A household energy storage battery safety intelligent early warning system and method

Abstract

The application relates to the technical field of energy storage batteries, and discloses a household energy storage battery safety intelligent early warning system and method, wherein the method comprises the following steps: collecting battery operation data of a household energy storage battery; substituting the battery operation data into an electrochemical state space model to calculate an electrode reaction rate, obtaining an exchange current density and a transfer coefficient, and calculating electrochemical state parameters; calculating a temperature deviation value, a gas pressure deviation value and a potential deviation value according to the electrochemical state parameters, and performing adaptive noise parameter adjustment to obtain a safety evaluation value; calculating a Gibbs free energy change value and a side reaction rate according to the safety evaluation value, and performing electrochemical thermodynamic equilibrium condition judgment based on the Gibbs free energy change value and the side reaction rate to generate a warning control instruction. The method improves the state estimation precision under the conditions of battery aging and temperature gradient change, and ensures the accuracy and robustness of the energy storage battery safety warning.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of energy storage batteries, in particular to a household energy storage battery safety intelligent early warning system and method. BACKGROUND

[0002] The operation environment of household energy storage batteries faces multiple challenges such as temperature fluctuations, load changes, and cycle aging. These factors interact with the electrochemical reaction process inside the battery, affecting the safety state of the battery. The robustness of the traditional filtering algorithm is severely insufficient under the influence of battery aging effect and temperature gradient change. In addition, the fixed threshold early warning strategy cannot adapt to the time-varying characteristics of electrochemical parameters and the dynamic decay of battery health state, resulting in low early warning accuracy. SUMMARY

[0003] The present application provides a household energy storage battery safety intelligent early warning system and method, which improves the state estimation accuracy under the conditions of battery aging and temperature gradient change, and ensures the accuracy and robustness of the safety early warning of energy storage batteries.

[0004] In a first aspect, the present application provides a household energy storage battery safety intelligent early warning method, which comprises:

[0005] Collecting battery operation data of a household energy storage battery;

[0006] Substituting the battery operation data into an electrochemical state space model to calculate the electrode reaction rate, obtaining the exchange current density and the transfer coefficient, and calculating the electrochemical state parameters based on the exchange current density and the transfer coefficient;

[0007] According to the electrochemical state parameters, calculate the temperature deviation value, the gas pressure deviation value and the potential deviation value, and based on the temperature deviation value, the gas pressure deviation value and the potential deviation value, adjust the adaptive noise parameters to obtain the safety evaluation value;

[0008] According to the safety evaluation value, calculate the Gibbs free energy change value and the side reaction rate, and based on the Gibbs free energy change value and the side reaction rate, execute the electrochemical thermodynamic equilibrium condition judgment, and generate the early warning control instruction.

[0009] In combination with the first aspect, in a first implementation manner of the first aspect of the present application, the battery operation data of the household energy storage battery is collected, comprising:

[0010] Obtaining the original sensor signal from the voltage sensor, current sensor, temperature sensor and gas pressure sensor of the household energy storage battery;

[0011] Analog-to-digital conversion and timing synchronization calibration are performed on the original sensor signal to obtain battery operation data.

[0012] In a second implementation form of the first aspect, the battery operation data is substituted into an electrochemical state space model to calculate electrode reaction rate, to obtain exchange current density and transfer coefficient, and the electrochemical state parameters are calculated based on the exchange current density and the transfer coefficient, including:

[0013] The battery operation data is substituted into an electrochemical state space model to calculate electrode reaction rate, to obtain exchange current density and transfer coefficient.

[0014] The lithium ion distribution concentration and the solid phase potential are calculated based on the exchange current density and the transfer coefficient, and the electrochemical reaction heat generation and gas evolution are coupled to calculate the temperature field and the gas evolution pressure.

[0015] The electrochemical state parameters are obtained by combining the lithium ion distribution concentration, the solid phase potential, the temperature field and the gas evolution pressure.

[0016] In a third implementation form of the first aspect, the battery operation data is substituted into an electrochemical state space model to calculate electrode reaction rate, to obtain exchange current density and transfer coefficient, including:

[0017] The voltage data, current data and temperature data in the battery operation data are input into a Butler-Volmer electrode reaction kinetics state equation of the electrochemical state space model for initialization, to obtain an electrochemical reaction state vector.

[0018] The equilibrium potential and the actual potential in the electrochemical state space model are calculated based on the difference of the electrochemical reaction state vector, to obtain an electrode overpotential.

[0019] The electrode overpotential is substituted into a Butler-Volmer equation of the electrochemical state space model for parameter identification, to obtain an electrode reaction kinetics parameter group, and the exchange current density and the transfer coefficient are separated and extracted from the electrode reaction kinetics parameter group.

[0020] In a fourth implementation form of the first aspect, the temperature deviation value, the gas pressure deviation value and the potential deviation value are calculated according to the electrochemical state parameters, and the adaptive noise parameter adjustment is performed based on the temperature deviation value, the gas pressure deviation value and the potential deviation value, to obtain a safety evaluation value, including:

[0021] The temperature deviation value, the gas pressure deviation value and the potential deviation value are calculated according to the electrochemical state parameters.

[0022] The initial evaluation value is obtained by weighted summing of the temperature deviation value, the gas pressure deviation value, and the potential deviation value.

[0023] Based on the initial evaluation values ​​and the electrochemical state parameters, a Sigma point set is constructed, and the noise covariance parameter is obtained by adaptively adjusting the Sigma point set.

[0024] The safety assessment value is calculated based on the Sigma point set and the noise covariance parameter.

[0025] In conjunction with the first aspect, in a fifth implementation of the first aspect of the present invention, the step of constructing a Sigma point set based on the initial evaluation value and the electrochemical state parameters, and adaptively adjusting the noise parameters of the Sigma point set to obtain a noise covariance parameter, includes:

[0026] The initial evaluation value is combined with the electrochemical state parameters to form an extended state vector, and the Sigma point generation parameters are calculated based on the extended state vector.

[0027] Based on the Sigma point generation parameters, the center point of the extended state vector is set and the positive and negative offset points are calculated to obtain the Sigma point set.

[0028] Substitute the Sigma point set into the posterior distribution function of the electrochemical parameters to perform log-likelihood calculation and expectation update to obtain the variational Bayesian update parameters.

[0029] Based on the variational Bayesian update parameters, the covariance matrix of the safety state noise variance and the electrochemical parameter noise variance is reconstructed to obtain the noise covariance parameters.

[0030] In conjunction with the first aspect, in a sixth implementation of the first aspect of the present invention, the step of calculating the security assessment value based on the Sigma point set and the noise covariance parameter includes:

[0031] The measurement residual vector is obtained by calculating the residual vector based on the residual vector of each Sigma point in the Sigma point set and the actual electrochemical measurement data.

[0032] The measurement residual vector is substituted into multiple Gaussian kernel functions of different scales to calculate the kernel value, and then linearly combined according to the preset kernel weight coefficients to obtain the multi-kernel weight coefficients.

