A performance evaluation system in a lithium battery cathode material repairing process

Abstract

The application discloses a performance evaluation system in a lithium battery positive electrode material repairing process, and belongs to the technical field of electrochemical material repairing and intelligent control. The system comprises an acquisition module which acquires a reference impedance spectrum and a characteristic peak reference position of a positive electrode material sample before repairing; a monitoring module which applies an alternating current impedance detection signal in situ and collects a dynamic impedance spectrum in the whole repairing process, and generates an evolution track of a real-time position of a characteristic peak; and a pre-judgment module which monitors a change trend of the evolution track, and outputs a pre-judgment signal when the change changes from being violent to being gentle and continuously enhancing. The application changes the fixed time length control of the repairing process from experience driving to individualized dynamic end point control of data driving by finding a monotone mapping rule between the crystal structure recovery degree and the characteristic peak frequency of the impedance spectrum, eliminates the thermal damage and information distortion of sampling cooling measurement, and improves the performance consistency of the repaired material and the economy of the repairing process.

Description

Technical Field

[0001] This invention belongs to the field of electrochemical material repair and intelligent control technology, specifically a performance evaluation system for the repair process of lithium battery cathode materials. Background Technology

[0002] After long-term use, the crystal structure of lithium-ion battery cathode materials (such as lithium iron phosphate and ternary materials) degrades to varying degrees, manifesting as blockage of lithium-ion insertion / extraction channels, lattice distortion, and deposition of surface by-products, leading to capacity decay and increased internal resistance. To restore their electrochemical performance, high-temperature solid-state repair processes are commonly used in industry. This involves heating the degraded cathode material to a specific temperature and holding it for a certain period of time in an inert or weakly reducing atmosphere, allowing the crystal structure to regain its order and surface impurities to decompose and volatilize.

[0003] However, current repair processes primarily rely on empirical time-temperature curves, with the repair endpoint determined by a preset holding time, making dynamic adjustments impossible based on the actual degree of crystal structure restoration. This results in several consequences: insufficient repair leads to incomplete crystal structure restoration and limited capacity improvement; excessive repair can cause excessive lattice rearrangement and particle sintering and growth, ultimately worsening electrochemical performance. Furthermore, existing impedance testing methods require removing the sample from the repair equipment and cooling it to room temperature before measurement. This process not only interrupts the repair process but also introduces additional thermal stress damage due to the drastic temperature change, and the measurement results fail to reflect the material's structural state at the actual repair temperature. Summary of the Invention

[0004] The purpose of this invention is to provide a performance evaluation system for the repair process of lithium battery cathode materials, so as to solve the problems mentioned in the background art.

[0005] A performance evaluation system for the repair process of lithium battery cathode materials includes: The acquisition module is used to acquire the reference impedance spectrum of the cathode material sample before repair within a preset frequency range, and extract the reference position of the characteristic peak in the reference impedance spectrum. The characteristic peak is the characteristic frequency response peak in the reference impedance spectrum that migrates to the reference position as the crystal structure of the cathode material is restored. The monitoring module is used to continuously apply an AC impedance detection signal with the same preset frequency range to the cathode material sample in the repair environment during the entire repair process of the cathode material sample, without sampling or shutting down the furnace. It also collects the dynamic impedance spectrum of the cathode material sample in the repair environment in real time, extracts the real-time position of the characteristic peak in the dynamic impedance spectrum, and generates the evolution trajectory of the real-time position over time. The prediction module is used to monitor the changing trend of the real-time position in the evolution trajectory. When the change of the real-time position changes from drastic to gradual and the gradual trend continues to strengthen, a prediction signal that the repair is nearing completion is output. The confirmation module, in response to the start of the prediction signal, is used to successively obtain the deviation between the real-time position and the reference position starting from the moment the prediction signal is output. When the deviation first falls into the preset qualified range, the module outputs a repair termination command and records the current moment as the repair completion moment. The evaluation module is used to determine the degree of matching between the evolutionary trajectory and the preset standard evolutionary trajectory based on the full process data of the evolutionary trajectory from the start of repair to the completion of repair, and output the evaluation result of the repair effect based on the degree of matching.

[0006] In some possible implementations, the monitoring module continuously acquires the dynamic impedance spectrum throughout the entire repair process of the cathode material sample, and the application of the detection signal and the acquisition of the dynamic impedance spectrum are both completed inside the repair equipment where the cathode material sample is located. All data acquisition operations of the monitoring module do not cause the repair equipment to stop or cool down.

[0007] In some possible implementations, the prediction module determines whether the change in the real-time position changes from drastic to gradual, and whether this gradual trend continues to strengthen, in the following manner: Obtain the real-time position of each of the multiple consecutive sampling points in the evolution trajectory, and calculate the position change between adjacent sampling points in sequence; When the position change corresponding to N consecutive sampling points shows a decreasing trend, it is determined that the real-time position change has changed from drastic to gradual. After determining that the change in the real-time position has changed from drastic to gradual, the real-time positions of multiple subsequent sampling points are acquired, and the positional change between each subsequent adjacent sampling point is calculated. If the position change of M consecutive sampling points in the subsequent sampling points is less than the preset stability threshold, then the smooth trend is confirmed to continue, and the prediction signal is output.

[0008] In some possible implementations, the confirmation module determines whether the deviation falls within a preset acceptable range for the first time in the following manner: Starting from the moment the predicted signal is output, the real-time position at each sampling moment is obtained sequentially according to the preset sampling interval, and the deviation between the real-time position at each sampling moment and the reference position is calculated respectively. The deviation at the current sampling time is compared with the upper and lower limits of the acceptable range to determine whether the deviation at the current sampling time is within the acceptable range. If the deviation at the current sampling time is within the acceptable range, then retrieve the deviation at the previous sampling time and determine whether the deviation at the previous sampling time is outside the acceptable range. If the deviation at the previous sampling time is outside the acceptable range, then the current sampling time is determined as the moment when the deviation first falls into the acceptable range, and the repair termination command is output. If the deviation at the previous sampling time is also within the acceptable range, then the deviation at the next sampling time is obtained for judgment until the first sampling time in which the deviation is within the acceptable range and the deviation at the previous sampling time is outside the acceptable range is found.

[0009] In some possible implementations, after determining that the deviation amount falls within the acceptable range for the first time, the confirmation module also performs the following confirmation operation before outputting the repair termination command: Continue to acquire the real-time position at the next K consecutive sampling times, and calculate the deviation between the real-time position at each subsequent sampling time and the reference position; If the deviation at each of the K consecutive sampling times remains within the acceptable range, then the repair termination command is output. If any deviation exceeds the acceptable range during any of the K consecutive sampling times, the repair termination command is revoked, and a restart signal is sent to the prediction module, which then restarts monitoring of the real-time position change trend.

[0010] In some possible implementations, after determining that the change in the real-time position has changed from drastic to gradual and before outputting the prediction signal, the prediction module also performs the following operations: Obtain the complete evolution trajectory from the start of repair to the current moment, and determine the starting moment when the rate of change of the real-time position in the evolution trajectory first shows a continuous downward trend; Calculate the time length from the start time to the current time. If the time length is less than the preset minimum stabilization duration, then the prediction signal will not be output for the time being, and the real-time position of the subsequent sampling points will be obtained for judgment.

[0011] In some possible implementations, after determining that the change in the real-time position has changed from drastic to gradual, the prediction module obtains the real-time positions of multiple sampling points within a preset time period before the current moment and calculates the fluctuation amplitude of the real-time positions of the multiple sampling points. If the fluctuation amplitude is greater than the preset fluctuation threshold, it is determined that there is an abnormal disturbance in the signal at the current moment. The predicted signal is not output for the time being, and the real-time position of the subsequent sampling points is obtained for judgment. If the fluctuation amplitude is less than or equal to the fluctuation threshold, then a confirmation operation is performed to confirm the continued existence of the smooth trend.

[0012] In some possible implementations, before the confirmation module successively acquires the deviation at each sampling time starting from the time the predicted signal is output, it also performs the following operations: Obtain the upper and lower limits of the qualified range; Based on the material type and repair process type of the cathode material sample, the corresponding upper limit adjustment coefficient and lower limit adjustment coefficient are retrieved from the preset parameter configuration table; The upper limit value is multiplied by the upper limit adjustment coefficient to obtain the updated upper limit value, and the lower limit value is multiplied by the lower limit adjustment coefficient to obtain the updated lower limit value. The qualified range is then redefined using the updated upper limit value and the updated lower limit value.

