Battery power state prediction method and device, electronic equipment and storage medium

Abstract

The invention provides a battery power state prediction method and device, electronic equipment and a storage medium, and the method comprises the steps: carrying out the data preprocessing of the working condition data of a battery, and determining the processed working condition data; constructing a battery model based on the microscopic electrochemical process, the mesoscopic equivalent circuit characteristic, the macroscopic thermal behavior and the processed working condition data; key parameters of the battery model are identified and updated based on a hierarchical online updating mechanism, so that the battery model determines internal state parameters of the battery; determining a dynamic security boundary in real time based on the internal state parameters; and by taking the dynamic safety boundary as a constraint condition, optimizing and solving the real-time available charging and discharging power of the battery through a predictive control algorithm. By constructing a multi-scale hybrid model, the precision and robustness of battery power state prediction are improved; and a hierarchical online parameter identification mechanism is adopted, so that the dynamic adaptive capacity of the model to battery aging and environment change is enhanced, and the prediction performance is ensured not to be degraded along with the use time.

Description

Technical Field

[0001] This application relates to the field of battery parameter estimation technology, and in particular to methods, apparatus, electronic devices and storage media for predicting battery power state. Background Technology

[0002] Battery State of Power (SOP) is a key state parameter in the Battery Management System (BMS), used to assess in real time the maximum power that the battery can safely release or absorb under current operating conditions. It directly affects vehicle dynamic response, energy recovery efficiency, and system operational safety. Currently, mainstream SOP prediction methods are mainly based on the Equivalent Circuit Model (ECM), establishing a dynamic response model of the battery through external observables such as voltage, current, and temperature, and combining empirical lookup table methods or simple boundary constraints for power estimation. However, these methods based on empirical models or simplified circuit models have significant limitations. First, the equivalent circuit model is essentially a "black box" or "gray box" modeling approach, only fitting the external electrical characteristics of the battery and not considering internal electrochemical state variables such as electrode potential and lithium-ion concentration distribution. This leads to a sharp drop in prediction accuracy under extreme conditions such as low temperature, high rate, or severe aging, with errors often exceeding 15%, making it difficult to meet high-precision control requirements. More critically, existing SOP prediction technologies have serious deficiencies in safety modeling. Most methods only use macroscopic parameters such as voltage, current, and temperature as safety boundary criteria, and fail to take into account microscopic failure mechanisms such as lithium plating and thermal runaway.

[0003] In summary, existing battery power state prediction technologies still have many bottlenecks in terms of model accuracy, parameter adaptability, safety boundary modeling, computational efficiency, and applicability throughout the entire battery lifecycle. There is an urgent need for a new SOP prediction method that can take into account the depth of physical mechanisms, real-time computing performance, and dynamic safety protection capabilities, so as to achieve synergistic optimization of high accuracy, high robustness, and high safety, and support the development of the next generation of intelligent battery management systems. Summary of the Invention

[0004] In view of this, the purpose of this application is to provide a method, device, electronic device and storage medium for predicting battery power state. By constructing a multi-scale hybrid model, it effectively overcomes the large prediction error of traditional equivalent circuit models under extreme conditions such as low temperature and high rate, and improves the accuracy and robustness of battery power state prediction. By adopting a hierarchical online parameter identification mechanism, it enhances the model's dynamic adaptability to battery aging and environmental changes, and ensures that the prediction performance does not degrade with the use time.

[0005] This application provides a method for predicting the state of power of a battery, the prediction method including: The acquired battery operating condition data is preprocessed to determine the processed operating condition data; A battery model was constructed based on microscopic electrochemical processes, mesoscopic equivalent circuit characteristics, macroscopic thermal behavior, and processed operating condition data. The key parameters of the battery model are identified and updated based on a hierarchical online update mechanism, so that the battery model can determine the internal state parameters of the battery. The dynamic safety boundary is determined in real time based on the internal state parameters; wherein, the dynamic safety boundary includes a lithium plating risk index and a thermal runaway risk index; Using the dynamic safety boundary as a constraint, the real-time available charge and discharge power of the battery is optimized and solved through a predictive control algorithm.

[0006] In one possible implementation, the step of preprocessing the acquired battery operating condition data to determine the processed operating condition data includes: The battery's operating condition data is subjected to real-time filtering and data synchronization correction to determine the initial operating condition data. The initial operating condition data is processed by key feature calculation to determine the dynamic internal resistance, capacity decay rate and temperature gradient of the battery. Based on the battery's dynamic internal resistance, capacity decay rate, temperature gradient, and preset evaluation criteria, determine whether there are data quality issues with the initial operating condition data; If so, the initial operating condition data is processed based on a compensation mechanism to determine the processed industrial control data; otherwise, the initial operating condition data is used as the processed industrial control data.

[0007] In one possible implementation, the construction of the battery model based on microscopic electrochemical processes, mesoscopic equivalent circuit characteristics, macroscopic thermal behavior, and processed operating data includes: For microscopic electrochemical processes, a pseudo two-dimensional electrochemical model with reduced order is adopted, and a microscopic electrochemical model is constructed by describing the lithium-ion diffusion and interfacial reaction kinetics through solid-phase diffusion equation and Butler-Volmer equation. To address the characteristics of the mesoscopic equivalent circuit, an extended equivalent circuit model including electrode potential observation nodes is constructed. A three-dimensional thermoelectric coupling model is established for macroscopic thermal behavior; The battery model is constructed by using an iterative coupling algorithm to analyze the state variable interactions between the microscopic electrochemical model, the extended equivalent circuit model, and the three-dimensional thermoelectric coupling model, as well as based on the processed operating data.

