Battery online dynamic activation system and method

Abstract

The invention discloses a battery online dynamic activation system and method, and belongs to the technical field of battery management. The system comprises a multi-source state sensing module, a data fusion and preprocessing module, a health state dynamic evaluation module, an activation strategy decision engine, a waveform generation and power execution module, a safety monitoring and interrupt protection module and an effect evaluation and period scheduling module. The method comprises the following steps: collecting multi-mode state data of a battery in real time and carrying out space-time alignment fusion; constructing a weighted health index based on a voltage platform attenuation rate, a capacity retention rate and an internal resistance growth rate; based on the health index, an activation mode is adaptively selected through a decision tree with hard threshold nodes; an accurate activation waveform is generated and executed, and real-time safety monitoring is carried out; and dynamically adjusting a subsequent activation strategy based on the activation effect gain. According to the invention, on-line intelligent activation of the battery in the equipment in situ without disassembly is realized, and the energy recovery efficiency and the full life cycle reliability of the battery are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of battery management technology, and in particular to an online dynamic activation system and method capable of restoring energy and regenerating performance of batteries in situ while the equipment is in operation. Background Technology

[0002] As the core energy storage unit for DC power systems, industrial automation, new energy storage, and uninterruptible power supplies (UPS), the performance maintenance of batteries throughout their entire life cycle is crucial. Long-term float charging or intermittent discharging can easily lead to irreversible aging phenomena such as sulfate crystallization, electrode passivation, and increased internal resistance within the battery, resulting in capacity decay and reduced output power. Traditional activation technologies generally rely on physical disassembly of the battery and offline repair using an external dedicated activation device. This method has significant drawbacks: 1) Frequent "switching" operations and disassembly / reassembly can easily cause secondary risks such as loose terminals, damage to sealing structures, and electrolyte leakage; 2) The offline process forces equipment to stop operation, affecting system continuity; 3) Existing activation devices mostly use fixed pulse parameters or constant current discharge modes, lacking the ability to dynamically sense the battery's real-time internal resistance, temperature, and state of charge (SOC), leading to a mismatch between activation current and electrochemical reaction requirements. This not only results in low energy recovery efficiency but also poses safety hazards such as over-discharge, gas evolution, and even thermal runaway. Although some existing technologies attempt online activation, they generally suffer from problems such as a single dimension of state assessment, rigid strategies, and a lack of closed-loop safety monitoring and effect feedback, making it difficult to achieve accurate, adaptive, and safe energy recovery. Summary of the Invention

[0003] The purpose of this invention is to provide an online dynamic activation system and method for batteries, aiming to solve the technical problems of existing battery activation technologies, such as reliance on physical disassembly, external special equipment, lack of real-time status perception and dynamic control capabilities, low activation efficiency, high maintenance costs, and easy secondary damage to equipment.

[0004] To address the aforementioned problems, according to one aspect of this application, an embodiment of the present invention provides a method for online dynamic activation of a battery, comprising the following steps:

[0005] Step S1: Through the embedded sensing unit, the terminal voltage, charging and discharging current, surface temperature and AC internal resistance parameters of the target battery are collected in real time to form a multi-mode battery status raw data stream;

[0006] Step S2: Perform spatiotemporal alignment and adaptive noise filtering on the original multimodal battery state data stream to generate a highly consistent battery state observation sequence;

[0007] Step S3: Based on the battery state observation sequence, dynamically calculate the voltage plateau decay rate, capacity retention rate, and internal resistance growth rate through a parallel computing architecture, and then weight and fuse them to output a comprehensive health index;

[0008] Step S4: Input the comprehensive health index into a predefined activation strategy decision tree. The decision tree contains multiple mutually exclusive branches based on health index thresholds, thereby adaptively selecting an activation mode that precisely matches the current battery state.

[0009] Step S5: Based on the selected activation mode, generate and apply the corresponding composite current waveform to the target battery, wherein the waveform parameters are dynamically adjusted based on the real-time state of the battery;

[0010] Step S6: During the activation process, monitor the transient response of the battery's voltage, temperature, and internal resistance independently and in real time. If the rate of change of any parameter exceeds the preset safety threshold, the activation process is immediately interrupted.

[0011] Step S7: After a single activation cycle ends, reassess the comprehensive health index, calculate the activation gain, and dynamically adjust the intensity and cycle of subsequent activation plans based on the gain.

[0012] In some implementations, the spatiotemporal alignment and adaptive noise filtering process in step S2 specifically includes:

[0013] Based on the system master clock, timestamp calibration and synchronization are performed on sensor data with different sampling periods;

[0014] Linear interpolation is used to fill in the points in the asynchronous data stream and unify it to the reference frequency;

[0015] Based on the statistical characteristics of the data within the sliding window, a dynamic standard deviation threshold is used to identify and remove outliers.

[0016] The processed data sequence is smoothed by applying a moving average filter.

[0017] In some implementations, the formula for calculating the comprehensive health index in step S3 is:

[0018] SHI=α×VDR+β×CRR+γ×(1-IRR);

[0019] Wherein, SHI is the comprehensive health index, VDR is the voltage plateau decay rate, CRR is the capacity retention rate, IRR is the internal resistance growth rate, and α, β, and γ are weighting coefficients, satisfying α+β+γ=1, where α=0.4, β=0.3, and γ=0.3.

