Environment adaptive SOC estimation method and system

Abstract

The invention discloses an environment adaptive SOC estimation method and system, and the method comprises the steps: collecting terminal voltage, load control quantity and charge and discharge states, mapping the terminal voltage, the load control quantity and the charge and discharge states into energy agents, and continuously accumulating the energy agents; taking display stepping or full charge / undervoltage as an event anchor point, and extracting adjacent interval proxy increments to establish correspondence with SOC changes; the SOC calculation parameters are fused and updated to fit the current environment; the SOC is converted and output in real time according to the updated parameters during operation; respectively maintaining charging / discharging parameters and aligning the charging / discharging parameters at end points; and when the voltage is suddenly reduced, the SOC jumps or the tail end is overtime, weight lifting, sample elimination or rapid reestimation are executed, error updating is inhibited, and convergence is accelerated. According to the technical scheme, accurate, stable and fast convergence SOC estimation of the current environment is achieved under the conditions that current is not used as input and a historical model is not depended on, and meanwhile dependence on a high-precision ADC and large-capacity storage is reduced.

Description

TECHNICAL FIELD

[0002] The present application relates to the technical field of SOC estimation, in particular to an environment adaptive SOC estimation method and system. BACKGROUND

[0003] The existing terminal product usually takes coulomb counting as the core of the estimation of the state of charge (SOC) of the battery, supplemented by the open circuit voltage-temperature / aging model and the cycle history parameters. However, under the working conditions such as low temperature, sudden power surge or switching of different working gears, the terminal voltage will rapidly drop or rebound, which is superimposed on the discrete step strategy of the display layer, and often causes problems such as "power-off / rebound", abnormal end time, and difficulty in aligning the full charge / under-voltage endpoints. For example, after the low-voltage threshold is triggered when there is a load, the jump flag is easily set (typically, the power supply voltage is lower than the predetermined lower limit and still has PWM output), which shows that the SOC suddenly drops. In some scenarios, the "long end time" exception is triggered when there is no effective step for a long time. When approaching the full charge threshold, the display end often needs additional flags or alignment actions to stabilize at 100% in the charging path. These phenomena introduce error accumulation and uncertainty from user experience to energy estimation. At the same time, in order to maintain the accuracy across environments, the traditional scheme also needs high-precision ADC and large historical parameter storage, which increases the BOM and resource occupation.

[0004] Therefore, there is an urgent need for a SOC estimation framework that can only rely on a period of charging / discharging behavior under the current environment to complete adaptive correction without using current as an estimation input and without relying on historical models such as capacity decay / temperature, and that can maintain endpoint consistency in both charging and discharging directions and has robustness to abnormal fluctuations, so as to suppress the jump, shorten the convergence time and reduce the device and storage costs in harsh scenarios such as low temperature and high power. SUMMARY

[0005] The purpose of the embodiments of the present application is to provide an environment adaptive SOC estimation method and system to solve the technical problems of long convergence time of the SOC estimation framework and high device and storage costs.

[0006] In order to solve the above technical problems, the embodiments of the present application provide an environment adaptive SOC estimation method, which adopts the following technical solutions: An environment adaptive SOC estimation method, comprising the following steps: Collecting one or more of the battery terminal voltage, the load control quantity and the charging / discharging state, converting the collected quantities into proxy indicators for representing energy balance through a preset mapping, and continuously accumulating in the running process; The step-up event is set by the discretization strategy based on the display layer, which is triggered by a discrete change of a predetermined amplitude of the displayed SOC or by a full charge / under-voltage boundary condition, and at each step-up event, the cumulative change of the proxy indicator since the last event is obtained and is correlated with the SOC change in the interval; According to the correlation, the SOC calculation parameters for describing the current environment are obtained, and the existing parameters are updated according to the preset fusion rule to form parameters matched with the current temperature, load and battery state; In subsequent operation, the real-time cumulative amount of the proxy indicator is converted into SOC according to the updated SOC calculation parameters, and the display output is driven to realize adaptive correction and continuous convergence of SOC under the current environment; Independent SOC calculation parameters are maintained for discharge and charge processes respectively, and boundary alignment is performed when full charge or under-voltage is detected to ensure endpoint consistency; When an abnormal situation including voltage drop, SOC jump or end continuous timeout is detected, at least one of the following is performed: increasing the parameter update weight, excluding abnormal samples or triggering fast re-estimation, to suppress false updates and accelerate convergence.

[0007] In a possible implementation, the step of collecting one or more of the battery terminal voltage, the load control amount and the charge-discharge state, and converting the collected amount into a proxy indicator for representing energy consumption through a preset mapping, and continuously accumulating during operation, specifically includes: The battery terminal voltage, load control amount and charge-discharge state are read at a fixed sampling period, and the current working gear is read synchronously; The collected amount is standardized according to the hardware configuration or working gear to establish a single / double-sided or different power level coefficient branch; The accumulation direction and single-period accumulation step of the proxy indicator are set according to the charge or discharge state; The step is added to the internal accumulation value, and an internal accumulation event is output once the accumulation value reaches a preset threshold for subsequent use.