[0033] Based on the multi-core weighting coefficients, a negative log-correlation entropy cost function is constructed, and the cost function is minimized by combining the noise covariance parameter to obtain the wave gain coefficient.

[0034] The wave gain coefficient is weighted with the set of Sigma points to update a state to obtain a safety evaluation value.

[0035] In a seventh implementation form of the first aspect, the Gibbs free energy change value and the side reaction rate are calculated according to the safety evaluation value, and an electrochemical thermodynamic equilibrium condition judgment is performed based on the Gibbs free energy change value and the side reaction rate to generate a pre-warning control instruction, including:

[0036] The Gibbs free energy change value is calculated according to the safety evaluation value and a gas precipitation pressure in the electrochemical state parameter;

[0037] The side reaction rate is calculated in combination with a temperature field in the electrochemical state parameter;

[0038] When the Gibbs free energy change value is less than a critical threshold value and the side reaction rate is less than a safety limit value, an equilibrium condition judgment result is output as an equilibrium state signal; when the Gibbs free energy change value is greater than the critical threshold value and the side reaction rate is greater than the safety limit value, the equilibrium condition judgment result is output as an unbalanced state signal;

[0039] A corresponding condition control strategy mode is selected based on the equilibrium condition judgment result to generate a corresponding pre-warning mode switching signal;

[0040] A pre-warning control instruction is generated based on the pre-warning mode switching signal.

[0041] In an eighth implementation form of the first aspect, the pre-warning control instruction is generated based on the pre-warning mode switching signal, including:

[0042] When the pre-warning mode switching signal is the equilibrium state signal, a pre-warning strategy is selected as an adaptive threshold pre-warning mode; when the pre-warning mode switching signal is the unbalanced state signal, the pre-warning strategy is selected as a fixed threshold pre-warning mode;

[0043] The temperature field and the gas precipitation pressure in the electrochemical state parameter are matched with a protection measure based on the safety evaluation value and the pre-warning strategy to generate the pre-warning control instruction.

[0044] In a second aspect, the present application provides a household energy storage battery safety intelligent pre-warning system, including:

[0045] A collection module is configured to collect battery operation data of a household energy storage battery.

[0046] an internal state analysis module, configured to substitute the battery operation data into an electrochemical state space model to calculate electrode reaction rates, to obtain an exchange current density and a transfer coefficient, and to calculate an electrochemical state parameter based on the exchange current density and the transfer coefficient;

[0047] a calculation module, configured to calculate a temperature deviation value, a gas pressure deviation value and a potential deviation value according to the electrochemical state parameter, and to perform adaptive noise parameter adjustment based on the temperature deviation value, the gas pressure deviation value and the potential deviation value to obtain a safety evaluation value;

[0048] a condition judgment module, configured to calculate a Gibbs free energy change value and a side reaction rate according to the safety evaluation value, and to perform electrochemical thermodynamic equilibrium condition judgment based on the Gibbs free energy change value and the side reaction rate to generate a pre-warning control instruction.

[0049] In the technical solution provided by the application, an electrochemical state space model based on Butler-Volmer electrode reaction kinetics and Fick diffusion law is constructed, a safety evaluation value is taken as a state variable to be uniformly modeled, an electrochemical constrained variational Bayesian multi-kernel correlation entropy unscented Kalman filtering algorithm is adopted, parameter adaptive updating is realized through a variational Bayesian strategy, non-Gaussian measurement noise is effectively processed in combination with a multi-kernel correlation entropy criterion, and the state estimation precision under the conditions of battery aging and temperature gradient change is improved. A condition control strategy framework based on an electrochemical thermodynamic equilibrium condition is established, a pre-warning mode is dynamically adjusted according to a Gibbs free energy change and a side reaction rate, a pre-warning decision space is expanded from a monotonous fixed threshold to a non-monotonous adaptive threshold, and the electrochemical parameter time-varying characteristics are effectively adapted. The application realizes non-monotonous adaptive pre-warning decision and hierarchical pre-warning control of multi-parameter electrochemical coupling, all estimation results and control outputs meet the physical constraints of the charge conservation law and the energy conservation law, and the physical rationality, accuracy and robustness of the pre-warning system are ensured.

[0050] Other features and advantages of the present application will become apparent from the following specification, taken in conjunction with the accompanying drawings, wherein like reference characters refer to like elements in the several views. The features and advantages of the present application will become better understood with reference to the detailed description and drawings, wherein:

[0051] To make the above objectives, features and advantages of the present application more obvious and easy to understand, the following preferred embodiments are specifically described below, and the accompanying drawings are referred to for detailed description. BRIEF DESCRIPTION OF DRAWINGS

[0052] Figure 1 FIG. 1 is a schematic diagram of one embodiment of a household energy storage battery safety intelligent pre-warning method in the embodiments of the present application;

[0053] Figure 2An embodiment schematic diagram of a household energy storage battery safety intelligent early warning system in the present application. DETAILED DESCRIPTION

[0054] To make the objectives, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions of the present application will be described below in connection with the drawings. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of protection of the present application.

[0055] The terms "comprising" and "having" and any variations thereof mentioned in the embodiments of the present application are intended to cover the inclusions without limitation. For example, a process, method, system, product or device comprising a series of steps or units is not limited to the listed steps or units, but can optionally further comprise other steps or units not listed, or can optionally further comprise other steps or units inherent to the process, method, product or device.

[0056] To facilitate the understanding of the embodiments, first, a household energy storage battery safety intelligent early warning method disclosed in the embodiments of the present application will be described in detail. As shown in Figure 1 The method comprises the following steps:

[0057] 101. Collecting battery operation data of a household energy storage battery;

[0058] In the embodiments, multiple types of state sensing elements are arranged in the household energy storage battery system, including a voltage sensor for monitoring the voltage change of a battery cell or a battery cluster in real time, a current sensor for obtaining the current flow characteristics in the charging and discharging process, a temperature sensor for sensing the heat accumulation state, and a gas pressure sensor for capturing the pressure change in the sealed battery pack due to the side reaction or abnormal gas emission behavior. The original signals are input into the front-end data acquisition module in an independent channel mode through the battery management system (BMS), and all analog electrical signals are converted into digital signals by an analog-to-digital converter (ADC) with multi-input capability to form standardized digital signal output. The data flow between different channels is time-aligned and synchronized by the clock reference in the system main control unit to eliminate the multi-channel data asynchronous problems caused by sampling time delay, communication delay or sensing response time difference, and to obtain the battery operation data vector marked with the same time stamp.