[0013] In some possible implementations, when determining the degree of matching between the evolutionary trajectory and the standard evolutionary trajectory, the evaluation module divides the evolutionary trajectory into multiple temperature segment trajectories according to the repair temperature range, and calculates the segment matching degree between each temperature segment trajectory and the segment trajectory of the corresponding temperature range in the standard evolutionary trajectory. The weighted overall matching degree is calculated as the matching degree based on the duration of each temperature segment trajectory and the segment matching degree. When the segment matching degree of any temperature segment trajectory is lower than the preset segment qualification threshold, the temperature range corresponding to the temperature segment trajectory is marked in the evaluation result, and the repair process adjustment suggestion associated with the temperature range is output.

[0014] In some possible implementations, the system further includes an exception response module, which is used to: During the process of the prediction module monitoring the changing trend of the real-time position, it simultaneously monitors whether the changes in the real-time position in the evolution trajectory exhibit at least one of the following abnormal situations: The real-time location change remains stagnant for a preset duration. The direction of change in the real-time position is opposite to the expected direction of change during the repair process; The real-time positions of multiple consecutive sampling points in the evolution trajectory are obtained, and the position change between adjacent sampling points is calculated. When the position change shows irregular and repeated fluctuations in multiple consecutive sampling intervals, and the position change does not show a trend of continuous decrease over time, it is determined that the change in the real-time position is abnormal. When the abnormal situation is detected, an abnormality warning signal is output to the outside. The abnormality warning signal includes the type identifier of the abnormal situation and the repair temperature and repair pressure information at the current moment. Meanwhile, the abnormal response module suspends the prediction signal output of the prediction module and the start of the confirmation module until the abnormal situation is eliminated and then resumes.

[0015] The technical solutions provided by the embodiments of this disclosure have at least the following beneficial effects: This invention discovers a monotonic mapping law between the degree of crystal structure restoration and the characteristic peak frequency of the impedance spectrum during the repair process of cathode materials. Based on this law, it pioneers a dynamic endpoint control technology for high-temperature in-situ impedance tracking. By directly applying an AC impedance detection signal inside the repair equipment and acquiring the response in situ, the characteristic peak frequency is used as a real-time quantitative indicator of the degree of crystal structure restoration without sampling or furnace shutdown. Based on this, the repair process is dynamically predicted, the repair endpoint is accurately determined, and the repair effect of the entire process is quantitatively evaluated. This transforms the repair process from experience-driven fixed-duration control to data-driven individualized dynamic endpoint control, completely eliminating thermal damage and information distortion caused by sampling and cooling measurements. It ensures sufficient repair while avoiding over-repair, improving the performance consistency of the repair material and the economy of the repair process. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the system framework structure of the present invention. Detailed Implementation

[0017] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] After long-term use, the crystal structure of lithium-ion battery cathode materials (such as lithium iron phosphate and ternary materials) degrades to varying degrees, manifesting as blockage of lithium-ion insertion / extraction channels, lattice distortion, and deposition of surface by-products, leading to capacity decay and increased internal resistance. To restore their electrochemical performance, high-temperature solid-state repair processes are commonly used in industry. This involves heating the degraded cathode material to a specific temperature and holding it for a certain period of time in an inert or weakly reducing atmosphere, allowing the crystal structure to regain its order and surface impurities to decompose and volatilize.

[0019] However, current repair processes primarily rely on empirical time-temperature curves, with the repair endpoint determined by a preset holding time, making dynamic adjustments impossible based on the actual degree of crystal structure restoration. This results in several consequences: insufficient repair leads to incomplete crystal structure restoration and limited capacity improvement; excessive repair can cause excessive lattice rearrangement and particle sintering and growth, ultimately worsening electrochemical performance. Furthermore, existing impedance testing methods require removing the sample from the repair equipment and cooling it to room temperature before measurement. This process not only interrupts the repair process but also introduces additional thermal stress damage due to the drastic temperature change, and the measurement results fail to reflect the material's structural state at the actual repair temperature.

[0020] This invention addresses the aforementioned problems by integrating an online AC impedance detection system within the repair equipment. This system tracks the evolution of characteristic peaks in the impedance spectrum of the cathode material at the repair temperature in real time, without sampling or shutting down the furnace. Based on this, the repair process can be dynamically predicted, the repair endpoint can be accurately determined, and the overall repair effect can be quantitatively evaluated.

[0021] Please see Figure 1 This application provides a performance evaluation system for the repair process of lithium battery cathode materials, including: The acquisition module is used to acquire the reference impedance spectrum of the cathode material sample before repair within a preset frequency range, and extract the reference position of the characteristic peak in the reference impedance spectrum. The characteristic peak is the characteristic frequency response peak in the reference impedance spectrum that migrates to the reference position as the crystal structure of the cathode material is restored.

[0022] In this embodiment, the acquisition module is deployed at the offline measurement station before the repair process starts, and is set separately from the repair equipment body to avoid damage to the precision electrochemical measurement circuit due to the high temperature environment of the repair equipment.

[0023] The cathode material sample was a retired lithium iron phosphate cathode sheet. After disassembly and peeling off the aluminum foil, cathode powder was obtained. The powder was pressed into circular samples with a diameter of 10 mm and a thickness of 1 mm. Both sides of the sample were coated with conductive silver paste as electrodes. The coating method was screen printing, and the silver paste had a solid content of 70%. After coating, it was dried at 120℃ for 30 minutes to form a conductive layer with a thickness of 20 μm.

[0024] The preset frequency range is 10 MHz to 100 kHz. This range is determined based on the following: the 10 MHz low-frequency end can capture the Warburg impedance behavior related to lithium-ion solid-state diffusion; the 100 kHz high-frequency end can cover the capacitive arc characteristics related to charge transfer processes; and the mid-frequency band includes characteristic responses related to processes such as lattice relaxation and interface polarization. For lithium iron phosphate materials, the most sensitive characteristic peak during the repair process is located in the middle region of this frequency band.

[0025] The baseline impedance spectrum was acquired before the remediation process began. Measurements were taken at room temperature (25℃±2℃) using an electrochemical workstation. The AC excitation signal amplitude was 10 mVrms, and the frequency scan range was 10 mHz to 100 kHz, with a logarithmic interval distribution. Ten points were taken every ten octaves, for a total of 60 frequency points across the entire frequency band. After measurement, the impedance spectrum data underwent Kramers-Kronig consistency verification. Subsequent analysis was only permitted after successful verification. The Kramers-Kronig verification pass criterion was: the root mean square error between the converted imaginary part and the measured imaginary part was less than 5% across the entire frequency band.

[0026] Characteristic peaks are identified using an extremum search method based on derivative spectra. The first derivative of the imaginary part of the impedance with respect to frequency is calculated; the zero point where the derivative changes from positive to negative corresponds to the maximum point of the imaginary part of the impedance, i.e., the location of the characteristic peak. In this embodiment, for lithium iron phosphate materials, the reference impedance spectrum exhibits three characteristic peaks in the range of 10 mHz to 100 kHz, located at 0.1 Hz, 10 Hz, and 1 kHz, respectively.

[0027] The characteristic peak at 10 Hz is related to the solid-state diffusion process of lithium ions within the cathode material particles, and its peak frequency corresponds to the lithium ion diffusion coefficient. During the repair process, this characteristic peak migrates to higher frequencies as the degree of crystal structure ordering increases, and the migration amplitude is positively correlated with the degree of repair. Therefore, in this embodiment, the 10 Hz characteristic peak is selected as the tracking target, and its reference position before repair is denoted as f0, in Hz.

[0028] The specific identification criteria for this characteristic peak are as follows: In the derivative spectrum of the reference impedance spectrum, the zero point where the derivative changes from positive to negative within the 5Hz to 15Hz range, and the maximum value of the imaginary part of the impedance corresponding to this zero point is second only to the 0.1Hz main peak across the entire frequency band. The ratio of the maximum value of the imaginary part of the impedance corresponding to this zero point to the amplitude of the 0.1Hz main peak is greater than 0.3 and less than 1.0. A ratio greater than 0.3 ensures that the peak has a sufficient signal-to-noise ratio, avoiding misjudgment of noise spikes; a ratio less than 1.0 ensures that its amplitude is indeed lower than the main peak. If multiple zero points satisfying the conditions exist within the 5Hz to 15Hz range, the one with the highest frequency is selected as the tracking target to avoid confusion with low-frequency Warburg impedance behavior. In this embodiment, the reference position f0 of a certain batch of decommissioned lithium iron phosphate materials was measured to be 12.3Hz.