[0008] In a possible implementation, the step of setting a step-up event by the discretization strategy based on the display layer, which is triggered by a discrete change of a predetermined amplitude of the displayed SOC or by a full charge / under-voltage boundary condition, and at each step-up event, the cumulative change of the proxy indicator since the last event is obtained and is correlated with the SOC change in the interval, the identification and confirmation of the step-up event specifically includes: The discrete change of SOC is pre-judged by configuring programmable step and time latch conditions in the display layer; When the discrete changes of SOC continuously satisfy the step size and continuously reach the time latching condition, it is confirmed as a valid step event; The cross-event interval is defined by two consecutive valid step events, and the correspondence between the internal cumulative value change and the SOC change is established only by the internal cumulative value change within the interval; If no valid step event occurs within the preset maximum waiting time, it is determined as a step timeout and the exception handling process is initiated.

[0009] In one possible implementation, the step of maintaining independent SOC calculation parameters for the discharge and charging processes respectively, and performing boundary alignment upon detecting full charge or undervoltage to ensure endpoint consistency, specifically includes: In the discharge state, the discharge parameters are updated only using the data of the discharge direction; in the charging state, the charging parameters are updated only using the data of the charging direction. When a full charge condition or the upper boundary is detected, the display layer SOC will be aligned to the upper boundary, and the charging parameters will be anchored at the current boundary point. When an undervoltage condition is detected and there is a load output, the display layer SOC will be aligned to the lower boundary, and the discharge parameters will be anchored at the current boundary point. After aligning any boundary, set an endpoint consistency constraint range for the parameters in the other direction to ensure that the two directions are aligned at the endpoints.

[0010] In one possible implementation, after the step of setting an endpoint consistency constraint interval for the parameters in the other direction to ensure alignment of the two directions at the endpoints after completing alignment of any boundary, the method further includes: After completing the upper or lower boundary alignment, the SOC calculation parameters for the corresponding direction and the boundary SOC are written to non-volatile storage and the version identifier is recorded. When the device is powered on or reset, it first loads the most recently valid stored parameters and boundary points as initialization values ​​and enters fast convergence mode; In fast convergence mode, it will automatically switch to regular update mode after accumulating no less than two valid step events. When a storage version is detected to be inconsistent with the current hardware configuration, parameter downgrade initialization is performed and boundary anchors are re-established.

[0011] In one possible implementation, when an abnormal situation including a voltage drop, a SOC jump, or a persistent timeout is detected, at least one of the following steps is performed to suppress erroneous updates and accelerate convergence: increasing the parameter update weight, removing abnormal samples, or triggering a fast reassessment. Specifically, this includes: If the terminal voltage is lower than the threshold and the display layer SOC experiences an abnormal drop in level within a short period of time, the jump flag is set, provided there is a load output. If no valid step event occurs within the preset maximum waiting time, the end timeout flag is set. If the display layer SOC bounces back or changes rapidly in both directions within a short period of time, the abnormal display flag will be set. For any anomaly flag set, at least one of the following actions should be performed: temporarily increase the parameter update weight, discard the current data sample that crosses the event interval, start the fast reassessment branch and reset the relevant counters.

[0012] In one possible implementation, for any anomaly flag being set, at least one of the following steps is performed: temporarily increasing the parameter update weight, discarding the current data sample across the event interval, initiating a fast reassessment branch, and resetting the relevant counters, specifically including: Select the weight level based on the type and severity of the anomaly indicator, using higher weights during rapid reassessment and lower weights during normal periods; At the end of each cross-event interval, a consistency judgment is performed on the samples within the interval, and abnormal samples that deviate from the interval midpoint by more than a threshold are removed. When two or more anomaly flags appear consecutively, the freeze parameters are updated until a new boundary alignment is completed or the next valid step event is obtained before unfreezing. When several valid step events without anomalies occur consecutively, the update weight is gradually reduced until it returns to the normal level.

[0013] To address the aforementioned technical problems, this application also provides an environment-adaptive SOC estimation system, which employs the following technical solution: An environment-adaptive SOC estimation system includes: The data acquisition module is used to collect one or more of the following: battery terminal voltage, load control quantity, and charging / discharging status. The collected quantities are converted into proxy indicators for characterizing energy balance through a preset mapping and are continuously accumulated during operation. The acquisition module is used to set step events based on the discretization strategy of the display layer. The step events are triggered by a predetermined amplitude discretization change in the displayed SOC or by a full charge / undercharge boundary condition. When each step event occurs, the module acquires the cumulative change of the proxy index since the previous event and establishes a corresponding relationship with the SOC change in that interval. The update module is used to obtain SOC calculation parameters for characterizing the current environment according to the correspondence, and update them with existing parameters according to a preset fusion rule to form parameters that match the current temperature, load and cell status. The conversion module is used to convert the real-time cumulative amount of the proxy indicator into SOC based on the updated SOC calculation parameters in subsequent operation, and drive the display output so that SOC can achieve adaptive correction and continuous convergence in the current environment. The alignment module is used to maintain independent SOC calculation parameters for the discharge and charging processes respectively, and to perform boundary alignment when full charge or undervoltage is detected to ensure endpoint consistency. The detection module is used to perform at least one of the following actions when abnormal conditions are detected, including voltage drop, SOC jump, or continuous timeout at the end, in order to suppress erroneous updates and accelerate convergence: increase parameter update weight, remove abnormal samples, or trigger fast reassessment.

