A quick charging safety control method for a mobile power supply

Abstract

The application discloses a kind of quick charging safety control methods of mobile power supply, specifically related to the charging control technical field of electrochemical energy storage device, the method is started when mobile power supply is connected with charging adapter after completing high-power discharge, obtains ion relaxation index by open-circuit detection, enters the rhythm cycle control of charging-stop-measurement, and current limiting according to risk level;In the cycle, ion relaxation index and terminal voltage change rate are adapted to adjust charging current and stop charging duration, finally smoothly switch to constant current and constant voltage charging phase when index is stable in safe range, realize the safety control of quick charging;The application accurately detects the ion distribution state of battery before charging, limits current according to risk level in rhythm slow start, dynamically optimizes charging current and stop charging duration by adaptive adjustment, and smoothly switches to constant current and constant voltage phase when the state is stable, safety and efficiency are considered throughout.

Description

Technical Field

[0001] This invention relates to the field of charging control technology for electrochemical energy storage devices, and more specifically, to a fast charging safety control method for mobile power supplies. Background Technology

[0002] With the widespread use of smartphones, tablets, laptops, and outdoor energy storage devices, the demand for high-capacity and fast-charging capabilities in power banks is constantly increasing. While high-rate charging technology significantly shortens charging time, it also brings considerable safety hazards and lifespan reduction issues. Especially when a power bank immediately enters fast-charging mode after completing a high-power discharge, uneven ion distribution often occurs within the battery cell, manifesting as a significant concentration gradient between the electrode surface and the internal region. Continuing to apply high current under these conditions can easily lead to excessively high local current density, causing surface metal deposition and even dendrite growth. This not only increases the risk of internal short circuits but also accelerates the decay of active materials, shortening the cell's lifespan.

[0003] Existing charging safety control methods mostly rely on external macroscopic parameters such as temperature, voltage, and current for overvoltage, overcurrent, or overtemperature protection. While these methods can cut off charging in extreme situations, they lack direct characterization and dynamic adjustment mechanisms for the migration state of ions within the battery cell, and cannot proactively intervene in the worsening trend of uneven ion distribution in the early or middle stages of charging. Furthermore, due to space and cost constraints in power banks, the addition of high-precision internal sensors is difficult to popularize in practical applications, further limiting the safety control effectiveness of existing technologies. Therefore, this invention proposes a fast charging safety control method for power banks to address the aforementioned problems. Summary of the Invention

[0004] To achieve the above objectives, the present invention provides the following technical solution:

[0005] A fast charging safety control method for a power bank includes the following steps:

[0006] Trigger detection: When the power bank is detected to be connected to the charging adapter after completing a high-power discharge operation, the fast charging safety control process is initiated and the ion distribution uniformity detection stage is entered.

[0007] Open circuit detection: By controlling the metal oxide semiconductor device in the charging and discharging circuit to be instantaneously disconnected within a set time window to establish an open circuit state, the rebound change curve of the cell's open circuit voltage is collected in the open circuit state, the first slope and second derivative of the rebound change curve are calculated, and the ion relaxation index used to characterize the degree of ion distribution balance inside the cell is obtained based on the calculation results.

[0008] Rhythmic soft start: Based on the ion relaxation index, the rhythmic cycle control of charging-stop charging-measurement is performed. The rhythmic cycle control includes constant current charging according to the preset charging time, stopping charging according to the preset stop charging time to promote ion diffusion equilibrium, and collecting the cell voltage change according to the preset measurement time to update the ion relaxation index.

[0009] Tiered current limiting: In rhythmic cycle control, the charging current is divided into multiple preset levels and limited to the corresponding current range based on the risk level of the ion relaxation index obtained in real time, so as to reduce the risk of surface metal deposition under uneven ion distribution.

[0010] Adaptive adjustment: Within each preset cycle of rhythmic cyclic control, the charging current and the charging pause duration are dynamically adjusted based on the latest calculated ion relaxation index and the rate of change of terminal voltage. When the ion relaxation index continues to decrease, the charging current is gradually increased and the charging pause duration is shortened. When the ion relaxation index increases, the charging current is immediately reduced and the charging pause duration is extended.

[0011] Safe Exit: When the ion relaxation index and the rate of change of terminal voltage rebound are stable within the preset safe range for several consecutive cycles, the charging process is smoothly switched to the constant current charging stage and the constant voltage charging stage.

[0012] In a preferred embodiment, trigger detection includes the following steps:

[0013] After detecting that the power bank is connected to the charging adapter, the system continuously collects the timestamp of the connection event, the output load current value, the average power value and discharge duration of the most recent high-power discharge, the surface temperature of the casing and the rate of temperature rise. The connection event is confirmed by the insertion status signal of the charging interface. The output load current is collected by the built-in current sampling circuit at least 5 times before and after connection and the average value is taken. The surface temperature of the casing is collected by the temperature sensor attached to the outer surface of the battery cell at 10-millisecond intervals to form a temperature curve.

[0014] The connection event is processed for contact jitter filtering and duration determination. The trigger level flag is calculated based on the power-time integral of the most recent high-power discharge, the ambient temperature gradient, and the rate of change of the housing temperature rise. The power-time integral is obtained by summing the instantaneous power values ​​during the discharge process over time. The ambient temperature gradient is obtained by dividing the change in ambient temperature in the 30 seconds before and 10 seconds after connection by the corresponding time difference. The rate of change of the housing temperature rise is the slope of the linear fitting of the temperature curve for 5 consecutive seconds after connection.

[0015] The power-time integral of the most recent high-power discharge, the ambient temperature gradient, and the rate of change of the shell temperature rise are normalized and then weighted and summed according to preset weighting coefficients. The resulting comprehensive value is divided into three levels: high, medium, and low. The initial sampling voltage resolution, sampling time interval, and total sampling duration are set accordingly for the ion distribution uniformity detection stage. The high level uses the highest resolution, shortest interval, and longest duration; the medium level uses the median value; and the low level uses the lowest resolution, longest interval, and shortest duration. The trigger level mark and the original calculation data are written into the control storage unit for use in the ion distribution uniformity detection stage.

[0016] In a preferred embodiment, contact jitter filtering involves continuously monitoring the amplitude of the input voltage signal fluctuation within 100 milliseconds after connection confirmation. If the voltage change exceeds ±5% of the stable voltage before connection within any 5-millisecond window, it is determined as a jitter event and the connection confirmation is discarded. The duration is determined as follows: after connection confirmation, the voltage and current remain within ±2% of the stable value before connection and the duration is not less than 50 milliseconds to be considered a valid connection event.

[0017] In a preferred embodiment, open-circuit detection includes the following steps:

[0018] Within the first time window after detecting the connection of the charging adapter, the metal oxide semiconductor device in the charging and discharging circuit is instantaneously disconnected to establish an open circuit state. The disconnection time is set to 50 microseconds to 200 microseconds. The baseline open circuit voltage is collected 10 milliseconds before disconnection for zero-point calibration. During the disconnection period, no less than 20 open circuit voltage samples are continuously acquired at a sampling interval of no more than 5 microseconds, and the surface temperature of the casing and the ambient temperature are recorded simultaneously.

[0019] The rebound curve is constructed using the discrete sequence of open-circuit voltage obtained during the disconnection period. Three-point median denoising is performed first, followed by third-order polynomial smoothing fitting. The first-order slope and second-order derivative of the rebound curve are calculated by fitting the polynomial. Three types of feature quantities are extracted: maximum rebound slope, peak value of second-order derivative, and time required to reach half-amplitude rebound. At the same time, the three types of feature quantities are standardized according to the baseline open-circuit voltage and temperature state.

[0020] The standardized maximum rebound slope, the peak value of the second derivative, and the half-amplitude rebound time are weighted and summed according to a preset influence ratio to obtain the ion relaxation index. The risk level is then divided into high-risk, medium-risk, and low-risk levels according to the ion relaxation index threshold.