[0029] It should be noted that the directional basis for the migration to the reference position here is as follows: before repair, the crystal structure of the material degrades, lithium ion diffusion is hindered, the diffusion time constant increases, and the corresponding characteristic peak frequency is low; during the repair process, the crystal structure is restored, the diffusion channel is unblocked, the diffusion time constant decreases, the characteristic peak frequency moves towards the high frequency direction, and gradually approaches the characteristic peak frequency of the new material or the fully repaired material, that is, the reference position.

[0030] Therefore, the physical meaning of the reference position f0 is: the target frequency value that the 10Hz characteristic peak of the cathode material sample should reach after sufficient repair. This value is obtained by impedance spectroscopy measurement of the same batch of new material samples under the same measurement conditions, or determined based on the statistical values ​​of similar materials after sufficient repair in the historical repair database. The method for determining the statistical values ​​of similar materials after sufficient repair in the historical repair database is as follows: retrieve no less than 10 sets of characteristic peak frequency data of similar materials after sufficient repair from the database, remove outliers, and take the arithmetic mean as f0. The outlier removal standard is that the deviation from the mean exceeds 2 times the standard deviation. In this embodiment, the reference position f0 of a certain batch of retired lithium iron phosphate material was measured to be 12.3Hz.

[0031] By limiting the tracking target to the second largest characteristic peak in the 5Hz to 15Hz range, and supplementing it with the priority rule of the highest frequency, the interference of low-frequency Warburg impedance and high-frequency charge transfer process is effectively avoided, ensuring that the core indicator reflecting solid-phase diffusion is always tracked during the repair process, thereby ensuring the pertinence and reliability of subsequent prediction and endpoint determination.

[0032] The prior acquisition of the reference impedance spectrum provides a benchmark for subsequent online monitoring, making it possible to quantitatively track the repair process. Without this benchmark, the migration of characteristic peaks during the repair process would lack a target guidance, making it impossible to determine when the repair is complete.

[0033] The monitoring module is used to continuously apply an AC impedance detection signal with the same preset frequency range to the cathode material sample in the repair environment during the entire repair process of the cathode material sample, without sampling or shutting down the furnace. It also collects the dynamic impedance spectrum of the cathode material sample in the repair environment in real time, extracts the real-time position of the characteristic peak in the dynamic impedance spectrum, and generates the evolution trajectory of the real-time position over time.

[0034] In this embodiment, the repair equipment is a tubular atmosphere furnace with an inner diameter of 60 mm and a constant temperature zone length of 200 mm. The positive electrode material sample is placed in an alumina crucible, which is positioned at the center of the constant temperature zone. The AC impedance detection unit of the monitoring module is integrated inside the furnace structure. The detection electrode is introduced into the furnace through a corundum protective sleeve with an outer diameter of 6 mm and an inner diameter of 2 mm. The electrode lead is a platinum-rhodium alloy wire. The corundum protective sleeve is made of 99% alumina with a long-term operating temperature limit of 1200℃ and a short-term withstand temperature of 1400℃. The selection redundancy is 1.7 times (based on a maximum repair temperature of 700℃) to withstand possible temperature control overshoot (set to not exceed 850℃) and local temperature rise due to Joule heating of the electrode lead during the repair process. The contact method between the detection electrode and the sample is point contact, with the contact pressure provided by a spring mechanism. The spring material is Inconel 718 high-temperature alloy, and the elastic modulus decreases by no more than 15% at 600℃.

[0035] The application and acquisition of the AC impedance detection signal are performed by the same hardware, which includes a signal generator, a potentiostat, and a spectrum analyzer. The signal generator produces an AC excitation signal with an amplitude of 50 mVrms, and the frequency sweep range is consistent with that used in the reference impedance spectrum measurement, i.e., 10 mHz to 100 kHz. The signal amplitude is set to 50 mVrms instead of the 10 mVrms used in room temperature measurements because: at the repair temperature (typically 500℃ to 700℃), the material conductivity increases, and the impedance modulus decreases significantly; therefore, the excitation amplitude is appropriately increased to ensure the signal-to-noise ratio.

[0036] 50 mVrms is still far below the nonlinear threshold that induces the electrochemical reaction. This nonlinear threshold was determined by applying an AC excitation signal with an amplitude gradually increased from 10 mVrms to 500 mVrms to a standard lithium iron phosphate sample at a recovery temperature of 600 °C. The Kramers-Kronig pass rate of the impedance spectrum was monitored. When the excitation amplitude reached 180 mVrms, the pass rate began to fall below 95%, and when it reached 220 mVrms, the pass rate dropped below 80%. The nonlinear threshold was determined to be 200 mVrms (the midpoint between 180 mVrms and 220 mVrms). 50 mVrms is only 25% of this threshold, falling within a clearly defined linear response region.

[0037] The application of the detection signal and the acquisition of the dynamic impedance spectrum are both completed inside the repair equipment containing the cathode material sample. Specifically, the signal is directly applied to the sample surface in the high-temperature repair environment via a platinum-rhodium electrode inside the corundum protective sleeve. The response signal is acquired by the same pair of electrodes, amplified and filtered by the signal conditioning circuit outside the furnace, and then sent to a spectrum analyzer for impedance calculation. The entire process does not require opening the furnace door, removing the sample, or interrupting the heating program, thus completely avoiding interference with the repair process and thermal shock damage to the sample caused by the sampling, cooling, measurement, and reloading procedures.

[0038] By directly introducing the detection electrode into the furnace through the corundum sleeve, the in-situ application of the excitation signal and the in-situ acquisition of the response signal are realized. The sample is always in the repair temperature environment, eliminating the thermal stress damage introduced by the sampling and cooling process, and avoiding the change of crystal structure state caused by drastic temperature changes. This allows the measurement results to truly reflect the actual structural state of the material under repair conditions.

[0039] All data acquisition operations of the monitoring module do not cause shutdown or cooling of the repair equipment. The signal generator output is electrically isolated from the temperature control system of the atmosphere furnace, and the operating status of the impedance detection unit is unaffected by the furnace temperature program; conversely, the rate of temperature rise and fall (typically 5℃ / min) will not trigger the pause or restart of impedance detection. The monitoring module thus achieves true parallel operation with the repair process. The dynamic impedance spectrum acquisition frequency is set to a full-band scan every 10 minutes, with each scan taking 3 minutes. During the scan interval, the detection system is in a low-power standby state, waiting for the next trigger.

[0040] The electrical isolation design between the detection system and the temperature control system ensures that impedance detection and heating procedures do not interfere with each other. The repair process does not need to be interrupted or cooled for measurement, maintaining the continuity of the repair process and the integrity of the thermal history. For repair processes that require precise control of thermal history, this design avoids the adverse effects of temperature fluctuations caused by frequent furnace shutdowns on repair kinetics.

[0041] After the dynamic impedance spectrum acquisition is completed, the extraction of the real-time position of the characteristic peak adopts the same derivative spectrum extremum search method as the reference impedance spectrum. Due to the large baseline drift of the impedance spectrum under high temperature environment, the original impedance imaginary part data is first smoothed by Savitzky-Golay before differentiation. The window width is 11 points and the polynomial order is 3 to suppress the interference of high-frequency noise on the extremum search. Boundary points are processed by the mirror extension method, that is, the first 5 points of the sequence are mirrored with the 6th point as the axis of symmetry, and the last 5 points are mirrored with the 6th point from the end as the axis of symmetry, to ensure that the length of the whole sequence remains unchanged after smoothing. The real-time position of the extracted 10Hz characteristic peak is denoted as ft, in Hz, where t is the sampling time. Connecting the ft of each time step in chronological order forms the evolution trajectory of the real-time position changing with time, denoted as {f(t), t∈[0,T]}, where T is the current cumulative repair time.