[0014] Compared with the prior art, the embodiments of this application have the following main advantages: The environmentally adaptive SOC estimation method disclosed in this application maps terminal voltage and load control quantities to accumulative energy proxies and uses step events as learning anchors for adaptive parameter updates. It maintains parameters separately for the charging and discharging process and performs boundary alignment at full charge / undervoltage points. Furthermore, it controls the update intensity and sample quality through anomaly identification and governance (suppressing jumps and tails, and maintaining convergence robustness). Thus, it achieves accurate, stable, and fast convergence SOC estimation for the current environment without using current as input or relying on historical models, while reducing dependence on high-precision ADCs and large-capacity storage (reducing cost and resource consumption). Attached Figure Description

[0015] To more clearly illustrate the solutions in this application, the accompanying drawings used in the description of the embodiments of this application will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a flowchart of an embodiment of the environment-adaptive SOC estimation method according to this application; Figure 2 This is a schematic diagram of an embodiment of the environment-adaptive SOC estimation system according to this application. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0018] refer to Figure 1A flowchart of an embodiment of the environment-adaptive SOC estimation method according to this application is shown. The environment-adaptive SOC estimation method includes the following steps: Step S101: Collect one or more of the following: battery terminal voltage, load control quantity, and charging / discharging state. Convert the collected quantities into proxy indicators for characterizing energy balance through a preset mapping and continuously accumulate them during operation.

[0019] In this embodiment, the system first synchronously collects battery terminal voltage, load control quantities (e.g., duty cycle, power level, or operating mode), and charging / discharging status, and performs debouncing and necessary filtering on the input side to suppress transient glitches. Subsequently, based on hardware calibration parameters and mode configuration, the collected quantities from different sources are normalized to a unified scale and mapped into a proxy indicator representing energy balance. This indicator selects the opposite accumulation polarity according to the charging / discharging direction in each processing cycle, and is superimposed on the internal accumulation amount at a predetermined step size. Upper and lower limits and overflow protection are set to ensure that the accumulation process remains numerically stable and traceable during long-term operation.

[0020] Step S102: Set step events based on the discretization strategy of the display layer. The step events are triggered by a predetermined amplitude discretization change in the displayed SOC or by full charge / undercharge boundary conditions. When each step event occurs, obtain the cumulative change of the proxy index since the previous event and establish a correspondence with the SOC change in that interval.

[0021] In this embodiment, the system performs de-jitter determination by applying time latching to a predetermined amplitude change in the display SOC. A step event is confirmed only when the discrete changes continuously satisfy both amplitude and duration conditions. Furthermore, step events are also generated when boundary conditions such as full charge or undervoltage are met. Each time a step event occurs, the system records the event timestamp and the current internal cumulative value, and pairs it with the record of the previous event to form a cross-event interval. This is done to strictly limit parameter learning to two observable and reproducible display nodes, avoiding being misled by short-term load disturbances or display layer anti-jitter mechanisms.

[0022] Step S103: Obtain the SOC calculation parameters for characterizing the current environment according to the correspondence, and update them with the existing parameters according to the preset fusion rules to form parameters that match the current temperature, load and cell status.

[0023] In this embodiment, after obtaining the cross-event interval, the system reads the cumulative changes of the proxy indicators within the interval and establishes a one-to-one correspondence with the displayed changes in the SOC within the interval. To improve reliability, the data within the interval undergoes a consistency check: when outliers caused by load surges, sensor anomalies, or display layer rebound are detected, these segments are not included in the statistics; if necessary, the data is also grouped according to operating level or temperature label, so that subsequent parameter updates can reflect the characteristics of different operating conditions. Through this "event-anchored" approach, the system robustly correlates measurable cumulative changes with visible SOC transitions, providing clean and verifiable samples for subsequent parameter updates.

[0024] In step S104, during subsequent operation, the real-time cumulative amount of the proxy indicator is converted into SOC based on the updated SOC calculation parameters, and the display output is driven so that SOC can achieve adaptive correction and continuous convergence in the current environment.