[0059] 102. Substituting the battery operation data into an electrochemical state space model to calculate the electrode reaction rate, obtaining the exchange current density and transfer coefficient, and calculating the electrochemical state parameters based on the exchange current density and transfer coefficient;

[0060] In this embodiment, the battery operating data is substituted into the electrochemical state space model established on the basis of the Butler-Volmer electrode kinetics equation and the Fick diffusion law, and the reaction rate at the electrode interface is solved according to the reaction kinetics framework therein to obtain the exchange current density representing the charge transfer capability and the transfer coefficient describing the reaction symmetry. With the reaction rate parameter group as input, the lithium ion distribution concentration function reflecting the evolution of the concentration distribution inside the battery is derived by solving the lithium ion migration equation under diffusion control, and the solid-phase potential distribution at each point in space is obtained by simultaneously solving the charge conservation equation, which is used to describe the potential change under the electron conduction path. The charge transfer behavior and the reaction heat release mechanism are physically coupled, the heat generation function is constructed according to the ohmic heating term and the reaction heat term, and the temperature field distribution is solved in space and time by combining the heat conduction equation, and a gas release model based on the side reaction rate is introduced, the gas generation rate is fitted by the temperature sensitive index and the side reaction constant, and the gas volume change is calculated by the state integral to obtain the gas precipitation pressure. The four types of physical quantities of lithium ion distribution concentration, solid-phase potential, temperature field and gas precipitation pressure obtained by analysis are combined in the form of a state vector to form an electrochemical state parameter set. Among them, the calculation of the electrode reaction rate to obtain the exchange current density and the transfer coefficient includes: inputting the voltage V, the current I and the temperature T data into the state space representation of the Butler-Volmer equation. The calculation process is: first, the overpotential is obtained according to the difference between the actual voltage of the battery and the theoretical equilibrium voltage, then the overpotential, the current I and the temperature T are substituted into the Butler-Volmer kinetics equation, and the exchange current density and the transfer coefficient are identified by a nonlinear parameter fitting algorithm. Specifically, the least squares method or the gradient descent method is used to minimize the error between the actual current and the theoretical current calculated by the Butler-Volmer equation, so as to determine the two key electrochemical parameters.

[0061] 103. Calculate temperature deviation value, gas pressure deviation value and potential deviation value according to the electrochemical state parameters, and adjust the adaptive noise parameters based on the temperature deviation value, the gas pressure deviation value and the potential deviation value to obtain a safety evaluation value;

[0062] In this embodiment, the temperature field, gas evolution pressure, solid phase potential and lithium ion distribution concentration and other state parameters are taken as inputs, and the temperature deviation value, gas pressure deviation value and potential deviation value are constructed according to the thermodynamic and electrochemical safety stability index. The temperature deviation value is calculated by the normalized deviation of the current temperature value and the reference safety temperature, the gas pressure deviation value is obtained based on the relative increase of the gas evolution pressure relative to the upper limit of the safety pressure of the sealed cavity, and the potential deviation value is determined by the deviation degree of the electrochemical window derived from the difference between the solid phase potential and the electrolyte potential. The temperature deviation value, gas pressure deviation value and potential deviation value are multiplied by the preset weight coefficient for weighted fusion to obtain the initial evaluation value. Based on the initial evaluation value and the current electrochemical state parameter, a Sigma point set is generated with an unscented Kalman filter structure as the carrier, the Sigma point set includes state mean points and positive and negative direction disturbance points, which are used to represent the distribution characteristics of the system state under uncertain conditions, and on this basis, a variational Bayesian inference mechanism is introduced. The state prediction error and measurement residual under each Sigma point are jointly statistically analyzed to dynamically estimate the state process noise covariance matrix and the measurement noise covariance matrix, and the noise covariance parameters reflecting the current running environment and the state disturbance degree are updated. The updated noise covariance parameters and the Sigma point set are used together for fusion calculation of the expected state trajectory, and the nonlinear correction of the estimation deviation is combined with the multi-core related entropy function to obtain the safety evaluation value output. Wherein, the Sigma point set is constructed and the adaptive noise parameter adjustment is as follows: the initial evaluation value S and the electrochemical state parameter form an extended state vector X, which includes lithium ion concentration, solid phase potential, temperature field, gas pressure and safety evaluation value. According to the dimension n of the extended state vector, the number of Sigma points is 2n+1. Taking the mean value of the extended state vector as the center point, symmetrically distributed Sigma points are generated around it according to the principal component direction of the covariance matrix. The weight of each Sigma point is calculated according to the standard formula of unscented transformation, and then the noise covariance matrix is updated in real time through the variational Bayesian method, so that it can adapt to battery aging and environmental changes. The Sigma point set is substituted into the posterior probability density function of the electrochemical parameter to calculate the log-likelihood value corresponding to each Sigma point. Through the variational inference method, it is assumed that the electrochemical parameter obeys a certain prior distribution, and then the parameters of the posterior distribution are updated according to the observed data. The specific process is to calculate the gradient of the likelihood function with respect to the parameters, and update the mean and variance parameters of the variational distribution, so that the variational distribution approximates the true posterior distribution. The expectation value is updated by Monte Carlo sampling or analytical calculation to obtain the optimal estimation value of the electrochemical parameter. Based on the parameters updated by the variational Bayesian method, the safety state noise variance and the electrochemical parameter noise variance are calculated respectively. These variances are combined to form a block diagonal covariance matrix Q, in which the diagonal blocks correspond to the noise characteristics of different state variables respectively.The reconstruction process takes into account the correlation between state variables, and ensures the positive definiteness of the covariance matrix through Cholesky decomposition or eigenvalue decomposition. The final noise covariance parameters can accurately describe the uncertainty level of the system. Each Sigma point in the Sigma point set is mapped to a predicted measurement value through the electrochemical observation equation, and then compared with the actual sensor measurement data. The calculation process is as follows: for each Sigma point Xi, the predicted measurement value Yi is obtained through the observation function h, and then the residual error ri is calculated, which is equal to the actual measurement value Z minus the predicted measurement value Yi. The residuals of all Sigma points are combined to form a measurement residual vector R, which reflects the degree of deviation between the model prediction and the actual observation.

[0063] 104. Calculate the Gibbs free energy change value and the side reaction rate according to the safety evaluation value, and perform electrochemical thermodynamic equilibrium condition judgment based on the Gibbs free energy change value and the side reaction rate to generate early warning control instructions.

