A current curve control method based on battery detection constant voltage power supply

Abstract

The application discloses a current curve control method based on a constant-voltage power supply of battery detection, and particularly relates to the technical field of battery detection, and comprises the following steps: after receiving detection process information and loading a control strategy, the output current of the power supply is controlled to increase the voltage to a target value and stably keep, and a constant-voltage control stage is entered; in the process of continuously decreasing the current, voltage and current change data are collected, a voltage response coefficient and a current smoothing coefficient are constructed, and a control factor is generated based on the two, which is used for dynamically correcting the current decrease rate, and finally the detection process is ended after the current is lower than a termination threshold; by setting a target voltage and a current value and loading a control strategy, the application realizes accurate regulation and control of the output of the power supply; in the constant-voltage stage, a response coefficient and a smoothing coefficient are introduced, a control factor is generated by reasoning, the current decrease rate is dynamically corrected, the output current is smoothly decreased in an exponential manner, the voltage stability and the current curve continuity are improved, and sharp fluctuations are avoided.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of battery detection, more particularly, the present application relates to a current curve control method based on a battery detection constant voltage power supply. BACKGROUND

[0002] At present, in the battery production and detection link, the constant voltage charging process has been widely used in battery capacity analysis, internal resistance evaluation and life prediction and other key scenes. The traditional battery detection constant voltage power supply system usually relies on the feedback loop composed of PWM control chip, DAC digital analog output, ADC sampling circuit and MCU controller, and realizes the automatic adjustment of the current by hardware setting or fixed strategy, so that the battery terminal voltage is maintained in the preset target voltage range.

[0003] In the existing control method, the constant voltage control stage generally adopts the following process: the system first raises the output voltage to the set target voltage value, then maintains the constant voltage state, and tries to reduce the output current in a slow way, so that the current curve gradually decreases to meet the standard battery charging / detection curve requirements. This method can theoretically compare the real-time value of the detection current with the standard current, and rely on the MCU to adjust the DAC output to realize the dynamic balance of the current.

[0004] In actual application, due to the inaccurate control of current drop gradient, limited resolution of DAC response and other problems, often leading to burr, peak or mutation in the current drop process. This kind of non-continuous change not only interferes with the accurate identification of the battery state by the host computer, but also hides risks in high rate, high energy density battery detection, which may induce heat runaway, trigger protection mechanism, and even cause safety accidents. In addition, the traditional method lacks the ability to analyze and control the voltage fluctuation and current curve change, and often cannot realize the fine control of the "overall smoothness" of the current curve under constant voltage state.

[0005] More importantly, most of the current methods set the current regulation speed as linear or fixed rate, ignoring the response difference of the battery at different stages to the current change, and failing to adaptively adjust the current drop strategy according to the real-time sampling results. Therefore, the "non-smooth section" in the current change curve is difficult to avoid, which affects the final detection accuracy and data availability. Therefore, the present application proposes a current curve control method based on a battery detection constant voltage power supply to solve the above problems. SUMMARY

[0006] To achieve the above purpose, the present application provides the following technical scheme:

[0007] A current curve control method based on a battery detection constant voltage power supply, comprising the following steps:

[0008] Receiving the detection procedure information and start instruction sent by the host computer, setting the target voltage value and target current value, and loading the corresponding control strategy;

[0009] Before starting the detection procedure, integrity verification is performed on the received procedure information, and the system running state is detected to ensure the safety and executability of the procedure;

[0010] The output current of the power supply is controlled to increase the output voltage until the set target voltage value is reached and stably maintained, entering the constant voltage control phase;

[0011] In the state of the constant voltage control phase, the current output is started to be reduced, and the initial adjustment strategy is to reduce the current reference value according to the set current reduction gradient in the initial stage;

[0012] During the continuous reduction of the current, the change data of the output voltage and current are collected in real time, wherein, according to the amplitude and direction of the voltage fluctuation, a response coefficient for representing the voltage stability is determined; at the same time, according to the continuity and smoothness of the current change, a smoothing coefficient for representing the current fluctuation characteristics is determined;

[0013] The response coefficient and the smoothing coefficient are used as input parameters to generate a control factor through an inference process, which is used to comprehensively reflect the current voltage stability state and current change characteristics, so as to dynamically correct the rhythm of the current reduction;

[0014] According to the control factor, the change rate of the current reference value is continuously adjusted, so that the output current is gradually reduced in an exponential manner that meets the set smoothing requirement, which not only maintains the voltage within the target range, but also ensures that the current change process has no sharp and abrupt points, thereby obtaining a smooth and continuous current curve;

[0015] When the current is reduced to below the preset termination threshold, the control output of the voltage and current is terminated, the corresponding execution circuit is closed, and the battery detection procedure is ended.

[0016] In a preferred embodiment, the integrity verification of the received procedure information means performing the following operations according to the preset standards respectively:

[0017] Procedure parameter verification for preventing instruction errors;

[0018] Communication link verification for preventing data distortion;

[0019] Hardware running detection for preventing abnormal output of equipment.

[0020] In a preferred embodiment, entering the constant voltage control phase means:

[0021] After the detection process starts, the controller calculates an initial target output current value according to the difference between the preset target voltage value and the current battery terminal voltage; the target current value is converted into a continuously changing analog control voltage signal by a digital-to-analog converter and sent to a power stage control unit in the power supply system;

[0022] The power stage control unit converts the control voltage into a duty cycle adjustment amount of the PWM signal based on a preset linear mapping relationship after receiving the control voltage signal, controls the on-time of the power switch device; as the PWM duty cycle gradually increases, the voltage at the output terminal continues to rise;

[0023] When the output voltage first reaches the set target voltage value and enters its allowable deviation range, the system determines that the constant voltage condition is met and enters the constant voltage control phase.

[0024] In a preferred embodiment, during the continuous current drop, the output voltage trend in each sampling period is dynamically monitored, and a voltage response coefficient is constructed based on the collected historical voltage sequence data to characterize the stability of the current voltage state. The construction method of the response coefficient includes:

[0025] First, the difference between the current period voltage value and the previous period voltage value is obtained, and the difference is allocated to the corresponding fluctuation grade score in the preset voltage tolerance interval;

[0026] Then, the mean and variance of the voltage fluctuation grade score are calculated in multiple adjacent sampling periods to characterize the consistency and severity of voltage changes;

[0027] The mean value of the fluctuation grade score is multiplied by the fluctuation variance value to form a voltage fluctuation severity factor;

[0028] A sliding rate measurement value based on the historical voltage sampling window is introduced to characterize the slow change trend of the voltage;

[0029] The voltage fluctuation severity factor is divided by the sliding rate measurement value, and a minimum limit value is set to avoid division by zero, and the voltage response coefficient of the current period is calculated.