[0021] Finally, the ion relaxation index, three types of characteristic quantities, disconnection duration, sampling interval, and corresponding temperature data are written into the control storage unit and used as input parameters for subsequent rhythmic soft start-up steps and graded current limiting steps.

[0022] In a preferred embodiment, the influence ratio in the ion relaxation index calculation is adaptively configured in segments according to the shell surface temperature as 40%, 30% and 30%, or 30%, 40% and 30%, or 30%, 30% and 40%.

[0023] In a preferred embodiment, rhythm soft start includes the following steps:

[0024] Based on the risk level classification results corresponding to the ion relaxation index, a rhythmic cycle control of charging-stopping-measuring is executed. Different risk levels correspond to different preset parameter datasets. Each preset parameter dataset includes charging duration, stopping duration, measurement duration, and constant current charging current grading parameters.

[0025] In the rhythmic cycle control, constant current charging is performed according to the preset charging time, and charging is stopped according to the preset stop charging time to promote ion diffusion equilibrium. The cell voltage change is collected within the measurement time and the ion relaxation index is updated.

[0026] In a preferred embodiment, tiered current limiting refers to:

[0027] Obtain the updated ion relaxation index, and based on the risk level to which the ion relaxation index belongs, retrieve a set of parameter data corresponding to the risk level from multiple pre-defined datasets.

[0028] When executing the next rhythmic cycle control, constant current charging and charging stop operations are performed according to the charging duration and charging stop duration corresponding to the current risk level. During the constant current charging phase, the current value is kept below the upper limit of the current range, and during the charging stop phase, a zero current state is maintained to reduce surface metal deposition under uneven ion distribution.

[0029] In a preferred embodiment, adaptive adjustment includes the following steps:

[0030] After each rhythmic cycle control cycle ends, the trend slope is obtained by linear regression based on the ion relaxation index sequence calculated from the previous cycle and the current cycle. The trend slope is then normalized according to the median of the absolute values ​​of the trend slopes over the past few cycles. At the same time, the average rate of change and the variance of the window rate of change are calculated for the terminal voltage time series collected during the measurement phase. The normalized trend slope, the average rate of change of the window, and the variance of the window rate of change are then weighted and summed to form a joint trend factor. The weighting coefficients are preset during the manufacturing phase or the algorithm initialization phase.

[0031] Consistency confidence is constructed based on the proportion of the joint trend factor and the trend slope sign of multiple consecutive periods. A monotonically increasing mapping function is established using the two as input variables. The monotonically increasing mapping function is a binary linear piecewise function or an S-shaped continuous function. Its output value is used as an adjustment indicator. The numerical range of the adjustment indicator is mapped to multiple charging current and charging stop duration combination indices in the parameter dataset.

[0032] After completing the parameter combination selection, boundary constraint rules are applied to the charging current and the charging stop duration. The boundary constraint rules include limiting the charging current between the highest and lowest tier values ​​of the risk level, and limiting the charging stop duration between the shortest and longest tier values ​​of the risk level.

[0033] In a preferred embodiment, when the adjustment indication value is higher than the upper tangent of the current index, a higher charging current level and a shorter charging stop duration level are selected from the parameter dataset corresponding to the same risk level according to a preset scaling factor. When the adjustment indication value is lower than the lower tangent of the current index, a lower charging current level and a longer charging stop duration level are selected according to a preset scaling factor. When the adjustment indication value is between the upper and lower tangents of the current index and the variance of the window change rate decreases, the parameter combination of the previous cycle is maintained and the hysteresis state is maintained. The scaling factor is a preset percentage offset of the charging current and the charging stop duration relative to the current level value.

[0034] In a preferred embodiment, controlling the smooth transition of the charging process to the constant current charging stage and the constant voltage charging stage refers to:

[0035] During the rhythmic cycle control process, the ion relaxation index and the rate of change of the terminal voltage rebound are continuously acquired for multiple cycles. After each cycle, the fluctuation amplitude and the trend of change of the two are calculated respectively. When the fluctuation amplitude is lower than the preset stability threshold and the trend of change is maintained within the positive and negative tolerance range for multiple cycles, it is determined that the ion distribution state and voltage rebound characteristics are within the preset safety range.

[0036] When the determination result meets the safety conditions, the charging current is gradually increased to the target current of the constant current charging stage and kept constant according to the preset smooth switching strategy until the cell voltage reaches the conversion voltage point of the constant voltage charging stage. Finally, the charging voltage is kept constant at the conversion voltage point and the charging current is dynamically reduced until charging is completed.

[0037] The technical effects and advantages of this invention are as follows:

[0038] This invention combines trigger detection and open-circuit detection to ensure that the fast charging safety control process is initiated the instant the power bank completes a high-power discharge operation and is connected to the charging adapter. In the open-circuit state, the rebound curve of the cell's open-circuit voltage is acquired, and the first-order slope and second-order derivative are calculated to obtain the ion relaxation index. This process accurately reflects the balance of ion distribution within the cell without applying additional stress, allowing the charging control system to grasp crucial internal state information at the initial stage. This enables the system to formulate charging strategies based on the actual state, avoiding safety risks such as metal deposition on the electrode surface and increased polarization caused by blindly entering high-current charging. This refined state perception design before charging ensures that fast charging possesses both high responsiveness and a safety foundation, providing precise input conditions for subsequent rhythmic soft start and tiered current limiting.

[0039] This invention utilizes preset charging, stopping, and measurement durations to form a rhythmic cyclical control mode of charging-stopping-measuring during the rhythmic soft start and tiered current limiting stages. Combined with real-time calculated ion relaxation index corresponding to the risk level, the charging current is tiered and limited to the corresponding range. A stable constant current is provided during the charging stage, ion diffusion equalization is promoted during the stopping stage, and the latest cell status is obtained during the measurement stage, achieving closed-loop safety control of the charging process. The tiered current limiting design ensures that the charging current does not exceed the safety threshold under different risk levels, effectively preventing structural damage to the cell caused by overcurrent in high-risk states, while improving charging efficiency in low-risk states. This strategy of allocating current levels according to risk level allows for dynamic matching of charging rate and safety, avoiding the risk accumulation and efficiency waste that may occur with traditional fixed current modes, thus balancing safety, stability, and speed throughout the entire charging process.

[0040] This invention dynamically adjusts the charging current and charging pause duration during the adaptive adjustment and safe exit phases, utilizing the updated ion relaxation index and terminal voltage change rate after each rhythmic cycle. When a continuous decrease in the ion relaxation index is detected, the charging current is gradually increased and the charging pause duration is shortened to accelerate the charging speed; when the ion relaxation index rises, the charging current is immediately reduced and the charging pause duration is extended to suppress the deterioration of ion distribution imbalance. Within multiple consecutive cycles, if the ion relaxation index and terminal voltage rebound rate remain stable within a preset safe range, a smooth switching strategy is triggered, allowing the charging process to smoothly transition to the constant current charging stage. After the cell voltage reaches the constant voltage stage transition point, it switches to constant voltage charging, ultimately dynamically reducing the charging current until charging is complete. This design avoids thermal and stress shocks caused by abrupt switching, while ensuring high-precision tracking of the cell state even near full charge, maintaining a balance between safety and efficiency throughout the entire charging process. Attached Figure Description

[0041] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings;

[0042] Figure 1 This is a schematic diagram of a fast charging safety control method for a mobile power bank according to the present invention. Detailed Implementation

[0043] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0044] Reference Figure 1 The following examples were obtained:

[0045] Example 1: A fast charging safety control method for a mobile power bank, comprising the following steps:

[0046] Trigger Detection: When the power bank is detected to have been connected to a charging adapter after completing a high-power discharge operation, the fast charging safety control process is initiated, and the ion distribution balance detection stage begins. The significance of this step lies in timely identification of this special operating condition when the power bank has just experienced high power output and the internal battery cells are in a high-ion migration state. Monitoring and determining the adapter connection event serves as a prerequisite for triggering the safety control process. At this time, the ion distribution within the battery cells may exhibit a significant imbalance. Directly entering the fast charging stage could lead to safety hazards such as metal deposition and dendrite growth. Therefore, upon detecting this state, the ion distribution balance detection stage is initiated immediately to allow for subsequent targeted adjustments to the charging strategy.