[0042] The core innovation of this invention lies in discovering the monotonic mapping law between the degree of crystal structure recovery and the characteristic peak frequency of the impedance spectrum during the repair process of cathode materials, and the measurability of this mapping relationship at the repair temperature. Existing technologies rely on preset time-temperature curves to determine the repair endpoint, implicitly assuming that different samples have the same degree of repair under the same time and temperature conditions. This assumption ignores individual differences in material degradation and reaction rate differences caused by atmospheric fluctuations, leading to widespread problems of insufficient or excessive repair. More importantly, existing impedance detection must be performed at room temperature, and the measurement results reflect the structural state after cooling, not the true state at the repair temperature. Drastic temperature changes themselves can introduce additional structural damage. Based on these physical laws, this invention allows impedance detection to be performed directly at the repair temperature, using the characteristic peak frequency ft as a real-time quantitative indicator of the degree of crystal structure recovery, and dynamically determining the repair progress by tracking the approach of ft to the reference position f0. By using high-temperature in-situ impedance tracking design, the degree of structural recovery of the material under actual repair conditions can be directly read by leveraging the monotonic mapping relationship of characteristic peak frequencies. This avoids thermal damage and information distortion caused by room temperature offline measurements. At the same time, it enables individualized dynamic determination of the repair endpoint, replacing empirical fixed-duration control and improving the accuracy of the repair process and the consistency of material properties.

[0043] The prediction module is used to monitor the real-time position change trend in the evolution trajectory. When the real-time position change changes from drastic to gradual and the gradual trend continues to strengthen, it outputs a prediction signal that the repair is nearing completion.

[0044] In this embodiment, the prediction module runs in the embedded processor of the monitoring module and is connected to the impedance detection hardware via an SPI bus to receive the real-time position ft of the characteristic peak at each sampling time.

[0045] The prediction module determines whether the real-time position change changes from drastic to gradual, and whether this gradual trend continues to strengthen, using the following method: First, the real-time positions of multiple consecutive sampling points in the evolution trajectory are obtained, and the positional changes between adjacent sampling points are calculated sequentially. Before calculating Δfi, the fi sequence is first subjected to median filtering with a window width of 3 points, i.e., fi' is equal to the median of fi-1, fi, and fi+1. Δfi is calculated by replacing fi with fi' to suppress the interference of single-point anomalous jumps on trend determination. Let the sampling point sequence be t1, t2, ..., tn, and the corresponding real-time positions be f1, f2, ..., fn. Then, the positional change Δfi between adjacent sampling points is equal to fi+1 minus fi, in Hz. In this embodiment, the sampling interval is 10 minutes, so Δfi reflects the characteristic peak frequency change within the 10-minute time window.

[0046] When the positional changes corresponding to N consecutive sampling points show a decreasing trend, it is determined that the real-time positional change has changed from drastic to gradual. In this embodiment, N is set to 5, and this value was determined through the following comparative experiment: 10 groups of samples from the same batch were subjected to offline tracking detection. That is, during the repair process, samples were taken every 10 minutes and cooled to room temperature before measuring the impedance spectrum. At the same time, the monitoring module recorded the data synchronously. The offline measurement results were used as the gold standard to determine the actual repair progress and verify the accuracy of online prediction. Prediction accuracy tests were conducted with N=3, 5, and 7 respectively. When N=3, there were two premature predictions due to the low trend threshold. When N=7, there was one delayed prediction due to the long waiting time, and the average waiting time increased by 40 minutes. When N=5, there were no premature predictions and no delayed predictions. The overall response speed and accuracy were optimal, so N=5 was adopted. That is, when the positional changes of 5 consecutive sampling points (corresponding to 50 minutes) Δf1>Δf2>Δf3>Δf4>Δf5, it is determined that the characteristic peak migration rate has entered a decreasing channel, and the repair response has changed from a rapid stage to a slow stabilization stage. The physical basis for this judgment is that in the early stage of repair, the material's crystal structure is highly disordered, the defect density is large, the driving force of the repair reaction is strong, and the characteristic peak frequency changes drastically; as the repair progresses, the defects are gradually eliminated, the driving force of the reaction weakens, and the characteristic peak frequency changes tend to be flat.

[0047] After determining that the change in real-time position has changed from drastic to gradual, the real-time positions of multiple subsequent sampling points are acquired, and the positional change between adjacent sampling points is calculated. If the positional change of M consecutive sampling points in the subsequent sampling points is less than a preset stability threshold, the gradual trend is confirmed to continue, and a prediction signal is output. In this embodiment, M is set to 3. This value was determined through the following comparative experiments: stability was verified by setting M=2, 3, and 5 respectively. When M=2, the confirmation window was too short, resulting in one false triggering of instantaneous fluctuation. When M=5, the confirmation window was too long, resulting in an average delay of 20 minutes in outputting the prediction signal. When M=3, there were no false triggering and the delay was controllable. Therefore, M=3 was adopted. The stabilization threshold was set to 0.5 Hz / 10 min, which is 5% of the typical rate of change (10 Hz / 10 min) in the initial stage of repair. Experiments showed that a threshold of 0.3 Hz was too strict, causing three normal stabilization processes to be judged as unstable; a threshold of 0.8 Hz was too lenient, causing two unstable processes to be misjudged as stable; and a threshold of 0.5 Hz resulted in no missed or false judgments, hence it was adopted. Specifically, when the positional change of three consecutive sampling points (30 minutes) is less than 0.5 Hz, it is confirmed that the characteristic peak has essentially stopped migrating, the repair is nearing completion, and a predictive signal is output.

[0048] The two-level mechanism of determining the decrease in change and confirming the stability threshold has the advantage of filtering out the interference of instantaneous fluctuations through the trend threshold compared to the single threshold determination. It only allows the confirmation process to begin after the repair response has truly entered the deceleration phase, reducing the risk of misjudgment. At the same time, it does not introduce significant response delay, thus achieving a balance between prediction accuracy and timeliness.

[0049] After determining that the change in real-time position has changed from drastic to gradual, and before outputting the prediction signal, the prediction module performs the following operations: Obtain the complete evolution trajectory from the start of the repair process to the current moment, and determine the starting moment when the rate of change of real-time position in the evolution trajectory first shows a continuous downward trend. The method for determining this starting moment is as follows: starting from the starting point of the evolution trajectory, check whether the position change of 5 consecutive sampling points shows a decreasing trend. The starting point of the first 5-point window that meets this condition is the starting moment.

[0050] The time from the start time to the current time is calculated. If the time is less than the preset minimum stabilization duration, the prediction signal is not output temporarily, and the real-time position of subsequent sampling points is acquired for judgment. In this embodiment, the minimum stabilization duration is set to 120 minutes. The minimum stabilization duration is negatively correlated with the repair temperature; for every 50°C decrease in temperature, this time increases by 40%. It is recommended to use 168 minutes at 550°C and 86 minutes at 650°C. This conversion relationship is obtained by linear fitting of three sets of temperature gradient experiments (5 sets of samples each at 550°C, 600°C, and 650°C), with a fitting determination coefficient R² = 0.94. The parameter was determined based on the following: five groups of lithium iron phosphate samples were repaired at 600℃, and the continued change of their characteristic peak frequencies after stabilization was monitored. The results showed that after stabilization, the characteristic peak frequencies continued to change by an average of 0.8 Hz within 60 minutes (accounting for 12% of the total), an average of 0.3 Hz within 90 minutes (accounting for 4% of the total), and an average change of less than 0.1 Hz after 120 minutes (accounting for less than 1.5% of the total, within the measurement error range). Therefore, 120 minutes is a sufficient duration to cover the residual relaxation process; a duration shorter than this carries the risk of incomplete repair.

[0051] The introduction of the shortest stabilization duration allows sufficient time for the slow defect relaxation and lattice fine-tuning processes within the material, avoiding incomplete repair caused by premature termination of repair due to surface stabilization, and improving the sufficiency of repair.

[0052] After determining that the change in real-time position has changed from drastic to gradual, the prediction module obtains the real-time positions of multiple sampling points within a preset time period before the current moment and calculates the fluctuation amplitude of the real-time positions of the multiple sampling points.

