A method, equipment and medium for detecting the charge level of a new energy vehicle charging battery

Abstract

This invention discloses a method, device, and medium for detecting the charge level of a new energy vehicle rechargeable battery, relating to the field of battery detection technology. The method includes: real-time acquisition of battery operating data; inputting the individual cell voltage and temperature change rates from the battery operating data into a multi-scale differential processor; performing voltage feature decoupling through voltage differentiation and temperature compensation operations within a sliding window to generate a decoupled voltage feature curve; loading dynamic compensation coefficients into the electrochemical equation of state to solve for a real-time state of charge (SOC) estimate; transmitting the real-time SOC estimate to the vehicle energy management platform via the vehicle control bus, generating a transmission result, and updating the historical SOC feature curve based on the transmission result. This invention achieves precise frequency domain stripping of the battery aging and degradation layer structure through a synergistic mechanism of a geological stratified rock layering model and lightweight wavelet decomposition.

Description

Technical Field

[0001] This invention relates to the field of battery testing technology, and in particular to a method, equipment and medium for testing the charge capacity of rechargeable batteries for new energy vehicles. Background Technology

[0002] In recent years, the state-of-charge (SOC) estimation technology for new energy vehicle batteries has gradually evolved from traditional methods based on equivalent circuit models to multi-physics coupling analysis. With the application of electrochemical impedance spectroscopy and frequency domain decomposition techniques, the accuracy of extracting battery aging characteristics has improved. In particular, the introduction of wavelet decomposition and machine learning methods has enhanced the ability to separate short-term operating fluctuations from long-term aging degradation, providing a new technical path for high-precision SOC estimation.

[0003] Existing solutions still have shortcomings in handling the coupled interference of aging features across multiple time scales. Current methods rely on fixed-band wavelet decomposition or recurrent neural networks to classify aging features, but these methods struggle to effectively separate cross-scale aliasing noise caused by temperature cycling and material phase transitions. Particularly in the mid-stage of battery aging, the coupling effect between microscopic lithium deposition and macroscopic capacity decay can trigger frequency domain energy leakage, leading to distortion of aging weight parameters and consequently, accumulated errors in SOC estimation. This is particularly pronounced under dynamic operating conditions, becoming a bottleneck restricting the accuracy of battery management. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a method for detecting the charge level of a new energy vehicle charging battery to solve the problem of decreased accuracy in state of charge estimation caused by the aliasing of multi-scale features due to battery aging.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides a method for detecting the charge level of a rechargeable battery in a new energy vehicle, comprising,

[0008] Real-time acquisition of battery operating data;

[0009] The cell voltage and temperature change rate in the battery operating data are input into the multi-scale differential processor. Voltage feature decoupling is performed through voltage differentiation and temperature compensation calculations within a sliding window to generate the decoupled voltage feature curve.

[0010] A geological stratification model is applied to the historical state of charge characteristic curves in the battery operating data to generate an aging attenuation layer structure with time scales. A lightweight wavelet decomposition algorithm is used to strip the aging attenuation layer structure with frequency domains and output the aging layer weight parameters at different time scales.

[0011] The decoupled voltage characteristic curves and aging layer weight parameters at different time scales are integrated, and the dynamic compensation coefficient is calculated through a cross-entropy decision mechanism.

[0012] The dynamic compensation coefficient is loaded into the electrochemical equation of state to solve for the real-time state of charge estimate.

[0013] The vehicle control bus transmits real-time state of charge (SOC) estimates to the vehicle energy management platform, generates transmission results, and updates historical SOC characteristic curves based on the transmission results.

[0014] As a preferred embodiment of the new energy vehicle charging battery power detection method of the present invention, the battery operating data includes single cell voltage, temperature change rate and historical state of charge characteristic curve.

[0015] In a preferred embodiment of the new energy vehicle rechargeable battery power detection method of the present invention, the specific steps for generating the decoupled voltage characteristic curve are as follows:

[0016] Align the individual unit voltage and temperature change rates according to millisecond time sequences, and construct a spatiotemporal diagram within a preset time window;

[0017] The spatiotemporal graph is input into the random diffusion processor, and an optimized battery state vector is generated through the neighborhood node state propagation and noise suppression mechanism.

[0018] Differential geometric projection is performed on the optimized battery state vector to extract orthogonal decoupling features in the voltage-temperature interaction space;

[0019] The orthogonal decoupling characteristic quantities are filtered by frequency domain energy and reconstructed to generate the decoupled voltage characteristic curve.

[0020] In a preferred embodiment of the new energy vehicle rechargeable battery power detection method of the present invention, the specific steps for generating the time-scale layered aging degradation layer structure are as follows:

[0021] Simulate geological sedimentary stratification and decompose historical charge state characteristic curves into three time scales: short-term, medium-term, and long-term.

[0022] Aging rate features are extracted for each time scale layer, and the aging rate features are enhanced by associating them with the temperature change rate, thus generating enhanced aging rate features.

[0023] Based on the enhanced aging rate characteristics, aging attenuation layer thickness parameters for three time scales—short-term, medium-term, and long-term—are generated.

[0024] The thickness parameter of the aging degradation layer is mapped to three-dimensional space. By analyzing the stability of the decoupled voltage characteristic curve, the layering consistency is verified, and the aging degradation layer structure layered according to the time scale is output.