[0025] In this embodiment, the system generates SOC calculation parameters to characterize the current environment and performs a controlled fusion update with existing parameters. The fusion process employs an adaptive strategy with memory: when the cross-event sample quality is high and the environment is stable, the update magnitude remains moderate to maintain display stability; when signs such as low temperature and high load, end-of-line tailing, or abrupt changes are detected, the update impact is temporarily increased to allow the parameters to adapt to the new environment more quickly. To ensure rollback and diagnostics, the system retains the most recent parameter versions and their generated metadata (operating condition labels, timestamps, sample size), and can roll back to a more robust historical parameter set when necessary, thereby avoiding the long-term impact of a single incidental anomaly on global convergence.

[0026] Step S105: Maintain independent SOC calculation parameters for the discharge process and the charging process respectively, and perform boundary alignment when full charge or undervoltage is detected to ensure endpoint consistency.

[0027] In this embodiment, after the parameter update is completed, the runtime SOC conversion no longer relies on historical cycle information or high-precision current measurement. Instead, it directly reads the currently accumulated proxy index, combines it with the latest parameters for conversion, and drives the display output. The internal calculation maintains a continuous SOC state variable to meet the needs of refined control and diagnosis. The display layer adopts an discrete step-by-step presentation method and uses hysteresis and clamping to suppress flicker when approaching the upper and lower boundaries. When there is a large deviation between the internal SOC and the displayed value, the display side completes the alignment through a gradual transition rather than an instantaneous jump to avoid visible back-and-forth jumps for the user, while ensuring that the continuity of the internal estimation is not interrupted by the display strategy.

[0028] Step S106: When an abnormal situation is detected, including voltage drop, SOC jump or continuous timeout at the end, at least one of the following is performed: increase parameter update weight, remove abnormal samples or trigger fast re-evaluation, in order to suppress false updates and accelerate convergence.

[0029] In this embodiment, considering the significant differences in electrochemical and thermal characteristics between the charging and discharging processes, the system maintains independent SOC calculation parameters for each process and sets up a strongly constrained boundary alignment mechanism at the endpoints: when full charge is detected, the display layer aligns to the upper boundary, and the reference point is fixed as the anchor point for the charging direction; when undervoltage is detected and there is load output, the display layer aligns to the lower boundary, and the reference point is fixed to the discharging direction. After endpoint alignment, the system applies consistency constraints to the parameters in both directions at the endpoints, ensuring seamless SOC transition at the boundaries regardless of whether switching from charging to discharging or vice versa. To shorten the initial convergence time after environmental changes or power outages and restarts, these anchor points and corresponding parameters are saved in a non-volatile manner and restored preferentially during power-on initialization.

[0030] It should be noted that anomaly detection and mitigation are integrated throughout the entire process. The system continuously monitors whether the terminal voltage drops rapidly when there is load output, whether the displayed SOC experiences unexpected jumps within a short period, and whether a valid step event has not been formed for an extended period. Once any of the above anomalies is confirmed, the system immediately enters the mitigation path, taking at least one of three measures: First, temporarily increasing the impact of parameter updates to allow the model to quickly absorb new environmental information and accelerate regression; second, discarding current learning samples that cross event intervals to prevent abnormal fragments from contaminating the parameters; third, triggering a rapid re-estimation branch, resetting relevant counters, and restarting convergence with the most recent boundary anchor point as a reference. After the anomaly subsides, the system automatically returns to the regular update channel and continues to rely on the step event-driven iterative approach to achieve stable convergence in the current environment.

[0031] This application achieves accurate, stable, and fast convergence of SOC estimation for the current environment without using current as input or relying on historical models. It also reduces the dependence on high-precision ADCs and large-capacity storage (reducing cost and resource consumption). By mapping terminal voltage and load control quantities to an accumulative energy proxy and using step events as learning anchors for adaptive parameter updates, maintaining parameters separately for the charging and discharging process and performing boundary alignment at full charge / undervoltage points, and controlling update intensity and sample quality through anomaly identification and governance (suppressing jumps and tailing, and maintaining convergence robustness), this application achieves accurate, stable, and fast convergence of SOC estimation for the current environment without using current as input or relying on historical models.

[0032] In some optional implementations of this embodiment, the steps of collecting one or more of the following—battery terminal voltage, load control quantity, and charge / discharge state—converting the collected quantities into proxy indicators for characterizing energy balance through a preset mapping, and continuously accumulating them during operation, specifically include: The battery terminal voltage, load control quantity, and charging / discharging status are read at a fixed sampling period, and the current operating level is read simultaneously. The collected data is standardized according to the hardware configuration or working mode to establish coefficient branches for single / double-sided or different power levels. Set the cumulative direction and single-cycle cumulative step size of the agent indicator according to the charging or discharging state; The step size is added to the internal cumulative value, and an internal cumulative event is output when the cumulative value reaches a preset threshold for subsequent use.