[0008] In one possible implementation, the hierarchical online update mechanism identifies and updates key parameters of the battery model to enable the battery model to determine the internal state parameters of the battery, including: The global parameter identification module based on the improved particle swarm optimization algorithm performs initial parameter calibration on the battery model. The ohmic internal resistance, polarization resistance, and polarization capacitance of the battery model are rapidly tracked in minutes using a recursive least squares method with a variable forgetting factor. The battery model is processed using a strong tracking unscented Kalman filter algorithm to predict the state of charge and health of the battery model. The long-term aging trend of the battery model is predicted using a physical information-based long short-term memory network.

[0009] In one possible implementation, the lithium plating risk index is determined using the following formula:

[0010] in, This is a lithium plating risk index. The theoretical lithium deposition potential, This is the real-time estimated negative electrode potential. This refers to the dynamic safety margin related to temperature, state of charge, and health status.

[0011] In one possible implementation, using the dynamic safety boundary as a constraint and optimizing the real-time available charge / discharge power of the battery through a predictive control algorithm includes: Based on the internal state parameters, a theoretical power boundary that simultaneously satisfies the voltage boundary and the electrochemical boundary is determined; wherein, the electrochemical boundary is determined by the difference between the real-time negative electrode potential and the lithium plating critical potential. The theoretical power boundary is combined with the dynamic safety boundary to generate a dynamic safety power range; Within the dynamic safe power range, the predictive control algorithm aims to track the target state of charge and minimize current fluctuations. It determines the real-time available charge and discharge power of the battery by solving for the optimal power command through rolling time-domain optimization.

[0012] This application embodiment also provides a battery power state prediction device, the prediction device comprising: The data preprocessing module is used to preprocess the acquired battery operating condition data and determine the processed operating condition data. The multi-scale model building module is used to build battery models based on microscopic electrochemical processes, mesoscopic equivalent circuit characteristics, macroscopic thermal behavior, and processed operating condition data. The intelligent parameter identification module is used to identify and update the key parameters of the battery model based on a hierarchical online update mechanism, so that the battery model can determine the internal state parameters of the battery. A safety boundary determination module is used to determine the dynamic safety boundary in real time based on the internal state parameters; wherein, the dynamic safety boundary includes a lithium plating risk index and a thermal runaway risk index; The power determination module is used to use the dynamic safety boundary as a constraint condition and optimize the solution of the real-time available charge and discharge power of the battery through a predictive control algorithm.

[0013] In one possible implementation, the data preprocessing module is used to preprocess the acquired battery operating condition data to determine the processed operating condition data: The battery's operating condition data is subjected to real-time filtering and data synchronization correction to determine the initial operating condition data. The initial operating condition data is processed by key feature calculation to determine the dynamic internal resistance, capacity decay rate and temperature gradient of the battery. Based on the battery's dynamic internal resistance, capacity decay rate, temperature gradient, and preset evaluation criteria, determine whether there are data quality issues with the initial operating condition data; If so, the initial operating condition data is processed based on a compensation mechanism to determine the processed industrial control data; otherwise, the initial operating condition data is used as the processed industrial control data.

[0014] This application also provides an electronic device, including: a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor communicates with the memory via the bus. When the machine-readable instructions are executed by the processor, the steps of the battery power state prediction method described above are performed.

[0015] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, performs the steps of the battery power state prediction method described above.

[0016] This application provides a method, apparatus, electronic device, and storage medium for predicting battery power state. The prediction method includes: preprocessing acquired battery operating condition data to determine processed operating condition data; constructing a battery model based on microscopic electrochemical processes, mesoscopic equivalent circuit characteristics, macroscopic thermal behavior, and processed operating condition data; identifying and updating key parameters of the battery model based on a hierarchical online update mechanism to determine the battery's internal state parameters; determining a dynamic safety boundary in real time based on the internal state parameters; wherein the dynamic safety boundary includes a lithium plating risk index and a thermal runaway risk index; and using the dynamic safety boundary as a constraint, optimizing the real-time available charge and discharge power of the battery through a predictive control algorithm. By constructing a multi-scale hybrid model, the large prediction error of traditional equivalent circuit models under extreme conditions such as low temperature and high rate is effectively overcome, improving the accuracy and robustness of battery power state prediction; the hierarchical online parameter identification mechanism enhances the model's dynamic adaptability to battery aging and environmental changes, ensuring that prediction performance does not degrade with usage time.

[0017] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 A flowchart illustrating a method for predicting the state of power of a battery provided in an embodiment of this application; Figure 2 This is a flowchart illustrating the construction of a battery model provided in an embodiment of this application; Figure 3 A schematic diagram of the structure of a battery power state prediction device provided in an embodiment of this application; Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of this application. Based on the embodiments of this application, every other embodiment obtained by those skilled in the art without inventive effort falls within the scope of protection of this application.

[0021] First, the applicable scenarios for this application will be introduced. This application can be applied to the field of battery parameter estimation technology.

[0022] Research has revealed that existing battery power state prediction technologies still face numerous bottlenecks in terms of model accuracy, parameter adaptability, safety boundary modeling, computational efficiency, and lifecycle applicability. There is an urgent need for a novel SOP prediction method that can balance the depth of physical mechanisms, real-time computational performance, and dynamic safety protection capabilities to achieve synergistic optimization of high accuracy, high robustness, and high safety, thereby supporting the development of next-generation intelligent battery management systems.