[0025] As a preferred embodiment of the new energy vehicle rechargeable battery power detection method of the present invention, the specific steps for outputting aging stratification weight parameters at different time scales are as follows:

[0026] Short-term, medium-term, and long-term time-series signals were extracted from the time-scale layered aging and decay structure.

[0027] Independent lightweight wavelet decomposition is performed on the short-term, medium-term, and long-term time-series signals respectively. After decomposition, the frequency domain energy components of each layer within the preset frequency band are extracted to generate the frequency domain energy components of each layer.

[0028] Based on the proportion of each frequency domain energy component in the total decomposed energy, the aging layer weight parameters for the short-term, medium-term, and long-term layers are calculated respectively, and the aging layer weight parameters for different time scales are output.

[0029] In a preferred embodiment of the new energy vehicle rechargeable battery power detection method of the present invention, the specific steps for calculating the dynamic compensation coefficient through the cross-entropy decision mechanism are as follows:

[0030] Time difference is performed on the decoupled voltage characteristic curve to generate a voltage time change rate signal;

[0031] Based on the decoupled voltage characteristic curve and the aging layer weight parameters at different time scales, a three-layer independent decision path is established.

[0032] The synergy of the three independent decision-making paths is quantified by the cross-entropy decision-making mechanism, and the dynamic compensation coefficient is calculated by combining the voltage time change rate signal.

[0033] As a preferred embodiment of the new energy vehicle rechargeable battery power detection method of the present invention, the specific steps for solving the real-time state of charge estimate are as follows:

[0034] By combining the dynamic compensation coefficient with the decoupled voltage characteristic curve, an electrochemical equation of state is generated.

[0035] Based on the battery operating current corresponding to the single cell voltage and the electrochemical state equation, a prediction calculation is performed to output a real-time estimated value of the state of charge.

[0036] The voltage deviation between the decoupled voltage characteristic curve and the real-time state of charge estimate is compared. If the deviation exceeds the preset voltage threshold, the electrochemical state equation is regenerated; otherwise, the real-time state of charge estimate is output.

[0037] In a preferred embodiment of the new energy vehicle rechargeable battery power detection method of the present invention, the specific steps of updating the historical state of charge characteristic curve based on the transmission results are as follows:

[0038] A hash digest is performed on the aging and decaying layer structure to generate a structural feature digest value. The real-time state of charge estimate is bound to the structural feature digest value and transmitted to the vehicle energy management platform through the vehicle control bus to generate a transmission result with a timestamp.

[0039] The vehicle energy management platform uses a digital twin model to verify the consistency between the real-time state of charge estimate and the structural feature summary value in the timestamped transmission results;

[0040] If the verification passes, the historical charge state characteristic curve is updated with time weighting based on the timestamp of the transmission result, and the updated historical charge state characteristic curve is generated. If the verification fails, the historical charge state characteristic curve is decomposed again.

[0041] In a second aspect, the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, wherein when the computer program is executed by the processor, it implements any step of the new energy vehicle charging battery power detection method described in the first aspect of the present invention.

[0042] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the new energy vehicle charging battery power detection method as described in the first aspect of the present invention.

[0043] The beneficial effects of this invention are as follows: Through the synergistic mechanism of a geological stratification model and lightweight wavelet decomposition, accurate frequency domain stripping of the battery aging degradation layer structure is achieved. The historical state-of-charge characteristic curve is decomposed into three independent time scales: short-term, medium-term, and long-term. The resulting aging layer weight parameters directly correspond to physical degradation indicators such as SEI film growth rate, lithium dendrite accumulation, and active lithium loss rate, providing a quantitative basis for lifetime prediction. Furthermore, a lightweight wavelet decomposition algorithm is used to perform directional frequency band energy extraction on the time-series signals of each layer, meeting the real-time requirements of automotive-grade MCUs. Attached Figure Description

[0044] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0045] Figure 1 A flowchart for a method to detect the charge level of rechargeable batteries in new energy vehicles.

[0046] Figure 2A flowchart for generating the decoupled voltage characteristic curve.

[0047] Figure 3 A flowchart for generating aging stratification weight parameters.

[0048] Figure 4 A flowchart for solving the real-time state of charge estimate. Detailed Implementation

[0049] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0050] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0051] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0052] Reference Figures 1-4 This is one embodiment of the present invention, which provides a method for detecting the charge level of a new energy vehicle charging battery, including the following steps:

[0053] S1. Real-time acquisition of battery operating data.

[0054] S1.1: Battery operating data includes cell voltage, temperature change rate, and historical state of charge characteristic curves.

[0055] Specifically, the voltage value of the battery cells in new energy vehicles is measured by a voltage sensor, and the voltage data is continuously recorded at a millisecond sampling rate. The voltage sensor is directly connected to the battery cell terminals to ensure that the voltage data is captured in real time.

[0056] The temperature of the battery in a new energy vehicle is measured by a temperature sensor. Temperature data is continuously recorded at a sampling rate of milliseconds. The ratio of the temperature difference between adjacent sampling points to the sampling time interval is used as the rate of temperature change. The temperature sensor is attached to the battery surface to capture temperature changes in real time.

[0057] The historical state of charge (SOC) characteristic curve data, which was previously generated and stored, is read from non-volatile memory. The historical SOC characteristic curve is time series data that records the SOC change characteristics over multiple charge and discharge cycles, ensuring data integrity.

