A battery pack charging equalization control system based on dynamic impedance matching

Abstract

The application discloses a battery pack charging equalization control system based on dynamic impedance matching, comprising a data acquisition module, a contact fault diagnosis module connected to the data acquisition module, a system state decision module connected to the contact fault diagnosis module, a dynamic impedance online identification module connected to the data acquisition module, an intelligent equalization decision module connected to the contact fault diagnosis module, the system state decision module and the dynamic impedance online identification module respectively, and an equalization execution module connected to the intelligent equalization decision module. In the application, the sampling contact fault can be early and accurately identified, and graded alarm and system degraded operation can be automatically triggered according to the fault severity, so that safety accidents caused by measurement errors are prevented from the control strategy level, and the tolerance of the system to hardware defects is greatly improved.

Description

Technical Field

[0001] This invention relates to the field of battery testing technology, specifically to a battery pack charging equalization control system based on dynamic impedance matching. Background Technology

[0002] Active balancing technology is a core means to improve the usable capacity and lifespan of battery packs. Currently, most mainstream balancing strategies rely on the terminal voltage or state of charge (SOC) of individual battery cells. However, in practical applications, the terminals of the voltage sampling harness may develop abnormal contact resistance due to oxidation, loosening, or other reasons. This introduces significant measurement errors, causing the battery management system (BMS) to make misjudgments based on incorrect voltage data. On the one hand, it may fail to identify cells with truly high internal resistance that require balancing; on the other hand, it may perform unnecessary balancing operations on healthy cells. This not only reduces balancing efficiency and wastes energy but may also lead to safety hazards by masking the true risk of overcharging. Therefore, this paper proposes a battery pack charging balancing control system based on dynamic impedance matching to address these issues. Summary of the Invention

[0003] The purpose of this invention is to provide a battery pack charging equalization control system based on dynamic impedance matching to solve the problems mentioned in the background art.

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

[0005] A battery pack charging equalization control system based on dynamic impedance matching includes:

[0006] The data acquisition module is configured to synchronously acquire the terminal voltage of each battery cell in the battery pack, the total current flowing through the battery pack, and the temperature of each battery cell.

[0007] The contact fault diagnosis module is connected to the data acquisition module and is configured to calculate the real-time contact fault confidence of each voltage sampling channel based on the acquired voltage, current and temperature signals by fusing the outputs of steady-state voltage consistency analysis, dynamic impedance anomaly analysis and multivariate statistical process control.

[0008] The system status decision module is connected to the contact fault diagnosis module and is configured to dynamically switch the system operating status to one of the normal state, degraded state, safe state or failure protection state according to the contact fault confidence level, and generate system control parameters including the maximum allowable charge and discharge current limit accordingly.

[0009] The dynamic impedance online identification module is connected to the data acquisition module and is configured to apply the recursive least squares algorithm to identify and update the impedance model parameters of each battery cell in real time. The parameters include at least the ohmic internal resistance and the charge transfer impedance.

[0010] The intelligent balancing decision module is connected to the contact fault diagnosis module, the system status decision module, and the dynamic impedance online identification module, and is configured to perform the following operations:

[0011] a) Calculate the charge acceptance capability and health index of each battery cell based on the impedance model parameters;

[0012] b) To minimize the difference in state of charge between individual battery cells and the energy consumption for equalization, a constrained optimization problem model is constructed to generate an initial equalization current command.

[0013] c) Based on the contact fault confidence level, the initial equalization current command is corrected by confidence weighting, and a conservative equalization strategy is applied to the high fault confidence channel to generate the final equalization current command.

[0014] The equalization execution module is connected to the intelligent equalization decision module and is configured to drive the active equalization topology circuit to perform energy transfer operations on the battery pack according to the final equalization current command.

[0015] Preferably, the contact fault diagnosis module specifically includes:

[0016] The steady-state analysis unit is configured to calculate the temperature-compensated deviation between the voltage of each individual unit and the average value within the group when the system is stationary, and output a steady-state voltage consistency score after exponential weighted moving average filtering.

[0017] The dynamic analysis unit is configured to estimate the dynamic resistance of each channel and compare it with the benchmark within the group when the current changes, and output a dynamic impedance anomaly score.