[0030] In a preferred embodiment, during the continuous current drop, a current smoothing coefficient is constructed for each sampling period to measure the continuity and fluctuation characteristics of the current change, and the coefficient is used to adjust the current drop amplitude of the current period to avoid curve glitches, spikes or cliff-like mutations. The calculation method of the smoothing coefficient includes:

[0031] The actual current sampling values ​​of multiple consecutive cycles are extracted within the sliding time window to construct the original current sequence. By performing linear fitting or local trend extraction on the sequence, the trend curve of current changing with time is obtained. The original sampling values ​​and the corresponding trend values ​​are then differentially calculated to form a residual sequence representing the part of the current that deviates from the trend.

[0032] A peak distribution map is constructed based on the residual sequence, and the number and amplitude of peak positions are identified and scored. The entropy value of the residual sequence is then calculated as a measure of the complexity of current change. The peak scores and entropy values ​​are weighted and fused to form a fluctuation evaluation factor. The fluctuation evaluation factor is normalized with the average current decrease within the sampling period to form the final current smoothing coefficient.

[0033] When calculating the entropy value of the residual sequence, the entropy value is obtained by calculating the information entropy of the probability distribution of the residual sequence, and is used to characterize the overall fluctuation complexity of the residual sequence in the current period.

[0034] In a preferred embodiment, constructing a peak distribution map refers to:

[0035] In the residual sequence, the difference between each sampling period and the previous period is calculated as the instantaneous fluctuation amplitude of the current.

[0036] All fluctuation amplitudes in the residual sequence are grouped and statistically analyzed according to preset amplitude intervals. A fluctuation frequency histogram model representing the fluctuation intensity distribution is constructed to determine the frequency of occurrence of outliers with large fluctuation amplitudes in the overall sequence. The number and amplitude intervals of peak positions are identified and subjected to concentration analysis. Scoring is performed based on preset multi-level scoring rules to quantify the spike density and severity of current changes in the current cycle.

[0037] In a preferred embodiment, generating a control factor through a reasoning process means:

[0038] The voltage response coefficient and current smoothing coefficient are normalized to fall into a unified evaluation scale range. Then, the voltage response coefficient and current smoothing coefficient are used as input variables for fuzzy inference and mapped to a preset fuzzy level range through the membership function built into the fuzzy inference to form the input state vector of the current cycle.

[0039] Based on the matching result of the input vector in the rule matrix, the state type corresponding to the current operating state is determined by fuzzy rule reasoning. The preset current reduction adjustment ratio is extracted from the current adjustment strategy mapping table that is compatible with the state type. The current reduction adjustment ratio is applied to the current reference value adjustment range of the next cycle to form the current control factor of the next cycle.

[0040] In a preferred embodiment, continuously adjusting the rate of change of the current reference value means:

[0041] After the sampling of the current control cycle ends, the current drop adjustment factor obtained by reasoning in the current cycle is passed to the reference control logic of the next cycle, and dynamically superimposed with the original preset current drop control curve to generate a current regulation factor with time correction significance.

[0042] The rate of change of the target current reference value in the next cycle is continuously corrected exponentially using the current regulation factor. Specifically, this includes multiplying the original current descent gradient with a nonlinear convergence function constructed with the current regulation factor as the core parameter, and calculating the reference change correction magnitude for current control in the next cycle.

[0043] In the subsequent real-time sampling and feedback process, the correction magnitude of the reference change is applied to the power stage control unit, so that the actual output current exhibits a feedback-driven dynamic correction change over multiple control cycles.

[0044] The technical effects and advantages of this invention are as follows:

[0045] This invention sets target voltage and current values ​​and loads corresponding control strategies before the detection process begins, enabling the system to have clear control objectives and operational expectations from the initial detection stage. This effectively avoids abnormal output voltage or current fluctuations caused by improper initial parameter settings. Furthermore, by receiving detection process information and start commands from a host computer, the entire system can operate collaboratively with an external management platform, exhibiting excellent programmability and remote control capabilities. With the help of preset control strategies, the system can automatically adjust parameters for different battery types or application scenarios during the detection process, improving the targeting and flexibility of the detection process. Through the coordinated operation of the above mechanisms, the battery detection initialization process is more stable and reliable, effectively reducing the risk of misoperation and system failure probability during the start-up phase, laying the foundation for subsequent high-precision constant voltage control and current curve optimization.

[0046] This invention sets a specific target voltage value during the stage of controlling the power supply output current to increase the output voltage. It uses the stability of the output voltage within this target range as a prerequisite for entering the constant voltage control stage, thus achieving effective management of the voltage response process during the initial charging phase. This entry strategy based on target voltage setting and real-time judgment effectively suppresses abnormal voltage surges caused by external load changes, sudden changes in battery internal resistance, etc., avoiding battery damage or decreased charging efficiency due to premature entry into the constant voltage state. Simultaneously, by gradually increasing the output voltage through controlled output current, the system can gently guide the electrochemical reaction state inside the battery, reducing safety hazards caused by sudden voltage increases. The stable achievement of the constant voltage condition provides a good voltage platform for the smoothness of subsequent current curve control, further improving the safety and scientific rigor of the entire battery testing process.

[0047] This invention introduces a dynamic sampling mechanism based on output voltage and current change data during the constant voltage control stage. By calculating the voltage response coefficient and the current smoothness coefficient, a two-factor control model reflecting the current voltage stability and current change smoothness is constructed. Based on this, the system inputs these two dynamic coefficients into the inference process to generate a control factor used to correct the rate of decrease of the current reference value in real time. This mechanism breaks through the limitations of traditional single static current control methods, enabling dynamic adjustment of the current change strategy based on real-time feedback from the sampled data. Through the continuous action of this control factor, the system achieves an exponential smooth decrease in the current curve. Even in the presence of external disturbances or battery parameter fluctuations, it effectively avoids the generation of abnormal waveforms such as spikes, abrupt changes, and glitches, thereby ensuring the data continuity and physical stability of the entire detection process, improving the analyzability of the curve and the safety redundancy of system operation. Attached Figure Description

[0048] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings;

[0049] Figure 1 This is a schematic diagram of a current curve control method based on a battery-detected constant voltage power supply in this invention. Detailed Implementation

[0050] 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.

[0051] Reference Figure 1 The following examples were obtained:

[0052] Example 1: A current curve control method based on a constant voltage power supply for battery testing, comprising the following steps: receiving testing process information and start command sent by a host computer, setting target voltage and target current values, and loading corresponding control strategies; the significance of this step is to establish a communication mechanism between the system and the host computer, realizing the parameter distribution and initialization configuration of the testing task. Setting the target voltage and target current values ​​as the core control targets of this test, while loading the control strategy provides the logical basis for subsequent voltage regulation, current control, and feedback correction, ensuring that the entire process is executed according to the predetermined mode.

[0053] Before initiating the detection process, the received process information is verified for integrity, and the system's operational status is checked to ensure the process is safe and executable. This step ensures the accuracy of the process information and the reliability of the system operation. Integrity verification is used to eliminate problems such as communication errors and data transmission interruptions, preventing control logic abnormalities caused by missing instructions or abnormal formats. System operational status detection, on the other hand, detects the response status of hardware and software modules to avoid initiating the detection process under conditions of equipment malfunction, power malfunction, or communication malfunction, thereby preventing safety risks such as accidental output triggering or thermal runaway.