[0023] Based on this, the embodiments of this application provide a method for predicting the state of power of a battery. By constructing a multi-scale hybrid model, it effectively overcomes the large prediction error of the traditional equivalent circuit model under extreme conditions such as low temperature and high rate, and improves the accuracy and robustness of battery state of power prediction. By adopting a hierarchical online parameter identification mechanism, it enhances the model's dynamic adaptability to battery aging and environmental changes, and ensures that the prediction performance does not degrade with the use time.

[0024] Please see Figure 1 , Figure 1 This is a flowchart illustrating a method for predicting the state of power of a battery, as provided in an embodiment of this application. Figure 1 As shown in the embodiments of this application, the prediction method includes: S101: Perform data preprocessing on the acquired battery operating condition data to determine the processed operating condition data.

[0025] It should be noted that for voltage acquisition, the system employs a differential amplifier circuit design to effectively suppress common-mode interference, and uses a high-precision analog-to-digital converter to digitize the voltage signal. Current acquisition utilizes a Hall effect sensor based on the zero-flux principle, coupled with a 16-bit ADC to achieve high-precision measurement. The system monitors the current direction in real time and automatically switches ranges to ensure accurate current data under various operating conditions. The temperature monitoring network adopts a distributed design, with multiple temperature sensors placed at key locations on the battery surface.

[0026] Here, data preprocessing includes real-time filtering, data synchronization correction, and determination of key feature parameters.

[0027] In one possible implementation, the step of preprocessing the acquired battery operating condition data to determine the processed operating condition data includes: (1): The battery's operating condition data is filtered in real time and the data is synchronized and corrected to determine the initial operating condition data.

[0028] Here, an adaptive digital filter is used to filter the operating data.

[0029] Among them, the time-scale compensation model is established as follows: ,in, This is the timestamp after compensation and correction. This is the original timestamp. For transmission line delay, To handle delays.

[0030] (2): Perform key feature calculations on the initial operating data to determine the dynamic internal resistance, capacity decay rate and temperature gradient of the battery.

[0031] Here, the dynamic internal resistance is determined using the following formula. :

[0032] Where ΔV is the voltage change and ΔI is the current step change.

[0033] Here, the capacity decay rate C is determined using the following formula:

[0034] in, This is the initial rated capacity. This represents the current actual capacity.

[0035] Here, the temperature gradient is determined using the following formula. :

[0036] in, This is the temperature value.

[0037] (3): Based on the dynamic internal resistance of the battery, the capacity decay rate, the temperature gradient and the preset index evaluation standard, determine whether there is a data quality problem in the initial operating condition data; if so, process the initial operating condition data based on the compensation mechanism to determine the processed industrial control data; otherwise, use the initial operating condition data as the processed industrial control data.

[0038] Here, the battery's dynamic internal resistance, capacity decay rate, and temperature gradient are compared with preset evaluation standards to determine whether there are data quality issues with the initial operating condition data. If so, the initial operating condition data is processed according to the compensation mechanism to determine the processed industrial control data; otherwise, the initial operating condition data is used as the processed industrial control data.

[0039] The evaluation criteria include standard values ​​for dynamic internal resistance, capacity decay rate, and temperature gradient.

[0040] S102: A battery model was constructed based on microscopic electrochemical processes, mesoscopic equivalent circuit characteristics, macroscopic thermal behavior, and processed operating condition data.

[0041] In this step, a battery model is constructed based on the microscopic electrochemical process, mesoscopic equivalent circuit characteristics, macroscopic thermal behavior, and processed operating data.

[0042] Here, by organically combining microscopic electrochemical processes, mesoscopic equivalent circuit characteristics, and macroscopic thermodynamic behavior, a comprehensive characterization of the battery's internal state is achieved. This modeling framework comprises three key levels: Microscopic electrochemical processes, which establish a particle-level electrochemical model based on first-principles calculations to describe ion diffusion and charge transfer processes within the electrode materials; Mesoscopic equivalent circuit characteristics, which construct an improved equivalent circuit network to accurately reflect the battery's dynamic response characteristics; and Macroscopic thermodynamic behavior, which establishes a three-dimensional thermo-electric coupling model to predict the battery's temperature field distribution and thermal evolution.

[0043] Please see Figure 2 , Figure 2 This is a flowchart illustrating the construction of a battery model provided in an embodiment of this application. Figure 2 As shown, it includes: S201: For microscopic electrochemical processes, a pseudo-two-dimensional electrochemical model with reduced order is adopted. The microscopic electrochemical model is constructed by describing the lithium-ion diffusion and interfacial reaction kinetics through solid-phase diffusion equations and Butler-Volmer equations.

[0044] Here, the solid-state diffusion equation (PDE) is used to describe lithium-ion diffusion within the active particles using an improved single-particle model:

[0045] Boundary conditions:

[0046] Electrochemical reaction kinetics, based on the Butler-Volmer equation to describe interfacial reactions:

[0047] Overpotential:

[0048] Model reduction is performed by converting the PDE into an ODE using the Galerkin projection method.

[0049] in, This refers to the concentration of lithium ions in the solid phase. The time it takes for the diffusion process to proceed. The solid-phase diffusion coefficient is... Let be the radius distance from the center of the active material particle to a point inside the particle. J is the reaction current density. For exchange current density, , The transfer coefficients of the anode and cathode are respectively, where F is a constant. Let T be the gas constant and T be the absolute temperature. This is an overpotential. For solid-state potential, This represents the potential in the electrolyte. To be at equilibrium potential, For SEI film resistance. Let A be the reduced state vector, A be the state matrix, and B be the input matrix.

