A battery pack core temperature and thermal parameter joint estimation method based on sparse surface temperature measurement

Abstract

The present application relates to the technical field of battery management, and particularly relates to a battery pack core temperature and thermal parameter joint estimation method based on sparse surface temperature measurement, comprising the following steps: S101. Establishing a lithium battery thermal mathematical model; S102. Establishing an observation equation, determining the installation position of a temperature sensor and installing the temperature sensor; S103. Data acquisition, acquiring the current and temperature sensor data of the battery at the installation position of the temperature sensor, and recording the environmental temperature; S104. Establishing an extended random discrete state space model; S105. Parameter-state joint estimation algorithm, jointly estimating the core temperature and parameters of each battery of the battery pack. The present method significantly reduces the dependence on hardware temperature sensors, reduces the material cost and system integration difficulty, solves the problems of high cost, complex wiring and reduced reliability caused by the fact that the prior art needs to equip each battery in the battery pack with a temperature sensor, and has excellent engineering practicability and deployability.

Description

Technical Field

[0001] This invention relates to the field of battery management technology, and more specifically, to a method for jointly estimating the core temperature and thermal parameters of a battery pack based on sparse surface temperature measurement. Background Technology

[0002] Lithium-ion batteries are the core energy storage components of electric and hybrid vehicles, and their operating temperature directly determines the vehicle's performance, range, lifespan, and safety. During charging and discharging, a large amount of heat is generated inside the battery, causing its core temperature to be much higher than the easily measurable surface temperature. Because the temperature in the battery core area is closer to the critical point of thermal runaway, failure to accurately monitor it will pose serious safety hazards. Therefore, real-time and accurate estimation of the core temperature inside the battery pack is crucial for the design and control of the battery thermal management system and is a key technology for ensuring safe battery operation and extending battery lifespan.

[0003] However, the industry still faces some key technical bottlenecks in the accurate and low-cost estimation of multiple states (temperature, parameters) within battery packs containing multiple battery cells. These bottlenecks mainly include two aspects: First, in order to obtain sufficient information, traditional model-based estimation methods usually require the installation of temperature sensors on the surface of each battery, resulting in high hardware costs, complex wiring harnesses, and reduced system reliability, making it difficult to deploy on a large scale in commercial applications. Second, the thermophysical parameters of the battery (such as thermal resistance and thermal capacity) are not constant and will drift with battery aging, state of charge (SOC), and changes in operating conditions, leading to a decrease in the accuracy of fixed parameter models. Most existing methods perform parameter identification and temperature estimation in separate steps, which can easily lead to the accumulation of temperature estimation errors.

[0004] Currently, various methods have been proposed for estimating the core temperature of batteries. The mainstream methods generally employ a two-step approach, typically involving separate steps: thermal model parameter identification and core temperature estimation. For example, for a single battery, the least squares method is used to identify its thermal model parameters, and then a Kalman filter is constructed based on these parameters to estimate the core temperature. However, an inherent drawback of this two-step core temperature estimation method is the unidirectional propagation and accumulation of errors. Errors generated in the first step of thermal parameter identification will inevitably propagate and amplify into the second step of temperature estimation, thus reducing the long-term accuracy and reliability of the estimation results. Another approach is to derive the thermal mathematical model of the battery pack, estimate the battery pack's thermal parameters using the least squares method, and construct a Kalman filter based on the parameter identification results to achieve accurate estimation of the battery pack's core temperature. However, this method requires a temperature sensor to be placed on the surface of each battery within the battery pack.

[0005] Existing methods generally suffer from the following drawbacks in multi-cell battery pack scenarios: First, they are highly dependent on the deployment of full-scale temperature sensors, which requires temperature measurement points on the surface of each or most of the cells, which is impractical in actual battery packs consisting of hundreds or thousands of cells; Second, parameter identification and core temperature estimation for battery packs have not been separated, leading to error accumulation and reducing the accuracy and long-term robustness of the estimation.

[0006] Therefore, how to break the framework of the traditional two-step method and achieve high-precision, synchronous estimation of the core temperature and key thermal parameters of all batteries in the battery pack under the constraint of an extremely limited number of temperature sensors (i.e., sparse arrangement) is a technical problem that urgently needs to be solved in the field of battery thermal management. Summary of the Invention

[0007] This invention addresses the technical problems of existing battery pack core temperature estimation methods, such as excessive deployment of temperature sensors, high costs, and error accumulation caused by the separation of thermal parameter identification and temperature estimation processes. It proposes a joint estimation method for battery pack core temperature and thermal parameters based on sparse surface temperature measurement.