[0054] The power supply output current is controlled to increase the output voltage until the set target voltage value is reached and maintained stably, entering the constant voltage control stage. The purpose of this step is to drive the output voltage to gradually increase through current regulation, achieving precise approach and dynamic locking to the target voltage value. Maintaining stability means that after the output voltage first reaches the target value, the control system enters a voltage feedback regulation state, keeping the voltage stable within the allowable deviation range, thus forming the starting condition for the constant voltage control stage. This creates a stable voltage foundation for subsequent precise control of the current decrease behavior.

[0055] During the constant voltage control phase, the current output begins to decrease. Initially, the current reference value is reduced according to a set current decrease gradient as the initial adjustment strategy. This step signifies the formal entry into the current control phase, where the current reference value is gradually reduced based on the system's preset control gradient, forming the initial curve control structure. This initial adjustment strategy can be pre-set based on the load model, battery characteristics, or empirical parameters, leaving room for subsequent dynamic corrections. Simultaneously, reducing the current helps avoid overcharging risks and achieves stable energy release.

[0056] During the continuous decrease of current, real-time data on the changes in output voltage and current are collected. A response coefficient characterizing voltage stability is determined based on the amplitude and direction of voltage fluctuations; simultaneously, a smoothing coefficient characterizing current fluctuations is determined based on the continuity and stability of current changes. This step utilizes a dual-parameter extraction mechanism to acquire key state variables of the system output within each control cycle. The voltage response coefficient is used to determine whether there are drastic voltage jumps or overshoot, while the smoothing coefficient is used to identify whether spikes or abrupt fluctuations occur during the current decrease. Both coefficients are constructed from historical sampling data, reflecting the instantaneous state and trend characteristics of output changes, and serve as the foundational input for subsequent intelligent control.

[0057] By using the response coefficient and smoothing coefficient as input parameters, a control factor is generated through an inference process. This factor comprehensively reflects the current voltage stability and current variation characteristics, thereby dynamically correcting the rate of current decrease. This step enables intelligent judgment and adaptive adjustment of the control logic. The control factor, as a fusion output of the response coefficient and smoothing coefficient, is used by the inference model to analyze and evaluate the current system operating state. This control factor is targeted and timely, dynamically adjusting the rate of decrease of the current reference value to adapt to fluctuations in the actual output state, achieving more refined closed-loop control.

[0058] Based on the control factor, the rate of change of the current reference value is continuously adjusted so that the output current decreases exponentially in a manner that meets the set smoothness requirements. This maintains the voltage within the target range while ensuring that the current change process is free of spikes and abrupt changes, thus obtaining a smooth and continuous current curve. The significance of this step lies in reducing the current in a non-constant but predictable exponential manner to achieve the goal of controlling the "smoothness" of the current curve. Through the continuous adjustment of the control factor, the output current is prevented from dropping too quickly or exhibiting abnormal changes such as jumps or rebounds, ensuring the continuity and controllability of the curve recognized by the host computer. At the same time, maintaining the voltage within the target range avoids voltage fluctuations caused by current regulation, thereby improving the overall detection stability and equipment response quality.

[0059] When the current drops below the preset termination threshold, the control output of voltage and current is terminated, the corresponding execution circuit is shut down, and the battery detection process ends. This step serves as the final control segment of the detection process, achieving orderly termination of the current control process through threshold judgment. When the current drops below the termination set value, the system uniformly disconnects the control output, simultaneously shutting down the power stage or other execution circuits, thus disconnecting the output loop and ensuring complete and safe termination of the detection. This termination mechanism prevents residual output or false triggering, helping to extend equipment life and improve battery safety management.

[0060] The system receives detection process information and start commands from the host computer, sets target voltage and current values, and loads corresponding control strategies. Specifically, this includes: First, establishing a stable command transmission channel with the host computer through a preset data communication interface, and monitoring the issuance of detection tasks in real time through this channel. When the host computer sends detection process information and start commands, the system immediately parses the information packet and extracts the task parameters contained within. The process information includes at least: detection type, execution step sequence, target voltage and current values ​​required for each stage, constant voltage time limit, termination threshold, and control strategy identifier. After receiving the command information, the system automatically completes process caching and generates a unique process identifier for the current detection task to ensure the traceability and execution continuity of the control logic at each stage.

[0061] Upon receiving the process information, the system will simultaneously set the corresponding target voltage and target current values. This setting process maps the preset values ​​to the parameters based on the step configuration items in the process instructions. For example, in a standardized testing scenario for a single lithium battery, the target voltage can be set to 4.20 volts and the target current to 3.0 amperes. The system will then combine the currently set target parameters and load the corresponding control strategy. Loading the corresponding control strategy actually means that after the control system receives the testing process information and start command from the host computer, it selects and loads a set of preset control behavior logic, i.e., the "control strategy," based on parameters such as the specific testing scenario, battery type, cell status, and voltage target. For example, different control strategies are applied depending on the battery type (lithium-ion, lithium iron phosphate, nickel-metal hydride, etc.), cell capacity, internal resistance, testing conditions (room temperature, high temperature, low temperature), charge / discharge rate setting, whether it is an aging test, or a cycle test. For instance, when testing a new high-energy-density battery, the system may apply a strategy of "gradually increasing voltage, rapidly decreasing current, and tolerating fluctuations"; while for a severely aged battery, the system may apply a strategy of "slowly increasing voltage, gently decreasing current, and strictly limiting voltage".

[0062] The host computer is responsible for the macro-level scheduling and control management of the entire testing process, including but not limited to: defining testing steps (such as the voltage boost stage, constant voltage stage, current decrease stage, etc.); configuring target parameters (voltage, current, time) for each stage; issuing start or stop commands; and reading, recording, analyzing, and displaying feedback data. The host computer typically communicates with the slave computer via serial communication (UART), CAN bus, USB, or RS485 to send control commands and receive data feedback. In practical applications, the host computer can be: an industrial PC (IPC); an embedded computing platform (such as a Raspberry Pi or Jetson Nano); a general-purpose laptop or PC; or a tablet device with control software.

[0063] The software system running within the host computer is often: a real-time control platform (such as LabVIEW or PLC control interface); customer-customized software (such as a control system developed using C++ / Python); a data acquisition and analysis system (such as a SCADA interface); or a main control module embedded in a specific battery management system (BMS). In the technical background of this invention, the "host computer" is the initiator and controller of the entire detection and control process. It transmits the detection strategy and target parameters to the intermediate computer or lower-level control system via data communication, which then executes specific power regulation and feedback control. Therefore, the "host computer" typically does not directly control hardware execution but rather undertakes the control logic at the instruction and process levels, and can be considered the "main brain" of the system.