[0018] The statistical analysis unit is configured to perform principal component analysis on all voltage measurements and calculate the Hotelling T for each channel. 2 Statistical parameters and squared prediction error are used to output a multivariate statistical process control score.

[0019] The confidence fusion unit is configured to adaptively adjust the weights according to the real-time current magnitude, and perform weighted fusion of the steady-state voltage consistency score, dynamic impedance anomaly score, and multivariate statistical process control score to output the contact fault confidence.

[0020] Preferably, the system state decision module is preset with a first threshold, a second threshold, and a third threshold, wherein the third threshold > the second threshold > the first threshold, specifically set as follows:

[0021] When the contact failure confidence of all channels is lower than the first threshold, maintain or switch to normal state;

[0022] When the contact failure confidence level of any channel reaches or exceeds the first threshold but is lower than the second threshold, the system switches to a degraded state and limits the maximum charging current to 80% of the rated value.

[0023] When the contact failure confidence level of any channel reaches or exceeds the second threshold but is lower than the third threshold, switch to a safe state, limit the maximum charging current to 50% of the rated value, and disable the fast charging function.

[0024] When the contact fault confidence level of any channel reaches or exceeds the third threshold, and the instantaneous current is greater than 50% of the rated current, the system switches to the fail-safe state and issues a command to cut off the main relay.

[0025] Preferably, the online dynamic impedance identification module employs a recursive least squares algorithm with a forgetting factor, wherein the forgetting factor... The system adaptively adjusts based on battery temperature, using the following formula:

[0026]

[0027] in, Basic forgetting factor, For temperature coefficient, The current temperature. The reference temperature is used to accelerate parameter updates when temperature changes drastically.

[0028] Preferably, the charging acceptance capability of the individual battery cells in the intelligent balancing decision module is specifically as follows:

[0029]

[0030] in, For the first monomers in Charging acceptance factor at any given time At maximum state of charge, To estimate the current state of charge, and These are the ohmic internal resistance and charge transfer impedance provided by the dynamic impedance online identification module, respectively.

[0031] Preferably, the constrained optimization problem model in the intelligent equilibrium decision module mainly includes establishing and solving the optimization objective function. Specifically:

[0032]

[0033] The constraints of the optimization problem include: energy conservation constraints. Balanced current amplitude constraint Charged state boundary constraints

[0034] ;

[0035] in, For the first The equalization current of each individual cell The change in state of charge caused by this equilibrium current. For the penalty weighting coefficient, The target average state of charge.

[0036] Preferably, the logic of the intelligent balancing decision module for weighted correction of the initial balancing current command is as follows:

[0037] First, the measurement confidence weight for each channel is calculated, specifically as follows:

[0038] ,in, For the preset attenuation coefficient, For contact fault confidence;

[0039] Then, the initial optimal equilibrium current obtained by solving the optimization problem is used. Generate the final equalization current command .

[0040] Preferably, the data acquisition module includes:

[0041] A high-precision analog-to-digital converter for synchronous sampling of voltage and temperature signals;

[0042] A bidirectional Hall effect current sensor is used to measure the total current.

[0043] Redundant sampling circuitry is configured to switch to an alternative sampling path to obtain alternative voltage data for sampling channels diagnosed as having a high fault confidence level.

[0044] Preferably, a battery pack charging equalization control method based on dynamic impedance matching mainly includes the following steps:

[0045] S1: Synchronously collects voltage, current, and temperature data of the battery pack;

[0046] S2: Perform contact fault diagnosis and calculate the fault confidence of each sampling channel;

[0047] S3: Determine the current system operating status based on the confidence level and adaptively adjust the system power limit;

[0048] S4: Identify the dynamic impedance parameters of each battery cell online;

[0049] S5: Evaluate the charging acceptance capability of each cell based on impedance parameters, and calculate the optimal equalization current for anti-interference by solving the optimization model, taking into account the fault confidence weighting.

[0050] S6: Executes the equalization current command to control the operation of the equalization circuit.

