each charge-discharge cycle, the ohmic internal resistance of the battery is measured, and the change in ohmic internal resistance in this cycle relative to the previous cycle is calculated and recorded as the descent slope. After the charging and discharging of the same cycle are completed, the resistance value of the piezoresistive sensor on the back of the negative electrode current collector is read, and the linkage difference between the charging resistance and the discharging resistance is calculated. After the discharging is completed, the open-circuit voltage is recorded and the voltage change rate curve is obtained by differential calculation. The time interval between the two relaxation troughs and the peak width of the second trough are extracted, and the time interval and the peak width are multiplied to obtain the correlation product. The difference in descent slope between adjacent cycles, the difference in linkage difference between adjacent cycles, and the difference in correlation product between adjacent cycles are calculated and recorded as the first difference, the second difference, and the third difference, respectively. When at least two of the first difference, the second difference, and the third difference remain the same sign in three consecutive valid judgment cycles, the health of the lithium-sulfur battery is determined to have degraded.

[0008] As a preferred embodiment of the battery health estimation method for a lithium-sulfur battery according to the present invention, the calculation of the change in ohmic internal resistance of the current cycle relative to the previous cycle, denoted as the descent slope, includes: in each charge-discharge cycle, defining the start time of the cycle as the start time of the charging current and the end time of the cycle as the time when the discharge cutoff voltage is reached; using the current jump at the moment of end of charging and the moment of start of discharging, reading the first rebound value of the battery terminal voltage at a preset delay after the charging current is cut off, and calculating the ohmic internal resistance at the moment of end of charging in combination with the charging current; reading the second rebound value of the battery terminal voltage at a preset delay after the discharge current is cut off, and calculating the ohmic internal resistance at the moment of end of discharging in combination with the discharge current; taking the average of the two calculated ohmic internal resistances as the representative value of the ohmic internal resistance of the current cycle; for the nth cycle and n≥2, the descent slope of the current cycle is the difference between the representative value of the ohmic internal resistance of the (n-1)th cycle and the representative value of the ohmic internal resistance of the nth cycle. If the difference is positive, the descent slope is valid; otherwise, it is marked as invalid.

[0009] As a preferred embodiment of the battery health estimation method for a lithium-sulfur battery according to the present invention, the calculation of the resistance linkage difference between the charging and discharging states includes: attaching a piezoresistive sensor to the center of the back of the negative electrode current collector of the battery; the two leads of the piezoresistive sensor are connected to the data acquisition channel after insulation protection, and the lead wires are provided with a margin of movement when assembling the battery; at the moment when the constant current charging ends in the same cycle, i.e., when the charging cutoff voltage is reached, the resistance value of the piezoresistive sensor is read after a preset short time of stabilization; at the moment when the constant current discharging ends in the same cycle, i.e., when the discharging cutoff voltage is reached, the resistance value of the piezoresistive sensor is read; calculating the linkage difference for this cycle, wherein the linkage difference is the resistance value read at the moment when charging ends minus the resistance value read at the moment when discharging ends, and negative values ​​are allowed and the sign is retained; for the nth cycle and n≥2, calculating the linkage difference between adjacent cycles, wherein the linkage difference is the linkage difference of the nth cycle minus the linkage difference of the (n-1)th cycle, and the linkage difference is always valid.

[0010] As a preferred embodiment of the battery health estimation method for lithium-sulfur batteries described in this invention, the correlation product includes: after the discharge ends, a preset short delay is made to ensure that the piezoresistive sensor reading is completed, and a resting phase begins. The open-circuit voltage change sequence over time is recorded at a preset sampling frequency, with the time origin being the start of the resting phase. The open-circuit voltage sequence is subjected to first-order difference to obtain a voltage change rate curve, and the voltage change rate curve is filtered by a moving average with a preset window length to remove high-frequency noise. On the filtered voltage change rate curve, all local minima are identified as relaxation troughs, where a local minima is defined as a point whose value is less than the values ​​of its two adjacent points on either side. The time of occurrence of each relaxation trough is recorded. If at least two relaxation troughs exist, the earliest occurring relaxation trough and the second earliest occurring relaxation trough are used to calculate the two relaxation troughs. The time interval between the two relaxation troughs is calculated. If the time interval is less than two relaxation troughs, the time interval is set to equal the duration of the entire resting phase. The peak width of the second relaxation trough is taken. Based on the voltage change rate value of the second relaxation trough, the change rate threshold is calculated as the voltage change rate value plus half of the absolute value of the voltage change rate value. The search proceeds to the left from the second relaxation trough until the voltage change rate is greater than the change rate threshold for the first time, and then to the right until the voltage change rate is greater than the change rate threshold for the first time. The time length between the left and right boundaries is the peak width. If the time interval is less than two relaxation troughs, the peak width is set to zero. The correlation product of the current cycle is calculated. The correlation product is the product of the time interval and the peak width. For the nth cycle and n≥2, the correlation product difference between adjacent cycles is calculated. The correlation product difference is the correlation product of the nth cycle minus the correlation product of the (n-1)th cycle.

[0011] As a preferred embodiment of the battery health estimation method for lithium-sulfur batteries according to the present invention, the following steps are taken: when at least two of the first difference, second difference, and third difference values ​​maintain the same sign in three consecutive valid judgment cycles, the health of the lithium-sulfur battery is determined to have deteriorated. This includes: for the nth cycle and n≥2, if the first difference is valid, then this cycle is a valid judgment cycle, and a triplet consisting of the first difference, second difference, and third difference is obtained; three first-in-first-out queues are maintained, each queue having a capacity of three valid judgment cycles; the first difference and second difference are pushed into the corresponding queue, and the third difference is pushed in when it is valid and not pushed in when it is invalid; when the queue is full, the earliest value is popped; when the first difference is invalid, the cycle is skipped and not updated; when all three queues are full of three values, each time a new valid judgment cycle is obtained, the queues are updated and the following is determined: check whether the signs of the three values ​​in each queue are completely the same and all are non-zero; if the signs of at least two queues are all the same, then a deterioration is determined, an early warning signal is issued, and the current cycle number is recorded.