[0064] In this embodiment, the safety evaluation value is combined with the gas evolution pressure in the electrochemical state parameter to be substituted into the free energy change function established based on the electrochemical thermodynamic mechanism, the reaction enthalpy change is estimated through the quantitative relationship between heat release and pressure rise in the gas evolution process, and the reaction entropy term is extracted by combining the temperature distribution and the change of gas production trend to construct the real-time Gibbs free energy change value; at the same time, the temperature field information in the electrochemical state parameter is used to establish a side reaction rate model, the actual reaction rate under the current side reaction path is calculated by means of the exponential influence law of temperature change on the reaction rate and the optimal fitting parameters obtained in the parameter identification module, and the actual reaction rate is compared with the safety limit value in real time. According to the judgment logic, when the Gibbs free energy change value is less than the preset critical threshold and the side reaction rate does not exceed the safety upper limit, it is judged that the current running state of the system is still in the thermodynamic controllable interval, and the output balance condition judgment result is the balance state signal; when the Gibbs free energy change value exceeds the critical threshold accompanied by the acceleration of the side reaction rate, it is judged that the current running state of the system is in the unbalanced state interval, and the unbalanced state signal is output, and on this basis, the control strategy mode of a higher response level is switched to, the adaptive threshold strategy is called in the balanced state to maintain high sensitivity and low false alarm response, and the fixed strict threshold mode is switched to in the unbalanced state to ensure rapid intermittent control response. The judgment result is logically decoded and matched with the corresponding response template in the control strategy library to generate a warning mode switching signal, and the specific warning control instruction is generated according to the downstream warning control module, which includes multiple response mechanisms such as active heat dissipation power adjustment, charge and discharge current limiting, state of charge balancing, emergency power-off switching, etc. According to the multi-core weight coefficient, a cost function J is constructed, which is the negative logarithm of the weighted combination of the values of each kernel function. The measurement residual vector is substituted into multiple Gaussian kernel functions with different bandwidths to calculate the output values of each kernel function, and then the total correlation entropy value is obtained by summing the weight coefficients. The cost function minimization is solved by Newton method or quasi-Newton method, the goal is to find the filter gain K that minimizes the cost function. In the solving process, the noise covariance parameter is combined to constrain the value range of the filter gain, to ensure numerical stability. According to the safety evaluation value S and the gas evolution pressure P, the Gibbs free energy change of the electrochemical reaction is calculated. The calculation process is as follows: first, determine the chemical potential change of the reaction according to the gas evolution pressure, then combine the safety evaluation value to reflect the degree of system deviation from the equilibrium state, and calculate the Gibbs free energy change through the thermodynamic relationship. Specifically, substitute the pressure P into the chemical potential equation, and then modify the reaction progress according to the safety evaluation value to finally obtain the quantitative value of the degree of system deviation from thermodynamic equilibrium. Combine the temperature field T data to calculate the side reaction rate r through the Arrhenius equation. The calculation process is as follows: first, determine the activation energy Ea and the frequency factor A of the side reaction, then substitute the temperature T into the Arrhenius formula r equals A times exp minus Ea divided by RT, where R is the gas constant.The spatial distribution characteristics of the temperature field are reflected in the side reaction rate by weighted average or maximum value selection, and the high temperature area has greater contribution to the side reaction rate, thereby accurately evaluating the safety risk degree of the battery.

[0065] In a specific embodiment, the process of performing step 101 can specifically include the following steps:

[0066] Obtain raw sensor signals from the voltage sensor, current sensor, temperature sensor and gas pressure sensor of the household energy storage battery.

[0067] Analog-to-digital conversion and timing synchronization calibration are performed on the raw sensor signals to obtain battery operation data.

[0068] In this embodiment, a multi-source state perception channel is constructed in the system structure of the household energy storage battery, wherein the voltage sensor is used to collect the terminal voltage information of each cell or battery module to capture the potential fluctuation characteristics in the electrochemical reaction process, the current sensor is deployed on the main loop path to monitor the current direction and amplitude in the charging and discharging process in real time, and reflects the dynamic changes of load fluctuation and working state switching, the temperature sensor is distributed around the cell interior, module shell and heat dissipation structure, and is used to obtain thermal dynamic parameters such as heat accumulation, heat diffusion and temperature rise rate, and the gas pressure sensor is arranged in the sealed battery package shell or near the safety valve channel, and is used to detect the pressure change behavior in the gas precipitation process caused by side reactions, thermal runaway or overcharge and other causes. The raw perception data of these sensors are transmitted to the multi-channel interface circuit of the data acquisition module through the analog signal output mode, and all sensor signals are subjected to bandwidth filtering, anti-interference level isolation and input protection circuit to ensure the stability and anti-interference ability of the analog signal transmission process. The analog electrical signals of each channel are converted into digital quantity representation in real time by a high-precision analog-to-digital conversion unit, the analog-to-digital converter has a precision of 12 bits or more and has an independent sampling and holding function to meet the high-frequency dynamic sampling requirement. Considering the response time difference, analog signal rising delay and channel switching offset of different sensors, the sampling events of all analog-to-digital conversion channels are synchronously controlled by a unified clock source inside the system, a unified trigger sampling instruction is used, a hardware-level sampling synchronization trigger link is established, or a data channel timing compensation logic is configured to realize time reference alignment. After the analog-to-digital conversion is completed, all channel data are labeled with a unified time stamp and packaged to form a structured battery operation data frame.

[0069] In a specific embodiment, the process of performing step 102 can specifically include the following steps:

[0070] Substitute the battery operation data into the electrochemical state space model to calculate the electrode reaction rate to obtain the exchange current density and transfer coefficient.

[0071] The lithium ion distribution concentration and solid phase potential are calculated based on the exchange current density and transfer coefficient, and the electrochemical reaction heat production and gas evolution are coupled to calculate the temperature field and gas evolution pressure;

[0072] The electrochemical state parameters are obtained by combining the lithium ion distribution concentration, the solid phase potential, the temperature field and the gas evolution pressure.

[0073] In this embodiment, the battery operation data is mapped into the boundary conditions and input variables of the electrochemical state space model according to the input requirements, including voltage data, current data, temperature data and gas pressure data, and the electrochemical state space model takes the description of the internal kinetic evolution process of the lithium ion battery as the core, and the basic structure includes the charge transfer path, the ion migration channel, the energy conversion process and the side reaction channel. Through multi-time step input of the operation data and introduction of the state propagation function, the quantitative reconstruction of the electrode reaction process is realized. In this process, the reaction rate on the electrode interface is inversely calculated according to the current and voltage change trend, and the key parameters of the charge transfer capacity, i.e. the exchange current density, are calculated. At the same time, the transfer coefficient representing the directionality of electron migration is derived by fitting the asymmetric behavior of the positive and negative electrode materials in the reaction process. These two parameters together constitute the basic constraint factor of the reaction kinetics. The reaction rate parameter group is taken as the input of the diffusion process to drive the evolution process of the lithium ion concentration field. The mapping relationship between the concentration gradient field and the ion flux is established to calculate the distribution concentration of lithium ions inside the electrode sheet. At the same time, the electron conduction path is introduced to model the spatial distribution state of the solid phase potential, and the potential difference on the electron-ion interface is embedded in the model to reflect the internal resistance and polarization degree. Based on the above charge and ion migration state, the heat production term module in the heat model is called, and the heat source distribution function is established according to the reaction heat, the ohmic heat caused by current conduction and the self-heating process caused by the fluctuation of temperature-dependent parameters. Then, the temperature field distribution in the whole structure is calculated by combining the boundary heat exchange coefficient and the material thermal diffusion characteristics. At the same time, according to the lithium ion concentration and temperature state, the side reaction starting condition is judged. When entering the side reaction sensitive interval, the gas evolution model is called to calculate the gas generation rate per unit time. Combined with the internal finite volume condition of the battery and the gas diffusion rate, the gas pressure evolution value in the closed space is solved. The four types of physical state variables, i.e. the lithium ion distribution concentration, the solid phase potential, the temperature field and the gas evolution pressure, are packaged according to the unified data structure to form the electrochemical state parameter set with time sequence identifier, spatial characteristics and physical dimension marker.