[0008] The objective of this invention is achieved through the following technical solution:

[0009] A method for jointly estimating the core temperature and thermal parameters of a battery pack based on sparse surface temperature measurement includes the following steps:

[0010] S101. Establish a thermal mathematical model for lithium batteries;

[0011] S102. Establish the observation equation, determine the installation location of the temperature sensor, and install it;

[0012] S103. Data acquisition: Obtain the current of the battery at the temperature sensor installation location and the temperature sensor data, and record the ambient temperature;

[0013] S104. Establish an extended stochastic discrete state-space model;

[0014] S105. Parameter-state joint estimation algorithm, which jointly estimates the core temperature and parameters of each cell in the battery pack.

[0015] Furthermore, in step S101, the mathematical model is:

[0016]

[0017] in, It is the core temperature of the i-th battery. It is the surface temperature of the i-th battery. It is the ambient temperature. It is the heat capacity of the gel roll inside the battery. It is the thermal resistance between the core and the surface for heat exchange. It is the heat capacity of the battery casing. It is the thermal resistance of the surface during heat exchange with the air. It is the thermal resistance between batteries. It is the battery's internal resistance. It is the current flowing through each battery (the batteries are connected in series, so the current in each battery is equal). It is the energy received by the i-th battery from its neighboring batteries. It is the heat generated by the battery core. This refers to the number of batteries. Formula (1) can be transformed into a continuous-time state-space equation:

[0018]

[0019] in,

[0020]

[0021]

[0022]

[0023]

[0024]

[0025] in, The system matrix consists of the thermal parameters of the battery, describing how the current temperature state of the system evolves to the temperature state at the next moment. A1, A2, A3, A4, and A5 are elements in matrix A, which are also composed of battery thermal parameters. The input matrix consists of the physical parameters of the battery and describes how external inputs affect the changes in the system state. This is a state vector, which consists of the temperature of each battery. and constitute; The input vector is composed of external excitations (current and ambient temperature); in formula (4), and The quantity is .

[0026] Furthermore, in step S102, the observation equation is expressed as:

[0027]

[0028] Where y is the number of rows formed by the actual temperature sensor that can be measured. An array with 1 column, where This represents the total number of temperature sensors installed. The battery number containing the temperature sensor in the battery pack is recorded as follows: ,but For example, if four temperature sensors are placed on the surface of batteries 1, 2, 4, and 7 in a battery pack containing n batteries, then... .

[0029] The output matrix is ​​also determined by the location of the temperature sensor; the matrix It is OK A matrix of columns. The i-th row of the matrix represents the temperature sensor number, which is installed at the position numbered... On the surface of the battery. The construction rule for the i-th row is: the element in the i-th row and j-th column is... When column index hour, For indexes on other columns , For example, if four temperature sensors are placed on the surface of batteries 1, 2, 4, and 7 in a battery pack containing seven batteries, then...

[0030]

[0031] Furthermore, in step S102, the number and deployment locations of temperature sensors are determined through observability condition analysis. First, an observability matrix OB is constructed:

[0032]

[0033] in, It is the order of the system, i.e., the number of batteries. twice that of the matrix if and only if the matrix The rank is equal to Only when the model can be fully observed can the core and surface temperatures of each battery in the battery pack be accurately estimated. From formula (9), it can be seen that since OB is determined by H, its rank is related to the number and placement of the temperature sensors. By calculating the rank of the observability matrix OB corresponding to different combinations of temperature sensors with different numbers and placements, the rank of the observability matrix OB is chosen to be equal to... The combination with the fewest temperature sensors is used to determine the number and installation location of temperature sensors.

[0034] Furthermore, in step S104, the method for establishing the extended stochastic discrete state-space model is as follows:

[0035] The continuous-time differential equation (2) is transformed into a discrete-time difference equation, and the thermal parameters to be identified are added to form an extended stochastic discrete state-space model:

[0036] (10)

[0037] (11)

[0038] (12)

[0039] (13)

[0040] Where k is a discrete time point, and generally the smaller the discrete time interval, the more accurate the discrete model; It is a parameter vector composed of the five battery parameters to be identified; A d Let A be an extended matrix of the system matrix A, where It is the identity matrix, and 0 is the zero matrix. The bottom right corner of the 0 matrix indicates its row and column numbers; B d H is the extended matrix of the input matrix B; a This is an extension matrix of the output matrix H; This represents the extended state vector at the k-th discrete time step, which is composed of the state vector... and parameter vector It is pieced together; This represents the output vector at the k-th discrete time step. It is the system noise at the k-th discrete time, characterizing the imperfections of the battery equivalent circuit model and the random fluctuations of the internal electrochemical reaction; The measurement noise at the k-th discrete time step represents the measurement error and sampling quantization error of each temperature sensor; system noise. and measuring noise Assuming all noise is zero-mean white Gaussian noise, their covariances are respectively and , and The value is an adjustable hyperparameter.