[0012] As a preferred embodiment of the battery health estimation method for lithium-sulfur batteries described in this invention, the method further includes a measurement anomaly warning step: when there are no valid judgment cycles for a preset number of consecutive cycles, a measurement anomaly warning is issued, and subsequent cycles are monitored.

[0013] As a preferred embodiment of the battery health estimation method for lithium-sulfur batteries described in this invention, the method further includes a degradation confirmation step: after issuing an early warning signal, the method continues to monitor subsequent effective judgment cycles; if, within a preset number of consecutive effective judgment cycles after issuing the early warning, the number of parameters counted as valid is greater than or equal to two again, the degradation is confirmed to be irreversible, and a final alarm is issued.

[0014] Secondly, the present invention provides a battery health estimation system for lithium-sulfur batteries, comprising: an internal resistance slope extraction module, which measures the ohmic internal resistance of the battery in each charge-discharge cycle and calculates the change in ohmic internal resistance in the current cycle relative to the previous cycle, denoted as the falling slope; a piezoresistive difference calculation module, which reads the resistance value of the piezoresistive sensor on the back of the negative electrode current collector after charging and discharging in the same cycle, and calculates the linkage difference between the charging resistance and the discharging resistance; a relaxation product acquisition module, which records the open-circuit voltage and obtains the voltage change rate curve by difference after discharging, extracts the time interval between two relaxation troughs and the peak width of the second trough, and multiplies the time interval and the peak width to obtain the correlation product; and a determination module, which calculates the difference in falling slope between adjacent cycles, the difference in linkage difference between adjacent cycles, and the difference in correlation product between adjacent cycles, denoted as the first difference, the second difference, and the third difference, respectively; when at least two of the first difference, the second difference, and the third difference remain the same sign in three consecutive valid determination cycles, the health of the lithium-sulfur battery is determined to have deteriorated.

[0015] Thirdly, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program instructions are executed by the processor, they implement the steps of a battery health estimation method for a lithium-sulfur battery as described in the first aspect of the present invention.

[0016] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program instructions are executed by a processor, they implement the steps of a battery health estimation method for a lithium-sulfur battery as described in the first aspect of the present invention.

[0017] The beneficial effects of this invention are as follows: This invention can comprehensively capture early degradation signals of lithium-sulfur batteries during cycle aging from three perspectives: electrical, mechanical, and relaxation kinetics, significantly improving the sensitivity and reliability of health degradation early warning. The judgment mechanism, which combines adjacent cycle difference with the same sign test of three consecutive valid judgment cycles, effectively eliminates common-mode interference such as individual battery differences, temperature drift, and sensor zero drift. Simultaneously, the three-to-two voting mechanism significantly reduces the risk of false alarms caused by fluctuations in a single parameter. Furthermore, this invention does not require the construction of complex electrochemical models or extensive prior calibration experiments; the calculations are simple and easy to implement online. It can provide accurate health status feedback for the battery management system without increasing additional hardware costs, thereby effectively extending the battery pack's lifespan and improving system safety. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart of a method for estimating the battery health of a lithium-sulfur battery.

[0020] Figure 2 This is a structural diagram of a battery health estimation system for lithium-sulfur batteries. Detailed Implementation

[0021] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0022] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0023] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0024] Figure 1 This is a flowchart of a battery health estimation method for lithium-sulfur batteries according to an embodiment of the present invention. Figure 1As shown, a battery health estimation method for lithium-sulfur batteries includes: S1: In each charge-discharge cycle, measure the ohmic internal resistance of the battery and calculate the change in ohmic internal resistance in this cycle relative to the previous cycle, which is denoted as the decreasing slope.

[0025] First, in this embodiment, each charge-discharge cycle of the lithium-sulfur battery is executed according to a preset charge-discharge regime. The start time of a single cycle is the moment when the charging current begins to be applied to the battery, and the end time of the cycle is the moment when the discharge cutoff voltage is reached. A complete cycle includes, in sequence, a constant current charging stage, a constant current discharging stage, and after charging, there is no resting stage; the battery directly enters the discharging stage. In this embodiment, the ohmic internal resistance is measured only during the discharging stage to avoid measurement errors introduced during the charging stage due to uneven electrode polarization.

[0026] S1.1: In this embodiment, in order to avoid the impact of frequent current interruptions on actual electrical equipment, the instantaneous measurement method of charge-discharge switching is adopted.

[0027] Specifically, during the interval between the end of constant current charging and the start of discharging, the BMS typically controls the timing to ensure a switching interval of 50~200ms. After the charging current is cut off, the battery terminal voltage rebound value is read with a 1ms delay. Simultaneously record the charging current before disconnection. Calculate the internal resistance in ohms at the end of charging: ; Similarly, at the end of the constant current discharge, the reading is taken at the instant after the discharge current is cut off. Calculate the ohmic internal resistance at the end of discharge. .Will and The average value is used as the ohmic internal resistance characteristic value of this cycle and is used for subsequent slope calculation. If the system allows for current interruption with an extremely low duty cycle (<0.1%), 1 to 2 additional measurement points can be added during the discharge phase to form a sparse sequence.

[0028] S1.2: After obtaining the representative value of the ohmic internal resistance for each cycle, it is denoted as... That is, the average value of two internal resistance measurements, one at the end of charging and one at the end of discharging, for the first... One cycle ( The descent slope for this cycle is calculated using the following formula: ; Since the difference in the sequence number of adjacent cycles is 1, the descending slope is the difference in the representative value of the internal resistance of adjacent cycles. If the difference is positive, it means that the ohmic internal resistance decreases as the cycle increases, and the descending slope of the cycle is marked as valid; if the difference is negative or zero, it is marked as invalid.