[0064] Integrity verification of the received process information refers to performing the following operations according to preset standards: For each instruction parameter included in the detection process, process parameter verification is performed to prevent instruction errors. This step compares key parameters such as the target voltage value, target current value, duration, step number, and sampling period carried by the detection process with the legal ranges in the preset configuration template one by one to ensure that the parameters are not empty, do not exceed limits, do not conflict, and do not jump. For example, if the set target voltage value exceeds the maximum load capacity of the equipment (e.g., higher than 60 volts), the system will immediately prevent the process from loading and throw a high-priority exception to prevent the equipment from burning out due to overvoltage.

[0065] After the process parameter verification is completed, the system immediately initiates communication link verification to prevent data distortion. This step involves comparing the integrity and redundancy of the original instruction data packets from the host computer, including but not limited to CRC cyclic redundancy check verification, data frame header and footer integrity detection, and data length consistency judgment, to ensure that no abnormalities such as bit flipping, frame loss, or pseudo-instruction embedding occur during data transmission. For example, if a data bit verification error is detected in the received complete process configuration packet, the process will not be written to the task queue, the system will record an exception log, and request the host computer to reissue the process instructions.

[0066] After the communication link verification is successful, the system continues to perform hardware operation checks to prevent abnormal device output. This check process covers real-time self-test logic for the power stage module, current sampling circuit, voltage detection loop, and control chip status. In particular, before the system enters the execution process, it simulates output control commands and reads feedback signals to determine whether the response characteristics of the execution circuit are stable and whether the deviation is within the allowable range. For example, the system will simulate sending a zero-current output command and read the return value from the current feedback channel. If the returned current is not zero or has a significant offset (such as an offset exceeding 1 ampere), the system determines that there may be hardware faults such as leakage, short circuit, or bias imbalance at the output end and immediately terminates the startup process.

[0067] Only after all three verifications are completed and the results meet the system's preset standards is the system considered to have completed the integrity verification of the received process information. Then, the process start permission is unlocked, and the detection process is officially transferred to the control logic execution stage. This ensures that the subsequent control process starts and runs in a safe, reliable, and predictable system state, minimizing the risk of detection accidents caused by misoperation, instruction mismatch, or sudden circuit abnormalities.

[0068] Entering the constant voltage control stage refers to the following: After the detection process officially begins, the system calculates the difference between the target voltage value set in the detection process information and the battery terminal voltage currently collected in real time through the voltage sampling loop. Based on this difference, the system internally executes a current output model initialization process. This process determines an initial target output current value for driving the voltage rise by looking up a table or calculating a formula. For example, when the system detects a target voltage value of 4.20 volts and the current battery voltage is 3.850 volts, the system determines the difference to be 0.350 volts. Combining this with the output current corresponding to this voltage difference in the set boost response model, which is 3.0 amperes, the system sets the initial target output current to 3.0 amperes. This target output current is not directly output but is first transmitted to the digital-to-analog converter module inside the controller, converted into a continuously changing analog control voltage signal. This voltage signal serves as a key control quantity for driving the power output of the power supply system and will be transmitted to the power stage control unit for response adjustment in subsequent stages.

[0069] The target current is converted into an analog control voltage signal by a digital-to-analog converter and immediately sent to the power stage control unit in the power supply system. The power stage control unit establishes a mapping connection between the control voltage signal and the duty cycle change of the PWM signal based on a preset linear mapping relationship. PWM (Pulse Width Modulation) signal refers to a control method that adjusts the average energy output of the control signal by controlling the proportion of the high-level duration within one cycle to the total cycle time (i.e., the "duty cycle"). This mapping relationship can be set through system calibration; for example, a PWM duty cycle of 10% corresponds to a control voltage of 0.5 volts, and a PWM duty cycle of 60% corresponds to a control voltage of 2.5 volts, increasing linearly. In actual execution, the control voltage signal is sampled in real time and input to the PWM signal modulation module. The PWM signal output by this module dynamically adjusts its duty cycle according to the value of the control voltage. The duty cycle of the PWM signal refers to the proportion of time the PWM waveform remains high within each modulation cycle, which determines the conduction time of the power switching device. By adjusting the PWM duty cycle, the system controls the conduction state of the power transistors in the power stage output circuit, causing the output voltage to gradually increase at a predetermined rhythm. For example, when the control voltage increases to one volt, the duty cycle rises to twenty percent, and the power supply output voltage begins to climb from 3.850 volts to 3.950 volts.

[0070] As the control voltage signal continuously increases and the PWM duty cycle gradually rises, the power transistor's on-time extends, and the output voltage linearly increases accordingly. During this process, the system samples the output voltage in real time each cycle and continuously compares the sampled voltage value with the target voltage value. When the system detects that the output voltage first reaches the set target voltage value and enters its set allowable deviation range, it considers the constant voltage condition achieved. This allowable deviation range is typically set based on the battery type and detection requirements. For example, when the target voltage is 4.20 volts, the allowable deviation range can be set to ±0.020 volts, meaning the voltage is between 4.180 volts and 4.220 volts. If the system detects that the output voltage remains stable at 4.200 volts in a sampling, the system determines that it has entered the constant voltage control stage and locks the target voltage value as the reference for subsequent adjustments. This determination process not only relies on a single sampling result but can also enhance the robustness of stability judgment by setting the number of consecutive samplings that meet the condition (e.g., three consecutive cycles of sampled values ​​are all within the set range).

[0071] After determining that the system has entered the constant voltage control phase, it locks the current PWM duty cycle as the voltage holding reference and gradually shifts control to the current control logic to achieve a gradual decrease in the output current. In this phase, the system no longer actively increases the control voltage or changes the duty cycle to boost the voltage. Instead, it switches voltage control to a holding state, fine-tuning through real-time voltage feedback. Small corrections are only performed when the voltage fluctuates slightly, ensuring the voltage remains within the set target value and its allowable range. Simultaneously, the system initiates a current reference value descent control program, using a "set current descent gradient" as the initial control strategy. This allows the current output to decrease systematically while maintaining a constant voltage, creating the input prerequisites for subsequent dynamic coefficient adjustments, exponential descent control, and spike suppression mechanisms. The activation of current control signifies that the system has successfully transitioned from the stage of "controlling the power supply output current to increase the output voltage" to the constant voltage control stage of "stable output voltage maintenance and controlling the rhythm of current changes," laying the foundation for a smooth and safe detection process.

[0072] During the constant voltage control phase, the system begins to reduce the current output to achieve dynamic convergence control of the battery charging or detection process. To ensure sufficient smoothness and controllability of the current decrease process, the system pre-sets a set of current decrease gradient parameters for the initial stage. The current decrease gradient is a static setpoint representing the magnitude by which the current reference value should decrease per unit time, used to construct the reference control curve in the initial stage of the decrease. This gradient is typically calculated offline based on factors such as target current, target voltage, battery capacity, and internal impedance, and serves as the initial adjustment basis for the constant voltage phase. Its unit can be amperes per second or amperes per control cycle.