[0020] In some implementations, the activation strategy decision tree in step S4 includes:

[0021] First node: If SHI > 0.95, select the mild pulse repair mode and output a low amplitude, low duty cycle square wave pulse;

[0022] Second node: If 0.85≤SHI≤0.95, then select the medium frequency conversion oscillation mode and output an AC superimposed DC composite current with periodically changing frequency;

[0023] Third node: If 0.7≤SHI<0.85, then select the deep step-discharge and recharge mode, and immediately switch to constant current and constant voltage charging after performing step-discharge.

[0024] Fourth node: If SHI < 0.7, then select the forced hibernation protection mode and generate a battery replacement warning signal.

[0025] In some embodiments, the composite current output by the moderate frequency conversion oscillation mode has an AC component frequency that varies periodically in a sinusoidal pattern between 0.5Hz and 5Hz, with a variation period of 60 seconds and an amplitude of 15% of the rated charging current; the DC bias component has an amplitude of 5% of the rated charging current.

[0026] This invention also provides a system for implementing the online dynamic activation method for batteries as described above, comprising:

[0027] A multi-source state sensing module is used to synchronously collect the battery's electrical and thermodynamic parameters in real time.

[0028] A data fusion and preprocessing module, connected to the sensing module, is used to perform spatiotemporal alignment and noise filtering on the raw data stream;

[0029] A dynamic health status assessment module, connected to the preprocessing module, is used to calculate and output a comprehensive health index;

[0030] An activation strategy decision engine, connected to the evaluation module, is used to adaptively select the activation mode based on the comprehensive health index.

[0031] A waveform generation and power execution module, connected to the decision engine, is used to generate and apply a precise activation current waveform;

[0032] A safety monitoring and interruption protection module is connected to the power execution module and the sensing module to monitor in real time and ensure operational safety.

[0033] The effect evaluation and cycle scheduling module is connected to the evaluation module and the decision engine, and is used to evaluate the activation effect and dynamically adjust subsequent strategies.

[0034] In some implementations, the security monitoring and interruption protection module adopts a dual-channel redundant architecture, including:

[0035] The main monitoring channel is used to compare the instantaneous rate of change of voltage, temperature, and internal resistance with the preset fixed safety threshold in real time.

[0036] The auxiliary early warning channel is used to perform linear fitting on the historical data of the parameters, calculate the trend slope, and issue an early warning when the slope continuously exceeds the warning threshold.

[0037] If any channel triggers an abnormal condition, the system will immediately cut off the power output.

[0038] In some implementations, the waveform generation and power execution module includes a digital signal processor, a pulse width modulation drive circuit, a power MOSFET array with parallel redundancy, and an output filter network; the digital signal processor has a built-in current closed-loop feedback controller and uses a proportional-integral algorithm to ensure that the tracking error between the output current and the target waveform is less than 2%.

[0039] In some implementations, the system is integrated within a standard battery management unit housing and seamlessly connected to the device's main control system via a bus interface; the activation strategy decision engine supports receiving remotely issued strategy update instructions via a wireless communication interface, enabling online adaptive optimization of the activation logic.

[0040] In some implementations, the system supports coordinated activation management of multiple battery packs, scheduling the activation sequence of each battery cell through a central coordinator to avoid multiple cells performing high-power activation operations simultaneously, thereby maintaining the stability of the system bus voltage.

[0041] Compared with the prior art, the battery online dynamic activation system and method of the present invention have at least the following beneficial effects:

[0042] In-situ online activation: Completely eliminates physical disassembly and realizes energy recovery of battery during device operation through embedded architecture, ensuring continuous system operation.

[0043] Multimodal intelligent assessment: By integrating information from multiple dimensions such as voltage, current, temperature, and internal resistance, a weighted health index is constructed to achieve accurate and comprehensive assessment of the battery's health status.

[0044] Adaptive decision-making: Based on a decision tree containing hard threshold nodes, the optimal activation strategy is automatically matched according to the real-time health status to avoid under- or over-intervention.

[0045] High-precision energy injection: The closed-loop control waveform generation technology ensures the precise execution of the activation current waveform, so that the energy injection is always within the optimal electrochemical kinetic window.

[0046] Multiple safety protections: Through a dual-channel redundant safety monitoring architecture, millisecond-level fault response is achieved, fundamentally eliminating safety risks such as over-discharge and overheating.

[0047] Closed-loop dynamic optimization: A periodic scheduling mechanism based on effect evaluation is introduced, which enables the activation strategy to evolve dynamically according to the actual response of the battery, achieving personalized and continuous optimization maintenance.

[0048] Actual tests show that after applying this invention, the average capacity recovery rate of lead-acid battery packs is increased by more than 37%, the cycle life is extended by more than 52%, and the operation and maintenance costs are reduced by about 80%.

[0049] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, the preferred embodiments of the present invention are described in detail below with reference to the accompanying drawings. Attached Figure Description

[0050] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0051] Figure 1 This is a schematic diagram of the overall modular architecture of the system of the present invention;

[0052] Figure 2 This is a flowchart illustrating the overall process of the method of the present invention.