[0029] It should be noted that the decreasing slope is not the rate of change of ohmic internal resistance over time within a single cycle, but rather the chain-reaction change in the representative value of ohmic internal resistance between two adjacent charge-discharge cycles, i.e., the rate of decrease. The ohmic internal resistance of the cycle is represented by the value minus the first... The ohmic resistance of the loop is represented by this value. A positive value indicates that the ohmic resistance decreases as the loop increases, and this is defined as valid. A negative or zero value indicates that the resistance does not decrease or increases, and this is marked as invalid. The term "decreasing slope" is used here to emphasize directionality and validity judgment, distinguishing it from the conventional concept of slope.

[0030] S1.3: If the first loop or the first If the ohmic internal resistance value of any loop in the loop cannot be obtained due to measurement failure, such as abnormal voltage rebound value reading, current interruption timing error, sensor signal saturation, etc., the falling slope of the corresponding loop is marked as invalid, and the falling slope difference value of subsequent loops will no longer be calculated.

[0031] S1.4: For the first One cycle ( ), only when the first Loop and the first When all the descending slopes of the loops are valid, calculate the difference in descending slopes between adjacent loops, and record it as the first difference. : ; in, This is denoted as the first difference. If the slope of any of the cycles is invalid, it is marked as invalid. Invalid and will not be included in subsequent decline trend statistics.

[0032] S2: After the charging and discharging cycles are completed, read the resistance value of the piezoresistive sensor on the back of the negative current collector and calculate the linkage difference between the charging resistance and the discharging resistance.

[0033] It should be noted that in this embodiment, a piezoresistive thin-film sensor (e.g., a piezoresistive film based on carbon nanotubes or conductive polymer composite materials, whose resistance monotonically decreases with increasing pressure) is used as the deformation sensing element. This sensor is attached to the center of the back side of the battery's negative electrode current collector. The negative electrode current collector is typically copper foil or carbon-coated copper foil. During battery charging and discharging, the back side (i.e., the side not coated with the negative electrode active material) experiences normal compressive stress due to the volume expansion of the negative electrode active material.

[0034] The bonding method uses a flexible adhesive layer, such as acrylic pressure-sensitive adhesive or silicone, with a thickness of 10–50 μm. After the sensor is bonded, its surface and lead solder joints require the following protective treatments in sequence: First, apply a layer of parylene or polyimide insulating coating, 5–10 μm thick, to isolate it from electrolyte vapor corrosion; then cover it with a layer of electrolyte-resistant fluororubber protective tape, completely wrapping the sensor edges; the leads use 0.05 mm thick nickel strip or nickel-plated copper braided wire, and the lead solder joints are additionally reinforced with UV adhesive. After insulation protection, the leads at both ends of the sensor are connected to the external data acquisition channel. When assembling the battery, sufficient lead movement allowance must be reserved between the cell and the casing, for example, by making an S-shaped bend or a spiral coil, to prevent the sensor leads from breaking or the solder joints from falling off due to stretching during hundreds of cycles of expansion and contraction of the battery.

[0035] A 500-cycle test was conducted on a lithium-sulfur pouch battery assembled with a piezoresistive sensor using the above packaging method. After every 50 cycles, the sensor's zero-point resistance (discharged state, after 1 hour of rest) and full-charge resistance were measured at room temperature. The test results showed that in the first 200 cycles, the zero-point resistance drift was <±3%, and the full-charge resistance drift was <±5%. From 200 to 500 cycles, the drift gradually increased to ±8% and ±12%, but the inter-cycle trend of the linkage difference remained monotonic, without sign reversal. This result indicates that the packaged sensor can provide effective relative change information within 500 cycles, meeting the requirements for health estimation.

[0036] It should be noted that, in order to avoid the influence of lead capacitance on fast signals, the leads should be as short as possible and twisted-pair shielded cables should be used.

[0037] Furthermore, during the same charge-discharge cycle, when the constant current charging phase of the battery ends, i.e., when the battery terminal voltage first reaches the preset charging cutoff voltage (e.g., 2.8V set according to the characteristics of lithium-sulfur batteries), the controller issues a charging stop command. At this time, due to the sudden interruption of the charging current, the charge distribution at the electrode / electrolyte interface inside the battery needs to relax briefly, and directly reading the sensor resistance may contain spike noise caused by electromagnetic interference at the moment of current cutoff.

[0038] Therefore, in this embodiment, after the charging cutoff voltage is reached, a preset short delay, such as 0.5 seconds, is made sufficient to allow the inductive spike to decay without causing significant concentration polarization relaxation. Then, the resistance value of the piezoresistive sensor is read and recorded as follows: The delay time was calibrated experimentally: a large current was suddenly applied or cut off when the battery was at rest, and the stabilization time of the sensor signal was observed. The result was taken as twice the delay value.

[0039] Furthermore, at the end of the constant current discharge phase of the same cycle, i.e., when the battery terminal voltage drops to the preset discharge cutoff voltage (e.g., 1.5V or as set according to the specific battery system), the resistance value of the same piezoresistive sensor is immediately (without additional delay) read and recorded as follows. It should be noted that the reason for not adding a delay at the end of discharge is that after discharge, the battery will enter a resting phase, during which the current collector pressure will rapidly relax. Adding a delay would result in the loss of the true pressure state at the end of discharge.

[0040] To ensure the accuracy of readings at the end of the discharge, the data acquisition system should have sufficient response speed (sampling frequency not less than 100Hz) to capture the resistance value at the instant the cutoff voltage is reached. In actual control, synchronous reading is usually achieved by triggering an interrupt using a voltage comparator.

[0041] Furthermore, the linkage difference is calculated based on the charging and discharging resistance values ​​obtained in the same cycle. The difference can be positive or negative, and retains its original sign.