[0041] Furthermore, in step S105, the quantity to be estimated is first initialized:

[0042] First, define an extended state vector containing temperature state estimates and thermal parameter estimates. , and The internal composition is the same, and the mathematical definition is:

[0043] (14)

[0044] in, and These are the core temperature and surface temperature of the i-th battery estimated by this method, respectively. These are the battery thermal parameters estimated by this method. During initialization, , To set a specific value, it can be roughly estimated in advance through experience, measurement, or numerical calculation.

[0045] Furthermore, in step S105, after initializing the quantity to be estimated, the core temperature and parameters are iteratively updated, and the iteration termination condition is judged and the quantity is output.

[0046] The process of updating the core temperature and parameters requires iteratively executing formulas (15) to (18). Formula (15) is for calculating the Jacobian matrix, formula (16) is for calculating the gain matrix, formula (17) is for updating the covariance matrix, and formula (18) is for updating the state (temperature) and thermal parameters.

[0047] (15)

[0048] (16)

[0049] (17)

[0050] (18)

[0051] in, It is the Jacobian matrix at the k-th discrete time, which is the intermediate computational product necessary for estimating battery temperature and thermal parameters; It is the gain matrix at the (k+1)th discrete time, which is also a necessary intermediate calculation product for estimating battery temperature and thermal parameters; It is the covariance matrix of the estimation error at the k-th discrete time. It is not only a necessary intermediate calculation product for estimating battery temperature and thermal parameters, but can also be used to determine whether the algorithm has converged. It is the temperature data measured by the battery temperature sensor at the (k+1)th discrete time.

[0052] Furthermore, the iteration termination condition judgment and estimate output process in step S105 is as follows: when the absolute value of the residual between the estimated battery temperature and the corresponding battery temperature measured by the temperature sensor is less than ε in D consecutive iterations, or If the trace of the matrix is ​​less than ε in D consecutive iterations, the algorithm is considered converged, the iteration stops, and the estimated core temperatures and thermal parameters of each battery are determined from... It was extracted from [the source].

[0053] Compared with the prior art, the beneficial effects of the present invention are:

[0054] The method for jointly estimating the core temperature and thermal parameters of a battery pack based on sparse surface temperature measurement in this invention demonstrates and implements a sparse sensing scheme that accurately reconstructs the core temperature distribution of the entire battery pack by installing temperature sensors in only a few key locations through system observability analysis of the battery pack thermal model. This method significantly reduces the dependence on hardware temperature sensors, reduces material costs and system integration difficulty, and solves the problems of high cost, complex wiring, and reduced reliability caused by equipping every cell in the battery pack with a temperature sensor in existing technologies. It has excellent engineering practicality and deployability.

[0055] The parameter-state joint estimation algorithm proposed in this invention constructs an augmented state vector by combining the core temperature to be estimated and key thermal parameters, achieving synchronous online joint estimation of both. By performing real-time correction and updating of the state and parameters within a unified filtering framework, it effectively avoids error accumulation caused by separate identification and can dynamically track parameter changes, thereby significantly improving the robustness and long-term accuracy of the estimation model. This solves the error propagation and accumulation problem caused by the two-step method in existing technologies that first identify thermal parameters and then estimate the core temperature. Attached Figure Description

[0056] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0057] Figure 1 This is a flowchart of the method for jointly estimating the core temperature and thermal parameters of a battery pack based on sparse surface temperature measurement in this invention.

[0058] Figure 2 This is a flowchart of the specific method of step S105, parameter-state joint estimation algorithm in this invention.

[0059] Figure 3 This is a schematic diagram of the thermal model of a battery pack containing seven parallel cylindrical lithium battery cells in this embodiment.

[0060] Figure 4 This is a schematic diagram showing the installation location of the temperature sensor in this embodiment.

[0061] Figure 5 This is a comparison chart of the temperature estimation results and actual measurement results of battery 3 in the embodiment.

[0062] Figure 6 This is a comparison chart of the temperature estimation results and actual measurement results of battery 4 in the embodiment.