[0050] S202: Construct an extended equivalent circuit model that includes electrode potential observation nodes for the characteristics of the mesoscopic equivalent circuit.

[0051] Extended equivalent circuit structure:

[0052] Wherein the polarization voltage is:

[0053] Negative electrode potential calculation:

[0054] Positive electrode potential calculation:

[0055] Dynamic parameter model:

[0056] Where g(SOC) is the SOC dependency function, Terminal voltage, Open circuit voltage, Positive potential The negative electrode potential is... For ohmic internal resistance, This is the operating current. This is the polarization voltage. Let be the voltage of the i-th RC circuit. as well as All are time constants. as well as All are electrode reaction resistances. For charge transfer resistor, For activation energy, Here, T is the universal gas constant, and T is the current absolute temperature. For reference absolute temperature.

[0057] S203: A three-dimensional thermoelectric coupling model is established for macroscopic thermal behavior.

[0058] Three-dimensional thermoelectric coupling model equations:

[0059] Heat source item:

[0060] Boundary condition handling:

[0061] Discretization is performed using the finite volume method to generate a computational mesh:

[0062] Where ρ is the material density, For isobaric specific heat capacity, Let k be the rate of change of temperature over time, and k be the thermal conductivity. For volumetric heat production rate, The heat of reaction, It is Joule heat. The coefficient of convective heat transfer is given by T, where T is the surface temperature of the battery. Let k be the ambient temperature, and k∂T / ∂n be the surface heat flux density. Let $T$ be the temperature at node $i$ at (n+1)$, and $Δt$ be the time step. Let i be the equivalent thermal conductivity between nodes i and j. Let i be the contact area between nodes i and j. Let j be the temperature of the neighboring node j in the current era. Let be the distance between the centroids of nodes i and j. To control the volumetric heat generation rate within volume i.

[0063] S204: The battery model is constructed by using an iterative coupling algorithm to analyze the state variable interactions between the microscopic electrochemical model, the extended equivalent circuit model, and the three-dimensional thermoelectric coupling model, as well as based on the processed operating condition data.

[0064] In this step, a battery model is constructed by using an iterative coupling algorithm to analyze the interaction of state variables between the microscopic electrochemical model, the extended equivalent circuit model, and the three-dimensional thermoelectric coupling model, as well as based on the processed operating data.

[0065] This application improves the accuracy and robustness of battery power state prediction. By constructing a multi-scale hybrid model that integrates microscopic electrochemical mechanisms, mesoscopic equivalent circuit characteristics, and macroscopic thermo-electric coupling behavior, the problem of large prediction errors (typically exceeding 15%) of traditional equivalent circuit models under extreme conditions such as low temperature and high rate is effectively overcome. Experiments show that within a wide temperature range of -30°C to 60°C, the power prediction error of this invention is controlled within 3%; under 4C and above fast charging conditions, the prediction accuracy is improved by more than 40% compared with traditional methods.

[0066] S103: Based on a hierarchical online update mechanism, the key parameters of the battery model are identified and updated so that the battery model can determine the internal state parameters of the battery.

[0067] In this step, the key parameters of the battery model are identified and updated according to the hierarchical online update mechanism so that the battery model can determine the internal state parameters of the battery.

[0068] Here, a hierarchical online update mechanism is designed: the bottom layer uses recursive least squares with a variable forgetting factor to track rapidly changing polarization parameters; the middle layer uses unscented Kalman filtering to estimate SOC and SOH; and the top layer uses a deep learning network to predict long-term aging trends. Experimental results show that this scheme improves parameter identification accuracy by 60% and reduces computation time by 40%.

[0069] In one possible implementation, the hierarchical online update mechanism identifies and updates key parameters of the battery model to enable the battery model to determine the internal state parameters of the battery, including: A: The global parameter identification module based on the improved particle swarm optimization algorithm performs initial parameter calibration on the battery model.

[0070] Here, the global parameter identification module of the improved particle swarm optimization algorithm is used to calibrate the initial parameters of the battery model, and the quantum behavior particle swarm optimization algorithm is used to solve the problem that traditional methods are prone to getting trapped in local optima.

[0071] B: The ohmic internal resistance, polarization resistance, and polarization capacitance of the battery model are rapidly tracked in minutes using a recursive least squares method with a variable forgetting factor.

[0072] Here, a novel Dynamically Weighted Multiple-Inspiration Recursive Least Squares (DW-MI-RLS) method is proposed to address rapidly changing parameters such as ohmic internal resistance and charge transfer resistance. This method extends the traditional single-inspiration update by introducing a sliding window mechanism.

[0073]

[0074] In the formula, Φ(k)=[φ(k),φ(k-1),φ(k-2)] is the three-innovation matrix, and Y(k)=[y(k),y(k-1),y(k-2)]^T is the observation vector. The dynamic weight λ is adaptively adjusted according to the current change rate, and the response speed is improved by 60% under a current step of 10A / s. k is the index of the current time, i.e., the kth sampling time. Let K be the parameter vector to be identified, K(k) be the gain matrix, and Y(k) be the augmented observation vector. Let P(k-1) be the regression vector, P(k-1) be the covariance matrix, λ be the forgetting factor, and I be the identity matrix.

[0075] In this application, the model's dynamic adaptability to battery aging and environmental changes is enhanced. A hierarchical online parameter identification mechanism is employed, combining Dynamically Weighted Multiple Innovation Recursive Least Squares (DW-MI-RLS), Strong Tracking Unscented Kalman Filter (ST-UKF), and Physical Information Embedded LSTM Network (PI-LSTM) to achieve full lifecycle parameter updates. This scheme significantly suppresses model parameter drift caused by cyclic aging and temperature fluctuations, keeping parameter estimation bias stable within 5% over the long term and ensuring that prediction performance does not degrade with usage time.