[0033] In this embodiment, the system synchronously reads the terminal voltage, load control quantity, and charging / discharging status at a fixed sampling period, while simultaneously pulling the current operating level to ensure data comparability at the same time. Then, according to the hardware configuration and level, the system selects the corresponding coefficient branch in the software to standardize inputs from different sources and with different dimensions, ensuring that subsequent accumulation only processes values ​​under a unified scale. Next, the positive / negative polarity of the accumulation and the single-cycle step size are determined based on the charging / discharging direction, allowing the same "surrogate indicator" to be naturally distinguished in the two energy directions. Finally, the step size is superimposed on the controlled internal accumulation quantity, and a programmable threshold is set. When the accumulation quantity reaches the threshold, an internal accumulation event is output, preparing clean sample boundaries for subsequent "step event" matching and parameter learning. The direct benefits of this implementation are noise resistance, ease of parameter tuning, and auditability: the fixed period avoids drift caused by asynchronous sampling; the standardized coefficients converge hardware differences into the software table for easy mass production; and the threshold event creates a "clearly granular" step in the accumulation process, facilitating alignment with the discrete step on the display side and reducing jitter-induced false triggers.

[0034] In some optional implementations of this embodiment, the step of setting step events based on the discretization strategy of the display layer, wherein the step event is triggered by a predetermined amplitude discretization change in the displayed SOC or by a full charge / undercharge boundary condition, and in the step of obtaining the cumulative change of the proxy indicator since the previous event and establishing a correspondence with the SOC change in that interval at each step event, the identification and confirmation of the step event specifically includes: Configure programmable step size and time latch conditions in the display layer to predict discrete changes in the SOC; When the discrete changes of SOC continuously satisfy the step size and continuously reach the time latching condition, it is confirmed as a valid step event; The cross-event interval is defined by two consecutive valid step events, and the correspondence between the internal cumulative value change and the SOC change is established only by the internal cumulative value change within the interval; If no valid step event occurs within the preset maximum waiting time, it is determined as a step timeout and the exception handling process is initiated.

[0035] In this embodiment, the display layer first sets a programmable step size and time latch. The system predicts discrete changes in SOC and applies time consistency constraints. Only when both the amplitude and duration are satisfied is it confirmed as a valid step, thereby generating a timestamp and event index. Subsequently, a cross-event interval is defined using two adjacent valid step events, and the correspondence between the internal cumulative changes within this interval and the SOC changes is established only, ensuring that the parameter update data comes from clean samples "between two stable display nodes" rather than short-term disturbances. If there is no valid step within the preset maximum waiting time, the system determines that it is currently in the end-of-line tail or atypical working condition, marks it as a step timeout, and switches to anomaly management instead of continuing to update with unreliable samples. This identification and interval processing not only eliminates false steps caused by display jitter and short-term load fluctuations in engineering, but also statistically ensures that "a sample is a stable and reproducible running segment," making parameter learning traceable and resilient to outliers.

[0036] In some optional implementations of this embodiment, the steps described above, which involve maintaining independent SOC calculation parameters for the discharge and charging processes respectively, and performing boundary alignment upon detecting full charge or undervoltage to ensure endpoint consistency, specifically include: In the discharge state, the discharge parameters are updated only using the data of the discharge direction; in the charging state, the charging parameters are updated only using the data of the charging direction. When a full charge condition or the upper boundary is detected, the display layer SOC will be aligned to the upper boundary, and the charging parameters will be anchored at the current boundary point. When an undervoltage condition is detected and there is a load output, the display layer SOC will be aligned to the lower boundary, and the discharge parameters will be anchored at the current boundary point. After aligning any boundary, set an endpoint consistency constraint range for the parameters in the other direction to ensure that the two directions are aligned at the endpoints.

[0037] In this embodiment, discharge parameters are updated only by discharge data during the discharge state and charging parameters are updated only by charging data during the charging state, avoiding mutual interference between the two paths. Simultaneously, full charge and undervoltage are used as strong anchor points: when full charge or upper boundary conditions are detected, the display layer directly aligns to the upper limit and solidifies the current alignment point as the boundary anchor for the charging parameters; when undervoltage is detected under load output, the display layer aligns to the lower limit and solidifies that point as the boundary anchor for the discharge parameters. After anchoring any boundary, an endpoint consistency constraint range is applied to the parameters in the other direction, ensuring that the SOC is continuously connected at the endpoints without any "steps" regardless of whether switching from charging to discharging or vice versa. This refinement transforms "endpoint consistency" from a slogan into concrete action: a three-step closed loop of separate updates, boundary anchoring, and cross-directional constraints, which improves the experience of cross-directional switching and reduces long-term drift.

[0038] In some optional implementations of this embodiment, after completing the alignment of any boundary and setting a constraint interval for endpoint consistency of the parameters in the other direction to ensure alignment of the two directions at the endpoints, the method further includes: After completing the upper or lower boundary alignment, the SOC calculation parameters for the corresponding direction and the boundary SOC are written to non-volatile storage and the version identifier is recorded. When the device is powered on or reset, it first loads the most recently valid stored parameters and boundary points as initialization values ​​and enters fast convergence mode; In fast convergence mode, it will automatically switch to regular update mode after accumulating no less than two valid step events. When a storage version is detected to be inconsistent with the current hardware configuration, parameter downgrade initialization is performed and boundary anchors are re-established.