[0058] S2. Input the cell voltage and temperature change rate from the battery operating data into the multi-scale differential processor. Perform voltage feature decoupling through voltage differentiation and temperature compensation operations within the sliding window to generate the decoupled voltage feature curve.

[0059] S2.1: Align the individual unit voltage and temperature change rates according to millisecond time sequence, and construct a spatiotemporal diagram within a preset time window;

[0060] Specifically, both the individual cell voltage and temperature change rate are recorded at a millisecond sampling rate to generate individual cell voltage sequences and temperature change rate sequences;

[0061] Linear interpolation was used to align the individual voltage sequence and temperature change rate sequence to a unified millisecond-level time axis;

[0062] Construct a spatiotemporal graph within a preset time window:

[0063] The spatiotemporal diagram is a two-dimensional matrix structure, with rows representing time points, the first column storing the unit voltage value, and the second column storing the temperature change rate value. The size of the spatiotemporal diagram is determined by a preset time window.

[0064] It should be noted that the preset time window refers to a fixed duration (e.g., 500 milliseconds) used to align the individual unit voltage and temperature change rate data.

[0065] S2.2: Input the spatiotemporal graph into the random diffusion processor, and generate an optimized battery state vector through the neighborhood node state propagation and noise suppression mechanism;

[0066] Specifically, in the spatiotemporal diagram, each point in time is defined as a node, and the node state includes the unit voltage value and temperature change rate value at that point in time.

[0067] Perform neighbor node state propagation for each node: obtain neighbor nodes within a preset time range before and after the current node, generate neighbor node state values, and adjust the current node state value according to the distribution trend of neighbor node state values;

[0068] Noise suppression: Detect abrupt changes in the current node's state value. When the abrupt change exceeds a preset fluctuation threshold, replace the abrupt change in the current node's state value with the trend value of the neighboring node's state value.

[0069] After state propagation and noise suppression, each node generates an optimized battery state vector.

[0070] It should be noted that the preset time range refers to the set fixed neighborhood time span (e.g., ±50 milliseconds), which is used to determine the time boundaries before and after the current node in order to select neighboring nodes;

[0071] The preset fluctuation threshold is an empirical value set based on the normal fluctuation range of the single cell voltage of new energy vehicle batteries under dynamic operating conditions. The range is 0.05-0.15 volts. 0.05 volts is used to filter out small disturbances to ensure the smoothness of the curve, and 0.15 volts is used to effectively capture real voltage changes and avoid misjudging normal charging and discharging fluctuations as noise.

[0072] S2.3: Perform differential geometric projection on the optimized battery state vector to extract orthogonal decoupling features in the voltage-temperature interaction space;

[0073] Specifically, an orthogonal coordinate system is established within the voltage-temperature interaction space, with the horizontal axis representing the individual voltage value and the vertical axis representing the rate of temperature change.

[0074] Each data point of the optimized battery state vector is mapped to the voltage-temperature interaction space to form a spatially distributed point set.

[0075] By performing coordinate transformation on the spatially distributed point set through the device's built-in orthogonal coordinate system framework, the original coordinate system is rotated to the main direction of data distribution, generating the voltage characteristic axis and temperature compensation axis in the new coordinate system.

[0076] The projection value of each data point on the voltage characteristic axis is read as the voltage dominant component, and the projection value on the temperature compensation axis is read as the temperature disturbance component. The output voltage dominant component is used as the orthogonal decoupling characteristic quantity.

[0077] S2.4: The orthogonal decoupling characteristic quantity is filtered by frequency domain energy and reconstructed to generate the decoupled voltage characteristic curve.

[0078] Specifically, the voltage-dominant component is input into the signal processor, and the built-in frequency band decomposition engine decomposes the voltage-dominant component into multiple frequency band components, each of which contains frequency domain signal energy.

[0079] The signal processor automatically identifies and retains the dominant frequency band components whose frequency domain signal energy is higher than a preset energy threshold, while discarding the noise frequency band components whose frequency domain signal energy is lower than a preset energy threshold.

[0080] The retained dominant frequency band component is input to the signal reconstructor, and the reconstructed time-domain voltage signal is synthesized through physical circuitry. The reconstructed time-domain voltage signal serves as the decoupled voltage characteristic curve.

[0081] It should be noted that the preset energy threshold is an empirical value set based on the difference in energy distribution between signal and noise in the frequency domain. The value range is 85%-95% of the total frequency band energy. 85% ensures that most of the effective voltage characteristic information is retained to avoid signal loss, and 95% is used to strictly filter out low-frequency drift and high-frequency random noise to achieve high-fidelity reconstruction.

[0082] A superior approach employs a multi-scale differential processor to achieve millisecond-level spatiotemporal graph construction and a neighborhood node state propagation mechanism using a random diffusion processor. This suppresses dynamic operating condition interference in real time within a sliding window, solving the response lag problem caused by the hundreds of milliseconds of latency in traditional recurrent neural networks. Differential geometric projection is used to extract orthogonal decoupling features in the voltage-temperature interaction space, eliminating temperature coupling interference and avoiding effective signal loss due to frequency band aliasing in fixed-band wavelet decomposition. Furthermore, frequency domain screening with a preset energy threshold selectively preserves aging-sensitive frequency bands, improving the signal-to-noise ratio of the feature curve and providing high fidelity for aging stratification.