[0047] Open-circuit detection: An open-circuit state is established by momentarily disconnecting the metal-oxide-semiconductor (MOS) components in the charging / discharging circuit within a set time window. During this open-circuit state, the rebound curve of the cell's open-circuit voltage is acquired. The first slope and second derivative of the rebound curve are calculated, and the ion relaxation index, used to characterize the balance of ion distribution within the cell, is obtained based on the calculation results. The significance of this step lies in using the electrochemical relaxation process formed by the momentary open circuit to reflect the ion distribution within the cell. By briefly disconnecting the charging / discharging circuit, the cell is placed in a state free from external current interference, thus enabling accurate observation of the natural voltage rebound characteristics. The first slope of the rebound curve reflects the rate of voltage change, and the second derivative reveals the acceleration characteristics of the voltage change; the combination of these two factors can effectively determine the degree of ion distribution imbalance. The final ion relaxation index, as a quantitative indicator, will provide a data basis for subsequent charging strategy tiering, current limiting, and dynamic adjustment.

[0048] Rhythmic soft start: Based on the ion relaxation index, a rhythmic cycle control of charging-stopping-measuring is implemented. This cycle control includes constant current charging for a preset charging duration, stopping charging for a preset stopping duration to promote ion diffusion equalization, and collecting cell voltage changes for a preset measurement duration to update the ion relaxation index. The significance of this step is to avoid continuous high-power charging before ion distribution is equalized. Instead, intermittent rhythmic control, through a "charging-stopping-measuring" cycle, allows ions time to diffuse naturally during the stopping phase to improve distribution balance. During the charging phase, a stable current is provided to ensure controlled charging; during the stopping phase, external current influences are eliminated to promote ion redistribution; and during the measurement phase, voltage change data is collected to update the ion relaxation index, thereby dynamically reflecting the internal state of the cell and ensuring that the charging strategy continuously matches the actual state.

[0049] Tiered Current Limiting: In rhythmic cycle control, based on the risk level of the ion relaxation index acquired in real time, the charging current is divided into multiple preset levels and limited to the corresponding current range to reduce the risk of surface metal deposition under uneven ion distribution. The significance of this step lies in dividing the charging current into different levels according to the current risk state of the battery cell and strictly limiting it within the corresponding safe range, preventing the application of excessively high current when ion distribution is uneven, thereby mitigating the tendency for oversaturation deposition of surface metal ions. Through the correspondence between risk level and current level, a balance can be achieved between maintaining charging efficiency and ensuring safety, and the current limiting strategy can be adjusted in real time according to the battery cell status.

[0050] Adaptive Adjustment: Within each preset cycle of the rhythmic cyclic control, the charging current and charging pause duration are dynamically adjusted based on the latest calculated ion relaxation index and the rate of change of terminal voltage. When the ion relaxation index continues to decrease, the charging current is gradually increased and the charging pause duration is shortened; when the ion relaxation index increases, the charging current is immediately decreased and the charging pause duration is extended. The significance of this step is to utilize the latest cell state information obtained at the end of each rhythmic cycle to perform real-time adaptive optimization of charging parameters. A decrease in the ion relaxation index indicates that the ion distribution is becoming more balanced, and the charging current can be appropriately increased and the charging pause time reduced to improve efficiency. Conversely, when the ion relaxation index increases, it means that the ion imbalance is aggravated, and measures should be taken immediately to reduce the charging current and extend the charging pause time to prevent the electrochemical reaction from entering a high-risk range. Through this dynamic adjustment, the charging strategy always adapts to the real-time state changes of the cell.

[0051] Safe Exit: When the ion relaxation index and the rate of change of the terminal voltage rebound remain stable within a preset safe range for several consecutive cycles, the charging process is smoothly switched to the constant current charging stage and the constant voltage charging stage. The significance of this step is to smoothly transition the intermittent rhythm control to the normal fast charging curve, provided that the cell state has been confirmed to be stable over a long period and safety risks have been eliminated. By judging the stability of the ion relaxation index and the rate of change of the voltage rebound over multiple cycles, the timing of the switch is ensured to be reasonable. During the smooth switch, the charging current is gradually increased until the target value of the constant current stage is reached, and the process switches to the constant voltage stage when the cell voltage reaches the transition point of the constant voltage stage, ensuring that the charging process is completed safely and efficiently.

[0052] Trigger detection includes the following steps: After detecting that the power bank is connected to the charging adapter, the connection event is confirmed by the insertion status signal of the charging port, and the data acquisition process is immediately initiated. Data acquisition includes the following: First, the timestamp of the connection event is acquired. The timestamp records the precise moment of the connection action in milliseconds for easy time alignment with other measurement data. Second, the output load current value is acquired. The load current is measured by a current sampling circuit installed in the output circuit. This current sampling circuit acquires at least five current values ​​at fixed sampling intervals before and after connecting the charging adapter and takes the arithmetic mean. The average value is calculated by adding the acquired current samples and dividing by the total number of samples. For example, the current values ​​acquired before connection are 0.85 amps, 0.87 amps, 0.86 amps, 0.88 amps, etc. If the current is 0.86 amperes, then the average value is (0.85 + 0.87 + 0.86 + 0.88 + 0.86) / 5 = 0.864 amperes. Third, collect the average power value and duration of the most recent high-power discharge. The average power value is obtained by taking the arithmetic mean of the instantaneous power sequence during the discharge. The instantaneous power is the product of the real-time voltage and the real-time current. For example, if the power measured at ten sampling points during the discharge phase is 16.5 watts, 17.0 watts, 16.8 watts, 17.2 watts, 16.9 watts, 17.1 watts, 16.8 watts, 17.0 watts, 16.9 watts, and 17.1 watts respectively, then the average power value is:

[0053] (16.5+17.0+16.8+17.2+16.9+17.1+16.8+17.0+16.9+17.1) / 10=16.93 watts; the discharge duration is obtained by recording the difference between the timestamps of the start and end of the discharge;

[0054] The surface temperature and rate of temperature rise of the casing are collected. Temperature measurement is performed by a temperature sensor attached to the outer surface of the battery cell. At least ten sets of temperature data are continuously collected at ten-millisecond intervals to form a temperature curve, for example, the temperature sequence is 28.2 degrees Celsius, 28.3 degrees Celsius, 28.4 degrees Celsius... up to the tenth set of data, which can be used to analyze the temperature rise trend. Contact jitter filtering and duration determination are performed on connection events. Contact jitter filtering refers to continuously monitoring the fluctuation amplitude of the input terminal voltage signal within a 100-millisecond time window after connection confirmation. The monitored input terminal voltage sequence is compared with the average stable voltage before connection. If the voltage change exceeds ±5% of the stable voltage within any consecutive 5-millisecond detection window, the connection event is determined to have mechanical jitter or poor contact, and the connection confirmation record is discarded. Duration determination is that after connection confirmation, the input terminal voltage and output terminal current must simultaneously remain within ±2% of the stable value before connection, and this state must last for at least 50 milliseconds to be considered a valid connection event.