[0063] Figure 7 This is a comparison chart of the temperature estimation result and the actual measurement result of battery 7 in the embodiment. Detailed Implementation

[0064] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0065] Example:

[0066] This invention proposes a method for jointly estimating the core temperature and thermal parameters of a battery pack based on sparse surface temperature measurement. The implementation process is as follows: Figure 1 As shown, it includes the following steps:

[0067] S101. Establish a thermal mathematical model for lithium batteries:

[0068] The implementation process of this invention is illustrated using a battery pack containing seven parallel cylindrical lithium batteries as an example. The seven parallel cylindrical lithium batteries are numbered sequentially. Their thermal model is as follows: Figure 3 As shown, the mathematical model is expressed as:

[0069] (1)

[0070] in, It is the core temperature of the i-th battery. It is the surface temperature of the i-th battery. It is the ambient temperature. It is the heat capacity of the gel roll inside the battery. It is the thermal resistance between the core and the surface for heat exchange. It is the heat capacity of the battery casing. It is the thermal resistance of the surface during heat exchange with the air. It is the thermal resistance between batteries. It is the battery's internal resistance. It is the current flowing through each battery (the batteries are connected in series, so the current in each battery is equal). It is the energy received by the i-th battery from its neighboring batteries. It is the heat generated by the battery core. This refers to the number of batteries. Formula (1) can be transformed into a continuous-time state-space equation:

[0071] (2)

[0072] in,

[0073] (3)

[0074] (4)

[0075] (5)

[0076] (6)

[0077] (7)

[0078] in, The system matrix consists of the thermal parameters of the battery, describing how the current temperature state of the system evolves to the temperature state at the next moment. A1, A2, A3, A4, and A5 are elements in matrix A, which are also composed of battery thermal parameters. The input matrix consists of the physical parameters of the battery and describes how external inputs affect the changes in the system state. This is a state vector, which consists of the temperature of each battery. and constitute; The input vector is composed of external excitations (current and ambient temperature); in formula (4), and The quantity is .

[0079] S102. Establish the observation equation:

[0080] The observation equation can be expressed as:

[0081] (8)

[0082] Where y is the number of rows formed by the actual temperature sensor that can be measured. An array with 1 column, where This represents the total number of temperature sensors installed. The battery number containing the temperature sensor in the battery pack is recorded as follows: ,but For example, if four temperature sensors are placed on the surface of batteries 1, 2, 4, and 7 in a battery pack containing n batteries, then... .

[0083] The output matrix is ​​also determined by the location of the temperature sensor; the matrix It is OK A matrix of columns. The i-th row of the matrix represents the temperature sensor number, which is installed at the position numbered... On the surface of the battery. The construction rule for the i-th row is: the element in the i-th row and j-th column is... When column index hour, For indexes on other columns , .

[0084] Observability analysis can be used to optimize the deployment of temperature sensors in specific cells of the battery pack. The observability of the model can be verified by constructing its observability matrix.

[0085] (9)

[0086] in, It is the order of the system, i.e., the number of batteries. Twice that. If and only if the matrix The rank is equal to Only when the model is fully observable can the core and surface temperatures of each cell in the battery pack be accurately estimated. By calculating the rank of the observability matrix OB for different combinations of the number and placement of temperature sensors, it was found that the minimum condition for estimating the core temperatures of all cells is met when the temperature sensors are placed on the surfaces of cells 1, 2, 5, and 7. Figure 4 As shown, it was ultimately determined that temperature sensors would be deployed on the surfaces of batteries 1, 2, 5, and 7, and the output matrix would be... for:

[0087] (10)

[0088] S103. Data Acquisition:

[0089] Acquire the current and casing temperature of lithium-ion batteries No. 1, 2, 5, and 7, and simultaneously record the ambient temperature and battery current. .

[0090] S104. Establish an extended stochastic discrete state-space model:

[0091] The continuous-time differential equation (2) is transformed into a discrete-time difference equation, and the thermal parameters to be identified are added to form an extended stochastic discrete state-space model:

[0092] (11)

[0093] (12)

[0094] (13)

[0095] (14)

[0096] Where k is a discrete time point, and generally the smaller the discrete time interval, the more accurate the discrete model; It is a parameter vector composed of the five battery parameters to be identified; A d Let A be an extended matrix of the system matrix A, where It is the identity matrix, and 0 is the zero matrix. The bottom right corner of the 0 matrix indicates its row and column numbers; B d H is the extended matrix of the input matrix B; a This is an extension matrix of the output matrix H; This represents the extended state vector at the k-th discrete time step, which is composed of the state vector... and parameter vector It is pieced together; This represents the output vector at the k-th discrete time step. It is the system noise at the k-th discrete time, characterizing the imperfections of the battery equivalent circuit model and the random fluctuations of the internal electrochemical reaction; The measurement noise at the k-th discrete time step represents the measurement error and sampling quantization error of each temperature sensor; system noise. and measuring noise Assuming all noise is zero-mean white Gaussian noise, their covariances are respectively and The covariance value is an adjustable hyperparameter. In this embodiment, It has 19 rows and 19 columns, with 1e as the diagonal element. -8 A diagonal matrix where all other elements are 0. It is a diagonal matrix with 4 rows and 4 columns, with the diagonal element being 0.001 and the other elements being 0.