[0042] when A value greater than 0 indicates that the resistance in the charging state is greater than the resistance in the discharging state. Since the resistance of a piezoresistive sensor decreases with increasing pressure, a higher resistance in the charging state means that the pressure on the back of the current collector during charging is less than during discharging (i.e., the electrode expands more during discharging), which aligns with the volume expansion characteristics caused by lithium-ion insertion into the negative electrode. For healthy lithium-sulfur batteries, the volume expansion of the negative electrode due to lithium deposition in the charging state is usually greater than that in the discharging state; therefore, the resistance in the charging state is less than that in the discharging state, and the difference is usually negative. As the battery begins to degrade, this difference may continue to increase (changing in the positive direction), manifesting as a positive difference between adjacent cycles.

[0043] when When the value is less than 0, it means that the resistance in the charging state is less than the resistance in the discharging state, that is, the pressure during charging is greater than that during discharging. This may indicate an abnormality inside the battery, such as lithium plating causing uneven volume expansion during charging, or local blockage of the separator.

[0044] when When the value is approximately 0, it indicates that there is no significant difference in pressure between the charging and discharging states, which may mean that the active material of the electrode has been severely pulverized and detached, losing its expansion capacity.

[0045] It can be seen that, through a single parameter It comprehensively reflects the difference in mechanical deformation of the battery under charging and discharging states. Compared with monitoring only the resistance of a single state or the absolute resistance value, this difference eliminates common-mode interference from individual sensor zero drift, temperature drift, and initial assembly stress between batteries.

[0046] Preferably, to further filter out low-frequency drift, such as sensor aging and long-term changes in the overall average battery pressure, this invention uses the difference between adjacent cycles as a health characteristic. For the first... One cycle ( ), calculate the change in linkage difference between adjacent loops: ; in, This is denoted as the second difference. Because... It can be measured in each cycle regardless of whether there is an anomaly or not; even if the sensor fails, it will still output an abnormal value. It is always defined as valid and does not require an invalidation flag. However, in actual algorithms, if multiple consecutive loops... If the sensor's measurement range is exceeded, for example, if an open circuit or short circuit occurs, the system should issue a separate sensor fault alarm, but this alarm is not within the scope of the health decline judgment of this invention.

[0047] It can be seen that by differentiating between adjacent cycles, the trend of charge-discharge pressure differences over time can be amplified. For example, when the battery begins to exhibit irreversible expansion, the discharge pressure increases with each cycle while the charge pressure changes relatively little, leading to... Decrease cyclically, thereby A consistently negative result provides a clear signal for subsequent same-sign detection.

[0048] It should be noted that the piezoresistive sensor should be pre-compressed (e.g., applied with pressure from 0 to 500 kPa several times) before assembly to eliminate mechanical hysteresis effect, and the data acquisition channel should be equipped with a constant current source (e.g., 100 μA) for power supply, and the resistance change should be measured by a bridge or proportional method to improve noise immunity.

[0049] S3: After the discharge is completed, record the open circuit voltage and obtain the voltage change rate curve by difference. Extract the time interval between the two relaxation troughs and the peak width of the second trough. Multiply the time interval and the peak width to obtain the correlation product.

[0050] Before performing this step, it is first determined whether the system has been continuously stationary for a preset duration (e.g., 5 minutes) since the end of self-discharge and whether the current fluctuation during this period is less than the fluctuation threshold (e.g., 0.01C). If this condition is not met, the relaxation feature calculation for the current cycle is skipped, and the associated product of that cycle is marked as not obtained. In the subsequent calculation of the difference between adjacent cycles, if the associated product of the current cycle or the previous cycle is not obtained, the corresponding third difference is considered invalid, and the decay judgment is based solely on the first and second differences.

[0051] It should be noted that after reading the piezoresistive sensor resistance value at the moment the discharge ends in the above steps, to avoid interference from the sensor reading process during the initial voltage relaxation phase, this embodiment first sets a preset short delay (e.g., 0.2 seconds). This delay is used to ensure that the discharge current has been completely cut off and the data acquisition system has completed the latching of the piezoresistive signal. After the delay ends, the formal resting phase begins, with the resting time fixed at a preset value, such as 5 minutes. This resting time should be long enough to cover the main process of the open-circuit voltage relaxing from the discharge end value to the thermodynamic equilibrium potential, while avoiding excessive length that would reduce test efficiency.

[0052] During the resting phase, the battery open-circuit voltage is continuously recorded at a preset sampling frequency (e.g., 10Hz). The origin of time is defined as the start of the rest period, i.e., the instant the delay ends. The selection of the sampling frequency needs to take into account both the resolution of the voltage change rate curve and the amount of data storage: 10Hz is sufficient to capture the relaxation details corresponding to millivolt-level voltage fluctuations without generating excessive computational burden.

[0053] S3.1: To highlight the dynamic characteristics during the relaxation process, the open-circuit voltage sequence is subjected to first-order differential processing to obtain the voltage change rate curve. : ; in, The sampling interval is the time difference between two adjacent sampling points, measured in seconds; its reciprocal is the sampling frequency. The voltage change rate reflects the rate of change of the battery open-circuit voltage over time, measured in volts per second (V / s). Immediately after discharge, due to the rapid decay of concentration polarization, the voltage typically rises rapidly initially, with a positive and relatively large change rate. It then enters a slow recovery phase, with the change rate approaching zero. Secondary voltage fluctuations may occur due to the lithium polysulfide shuttle effect, resulting in multiple local minima on the rate of change curve, known as relaxation troughs.

[0054] Since the original voltage signal may contain microvolt-level noise, direct differential processing would amplify high-frequency noise components. Therefore, this embodiment applies a moving average filter to the obtained voltage change rate curve, with a preset filter window length of 5, i.e., the arithmetic mean of 5 consecutive data points. This window length effectively suppresses noise while preserving the relaxation trough shape: a window that is too short will result in insufficient filtering, while a window that is too wide will smooth out the trough characteristics. The filtered curve is denoted as... .

[0055] It should be noted that the first-order differential transforms the voltage signal into a more sensitive rate signal, making the voltage plateau or inflection point, which was originally difficult to observe directly, a clear extreme point.

[0056] S3.2: On the filtered voltage change rate curve, identify all local minimum points, which are the relaxation troughs.