[0076] C: The battery model is processed using a strong tracking unscented Kalman filter algorithm to predict the state of charge and health of the battery model.

[0077] Here, the estimation lag problem of traditional methods during SOC abrupt changes is solved based on the strong tracking unscented Kalman filter (ST-UKF) algorithm.

[0078] D: The long-term aging trend of the battery model is predicted using a physical information-based long short-term memory network.

[0079] Here, a hybrid physical information LSTM network (PI-LSTM) is constructed. The basic equations of the battery are embedded in the input layer of the network to preserve physical interpretability. The hidden layer adopts a combination of gated recurrent units and attention mechanism. The output layer is constrained to satisfy the second law of thermodynamics.

[0080] S104: The dynamic safety boundary is determined in real time based on the internal state parameters; wherein, the dynamic safety boundary includes the lithium plating risk index and the thermal runaway risk index.

[0081] In this step, the dynamic safety boundary is determined in real time based on the internal state parameters.

[0082] In one possible implementation, the lithium plating risk index is determined using the following formula:

[0083] in, This is a lithium plating risk index. The theoretical lithium deposition potential, This is the real-time estimated negative electrode potential. This refers to the dynamic safety margin related to temperature, state of charge, and health status.

[0084] Here, the thermal runaway risk index is determined using the following formula. :

[0085] This application, for the first time, deeply embeds the electrochemical safety boundary into the power prediction process to achieve proactive safety protection. By estimating the negative electrode potential in real time and establishing a dynamic lithium plating criterion, a quantitative evaluation mechanism of the lithium plating risk index (LRI) and safety index (Safety_Index) is proposed. An early warning is triggered when the LRI ≥ 80%, and charging is immediately terminated when it ≥ 100%. This mechanism can reduce the risk of lithium plating during fast charging by more than 90%, effectively avoid internal short circuits and thermal runaway caused by lithium dendrite growth, and significantly improve the operational safety of the battery system.

[0086] S105: Using the dynamic safety boundary as a constraint, the real-time available charge and discharge power of the battery is optimized and solved through a predictive control algorithm.

[0087] In this step, the dynamic safety boundary is used as a constraint, and the real-time available charge and discharge power of the battery is optimized and solved through predictive control algorithms.

[0088] Here, the system achieves closed-loop control of the entire power prediction process through a three-level optimization architecture: first, the theoretical power boundary is calculated in real time based on an enhanced equivalent circuit model; then, safety constraints are embedded to generate a dynamic safe power range; finally, the optimal power command is solved through a model predictive control algorithm. This architecture design ensures both the physical rationality of the prediction results and the safety of system operation.

[0089] In one possible implementation, using the dynamic safety boundary as a constraint and optimizing the real-time available charge / discharge power of the battery through a predictive control algorithm includes: a: Based on the internal state parameters, a theoretical power boundary that simultaneously satisfies the voltage boundary and the electrochemical boundary is determined; wherein, the electrochemical boundary is determined by the difference between the real-time negative electrode potential and the lithium plating critical potential.

[0090] Here, the theoretical power boundary that simultaneously satisfies the voltage boundary and the electrochemical boundary is determined based on the internal state parameters.

[0091] Among them, the dynamic power equation based on electrode potential is:

[0092]

[0093] This equation considers both electrical and electrochemical boundaries. A key innovation lies in the introduction of the potential sensitivity term ∂V / ∂I, which physically represents the negative electrode potential shift caused by a unit change in current. A safety constraint embedding layer constructs a multi-dimensional safety protection network. For maximum available charging power, For the maximum available discharge power, This is the charging / discharging cutoff voltage. This is the current segment voltage. This represents the real-time negative electrode potential. This is the critical potential for lithium plating. It is the internal resistance of the Ohm.

[0094] b: Combine the theoretical power boundary with the dynamic safety boundary to generate a dynamic safety power range.

[0095] c: Within the dynamic safe power range, the predictive control algorithm aims to track the target state of charge and minimize current fluctuations. It determines the real-time available charge and discharge power of the battery by solving for the optimal power command through rolling time-domain optimization.

[0096] Here, the optimization algorithm layer employs an improved model predictive control framework. Its innovation lies in decomposing the long-time domain optimization problem into a rolling time-domain solution:

[0097] In the formula, the prediction time domain N_p = 20 steps (corresponding to 20s), and the control time domain N_c = 5 steps. The solver uses the dual interior-point method, combined with warm-start technology to compress the computation time to within 5ms. For the predicted SOC at time k, This is the SOC reference value. As the weight of the SOC tracking error, Let R be the control variable (current) at time k, and S be the weight of the current amplitude and the current rate of change. The predicted negative electrode potential at time k. This is the critical potential for lithium plating. This is the SOC working window.

[0098] Here, based on the Model Predictive Control (MPC) framework, the electrochemical safety boundary is strictly constrained while maximizing available power, enabling the battery to maintain a good health state even under high-intensity usage scenarios. Real-world test results show that, under the same fast-charging strategy, the battery cycle life can reach 1.8 times that of conventional methods; in energy storage applications, the system's usable capacity is increased by more than 15%.