[0055] The trigger level flag is calculated using the power-time integral of the most recent high-power discharge, the ambient temperature gradient, and the rate of change of the casing temperature. The power-time integral is the cumulative sum of the instantaneous power values ​​over time during the discharge process. For example, if the power value p(t) is sampled every 0.1 seconds, the integral value over the total discharge duration T can be expressed as Σ[p(ti)×Δt], where Δt is 0.1 seconds. If the power values ​​are successively 16.8 W, 17.0 W, 17.2 W, 16.9 W… until the end of T seconds, then each value is multiplied by 0.1 seconds and accumulated to obtain the total integral. The ambient temperature gradient is the change in ambient temperature between the 30 seconds before and the 10 seconds after access, divided by the corresponding time difference. For example, if the average ambient temperature is 27.5 degrees Celsius in the 30 seconds before access and 28.0 degrees Celsius in the 10 seconds after access, then the gradient is (28.0−27.5) / (30+10)=0.5 / 40=0.0125 degrees Celsius per second. The rate of change of shell temperature is the slope of the linear fitting of the temperature curve for 5 consecutive seconds after connection. The least squares method is used to fit the temperature and time to obtain the value of k in the straight line y=kx+b as the rate of change. For example, if the temperature rises from 28.2 degrees Celsius to 28.7 degrees Celsius in 5 seconds, then the rate of change k=(28.7−28.2) / 5=0.1 degrees Celsius per second.

[0056] The power-time integral, ambient temperature gradient, and shell temperature rise rate of the most recent high-power discharge are normalized and then weighted and summed using preset weighting coefficients. Normalization involves subtracting the minimum value of each indicator from its historical value and dividing by its range, mapping the result to a value between 0 and 1. For example, if an indicator has a minimum value of 10, a maximum value of 50, and a current value of 30, the normalized result would be (30−10) / (50−10) = 0.5. The weighted summation formula is: Comprehensive Value = α × Normalized Power-Time Integral + β × Normalized Ambient Temperature Gradient + γ × Normalized Shell Temperature Rise Rate, where α, β, and γ are preset weighting coefficients. The obtained comprehensive value is compared with preset grading thresholds and classified into three levels: high, medium, and low. The thresholds for high, medium, and low levels can be determined through experimental statistics during the design phase. Corresponding to different risk levels, the initial sampling voltage resolution, sampling time interval, and total sampling duration are set for the ion distribution uniformity detection stage. High-level stages use the highest resolution, shortest interval, and longest duration; medium-level stages use medium parameters; and low-level stages use the lowest resolution, longest interval, and shortest duration. Trigger level markers and raw calculation data are recorded in the control storage unit after classification, used for calling the initial control parameters for the ion distribution uniformity detection stage, enabling the detection process to adaptively optimize the sampling strategy based on the initial risk situation.

[0057] The determination of α, β, and γ can be achieved through the following parameter optimization process based on existing technologies: Experimental data acquisition stage: Prepare multiple sets of mobile power bank samples in different health states (healthy, sub-healthy, and deteriorated) in a temperature-controlled laboratory environment, covering various cell types (such as ternary lithium and lithium iron phosphate). Apply a high-power discharge → fast charging switching condition simulating user scenarios to each set of samples, and collect three raw indicators: power time integral, ambient temperature gradient, and shell temperature rise rate, as well as the corresponding ion distribution uniformity detection results (the "true degree of imbalance" can be obtained as a label value through impedance spectroscopy measurement or open-circuit voltage recovery analysis). Repeat the test multiple times for each condition to ensure a sufficient sample size (e.g., no less than 30 sets of samples for each level).

[0058] Data normalization and correlation analysis: The three indicators are normalized (Min-Max or Z-score standardization) to eliminate dimensional differences. Correlation coefficients (such as Pearson correlation coefficient or Spearman rank correlation coefficient) are calculated between each indicator and the label value to preliminarily assess the strength of each indicator's influence on the degree of ion imbalance. Weight optimization calculation: Using the three normalized indicators as input features and the ion imbalance label as the target value, a coefficient vector [α',β',γ'] is obtained by fitting multiple linear regression, ridge regression, or least squares method.

[0059] The obtained coefficient vector is normalized, for example: α=α' / (α'+β'+γ'), β=β' / (α'+β'+γ'), γ=γ' / (α'+β'+γ'). If the model's generalization ability is considered during the fitting process, cross-validation can be used to verify the stability of the weights on different datasets.

[0060] Validation and Consolidation: Apply the obtained weight coefficients to other test samples that were not used in the training to verify the accuracy of the comprehensive value in classifying risk levels (the accuracy must meet the design requirements, such as ≥90%). After successful validation, solidify the set of weights into the control firmware as preset values, and allow adjustment in subsequent batches via firmware upgrades.

[0061] Contact jitter filtering involves continuously monitoring the amplitude of the input voltage signal fluctuation within 100 milliseconds after connection confirmation. If the voltage change exceeds ±5% of the stable voltage before connection within any 5-millisecond window, it is determined as a jitter event and the connection confirmation is discarded. The duration is determined as follows: after connection confirmation, the voltage and current remain within ±2% of the stable value before connection and the duration is not less than 50 milliseconds to be considered a valid connection event.

[0062] Open-circuit detection includes the following steps: Within the first time window after detecting the connection of the charging adapter, the metal-oxide-semiconductor (MOS) components in the charging / discharging circuit are instantaneously disconnected within a very short time to form an open-circuit state. The disconnection duration is set between 50 and 200 microseconds to ensure that the ion distribution within the cell does not change significantly during measurement, while also avoiding instability in the measurement signal due to excessively short disconnection time. During the 10-millisecond period before the disconnection operation, a baseline open-circuit voltage is acquired. This baseline open-circuit voltage serves as a zero-point calibration reference to eliminate errors introduced by measurement system offsets or environmental noise in subsequent calculations. During the disconnection period, at least 20 open-circuit voltage samples are continuously acquired at sampling intervals of no more than 5 microseconds to ensure that the time resolution can capture the rapid voltage change characteristics at the instant of open-circuit. Simultaneously, the casing surface temperature and ambient temperature are recorded. The casing surface temperature is acquired by a temperature sensor installed on the outer surface of the cell, and the ambient temperature is acquired by an external ambient temperature sensor to compensate for temperature factors during subsequent data processing.

[0063] Using the discrete open-circuit voltage sequence obtained during the disconnection period, a voltage rebound curve is constructed in time sequence. The rebound curve reflects the natural voltage recovery characteristics of the battery cell caused by the electrochemical process after the removal of the external current load. To reduce high-frequency noise interference, the voltage sequence is first subjected to three-point median denoising, i.e., the value of each sampling point is replaced with the median of its three adjacent sampling values ​​to remove isolated abnormal pulse points. Then, a third-order polynomial smoothing fit is applied to the denoised data, and the polynomial coefficients are obtained by fitting using the least squares method, thus obtaining the smooth rebound curve function expression. Based on the fitted curve, its first-order slope is calculated as the voltage change rate, and the second-order derivative is used as a measure of the voltage change acceleration. The first-order slope can be obtained by calculating the first-order derivative of the polynomial at each sampling point, and the second-order derivative is obtained similarly. Furthermore, three characteristic quantities are extracted from the curve: the maximum rebound slope, which is the maximum value of the first slope, representing the moment when the voltage recovers the fastest; the peak value of the second derivative, which is the maximum value of the second derivative, representing the extreme point of the acceleration change; and the time required to reach half-amplitude rebound, which is the time it takes for the voltage to recover to half of the total rebound amplitude from the moment of disconnection, and this time reflects the speed of the rebound process.

[0064] The three types of characteristic quantities mentioned above are standardized according to the baseline open-circuit voltage and temperature conditions. The standardization method is to subtract the mean of the characteristic value corresponding to the baseline voltage or baseline temperature from each characteristic value and divide by the standard deviation of the characteristic value at that temperature condition, thereby eliminating the systematic differences of characteristic quantities under different temperatures or initial voltages. Then, the standardized maximum rebound slope, peak value of the second derivative, and half-amplitude rebound time are weighted and summed according to preset influence ratios. The weighted summation formula is: Ion relaxation index = A × Standardized maximum rebound slope + B × Standardized peak value of the second derivative + C × Standardized half-amplitude rebound time, where A, B, and C are preset influence ratios. The obtained ion relaxation index is used as a quantitative indicator for risk level classification. By comparing it with two pre-set thresholds, the risk level is divided into three levels: high risk, medium risk, and low risk. High risk level indicates a significantly uneven ion distribution, medium risk level indicates a slightly uneven ion distribution, and low risk level indicates a near-equilibrium ion distribution.