[0051] Compared with the prior art, the beneficial effects of the present invention are:

[0052] This invention utilizes model fusion diagnostics to identify sampling contact faults early and accurately, and automatically triggers tiered alarms and system degradation operations (such as current limiting and switching redundant channels) based on the severity of the fault. This prevents safety accidents caused by measurement errors from a control strategy perspective, greatly improving the system's tolerance to hardware defects. For channels with high fault confidence, by reducing the weight of their data in decision-making, using model predictions as substitutes, and applying balanced current limits, the system can still perform relatively safe and effective actions even in cases of partial "sensory inaccuracy," ensuring the availability of core functions. Attached Figure Description

[0053] Figure 1 This is a schematic diagram of the system structure of the present invention. Detailed Implementation

[0054] 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 skilled in the art without creative effort are within the scope of protection of the present invention.

[0055] Please see Figure 1 The present invention provides a technical solution:

[0056] A battery pack charging equalization control system based on dynamic impedance matching comprises six core modules. These modules interact with each other via shared memory and event flags. Specifically:

[0057] Data acquisition module: responsible for driving the underlying hardware and acquiring raw signals;

[0058] The contact fault diagnosis module, dynamic impedance online identification module, and intelligent balancing decision module are the core algorithm modules that realize state perception and decision-making.

[0059] System status decision module: acts as the overall scheduler for system operation;

[0060] Equalizer module: responsible for converting digital instructions into actual analog control signals.

[0061] The data acquisition module is implemented through hardware circuitry and underlying driver software. The hardware mainly consists of voltage and temperature sampling, current sampling, and redundant sampling design.

[0062] Voltage and temperature sampling uses the BQ79616 chip as the analog front-end. Six of these chips are daisy-chained together to achieve synchronous acquisition of voltage and temperature data from 96 individual cells, with a fixed sampling period of 10ms. The positive and negative input terminals of each sampling channel are directly soldered to the electrodes of the individual cells using Kelvin connections to minimize the influence of lead resistance.

[0063] The current sampling uses a HAS 200-S bidirectional Hall current sensor with a range of ±300A and a bandwidth of 10kHz, outputting an analog voltage signal. This signal is connected to the high-precision 16-bit ADC dedicated channel of the main control MCU.

[0064] For every module consisting of 8 battery cells, an additional set of spare sampling harnesses is deployed, connected to a redundant channel of another BQ79616 chip. The spare channel is disabled by default and is only switched on by software commands when a fault is diagnosed in the main channel.

[0065] The specific implementation steps of the underlying driver software of the data acquisition module are as follows: 1. Configure the ADC to use a timer-triggered synchronous sampling mode to ensure that the sampling time of voltage and current is strictly aligned; 2. Apply a first-order low-pass digital filter (cutoff frequency 100Hz) to the raw sampled data to suppress switching noise; 3. Implement the "zero current calibration" function: when the system is powered on and initialized, if the total current is detected to be less than 0.5A for 3 seconds, the output value of the current sensor at this time is recorded as "zero bias" and subtracted in real time in subsequent measurements.

[0066] The contact fault diagnosis module performs internal processing through its steady-state analysis unit, dynamic analysis unit, statistical analysis unit, and confidence fusion unit, with an operating cycle of 50ms. The steady-state analysis unit process is as follows:

[0067] When the system current If <5A (considered static) continues for more than 5 seconds, initiate steady-state analysis;

[0068] Calculate the average voltage within the group. ;

[0069] Calculate the deviation after temperature compensation for each individual unit. ,in ;

[0070] Applying EWMA filtering ;

[0071] Perform steady-state scoring ,in This is the standard deviation of the long-term historical steady-state deviation of the monomer (initial value is 1mV).

[0072] The workflow of the dynamic analysis unit is as follows:

[0073] Real-time monitoring of current changes. When a step change in current is detected... At the same time, record the voltage and current at 5 sampling points before and after the event;

[0074] Calculate the dynamic resistance estimate ;

[0075] Calculate the dynamic resistance reference within the calculation group (Use the median to resist outliers);

[0076] Perform dynamic scoring ,in .