[0073] The current descent gradient, as part of the initial regulation strategy, has been invoked and bound by the system during the loading phase of the corresponding control strategy, and is part of the controller's preset control strategy. The initial regulation strategy includes not only the specific value of the current descent gradient, but also parameter settings such as application timing and control cycle interval. After the voltage reaches the set target, the system automatically enters the constant voltage control phase and prioritizes the initial downward adjustment of the current reference value according to this initial regulation strategy. This lays the foundation for subsequent dynamic fine-tuning control driven by the response coefficient and smoothness coefficient, ensuring that the overall current curve has good stability and continuity in the initial descent phase.

[0074] During the continuous decrease of current, the output voltage change trend within each sampling period is dynamically monitored, and a voltage response coefficient is constructed based on the collected historical voltage sequence data to characterize the stability of the current voltage state. The construction method of this response coefficient includes: the system synchronously collects the current value of the output voltage within each current sampling period, and calculates the difference between the current period's voltage value and the previous period's voltage value to obtain the instantaneous voltage change amplitude within that period. Subsequently, the system assigns a fluctuation level score to this difference based on its proportion within a preset voltage tolerance range. For example, if the target voltage is 4.200 volts and the allowable tolerance range is ±0.020 volts, the system divides the entire tolerance range into five levels, corresponding to different intervals such as the absolute value of the difference between 0 and 0.004 volts, and 0.004 volts to 0.008 volts. When the voltage change in a certain period is 0.007 volts, the system can determine its fluctuation level as level three and assign it a fluctuation level score of three. Through this mechanism, the voltage change in each period will be classified into a specific level, obtaining discrete quantization characteristics for subsequent statistical processing.

[0075] Within each current period, statistical analysis is performed on the fluctuation level score sequences from multiple adjacent sampling periods that have been categorized, extracting their mean and variance. This step aims to quantitatively characterize the consistency and severity of voltage fluctuation trends. The mean of the fluctuation level score reflects the average severity of voltage changes within a certain time window, while the variance reveals the dispersion and abrupt changes in voltage fluctuations. For example, within a sliding window (such as six periods), if the obtained fluctuation level sequence is "3, 2, 3, 4, 3, 2", its mean is 2.83, and its variance is approximately 0.39. The system multiplies this mean and variance to calculate a comprehensive index representing the degree of voltage fluctuation within this time window, called the voltage fluctuation severity factor. The larger this factor, the more significant and unstable the voltage changes within that period, providing a basic parameter for the sensitivity control of subsequent current regulation.

[0076] After obtaining the voltage fluctuation severity factor, the system further introduces a new dynamic quantity, namely the sliding rate measurement, to characterize the slow trend of voltage change, thus eliminating abnormal response judgments that may be caused by minor disturbances during the cycle. The system performs first-order derivative processing on the voltage value within the sliding window, calculating the average voltage change rate between consecutive cycles to obtain the overall voltage change rate within that time period. For example, if the current six-cycle voltage sampling values ​​are "4.201, 4.200, 4.199, 4.198, 4.200, 4.202", the system obtains the change rate sequence "-0.001, -0.001, -0.001, 0.002, 0.002" through differential processing, with an average rate of zero, indicating that the voltage remains generally stable. This sliding rate measurement will serve as a normalization factor between the voltage fluctuation and trend of the current cycle, introduced into the next stage for the final construction of the voltage response coefficient.

[0077] After ensuring that the voltage fluctuation severity factor and the sliding rate measurement value are constructed separately, the final fusion calculation step is performed. That is, the voltage fluctuation severity factor is divided by the sliding rate measurement value to form the final numerical index characterizing the voltage state of the current cycle, which is the voltage response coefficient. A larger coefficient indicates more severe and rapid voltage fluctuations, requiring the system to improve its ability to suppress the current drop rhythm; a smaller coefficient indicates a more stable voltage, allowing for a more relaxed current drop restriction and improved control efficiency. During this process, to avoid mathematical instability caused by the sliding rate measurement value approaching zero and the denominator approaching zero, the system introduces a set minimum limit (e.g., 1 x 10^-4 volts per cycle). When the rate is below this value, the limit is forcibly used as the denominator in the calculation, ensuring computational stability and logical safety. The finally calculated voltage response coefficient, along with the current smoothing coefficient, is input into the subsequent inference model to complete the joint decision on the current control rhythm.

[0078] As the current reference value continuously decreases, a current smoothing coefficient is constructed for each sampling period to measure the continuity and fluctuation characteristics of current changes. This coefficient adjusts the current decrease amplitude in the current period to avoid spikes, jarring changes, or abrupt drops in the curve. The calculation method for this smoothing coefficient includes: after the system controller enters the constant voltage control phase and begins executing the current reference value reduction logic, the system extracts actual current sampling values ​​from multiple consecutive periods within a sliding time window, constructing a set of original current sequences arranged by time. To analyze the changing trend of this original current sequence, the system uses linear fitting or local polynomial curve fitting to model the relationship between current and time in the sequence, thereby obtaining a trend curve describing the overall current trend in the current period. Subsequently, the system performs a difference operation between each sampling value in the original current sequence and its corresponding trend curve value to calculate the deviation at that point. All these deviations are arranged in sampling order to form a residual sequence describing the deviation characteristics of current fluctuations. This residual sequence is used to capture abnormal current behavior that changes drastically in a short time and serves as the basic data source for subsequent fluctuation analysis and complexity evaluation.

[0079] A peak distribution map is constructed based on the residual sequence to identify and quantify spike characteristics in the current curve. Specifically, the system first calculates the difference between the residual value of each sampling period and the residual value of the previous period to obtain the instantaneous current fluctuation amplitude. Then, all fluctuation amplitudes are grouped and statistically analyzed according to pre-set amplitude threshold intervals, such as "low fluctuation zone," "medium fluctuation zone," and "high fluctuation zone," and a fluctuation frequency histogram is constructed to reflect the frequency of occurrence of fluctuations of different intensities within the current sliding window. Based on this histogram, the system identifies the location and number of peaks with fluctuation amplitudes greater than the set threshold and records the amplitude interval of the peaks. Next, a concentration analysis is performed on all identified peaks to determine whether the peaks are continuously and densely distributed. Based on their number, amplitude level, continuity, and other dimensions, a peak score representing the severity of the current spike is generated according to a multi-level scoring rule. The higher the peak score, the more drastic the current change in the current period, requiring a rapid response from the subsequent control system.

[0080] After completing the peak scoring, to further quantify the overall complexity of current fluctuations, an entropy calculation process for the residual sequence is introduced. This process employs information entropy theory, normalizing the residual sequence and statistically analyzing its probability distribution density, then calculating the entropy value based on the dispersion of the probability distribution. For example, if the frequency distribution of residual values ​​is uniform within a sliding window, the entropy value is high, indicating that the fluctuation distribution has no obvious pattern and strong disorder; if the residual values ​​are concentrated around a few specific values, the entropy value is low, indicating that the current change has a certain regularity and stability. This entropy value, as a core metric reflecting the structural complexity of the residual sequence fluctuations, is complementary to the peak scoring value. The system merges the two according to a weighted factor (e.g., the peak scoring weight is 0.6, and the entropy value weight is 0.4) to obtain a fluctuation evaluation factor used to describe the overall severity of current changes.