[0065] The influence ratios in the ion relaxation index calculation are adaptively configured piecewise with respect to the shell surface temperature as follows: 40%, 30% and 30%, or 30%, 40% and 30%, or 30%, or 30% and 40%. These influence ratios correspond to the weighting coefficients of three characteristic quantities: the standardized maximum rebound slope, the peak value of the standardized second derivative, and the standardized half-amplitude rebound time, respectively, and the sum of these three ratios is 100%. The piecewise adaptive configuration of the influence ratios means dividing the shell surface temperature into multiple temperature ranges, using different weighting ratio combinations within each range to adapt to the impact of temperature changes on the cell's ion migration characteristics and electrochemical reaction kinetics. For example, when the shell surface temperature is in the first temperature range, the weight of the standardized maximum rebound slope is 40%, the weight of the standardized second derivative peak value is 30%, and the weight of the standardized half-amplitude rebound time is 30%; when the shell surface temperature is in the second temperature range, the weight of the standardized maximum rebound slope is 30%, the weight of the standardized second derivative peak value is 40%, and the weight of the standardized half-amplitude rebound time is 30%; when the shell surface temperature is in the third temperature range, the weight of the standardized maximum rebound slope is 30%, the weight of the standardized second derivative peak value is 30%, and the weight of the standardized half-amplitude rebound time is 40%.

[0066] The division of temperature ranges can be determined through experimental data analysis during the manufacturing or equipment calibration stages. For example, in a temperature-controlled laboratory, the ion relaxation behavior of the battery cell at different temperatures can be measured, and the correlation between the three characteristic quantities and the degree of unevenness in the actual ion distribution at different temperatures can be recorded. Then, the relative importance of each characteristic quantity within the temperature range can be determined using the correlation coefficient and regression analysis results, thereby determining the corresponding weighted ratio combination. The adjustment of the influencing ratio is achieved through a lookup table or piecewise function in the control firmware. During operation, the surface temperature of the casing is read in real time, and the corresponding ratio combination is automatically called according to the current temperature range to participate in the calculation of the ion relaxation index. In specific implementation, for example, when the surface temperature of the casing is 30 degrees Celsius, falling into the second temperature range, the calculation process will automatically call the ratio combination of 30%, 40%, and 30%, multiply the three characteristic quantities by the corresponding ratios respectively, and then sum them to obtain the ion relaxation index value under that temperature condition. This ensures that the ion relaxation index more accurately reflects the degree of uniformity of ion distribution inside the battery cell under different temperature conditions, reducing measurement deviations caused by temperature changes.

[0067] The rhythmic soft start includes the following steps: Based on the risk level classification corresponding to the ion relaxation index, a rhythmic cycle control of charging-stopping-measuring is executed. The ion relaxation index is a quantitative value characterizing the degree of ion distribution equilibrium within the cell, calculated through an open-circuit detection process. Risk levels are divided into high-risk, medium-risk, and low-risk levels. Different risk levels correspond to different preset parameter datasets. Each preset parameter dataset contains four types of parameters: charging duration, stop-charging duration, measurement duration, and constant-current charging current grading parameters. Charging duration refers to the length of time a constant current is continuously applied to the cell during the rhythmic cycle. Stop-charging duration refers to the length of time during the cycle when charging is stopped and zero current is maintained to allow ions to redistribute and equalize within the electrode material. Measurement duration refers to the length of time during or after stop-charging when cell voltage data is collected for analysis. Constant-current charging current grading parameters refer to different safe current levels set according to the risk level during the constant-current charging stage. The preset dataset corresponding to the high-risk level will set a shorter charging time, a longer charging stop time, and a smaller constant current charging current level to reduce the risk of ion imbalance; the parameters for the medium-risk level are set between high and low risk, balancing safety and charging efficiency; the parameter dataset for the low-risk level will set a longer charging time, a shorter charging stop time, and a higher constant current charging current level to improve charging efficiency.

[0068] In rhythmic cycle control, a constant current charging current is applied to the battery cell according to a preset charging duration. This constant current is controlled within the current range specified for the risk level, ensuring the charging process remains within a safe power range. Immediately after the charging phase ends, charging is stopped for a preset stop-charging duration, maintaining a zero-current state during this phase. This phase eliminates the ion concentration gradient generated during charging, promoting a uniform ion distribution within the electrodes through diffusion, reducing the likelihood of surface metal deposition during subsequent charging. During the measurement duration of the stop-charging phase, the battery cell voltage change is collected at fixed sampling intervals. This voltage change reflects the dynamic process of voltage rebound after charging stops. The characteristics of this process allow for the recalculation and updating of the ion relaxation index. Updating the ion relaxation index involves denoising and smoothing the measured voltage change data, extracting feature quantities, and weighting and summing them according to their influence ratios to obtain a new ion relaxation index value. The new ion relaxation index is used to determine the risk level of the next cycle, thereby determining the preset parameter dataset to be used in the next cycle, enabling the charging control to adapt to changes in the battery cell state in real time. This rhythmic slow-start method combines risk level classification management with periodic charging parameter adjustment. By alternating constant current charging and charging stop measurement at different stages, it dynamically balances charging efficiency and safety, ensuring that it avoids directly entering full-power fast charging before the ion distribution is fully balanced, thereby reducing the risk of dendrite formation and diaphragm puncture.

[0069] Tiered current limiting refers to: obtaining the updated ion relaxation index, which is an indicator used to quantify the degree of ion distribution balance within the battery cell, calculated from the characteristic quantity of the open-circuit voltage rebound curve during the open-circuit detection step. Based on the risk level to which the ion relaxation index belongs, a set of parameter data corresponding to that risk level is retrieved from a pre-set dataset of multiple parameter sets. These multiple parameter datasets are obtained through experimental calibration and algorithm optimization during the manufacturing or system initialization phases. Each set of parameter data includes information such as charging duration, charging stop duration, constant current charging current tier range, and measurement duration for that risk level, used to guide the selection of charging and discharging strategies in rhythmic cycle control. Risk levels are typically divided into three levels: high risk, medium risk, and low risk. A high risk level indicates a high degree of ion distribution imbalance within the battery cell, requiring a more conservative charging strategy; a medium risk level indicates an intermediate degree of imbalance, requiring a compromise between safety and efficiency in the charging strategy; and a low risk level indicates a near-equilibrium ion distribution, allowing for the use of higher charging power to improve efficiency.

[0070] When executing the next rhythmic cycle control, the system performs constant current charging according to the parameter dataset corresponding to the current risk level and the charging duration under that risk level. During the charging phase, the charging current value is kept below the upper limit of the constant current charging current range for the current risk level. For example, at a high risk level, if the constant current charging current range is 0.2C to 0.3C (where C is the charging rate corresponding to the cell's rated capacity), the actual charging current during the constant current charging phase will be limited to no more than 0.3C to avoid excessive current causing local ion oversaturation and metal deposition. After the charging phase ends, the system enters the de-charging phase according to the de-charging duration under the current risk level. During the de-charging phase, a zero-current state is maintained so that, with the external current load completely removed, the concentration diffusion and migration within the cell promote the redistribution and equalization of ions within the electrode material.