[0053] In this embodiment, the preset time period is 60 minutes prior to the current moment, i.e., 6 sampling points. The fluctuation amplitude is calculated as the maximum value minus the minimum value of the real-time position among these 6 sampling points. If the fluctuation amplitude is greater than the preset fluctuation threshold, it is determined that the signal at the current moment has an abnormal disturbance, the prediction signal is not output temporarily, and the real-time position of subsequent sampling points is continuously acquired for judgment. In this embodiment, the fluctuation threshold is set to 2.0Hz. The determination of this value is based on the following: statistics were performed on 10 groups of samples without external disturbances during normal repair, and the maximum fluctuation amplitude within the 60-minute window was 1.2Hz; statistics were performed on 5 groups of samples with external disturbances (atmosphere flow ±20% fluctuation, temperature control ±5℃ overshoot), and the minimum fluctuation amplitude was 2.5Hz. The median value of the two sets of statistical values, 2.0Hz, is taken as the threshold, which can effectively distinguish between normal fluctuations and abnormal disturbances. It has been verified that under this threshold, there are no misjudgments of normal as abnormal and no omissions of abnormal as normal. During initial system deployment, the system is initialized with a default threshold of 1.5Hz. For the first five furnace repair processes, normal fluctuation data is automatically collected. From the sixth furnace onwards, an adaptive threshold is used: the larger of 1.2 times the maximum normal fluctuation value of the first five furnaces and the default threshold of 1.5Hz. Abnormal disturbance data must be collected by artificially introducing standard disturbances (atmosphere flow ±20%, temperature control ±5℃) during the process debugging phase. If the fluctuation amplitude is less than or equal to 2.0Hz, the aforementioned confirmation operation of the continued existence of a smooth trend is performed.

[0054] The fluctuation amplitude check adds an environmental adaptability barrier to process prediction by quantifying the statistical difference between normal process fluctuations and external disturbances. This mechanism enables the system to remain robust in complex industrial environments such as atmosphere flow pulsation and temperature control overshoot, reducing the risk of misjudgment caused by abnormal disturbances.

[0055] The confirmation module, in response to the start of the prediction signal, is used to successively obtain the deviation between the real-time position and the reference position starting from the moment the prediction signal is output. When the deviation first falls into the preset qualified range, the module outputs a repair termination command and records the current moment as the repair completion moment.

[0056] In this embodiment, the deviation is defined as the absolute difference between the real-time position ft of the characteristic peak at the current moment and the reference position f0, i.e., |ft-f0|, in Hz.

[0057] Specifically, the acceptable range is defined as the closed interval formed by the deviation being less than or equal to the preset upper limit and greater than or equal to the preset lower limit. In this embodiment, the upper limit of the acceptable range is set to 2.0Hz, and the lower limit is set to -0.5Hz. The determination of this range is based on the following: statistical analysis of 20 fully repaired samples from the same batch shows that the deviation distribution of their characteristic peak frequencies from the reference position f0 has a mean of 0.3Hz and a standard deviation of 0.6Hz. The mean plus three times the standard deviation (2.1Hz) is rounded to the upper limit of 2.0Hz, and the mean minus one time the standard deviation (-0.3Hz) is rounded to the lower limit of -0.5Hz to account for a slight lack of tolerance. The 20 samples are taken from retired cathode materials from the same production batch and included in the statistics after offline full repair verification.

[0058] If a material batch is changed, the deviation distribution of the fully repaired samples in that batch must be re-measured, or a coefficient correction method can be used: the measured f0 value of the current batch is compared with the historical average f0 value in the database, and recalibration is triggered when the deviation exceeds 10%. Thus, the upper limit of 2.0Hz covers the normal fluctuation range of 99.7% of the fully repaired samples, while the lower limit of -0.5Hz allows for slight under-repair while preventing over-repair. A positive upper limit means that the characteristic peak frequency is allowed to be slightly lower than the reference position (i.e., slightly under-repaired), while a negative lower limit means that the characteristic peak frequency is allowed to be slightly higher than the reference position (i.e., slightly over-repaired). The basis for this asymmetric design is that when lithium iron phosphate material is slightly under-repaired, the crystal structure can continue to recover slowly during subsequent charge-discharge cycles, with limited impact on long-term performance; while over-repair leads to particle sintering and excessive lattice rearrangement, causing irreversible damage. Therefore, the acceptable range is appropriately widened towards slight under-repair and strictly tightened towards over-repair.

[0059] The statistical method for determining the acceptable range ensures that the endpoint judgment criteria are based on actual fluctuation data of fully repaired samples, rather than subjective experience values, thus improving the objectivity and repeatability of the judgment criteria. The asymmetric design utilizes the material properties of lithium iron phosphate, which allows for slight under-repair but is difficult to reverse, minimizing the risk of irreversible damage caused by over-repair while ensuring sufficient repair.

[0060] The confirmation module determines whether the deviation falls within the preset acceptable range for the first time using the following method: Starting from the moment the prediction signal is output, the real-time position at each sampling moment is acquired sequentially according to a preset sampling interval, and the deviation between the real-time position and the reference position at each sampling moment is calculated. In this embodiment, the sampling interval after the prediction signal is output is shortened to 5 minutes to improve the timeliness of endpoint determination.

[0061] The deviation at the current sampling time is compared with the upper limit of 2.0Hz and the lower limit of -0.5Hz of the acceptable range to determine whether the deviation at the current sampling time is within the acceptable range.

[0062] If the deviation at the current sampling time is within the acceptable range, the deviation at the previous sampling time is retrieved to determine if it is outside the acceptable range. If the deviation at the previous sampling time is outside the acceptable range, the current sampling time is determined as the moment when the deviation first falls into the acceptable range, a repair termination command is output, and the current time is recorded as the repair completion time.

[0063] If the deviation at the previous sampling time is also within the acceptable range, the deviation at the next sampling time is obtained for judgment until the first sampling time where the deviation is within the acceptable range and the deviation at the previous sampling time is outside the acceptable range is found. The purpose of this first-fall-in judgment mechanism is to ensure that the repair termination command is triggered only at the critical moment when the characteristic peak frequency enters the acceptable range from outside the acceptable range, avoiding repeated output of the termination command due to continuous being within the acceptable range.

[0064] The first-time fall-in mechanism accurately captures the critical point of qualitative change from never meeting the standard to just meeting the standard by tracing the state of the previous moment. Compared with the simple pass-and-stop strategy, it avoids multiple triggers caused by fluctuations in the pass zone, ensures the uniqueness and certainty of the termination command, and prevents the repair equipment from being misoperated due to repeated triggers.

[0065] Before the confirmation module acquires the deviation at each sampling time starting from the moment the predicted signal is output, it also performs the following operations: Obtain the upper limit of the acceptable range: 2.0Hz and the lower limit: -0.5Hz.

[0066] Based on the material type and repair process type of the cathode material sample, the corresponding upper and lower limit adjustment coefficients are retrieved from a preset parameter configuration table. In this embodiment, the parameter configuration table is stored in a two-dimensional matrix, with row indices representing material types (lithium iron phosphate, lithium nickel cobalt manganese oxide, lithium cobalt oxide, etc.) and column indices representing repair process types (solid-phase repair, liquid-phase repair, gas-phase repair, etc.). The parameter configuration table is stored as an SQLite database file in the local flash memory of the embedded processor, with the following table structure: material_type (TEXT, primary key), process_type (TEXT, primary key), upper_coeff (REAL), lower_coeff (REAL), and update_timestamp (INTEGER). Updates are performed using transactional writes; if a write fails, the original value is retained and an error log is recorded. For the solid-phase repair process of lithium iron phosphate materials, the upper limit adjustment coefficient is 1.0 and the lower limit adjustment coefficient is 1.0, meaning the qualified range remains unchanged. For the solid-phase repair process of lithium nickel cobalt manganese oxide materials, the upper limit adjustment coefficient is 1.2 and the lower limit adjustment coefficient is 0.8, meaning the qualified range is widened to [-0.4Hz, 2.4Hz]. This is because the crystal structure of ternary materials is more complex, and the characteristic peak frequency fluctuation range near the repair endpoint is larger.

[0067] The updated upper limit value is obtained by multiplying the upper limit value by the upper limit adjustment factor, and the updated lower limit value is obtained by multiplying the lower limit value by the lower limit adjustment factor. The qualified range is then redefined using the updated upper limit value and the updated lower limit value.

[0068] The two-dimensional matrix design of the parameter configuration table allows the same endpoint determination logic to adapt to the technical characteristics of different material systems and repair processes, eliminating the need to develop separate determination programs for each material. This unified framework plus parameter adaptation mode reduces the engineering complexity of the system when deployed on multiple material production lines and shortens the debugging cycle during production line switchover.

[0069] After the verification module determines that the deviation has fallen into the acceptable range for the first time, it also performs the following verification operation before outputting the repair termination command: The system continues to acquire the real-time position for the next K consecutive sampling times, and calculates the deviation between the real-time position and the reference position for each subsequent sampling time. In this embodiment, K is set to 3, meaning the sampling interval after the prediction signal output is 5 minutes, so 3 consecutive sampling times correspond to 15 minutes. The value of K is tied to the length of the confirmation time window rather than the number of sampling points, and the confirmation time window is fixed at 15 minutes. If the sampling interval is adjusted to ts minutes, K automatically takes the value of 15 / ts rounded up to ensure a constant confirmation time.