[0083] S3. Apply a geological stratification model to the historical state of charge characteristic curves in the battery operating data to generate an aging attenuation layer structure layered by time scale. Use a lightweight wavelet decomposition algorithm to perform frequency domain stripping on the aging attenuation layer structure layered by time scale, and output the aging layer weight parameters at different time scales.

[0084] S3.1: Simulate geological sedimentary stratification and decompose the historical charge state characteristic curve into three time scales: short-term, medium-term, and long-term.

[0085] Specifically, time scale layers are divided according to preset time periods;

[0086] Short-term layer: Extracts all data points within the most recent preset time period;

[0087] Intermediate layer: Extracts data points that are one preset time period earlier than the short-term layer but later than the long-term layer;

[0088] Long-term layer: Extracts historical data points earlier than the intermediate layer.

[0089] Example: Preset time periods: short-term layer 1 hour, medium-term layer 24 hours, long-term layer 30 days.

[0090] S3.2: Extract aging rate features for each time scale layer, and enhance the aging rate features by associating them with the temperature change rate, thereby generating enhanced aging rate features.

[0091] Specifically, historical state of charge characteristic curve data for each time scale layer are extracted, and the aging rate simulation signal is directly generated using the device's built-in differential circuit.

[0092] The temperature change rate at the same time point as each time scale layer is read from the battery operating data. The temperature change rate is directly mapped into an analog signal with the corresponding voltage amplitude through the built-in digital-to-analog converter of the device, forming a temperature change rate signal at the same time point.

[0093] The aging rate analog signal and the temperature change rate signal at the same time stamp are input into the analog multiplier circuit, and the output product signal is used as an enhanced aging rate feature.

[0094] S3.3: Based on the enhanced aging rate characteristics, generate aging attenuation layer thickness parameters for short-term, medium-term, and long-term time scales;

[0095] Specifically, the short-term, medium-term, and long-term enhanced aging rate characteristic signals are extracted.

[0096] The short-term, medium-term, and long-term enhanced aging rate characteristic signals are input into independent physical integrator circuits, and the signal energy is accumulated through capacitor charging and discharging.

[0097] The output of the physical integrator circuit is connected to the root mean square converter to generate short-term, medium-term, and long-term DC voltage signals, which are then read by the analog-to-digital converter as aging attenuation layer thickness parameters for the three time scales: short-term, medium-term, and long-term.

[0098] S3.4: Map the aging attenuation layer thickness parameter to three-dimensional space, analyze the stability of the decoupled voltage characteristic curve, verify the layering consistency, and output the aging attenuation layer structure layered according to the time scale.

[0099] Specifically, the aging degradation layer thickness parameters for the short-term, medium-term, and long-term timescales are input into a three-dimensional coordinate system, and fixed location points are set:

[0100] Place the short-term aging and attenuation layer thickness parameter value at X-axis coordinate 1, the medium-term aging and attenuation layer thickness parameter value at X-axis coordinate 2, and the long-term aging and attenuation layer thickness parameter value at X-axis coordinate 3. Fix the Y-axis and Z-axis coordinates to 0 to generate a three-dimensional point set.

[0101] The decoupled voltage characteristic curve is input into the hardware fluctuation detector, and the effective voltage is output through the capacitor charging and discharging mechanism. The effective voltage is then compared with the preset stability threshold.

[0102] If the effective voltage does not exceed the preset stability threshold, the verification is successful. After successful verification, a three-dimensional point set is output as an aging and decay layer structure layered according to the time scale.

[0103] If the effective voltage exceeds the preset stability threshold, the historical state of charge characteristic curve is re-decomposed into three time scales: short-term, medium-term, and long-term.

[0104] It should be noted that the preset stability threshold is an empirical value set based on the allowable fluctuation range of the decoupling voltage characteristic curve of the battery during the verification stage. The value range is 0.03-0.08 volts. 0.03 volts is used to ensure that the stratification can respond sensitively to small steady-state changes in voltage, and 0.08 volts is used to tolerate fluctuations under normal operating conditions and avoid frequent failures of stratification verification due to brief disturbances.

[0105] S3.5: Extract short-term, medium-term, and long-term time series signals from the time-scale layered aging and decay layer structure;

[0106] Specifically, access the contiguous address space of pre-partitioned memory physical block 1 to obtain the complete data sequence of historical charge state characteristic curves within the most recent time period (e.g., 1 hour);

[0107] Access the contiguous address space of pre-partitioned memory physical block 2 to obtain the complete data sequence of historical charge state characteristic curves within a medium time period (e.g., 24 hours);

[0108] Access the contiguous address space of pre-partitioned memory physical block 3 to obtain the complete data sequence of historical charge state characteristic curves over a long period of time (e.g., 30 days).

[0109] Short-term layer timing signal = complete data sequence stored in physical block 1;

[0110] Mid-level timing signal = complete data sequence stored in physical block 2;

[0111] Long-term layer timing signal = complete data sequence stored in physical block 3.

[0112] S3.6: Perform independent lightweight wavelet decomposition on the short-term, medium-term, and long-term time-series signals respectively, and extract the frequency domain energy components of each layer within the preset frequency band after decomposition to generate the frequency domain energy components of each layer.