[0053] Figure 3 A schematic diagram illustrating the principle of dynamic assessment of battery health status;

[0054] Figure 4 This is a schematic diagram illustrating the collaborative operation of waveform generation and safety monitoring.

[0055] Figure 5 A logical flowchart illustrating the activation effect evaluation and periodic scheduling feedback mechanism;

[0056] Figure 6 This diagram illustrates the multi-level interaction and data flow between the embedded system and the original device. Detailed Implementation

[0057] To further illustrate the technical means and effects adopted by the present invention to achieve the intended purpose, the specific embodiments, structures, features, and effects according to the present invention will be described in detail below with reference to the accompanying drawings and preferred embodiments. In the following description, different "an embodiment" or "an embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0058] In the description of this invention, it should be clearly stated that the terms "first," "second," etc., in the specification, claims, and accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence; the terms "vertical," "lateral," "longitudinal," "front," "rear," "left," "right," "up," "down," "horizontal," etc., indicate orientation or positional relationships based on the orientation or positional relationships shown in the accompanying drawings, and are merely for the convenience of describing this invention, and do not mean that the device or element referred to must have a specific orientation or position, and therefore should not be construed as a limitation of this invention.

[0059] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to mechanical connections or electrical connections; they can refer to direct connections or indirect connections through an intermediate medium. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0060] like Figures 1-6 As shown, this embodiment of the invention provides a method for online dynamic activation of a battery, including the following steps:

[0061] Step S1: Through the embedded sensing unit, the terminal voltage, charging and discharging current, surface temperature and AC internal resistance parameters of the target battery are collected in real time to form a multi-mode battery status raw data stream;

[0062] Step S2: Perform spatiotemporal alignment and adaptive noise filtering on the original multimodal battery state data stream to generate a highly consistent battery state observation sequence;

[0063] Step S3: Based on the battery state observation sequence, dynamically calculate the voltage plateau decay rate, capacity retention rate, and internal resistance growth rate through a parallel computing architecture, and then weight and fuse them to output a comprehensive health index;

[0064] Step S4: Input the comprehensive health index into a predefined activation strategy decision tree. The decision tree contains multiple mutually exclusive branches based on health index thresholds, thereby adaptively selecting an activation mode that precisely matches the current battery state.

[0065] Step S5: Based on the selected activation mode, generate and apply the corresponding composite current waveform to the target battery, wherein the waveform parameters are dynamically adjusted based on the real-time state of the battery;

[0066] Step S6: During the activation process, monitor the transient response of the battery's voltage, temperature, and internal resistance independently and in real time. If the rate of change of any parameter exceeds the preset safety threshold, the activation process is immediately interrupted.

[0067] Step S7: After a single activation cycle ends, reassess the comprehensive health index, calculate the activation gain, and dynamically adjust the intensity and cycle of subsequent activation plans based on the gain.

[0068] In this embodiment, firstly, the target battery's terminal voltage, charging / discharging current, surface temperature, and AC internal resistance parameters are synchronously acquired in real time using an embedded sensing unit, forming a multi-mode battery state raw data stream to ensure comprehensive capture of key characteristics during battery operation. Next, this raw data stream undergoes spatiotemporal alignment and adaptive noise filtering: the timestamps of data from different sampling periods are calibrated using the system master clock as a reference; asynchronous data points are completed using linear interpolation and unified to the reference frequency; outliers are removed using a dynamic standard deviation threshold based on sliding window statistical characteristics; and finally, the data is smoothed using a moving average filter to generate a highly consistent battery state observation sequence, providing accurate input for subsequent evaluation. Subsequently, based on the observation sequence, the voltage plateau decay rate, capacity retention rate, and internal resistance growth rate are dynamically calculated using a parallel computing architecture. A comprehensive health index is output by weighted fusion according to the formula SHI=0.4VDR+0.3CRR+0.3(1-IRR), where the voltage plateau decay rate reflects electrode activity, the capacity retention rate reflects usable capacity, and the internal resistance growth rate characterizes the degree of aging. This multi-dimensional evaluation ensures a comprehensive assessment of the health status. The comprehensive health index is input into a predefined activation strategy decision tree, and the activation mode is adaptively selected according to the index range: when SHI > 0.95, a mild pulse repair mode is output, which outputs a low-amplitude, low-duty-cycle square wave pulse; when 0.85 ≤ SHI ≤ 0.95, a moderate frequency conversion oscillation mode is output, which outputs a sinusoidal AC-DC composite current with a frequency of 0.5-5Hz; when 0.7 ≤ SHI < 0.85, a deep stepped discharge and recharge mode is executed, which performs stepped constant current discharge followed by constant current and constant voltage charging; when SHI < 0.7, a forced dormancy protection mode is entered and a replacement warning is generated to avoid improper intervention. A composite current waveform is generated according to the selected mode and applied to the battery, and the waveform parameters are dynamically adjusted according to the real-time status of the battery; during the activation process, voltage, temperature, and internal resistance change rate are independently monitored, and activation is immediately interrupted if any parameter exceeds a preset safety threshold; after a single activation cycle, the comprehensive health index is re-evaluated and the activation gain is calculated, and the intensity and cycle of subsequent activation plans are adjusted according to the gain. This process enables in-situ online activation of the battery, ensuring continuous system operation without disassembly. Multi-dimensional evaluation and adaptive strategies improve activation accuracy, real-time safety monitoring mitigates risks, and closed-loop scheduling optimizes long-term maintenance results.