[0081] The calculated fluctuation evaluation factor is normalized to the average current decrease within the current sampling period to eliminate the influence of units, magnitudes, and time spans. This normalization process includes: calculating the current decrease amplitude in the current period (e.g., the ampere difference divided by the period time), dividing the fluctuation evaluation factor by the current decrease amplitude, and adding a preset safety offset value (to avoid the risk of the denominator being zero), thus generating the final current smoothing coefficient. This current smoothing coefficient is a continuous value used for dynamic feedback control. A larger value indicates a less stable current curve, and the system will significantly slow down the current decrease rate in the next period; a smaller coefficient indicates lower fluctuation, allowing the current decrease rate to remain constant or moderately accelerate. This coefficient is ultimately fed into the subsequent fuzzy inference module, forming an input variable along with the voltage response coefficient to generate a control factor, thereby precisely adjusting the adjustment rhythm of the current reference value in the next period and achieving continuous optimization of the current output.

[0082] It should be noted that the multi-level scoring rules were configured by the system during the "loading corresponding control strategies" phase, and the strategy parameters were set specifically according to the differences in battery type, power response rate, and detection scenario. Multi-level scoring rules typically include multiple dimensions such as amplitude level classification (e.g., low amplitude is level one, medium amplitude is level two, and high amplitude is level three), peak quantity evaluation (e.g., more than three consecutive peaks indicate high risk), and peak concentration judgment (e.g., peaks appearing densely within a control cycle are given increased weight). In actual operation, the controller quantifies the peak behavior identified within each sliding time window according to this rule system, ultimately generating a peak score value representing the intensity of the current curve spike in the current cycle. This score value combines quantification and adaptability, serving as an important basis for subsequent fluctuation analysis.

[0083] After peak scoring is completed, entropy is introduced as a metric for the complexity of current fluctuations to achieve a comprehensive analysis of the regularity of the overall residual sequence. The system integrates the peak score and entropy value using a weighted factor to calculate a fluctuation evaluation factor that describes the overall severity of current changes. This weighting factor setting also originates from the system initialization phase of "loading the corresponding control strategy." The controller determines the weight ratio based on the characteristics of the actual application scenario. For example, in high-safety-level battery testing, where priority is given to the influence of peak scoring, the peak scoring weight can be set to 0.6 and the entropy weight to 0.4. If more emphasis is placed on the complex changing trends under small disturbances, the ratio can be adjusted in reverse.

[0084] Generating a control factor through the inference process involves normalizing the voltage response coefficient and current smoothing coefficient obtained from the previous cycle. The aim is to bring these two evaluation indicators, with different physical meanings and inconsistent numerical distribution ranges, into a unified evaluation scale, facilitating standardized input for subsequent inference processes. The normalization process employs a range standardization method, mapping sampled values ​​to a closed interval between zero and one. The maximum and minimum values ​​are determined by empirical ranges set during the initial "loading corresponding control strategies" phase of system operation. For example, if the voltage response coefficient's value distribution range is set to zero to five, then a normalized result of 0.5 for a current sampled value of 2.5 is equivalent to 0.5. Similarly, the current smoothing coefficient is normalized within its specific upper and lower limits, ensuring that both coefficients can be included in the same level interval for fuzzy logic membership mapping. This standardization step ensures that different input dimensions have fair influence weights in the fuzzy inference system, avoiding the partial derivative guidance effect caused by differences in numerical magnitudes.

[0085] The normalized voltage response coefficient and current smoothing coefficient are used as input variables for fuzzy inference. These are input into the membership functions built into the fuzzy inference system and mapped to preset fuzzy level intervals, thus forming the input state vector for the current cycle. The membership function can be a trigonometric function or a trapezoidal function. During the "loading corresponding control strategy" phase, the system selects the optimal model based on the target battery's response sensitivity and control accuracy requirements. For example, the fuzzy levels of the voltage response coefficient can be divided into three levels: "stable," "fluctuating," and "violent," with corresponding membership functions defined between zero and one. The current smoothing coefficient is divided into levels such as "stable," "unstable," and "violently fluctuating." The position of each input variable in the membership function determines its corresponding membership degree, which describes the confidence level of the input variable belonging to each level. Based on this, the system combines the levels of the voltage response coefficient and the current smoothing coefficient to form a two-dimensional input state vector, which is then used in the subsequent inference rule matrix matching process.

[0086] Based on the matching results of the input state vector in the rule matrix, the system determines the state type corresponding to the current operating state through a fuzzy rule inference process. The rule matrix is ​​constructed during system initialization based on extensive historical detection data and expert experience, forming a rule mapping set that maps input levels to output strategies. For example, when the voltage response coefficient is at the "severe" level and the current smoothing coefficient is at the "unstable" level, the system infers that the battery may be experiencing a sudden load change or internal instability, requiring immediate slowing of the current drop rate to prevent overshoot or thermal runaway. In this case, the state type retrieved from the rule matrix is ​​"high-risk wave dynamics." Based on this state type, the system extracts a preset current drop adjustment factor from the mapping table configured in the "loading corresponding control strategy" stage. This factor is typically a floating-point number between 0.5 and 0.9, used to correct the current drop rate in the next cycle. For example, an extraction factor of 0.6 indicates that the next cycle should slow the drop rate by 40% based on the original drop curve to improve stability.

[0087] The extracted current drop adjustment factor is applied to the current reference value adjustment range for the next cycle, forming the current control factor for the next cycle. This control factor not only retains the trend inertia of the previous cycle but also introduces state feedback correction capability, possessing the characteristics of time recursion and conditional adaptation. Before the start of the next control cycle, this current control factor is loaded into the current reference calculation module, participating together with the original current drop gradient to generate a new current reference change range. The controller uses this correction range to control the power stage, ultimately achieving dynamic adjustment of the current output, so that the actual output current continuously decreases in a way that better adapts to the target voltage control range and the smoothness requirements of the current curve. If the system judges the state as "stable and smooth" for several consecutive cycles, the current drop adjustment factor tends to one, allowing the system to resume the original drop rate and improve testing efficiency; conversely, if it is continuously in "violent dynamics," the factor remains at a low value to fully protect the safe operation of the battery cells.