[0099] In this application, the hardware acceleration system completely solves the problem of real-time execution of complex algorithms on embedded platforms through heterogeneous computing architecture and algorithm-level optimization. The system adopts an "FPGA+ARM" dual-core architecture, where the FPGA is responsible for time-sensitive tasks, and the ARM processor runs slowly changing parameter updates and decision logic. This division of labor fully leverages the parallel processing advantages of the FPGA and the flexible control characteristics of the ARM, reducing the measured data processing latency to 1 / 5 of the traditional solution. Three parallel pipelines are designed inside the FPGA: the first pipeline handles voltage and current signal acquisition and filtering, using a sliding window mean filtering algorithm to achieve zero-wait filtering output at a 100MHz clock; the second pipeline is dedicated to multi-scale model solving, implementing hardware acceleration of matrix multiplication and addition operations through a fixed arithmetic unit, with a single RC network solution requiring only 0.8μs; the third pipeline is responsible for real-time calculation of safety boundaries, with 48 built-in parallel comparators, capable of completing threshold judgments for 15 safety parameters within a single clock cycle.

[0100] The beneficial effects of this application are as follows: In terms of model architecture, it creatively integrates the equivalent circuit model and the quasi-electrochemical model to construct a physically interpretable "gray box" model architecture. By introducing a dynamic observation layer for electrode potential, this model achieves real-time estimation of key internal states such as negative electrode potential and lithium-ion surface concentration, enabling the model to maintain a prediction accuracy within ±3% over a wide temperature range of -30℃ to 60℃. Especially under high-rate conditions above 4C, it reduces the prediction error by more than 40% compared to traditional methods.

[0101] Regarding dynamic adaptability, this application developed an adaptive algorithm system based on multi-timescale parameter identification. By designing a hierarchical parameter update mechanism, accurate tracking of different stages of battery aging is achieved: short-term parameters are updated minute-by-minute using a recursive least squares method with a forgetting factor; medium-term parameters are corrected through daily capacity calibration; and long-term parameters are adjusted monthly in conjunction with health status assessment. Experimental data shows that this scheme enables the model to maintain stable predictive performance throughout the entire battery lifespan, with parameter drift error controlled within 5%.

[0102] In terms of safety and control, this application breaks through the technical limitations of traditional power prediction by innovatively embedding the electrochemical safety boundary deep into the prediction algorithm. Specifically, this includes: 1) establishing a lithium plating risk prediction model based on real-time estimation of the negative electrode potential, ensuring that the negative electrode potential is always higher than the critical lithium plating potential by dynamically adjusting the charging power; 2) developing a multi-parameter coupled thermal runaway early warning algorithm, comprehensively considering characteristic parameters such as temperature change rate and voltage relaxation characteristics; and 3) designing a lithium deposition uniformity evaluation index to prevent localized lithium plating. Experiments demonstrate that this scheme can reduce the lithium plating risk during battery fast charging by more than 90%, significantly improving system safety.

[0103] In terms of computational efficiency, this application achieves a significant breakthrough through innovative algorithm architecture. It employs model reduction techniques to simplify complex electrochemical equations into equivalent circuit forms while preserving key physical features; it develops a hardware-accelerated parallel computing framework, compressing the algorithm's computation time to less than 5ms; and it innovatively utilizes feature parameter database technology to significantly reduce the computational load of online parameter identification. These technological innovations improve the computational efficiency of this invention by two orders of magnitude compared to traditional electrochemical models, fully meeting the millisecond-level real-time requirements of BMS systems.

[0104] This application provides a method for predicting the state of power of a battery. The method includes: preprocessing the acquired battery operating condition data to determine the processed operating condition data; constructing a battery model based on microscopic electrochemical processes, mesoscopic equivalent circuit characteristics, macroscopic thermal behavior, and the processed operating condition data; identifying and updating the key parameters of the battery model based on a hierarchical online update mechanism to determine the internal state parameters of the battery; determining a dynamic safety boundary in real time based on the internal state parameters; wherein the dynamic safety boundary includes a lithium plating risk index and a thermal runaway risk index; and using the dynamic safety boundary as a constraint, optimizing the real-time available charge and discharge power of the battery through a predictive control algorithm. By constructing a multi-scale hybrid model, the large prediction error of traditional equivalent circuit models under extreme conditions such as low temperature and high rate is effectively overcome, improving the accuracy and robustness of battery power state prediction; the hierarchical online parameter identification mechanism enhances the model's dynamic adaptability to battery aging and environmental changes, ensuring that the prediction performance does not degrade with usage time.

[0105] Please see Figure 3 , Figure 3 This is a schematic diagram of a battery power state prediction device provided in an embodiment of this application. Figure 3 As shown, the battery power state prediction device 300 includes: The data preprocessing module 310 is used to preprocess the acquired battery operating condition data and determine the processed operating condition data. The multi-scale model building module 320 is used to build a battery model based on microscopic electrochemical processes, mesoscopic equivalent circuit characteristics, macroscopic thermal behavior, and processed operating condition data. The intelligent parameter identification module 330 is used to identify and update the key parameters of the battery model based on a hierarchical online update mechanism, so that the battery model can determine the internal state parameters of the battery. The safety boundary determination module 340 is used to determine the dynamic safety boundary in real time based on the internal state parameters; wherein, the dynamic safety boundary includes a lithium plating risk index and a thermal runaway risk index; The power determination module 350 is used to use the dynamic safety boundary as a constraint condition and optimize the solution of the real-time available charge and discharge power of the battery through a predictive control algorithm.