[0039] In this embodiment, the SOC calculation parameters and boundary SOC in the corresponding direction are further specified to be written to non-volatile storage and version information is recorded so that the system can quickly start with the latest and most reliable operating parameters after power failure or restart. When the device comes online, it prioritizes loading the most recently valid parameters and anchor points and enters a fast convergence mode, relying on high update sensitivity to quickly adapt to the current environment. After accumulating at least two valid step events, it automatically returns to the normal update mode to avoid short-term occasional events from having an excessive impact on long-term parameters. If a discrepancy between the storage version and the hardware configuration is detected (e.g., the battery cell has been replaced or the carrier board has a different number of faces), the system performs parameter downgrade initialization and re-establishes the boundary anchor points to prevent old parameters from being "carried" to the new hardware. Through this set of solidification-loading-fast convergence-return to normal process, the method achieves reliable cold start behavior and cross-batch consistency, significantly shortening the recovery time after environment switching.

[0040] In some optional implementations of this embodiment, when an abnormal situation including voltage drop, SOC jump, or continuous timeout is detected, at least one of the following steps is performed to suppress erroneous updates and accelerate convergence: increasing parameter update weight, removing abnormal samples, or triggering fast re-evaluation. Specifically, this includes: If the terminal voltage is lower than the threshold and the display layer SOC experiences an abnormal drop in level within a short period of time, the jump flag is set, provided there is a load output. If no valid step event occurs within the preset maximum waiting time, the end timeout flag is set. If the display layer SOC bounces back or changes rapidly in both directions within a short period of time, the abnormal display flag will be set. For any anomaly flag set, at least one of the following actions should be performed: temporarily increase the parameter update weight, discard the current data sample that crosses the event interval, start the fast reassessment branch and reset the relevant counters.

[0041] In this embodiment, if the terminal voltage rapidly drops below the threshold and the displayed SOC abnormally decreases in a short period of time when there is a load output, the jump flag is set, indicating that the current operating condition does not match the original parameters; if no effective step occurs for a long time, the end timeout flag is set, indicating that a slow change or end tailing has entered the stage; if the display rebounds or rapid changes in the forward and reverse directions occur in a short period of time, the display anomaly flag is set to intercept scenarios where the display layer strategy and the algorithm layer state are out of sync. For any flag set, the system executes at least one of three governance measures: increasing the parameter update weight to accelerate the absorption of new operating condition information, discarding the current learning samples across event intervals to prevent contamination, and initiating rapid reassessment and resetting the relevant counters to "start from the anchor point again". By realistically implementing the "discrimination-setting-governance" chain, anomalies are no longer handled manually based on experience, but enter a verifiable state machine, which protects the long-term stability of parameters and ensures rapid response in the event of sudden environmental changes.

[0042] In some optional implementations of this embodiment, for any exception flag being set, at least one of the following steps is performed: temporarily increasing the parameter update weight, discarding the current data sample across the event interval, initiating a fast reassessment branch, and resetting the relevant counters. Specifically, these steps include: Select the weight level based on the type and severity of the anomaly indicator, using higher weights during rapid reassessment and lower weights during normal periods; At the end of each cross-event interval, a consistency judgment is performed on the samples within the interval, and abnormal samples that deviate from the interval midpoint by more than a threshold are removed. When two or more anomaly flags appear consecutively, the freeze parameters are updated until a new boundary alignment is completed or the next valid step event is obtained before unfreezing. When several valid step events without anomalies occur consecutively, the update weight is gradually reduced until it returns to the normal level.

[0043] In this embodiment, the system selects weight levels based on the anomaly type and severity. Higher weights are used during rapid reassessment to improve adaptation speed, and weights are gradually reduced after returning to normal to restore robustness. At the end of each cross-event interval, a consistency check is performed on the samples, removing outliers with excessively large deviations from the relative interval representative value to avoid one-time extreme perturbations "distorting" the parameters. When two or more anomaly markers appear consecutively, the system temporarily freezes parameter updates until a new boundary alignment or the next effective step occurs, thus systematically preventing the positive feedback loop of "anomaly—update—more anomaly." When several consecutive effective step events without anomalies are obtained, the weights are gradually reduced back to normal levels, completing the transition from "rapid alignment" to "smooth tracking." This refinement provides the "brake and accelerator" for parameter learning, enabling the algorithm to quickly follow changes in operating conditions without excessive fluctuations under stable conditions, thereby achieving the engineering implementation of "adaptive correction and continuous convergence" in your system.

[0044] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by instructing related hardware through computer-readable instructions. These computer-readable instructions can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above methods. The aforementioned storage medium can be a non-volatile storage medium such as a magnetic disk, optical disk, or read-only memory (ROM), or random access memory (RAM).