[0088] Continuous adjustment of the rate of change of the current reference value refers to the following: after the sampling of the current control cycle ends, the system transmits the current decrease adjustment factor obtained by the fuzzy inference module in that cycle to the reference control logic of the next cycle. This current decrease adjustment factor is a dynamic factor inferred from the comprehensive judgment of the voltage response coefficient and the current smoothing coefficient in the previous cycle, reflecting the regulation requirements of the current change rate under the current system operating state. After transmitting this factor to the control logic of the next cycle, the system does not directly replace the original control command, but dynamically superimposes it with the preset current decrease control curve. This superposition process introduces the concept of "time correction significance," that is, while maintaining the consistency of the overall decreasing trend, short-term fluctuations are allowed to feedback and correct the control output, enabling the system to have disturbance rejection capability and adaptability, thereby establishing a feedback-adjustment-response closed-loop control chain between consecutive cycles.

[0089] By utilizing a current regulation factor, the rate of change of the target current reference value for the next cycle is continuously corrected exponentially. This correction process not only references the originally set current descent gradient but also introduces a nonlinear convergence function to adjust the specific intensity of the descent rate, thereby achieving a smoother and more flexible control over the actual current output. The nonlinear function model can be configured in various ways according to the battery's performance characteristics and operating scenarios, mainly including the following three types:

[0090] The first function type is the negative exponential function model. The basic behavior of this function is: when the input value is small, the output is close to zero; as the input value gradually increases, the output rises rapidly, eventually approaching one, but never exceeding one. Its functional relationship can be described in words as: "one minus the natural logarithm function multiplied by a negative constant factor on the base of the input value." For example, when the current regulation factor is 0.75, substituting it into this function yields an actual adjustment coefficient of approximately 0.63. This model has natural convergence characteristics and is suitable for gradually slowing down the rate of decrease in voltage to maintain stability in scenarios with a high risk of sudden current changes.

[0091] The second function type is the hyperbolic tangent function model. The main characteristic of this function is its symmetry and "S"-shaped trend. Its behavior is as follows: when the input value is zero, the output value is zero; as the input value increases in the positive direction, the output gradually approaches one; as it increases in the negative direction, the output approaches negative one. In practical applications, by multiplying the function's input value by a specific scaling factor, its output value varies between zero and one, thus adapting to the range of current regulation. This function is suitable for scenarios requiring rapid convergence in the middle stage and smooth transitions at the beginning and end, such as for testing high-energy-density batteries, enabling differentiated rate control of current changes at different stages.

[0092] The third function type is the weighted adaptive convergence model: this function is a nonlinear response model dynamically constructed based on external environmental variables. Its form can be described as: "Multiplying the current input value by a weighting factor constructed from feedback data such as battery temperature, voltage fluctuation rate, and current change rate, and then processing it through an exponential or logistic function to obtain the final output result." For example, when the battery temperature shows a significant upward trend, the system automatically reduces the weighting factor and decreases the current adjustment range to avoid overheating risks; while in a stable state with low volatility, the weighting factor is appropriately increased to accelerate the adjustment pace. This function has high adaptability and is suitable for complex and variable battery testing processes.

[0093] The correction result of the selected function output is multiplied by the current preset current descent gradient to obtain the current reference value correction magnitude for the next control cycle. This magnitude will be applied to the power output control module in subsequent control cycles, causing the output current to gradually decrease along a dynamically softened curve, achieving stable voltage, smooth current, and a curve without spikes or jolts, significantly improving the safety and reliability of the battery detection process during the constant voltage phase.

[0094] To implement the control strategy at the hardware level, the system applies the obtained current reference change correction amplitude to the power stage control unit and acts on the generation process of the analog control voltage signal in real time. This analog control voltage signal, as the direct control quantity output by the digital-to-analog converter (DAC), maps to the PWM (Pulse Width Modulation) control signal of the power switching devices in the power supply system, driving fine adjustment of the duty cycle. This allows for fine-tuning of the output current without changing the total control cycle length. For example, if the correction amplitude obtained after a certain cycle sampling phase is 0.15 amps, and based on a historical cycle current value of 6.3 amps, the target reference current will be updated to 6.15 amps in the next cycle. This is then converted to a 3.8-volt control signal by the DAC circuit and finally output to the power switching circuit via the PWM control signal, controlling the MOSFET on-time and thus causing the output current to converge towards the new target.

[0095] In subsequent control cycles, the system recursively applies the correction magnitude of the benchmark change to subsequent cycles, achieving a continuous and dynamic control process. After sampling in each cycle, the system inputs the new voltage response coefficient and current smoothing coefficient into the fuzzy inference module to form the current cycle state judgment and continuously updates the current drop adjustment ratio, enabling the control system to have the capability of "multi-cycle information fusion + rolling update". This iterative process causes the output current to exhibit an exponentially decreasing trend throughout the constant voltage phase. Furthermore, because state feedback judgment is performed in each cycle, this trend has the ability to automatically correct for sudden abnormal states. In the testing of high-power lithium batteries or high-energy-density cells, this control strategy can effectively avoid potential safety risks such as voltage fluctuations and heat surges caused by sudden current drops, improving the stability and safety margin of the entire testing process. Simultaneously, because the curve changes continuously and smoothly, without burrs or spikes, it greatly optimizes the readability of the test data and the analytical value of the host computer records, providing a more accurate current curve basis for subsequent health status assessment and aging modeling.

[0096] When the current drops below the preset termination threshold, the system immediately initiates the termination process's judgment mechanism and verifies the continuity of the currently collected current data. Once it confirms that the current has stably fallen below the termination threshold and is not due to instantaneous fluctuations or sampling errors, the control logic unit issues a termination control command, terminating the voltage and current control outputs to ensure the power stage no longer maintains its output state. Simultaneously, the system shuts down execution circuits associated with the power output, such as power switching devices, digital-to-analog converter output channels, and pulse-width modulation drive units, achieving a safe closed-loop termination of the detection process. Throughout the termination operation, the system retains complete detection data and termination time parameters, uploading them to the host computer via the communication interface for subsequent data archiving and analysis. At this point, the battery detection process officially ends, and the system enters standby mode or prepares for the next detection task according to its configuration.

[0097] The control logic unit refers to the main control processor embedded in the power control system. Specifically, it can be composed of a microcontroller, digital signal processor, or programmable logic device with equivalent control capabilities. Its main functions are: receiving process parameters and control commands from the host computer, scheduling the output paths of voltage and current, and executing core logic tasks such as voltage sampling, current sampling, strategy loading, control factor inference, and current reference correction. Simultaneously, this control logic unit possesses computing resources for caching, filtering, and analyzing sampled data, and is responsible for information exchange with the digital-to-analog converter, analog front-end module, and communication interface. The power stage control unit refers to the power control module in the power system responsible for actually driving the load and regulating the output voltage and current. It typically includes power switching transistors (such as metal-oxide-semiconductor field-effect transistors), PWM waveform generation modules (pulse width modulators), driver chips, and inductor-capacitor energy storage circuits. This power stage control unit receives analog control signals or digital duty cycle adjustment signals from the control logic unit to achieve dynamic regulation of the output power, thereby supporting the entire physical execution process of constant voltage control and smooth current curve decline.

[0098] 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.

[0099] 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.

[0100] 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.

[0101] 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.