[0113] Specifically, by convolving the short-term time series signal with wavelet basis functions, the short-term time series signal is decomposed into multiple frequency components at different scales. The signal features of each frequency component are extracted to generate a short-term wavelet coefficient sequence. The short-term wavelet coefficient sequence corresponding to the preset frequency band is extracted. The sum of the squares of all coefficients in the short-term wavelet coefficient sequence corresponding to the preset frequency band is obtained by the standard energy integration method as the short-term frequency domain energy component.

[0114] After performing independent lightweight wavelet decomposition on the intermediate layer time series signal, an intermediate layer wavelet coefficient sequence is generated. The intermediate layer wavelet coefficient sequence corresponding to the preset frequency band is extracted. The sum of the squares of all coefficients in the intermediate layer wavelet coefficient sequence corresponding to the preset frequency band is obtained by the standard energy integration method as the intermediate layer frequency domain energy component.

[0115] After performing independent lightweight wavelet decomposition on the long-term time series signal, a long-term wavelet coefficient sequence is generated. The long-term wavelet coefficient sequence corresponding to the preset frequency band is extracted. The sum of the squares of all coefficients in the long-term wavelet coefficient sequence corresponding to the preset frequency band is obtained by the standard energy integration method as the long-term frequency domain energy component.

[0116] It should be noted that the preset frequency band is a low frequency band (0-5Hz). The basis for this setting is that the frequency band is determined by electrochemical impedance spectroscopy testing to characterize the battery's aging-sensitive features, and it is physically matched with the battery's relaxation time constant to meet automotive-grade real-time requirements.

[0117] S3.7: Specifically, based on the proportion of each layer's frequency domain energy component in the total decomposed energy, the aging layer weight parameters for the short-term, medium-term, and long-term layers are calculated respectively, and the aging layer weight parameters for different time scales are output. The expression is as follows:

[0118] ;

[0119] In the formula, This represents the stratification weight parameter for short-term aging. Represents the short-term layer frequency domain energy components. Indicates the short-term layer. Represents the total energy of the decomposition. This represents the stratification weight parameter for the mid-term aging layer. Represents the frequency domain energy components of the intermediate layer. Indicates the intermediate layer. This represents the weighting parameter for long-term aging layers. This represents the long-term layer frequency domain energy component. Indicates the long-term layer.

[0120] It should be noted that, , The dimensions of the sum are The dimension of E is Final output , and Since it is dimensionless, we maintain dimensional consistency.

[0121] The weighting parameter for short-term aging is set between 0.2 and 0.4. This value is chosen to accurately capture rapid dynamic aging characteristics such as SEI film growth. 0.2 is used to avoid excessive amplification of short-term noise, and 0.4 is used to ensure a sensitive response to initial capacity decay. The weighting parameter for medium-term aging is set between 0.4 and 0.6. This value is chosen to effectively characterize dominant decay mechanisms such as lithium dendrite accumulation. 0.4 is used to balance the impact of short-term disturbances, and 0.6 is used to strengthen the core contribution of medium-term decay to battery health. The weighting parameter for long-term aging is set between 0.1 and 0.3. This value is chosen to characterize slow, gradual aging processes such as active lithium loss. 0.1 is used to prevent long-term background noise interference, and 0.3 is used to ensure continuous tracking of the decay trend throughout the entire lifecycle.

[0122] Superiorly, compared to existing fixed-time-window Fourier transforms, this method transfers geological sedimentology to battery aging analysis. By dividing the physical timescales into short-term, medium-term, and long-term layers, it overcomes the limitation of traditional methods that cannot separate different aging mechanisms using fixed time windows. Through an aging rate feature enhancement mechanism correlated with temperature change rate, it utilizes the nonlinear coupling effect of temperature and aging to enhance the intensity of weak decay signals, improving the effective signal detection rate compared to conventional linear weighted fusion methods. Furthermore, by combining the frequency domain energy weights of the layered structure, it replaces the simple thresholding of the frequency domain principal components in existing technologies, increasing the correlation coefficient between the attenuation layer thickness parameter and the actual capacity decay of the battery, and overcoming the problem of long-term aging characteristics being overwhelmed by short-term noise.

[0123] S4. The decoupled voltage characteristic curves and aging layer weight parameters at different time scales are integrated, and the dynamic compensation coefficient is calculated through the cross-entropy decision mechanism.

[0124] S4.1: Perform time difference on the decoupled voltage characteristic curve to generate a voltage time change rate signal;

[0125] Specifically, the decoupled voltage characteristic curve is input into a hardware differential circuit group, which includes a series delay unit and a subtractor.

[0126] The delay unit applies a fixed time delay (e.g., 1 millisecond) to the decoupled voltage characteristic curve. The subtractor receives the current decoupled voltage characteristic curve signal value and the delayed decoupled voltage characteristic curve signal value, and outputs the voltage difference sequence between adjacent time points as the voltage time change rate signal.

[0127] S4.2: Based on the decoupled voltage characteristic curve and the aging layer weight parameters at different time scales, establish a three-layer independent decision path;

[0128] Specifically, the decoupled voltage characteristic curve is connected to input terminal 1 of the short-term layer analog multiplier circuit, and the short-term layer aging layer weight parameters are connected to input terminal 2 of the short-term layer analog multiplier circuit. A short-term decision path is formed from the input terminal to the output terminal of the short-term layer analog multiplier circuit.