[0045] It should be understood that although the steps in the flowcharts of the accompanying figures are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the accompanying figures may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0046] Further reference Figure 2 As a response to the above Figure 1 To implement the method shown, this application provides an embodiment of an environment-adaptive SOC estimation system, which is similar to... Figure 1 Corresponding to the method embodiments shown, the system can be specifically applied to various electronic devices.

[0047] like Figure 2As shown, the environment-adaptive SOC estimation system 200 described in this embodiment includes: an acquisition module 201, an identification module 202, a calculation module 203, a training module 204, and a processing module 205. Wherein: The data acquisition module 201 is used to acquire one or more of the following: battery terminal voltage, load control quantity, and charging / discharging status. The acquired quantities are converted into proxy indicators for characterizing energy balance through a preset mapping and are continuously accumulated during operation. The acquisition module 202 is used to set step events based on the discretization strategy of the display layer. The step events are triggered by a predetermined amplitude discretization change in the displayed SOC or by a full charge / undercharge boundary condition. When each step event occurs, the module acquires the cumulative change of the proxy index since the previous event and establishes a corresponding relationship with the SOC change in that interval. The update module 203 is used to obtain the SOC calculation parameters for characterizing the current environment according to the correspondence, and update them with the existing parameters according to the preset fusion rules to form parameters that match the current temperature, load and cell status. The conversion module 204 is used to convert the real-time cumulative amount of the proxy indicator into SOC based on the updated SOC calculation parameters in subsequent operation, and drive the display output so that SOC can achieve adaptive correction and continuous convergence in the current environment. Alignment module 205 is used to maintain independent SOC calculation parameters for the discharge process and the charging process respectively, and to perform boundary alignment when full charge or undervoltage is detected to ensure endpoint consistency. The detection module 206 is used to perform at least one of the following when an abnormal situation is detected, including voltage drop, SOC jump, or continuous timeout at the end, in order to suppress false updates and accelerate convergence.

[0048] The environmentally adaptive SOC estimation system provided in this embodiment of the invention can realize all the processes of the environmentally adaptive SOC estimation method in the above embodiments. The functions and technical effects of each module in the device are the same as those of the environmentally adaptive SOC estimation method in the above embodiments, and will not be repeated here.

[0049] The above are merely preferred embodiments of this application and are not intended to limit this application. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. An environment-adaptive SOC estimation method, characterized in that, Includes the following steps: Collect one or more of the following: battery terminal voltage, load control quantity, and charging / discharging status. Convert the collected quantities into proxy indicators for characterizing energy balance through a preset mapping and continuously accumulate them during operation. Step events are set based on the discretization strategy of the display layer. The step events are triggered by a predetermined amplitude discretization change in the displayed SOC or by a full charge / undercharge boundary condition. When each step event occurs, the cumulative change of the proxy index since the previous event is obtained and a corresponding relationship is established with the SOC change in that interval. Based on the correspondence, SOC calculation parameters for characterizing the current environment are obtained, and updated with existing parameters according to a preset fusion rule to form parameters that match the current temperature, load, and cell status. In subsequent operation, based on the updated SOC calculation parameters, the real-time cumulative amount of the proxy indicator is converted into SOC and the display output is driven, so that SOC can achieve adaptive correction and continuous convergence in the current environment. Independent SOC calculation parameters are maintained for the discharge and charging processes respectively, and boundary alignment is performed when full charge or undervoltage is detected to ensure endpoint consistency. When anomalies are detected, including voltage drops, SOC jumps, or persistent timeouts at the end, at least one of the following actions should be taken: increasing parameter update weights, removing abnormal samples, or triggering a fast reassessment, in order to suppress erroneous updates and accelerate convergence.

2. The environmentally adaptive SOC estimation method according to claim 1, characterized in that, The steps of collecting one or more of the following: battery terminal voltage, load control quantity, and charge / discharge state; converting the collected quantities into proxy indicators for characterizing energy balance through a preset mapping; and continuously accumulating these indicators during operation, specifically include: The battery terminal voltage, load control quantity, and charging / discharging status are read at a fixed sampling period, and the current operating level is read simultaneously. The collected data is standardized according to the hardware configuration or working mode to establish coefficient branches for single / double-sided or different power levels. Set the cumulative direction and single-cycle cumulative step size of the agent indicator according to the charging or discharging state; The step size is added to the internal cumulative value, and an internal cumulative event is output when the cumulative value reaches a preset threshold for subsequent use.

3. The environmentally adaptive SOC estimation method according to claim 1, characterized in that, The step of setting step events based on the discretization strategy of the display layer, wherein the step event is triggered by a predetermined amplitude discretization change in the displayed SOC or by a full charge / undercharge boundary condition, and in the step of obtaining the cumulative change of the proxy indicator since the previous event and establishing a correspondence with the SOC change in that interval at each step event, the identification and confirmation of the step event specifically includes: Configure programmable step size and time latch conditions in the display layer to predict discrete changes in the SOC; When the discrete changes of SOC continuously satisfy the step size and continuously reach the time latching condition, it is confirmed as a valid step event; The cross-event interval is defined by two consecutive valid step events, and the correspondence between the internal cumulative value change and the SOC change is established only by the internal cumulative value change within the interval; If no valid step event occurs within the preset maximum waiting time, it is determined as a step timeout and the exception handling process is initiated.