[0070] If the deviation remains within the acceptable range for K consecutive sampling times, a repair termination command is output. This confirmation operation prevents false triggering caused by momentary fluctuations in the characteristic peak frequency just crossing the acceptable range boundary, ensuring that the repair termination command is only output after the characteristic peak has stably entered the acceptable region.

[0071] If any deviation falls outside the acceptable range within K consecutive sampling times, the repair termination command is revoked, and a restart signal is sent to the prediction module, which then resumes monitoring the real-time position change trend. This rollback mechanism allows the system to automatically resume monitoring when the characteristic peak frequency rebounds, rather than terminating the repair at an incorrect moment.

[0072] The K-times confirmation rollback mechanism essentially adds a safety net to the endpoint determination. Even if the characteristic peak frequency happens to fall within the acceptable range due to some transient factor, the system will not immediately terminate the repair, but will wait for a period of time to observe its stability. This strategy of first qualifying, then confirming, and then reverting if unstable minimizes the risk of misjudgment and termination. For long-cycle, high-energy-consuming repair processes, it avoids incomplete repair and energy waste caused by premature termination.

[0073] The evaluation module is used to determine the degree of matching between the evolutionary trajectory and the preset standard evolutionary trajectory based on the data of the entire process from the start to the completion of the repair, and output the evaluation results of the repair effect based on the degree of matching.

[0074] In this embodiment, the standard evolution trajectory is derived from the impedance tracking database of new material samples from the same batch or historical fully repaired samples. The data is in the form of a standard curve fstd(t) showing the change of characteristic peak frequency with repair time. The sampling interval is consistent with the monitoring module and is 10 minutes.

[0075] The specific method for constructing the standard evolution trajectory is as follows: First, select no fewer than 15 groups of new materials from the same batch or similar decommissioned materials that have undergone sufficient offline repair verification as standard samples. Under the same repair process conditions, perform complete online impedance tracing to obtain the original curves of the characteristic peak frequencies of each standard sample changing over time. Second, align the time axes of each original curve. The alignment anchor point is the end of the low-temperature activation stage, i.e., the moment when the furnace temperature transitions from the 300℃ holding period to the reheating stage. This moment is automatically identified by the inflection point where the first derivative of the furnace temperature-time curve changes from zero to positive. Using this anchor point as the zero point of time, shift the time axes of each original curve so that the anchor points of all curves coincide.

[0076] Then, the aligned curves are fused using a point-by-point arithmetic mean method. For each sampling time point, the arithmetic mean of the characteristic peak frequencies of all standard samples at that time is calculated to form the fused standard curve. If a standard sample is missing data at a specific time due to anomalies, only the valid samples at that time are averaged, with a minimum of 10 valid samples. Finally, the fused standard curve is smoothed using cubic spline interpolation with a 10-minute interval between interpolation nodes, resulting in a continuous and smooth standard evolution trajectory fstd(t). The standard evolution trajectory is updated and validated quarterly using the production data from the most recent three months as samples. If the weighted average matching degree between the new sample and the current standard trajectory is lower than 0.90, the standard trajectory is recalibrated, following the same recalibration process as described above.

[0077] When determining the degree of matching between the evolutionary trajectory and the standard evolutionary trajectory, the evaluation module divides the evolutionary trajectory into multiple temperature segment trajectories according to the repair temperature range. In this embodiment, the repair temperature program is as follows: room temperature to 300°C is the heating stage, 300°C for 30 minutes is the low-temperature activation stage, 300°C to 600°C is the reheating stage, and 600°C for repair completion is the high-temperature repair stage. Accordingly, the evolutionary trajectory is divided into four temperature segment trajectories: the heating stage trajectory, the low-temperature activation stage trajectory, the reheating stage trajectory, and the high-temperature repair stage trajectory.

[0078] The segment matching degree between each temperature segment trajectory and the corresponding temperature range segment trajectory in the standard evolution trajectory is calculated. The matching degree is calculated using the Dynamic Time Warping (DTW) algorithm, which aligns the key nodes of characteristic peak frequency changes in the two trajectories through nonlinear time axis scaling, and calculates the root mean square error after alignment as the segment matching degree. The automatic identification method for key nodes is as follows: The first derivative of the temperature-time curve is calculated, and the point where the derivative changes from positive to negative (the inflection point where the temperature changes from rising to flat) is marked as the end of the heating segment / the beginning of the holding segment. Before the derivative calculation, the temperature data is smoothed using Savitzky-Golay, with a window width of 5 points and a polynomial order of 2. The zero-point threshold is set when the absolute value of the derivative is less than 0.1℃ / min. The first derivative of the characteristic peak frequency-time curve is calculated, and the abrupt change point where the derivative changes from large to small (the point where the rate of frequency change drops sharply) is marked as the acceleration point of the repair reaction. Before the derivative calculation, the frequency data is smoothed using Savitzky-Golay, with a window width of 7 points and a polynomial order of 3. The abrupt change point is determined when the derivative values ​​of three consecutive sampling points are all less than 50% of the minimum derivative value of the previous three sampling points. The DTW algorithm uses the automatically identified key nodes as anchor points to perform piecewise time axis scaling and alignment, rather than global uniform scaling, thereby accurately matching the heating lag or reaction start-up delay caused by differences in thermal inertia in different samples. The segment matching degree ranges from 0 to 1, with a value closer to 1 indicating a better match.

[0079] Based on the duration and matching degree of each temperature segment trajectory, a weighted overall matching degree is calculated as the degree of matching. The weighting formula is: the overall matching degree equals the matching degree of each segment multiplied by the proportion of the duration of that segment to the total time of the entire restoration process, and then all are added together. That is, the longer the duration of a segment, the greater its contribution to the overall result, which is consistent with the physical fact that in actual processes, the long-term heat preservation period plays a dominant role in the restoration effect.

[0080] When the segment matching degree of any temperature segment trajectory is lower than the preset segment qualification threshold, the temperature range corresponding to that temperature segment trajectory is marked in the evaluation results, and the repair process adjustment suggestions associated with that temperature range are output. In this embodiment, the segment qualification threshold is set to 0.85. When the matching degree of the heating segment is lower than 0.85, it is recommended to reduce the heating rate from 5℃ / min to 3℃ / min and add a 30-minute intermediate heat preservation platform at 400℃; when the matching degree of the low-temperature activation segment is lower than 0.85, it is recommended to adjust the heat preservation temperature within ±20℃, step by step in 10℃ for verification, or extend or shorten the heat preservation time by ±15 minutes; when the matching degree of the reheating segment is lower than 0.85, it is recommended to reduce the heating rate from 5℃ / min to 3℃ / min, or add a 20-minute intermediate heat preservation platform at 450℃; when the matching degree of the high-temperature repair segment is lower than 0.85, it is recommended to adjust the heat preservation temperature within ±20℃, step by step in 10℃ for verification, or extend / shorten the heat preservation time by ±30 minutes.

[0081] If the segment matching degree of the heating section is less than 0.85, it indicates that the heating rate may be too fast, and it is recommended to reduce the heating rate or add an intermediate insulation platform; if the segment matching degree of the low-temperature activation section is less than 0.85, it indicates that the low-temperature activation temperature or time may deviate from the optimal value, and it is recommended to adjust the insulation temperature or time; if the segment matching degree of the reheating section is less than 0.85, it indicates that the material's thermal response is abnormal during the reheating process, and it is recommended to reduce the heating rate or add an intermediate insulation platform; if the segment matching degree of the high-temperature repair section is less than 0.85, it indicates that the insulation temperature may deviate from the optimal value, and it is recommended to adjust the target temperature or extend / shorten the insulation time.

[0082] The independent evaluation of segmented matching degree and the linked output of process adjustment suggestions enable the effect assessment to go beyond a general judgment of good or bad, and instead pinpoint the deviation of specific process segments, providing directly actionable guidance for process optimization. This diagnostic and prescription-based approach shortens the feedback cycle of process optimization from several days for traditional batch testing, problem location, and parameter adjustment to immediate output after a single fix, thus improving the efficiency of process iteration.