[0129] The decoupled voltage characteristic curve is connected to input terminal 1 of the intermediate layer analog multiplier circuit, and the intermediate layer aging layer weight parameters are connected to input terminal 2 of the intermediate layer analog multiplier circuit. The intermediate decision path is formed from the input terminal to the output terminal of the intermediate layer analog multiplier circuit.

[0130] The decoupled voltage characteristic curve is connected to input terminal 1 of the long-term layer analog multiplier circuit, and the long-term layer aging layer weight parameters are connected to input terminal 2 of the long-term layer analog multiplier circuit. The long-term decision path is formed from the input terminal to the output terminal of the long-term layer analog multiplier circuit.

[0131] Short-run decision path = physical signal path of short-run layer analog multiplier circuit;

[0132] Intermediate decision path = physical signal path of the intermediate layer analog multiplier circuit;

[0133] Long-term decision path = physical signal path of long-term layer analog multiplier circuit;

[0134] This forms a three-tiered independent decision-making path.

[0135] S4.3: Specifically, the synergy of the three independent decision-making paths is quantified through a cross-entropy decision-making mechanism, and the dynamic compensation coefficient is calculated by combining the voltage time change rate signal. The expression is as follows:

[0136] ;

[0137] ;

[0138] ;

[0139] ;

[0140] In the formula, Represents the cross-entropy function. Indicates the current time The probability distribution of three independent decision paths at time , This represents the decision probability of the short-run layer at time t. This represents the decision probability of the intermediate layer at time t. This represents the decision probability of the long-term layer at time t. Indicates the current time The baseline distribution at that time This represents the baseline probability component of the short-term layer. Represents the baseline probability component of the intermediate layer. This represents the baseline probability component of the long-term layer. Indicates the current moment. Indicates the time-scale layer index. This indicates a summation of the short-term, medium-term, and long-term indexes. Indicates the first The baseline probability components of the layer, Represents the natural logarithm operation. Indicates the first The decision probability components of the layer, Indicates the dynamic compensation coefficient. Describes the differential operator. This represents the voltage-time rate of change signal. This indicates the operation of natural exponents.

[0141] It should be noted that the dynamic compensation coefficient is derived from the cross-entropy decision mechanism for the coupling calculation of the three-layer aging path synergy and voltage change rate. The value range is 0.5-2.0. 0.5 is used to implement conservative compensation to prevent SOC overestimation when there is path divergence or voltage drastic change. 2.0 is used to amplify the compensation effect to quickly correct the cumulative error of the electrochemical equation of state when the path is highly synergistic and the voltage is stable.

[0142] It should be noted that, , and Dimensionless, ultimately Dimensionless The dimension of s -1 ,final The dimension of s -1 .

[0143] S5. Load the dynamic compensation coefficient into the electrochemical equation of state and solve for the real-time state of charge estimate.

[0144] S5.1: Specifically, the dynamic compensation coefficient is combined with the decoupled voltage characteristic curve to generate the electrochemical state equation, the expression of which is:

[0145] ;

[0146] In the formula, Indicates the current time The equivalent voltage at that time, Indicates the current time The voltage characteristic curve after decoupling at that time. Indicates the dynamic compensation coefficient. Indicates the battery's internal resistance. Indicates the current time The battery operating current at that time.

[0147] It should be noted that, and The dimension of is V. The dimension of Ω is . The dimension is A. The dimension of the quantity is V, and the dimension is ultimately kept consistent.

[0148] S5.2: Specifically, based on the battery operating current corresponding to the single-cell voltage and the electrochemical state equation, a predictive calculation is performed to output a real-time estimated state of charge, expressed as:

[0149] ;

[0150] In the formula, Indicates the current time Real-time state of charge estimate at time, Indicates reference time The reference value of the state of charge at that time, Indicates the battery's rated capacity. This represents the reciprocal of the battery's rated capacity. Indicates to arrive Perform definite integrals, This indicates the difference in charge and discharge efficiency. express The battery operating current at any given time, Represents the integral time element. Indicates the current time The absolute value of the battery operating current at that time. Indicates the sampling time interval. This represents a time unit conversion constant, converting seconds to hours.

[0151] It should be noted that, The dimension of is %. The dimension of is %. The dimension is Ah. The dimension of is Ah -1 , Dimensionless The dimension is A. The dimension is s. The dimension is A. The dimension is s. The dimension of s -1 , Dimensionless Dimensionless, ultimately The dimension of is %, to maintain dimensional consistency.

[0152] S5.3: Compare the voltage deviation between the decoupled voltage characteristic curve and the real-time state of charge estimate. If the deviation exceeds the preset voltage threshold, regenerate the electrochemical state equation; otherwise, output the real-time state of charge estimate.

[0153] Specifically, each data point in the decoupled voltage characteristic curve is input to one input of the subtractor circuit, and the voltage value corresponding to the real-time state of charge estimate is input to the other input of the subtractor circuit. The absolute value of the difference between the decoupled voltage characteristic curve and the real-time state of charge estimate is output by the subtractor circuit as the voltage deviation value.

[0154] The voltage deviation value is compared with the preset voltage threshold. If the voltage deviation value is greater than the preset voltage threshold, the electrochemical state equation is regenerated. If the voltage deviation value is less than or equal to the preset voltage threshold, the real-time state of charge estimate is output.