4. The environment-adaptive SOC estimation method according to claim 1, characterized in that, The steps of maintaining independent SOC calculation parameters for the discharge and charging processes respectively, and performing boundary alignment upon detecting full charge or undervoltage to ensure endpoint consistency, specifically include: In the discharge state, the discharge parameters are updated only using the data of the discharge direction; in the charging state, the charging parameters are updated only using the data of the charging direction. When a full charge condition or the upper boundary is detected, the display layer SOC will be aligned to the upper boundary, and the charging parameters will be anchored at the current boundary point. When an undervoltage condition is detected and there is a load output, the display layer SOC will be aligned to the lower boundary, and the discharge parameters will be anchored at the current boundary point. After aligning any boundary, set an endpoint consistency constraint range for the parameters in the other direction to ensure that the two directions are aligned at the endpoints.

5. The environmentally adaptive SOC estimation method according to claim 4, characterized in that, After completing the alignment of any boundary, setting a constraint interval for endpoint consistency for the parameters in the other direction to ensure alignment of the two directions at the endpoints, the method further includes: After completing the upper or lower boundary alignment, the SOC calculation parameters for the corresponding direction and the boundary SOC are written to non-volatile storage and the version identifier is recorded. When the device is powered on or reset, it first loads the most recently valid stored parameters and boundary points as initialization values ​​and enters fast convergence mode; In fast convergence mode, it will automatically switch to regular update mode after accumulating no less than two valid step events. When a storage version is detected to be inconsistent with the current hardware configuration, parameter downgrade initialization is performed and boundary anchors are re-established.

6. The environmentally adaptive SOC estimation method according to claim 1, characterized in that, When an abnormal situation, including voltage drop, SOC jump, or continuous timeout at the end, is detected, at least one of the following steps is performed: increasing the parameter update weight, removing abnormal samples, or triggering rapid reassessment, in order to suppress erroneous updates and accelerate convergence. Specifically, this includes: If the terminal voltage is lower than the threshold and the display layer SOC experiences an abnormal drop in level within a short period of time, the jump flag is set, provided there is a load output. If no valid step event occurs within the preset maximum waiting time, the end timeout flag is set. If the display layer SOC bounces back or changes rapidly in both directions within a short period of time, the abnormal display flag will be set. For any anomaly flag set, at least one of the following actions should be performed: temporarily increase the parameter update weight, discard the current data sample that crosses the event interval, start the fast reassessment branch and reset the relevant counters.

7. The environmentally adaptive SOC estimation method according to claim 6, characterized in that, For any anomaly flag being set, at least one of the following steps must be performed: temporarily increasing the parameter update weight, discarding the current data sample spanning the event interval, or initiating a fast reassessment branch and resetting the relevant counters. Specifically, these steps include: Select the weight level based on the type and severity of the anomaly indicator, using higher weights during rapid reassessment and lower weights during normal periods; At the end of each cross-event interval, a consistency judgment is performed on the samples within the interval, and abnormal samples that deviate from the interval midpoint by more than a threshold are removed. When two or more anomaly flags appear consecutively, the freeze parameters are updated until a new boundary alignment is completed or the next valid step event is obtained before unfreezing. When several valid step events without anomalies occur consecutively, the update weight is gradually reduced until it returns to the normal level.

8. An environment-adaptive SOC estimation system, characterized in that, include: The data acquisition module is used to collect one or more of the following: battery terminal voltage, load control quantity, and charging / discharging status. The collected quantities are converted into proxy indicators for characterizing energy balance through a preset mapping and are continuously accumulated during operation. The acquisition module is used to set step events based on the discretization strategy of the display layer. The step events are triggered by a predetermined amplitude discretization change in the displayed SOC or by a full charge / undercharge boundary condition. When each step event occurs, the module acquires the cumulative change of the proxy index since the previous event and establishes a corresponding relationship with the SOC change in that interval. The update module is used to obtain SOC calculation parameters for characterizing the current environment according to the correspondence, and update them with existing parameters according to a preset fusion rule to form parameters that match the current temperature, load and cell status. The conversion module is used to convert the real-time cumulative amount of the proxy indicator into SOC based on the updated SOC calculation parameters in subsequent operation, and drive the display output so that SOC can achieve adaptive correction and continuous convergence in the current environment. The alignment module is used to maintain independent SOC calculation parameters for the discharge and charging processes respectively, and to perform boundary alignment when full charge or undervoltage is detected to ensure endpoint consistency. The detection module is used to perform at least one of the following actions when abnormal conditions are detected, including voltage drop, SOC jump, or continuous timeout at the end, in order to suppress erroneous updates and accelerate convergence: increase parameter update weight, remove abnormal samples, or trigger fast reassessment.

Images