[0077] The workflow of the statistical analysis unit is as follows:

[0078] This is performed once every 100 sampling cycles (i.e., 1 second);

[0079] The 10 sets of voltage measurements in the past second were used to construct a 10×n matrix, and PCA analysis was performed. The first 3 principal components (cumulative contribution rate > 95%) were retained.

[0080] Calculate each channel The statistical measure and the SPE statistic are compared with their control limits at a 99% confidence level.

[0081] Statistical scoring .

[0082] The process of the confidence fusion unit is as follows:

[0083] Based on the current absolute value Dynamically adjust weights:

[0084] like <5A, then ;

[0085] like >5A, then ;

[0086] Calculate the overall confidence level ;

[0087] Output Limit the amplitude to ensure it is within the range of [0, 1].

[0088] The system state decision module operates as a state machine, receiving data from the diagnostic module. The threshold is set to The transition process for each state is as follows:

[0089] Under normal conditions, all control parameters are unlimited, and the equalization function is fully enabled. When When the time comes, the system will be transferred to a downgraded state;

[0090] A yellow warning is issued in the downgraded state, limiting the charging current. .like If it lasts for 60 seconds, it will return to normal; if If so, then the system will transition to a safe state;

[0091] When the system is in a safe state, an orange alarm will be issued, a fault code will be recorded, and the charging current will be limited. Disable fast charging (i.e., reject charging requests higher than 0.5C). If If it lasts for 300 seconds, it will return to a downgraded state; if If the instantaneous current is greater than 0.5*I_rated, then the system will switch to fail-safe mode.

[0092] In the failover protection state, a red alarm and audible / visual alert are issued, and a "Power Down Now" request is sent via the CAN bus. If no response is received within 200ms, the hardware is directly driven to disconnect the main positive and main negative relays. This state requires manual reset to exit.

[0093] The dynamic impedance online identification module operates with a 1-second cycle and uses a first-order RC equivalent circuit model. First, the model is discretized, specifically... , ,in , , The model is then built using the RLS algorithm, and finally the health index is calculated. Initial internal resistance The battery is calibrated and stored during the factory capacity assessment. The RLS algorithm is used to establish the model as follows:

[0094] Introducing parameter vectors:

[0095] Regression vector:

[0096] Output: Enable adaptive forgetting factor and set a base value. Upper and lower limits The strategy is adjusted according to the following adjustment: if the battery temperature change rate is > 1°C / min, then... If the rate of change of SOC is > 1% / min, then Other cases This ensures rapid tracking under dynamic operating conditions and smooth estimation under steady-state conditions.

[0097] After each iteration, the result is obtained by inverse calculation. , .

[0098] The intelligent balancing decision module is triggered under two events: 1) entering the charging state; 2) a change in the system state. Its decision-making process is as follows:

[0099] Perform data preprocessing, for The channel uses interpolation between adjacent cells to obtain a replacement voltage value. According to the formula Calculate the confidence weights to ultimately derive the voltage used for decision-making: ;

[0100] The charge acceptance capability is calculated by using the preprocessed voltage and the identified internal resistance, and then estimated in real time through an extended Kalman filter (EKF). The acceptance coefficient is calculated from this. The larger this value is and the smaller the internal resistance, the more it can accept large current charging;

[0101] Establish the optimization objective function

[0102]

[0103] The optimal equilibrium current vector is obtained by solving the problem under the following constraints. The constraint is an energy conservation constraint. Balanced current amplitude constraint Charged state boundary constraints ;

[0104] The final execution current is obtained as follows: .

[0105] The hardware of the balancing execution module adopts a distributed active balancing architecture. Each battery cell is equipped with an independent bidirectional Buck-Boost circuit, with its primary side connected to the cell and all secondary sides connected in parallel to a common bus. The common bus voltage is controlled at half of the total battery pack voltage. The software of the balancing execution module receives input from the intelligent balancing decision module. The command adjusts the duty cycle of the corresponding switching transistors of each individual unit through the PID controller, so that the measured equal current tracks the command value. The PID parameters are fed forward compensation based on the bus voltage and the individual unit voltage.