[0155] It should be noted that the preset voltage threshold is an empirical value set based on the maximum allowable deviation range between the theoretical voltage of the electrochemical equation of state and the decoupled measured voltage. The value range is 0.1-0.3 volts. 0.1 volts is used to ensure that the SOC estimation performance is sensitive to correct small data mismatches, and 0.3 volts is used to tolerate sensor measurement noise and rapid fluctuations in operating conditions, avoiding unnecessary frequent reconstruction of the electrochemical equation of state.

[0156] S6. Transmit the real-time state of charge estimate to the vehicle energy management platform via the vehicle control bus, generate the transmission result, and update the historical state of charge characteristic curve based on the transmission result.

[0157] S6.1: Perform a hash digest on the aging and decaying layer structure to generate a structural feature digest value. Bind the real-time state of charge estimate with the structural feature digest value and transmit it to the vehicle energy management platform through the vehicle control bus to generate a transmission result with a timestamp.

[0158] Specifically, the three-dimensional point set of the aging decay layer structure layered by time scale is input into the hash processor, and the hash processor outputs a fixed-length hexadecimal string as the structural feature summary value.

[0159] The real-time state of charge estimate and structural feature summary value are encapsulated into a data frame and sent to the vehicle energy management platform through the vehicle control bus;

[0160] After receiving a data frame, the vehicle energy management platform appends the current UTC timestamp to generate a transmission result containing the timestamp.

[0161] S6.2: The vehicle energy management platform uses a digital twin model to verify the consistency between the real-time state of charge estimate and the structural feature summary value in the timestamped transmission results;

[0162] It should be noted that the digital twin model was trained by calling historical battery lifecycle data before deployment. The training process includes: collecting historical battery operating data (cell voltage, temperature change rate and state of charge value) to construct a spatiotemporal sequence, extracting the aging rate features of the spatiotemporal sequence, supervising the learning of the aging rate features and the actual aging degradation layer structure, and solidifying the relationship parameters between the aging rate features and the actual aging degradation layer structure into the vehicle energy management platform to obtain the digital twin model.

[0163] Specifically, the transmission results containing timestamps are input into the digital twin model built into the vehicle energy management platform. The digital twin model calls the historical battery state data corresponding to the timestamps to reconstruct the three-dimensional point set of the aging and degradation layer structure.

[0164] Perform a hash digest on the reconstructed aging attenuation layer structure's 3D point set to generate a digest value for verifying the structure's features;

[0165] The verification structure feature digest value is compared with the structure feature digest value in the transmission result containing the timestamp. The verification is successful if all bytes are completely identical; otherwise, the verification fails.

[0166] S6.3: If the verification passes, the historical charge state characteristic curve is updated by time weighting based on the timestamp of the transmission result, and the updated historical charge state characteristic curve is generated. If the verification fails, the historical charge state characteristic curve is decomposed again.

[0167] Specifically, when the digital twin model passes verification, the timestamp and real-time state of charge estimate are extracted from the timestamped transmission results;

[0168] Read the historical state of charge characteristic curve data segment corresponding to the timestamp from the memory, and update the historical state of charge characteristic curve data segment corresponding to the timestamp using the exponential weighting algorithm;

[0169] The real-time state of charge estimate is fused with the historical state of charge characteristic curve data segment according to the preset forgetting factor to generate the updated historical state of charge characteristic curve data segment and write it back to the memory.

[0170] If the verification fails, clear the current historical state of charge characteristic curve data and re-decompose the historical state of charge characteristic curve.

[0171] It should be noted that the preset forgetting factor is an empirical value set based on the balance between the stability of historical battery SOC data and the sensitivity of real-time updates. The value ranges from 0.7 to 0.95. 0.7 ensures that the digital twin model can quickly track sudden changes in battery status (such as rapid discharge), while 0.95 is used to smooth daily charge and discharge fluctuations to maintain the stability of long-term aging trends.

[0172] This embodiment also provides a computer device applicable to the method for detecting the charge level of a new energy vehicle's rechargeable battery, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the method for detecting the charge level of a new energy vehicle's rechargeable battery as proposed in the above embodiment.

[0173] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0174] This embodiment also provides a storage medium storing a computer program, which, when executed by a processor, implements the method for detecting the charge level of a new energy vehicle charging battery as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0175] In summary, this invention achieves precise frequency domain stripping of the battery aging degradation layer structure through a synergistic mechanism of a geological stratification rock layer model and lightweight wavelet decomposition. The historical state of charge characteristic curve is decomposed into three independent time scales: short-term, medium-term, and long-term. The resulting aging layer weight parameters directly correspond to physical degradation indicators such as SEI film growth rate, lithium dendrite accumulation, and active lithium loss rate, providing a quantitative basis for lifetime prediction. Furthermore, a lightweight wavelet decomposition algorithm is used to perform directional frequency band energy extraction on the time-series signals of each layer, meeting the real-time requirements of automotive-grade MCUs.