[0074] In a specific embodiment, the process of performing the step of substituting the battery operation data into the electrochemical state space model to calculate the electrode reaction rate to obtain the exchange current density and the transfer coefficient can specifically include the following steps:

[0075] The voltage data, current data, and temperature data in the battery operation data are input into the Butler-Volmer electrode reaction kinetics state equation of the electrochemical state space model for initialization to obtain an electrochemical reaction state vector;

[0076] The equilibrium potential and the actual potential in the electrochemical state space model are differentially calculated based on the electrochemical reaction state vector to obtain an electrode overpotential;

[0077] The electrode overpotential is substituted into the Butler-Volmer equation of the electrochemical state space model for parameter identification to obtain an electrode reaction kinetics parameter group, and the exchange current density and the transfer coefficient are separated and extracted from the electrode reaction kinetics parameter group.

[0078] In this embodiment, input variables are extracted from the battery operation data, including real-time sampled voltage data, current data, and temperature data, and the input variables are input as external observation boundary conditions of the model into the electrochemical state space model for initialization processing. The Butler-Volmer electrode reaction kinetics state equation is used as a core module to describe the rate relationship of charge transfer reaction and the nonlinear potential response mechanism. In the initialization process, the electrode overall potential state is calibrated according to the input voltage data, and the electrode reaction rate and current density are established by combining the current data, and the sensitivity of the reaction rate to the thermal environment change is corrected by using the temperature data, thereby constructing an electrochemical reaction state vector. The state vector internally contains multiple dynamic components describing reaction speed, charge accumulation, potential response, and thermal field coupling. Based on the electrochemical reaction state vector, the electrode equilibrium potential under the current reaction condition is solved through the thermodynamic database or state equation inside the state space model according to the electrode material type and the activity of the substance under the operating temperature, and the electrode actual potential obtained from the voltage data is differentially calculated to obtain the electrode overpotential on the current reaction interface. The electrode overpotential is an important parameter reflecting the degree of deviation of the electrode interface from the thermodynamic equilibrium, and is substituted into the Butler-Volmer nonlinear kinetic expression as the core input variable of the kinetic parameter identification. By constructing the nonlinear mapping relationship between the reaction rate, the overpotential, the current density, and the temperature, and based on the multi-stage sample space constructed based on the historical operation data, the unknown kinetic parameters in the nonlinear equation are iteratively solved by combining the minimum error optimization strategy or the Bayesian inversion method, and the electrode reaction kinetics parameter group including the exchange current density, the transfer coefficient, the activation energy, the charge transfer rate constant, and other contents is obtained. The exchange current density directly related to the electrode interface reaction rate and the transfer coefficient related to the symmetry of the positive and negative electrode reaction are extracted from the electrode reaction kinetics parameter group.

[0079] In a specific embodiment, the process of step 103 can specifically include the following steps:

[0080] calculating a temperature deviation value, a gas pressure deviation value and a potential deviation value according to the electrochemical state parameters;

[0081] weighting and summing the temperature deviation value, the gas pressure deviation value and the potential deviation value to obtain an initial evaluation value;

[0082] constructing a Sigma point set based on the initial evaluation value and the electrochemical state parameters, and performing adaptive noise parameter adjustment on the Sigma point set to obtain a noise covariance parameter;

[0083] calculating a safety evaluation value according to the Sigma point set and the noise covariance parameter.

[0084] In this embodiment, the temperature field, gas precipitation pressure and solid phase potential and other key variables are extracted from the electrochemical state parameters, and three sub-indices reflecting the risks of temperature rise, internal gas accumulation and potential deviation are constructed according to these variables. The temperature deviation value evaluates the thermal runaway trend by the deviation degree of the current temperature value relative to the safe operation temperature, the gas pressure deviation value reflects the internal gas inflation strength according to the relative distance between the current gas pressure and the shell pressure limit, and the potential deviation value reflects the degree of deviation of the current solid phase potential from the electrochemical window boundary for capturing electrode polarization and overcharge risk. The three threat components are respectively assigned to the weight coefficients matched with their risk levels, and the initial evaluation value is calculated by weighted fusion to reflect the overall physical safety state trend of the current battery system. The initial evaluation value and the corresponding electrochemical state parameters are used as the center point to construct a multi-dimensional Sigma point set, wherein each Sigma point represents a hypothetical state distribution point after disturbance in the state space. The Sigma points are generated by the unscented transformation rule and carry corresponding statistical weights. All Sigma points are input into the state propagation function, the prediction error distribution is constructed by combining the historical observation error, and the covariance matrix of the process noise and the measurement noise is dynamically updated by the variational Bayesian inference method to obtain the noise covariance parameter more consistent with the current state fluctuation characteristics. On the basis of fusing the Sigma point propagation results and the noise covariance parameter, the distribution characteristics of the current system in the threat space are jointly estimated, and the safety evaluation value after statistical filtering and Bayesian correction is output.

[0085] In a specific embodiment, the process of constructing a Sigma point set based on the initial evaluation value and the electrochemical state parameters, and performing adaptive noise parameter adjustment on the Sigma point set to obtain a noise covariance parameter can specifically include the following steps:

[0086] combining the initial evaluation value and the electrochemical state parameters to form an extended state vector, and calculating Sigma point generation parameters based on the extended state vector;

[0087] According to the Sigma point generation parameters, a central point setting and positive and negative offset point calculation are performed on the extended state vector to obtain a Sigma point set;

[0088] After substituting the Sigma point set into the electrochemical parameter posterior distribution function, log-likelihood calculation and expectation value updating are performed to obtain the variational Bayesian updating parameter;

[0089] Based on the variational Bayesian updating parameter, the covariance matrix reconstruction is performed on the safety state noise variance and the electrochemical parameter noise variance to obtain the noise covariance parameter.