[0057] The criteria for determining local minima are as follows: For a data point on the curve, if the index of the point from both the start and end points of the sequence is not less than 2 (excluding the first two and last two points), and the value of the point is less than the values ​​of its two adjacent points on either side, then the point is marked as a relaxation trough. Points within less than four neighborhoods of the start or end point of the sequence are not included in the local minima identification to avoid boundary effects.

[0058] The use of two points on each side, rather than adjacent points, is to improve recognition reliability in environments with small fluctuations and avoid misjudging minor shaking as a trough. If there are multiple consecutive points on the curve that meet the conditions, the position of the minimum value is taken as the trough.

[0059] Furthermore, the timing of each relaxation trough was recorded (relative to the start of the resting period). A typical lithium-sulfur battery relaxation process may exhibit two distinct troughs: The first trough corresponds to the turning point where the voltage rise rate changes from positive to negative during the rapid decline of concentration polarization after discharge. In practice, due to the multi-step reaction in lithium-sulfur batteries, the voltage change rate may be positive first, then negative, and then positive again, resulting in multiple extreme values.

[0060] The second trough often occurs later and is related to the redistribution of lithium polysulfides or internal side reactions. Morphology (e.g., depth, width) is sensitive to battery health.

[0061] It can be seen that the above operation adopts a strict five-neighbor minimum value determination, avoiding false alarms caused by single-point noise or small fluctuations.

[0062] S3.3: If the number of identified relaxation troughs is greater than or equal to two, then the earliest occurring trough is taken as the first trough, and the second earliest occurring trough is taken as the second trough. The time interval between the two is calculated. .

[0063] This interval reflects the timescale of ion redistribution and side reaction kinetics within the battery. For healthy batteries, the interval between the two troughs is typically stable within a certain range; however, when the separator becomes partially blocked or the electrolyte viscosity increases, the second trough is delayed, leading to… The second trough may increase; conversely, when the electrode structure collapses and the active material falls off, the second trough may occur earlier or disappear.

[0064] If fewer than two relaxation troughs are identified during the entire resting period, for example, if there is only one trough or no local minimum, then let This is equal to the entire resting period, i.e., a preset fixed duration, such as 300 seconds. This situation usually occurs when the battery is severely damaged and the voltage changes monotonically without fluctuations. Assigning the maximum interval value can cause its correlation product to deviate significantly from the normal range, thereby triggering a degradation judgment.

[0065] S3.4: Take the second relaxation trough, i.e., the second local minimum point, and calculate the peak width. The specific steps are as follows: (1) Let the voltage change rate at the trough be . It is usually a negative value, representing the rate of voltage drop, depending on the specific curve shape, but it is a minimum value at the trough.

[0066] (2) Calculate the threshold .like Then search left and right until the rate of voltage change first exceeds Otherwise, the peak width will be set directly to a preset minimum value (e.g., 0.01 seconds) or the original search rule will be used but changed to be less than the minimum value the first time. (Choose according to actual physical meaning). To avoid excessive complexity, a unified definition of half-width at half-maximum can be adopted: when... If the peak width is not calculated for the trough, the correlation product of the cycle is directly set to 0.

[0067] (3) Starting from the next adjacent point to the left of the trough, search for points on the voltage change rate curve to the left until the first voltage change rate greater than 1 is found. Record the time at that point. If the search reaches the left endpoint of the sequence but the sequence is not found, then the time of the left endpoint is used as the time. .

[0068] Similarly, search to the right (in the direction of increasing time) until you find the first voltage change rate greater than... Record the time at that point. If the search reaches the right endpoint of the sequence but no point is found, then the time of the right endpoint is taken as the starting point. .

[0069] (4) Then the peak width of the valley The unit is seconds.

[0070] If the number of relaxation troughs is less than two during the entire resting period, then let Among them, peak width This reflects the sharpness of the second relaxation trough, namely... The smaller the value, the narrower the trough, corresponding to a rapid recovery of the voltage change rate from negative, which is usually related to the rapid rebalancing process inside the battery. The larger the value, the longer the trough lasts, which may indicate the presence of persistent side reactions or ion retardation on the electrode surface.

[0071] It should be noted that defining peak width using half-width at half-maximum (with 50% of the absolute value as the threshold) is a mature method in the field of signal processing, which has clear physical meaning and is not affected by baseline drift.

[0072] S3.5: Multiply the time intervals and peak widths obtained above to obtain the correlation product of this cycle: ; when hour, .when Take the entire settling period and hour, It will be exceptionally large. The unit of the correlated product is the square of time (seconds). 2 It penalizes both excessively long time intervals and excessively wide second troughs. Experiments show that when lithium-sulfur batteries begin to experience irreversible degradation, the time interval between the two relaxation troughs tends to lengthen, while the peak width of the second trough also increases. The product of these two factors exhibits an accelerated trend, resulting in a higher signal-to-noise ratio and prediction lead time compared to a single parameter.

[0073] It should be noted that a composite health factor is constructed by multiplying two independent relaxation features (event interval and valley width). This composite factor amplifies the early, subtle changes in a single parameter, enabling earlier prediction of battery health degradation trends, and avoids the complex logic of setting two separate thresholds.

[0074] S3.6: The time interval of the relaxation trough and the peak width of the second trough both change with the ambient temperature.

[0075] To reduce the interference of temperature on the correlation product, a preset temperature compensation coefficient can be used to substitute the measured temperature, time interval, and peak width into the linear correction calculation to obtain the equivalent time interval and equivalent peak width after correction to the reference temperature (e.g., 25℃), and then multiply them to obtain the correlation product.

[0076] The temperature compensation method described above is only one feasible example provided by the present invention. In practical applications, other forms of compensation functions, such as polynomial fitting, table lookup interpolation, etc., can also be adopted according to the battery model, usage environment, or experimental data. Alternatively, the compensation step can be omitted when temperature measurement conditions are not available. However, it is recommended that this method be used only for constant temperature environments or comparative analysis of adjacent cycles.