[0176] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for detecting the charge level of a new energy vehicle charging battery, characterized in that: include, Real-time acquisition of battery operating data; The cell voltage and temperature change rate in the battery operating data are input into the multi-scale differential processor. Voltage feature decoupling is performed through voltage differentiation and temperature compensation calculations within a sliding window to generate the decoupled voltage feature curve. A geological stratification model is applied to the historical state of charge characteristic curves in the battery operating data to generate an aging attenuation layer structure with time scales. A lightweight wavelet decomposition algorithm is used to strip the aging attenuation layer structure with frequency domains and output the aging layer weight parameters at different time scales. The decoupled voltage characteristic curves and aging layer weight parameters at different time scales are integrated, and the dynamic compensation coefficient is calculated through a cross-entropy decision mechanism. The dynamic compensation coefficient is loaded into the electrochemical equation of state to solve for the real-time state of charge estimate. The vehicle control bus transmits real-time state of charge (SOC) estimates to the vehicle energy management platform, generates transmission results, and updates historical SOC characteristic curves based on the transmission results.

2. The method for detecting the charge level of a new energy vehicle charging battery as described in claim 1, characterized in that: The battery operating data includes cell voltage, temperature change rate, and historical state of charge characteristic curves.

3. The method for detecting the charge level of a new energy vehicle charging battery as described in claim 2, characterized in that: The specific steps for generating the decoupled voltage characteristic curve are as follows. Align the individual unit voltage and temperature change rates according to millisecond time sequences, and construct a spatiotemporal diagram within a preset time window; The spatiotemporal graph is input into the random diffusion processor, and an optimized battery state vector is generated through the neighborhood node state propagation and noise suppression mechanism. Differential geometric projection is performed on the optimized battery state vector to extract orthogonal decoupling features in the voltage-temperature interaction space; The orthogonal decoupling characteristic quantities are filtered by frequency domain energy and reconstructed to generate the decoupled voltage characteristic curve.

4. The method for detecting the charge level of a new energy vehicle charging battery as described in claim 3, characterized in that: The specific steps for generating the time-scale layered aging degradation layer structure are as follows. Simulate geological sedimentary stratification and decompose historical charge state characteristic curves into three time scales: short-term, medium-term, and long-term. Aging rate features are extracted for each time scale layer, and the aging rate features are enhanced by associating them with the temperature change rate, thus generating enhanced aging rate features. Based on the enhanced aging rate characteristics, aging attenuation layer thickness parameters for short-term, medium-term, and long-term time scales are generated. The thickness parameter of the aging degradation layer is mapped to three-dimensional space. By analyzing the stability of the decoupled voltage characteristic curve, the layering consistency is verified, and the aging degradation layer structure layered according to the time scale is output.

5. The method for detecting the charge level of a new energy vehicle charging battery as described in claim 4, characterized in that: The specific steps for outputting aging stratification weight parameters at different time scales are as follows. Short-term, medium-term, and long-term time-series signals were extracted from the time-scale layered aging and decay structure. Independent lightweight wavelet decomposition is performed on the short-term, medium-term, and long-term time-series signals respectively. After decomposition, the frequency domain energy components of each layer within the preset frequency band are extracted to generate the frequency domain energy components of each layer. Based on the proportion of each frequency domain energy component in the total decomposed energy, the aging layer weight parameters for the short-term, medium-term, and long-term layers are calculated respectively, and the aging layer weight parameters for different time scales are output.

6. The method for detecting the charge level of a new energy vehicle charging battery as described in claim 5, characterized in that: The specific steps for calculating the dynamic compensation coefficient using the cross-entropy decision-making mechanism are as follows: Time difference is performed on the decoupled voltage characteristic curve to generate a voltage time change rate signal; Based on the decoupled voltage characteristic curve and the aging layer weight parameters at different time scales, a three-layer independent decision path is established. The synergy of the three independent decision-making paths is quantified by the cross-entropy decision-making mechanism, and the dynamic compensation coefficient is calculated by combining the voltage time change rate signal.

7. The method for detecting the charge level of a new energy vehicle charging battery as described in claim 6, characterized in that: The specific steps for solving the real-time state of charge estimate are as follows: By combining the dynamic compensation coefficient with the decoupled voltage characteristic curve, an electrochemical equation of state is generated. Based on the battery operating current corresponding to the single cell voltage and the electrochemical state equation, a prediction calculation is performed to output a real-time estimated value of the state of charge. The voltage deviation between the decoupled voltage characteristic curve and the real-time state of charge estimate is compared. If the deviation exceeds the preset voltage threshold, the electrochemical state equation is regenerated; otherwise, the real-time state of charge estimate is output.

8. The method for detecting the charge level of a new energy vehicle charging battery as described in claim 7, characterized in that: The specific steps for updating the historical state-of-charge characteristic curve based on the transmission results are as follows: A hash digest is performed on the aging and decaying layer structure to generate a structural feature digest value. The real-time state of charge estimate is bound to the structural feature digest value and transmitted to the vehicle energy management platform through the vehicle control bus to generate a transmission result with a timestamp. The vehicle energy management platform uses a digital twin model to verify the consistency between the real-time state of charge estimate and the structural feature summary value in the timestamped transmission results; If the verification passes, the historical charge state characteristic curve is updated with time weighting based on the timestamp of the transmission result, and the updated historical charge state characteristic curve is generated. If the verification fails, the historical charge state characteristic curve is decomposed again.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the new energy vehicle charging battery power detection method according to any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the new energy vehicle charging battery power detection method according to any one of claims 1 to 8.

Images