[0106] Taking a 96-cell battery pack fast charging application as an example, the system is initially in normal condition. After charging begins, the current reaches 150A. Due to a subtle looseness at the sampling terminal of cell 23 (contact resistance of approximately 10mΩ), an abnormal voltage drop of 1.5V is generated under the influence of the current, and the following process is performed:

[0107] 1. The dynamic analysis unit quickly calculated that the dynamic resistance of cell 23 was abnormally high. The sudden increase, due to the large current, leads to the fusion weight β becoming dominant, resulting in... Within seconds .

[0108] 2. The system enters a degraded state, reducing the charging current limit from 200A to 160A and issuing a warning.

[0109] 3. The intelligent balancing decision module will assign measurement weights to cell 23. The SOC is estimated by relying on the interpolation voltage of adjacent cells 22 and 24. Based on this, the optimization algorithm generates balancing instructions to either charge the healthy cells more or perform very conservative discharge balancing on cell 23 to avoid decision-making errors.

[0110] 4. If the loosening worsens, If the voltage continues to rise above 0.9, the system enters a safe state, and the current is limited to 100A to forcibly reduce the risk of overheating. If the final contact point overheats, causing a loose connection in the sampling line, the voltage will jump, reaching 1.0 while the current remains high. If the current reaches 1.0 and is still relatively high, the system will enter a failure protection state, disconnect the circuit, and prevent the accident from escalating.

[0111] Contents not described in detail in this specification are existing technologies known to those skilled in the art. Standard parts used in this invention can be purchased commercially, and irregularly shaped parts can be custom-made according to the description and drawings. The specific connection methods for each part all employ conventional methods such as bolts, rivets, and welding, which are already mature technologies. The machinery, parts, and equipment all use conventional models from the prior art, and the circuit connections also employ conventional connection methods from the prior art, which will not be detailed here.

[0112] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A battery pack charging equalization control system based on dynamic impedance matching, characterized in that, include: The data acquisition module is configured to synchronously acquire the terminal voltage of each battery cell in the battery pack, the total current flowing through the battery pack, and the temperature of each battery cell. The contact fault diagnosis module is connected to the data acquisition module and is configured to calculate the real-time contact fault confidence of each voltage sampling channel based on the acquired voltage, current and temperature signals by fusing the outputs of steady-state voltage consistency analysis, dynamic impedance anomaly analysis and multivariate statistical process control. The system status decision module is connected to the contact fault diagnosis module and is configured to dynamically switch the system operating status to one of the normal state, degraded state, safe state or failure protection state according to the contact fault confidence level, and generate system control parameters including the maximum allowable charge and discharge current limit accordingly. The dynamic impedance online identification module is connected to the data acquisition module and is configured to apply the recursive least squares algorithm to identify and update the impedance model parameters of each battery cell in real time. The parameters include at least the ohmic internal resistance and the charge transfer impedance. The intelligent balancing decision module is connected to the contact fault diagnosis module, the system status decision module, and the dynamic impedance online identification module, and is configured to perform the following operations: a) Calculate the charge acceptance capability and health index of each battery cell based on the impedance model parameters; b) To minimize the difference in state of charge between individual battery cells and the energy consumption for equalization, a constrained optimization problem model is constructed to generate an initial equalization current command. c) Based on the contact fault confidence level, the initial equalization current command is corrected by confidence weighting, and a conservative equalization strategy is applied to the high fault confidence channel to generate the final equalization current command. The equalization execution module is connected to the intelligent equalization decision module and is configured to drive the active equalization topology circuit to perform energy transfer operations on the battery pack according to the final equalization current command.

2. The battery pack charging equalization control system based on dynamic impedance matching according to claim 1, characterized in that, The contact fault diagnosis module specifically includes: The steady-state analysis unit is configured to calculate the temperature-compensated deviation between the voltage of each individual unit and the average value within the group when the system is stationary, and output a steady-state voltage consistency score after exponential weighted moving average filtering. The dynamic analysis unit is configured to estimate the dynamic resistance of each channel and compare it with the benchmark within the group when the current changes, and output a dynamic impedance anomaly score. The statistical analysis unit is configured to perform principal component analysis on all voltage measurements and calculate the Hotelling T for each channel. 2 Statistical parameters and squared prediction error are used to output a multivariate statistical process control score. The confidence fusion unit is configured to adaptively adjust the weights according to the real-time current magnitude, and perform weighted fusion of the steady-state voltage consistency score, dynamic impedance anomaly score, and multivariate statistical process control score to output the contact fault confidence.