[0083] The system also includes an anomaly response module, which is used to simultaneously monitor whether the changes in real-time position in the evolution trajectory exhibit at least one of the following abnormal situations during the process of the prediction module monitoring the trend of real-time position changes: The real-time location changes remain stagnant for a preset duration; In this embodiment, the preset duration is set to 60 minutes. A standstill is defined as a state where the absolute value of the characteristic peak frequency change at six consecutive sampling points (with a 10-minute sampling interval) is less than 0.1 Hz. Possible causes for this include: poor contact of the detection electrode leading to signal interruption, severe atmosphere leakage causing termination of the repair reaction, and sample sintering deformation causing electrode detachment.

[0084] The direction of change in the real-time location is opposite to the expected direction of change during the repair process; For lithium iron phosphate materials, the expected trend is a shift in characteristic peak frequency towards higher frequencies (i.e., ft increases over time). If the position changes of three consecutive sampling points are all negative (i.e., ft continuously decreases), the trend is considered abnormal. Possible causes for this include: the atmosphere mistakenly changing from reducing to oxidizing, leading to further material degradation; or a malfunction in the temperature control system causing a drop in the actual temperature.

[0085] The real-time positions of multiple consecutive sampling points in the evolution trajectory are obtained, and the position change between adjacent sampling points is calculated. When the position change shows irregular and repeated fluctuations in multiple consecutive sampling intervals, and the position change does not show a trend of continuous decrease over time, it is determined that there is an anomaly in the real-time position change. In this embodiment, 10 sampling points (100 minutes) are taken in multiple consecutive sampling intervals. If the signs of the changes in these 10 locations alternate more than 5 times and their absolute values ​​do not show a monotonically decreasing trend, it is determined to be an irregular fluctuation anomaly. Possible causes of this situation include: atmosphere flow pulsation, temperature control system oscillation, external electromagnetic interference, etc.

[0086] When any of the abnormal conditions is detected, the abnormal response module outputs an abnormality alert signal to the outside. The data format of the abnormality alert signal is JSON, containing the following fields: abnormality type identifier (STALL, REVERSAL, FLUCTUATION correspond to the three abnormal conditions respectively), current repair temperature (read by the atmosphere furnace temperature control system), and current repair pressure (read by the furnace pressure sensor). The abnormality alert signal is sent to the operation and maintenance monitoring platform via the MQTT protocol. The MQTT protocol configuration parameters are as follows: broker address is the operation and maintenance monitoring platform IP (default 192.168.1.100), port 1883, topic structure is furnace / fault / {deviceID} / {faultType}, where deviceID is the atmosphere furnace number and faultType is the abnormality type identifier; QoS level is 1 (at least once), message retention flag is false, connection timeout is 30 seconds, and heartbeat interval is 60 seconds. The abnormality alert signal is sent to the operation and maintenance monitoring platform via the MQTT protocol, and a local audible and visual alarm is triggered simultaneously.

[0087] Simultaneously, the anomaly response module suspends the prediction signal output of the prediction module and the startup of the confirmation module until the anomaly is eliminated. The specific implementation of the suspension mechanism is as follows: the anomaly response module sends a pause flag to both the prediction and confirmation modules. When both modules detect that the flag is valid, they freeze their current state and stop outputting any signals. After the anomaly is eliminated (either through manual confirmation by maintenance personnel or automatic detection by the system that the anomaly indicators have returned to normal), the anomaly response module clears the pause flag, the prediction module continues monitoring from the current sampling point, and the confirmation module remains in standby mode until a new prediction signal arrives.

[0088] The automatic judgment criteria for eliminating abnormal situations are as follows: For stagnation anomalies, the absolute value of the position change at three consecutive sampling points (30 minutes) is greater than 0.2Hz and shows a monotonically increasing or decreasing trend; for change direction anomalies, the position change at three consecutive sampling points (30 minutes) is positive (restoring the expected high-frequency migration direction); for irregular fluctuation anomalies, the fluctuation amplitude at six consecutive sampling points (60 minutes) is less than 1.5Hz and the absolute value of the position change shows a monotonically decreasing trend. The automatic clearing condition is that the recovery criteria are met continuously for more than twice the anomaly judgment window duration, i.e., stagnation anomalies require 60 consecutive minutes of normal operation, change direction anomalies require 60 consecutive minutes of normal operation, and fluctuation anomalies require 120 consecutive minutes of normal operation. This double redundancy design prevents frequent pauses and recovery oscillations caused by repeated anomaly fluctuations. After meeting the above corresponding criteria, the system automatically clears the pause flag without manual confirmation; if the automatic recovery criteria are not met within 30 minutes, the system prompts maintenance personnel for manual confirmation.

[0089] The monitoring, alarm, pause, and recovery closed-loop design of the anomaly response module allows the system to intervene promptly when unexpected situations occur during the repair process, preventing misinterpretations of signals from leading to erroneous process operations. More importantly, the pause rather than termination approach preserves the continuity of the repair process. After the anomaly is cleared, the system can resume from the breakpoint without restarting the entire repair procedure, thus avoiding the scrapping of the entire furnace of materials and energy waste caused by the anomaly.

[0090] As a specific implementation method, the solid-phase repair process of a batch of retired lithium iron phosphate cathode materials is taken as an example. The initial capacity retention rate of this batch of materials is 68% of the rated capacity. After disassembly, representative samples were prepared into disc samples. Before repair, the reference impedance spectrum was measured at 25℃, and the reference position f0 of the 10Hz characteristic peak was 12.3Hz. The repair process is as follows: under an argon atmosphere, the temperature is increased to 300℃ at 5℃ / min, held for 30 minutes, and then increased to 600℃ at 5℃ / min to enter the high-temperature repair stage.

[0091] The monitoring module collects dynamic impedance spectra every 10 minutes from the start of heating. Repair started at 10:00, and the first acquisition was at 10:10. The real-time position of the characteristic peak, f1, was 3.2Hz (far lower than f0, reflecting severe material degradation). From 10:20 to 11:00, the characteristic peak frequency rapidly increased from 3.2Hz to 8.7Hz, with position changes of 1.8Hz, 1.5Hz, 1.2Hz, 0.9Hz, and 0.6Hz respectively. These five consecutive sampling points showed a decreasing trend, and the prediction module determined at 11:00 that the change had shifted from drastic to gradual. From 11:00 to 11:30, the subsequent three position changes were 0.4Hz, 0.3Hz, and 0.2Hz, all less than the stability threshold of 0.5Hz. Furthermore, the fluctuation amplitude over the previous 60 minutes was 1.5Hz (less than the 2.0Hz threshold), and the shortest stabilization period exceeded 120 minutes (starting from 10:20). The prediction module output a prediction signal at 11:30.

[0092] The confirmation module started at 11:30, and the sampling interval was shortened to 5 minutes. At 11:30, the deviation |8.7-12.3|=3.6Hz, exceeding the acceptable range [-0.5, 2.0]Hz. At 11:35, the deviation was 2.8Hz, still exceeding the acceptable range. At 11:40, the deviation was 1.9Hz, falling into the acceptable range, and the deviation of 2.8Hz at the previous moment (11:35) was outside the range, satisfying the initial falling-in condition. The module continued to acquire three more sampling points: 11:45 deviation of 1.6Hz, 11:50 deviation of 1.4Hz, and 11:55 deviation of 1.2Hz, all remaining within the acceptable range. The confirmation module output a repair termination command at 11:55, marking the repair completion time as 11:55, with a cumulative high-temperature repair time of 115 minutes.

[0093] The evaluation module compared the entire evolution trajectory with a standard evolution trajectory. The matching degrees for the four temperature segments were: heating segment 0.91, low-temperature activation segment 0.88, reheating segment 0.93, and high-temperature repair segment 0.95, with a weighted overall matching degree of 0.93. The evaluation results indicate good repair effect, with the highest matching degree in the high-temperature repair segment, suggesting that the repair kinetics of this batch of materials at 600℃ are consistent with the standard sample.

[0094] After being cooled and assembled into coin cells, the repaired sample was tested and its capacity was restored to 94% of the rated capacity, while its internal resistance was reduced to 42% of the value before repair, verifying the consistency between the online evaluation results and the offline electrochemical performance.

[0095] Therefore, this invention constructs a complete performance evaluation system for the repair of lithium battery cathode materials by establishing a reference impedance spectrum in the acquisition module, tracking high-temperature in-situ impedance in the monitoring module, determining two-stage stabilization in the prediction module, confirming the initial entry and exit points in the confirmation module, evaluating the segmented matching in the assessment module, and monitoring the entire process for safety in the anomaly response module. The collaborative work of these modules transforms the repair process from experience-driven, fixed-duration control to data-driven, dynamic endpoint control. This ensures sufficient repair while avoiding over-repair, and the accumulation of data throughout the process provides a quantitative basis for process optimization, thereby improving the performance consistency of the repair materials and the economic efficiency of the repair process.