[0077] S3.7: To eliminate the influence of initial differences between batteries and common-mode factors such as ambient temperature, and to maintain consistency with the differential characteristics of steps S1 and S2, this embodiment further calculates the change in the correlation product between adjacent cycles, denoted as the third difference. For the first... One cycle ( ): ; If the associated product of the current loop or the previous loop is marked as not obtained, then the corresponding third difference is marked. Record as invalid; otherwise, calculate according to the above formula. And this difference always has a numerical value (including zero) in calculation.

[0078] It should be noted that when multiple consecutive loops... hour, It is also 0. At this time, the handling of zero-value symbols will be managed by the unified logic of step S4 (zero values ​​are not included in the same sign statistics).

[0079] S4: Calculate the difference in the descent slope between adjacent loops, the difference in the linkage difference between adjacent loops, and the difference in the correlation product between adjacent loops, and denot them as the first difference, the second difference, and the third difference, respectively.

[0080] S4.1: In the preceding steps, the descent slope, linkage difference, and correlation product for each cycle have been obtained. To eliminate the interference of individual battery differences and absolute cycle values, and to unify the expression of the changing trends of the three physical quantities, this embodiment focuses on the [specific step / step]. One cycle ( The three differences between adjacent cycles are calculated separately, and the characteristic quantities of the three different physical dimensions (electricity, mechanics, and relaxation dynamics) are uniformly transformed into dimensionless changes, i.e., differences, and assigned the same positive and negative physical meanings—positive values ​​indicate that the characteristic has increased compared to the previous cycle, and negative values ​​indicate that it has decreased. This unification process lays the foundation for subsequent joint sign analysis.

[0081] S4.2: Since the first difference might be marked as invalid due to an invalid descending slope, to ensure that the analysis of the three differences is based on the same set of cyclic data: For the One cycle ( If the first difference If a loop is valid, meaning it was not marked as invalid in step S1, then the loop is a valid decision loop; each valid decision loop corresponds to a triplet. If the first difference is invalid, the loop will not generate a valid decision loop. This step will skip the loop directly, without performing any queue update or decay judgment, and the loop number will continue to increment.

[0082] It needs to be clarified that when When invalid, the third difference in the triplet of the loop is considered missing, and this parameter will not participate in the voting in subsequent sign consistency judgments.

[0083] It should be noted that by filtering the above-mentioned effective judgment loop, it is ensured that the three differences come from the same set of valid measurement data. This can avoid the misalignment of the loop numbers of the three features due to the failure of a certain sensor or algorithm, thereby preventing false sign consistency judgments.

[0084] S4.3: To detect the sign consistency of three consecutive valid decision loops, this embodiment maintains a First-In-First-Out (FIFO) queue for the first difference, second difference, and third difference, respectively, with a fixed capacity of three queues. The queue update rules are as follows: Whenever a valid loop is encountered, the loop's... Push them into their respective queues; if If a value is valid (not marked as invalid), it is pushed into the third difference queue. Otherwise, the third difference queue remains unchanged; no new value is pushed, and no old value is popped. If the queue is full (three values ​​are already stored) when pushing, the earliest value stored in the queue is popped first, and then the new value is pushed, always keeping the queue containing only the differences of the three most recent valid judgment loops. If an invalid judgment loop is encountered, no queue is updated, the queue contents remain unchanged, and no judgment is made.

[0085] It should be noted that the FIFO queue with a capacity of three described above can automatically maintain a sliding time window. The three valid decision loops within the window may not be adjacent in physical time (invalid loops are skipped in between), but they are consistent in data quality. This design ensures both the continuity of trend analysis and eliminates interference from invalid data.

[0086] S4.4: In the initial stage, if there are fewer than three valid judgment loops, only accumulate data without performing decay judgment until each of the three queues has three values.

[0087] When each of the three queues is filled with three values ​​for the first time (i.e., it has gone through three valid decision cycles), a decay judgment is immediately executed; thereafter, every time a new valid decision cycle is obtained and the queue is updated, a judgment step is also immediately executed.

[0088] Specifically, the judgment steps are as follows: First, for the three differences in each queue, determine the sign of each difference. If the difference is positive, record the sign as positive; if the difference is negative, record the sign as negative; if the difference is zero, record the sign as zero.

[0089] For a single parameter, such as the first difference, check if the signs of the three values ​​in the parameter queue are exactly the same. If all three values ​​are positive or all three values ​​are negative, the parameter is considered valid, i.e., it changes monotonically in the same direction, and the count value is incremented by one. If the three values ​​contain zero, such as positive, positive, and zero; or negative, zero, and negative; or all zero; or both positive and negative, the parameter is considered invalid, and the count value remains unchanged.

[0090] The total number of valid values ​​among the three parameters (first difference, second difference, and third difference) is denoted as . .like If at least two parameters show the same sign change in the most recent three valid judgment cycles, then the health of the lithium-sulfur battery is determined to be deteriorating, an early warning signal is issued, and the current cycle number is recorded.

[0091] It should be noted that the above-mentioned voting mechanism of selecting two out of three reduces the risk of false alarms from a single sensor or feature parameter. Even if a feature is temporarily insensitive due to physical reasons, as long as two other features simultaneously point to the same deteriorating trend, the system can issue an early warning, thus improving reliability. At the same time, zero values ​​are excluded from the same-sign statistics, avoiding false same-sign determinations due to excessively small changes.

[0092] S4.5: In actual operation, there may be situations where multiple consecutive cycles fail to produce a valid judgment cycle, such as the continuously invalidating downward slope in step S1. This could be due to drastic fluctuations in battery internal resistance or a measurement system malfunction. Therefore, this embodiment sets a preset threshold number, such as 5 cycles. When no valid judgment cycle is generated for a preset number of consecutive cycles, the system issues a measurement anomaly warning, prompting the operator to check the current interruption method measurement circuit or the ohmic internal resistance calculation logic. After issuing the warning, the system continues to monitor subsequent cycles, and once a valid judgment cycle reappears, the normal judgment process resumes.