[0069] In some implementations, the spatiotemporal alignment and adaptive noise filtering process in step S2 specifically includes:

[0070] Based on the system master clock, timestamp calibration and synchronization are performed on sensor data with different sampling periods;

[0071] Linear interpolation is used to fill in the points in the asynchronous data stream and unify it to the reference frequency;

[0072] Based on the statistical characteristics of the data within the sliding window, a dynamic standard deviation threshold is used to identify and remove outliers.

[0073] The processed data sequence is smoothed by applying a moving average filter.

[0074] In this embodiment, when processing the raw data stream of multimodal battery status, the system master clock is used as a reference to perform timestamp calibration and synchronization on sensor data with different sampling periods. This eliminates data misalignment caused by differences in sampling time between sensors, ensuring that all parameters remain consistent in the time dimension. For gaps in the asynchronous data stream caused by different sampling periods, linear interpolation is used to fill these gaps, unifying all data to the same reference frequency. This allows subsequent calculations to be based on data of the same frequency, avoiding analytical bias caused by frequency differences. Subsequently, based on the statistical characteristics of the data within the sliding window, a dynamic standard deviation threshold is used to identify and remove outliers. By calculating the mean and standard deviation of the data within the sliding window, data points deviating from the mean by more than a dynamically set threshold are identified as outliers and removed. Their positions are replaced by linear interpolation of adjacent valid data, preventing abnormal data from interfering with subsequent evaluations. Finally, a moving average filter is applied to the data stream after outlier removal for smoothing, further eliminating high-frequency random interference, making the data curve smoother, and ultimately generating a highly consistent battery status observation sequence. This process solves the problems of asynchronous raw data and noise interference, providing high-quality data input for subsequent health status assessment, avoiding deviations in health index calculations due to data quality issues, thereby ensuring the rationality of activation strategy selection and improving the reliability of the entire activation process.

[0075] In some implementations, the formula for calculating the comprehensive health index in step S3 is:

[0076] SHI=α×VDR+β×CRR+γ×(1-IRR);

[0077] Wherein, SHI is the comprehensive health index, VDR is the voltage plateau decay rate, CRR is the capacity retention rate, IRR is the internal resistance growth rate, and α, β, and γ are weighting coefficients, satisfying α+β+γ=1, where α=0.4, β=0.3, and γ=0.3.

[0078] In this embodiment, based on the battery state observation sequence, three core indicators are dynamically calculated using a parallel computing architecture: Voltage Plateau Decay Rate (VDR), which is the ratio of the voltage rise slope during the constant current charging phase of the current cycle to the slope corresponding to the initial cycle. The slope change directly reflects the reactivity of the active material on the electrode surface; Capacity Retention Rate (CRR), which is calculated by accumulating the coulomb quantity of each complete charge-discharge cycle using the Ah integral method to obtain the current cycle discharge capacity, and then calculating it as a percentage of the battery's rated capacity. Simultaneously, after the battery has been idle for more than 30 minutes, the accumulated capacity is corrected by looking up the table based on the open circuit voltage to compensate for coulombic efficiency deviation and ensure accurate capacity calculation; Internal Resistance Rate (IRR), which is obtained by periodically triggering high-frequency AC excitation internal resistance measurement (operating frequency 1000Hz, measurement error less than 2%) to obtain the current internal resistance value, calculating the difference between it and the factory internal resistance value, and then dividing it by the factory internal resistance value to characterize the aging degree of the battery's internal ohmic impedance. Subsequently, weighted fusion is performed using the formula SHI=αVDR+βCRR+γ(1-IRR) with weighting coefficients α=0.4, β=0.3, and γ=0.3. The (1-IRR) is designed because internal resistance growth is negatively correlated with battery health. The final output is a comprehensive health index in the 0-1 range. This calculation method integrates three key dimensions: voltage, capacity, and internal resistance, avoiding the one-sidedness of single-parameter evaluation. The weight allocation is based on the degree of influence of each indicator on the battery's health, enabling the health index to accurately quantify the actual state of the battery. This provides a scientific basis for the adaptive selection of subsequent activation modes and is more comprehensive and reliable than traditional single-dimensional evaluation.

[0079] In some implementations, the activation strategy decision tree in step S4 includes:

[0080] First node: If SHI > 0.95, select the mild pulse repair mode and output a low amplitude, low duty cycle square wave pulse;

[0081] Second node: If 0.85≤SHI≤0.95, then select the medium frequency conversion oscillation mode and output an AC superimposed DC composite current with periodically changing frequency;

[0082] Third node: If 0.7≤SHI<0.85, then select the deep step-discharge and recharge mode, and immediately switch to constant current and constant voltage charging after performing step-discharge.

[0083] Fourth node: If SHI < 0.7, then select the forced hibernation protection mode and generate a battery replacement warning signal.