[0090] In this embodiment, the initial evaluation value and the key state variables output by the electrochemical state space model, including lithium ion concentration distribution, solid phase potential, temperature field, and gas precipitation pressure, are normalized according to a unified scale, and are combined to form an extended state vector with a unified space-time reference. The extended state vector represents the coupling state characteristics of the current battery operating state in the thermal, electrical, and chemical three dimensions in the form of a multi-dimensional joint variable. According to the dimension number of the extended state vector and the state covariance structure, scale parameters required for Sigma point generation are calculated, including weighting factors, balance factors, and offset scale coefficients corresponding to the state dimension number. These parameters are used to control the discrete density and coverage range of the Sigma points in the state space. Taking the extended state vector as the central point, a set of symmetrically distributed Sigma points is calculated by adding and subtracting the positive and negative offset amounts of the covariance matrix vector under the weighting of the scale parameters. Each Sigma point represents a possible sampling point of the current state distribution under different direction perturbations, and is used to simulate the state propagation behavior of the nonlinear system. The Sigma points are substituted into the electrochemical parameter posterior distribution function constructed based on the historical observation sequence one by one. The likelihood value corresponding to each Sigma point is calculated through the log-likelihood function, and the expectation step updating operation in the expectation maximization process is performed based on the likelihood value, so as to derive the variational Bayesian updating parameter under the current state. The updating parameter includes the posterior mean estimation and the covariance structure correction term of the electrochemical parameter, and also includes the statistical characteristics of the state disturbance and the system observation uncertainty. The variational Bayesian updating parameter is used for covariance matrix construction of the threat state noise and the electrochemical model parameter noise. The safety state noise variance reflects the fluctuation degree of the threat index between the prediction and the observation, and the electrochemical parameter noise variance is used to describe the disturbance characteristics of the physical mechanisms such as material properties, reaction rate, and temperature response in the actual environment. By fusing the two variance terms, the covariance matrix structure in the state propagation process is reconstructed to form the noise covariance parameter.

[0091] In a specific embodiment, the process of calculating the safety evaluation value according to the Sigma point set and the noise covariance parameter can specifically include the following steps:

[0092] The residual error vector is calculated based on each Sigma point in the Sigma point set and the actual electrochemical measurement data, and a measurement residual error vector is obtained.

[0093] The measurement residual error vector is substituted into a plurality of Gaussian kernel functions with different scales to calculate kernel values, and linear combination is performed according to a preset kernel weight coefficient to obtain a multi-kernel weight coefficient.

[0094] A negative logarithmic correlation entropy cost function is constructed according to the multi-kernel weight coefficient, and the cost function is minimized to solve a wave gain coefficient in combination with a noise covariance parameter.

[0095] The wave gain coefficient is used for weighted state updating with the Sigma point set to obtain a safety evaluation value.

[0096] In the Sigma point set, the predicted observation of each Sigma point after mapping by the electrochemical state space propagation function is extracted, and the predicted observation is compared with the actual electrochemical measurement data collected by the sensor in real time point by point. The difference vector between the predicted observation and the actual observation corresponding to each Sigma point is solved to construct a measurement residual error vector set containing residual error information of all Sigma points, reflecting the distribution structure and deviation degree of the nonlinear state measurement error of the system at the current time. The measurement residual error vector is input into a plurality of Gaussian kernel functions with different scale parameters, and the Gaussian kernel function is used to describe the fuzziness and local uncertainty of the measurement error. Each kernel function has different response sensitivity to the error according to its scale size, so as to simultaneously capture the local error concentration and global error diffusion trend. On this basis, linear combination is performed according to the preset kernel weight coefficient, the kernel response results under multiple different scales are superimposed to form a unified multi-kernel weight coefficient, and the distribution consistency of the entire measurement error space under multiple resolution scales is described. A negative logarithmic correlation entropy cost function is constructed based on the multi-kernel weight coefficient. The negative logarithmic correlation entropy cost function has the characteristics of low sensitivity to non-Gaussian abnormal values and high stability to multi-modal error structure, effectively inhibiting isolated peak disturbances in the measurement error. A noise covariance parameter is introduced as a regularization term of the cost function, and the wave gain coefficient is solved by minimizing the cost function. The wave gain coefficient considers the uncertainty of state propagation and combines the distribution characteristics and parameter drift behavior of the measurement error in the construction process, and has high adaptability and nonlinear fault tolerance. The calculated wave gain coefficient is applied to the Sigma point state updating process, the state mean value is weighted and adjusted according to the state value of each Sigma point in the prediction stage and the error performance in the observation space, the current state estimation result is corrected and converged, and the iterative generation of the safety evaluation value is simultaneously completed in the updating process.

[0097] In a specific embodiment, the process of performing step 104 can specifically include the following steps:

[0098] According to the safety evaluation value and the gas evolution pressure in the electrochemical state parameter, the Gibbs free energy change value is calculated;

[0099] Combined with the temperature field in the electrochemical state parameter, the side reaction rate is calculated;

[0100] When the Gibbs free energy change value is less than the critical threshold value and the side reaction rate is less than the safety limit value, the output of the balance condition judgment result is the balanced state signal; when the Gibbs free energy change value is greater than the critical threshold value and the side reaction rate is greater than the safety limit value, the output of the balance condition judgment result is the unbalanced state signal;

[0101] Based on the balance condition judgment result, the corresponding condition control strategy mode is selected and the corresponding warning mode switching signal is generated;

[0102] Based on the warning mode switching signal, the warning control instruction is generated.

[0103] In this embodiment, the safety evaluation value is coupled with the gas evolution pressure data in the electrochemical state parameter, the influence of the gas accumulation rate on the stress evolution in the closed cavity is analyzed, the free energy change trend reflecting the reaction direction and the thermodynamic stability is derived, and the Gibbs free energy change value under the current working condition is estimated. The Gibbs free energy change value as an index reflecting whether the system is in a thermodynamic reversible interval can effectively capture the dynamic trend of whether the electrochemical system is close to the irreversible reaction boundary at the micro level. Combined with the temperature field information in the electrochemical state parameter, the distribution range of the high temperature region and the thermal gradient change rate are identified, and through the built-in side reaction rate mapping relationship in the model, the local reaction activation enhancement effect brought by temperature rise is quantified, and the side reaction rate under the current state is calculated, reflecting the activity degree of the side reaction between the battery material and the electrolyte in the non-main reaction path. By comparing the Gibbs free energy change value with its critical threshold value, and comparing the side reaction rate with its safety limit value, when the detection results show that both are within the acceptable range, that is, the free energy change does not exceed the instability point, and the side reaction rate does not exceed the safety upper limit, it is determined that the current is in a thermodynamic equilibrium state, and an equilibrium state signal is output. On the contrary, when both exceed the set threshold value, it is considered that the electrochemical system has entered an irreversible evolution channel or the side reaction is in a sustained acceleration trend, so an unbalanced state signal is output, and an emergency response mechanism is triggered. After obtaining the thermodynamic equilibrium condition judgment result, the control strategy management module calls the corresponding conditional control strategy mode according to the state result, selects the flexible threshold self-adaptive strategy in the equilibrium state to maintain the system operation efficiency and sensitivity, and switches to the rigid safety protection strategy in the unbalanced state to forcibly limit the load behavior and shorten the control response cycle. According to the selected strategy, a warning mode switching signal with control logic meaning is generated, which carries preset response level, control boundary and execution priority parameters, and is parsed, compiled and structured by a warning control instruction generation unit to generate a warning control instruction, which drives multiple safety control execution modules including power limitation, heat dissipation enhancement, active balancing and charge / discharge interruption.