[0102] 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 current curve control method based on a battery-detected constant voltage power supply, characterized in that, Includes the following steps: Receive detection process information and start command sent by the host computer, set target voltage and target current values, and load corresponding control strategies; Before starting the detection process, the integrity of the received process information is verified, and the system operation status is detected. The power supply output current is controlled to increase the output voltage until the set target voltage value is reached and maintained stably, thus entering the constant voltage control stage. During the constant voltage control phase, the current output begins to decrease. Initially, the current reference value is reduced according to the set current decrease gradient as the initial adjustment strategy. During the continuous decrease of current, the output voltage and current change data are collected in real time. Based on the amplitude and direction of voltage fluctuation, a response coefficient is determined to characterize voltage stability. At the same time, based on the continuity and stability of current change, a smoothing coefficient is determined to characterize current fluctuation characteristics. The response coefficient and smoothing coefficient are used together as input parameters, and a control factor is generated through the inference process to comprehensively reflect the current voltage stability and current change characteristics, thereby dynamically correcting the current current drop rate. Based on the control factor, the rate of change of the current reference value is continuously adjusted so that the output current decreases exponentially in a manner that meets the set smoothness requirements, thereby obtaining a smooth and continuous current curve; when the current drops below the preset termination threshold, the control output of voltage and current is terminated, the corresponding execution circuit is turned off, and the battery detection process ends. Generating a control factor through a reasoning process refers to: The voltage response coefficient and current smoothing coefficient are normalized to fall into a unified evaluation scale range. Then, the voltage response coefficient and current smoothing coefficient are used as input variables for fuzzy inference and mapped to a preset fuzzy level range through the membership function built into the fuzzy inference to form the input state vector of the current cycle. Based on the matching result of the input vector in the rule matrix, the state type corresponding to the current operating state is determined by fuzzy rule reasoning. The preset current reduction adjustment ratio is extracted from the current adjustment strategy mapping table that is compatible with the state type. The current reduction adjustment ratio is applied to the current reference value adjustment range of the next cycle to form the current control factor of the next cycle.

2. The current curve control method based on a battery-detected constant voltage power supply according to claim 1, characterized in that, Verifying the integrity of received process information involves performing the following operations according to preset standards: Parameter validation used to prevent instruction errors; Communication link verification used to prevent data distortion; Hardware operation detection used to prevent abnormal output from devices.

3. The current curve control method based on a battery-detected constant voltage power supply according to claim 2, characterized in that, Entering the constant pressure control stage refers to: After the detection process begins, the controller calculates an initial target output current value based on the difference between the preset target voltage value and the current battery terminal voltage. This target output current value is converted into a continuously changing analog control voltage signal by a digital-to-analog converter and sent to the power stage control unit in the power supply system. After receiving the analog control voltage signal, the power stage control unit converts the control voltage into a duty cycle adjustment of the pulse width modulation signal based on a preset linear mapping relationship, and controls the on-time of the power switching device to regulate the rise of the output voltage. When the output voltage reaches the set target voltage value for the first time and enters its allowable deviation range, the system determines that the constant voltage condition has been met and enters the constant voltage control stage.

4. The current curve control method based on a battery-detected constant voltage power supply according to claim 3, characterized in that, During the continuous decrease of current, the output voltage change trend within each sampling period is dynamically monitored, and a voltage response coefficient is constructed based on the collected historical voltage sequence data to characterize the stability of the current voltage state. The method for constructing this response coefficient includes: First, obtain the difference between the current cycle voltage value and the previous cycle voltage value, and then assign a corresponding fluctuation level score based on the proportion of this difference within the preset voltage tolerance range. Next, the mean and variance of the voltage fluctuation level scores are statistically analyzed over multiple adjacent sampling periods to characterize the consistency and severity of voltage changes. The voltage fluctuation severity factor is formed by multiplying the mean of the fluctuation level scores by the fluctuation variance value. A sliding rate measurement based on a historical voltage sampling window is introduced to characterize the slow trend of voltage change; The voltage fluctuation severity factor is divided by the measured sliding rate, and a minimum limit is set to avoid division by zero, to calculate the voltage response coefficient for the current cycle.

5. The current curve control method based on a battery-detected constant voltage power supply according to claim 4, characterized in that, As the current reference value continues to decrease, a current smoothing coefficient is constructed for each sampling period to measure the continuity and fluctuation characteristics of current changes. This coefficient is used to adjust the magnitude of the current decrease in the current period. The calculation method for this smoothing coefficient includes: The actual current sampling values ​​of multiple consecutive cycles are extracted within the sliding time window to construct the original current sequence. By performing linear fitting or local trend extraction on the sequence, the trend curve of current changing with time is obtained. The original sampling values ​​and the corresponding trend values ​​are then differentially calculated to form a residual sequence representing the part of the current that deviates from the trend. A peak distribution map is constructed based on the residual sequence, and the number and amplitude of peak positions are identified and scored. The entropy value of the residual sequence is then calculated as a measure of the complexity of current change. The peak scores and entropy values ​​are weighted and fused to form a fluctuation evaluation factor. The fluctuation evaluation factor is normalized with the average current decrease within the sampling period to form the final current smoothing coefficient.

6. The current curve control method based on a battery-detected constant voltage power supply according to claim 5, characterized in that, When calculating the entropy value of the residual sequence, the entropy value is obtained by calculating the information entropy of the probability distribution of the residual sequence, and is used to characterize the overall fluctuation complexity of the residual sequence in the current period.

7. The current curve control method based on a battery-detected constant voltage power supply according to claim 5, characterized in that, Constructing a peak distribution map refers to: In the residual sequence, the difference between each sampling period and the previous period is calculated as the instantaneous fluctuation amplitude of the current. All fluctuation amplitudes in the residual sequence are grouped and statistically analyzed according to preset amplitude intervals. A fluctuation frequency histogram model representing the fluctuation intensity distribution is constructed to determine the frequency of occurrence of outliers with large fluctuation amplitudes in the overall sequence. The number and amplitude intervals of peak positions are identified and subjected to concentration analysis. Scoring is performed based on preset multi-level scoring rules to quantify the spike density and severity of current changes in the current cycle.

8. The current curve control method based on a battery-detected constant voltage power supply according to claim 7, characterized in that, Continuous adjustment of the rate of change of the current reference value refers to: After the sampling of the current control cycle ends, the current drop adjustment factor obtained by reasoning in the current cycle is passed to the reference control logic of the next cycle, and dynamically superimposed with the original preset current drop control curve to generate a current regulation factor with time correction significance. The rate of change of the target current reference value in the next cycle is continuously corrected exponentially using the current regulation factor. Specifically, this includes multiplying the original current descent gradient with a nonlinear convergence function constructed with the current regulation factor as the core parameter, and calculating the reference change correction magnitude for current control in the next cycle. In the subsequent real-time sampling and feedback process, the correction magnitude of the reference change is applied to the power stage control unit, so that the actual output current exhibits a feedback-driven dynamic correction change over multiple control cycles.

Images