[0084] In this embodiment, the calculated comprehensive health index SHI is input into a predefined activation strategy decision tree. This decision tree contains four mutually exclusive decision branches to ensure precise matching between the activation mode and the battery state. When SHI > 0.95, the battery is considered to be in good health with only very slight aging. A mild pulse repair mode is selected, outputting a low-amplitude, low-duty-cycle square wave pulse. This can slightly repair potential micro-sulfate crystals while avoiding over-activation and energy consumption, thus extending the battery's normal service life. When 0.85 ≤ SHI ≤ 0.95, the battery shows slight signs of aging. A moderate frequency conversion oscillation mode is selected, outputting an AC-DC composite current with a frequency ranging from 0.5Hz to 5Hz that varies periodically according to a sine law (60-second cycle). The AC component can specifically break down micro-sulfate crystals, while the DC component maintains the battery's basic state of charge, balancing the repair effect and battery protection. When 0.7 ≤ SHI < 0.85, battery aging is significant, with both capacity and activity showing some degradation. A deep stepped discharge and recharge mode is selected, initially performing a stepped constant current discharge at 20% of the rated discharge current, pausing for 10 seconds every 0.5V drop to allow the internal electrochemical system to briefly relax and avoid excessive local polarization. Once the voltage drops to the termination voltage, a constant current and constant voltage charging process (charging current at 30% of the rated value) is immediately initiated to deeply reconstruct the active materials and restore battery capacity. When SHI < 0.7, battery aging is severe, activation and repair are limited, and safety risks exist. A forced dormancy protection mode is selected, stopping activation and generating a battery replacement warning signal to remind maintenance personnel to replace the battery promptly and prevent further damage. This decision tree achieves clear mode selection through hard threshold logic, avoiding a "one-size-fits-all" fixed strategy. It ensures that batteries with different aging levels receive a suitable activation solution while preventing damage from ineffective or over-activation, improving the targeting and safety of activation.

[0085] In some embodiments, the composite current output by the moderate frequency conversion oscillation mode has an AC component frequency that varies periodically in a sinusoidal pattern between 0.5Hz and 5Hz, with a variation period of 60 seconds and an amplitude of 15% of the rated charging current; the DC bias component has an amplitude of 5% of the rated charging current.

[0086] In this embodiment, after the activation strategy decision engine selects the moderate frequency conversion oscillation mode, the waveform generation and power execution module generates a specific composite current waveform according to preset parameters. This composite current consists of an AC component and a DC bias component: the frequency of the AC component varies periodically with a sinusoidal pattern between 0.5Hz and 5Hz, with a variation period set to 60 seconds. This frequency range can adapt to different degrees of slight sulfate crystallization inside the battery—the low-frequency range is beneficial for acting on deep crystals, while the high-frequency range is convenient for breaking down surface crystals. The sinusoidal variation can avoid excessive local reactions at a fixed frequency. The amplitude of the AC component is 15% of the rated charging current, which can provide sufficient energy to promote crystal dissolution without triggering a violent electrochemical reaction that would cause a temperature rise. The amplitude of the DC bias component is 5% of the rated charging current. Its function is to maintain the basic state of charge of the battery during the activation process, avoiding excessive decrease in the battery's state of charge due to simple AC activation, which would affect the power supply requirements of subsequent equipment. The generated composite current waveform is processed by a pulse width modulation drive circuit and a power MOSFET array, and then the current ripple is reduced by an output filter network (cutoff frequency 10Hz) to ensure stable output current before being applied to the target battery. Simultaneously, the system monitors the battery's voltage and temperature changes in real time, dynamically fine-tuning the frequency and amplitude of the AC component to ensure the waveform always adapts to the battery's real-time state. This composite current design balances repair effectiveness and battery protection; the frequency conversion characteristic improves crystallization breakdown efficiency, and the DC bias ensures charge safety. Compared to fixed frequency or single waveforms, it can more accurately address minor aging issues, improving activation efficiency while reducing battery wear.

[0087] This invention also provides a system for implementing the online dynamic activation method for batteries as described above, comprising:

[0088] A multi-source state sensing module is used to synchronously collect the battery's electrical and thermodynamic parameters in real time.

[0089] A data fusion and preprocessing module, connected to the sensing module, is used to perform spatiotemporal alignment and noise filtering on the raw data stream;

[0090] A dynamic health status assessment module, connected to the preprocessing module, is used to calculate and output a comprehensive health index;

[0091] An activation strategy decision engine, connected to the evaluation module, is used to adaptively select the activation mode based on the comprehensive health index.

[0092] The waveform generation and power execution module, connected to the decision engine, is used to generate and apply a precise activation current waveform;

[0093] A safety monitoring and interruption protection module is connected to the power execution module and the sensing module to monitor in real time and ensure operational safety.

[0094] The effect evaluation and cycle scheduling module is connected to the evaluation module and the decision engine, and is used to evaluate the activation effect and dynamically adjust subsequent strategies.