[0104] In a specific embodiment, the process of generating a warning control instruction based on the warning mode switching signal can specifically include the following steps:

[0105] When the warning mode switching signal is the equilibrium state signal, the warning strategy is selected as the adaptive threshold warning mode, and when the warning mode switching signal is the unbalanced state signal, the warning strategy is selected as the fixed threshold warning mode.

[0106] Based on the safety evaluation value and the warning strategy, the temperature field and the gas evolution pressure in the electrochemical state parameter are matched with the protection measures to generate a warning control instruction.

[0107] In the embodiment, when the early warning mode switching signal is a balanced state signal, it indicates that the current battery system is still in a controllable thermodynamic operating interval. At this time, according to the strategy matching logic, the adaptive threshold early warning mode with the ability of sensitive adjustment and dynamic adaptability is selected. The safety threshold in the adaptive threshold early warning mode is adjusted in real time according to the temperature field evolution rate, gas evolution pressure change trend, and historical characteristics of cycle number and state of charge in the electrochemical state parameter, allowing the early warning tolerance to be automatically increased when the system state fluctuation amplitude is small but within the normal range, so as to avoid unnecessary false alarms or excessive protection behavior. When the early warning mode switching signal is an unbalanced state signal, it indicates that the battery system has obvious thermodynamic imbalance signs, accompanied by rising side reaction rate or abnormal heat diffusion. At this time, the system switches to the fixed threshold early warning mode. This mode is based on the steady-state limit threshold and no longer allows the boundary to be relaxed based on the system adaptive logic, but through the forced execution of the conservative threshold limiting mechanism, it ensures that the early warning trigger condition is more stringent when there is a serious risk and the safety redundancy is prioritized. The current safety evaluation value and the selected early warning strategy are input as parameter groups into the early warning decision logic module, and the temperature field and gas evolution pressure in the electrochemical state parameter are jointly analyzed. When the adaptive threshold mode is used, the temperature early warning trigger line is adjusted according to the local temperature rise speed, maximum temperature gradient, and historical temperature fluctuation interval in the temperature field, and the gas pressure response bandwidth is dynamically set according to the trend and rate of the gas evolution pressure deviation from the normal working pressure, so as to match the appropriate protection measures, such as adjusting the heat dissipation power, optimizing the charging rate, or balancing the SOC deviation, to achieve mild intervention. In the fixed threshold mode, the response boundaries of temperature and pressure are set as non-adjustable values. If the current state breaks the boundary threshold, the forced control measures are immediately matched, including emergency power-off, charge cutoff, forced cooling, or battery cluster isolation operation. The instruction is structured and packaged into a pre-warning control instruction, which is sent to the execution layer safety subsystem through the control bus.

[0108] The above describes the household energy storage battery safety intelligent early warning method in the embodiment of the application. The household energy storage battery safety intelligent early warning system in the embodiment of the application is described below. Please refer to Figure 2 An embodiment of the household energy storage battery safety intelligent early warning system in the embodiment of the application includes:

[0109] The acquisition module 201 is configured to acquire the battery operation data of the household energy storage battery.

[0110] The internal state analysis module 202 is configured to substitute the battery operation data into the electrochemical state space model to calculate the electrode reaction rate, obtain the exchange current density and transfer coefficient, and calculate the electrochemical state parameter based on the exchange current density and transfer coefficient.

[0111] The computing module 203 is configured to calculate a temperature deviation value, a gas pressure deviation value and a potential deviation value according to the electrochemical state parameters, and to adjust the adaptive noise parameters based on the temperature deviation value, the gas pressure deviation value and the potential deviation value, so as to obtain a safety evaluation value;

[0112] The condition judging module 204 is configured to calculate a Gibbs free energy change value and a side reaction rate according to the safety evaluation value, and to execute electrochemical thermodynamic equilibrium condition judgment based on the Gibbs free energy change value and the side reaction rate, so as to generate a pre-warning control instruction.

[0113] Through the cooperation of the above-mentioned various components, by constructing an electrochemical state space model based on the Butler-Volmer electrode reaction kinetics and the Fick diffusion law, the safety evaluation value is used as a state variable to unify modeling, the electrochemical constrained variational Bayesian multi-kernel correlation entropy unscented Kalman filtering algorithm is adopted, the parameter adaptive update is realized through the variational Bayesian strategy, the multi-kernel correlation entropy criterion is combined to effectively process the non-Gaussian measurement noise, and the state estimation accuracy under the conditions of battery aging and temperature gradient change is improved. A condition control strategy framework based on the electrochemical thermodynamic equilibrium condition is established, the pre-warning mode is dynamically adjusted according to the Gibbs free energy change and the side reaction rate, the pre-warning decision space is expanded from a single fixed threshold to a non-monotonic adaptive threshold, and the electrochemical parameter time-varying characteristics are effectively adapted. The present application realizes the non-monotonic adaptive pre-warning decision and hierarchical pre-warning control of the multi-parameter electrochemical coupling, all the estimation results and control outputs meet the physical constraints of the charge conservation law and the energy conservation law, and the accuracy and robustness of the energy storage battery safety pre-warning are ensured.

[0114] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-mentioned system, system and unit can refer to the corresponding processes in the foregoing method embodiments, which will not be repeated here.

[0115] The integrated unit, if realized in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or the whole or part of the technical solutions can be embodied in the form of a software product, which is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in the embodiments of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various program code storage media.

[0116] The above-described embodiments are only used to illustrate the technical solutions of the present application, but not to limit the present application; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those skilled in the art that the technical solutions recorded in the foregoing embodiments can be modified, or some technical features thereof can be replaced by equivalent technical features; and these modifications or replacements 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 the present application.

Claims

1. A household energy storage battery safety intelligent early warning method, characterized in that, The method comprises the following steps: collecting battery operation data of a household energy storage battery; substituting the battery operation data into an electrochemical state space model to calculate electrode reaction rate, to obtain exchange current density and transfer coefficient, and to calculate electrochemical state parameters based on the exchange current density and the transfer coefficient; calculating temperature deviation value, gas pressure deviation value and potential deviation value according to the electrochemical state parameters, and performing adaptive noise parameter adjustment based on the temperature deviation value, the gas pressure deviation value and the potential deviation value to obtain a safety evaluation value; calculating Gibbs free energy change value and side reaction rate according to the safety evaluation value, and performing electrochemical thermodynamic equilibrium condition judgment based on the Gibbs free energy change value and the side reaction rate to generate a warning control instruction; specifically, substituting the safety evaluation value and the gas precipitation pressure in the electrochemical state parameters into a free energy change function established based on the electrochemical thermodynamic mechanism, estimating the reaction enthalpy change through the quantitative relationship between heat release and pressure rise in the gas precipitation process, and combining the temperature distribution and the gas production trend change to extract the reaction entropy term, to construct the real-time Gibbs free energy change value; at the same time, a side reaction rate model is established by using the temperature field information in the electrochemical state parameters, the actual reaction rate under the current side reaction path is calculated by means of the exponential influence law of temperature change on reaction rate and the optimal fitting parameters obtained in the parameter identification module.