[0071] The design philosophy of this tiered current limiting strategy is to directly link the risk level with the upper limit of the charging current and strictly adhere to this upper limit in each rhythmic cycle control, thereby achieving state-based current limiting control. Since the parameter datasets for different risk levels have been designed with multi-dimensional operating conditions such as temperature, power-time integral, ambient temperature gradient, and casing temperature rise rate in mind, tiered current limiting is not merely a simple current restriction but also embodies a comprehensive control approach based on multi-parameter decision-making. By implementing an upper limit during the constant current charging phase and maintaining a zero current state during the charging shutdown phase, the risk of surface metal deposition under uneven ion distribution conditions can be effectively reduced, dendrite growth can be slowed down, and the possibility of separator puncture can be decreased, thereby improving the safety of the entire charging process and the cell lifespan.

[0072] The adaptive adjustment includes the following steps: After each rhythmic cycle control period, based on the ion relaxation index sequence calculated from the previous and current cycles, a linear regression is used to obtain the trend slope. In this invention, the trend slope is used to represent the direction and speed of change of the ion relaxation index over time. The sign of the trend slope is monotonically positively correlated with the risk level. That is, when the trend slope is positive, it indicates that the ion relaxation index is increasing compared to the previous detection window, representing an aggravation of ion distribution imbalance and an increase in electrochemical risk. In this case, the charging current should be reduced and the charging stop time extended in parameter adjustment. When the trend slope is negative, it indicates that the ion relaxation index is decreasing, representing a trend towards ion distribution equilibrium and a decrease in electrochemical risk. In this case, the charging current should be increased and the charging stop time shortened in parameter adjustment. When the trend slope is zero or close to zero, it indicates that the ion relaxation index remains stable, and the current grading parameters should be maintained unchanged. To avoid misjudgment of direction caused by sampling jitter, the calculated trend slope result needs to be threshold filtered before entering the adjustment logic. Only when its absolute value exceeds the preset minimum effective rate of change is it considered an effective trend. To avoid abnormal interference caused by single fluctuations, the trend slope is normalized. The denominator of the normalization is the median of the absolute values ​​of the trend slope over a number of past cycles. For example, the absolute values ​​of the trend slope over the past five rhythmic cycles are taken and the median is calculated. The current trend slope is divided by this median to obtain the normalized trend slope, making it comparable and consistent in dimensions.

[0073] During the measurement phase, the system collects time-series data of the terminal voltage, which is the potential difference between the two ends of the cell within the measurement period. The sampling interval of the time series is on the order of milliseconds to ensure sufficient resolution. A fixed window is applied to the time series for sliding calculation to obtain the voltage change rate of each window and calculate its average value as the window average change rate. At the same time, the variance of these window change rates is calculated as the window change rate variance to reflect the stability and volatility of the terminal voltage change.

[0074] A joint trend factor is obtained by weighting and summing the normalized trend slope, the average rate of change of the window, and the variance of the window rate of change using weighting coefficients determined through experimental calibration during the manufacturing or algorithm initialization phases. For example, the weighting coefficients may be set to 0.4, 0.35, and 0.25, corresponding to the normalized trend slope, the average rate of change of the window, and the variance of the window rate of change, respectively. The joint trend factor is the weighted result of these three factors; a larger value indicates that the overall state is more stable and the charging conditions can be relaxed.

[0075] Next, the proportion of consecutive cycles with the same trend slope sign is calculated. This is the ratio of the number of times the trend slope has the same sign to the total number of cycles in the past few rhythmic cycles. A higher proportion indicates stronger trend consistency. Using the joint trend factor and the proportion of consistent trend slope signs as input variables, a monotonically increasing mapping function is established to map the input values ​​to adjustment indicator values. This monotonically increasing mapping function can be a bivariate linear piecewise function or an sigmoid continuous function (e.g., a logistic function). The parameters can be determined through data fitting during algorithm initialization. The numerical range of the adjustment indicator corresponds one-to-one with multiple preset charging current and charging stop duration combinations in the parameter dataset. For example, when the adjustment indicator is between 0 and 1, it can be divided into several intervals with equal or non-equal intervals. Each interval maps to a different charging current and charging stop duration combination, thereby achieving state-based adaptive strategy adjustment. After selecting the parameter combination, boundary constraint rules must be applied to the charging current and charging stop duration. Boundary constraint rules are set under the restrictions of risk level. For example, in a high-risk level, the charging current is limited to the highest level of 0.3C and the lowest level of 0.2C, and the charging stop time is limited to the shortest level of 5 seconds and the longest level of 10 seconds. In a low-risk level, the upper limit of the charging current is allowed to reach 0.8C, and the charging stop time is shortened to between 2 and 5 seconds, so as to maximize charging efficiency within a safe range.

[0076] In its implementation, this monotonically increasing mapping function can take two forms: a bivariate linear piecewise function or a smoothly transitioning sigmoid continuous function. When using a bivariate linear piecewise function, the range of values ​​for the joint trend factor and the proportion of the trend slope with consistent signs is first divided into several rectangular intervals. For example, the range of each variable can be divided into four intervals based on three dividing points: 20%, 50%, and 80%. Then, the intervals of the two variables are combined into multiple two-dimensional regions. For each two-dimensional region, a straight line with a pre-defined slope and intercept value is used. The output of the line is the weighted sum of the input values ​​within that region plus the intercept value. For example, in regions with low joint trend factors and low consistency ratios, the slope can be set to 0.05 and the intercept to 0.1; in regions with high joint trend factors and high consistency ratios, the slope can be set to 0.1 and the intercept to 0.3; intermediate regions are determined by interpolation based on experience or historical training data. Once the region containing the input value is determined, the calculation is performed according to the preset slope and intercept of that region. This involves multiplying the joint trend factor and the consistency ratio by their respective weights, summing them, multiplying by the slope, and adding the intercept to obtain the value of the adjustment indication.

[0077] When using an S-shaped continuous function, the calculation steps include: first, multiplying the joint trend factor and the consistency ratio by their respective weights and summing the results to obtain a total input value; then, calculating the output value according to the formula, which is expressed as follows: taking the value of the maximum output amplitude as the numerator, and adding the result of the exponential function as the denominator. The input of the exponential function is the product of the negative growth rate coefficient and the total input value, plus the offset coefficient. The growth rate coefficient determines the steepness of the middle stage of the curve; the larger the value, the faster the change. The offset coefficient determines the position of the input value at the center point of the curve. In specific calculations, for example, if the maximum output amplitude is set to 1, the growth rate coefficient to 5, and the offset coefficient to -2, then when the total input value is zero, the exponential part equals the sum of the growth rate coefficient multiplied by the total input value plus the offset coefficient, which is -2. The result of the exponential function is approximately 0.135, the denominator is 1 plus 0.135, which is approximately 1135, and the output value is approximately 0.88. When the total input value increases to 0.5, the exponential part is 5 multiplied by 0.5 plus -2, which equals 0.5. The result of the exponential function is approximately 1.65, the denominator is 2.65, and the output value is approximately 0.38. When the total input value increases to 2, the exponential part is 5 multiplied by 2 plus -2, which equals 8. The result of the exponential function is approximately 2968, the denominator is 2969, and the output value is close to 0.0034, showing a smoothing out in the high input value region. By adjusting the maximum output amplitude, growth rate coefficient, and offset coefficient, the shape and mapping characteristics of the function can be flexibly changed. In its implementation, this step first involves calculating the parameter combination for the previous cycle's rhythmic cyclic control, then obtaining the current cycle's adjustment indication value, and determining its position within the current parameter combination index range. The index range is defined by the upper and lower tangent values, which are determined during the algorithm initialization phase based on risk level, charging safety assessment results, and historical operating data fitting. These values ​​are used to determine whether an upward or downward adjustment of the parameter combination is necessary.