[0106] Furthermore, the data preprocessing module 310 is used to preprocess the acquired battery operating condition data to determine the processed operating condition data: The battery's operating condition data is subjected to real-time filtering and data synchronization correction to determine the initial operating condition data. The initial operating condition data is processed by key feature calculation to determine the dynamic internal resistance, capacity decay rate and temperature gradient of the battery. Based on the battery's dynamic internal resistance, capacity decay rate, temperature gradient, and preset evaluation criteria, determine whether there are data quality issues with the initial operating condition data; If so, the initial operating condition data is processed based on a compensation mechanism to determine the processed industrial control data; otherwise, the initial operating condition data is used as the processed industrial control data.

[0107] Furthermore, the multi-scale model building module 320 is used to construct the battery model based on the microscopic electrochemical processes, mesoscopic equivalent circuit characteristics, macroscopic thermal behavior, and processed operating condition data. For microscopic electrochemical processes, a pseudo two-dimensional electrochemical model with reduced order is adopted, and a microscopic electrochemical model is constructed by describing the lithium-ion diffusion and interfacial reaction kinetics through solid-phase diffusion equation and Butler-Volmer equation. To address the characteristics of the mesoscopic equivalent circuit, an extended equivalent circuit model including electrode potential observation nodes is constructed. A three-dimensional thermoelectric coupling model is established for macroscopic thermal behavior; The battery model is constructed by using an iterative coupling algorithm to analyze the state variable interactions between the microscopic electrochemical model, the extended equivalent circuit model, and the three-dimensional thermoelectric coupling model, as well as based on the processed operating data.

[0108] Furthermore, the intelligent parameter identification module 330 is used to identify and update the key parameters of the battery model using the hierarchical online update mechanism, so that the battery model can determine the internal state parameters of the battery. The global parameter identification module based on the improved particle swarm optimization algorithm performs initial parameter calibration on the battery model. The ohmic internal resistance, polarization resistance, and polarization capacitance of the battery model are rapidly tracked in minutes using a recursive least squares method with a variable forgetting factor. The battery model is processed using a strong tracking unscented Kalman filter algorithm to predict the state of charge and health of the battery model. The long-term aging trend of the battery model is predicted using a physical information-based long short-term memory network.

[0109] Furthermore, the safety boundary determination module 340 determines the lithium plating risk index using the following formula:

[0110] in, This is a lithium plating risk index. The theoretical lithium deposition potential, This is the real-time estimated negative electrode potential. This refers to the dynamic safety margin related to temperature, state of charge, and health status.

[0111] Furthermore, the power determination module 350 is used to optimize and solve for the real-time available charge and discharge power of the battery by using the dynamic safety boundary as a constraint condition and through a predictive control algorithm. Based on the internal state parameters, a theoretical power boundary that simultaneously satisfies the voltage boundary and the electrochemical boundary is determined; wherein, the electrochemical boundary is determined by the difference between the real-time negative electrode potential and the lithium plating critical potential. The theoretical power boundary is combined with the dynamic safety boundary to generate a dynamic safety power range; Within the dynamic safe power range, the predictive control algorithm aims to track the target state of charge and minimize current fluctuations. It determines the real-time available charge and discharge power of the battery by solving for the optimal power command through rolling time-domain optimization.

[0112] This application provides a battery power state prediction device, comprising: a data preprocessing module for preprocessing acquired battery operating condition data to determine processed operating condition data; a multi-scale model construction module for constructing a battery model based on microscopic electrochemical processes, mesoscopic equivalent circuit characteristics, macroscopic thermal behavior, and processed operating condition data; an intelligent parameter identification module for identifying and updating key parameters of the battery model based on a hierarchical online update mechanism to enable the battery model to determine the internal state parameters of the battery; a safety boundary determination module for determining a dynamic safety boundary in real time based on the internal state parameters; wherein the dynamic safety boundary includes a lithium plating risk index and a thermal runaway risk index; and a power determination module for using the dynamic safety boundary as a constraint condition and optimizing the real-time available charge and discharge power of the battery through a predictive control algorithm. By constructing a multi-scale hybrid model, the large prediction error of the traditional equivalent circuit model under extreme conditions such as low temperature and high rate is effectively overcome, thus improving the accuracy and robustness of battery power state prediction. The hierarchical online parameter identification mechanism enhances the model's dynamic adaptability to battery aging and environmental changes, ensuring that the prediction performance does not degrade with the use time.

[0113] Please see Figure 4 , Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 4 As shown, the electronic device 400 includes a processor 410, a memory 420, and a bus 430.

[0114] The memory 420 stores machine-readable instructions executable by the processor 410. When the electronic device 400 is running, the processor 410 communicates with the memory 420 via the bus 430. When the machine-readable instructions are executed by the processor 410, they can perform the operations described above. Figure 1 The steps of the battery power state prediction method in the method embodiment shown are described in detail in the method embodiment, and will not be repeated here.

[0115] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, can perform the above-described actions. Figure 1 The steps of the battery power state prediction method in the method embodiment shown are described in detail in the method embodiment, and will not be repeated here.

[0116] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0117] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the shown or discussed mutual couplings, direct couplings, or communication connections may be through some communication interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms.