[0095] In this embodiment, the system consists of a multi-source state perception module, a data fusion and preprocessing module, a health status dynamic assessment module, an activation strategy decision engine, a waveform generation and power execution module, a safety monitoring and interruption protection module, and an effect evaluation and cycle scheduling module. These modules work collaboratively via an internal bus to form a complete closed-loop process. The multi-source state perception module employs a high-precision voltage sensor (range 0-5V, resolution better than 1mV), a Hall effect current sensor (range -50A to +50A, linearity error less than 0.5%), a surface-mount platinum resistance temperature sensor (temperature range -40℃ to +125℃, accuracy ±0.5℃), and a high-frequency AC excitation internal resistance measurement circuit (operating frequency 1000Hz, error less than 2%) to synchronously acquire the battery's electrical and thermodynamic parameters in real time, forming a multi-modal raw data stream. After receiving the raw data, the data fusion and preprocessing module performs timestamp calibration (error <10ms), linear interpolation point filling, outlier removal using a 3x standard deviation criterion, and sliding window smoothing filtering to output a highly consistent observation sequence. The dynamic health status assessment module incorporates a calculator and a weighted fusion unit to calculate the voltage plateau attenuation rate, capacity retention rate, and internal resistance growth rate, and outputs a comprehensive health index based on weights. The activation strategy decision engine stores a predefined decision tree and outputs an activation mode identifier based on the health index. The waveform generation and power execution module, based on the identifier, generates control signals using a digital signal processor. These signals are then passed through a pulse width modulation drive circuit and a power MOSFET array (parallel redundancy, on-resistance <5mΩ) to output waveforms. A filtering network ensures ripple <5%. The safety monitoring and interruption protection module independently monitors parameter change rates and immediately cuts off power output if thresholds are exceeded. The effect evaluation and cycle scheduling module recalculates the health index after activation and adjusts subsequent strategies according to gain. The system achieves in-situ online activation, ensuring continuous equipment operation without disassembly. Modular design improves reliability, and closed-loop control throughout the process enhances activation accuracy and safety while reducing maintenance costs.

[0096] In some implementations, the security monitoring and interruption protection module adopts a dual-channel redundant architecture, including:

[0097] The main monitoring channel is used to compare the instantaneous rate of change of voltage, temperature, and internal resistance with the preset fixed safety threshold in real time.

[0098] The auxiliary early warning channel is used to perform linear fitting on the historical data of the parameters, calculate the trend slope, and issue an early warning when the slope continuously exceeds the warning threshold.

[0099] If any channel triggers an abnormal condition, the system will immediately cut off the power output.

[0100] In this embodiment, the safety monitoring and interruption protection module operates independently of the system's main control loop, employing a dual-channel redundant architecture of a main monitoring channel and an auxiliary early warning channel to provide dual protection for the activation process. The main monitoring channel receives battery voltage, temperature, and internal resistance data transmitted from the multi-source state sensing module in real time. It calculates the instantaneous change rate of each parameter through high-frequency sampling and compares it with preset fixed safety thresholds: if the voltage drops by more than 0.3V within 1 second, the temperature rises by more than 5°C within 1 minute, or the internal resistance increases by more than 10% within 1 minute, an anomaly is immediately identified, triggering the protection mechanism. The auxiliary early warning channel performs a sliding window linear fitting (window length 30 sampling points) on historical data of voltage, temperature, and internal resistance, calculating the slope of the parameter change trend. When the absolute value of the slope continuously exceeds the preset early warning threshold for three consecutive windows, an early warning signal is issued to predict potential risks. Regardless of whether the main channel triggers an anomaly protection or the auxiliary channel triggers an early warning, the module will immediately generate a hardware interrupt signal, directly cutting off the drive power supply to the power MOSFET array in the waveform generation and power execution module, physically disconnecting the activation current output path, and simultaneously sending an interrupt request to the system main controller. The main controller then terminates the activation program, enters a safe standby state, and logs the anomaly type, trigger time, and instantaneous parameter values ​​in non-volatile memory. This dual-channel design avoids safety vulnerabilities caused by single-channel failure. The main channel responds quickly to sudden anomalies, while the auxiliary channel provides early warnings of potential risks, fundamentally preventing battery damage or thermal runaway caused by over-discharge, overheating, or sudden changes in internal resistance during activation, significantly improving system operational safety.

[0101] In some implementations, the waveform generation and power execution module includes a digital signal processor, a pulse width modulation drive circuit, a power MOSFET array with parallel redundancy, and an output filter network; the digital signal processor has a built-in current closed-loop feedback controller and uses a proportional-integral algorithm to ensure that the tracking error between the output current and the target waveform is less than 2%.