[0078] When the adjustment indication value is higher than the upper tangent point value of the current index interval, it is determined that the current charging control can increase the charging intensity while ensuring safety. At this time, according to the preset proportional factor, a higher charging current level and a shorter charging stop time level are selected from the parameter dataset corresponding to the current risk level. The proportional factor is the percentage offset of the charging current and charging stop time relative to the current level value. For example, the charging current proportional factor can be 10% to 20%, indicating an increase of 10% to 20% based on the original level value; the charging stop time proportional factor can be 5% to 15%, indicating a decrease of 5% to 15% based on the original level value. In actual execution, the system will first search the parameter dataset of the current risk level for the next higher level combination that meets the requirements after the proportional factor adjustment. If there are multiple candidate combinations, the combination with a smaller increase in charging current and a moderate reduction in charging stop time will be selected first to balance charging speed and safety.

[0079] When the adjustment indicator value is lower than the lower cutoff point value of the current index range, it is determined that the current charging state shows a trend of increased uneven ion diffusion or a rising risk of metal deposition. At this time, a lower charging current level and a longer charging pause duration level are selected according to the preset scaling factor. The scaling factor is a negative offset at this time; for example, the charging current scaling factor can decrease by 10% to 30%, and the charging pause duration scaling factor can increase by 10% to 25%. During execution, the system will prioritize the combination that can significantly extend the ion diffusion time and has a moderate decrease in charging current to ensure that the loss of charging efficiency is minimized while reducing risk.

[0080] When the adjustment indicator value is between the upper and lower tangent values ​​of the current index interval, the system further checks the variance of the window rate of change of the terminal voltage time series acquired during the measurement phase. If the variance of the window rate of change shows a decreasing trend, it indicates that the ion distribution is gradually becoming more balanced and the stability of the charging process is improving. At this point, the parameter combination of the previous cycle is maintained, and the system enters a hysteresis state. The hysteresis state means that no parameter combination is adjusted upwards or downwards within multiple consecutive cycles, and even if the adjustment indicator value fluctuates within a certain range, it will not trigger a range switching. This design is used to avoid frequent parameter adjustments due to short-term fluctuations, thereby maintaining the stability and consistency of the charging process.

[0081] For example, in a specific application scenario, the current risk level is medium risk. The previous cycle's charging current was 2.0 amps, the charging pause duration was 10 seconds, the upper cutoff point was 0.6, the lower cutoff point was 0.4, and the scaling factor was 20%. If the current cycle adjustment indicator value is 0.7, which is higher than the upper cutoff point value of 0.6, the system will adjust the charging current to 2.4 amps (a 20% increase) and the charging pause duration to 8 seconds (a 20% decrease). If the current cycle adjustment indicator value is 0.35, which is lower than the lower cutoff point value of 0.4, the system will adjust the charging current to 1.6 amps (a 20% decrease) and the charging pause duration to 12 seconds (a 20% increase). If the current cycle adjustment indicator value is 0.55, and the window rate of change variance has decreased by 15% compared to the previous cycle, the combination of 2.0 amps and 10 seconds remains unchanged, and the system enters a hysteresis state, continuing for at least three rhythmic cycles before reassessing whether to adjust.

[0082] In rhythmic cycle control, the ion relaxation index and the terminal voltage rebound rate are continuously collected for each cycle. The ion relaxation index characterizes the uniformity of ion distribution in the cell during charging, while the terminal voltage rebound rate characterizes the dynamic characteristics of voltage recovery after charging stops. After each cycle, the absolute values ​​of the differences between the ion relaxation index and the terminal voltage rebound rate between the current and previous cycles are calculated, and this is used as the fluctuation amplitude. Simultaneously, the trend of these fluctuations over multiple consecutive cycles is calculated. The trend can be derived from the slope of a linear regression, with positive and negative signs indicating the direction of change, and the absolute value of the slope representing the rate of change. The system pre-sets stability thresholds, which can be determined during the manufacturing or algorithm initialization phase based on cell type, capacity, and historical test data. For example, the stability threshold for the fluctuation amplitude of the ion relaxation index can be within 3%, and the stability threshold for the fluctuation amplitude of the terminal voltage rebound rate can be within 5%. The system also sets positive and negative tolerance ranges to judge the stability of the trend. For example, a rate of change with an absolute value of the trend slope below 0.5% can be considered within the tolerance range.

[0083] When the fluctuation amplitude of the ion relaxation index and the rate of change of the terminal voltage rebound are both lower than the corresponding stability threshold in several consecutive cycles (e.g., three to five cycles), and the slope of the trend is always within the positive and negative tolerance range, the system determines that the current ion distribution is balanced and the voltage rebound characteristics are stable, that is, the charging safety conditions are met.

[0084] After safety conditions are met, a preset smooth switching strategy is executed. This smooth switching strategy is designed to avoid sudden changes in the charging phase that could impact the battery cell. It includes the following steps: First, the current charging current is gradually increased in segments to reach the target current for the constant current charging phase. For example, the current value is increased by 5% to 10% in each cycle until the target value is reached. The target constant current can be dynamically determined based on the battery cell's rated charging rate and risk level. For example, for a battery cell with a rated capacity of 100 amp-hours, the target constant current for a medium-risk level can be set to 50 to 70 amps.

[0085] The second step involves maintaining a constant current during the constant current charging phase, continuing charging until the cell terminal voltage reaches the preset constant voltage charging phase transition voltage point. This transition voltage point can be determined based on the cell's chemical system; for example, it can be set to 4.2 volts for ternary lithium cells and 3.65 volts for lithium iron phosphate cells.

[0086] The third step is to enter the constant voltage charging stage when the cell terminal voltage reaches the conversion voltage point. The terminal voltage is kept constant, and the charging current is gradually reduced through closed-loop control. The rate of current decrease can be controlled according to an exponential decay function or a linear decreasing function, for example, by reducing it by 2% to 5% per minute, until the charging current drops to the cutoff value (for example, 1% of the cell capacity), and then charging is terminated.

[0087] For example, in a specific application scenario, a lithium iron phosphate cell with a rated capacity of 100 Ah, during rhythmic cycle control, exhibits an ion relaxation index fluctuation of less than 2% and a terminal voltage rebound rate fluctuation of less than 3% over five consecutive cycles, with both trend slopes below a tolerance range of ±0.3%. At this point, the system determines that safety conditions are met and smoothly switches to the constant current charging stage. In each rhythmic cycle, the charging current is gradually increased from 30 Ah to 60 Ah. After constant current charging until the cell voltage reaches 3.65 V, the system switches to the constant voltage charging stage, reducing the charging current by 3% per minute until the current drops to 1 Ah, at which point the charging process ends.

[0088] The above algorithms or formulas are all dimensionless and numerical calculations, and the results are obtained by software simulation based on a large amount of collected data to obtain the most recent real-world results. The preset parameters are set by those skilled in the art according to the actual situation.