[0093] S4.6: To prevent false alarms caused by single noise or accidental fluctuations, this embodiment does not immediately perform final actions (such as shutdown or maintenance) after issuing the first degradation warning signal, but instead enters a confirmation phase.

[0094] Specifically, after issuing a warning signal, the system continues to monitor subsequent valid judgment cycles; if the warning occurs again within a preset number of valid judgment cycles after the warning is issued (e.g., 10 valid judgment cycles), If the situation is as described above, the degradation trend is confirmed to be irreversible, and a final alarm signal is issued. The final alarm signal can be used to trigger the battery management system (BMS) to take protective actions, such as limiting the charging and discharging power, marking that the cell in the battery pack needs to be replaced, or uploading to a remote monitoring platform.

[0095] If it does not reappear within a preset number of consecutive valid judgment cycles. If the warning signal is cancelled (considered to be an occasional disturbance), normal monitoring will continue.

[0096] It can be seen that lithium-sulfur batteries may experience temporary characteristic changes due to external factors such as temperature fluctuations and brief overcharging during actual use. Through the reproduction verification of multiple consecutive valid judgment cycles, it is possible to distinguish between irreversible trends caused by aging and temporary anomalies, thereby improving the reliability of the system's decision-making.

[0097] It should be noted that zero is neither considered positive nor negative in the sign consistency judgment. Specifically, when checking whether three values ​​in a parameter queue have the same sign, if at least one of them is zero, the queue does not meet the same sign condition (because zero is neither positive nor negative), and the parameter is counted as invalid. Zero values ​​are also not included in the statistics of monotonically changing values ​​in the same direction; that is, they are not voted on.

[0098] It should be noted that in actual operation, an anti-shake mechanism can be used. If the condition is met again within three consecutive valid judgment cycles, the warning will be maintained; otherwise, the warning will be cleared.

[0099] It should be noted that the open-circuit voltage relaxation analysis involved in this method requires the battery to be kept stationary after discharge. Therefore, it is more suitable for application scenarios with stationary conditions (such as periodic testing and laboratory aging tests). For situations where long-term stationary conditions are not allowed, only two parameters, the difference in internal resistance slope and the difference in voltage-resistance linkage, can be used for judgment. However, the confidence of the degradation criterion will decrease in this case.

[0100] Furthermore, such as Figure 2 As shown, this embodiment also provides a battery health estimation system for lithium-sulfur batteries, including, The internal resistance slope extraction module measures the battery's ohmic internal resistance in each charge-discharge cycle, calculates the change in ohmic internal resistance in the current cycle relative to the previous cycle, and records it as the falling slope. The piezoresistive difference calculation module reads the resistance value of the piezoresistive sensor on the back of the negative current collector after the completion of charging and discharging in the same cycle, and calculates the linkage difference between the charging resistance and the discharging resistance. The relaxation product acquisition module records the open-circuit voltage and differentially obtains the voltage change rate curve after the discharge is completed. It extracts the time interval between the two relaxation troughs and the peak width of the second trough, and multiplies the time interval and the peak width to obtain the correlation product. The determination module calculates the difference in the downward slope between adjacent cycles, the difference in the linkage difference between adjacent cycles, and the difference in the correlation product between adjacent cycles, which are respectively denoted as the first difference, the second difference, and the third difference. When at least two of the first difference, the second difference, and the third difference remain the same sign in three consecutive valid determination cycles, the health of the lithium-sulfur battery is determined to have deteriorated.

[0101] This embodiment also provides a computer device applicable to a battery health estimation method for lithium-sulfur batteries, including a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the battery health estimation method for lithium-sulfur batteries as proposed in the above embodiment.

[0102] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0103] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements the battery health estimation method for lithium-sulfur batteries as proposed in the above embodiments.

[0104] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for estimating the health of a lithium-sulfur battery, characterized in that: include: In each charge-discharge cycle, the ohmic internal resistance of the battery is measured, and the change in ohmic internal resistance in this cycle relative to the previous cycle is calculated and denoted as the descent slope. After the charging and discharging cycles are completed, the resistance value of the piezoresistive sensor on the back of the negative current collector is read, and the linkage difference between the charging resistance and the discharging resistance is calculated. After the discharge is completed, the open-circuit voltage is recorded and the voltage change rate curve is obtained by differential calculation. The time interval between the two relaxation troughs and the peak width of the second trough are extracted. The time interval and the peak width are multiplied to obtain the correlation product. The differences in the descent slope between adjacent cycles, the differences in the linkage differences between adjacent cycles, and the differences in the correlation products between adjacent cycles are calculated and denoted as the first difference, the second difference, and the third difference, respectively. When at least two of the first difference, the second difference, and the third difference remain the same sign in three consecutive valid judgment cycles, the health of the lithium-sulfur battery is determined to have deteriorated.

2. The battery health estimation method for lithium-sulfur batteries as described in claim 1, characterized in that: The calculation of the change in ohmic internal resistance in the current cycle relative to the previous cycle, denoted as the descending slope, includes: In each charge-discharge cycle, the start time of the cycle is defined as the start time of the charging current, and the end time of the cycle is defined as the time when the discharge cutoff voltage is reached. By utilizing the current jump at the moment of charging completion and the moment of discharging start, the first rebound value of the battery terminal voltage is read at a preset delay after the charging current is cut off, and the ohmic internal resistance at the moment of charging completion is calculated in combination with the charging current. The second rebound value of the battery terminal voltage is read at a preset delay time after the discharge current is cut off, and the ohmic internal resistance at the end of the discharge is calculated in combination with the discharge current. The average value of the ohmic internal resistance obtained from the two calculations is taken as the representative value of the ohmic internal resistance for this cycle. For the nth cycle and n≥2, the descending slope of this cycle is the difference between the representative value of the ohmic internal resistance of the (n-1)th cycle and the representative value of the ohmic internal resistance of the nth cycle. If the difference is positive, the descending slope is valid; otherwise, it is marked as invalid.