[0102] In this embodiment, the waveform generation and power execution module consists of a digital signal processor (DSP), a pulse width modulation (PWM) drive circuit, a parallel redundant power MOSFET array, and an output filter network. The core employs current closed-loop control to ensure accurate waveform output. Upon receiving the mode identifier from the activation strategy decision engine, the DSP retrieves the waveform parameters corresponding to the activation mode from its built-in parameter table (such as the amplitude, frequency, and duty cycle of a mild pulse, and the frequency range and component amplitude of a moderate frequency conversion) to generate a basic timing control signal. Simultaneously, the DSP incorporates a current closed-loop feedback controller, employing a proportional-integral (PI) adjustment algorithm to acquire the actual current signal at the power output terminal in real time. It calculates the deviation from the target waveform command and dynamically adjusts the pulse width and frequency of the control signal based on the deviation value, ensuring that the actual output current closely tracks the target waveform. The PWM drive circuit converts the adjusted control signal into a gate level that meets the driving requirements of the power MOSFET array, driving the devices in the array to turn on and off sequentially. The power MOSFET array uses a parallel redundant design, with a total on-resistance of less than 5mΩ, improving current output capability and module reliability, and preventing activation interruption due to a single device failure. The output filter network has a cutoff frequency set to 10Hz to filter the power output current, effectively suppressing high-frequency ripple and ensuring that the output current ripple is less than 5%. This results in a smooth and precise activation current waveform being applied to the target battery. The module uses closed-loop control to keep the error between the actual output and the target waveform within 2%. Combined with redundant power devices and the filter network, this ensures accurate injection of activation energy, avoiding low activation efficiency or localized battery damage caused by waveform distortion, thus improving activation performance and module reliability.

[0103] In some implementations, the system is integrated within a standard battery management unit housing and seamlessly connected to the device's main control system via a bus interface; the activation strategy decision engine supports receiving remotely issued strategy update instructions via a wireless communication interface, enabling online adaptive optimization of the activation logic.

[0104] In this embodiment, the entire system is encapsulated within a standard battery management unit housing. The housing dimensions and interfaces conform to industry-standard specifications. It seamlessly connects to the existing equipment control system via a CAN or RS485 bus interface, requiring no modification to the battery's physical structure, external wiring, or internal circuitry. This enables plug-and-play deployment and adapts to existing equipment in various scenarios, such as data center backup power, communication base station energy storage, and industrial automation, reducing the difficulty and cost of retrofitting. The activation strategy decision engine has a built-in wireless communication interface (supporting LoRa or 4G), supporting the reception of strategy update commands from a remote operation and maintenance platform. Operation and maintenance personnel can remotely adjust the decision tree's threshold parameters (such as SHI segment thresholds), waveform parameters for each activation mode (such as the duty cycle of a mild pulse, the frequency change period of a moderate frequency conversion, and the current ratio of deep discharge), and even update the decision tree branch logic based on the battery chemistry (such as lead-acid, lithium-ion), the load characteristics of the application scenario, and long-term activation data feedback, without needing to disassemble the system or replace hardware on-site. Upon receiving the command, the decision engine automatically verifies its integrity and validity. If correct, it updates the built-in policy parameters and generates an update log, which is then sent back to the remote platform, enabling online adaptive optimization of the activation logic. This design enhances system flexibility and versatility. Standard packaging and bus interfaces simplify deployment, and remote policy updates adapt to different battery types and scenario requirements, avoiding activation effect degradation due to fixed policies and extending system lifespan and applicability.

[0105] In some implementations, the system supports coordinated activation management of multiple battery packs, scheduling the activation sequence of each battery cell through a central coordinator to avoid multiple cells performing high-power activation operations simultaneously, thereby maintaining the stability of the system bus voltage.

[0106] In this embodiment, the system adopts a distributed architecture to support collaborative activation management of multiple battery packs. Each battery cell is equipped with an independent multi-source state sensing module and a local control unit, responsible for collecting its own state data, calculating the health index, and reporting activation requirements to the central coordinator. The central coordinator, as the core scheduling unit, receives real-time health status (comprehensive health index), activation mode requirements, and system bus voltage data from each battery cell, and formulates a collaborative activation timing plan based on preset scheduling logic. During scheduling, the central coordinator prioritizes battery cells with poor health status (e.g., SHI in the 0.7-0.85 range) to perform activation operations, ensuring that aging batteries are repaired in a timely manner. Simultaneously, it strictly staggers the time when multiple cells simultaneously execute high-power activation modes (e.g., deep stepped discharge and recharge mode)—because high-power modes consume large currents, simultaneous execution can easily cause significant fluctuations in the system bus voltage, affecting the power supply to the main equipment. For example, when cell 1 performs deep stepped discharge, cells 2, 3, etc., can only execute mild pulse or medium frequency conversion modes, or remain in standby mode, and the next cell requiring high-power activation is scheduled only after cell 1 has completed activation. During activation, the central coordinator monitors the system bus voltage in real time. If the voltage is detected to be lower than a preset threshold, the operation of the currently activated high-power unit is immediately suspended to prioritize the stability of the bus voltage. Once the voltage recovers, the timing is readjusted. This collaborative scheduling mechanism avoids bus voltage fluctuations caused by the simultaneous activation of multiple batteries, ensuring stable power supply to the main equipment. At the same time, it enables differentiated activation management for each battery unit, improving overall activation efficiency and battery pack lifespan.