2. The method of claim 1, wherein the method further comprises: The method for collecting battery operation data of a household energy storage battery comprises the following steps: obtaining original sensor signals from voltage sensors, current sensors, temperature sensors and gas pressure sensors of the household energy storage battery; performing analog-digital conversion and time sequence synchronization calibration on the original sensor signals to obtain battery operation data.

3. The method of claim 1, wherein the method further comprises: The method for substituting the battery operation data into an electrochemical state space model to calculate electrode reaction rate, to obtain exchange current density and transfer coefficient, and to calculate electrochemical state parameters based on the exchange current density and the transfer coefficient comprises the following steps: substituting the battery operation data into an electrochemical state space model to calculate electrode reaction rate, to obtain exchange current density and transfer coefficient; calculating lithium ion distribution concentration and solid phase potential based on the exchange current density and the transfer coefficient, and coupling calculation of electrochemical reaction heat generation and gas precipitation to obtain temperature field and gas precipitation pressure; combining the lithium ion distribution concentration, the solid phase potential, the temperature field and the gas precipitation pressure to obtain electrochemical state parameters.

4. The safety intelligent early warning method for household energy storage battery according to claim 3, characterized in that, The method for substituting the battery operation data into an electrochemical state space model to calculate electrode reaction rate, to obtain exchange current density and transfer coefficient comprises the following steps: inputting voltage data, current data and temperature data in the battery operation data into a Butler-Volmer electrode reaction kinetics state equation of the electrochemical state space model for initialization to obtain an electrochemical reaction state vector; performing difference calculation on the equilibrium potential and the actual potential in the electrochemical state space model based on the electrochemical reaction state vector to obtain electrode overpotential; The electrode overpotential is substituted into the Butler-Volmer equation of the electrochemical state space model to identify parameters, to obtain an electrode reaction kinetics parameter group, and to separate and extract an exchange current density and a transfer coefficient from the electrode reaction kinetics parameter group.

5. The method of claim 1, wherein the method further comprises: The temperature deviation value, the gas pressure deviation value and the potential deviation value are calculated according to the electrochemical state parameters, and adaptive noise parameter adjustment is performed based on the temperature deviation value, the gas pressure deviation value and the potential deviation value to obtain a safety evaluation value, including: A temperature deviation value, a gas pressure deviation value and a potential deviation value are calculated according to the electrochemical state parameters; The temperature deviation value, the gas pressure deviation value and the potential deviation value are weighted and summed to obtain an initial evaluation value; A Sigma point set is constructed based on the initial evaluation value and the electrochemical state parameters, and adaptive noise parameter adjustment is performed on the Sigma point set to obtain a noise covariance parameter; A safety evaluation value is calculated according to the Sigma point set and the noise covariance parameter.

6. The method of claim 5, wherein the method further comprises: The Sigma point set is constructed based on the initial evaluation value and the electrochemical state parameters, and adaptive noise parameter adjustment is performed on the Sigma point set to obtain a noise covariance parameter, including: The initial evaluation value is combined with the electrochemical state parameters to form an extended state vector, and Sigma point generation parameters are calculated based on the extended state vector; Center points are set and positive and negative offset points are calculated for the extended state vector according to the Sigma point generation parameters to obtain a Sigma point set; Variational Bayesian update parameters are obtained by substituting the Sigma point set into an electrochemical parameter posterior distribution function for log-likelihood calculation and expected value updating. Covariance matrix reconstruction is performed on the safety state noise variance and the electrochemical parameter noise variance based on the variational Bayesian update parameters to obtain a noise covariance parameter.

7. The method of claim 6, wherein the method further comprises: The safety evaluation value is calculated according to the Sigma point set and the noise covariance parameter, including: Residual vectors are calculated based on each Sigma point in the Sigma point set and actual electrochemical measurement data to obtain a measurement residual vector; Multiple kernel weight coefficients are obtained by substituting the measurement residual vector into multiple Gaussian kernel functions of different scales for kernel value calculation and linear combination according to a preset kernel weight coefficient. A negative log correlation entropy cost function is constructed according to the multiple kernel weight coefficients, and cost function minimization is solved in combination with the noise covariance parameter to obtain a wave gain coefficient; The safety evaluation value is obtained by weighting state updating of the wave gain coefficient and the Sigma point set.

8. The method of claim 1, wherein the method further comprises: The Gibbs free energy change value and the side reaction rate are calculated according to the safety evaluation value, and electrochemical thermodynamic equilibrium condition judgment is performed based on the Gibbs free energy change value and the side reaction rate to generate a warning control instruction, including: The Gibbs free energy change value is calculated according to the safety evaluation value and the gas precipitation pressure in the electrochemical state parameters; The side reaction rate is calculated in combination with the temperature field in the electrochemical state parameters; When the Gibbs free energy change value is less than a critical threshold value and the side reaction rate is less than a safety limit value, output a balance condition judgment result as a balanced state signal; when the Gibbs free energy change value is greater than the critical threshold value and the side reaction rate is greater than the safety limit value, output a balance condition judgment result as an unbalanced state signal; Select a corresponding condition control strategy mode based on the balance condition judgment result and generate a corresponding early warning mode switching signal; Generate a pre-warning control instruction based on the early warning mode switching signal.

9. The method of claim 8, wherein the method further comprises: The generation of the pre-warning control instruction based on the early warning mode switching signal includes: When the early warning mode switching signal is a balanced state signal, select a pre-warning strategy as an adaptive threshold early warning mode; when the early warning mode switching signal is an unbalanced state signal, select a pre-warning strategy as a fixed threshold early warning mode; Based on the safety evaluation value and the pre-warning strategy, match protection measures for the temperature field, gas precipitation pressure in the electrochemical state parameter, and generate a pre-warning control instruction.

10. A safety intelligent early warning system for household energy storage battery, characterized in that, A household energy storage battery safety intelligent early warning method for executing any one of claims 1-9, comprising: A collection module for collecting battery operation data of a household energy storage battery; An internal state analysis module for substituting the battery operation data into an electrochemical state space model to calculate electrode reaction rate, obtain exchange current density and transfer coefficient, and calculate electrochemical state parameters based on the exchange current density and the transfer coefficient; A calculation module for calculating temperature deviation value, gas pressure deviation value and potential deviation value according to the electrochemical state parameters, and performing adaptive noise parameter adjustment based on the temperature deviation value, the gas pressure deviation value and the potential deviation value to obtain a safety evaluation value; A condition judgment module for calculating Gibbs free energy change value and side reaction rate according to the safety evaluation value, and performing electrochemical thermodynamic balance condition judgment based on the Gibbs free energy change value and the side reaction rate to generate a pre-warning control instruction.

Images