[0089] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0090] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0091] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0092] The above description is merely a specific embodiment 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 fast charging safety control method for a mobile power bank, characterized in that, Includes the following steps: Trigger detection: When the power bank is detected to be connected to the charging adapter after completing a high-power discharge operation, the fast charging safety control process is initiated and the ion distribution uniformity detection stage is entered. Open circuit detection: By controlling the metal oxide semiconductor device in the charging and discharging circuit to be instantaneously disconnected within a set time window to establish an open circuit state, the rebound change curve of the cell's open circuit voltage is collected in the open circuit state, the first slope and second derivative of the rebound change curve are calculated, and the ion relaxation index used to characterize the degree of ion distribution balance inside the cell is obtained based on the calculation results. Rhythmic soft start: Based on the ion relaxation index, the rhythmic cycle control of charging-stop charging-measurement is performed. The rhythmic cycle control includes constant current charging according to the preset charging time, stopping charging according to the preset stop charging time to promote ion diffusion equilibrium, and collecting the cell voltage change according to the preset measurement time to update the ion relaxation index. Tiered current limiting: In rhythmic cycle control, the charging current is divided into multiple preset levels and limited to the corresponding current range based on the risk level of the ion relaxation index obtained in real time, so as to reduce the risk of surface metal deposition under uneven ion distribution. Adaptive adjustment: Within each preset cycle of rhythmic cyclic control, the charging current and the charging pause duration are dynamically adjusted based on the latest calculated ion relaxation index and the rate of change of terminal voltage. When the ion relaxation index continues to decrease, the charging current is gradually increased and the charging pause duration is shortened. When the ion relaxation index increases, the charging current is immediately reduced and the charging pause duration is extended. Safe Exit: When the ion relaxation index and the rate of change of terminal voltage rebound are stable within the preset safe range for several consecutive cycles, the charging process is smoothly switched to the constant current charging stage and the constant voltage charging stage. Open circuit testing includes the following steps: Within the first time window after detecting the connection of the charging adapter, the metal oxide semiconductor device in the charging and discharging circuit is instantaneously disconnected to establish an open circuit state. The disconnection time is set to 50 microseconds to 200 microseconds. The baseline open circuit voltage is collected 10 milliseconds before disconnection for zero-point calibration. During the disconnection period, no less than 20 open circuit voltage samples are continuously acquired at a sampling interval of no more than 5 microseconds, and the surface temperature of the casing and the ambient temperature are recorded simultaneously. The rebound curve is constructed using the discrete sequence of open-circuit voltage obtained during the disconnection period. Three-point median denoising is performed first, followed by third-order polynomial smoothing fitting. The first-order slope and second-order derivative of the rebound curve are calculated by fitting the polynomial, and three types of feature quantities are extracted: maximum rebound slope, peak value of second-order derivative, and time required to reach half-amplitude rebound. At the same time, the three types of feature quantities are standardized according to the baseline open-circuit voltage and temperature state. The standardized maximum rebound slope, the peak value of the second derivative, and the half-amplitude rebound time are weighted and summed according to a preset influence ratio to obtain the ion relaxation index. The risk level is then divided into high-risk, medium-risk, and low-risk levels according to the ion relaxation index threshold.

2. The fast charging safety control method for a mobile power bank according to claim 1, characterized in that, Trigger detection includes the following steps: After detecting that the power bank is connected to the charging adapter, the system continuously collects the timestamp of the connection event, the output load current value, the average power value and discharge duration of the most recent high-power discharge, the surface temperature of the casing and the rate of temperature rise. The connection event is confirmed by the insertion status signal of the charging interface. The output load current is collected by the built-in current sampling circuit at least 5 times before and after connection and the average value is taken. The surface temperature of the casing is collected by the temperature sensor attached to the outer surface of the battery cell at 10-millisecond intervals to form a temperature curve. The connection event is processed for contact jitter filtering and duration determination. The trigger level flag is calculated based on the power-time integral of the most recent high-power discharge, the ambient temperature gradient, and the rate of change of the housing temperature rise. The power-time integral is obtained by summing the instantaneous power values ​​during the discharge process over time. The ambient temperature gradient is obtained by dividing the change in ambient temperature in the 30 seconds before and 10 seconds after connection by the corresponding time difference. The rate of change of the housing temperature rise is the slope of the linear fitting of the temperature curve for 5 consecutive seconds after connection. The three indicators—power-time integral of the most recent high-power discharge, ambient temperature gradient, and shell temperature rise rate—are normalized and then weighted and summed according to preset weighting coefficients. The resulting comprehensive value is divided into three levels: high, medium, and low. Correspondingly, the initial sampling voltage resolution, sampling time interval, and total sampling duration are set for the ion distribution uniformity detection stage.

3. The fast charging safety control method for a mobile power bank according to claim 2, characterized in that, Contact jitter filtering involves continuously monitoring the amplitude of the input voltage signal fluctuation within 100 milliseconds after connection confirmation. If the voltage change exceeds ±5% of the stable voltage before connection within any 5-millisecond window, it is considered a jitter event. The duration is determined by the voltage and current remaining within ±2% of the stable value before connection after connection confirmation for a duration of not less than 50 milliseconds to be considered a valid connection event.

4. The fast charging safety control method for a mobile power bank according to claim 3, characterized in that, Rhythm slow start includes the following steps: Based on the risk level classification results corresponding to the ion relaxation index, a rhythmic cycle control of charging-stopping-measuring is executed. Different risk levels correspond to different preset parameter datasets. Each preset parameter dataset includes charging duration, stopping duration, measurement duration, and constant current charging current grading parameters. In the rhythmic cycle control, constant current charging is performed according to the preset charging time, and charging is stopped according to the preset stop charging time to promote ion diffusion equilibrium. The cell voltage change is collected within the measurement time and the ion relaxation index is updated.

5. The fast charging safety control method for a mobile power bank according to claim 4, characterized in that, Tiered traffic restriction refers to: Obtain the updated ion relaxation index, and based on the risk level to which the ion relaxation index belongs, retrieve a set of parameter data corresponding to the risk level from multiple pre-defined datasets. When executing the next rhythmic cycle control, constant current charging and charging stop operations are performed according to the charging duration and charging stop duration corresponding to the current risk level. During the constant current charging phase, the current value is kept below the upper limit of the current range, and during the charging stop phase, the current is kept at zero.

6. The fast charging safety control method for a mobile power bank according to claim 5, characterized in that, Adaptive adjustment includes the following steps: After each rhythmic cycle control period ends, the trend slope is obtained by linear regression based on the ion relaxation index sequence calculated from the previous period and the current period. The trend slope is then normalized according to the median of the absolute values ​​of the trend slopes over the past few periods. At the same time, the average rate of change and the variance of the window rate of change are calculated for the terminal voltage time series collected during the measurement phase. The normalized trend slope, the average rate of change of the window, and the variance of the window rate of change are then weighted and summed to form a joint trend factor. Consistency confidence is constructed based on the proportion of the joint trend factor and the trend slope sign of multiple consecutive periods. A monotonically increasing mapping function is established using the two as input variables. The monotonically increasing mapping function is a binary linear piecewise function or an S-shaped continuous function. Its output value is used as an adjustment indicator. The numerical range of the adjustment indicator is mapped to multiple charging current and charging stop duration combination indices in the parameter dataset. After completing the parameter combination selection, boundary constraint rules are applied to the charging current and the charging stop duration. The boundary constraint rules include limiting the charging current between the highest and lowest tier values ​​of the risk level, and limiting the charging stop duration between the shortest and longest tier values ​​of the risk level.

7. The fast charging safety control method for a mobile power bank according to claim 6, characterized in that, When the adjustment indication value is higher than the upper tangent of the current index, a charging current range higher than the current range and a charging stop duration range shorter than the current range are selected from the parameter dataset corresponding to the same risk level according to the preset scaling factor. When the adjustment indication value is lower than the lower tangent of the current index, a charging current range lower than the current range and a charging stop duration range longer than the current range are selected according to the preset scaling factor. When the adjustment indication value is between the upper and lower tangents of the current index and the variance of the window change rate decreases, the parameter combination of the previous cycle is maintained and the hysteresis state is maintained. The scaling factor is the preset percentage offset of the charging current and charging stop duration relative to the current range value.

8. The fast charging safety control method for a mobile power bank according to claim 6, characterized in that, Smoothly switching the charging process to the constant current charging stage and the constant voltage charging stage refers to: During the rhythmic cycle control process, the ion relaxation index and the rate of change of the terminal voltage rebound are continuously acquired for multiple cycles. After each cycle, the fluctuation amplitude and the trend of change of the two are calculated respectively. When the fluctuation amplitude is lower than the preset stability threshold and the trend of change is maintained within the positive and negative tolerance range for multiple cycles, it is determined that the ion distribution state and voltage rebound characteristics are within the preset safety range. When the determination result meets the safety conditions, the charging current is gradually increased to the target current of the constant current charging stage and kept constant according to the preset smooth switching strategy until the cell voltage reaches the conversion voltage point of the constant voltage charging stage. Finally, the charging voltage is kept constant at the conversion voltage point and the charging current is dynamically reduced until charging is completed.

Images