[0118] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0119] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0120] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a processor-executable, non-volatile, computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0121] Finally, it should be noted that the above-described embodiments are merely specific implementations of this application, used to illustrate the technical solutions of this application, and not to limit them. The scope of protection of this application is not limited thereto. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments, or make equivalent substitutions for some of the technical features, within the scope of the technology disclosed in this application. Such modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for predicting the state of power of a battery, characterized in that, The prediction method includes: The acquired battery operating condition data is preprocessed to determine the processed operating condition data; A battery model was constructed based on microscopic electrochemical processes, mesoscopic equivalent circuit characteristics, macroscopic thermal behavior, and processed operating condition data. The key parameters of the battery model are identified and updated based on a hierarchical online update mechanism, so that the battery model can determine the internal state parameters of the battery. The dynamic safety boundary is determined in real time based on the internal state parameters; wherein, the dynamic safety boundary includes a lithium plating risk index and a thermal runaway risk index; Using the dynamic safety boundary as a constraint, the real-time available charge and discharge power of the battery is optimized and solved through a predictive control algorithm.

2. The prediction method according to claim 1, characterized in that, The process of preprocessing the acquired battery operating condition data to determine the processed operating condition data includes: The battery's operating condition data is subjected to real-time filtering and data synchronization correction to determine the initial operating condition data. The initial operating condition data is processed by key feature calculation to determine the dynamic internal resistance, capacity decay rate and temperature gradient of the battery. Based on the battery's dynamic internal resistance, capacity decay rate, temperature gradient, and preset evaluation criteria, determine whether there are data quality issues with the initial operating condition data; If so, the initial operating condition data is processed based on a compensation mechanism to determine the processed industrial control data; otherwise, the initial operating condition data is used as the processed industrial control data.

3. The prediction method according to claim 1, characterized in that, The battery model, constructed based on microscopic electrochemical processes, mesoscopic equivalent circuit characteristics, macroscopic thermal behavior, and processed operating data, includes: For microscopic electrochemical processes, a pseudo two-dimensional electrochemical model with reduced order is adopted, and a microscopic electrochemical model is constructed by describing the lithium-ion diffusion and interfacial reaction kinetics through solid-phase diffusion equation and Butler-Volmer equation. To address the characteristics of the mesoscopic equivalent circuit, an extended equivalent circuit model including electrode potential observation nodes is constructed. A three-dimensional thermoelectric coupling model is established for macroscopic thermal behavior; The battery model is constructed by using an iterative coupling algorithm to analyze the state variable interactions between the microscopic electrochemical model, the extended equivalent circuit model, and the three-dimensional thermoelectric coupling model, as well as based on the processed operating data.

4. The prediction method according to claim 1, characterized in that, The hierarchical online update mechanism identifies and updates key parameters of the battery model to enable the battery model to determine the internal state parameters of the battery, including: The global parameter identification module based on the improved particle swarm optimization algorithm performs initial parameter calibration on the battery model. The ohmic internal resistance, polarization resistance, and polarization capacitance of the battery model are rapidly tracked in minutes using a recursive least squares method with a variable forgetting factor. The battery model is processed using a strong tracking unscented Kalman filter algorithm to predict the state of charge and health of the battery model. The long-term aging trend of the battery model is predicted using a physical information-based long short-term memory network.

5. The prediction method according to claim 1, characterized in that, The lithium plating risk index is determined using the following formula: in, This is a lithium plating risk index. The theoretical lithium deposition potential, This is the real-time estimated negative electrode potential. This refers to the dynamic safety margin related to temperature, state of charge, and health status.

6. The prediction method according to claim 1, characterized in that, The step of using the dynamic safety boundary as a constraint and optimizing the real-time available charge and discharge power of the battery through a predictive control algorithm includes: Based on the internal state parameters, a theoretical power boundary that simultaneously satisfies the voltage boundary and the electrochemical boundary is determined; wherein, the electrochemical boundary is determined by the difference between the real-time negative electrode potential and the lithium plating critical potential. The theoretical power boundary is combined with the dynamic safety boundary to generate a dynamic safety power range; Within the dynamic safe power range, the predictive control algorithm aims to track the target state of charge and minimize current fluctuations. It determines the real-time available charge and discharge power of the battery by solving for the optimal power command through rolling time-domain optimization.

7. A device for predicting the state of power of a battery, characterized in that, The prediction device includes: The data preprocessing module is used to preprocess the acquired battery operating condition data and determine the processed operating condition data. The multi-scale model building module is used to build battery models based on microscopic electrochemical processes, mesoscopic equivalent circuit characteristics, macroscopic thermal behavior, and processed operating condition data. The intelligent parameter identification module is used to identify and update the key parameters of the battery model based on a hierarchical online update mechanism, so that the battery model can determine the internal state parameters of the battery. A safety boundary determination module is used to determine the dynamic safety boundary in real time based on the internal state parameters; wherein, the dynamic safety boundary includes a lithium plating risk index and a thermal runaway risk index; The power determination module is used to use the dynamic safety boundary as a constraint condition and optimize the solution of the real-time available charge and discharge power of the battery through a predictive control algorithm.

8. The prediction device according to claim 7, characterized in that, The data preprocessing module is used to preprocess the acquired battery operating condition data to determine the processed operating condition data. The battery's operating condition data is subjected to real-time filtering and data synchronization correction to determine the initial operating condition data. The initial operating condition data is processed by key feature calculation to determine the dynamic internal resistance, capacity decay rate and temperature gradient of the battery. Based on the battery's dynamic internal resistance, capacity decay rate, temperature gradient, and preset evaluation criteria, determine whether there are data quality issues with the initial operating condition data; If so, the initial operating condition data is processed based on a compensation mechanism to determine the processed industrial control data; otherwise, the initial operating condition data is used as the processed industrial control data.

9. An electronic device, characterized in that, include: The device includes a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor communicates with the memory via the bus. The machine-readable instructions are executed by the processor to perform the steps of the battery power state prediction method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, performs the steps of the battery power state prediction method as described in any one of claims 1 to 7.

Images