[0107] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices, apparatuses, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0108] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for online dynamic activation of a battery, characterized in that, Includes the following steps: Step S1: Through the embedded sensing unit, the terminal voltage, charging and discharging current, surface temperature and AC internal resistance parameters of the target battery are collected in real time to form a multi-mode battery status raw data stream; Step S2: Perform spatiotemporal alignment and adaptive noise filtering on the original multimodal battery state data stream to generate a highly consistent battery state observation sequence; Step S3: Based on the battery state observation sequence, dynamically calculate the voltage plateau decay rate, capacity retention rate, and internal resistance growth rate through a parallel computing architecture, and then weight and fuse them to output a comprehensive health index; Step S4: Input the comprehensive health index into a predefined activation strategy decision tree. The decision tree contains multiple mutually exclusive branches based on health index thresholds, thereby adaptively selecting an activation mode that precisely matches the current battery state. Step S5: Based on the selected activation mode, generate and apply the corresponding composite current waveform to the target battery, wherein the waveform parameters are dynamically adjusted based on the real-time state of the battery; Step S6: During the activation process, monitor the transient response of the battery's voltage, temperature, and internal resistance independently and in real time. If the rate of change of any parameter exceeds the preset safety threshold, the activation process is immediately interrupted. Step S7: After a single activation cycle ends, reassess the comprehensive health index, calculate the activation gain, and dynamically adjust the intensity and cycle of subsequent activation plans based on the gain.

2. The online dynamic activation method for batteries according to claim 1, characterized in that, The spatiotemporal alignment and adaptive noise filtering process described in step S2 specifically includes: Based on the system master clock, timestamp calibration and synchronization are performed on sensor data with different sampling periods; Linear interpolation is used to fill in the points in the asynchronous data stream and unify it to the reference frequency; Based on the statistical characteristics of the data within the sliding window, a dynamic standard deviation threshold is used to identify and remove outliers. The processed data sequence is smoothed by applying a moving average filter.

3. The online dynamic activation method for batteries according to claim 1, characterized in that, The formula for calculating the comprehensive health index in step S3 is as follows: SHI=α×VDR+β×CRR+γ×(1-IRR); Wherein, SHI is the comprehensive health index, VDR is the voltage plateau decay rate, CRR is the capacity retention rate, IRR is the internal resistance growth rate, and α, β, and γ are weighting coefficients, satisfying α+β+γ=1, where α=0.4, β=0.3, and γ=0.

3.

4. The online dynamic activation method for batteries according to claim 1, characterized in that, The activation strategy decision tree described in step S4 includes: First node: If SHI > 0.95, select the mild pulse repair mode and output a low amplitude, low duty cycle square wave pulse; Second node: If 0.85≤SHI≤0.95, then select the medium frequency conversion oscillation mode and output an AC superimposed DC composite current with periodically changing frequency; Third node: If 0.7≤SHI<0.85, then select the deep step-discharge and recharge mode, and immediately switch to constant current and constant voltage charging after performing step-discharge. Fourth node: If SHI < 0.7, then select the forced hibernation protection mode and generate a battery replacement warning signal.

5. The online dynamic activation method for batteries according to claim 4, characterized in that, The composite current output by the medium frequency conversion oscillation mode has an AC component frequency that varies periodically in a sinusoidal pattern between 0.5Hz and 5Hz, with a variation period of 60 seconds and an amplitude of 15% of the rated charging current; the DC bias component has an amplitude of 5% of the rated charging current.

6. A system for implementing the online dynamic activation method for batteries according to any one of claims 1-5, characterized in that, include: A multi-source state sensing module is used to synchronously collect the battery's electrical and thermodynamic parameters in real time. A data fusion and preprocessing module, connected to the sensing module, is used to perform spatiotemporal alignment and noise filtering on the raw data stream; A dynamic health status assessment module, connected to the preprocessing module, is used to calculate and output a comprehensive health index; An activation strategy decision engine, connected to the evaluation module, is used to adaptively select the activation mode based on the comprehensive health index. A waveform generation and power execution module, connected to the decision engine, is used to generate and apply a precise activation current waveform; A safety monitoring and interruption protection module is connected to the power execution module and the sensing module to monitor in real time and ensure operational safety. The effect evaluation and cycle scheduling module is connected to the evaluation module and the decision engine, and is used to evaluate the activation effect and dynamically adjust subsequent strategies.

7. The system according to claim 6, characterized in that, The security monitoring and interruption protection module adopts a dual-channel redundant architecture, including: The main monitoring channel is used to compare the instantaneous rate of change of voltage, temperature, and internal resistance with the preset fixed safety threshold in real time. The auxiliary early warning channel is used to perform linear fitting on the historical data of the parameters, calculate the trend slope, and issue an early warning when the slope continuously exceeds the warning threshold. If any channel triggers an abnormal condition, the system will immediately cut off the power output.

8. The system according to claim 6, characterized in that, The waveform generation and power execution module includes a digital signal processor, a pulse width modulation drive circuit, a power MOSFET array with parallel redundancy design, and an output filter network; the digital signal processor has a built-in current closed-loop feedback controller and uses a proportional-integral algorithm to ensure that the tracking error between the output current and the target waveform is less than 2%.

9. The system according to claim 6, characterized in that, The system is integrated within a standard battery management unit housing and seamlessly connected to the device's main control system via a bus interface. The activation strategy decision engine supports receiving remotely issued strategy update commands via a wireless communication interface, enabling online adaptive optimization of the activation logic.

10. The system according to claim 6, characterized in that, The system supports collaborative activation management of multiple battery packs. It schedules the activation sequence of each battery cell through a central coordinator to avoid multiple cells performing high-power activation operations at the same time, thereby maintaining the stability of the system bus voltage.

Images