3. The battery health estimation method for lithium-sulfur batteries as described in claim 1, characterized in that: The calculation of the resistance difference between the charging and discharging states includes: A piezoresistive sensor is attached to the center of the back of the negative current collector of the battery. The two leads of the piezoresistive sensor are connected to the data acquisition channel after being insulated and protected. When assembling the battery, allowance is reserved for the movement of the leads. At the moment when the constant current charging cycle ends, i.e. when the charging cutoff voltage is reached, the resistance value of the piezoresistive sensor is read after a short time of stabilization. At the instant the constant current discharge of the same cycle ends, i.e. when the discharge cutoff voltage is reached, the resistance value of the piezoresistive sensor is read. Calculate the linkage difference for this cycle. The linkage difference is the resistance value read at the moment the charging ends minus the resistance value read at the moment the discharging ends. Negative values ​​are allowed, and the sign is retained. For the nth cycle and n≥2, calculate the linkage difference between adjacent cycles. The linkage difference is the linkage difference of the nth cycle minus the linkage difference of the (n-1)th cycle. The linkage difference is always valid.

4. The battery health estimation method for lithium-sulfur batteries as described in claim 1, characterized in that: The associated product includes: After the discharge is completed, a preset short delay is made to ensure that the piezoresistive sensor has finished reading and the resting phase begins. The battery open circuit voltage changes over time at a preset sampling frequency, with the time origin being the start of the resting phase. The open-circuit voltage sequence is subjected to first-order difference to obtain the voltage change rate curve, and the voltage change rate curve is filtered by moving average with a preset window length to remove high-frequency noise. On the filtered voltage change rate curve, all local minimum points are identified as relaxation troughs. The local minimum point is defined as a point whose value is less than the values ​​of the two adjacent points on the left and right. The time when each relaxation trough occurs is recorded. If there are at least two relaxation troughs, take the earliest relaxation trough and the second earliest relaxation trough, and calculate the time interval between the two relaxation troughs; if there are less than two relaxation troughs, then set the time interval equal to the duration of the entire resting phase. Take the peak width of the second relaxation trough. Using the voltage change rate value of the second relaxation trough as a benchmark, calculate the change rate threshold as the voltage change rate value plus half of the absolute value of the voltage change rate value. Search to the left of the second relaxation trough until the voltage change rate is greater than the change rate threshold for the first time, and search to the right until the voltage change rate is greater than the change rate threshold for the first time. The time length between the left and right boundaries is the peak width. If it is less than two relaxation troughs, set the peak width to zero. Calculate the correlation product for this iteration, where the correlation product is the product of the time interval and the peak width; For the nth cycle and n≥2, calculate the correlation product difference between adjacent cycles, where the correlation product difference is the correlation product of the nth cycle minus the correlation product of the (n-1)th cycle.

5. The battery health estimation method for lithium-sulfur batteries as described in claim 1, characterized in that: A lithium-sulfur battery is deemed to have experienced health degradation when at least two of the first, second, and third differences retain the same sign in three consecutive valid judgment cycles, including: For the nth loop and n≥2, if the first difference is valid, then this loop is a valid determination loop, and a triplet consisting of the first difference, the second difference, and the third difference is obtained; Maintain three first-in-first-out queues, each with a capacity of three; When a valid judgment loop is encountered, the first difference and the second difference are pushed into the corresponding queue. The third difference is pushed in when it is valid and not pushed in when it is invalid. When the queue is full, the oldest value is popped. When a loop with an invalid first difference is encountered, it is skipped and no update is performed. When all three queues are full of three values, each time a new valid judgment loop is obtained, the queues are updated and the following judgment is made: check whether the signs of the three values ​​in each queue are completely the same and all are non-zero. If the signs of at least two queues are completely the same, then a decay is judged, an early warning signal is issued, and the current loop number is recorded.

6. The battery health estimation method for lithium-sulfur batteries as described in claim 5, characterized in that: It also includes a measurement anomaly warning procedure: If no valid judgment cycle is found for a preset number of consecutive cycles, a measurement anomaly warning will be issued, and monitoring of subsequent cycles will continue.

7. The battery health estimation method for a lithium-sulfur battery as described in claim 6, characterized in that: It also includes the recession confirmation step: After issuing the warning signal, continue to monitor subsequent effective judgment cycles; If, within a preset number of valid judgment cycles after the warning is issued, the number of parameters counted as valid is again greater than or equal to two, then the decline is confirmed to be irreversible, and a final alarm is issued.

8. A battery health estimation system for lithium-sulfur batteries, based on the battery health estimation method for lithium-sulfur batteries according to any one of claims 1 to 7, characterized in that: Also includes: The internal resistance slope extraction module measures the battery's ohmic internal resistance in each charge-discharge cycle, calculates the change in ohmic internal resistance in the current cycle relative to the previous cycle, and records it as the falling slope. The piezoresistive difference calculation module reads the resistance value of the piezoresistive sensor on the back of the negative current collector after the completion of charging and discharging in the same cycle, and calculates the linkage difference between the charging resistance and the discharging resistance. The relaxation product acquisition module records the open-circuit voltage and differentially obtains the voltage change rate curve after the discharge is completed. It extracts the time interval between the two relaxation troughs and the peak width of the second trough, and multiplies the time interval and the peak width to obtain the correlation product. The determination module calculates the difference in the downward slope between adjacent cycles, the difference in the linkage difference between adjacent cycles, and the difference in the correlation product between adjacent cycles, which are respectively denoted as the first difference, the second difference, and the third difference. When at least two of the first difference, the second difference, and the third difference remain the same sign in three consecutive valid determination cycles, the health of the lithium-sulfur battery is determined to have deteriorated.

9. A computer device, characterized in that, include: Memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions, which, when executed by the processor, implement the steps of the battery health estimation method for a lithium-sulfur battery according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the battery health estimation method for a lithium-sulfur battery as described in any one of claims 1 to 7.

Images