3. The battery pack charging equalization control system based on dynamic impedance matching according to claim 1, characterized in that, The system state decision module is preset with a first threshold, a second threshold, and a third threshold, wherein the third threshold > the second threshold > the first threshold, and is specifically configured as follows: When the contact failure confidence of all channels is lower than the first threshold, maintain or switch to normal state; When the contact failure confidence level of any channel reaches or exceeds the first threshold but is lower than the second threshold, the system switches to a degraded state and limits the maximum charging current to 80% of the rated value. When the contact failure confidence level of any channel reaches or exceeds the second threshold but is lower than the third threshold, switch to a safe state, limit the maximum charging current to 50% of the rated value, and disable the fast charging function. When the contact fault confidence level of any channel reaches or exceeds the third threshold, and the instantaneous current is greater than 50% of the rated current, the system switches to the fail-safe state and issues a command to cut off the main relay.

4. A battery pack charging equalization control system based on dynamic impedance matching according to claim 1, characterized in that, The dynamic impedance online identification module employs a recursive least squares algorithm with a forgetting factor. The system adaptively adjusts based on battery temperature, using the following formula: ； in, Basic forgetting factor, For temperature coefficient, The current temperature. The reference temperature is used to accelerate parameter updates when temperature changes drastically.

5. A battery pack charging equalization control system based on dynamic impedance matching according to claim 1, characterized in that, The charging acceptance capability of the individual battery cells in the intelligent balancing decision module is specifically as follows: ； in, For the first monomers in Charging acceptance factor at any given time At maximum state of charge, To estimate the current state of charge, and These are the ohmic internal resistance and charge transfer impedance provided by the dynamic impedance online identification module, respectively.

6. A battery pack charging equalization control system based on dynamic impedance matching according to claim 1, characterized in that: The constrained optimization problem model in the intelligent equilibrium decision-making module mainly includes establishing and solving the optimization objective function. Specifically: ； The constraints of the optimization problem include: energy conservation constraints. Balanced current amplitude constraint Charged state boundary constraints ; in, For the first The equalization current of each individual cell The change in state of charge caused by this equilibrium current. For the penalty weighting coefficient, The target average state of charge.

7. A battery pack charging equalization control system based on dynamic impedance matching according to claim 1, characterized in that, The logic of the intelligent equalization decision module in weighting and correcting the initial equalization current command is as follows: First, the measurement confidence weight for each channel is calculated, specifically as follows: ,in, For the preset attenuation coefficient, For contact fault confidence; Then, the initial optimal equilibrium current obtained by solving the optimization problem is used. Generate the final equalization current command .

8. A battery pack charging equalization control system based on dynamic impedance matching according to claim 1, characterized in that, The data acquisition module includes: A high-precision analog-to-digital converter for synchronous sampling of voltage and temperature signals; A bidirectional Hall effect current sensor is used to measure the total current. Redundant sampling circuitry is configured to switch to an alternative sampling path to obtain alternative voltage data for sampling channels diagnosed as having a high fault confidence level.

9. A control method for a battery pack charging equalization control system based on dynamic impedance matching according to any one of claims 1-7, characterized in that, The main steps include: S1: Synchronously collects voltage, current, and temperature data of the battery pack; S2: Perform contact fault diagnosis and calculate the fault confidence of each sampling channel; S3: Determine the current system operating status based on the confidence level and adaptively adjust the system power limit; S4: Identify the dynamic impedance parameters of each battery cell online; S5: Evaluate the charging acceptance capability of each cell based on impedance parameters, and calculate the optimal equalization current for anti-interference by solving the optimization model, taking into account the fault confidence weighting. S6: Executes the equalization current command to control the operation of the equalization circuit.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program is used to implement the control logic of the battery pack charging equalization control system based on dynamic impedance matching as described in any one of claims 1 to 7.

Images