[0096] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A performance evaluation system for the repair process of lithium battery cathode materials, characterized in that, The system includes: The acquisition module is used to acquire the reference impedance spectrum of the cathode material sample before repair within a preset frequency range, and extract the reference position of the characteristic peak in the reference impedance spectrum. The characteristic peak is the characteristic frequency response peak in the reference impedance spectrum that migrates to the reference position as the crystal structure of the cathode material is restored. The monitoring module is used to continuously apply an AC impedance detection signal with the same preset frequency range to the cathode material sample in the repair environment during the entire repair process of the cathode material sample, without sampling or shutting down the furnace. It also collects the dynamic impedance spectrum of the cathode material sample in the repair environment in real time, extracts the real-time position of the characteristic peak in the dynamic impedance spectrum, and generates the evolution trajectory of the real-time position over time. The prediction module is used to monitor the changing trend of the real-time position in the evolution trajectory. When the change of the real-time position changes from drastic to gradual and the gradual trend continues to strengthen, a prediction signal that the repair is nearing completion is output. The confirmation module, in response to the start of the prediction signal, is used to successively obtain the deviation between the real-time position and the reference position starting from the moment the prediction signal is output. When the deviation first falls into the preset qualified range, the module outputs a repair termination command and records the current moment as the repair completion moment. The evaluation module is used to determine the degree of matching between the evolutionary trajectory and the preset standard evolutionary trajectory based on the full process data of the evolutionary trajectory from the start of repair to the completion of repair, and output the evaluation result of the repair effect based on the degree of matching.

2. The performance evaluation system for the lithium battery cathode material repair process according to claim 1, characterized in that, The monitoring module continuously acquires the dynamic impedance spectrum throughout the entire repair process of the cathode material sample. The application of the detection signal and the acquisition of the dynamic impedance spectrum are both completed inside the repair equipment where the cathode material sample is located. All data acquisition operations of the monitoring module do not cause the repair equipment to stop or cool down.

3. The performance evaluation system for the lithium battery cathode material repair process according to claim 1, characterized in that, The prediction module determines whether the change in the real-time position changes from drastic to gradual, and whether this gradual trend continues to strengthen, in the following manner: Obtain the real-time position of each of the multiple consecutive sampling points in the evolution trajectory, and calculate the position change between adjacent sampling points in sequence; When the position change corresponding to N consecutive sampling points shows a decreasing trend, it is determined that the real-time position change has changed from drastic to gradual. After determining that the change in the real-time position has changed from drastic to gradual, the real-time positions of multiple subsequent sampling points are acquired, and the positional change between each subsequent adjacent sampling point is calculated. If the position change of M consecutive sampling points in the subsequent sampling points is less than the preset stability threshold, then the smooth trend is confirmed to continue, and the prediction signal is output.

4. The performance evaluation system for the lithium battery cathode material repair process according to claim 1, characterized in that, The confirmation module determines whether the deviation falls within the preset acceptable range for the first time in the following manner: Starting from the moment the predicted signal is output, the real-time position at each sampling moment is obtained sequentially according to the preset sampling interval, and the deviation between the real-time position at each sampling moment and the reference position is calculated respectively. The deviation at the current sampling time is compared with the upper and lower limits of the acceptable range to determine whether the deviation at the current sampling time is within the acceptable range. If the deviation at the current sampling time is within the acceptable range, then retrieve the deviation at the previous sampling time and determine whether the deviation at the previous sampling time is outside the acceptable range. If the deviation at the previous sampling time is outside the acceptable range, then the current sampling time is determined as the moment when the deviation first falls into the acceptable range, and the repair termination command is output. If the deviation at the previous sampling time is also within the acceptable range, then the deviation at the next sampling time is obtained for judgment until the first sampling time in which the deviation is within the acceptable range and the deviation at the previous sampling time is outside the acceptable range is found.

5. The performance evaluation system for the lithium battery cathode material repair process according to claim 4, characterized in that, After determining that the deviation amount falls into the acceptable range for the first time, the confirmation module performs the following confirmation operation before outputting the repair termination command: Continue to acquire the real-time position at the next K consecutive sampling times, and calculate the deviation between the real-time position at each subsequent sampling time and the reference position; If the deviation at each of the K consecutive sampling times remains within the acceptable range, then the repair termination command is output. If any deviation exceeds the acceptable range during any of the K consecutive sampling times, the repair termination command is revoked, and a restart signal is sent to the prediction module, which then restarts monitoring of the real-time position change trend.

6. The performance evaluation system for the lithium battery cathode material repair process according to claim 3, characterized in that, After determining that the change in the real-time position has changed from drastic to gradual, and before outputting the prediction signal, the prediction module also performs the following operations: Obtain the complete evolution trajectory from the start of repair to the current moment, and determine the starting moment when the rate of change of the real-time position in the evolution trajectory first shows a continuous downward trend; Calculate the time length from the start time to the current time. If the time length is less than the preset minimum stabilization duration, then the prediction signal will not be output for the time being, and the real-time position of the subsequent sampling points will be obtained for judgment.

7. The performance evaluation system for the lithium battery cathode material repair process according to claim 3, characterized in that, After determining that the change in the real-time position has changed from drastic to gradual, the prediction module obtains the real-time positions of multiple sampling points within a preset time period before the current moment and calculates the fluctuation amplitude of the real-time positions of the multiple sampling points. If the fluctuation amplitude is greater than the preset fluctuation threshold, it is determined that there is an abnormal disturbance in the signal at the current moment. The predicted signal is not output for the time being, and the real-time position of the subsequent sampling points is obtained for judgment. If the fluctuation amplitude is less than or equal to the fluctuation threshold, then a confirmation operation is performed to confirm the continued existence of the smooth trend.

8. The performance evaluation system for the lithium battery cathode material repair process according to claim 4, characterized in that, Before the confirmation module acquires the deviation at each sampling time sequentially starting from the time the predicted signal is output, it also performs the following operations: Obtain the upper and lower limits of the qualified range; Based on the material type and repair process type of the cathode material sample, the corresponding upper limit adjustment coefficient and lower limit adjustment coefficient are retrieved from the preset parameter configuration table; The upper limit value is multiplied by the upper limit adjustment coefficient to obtain the updated upper limit value, and the lower limit value is multiplied by the lower limit adjustment coefficient to obtain the updated lower limit value. The qualified range is then redefined using the updated upper limit value and the updated lower limit value.

9. The performance evaluation system for the repair process of lithium battery cathode materials according to claim 1, characterized in that, When determining the degree of matching between the evolutionary trajectory and the standard evolutionary trajectory, the evaluation module divides the evolutionary trajectory into multiple temperature segment trajectories according to the repair temperature range, and calculates the segment matching degree between each temperature segment trajectory and the segment trajectory of the corresponding temperature range in the standard evolutionary trajectory. The weighted overall matching degree is calculated as the matching degree based on the duration of each temperature segment trajectory and the segment matching degree. When the segment matching degree of any temperature segment trajectory is lower than the preset segment qualification threshold, the temperature range corresponding to the temperature segment trajectory is marked in the evaluation result, and the repair process adjustment suggestion associated with the temperature range is output.

10. The performance evaluation system for the repair process of lithium battery cathode materials according to claim 1, characterized in that, The system further includes an exception response module, which is used for: During the process of the prediction module monitoring the changing trend of the real-time position, it simultaneously monitors whether the changes in the real-time position in the evolution trajectory exhibit at least one of the following abnormal situations: The real-time location change remains stagnant for a preset duration. The direction of change in the real-time position is opposite to the expected direction of change during the repair process; The real-time positions of multiple consecutive sampling points in the evolution trajectory are obtained, and the position change between adjacent sampling points is calculated. When the position change shows irregular and repeated fluctuations in multiple consecutive sampling intervals, and the position change does not show a trend of continuous decrease over time, it is determined that the change in the real-time position is abnormal. When the abnormal situation is detected, an abnormality warning signal is output to the outside. The abnormality warning signal includes the type identifier of the abnormal situation and the repair temperature and repair pressure information at the current moment. Meanwhile, the abnormal response module suspends the prediction signal output of the prediction module and the start of the confirmation module until the abnormal situation is eliminated and then resumes.

Images