[0097] S105. Parameter-State Joint Estimation Algorithm:

[0098] The core temperature and parameters of each cell in a lithium-ion battery pack are jointly estimated.

[0099] like Figure 2 As shown, the estimators are first initialized. An extended state vector containing temperature state estimates and thermal parameter estimates is defined first. The mathematical definition is:

[0100] (15)

[0101] This embodiment is designed When the initial value is, Battery thermal parameters The initial error is determined by numerical calculation and is set to deviate from the true value by 20%, 20%, 100%, 30%, and 50%, respectively.

[0102] Finish After initialization, the core temperature and parameters are iteratively updated, and formulas (16)-(19) are executed iteratively. Formula (16) is for calculating the Jacobian matrix, formula (17) is for calculating the gain matrix, formula (18) is for updating the covariance matrix, and formula (19) is for updating the state (temperature) and thermal parameters.

[0103] (16)

[0104] (17)

[0105] (18)

[0106] (19)

[0107] in, It is the Jacobian matrix at the k-th discrete time, which is the intermediate computational product necessary for estimating battery temperature and thermal parameters; It is the gain matrix at the (k+1)th discrete time, which is also a necessary intermediate calculation product for estimating battery temperature and thermal parameters; It is the covariance matrix of the estimation error at the k-th discrete time. It is not only a necessary intermediate calculation product for estimating battery temperature and thermal parameters, but can also be used to determine whether the algorithm has converged. It is the temperature data measured by the battery temperature sensor at the (k+1)th discrete time.

[0108] In this embodiment, if If the trace of the matrix is ​​less than 0.001 in 10 consecutive iterations, the algorithm is considered converged, the iteration stops, and the estimated core temperatures and thermal parameters of each battery are determined from... It was extracted from [the source].

[0109] The errors between the estimated and actual measured values ​​of the surface and core temperatures of each battery in this embodiment are shown in the table below:

[0110]

[0111] Figures 5-7 The table shows line graphs depicting the estimated and actual measured surface and core temperatures of batteries 3, 4, and 7 over time. Based on the above table and... Figure 5-7 As shown, the difference between the surface temperature and core temperature of the battery estimated using the method of the present invention and the actual measured values ​​of the temperature sensors deployed on each battery is extremely small.

[0112] The errors between the estimated and actual measured values ​​of the thermal parameters of the battery pack in this embodiment are shown in the table below:

[0113]

[0114] The method for jointly estimating the core temperature and thermal parameters of a battery pack based on sparse surface temperature measurement in this invention demonstrates and implements a sparse sensing scheme by installing temperature sensors only in a few key locations—for example, in this embodiment, only four temperature sensors are needed in seven batteries—to accurately reconstruct the core temperature distribution of the entire battery pack. This method significantly reduces reliance on hardware temperature sensors, reduces material costs and system integration difficulty, and solves the problems of high cost, complex wiring, and reduced reliability caused by equipping every battery in the pack with a temperature sensor in existing technologies. It possesses excellent engineering practicality and deployability.

[0115] The parameter-state joint estimation algorithm proposed in this invention constructs an augmented state vector by combining the core temperature to be estimated and key thermal parameters, achieving synchronous online joint estimation of both. By performing real-time correction and updating of the state and parameters within a unified filtering framework, it effectively avoids error accumulation caused by separate identification and can dynamically track parameter changes, thereby significantly improving the robustness and long-term accuracy of the estimation model. This solves the error propagation and accumulation problem caused by the two-step method in existing technologies that first identify thermal parameters and then estimate the core temperature.

[0116] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A battery pack core temperature and thermal parameter joint estimation method based on sparse surface temperature measurement, characterized in that, Comprising the following steps: S101. Establishing a thermal mathematical model of lithium battery; S102. Establishing an observation equation, determining the installation position of the temperature sensor and installing; constructing an observability matrix OB: wherein, is the order of the system, i.e. the number of batteries is twice the number of batteries, if and only if the matrix has a rank equal to The model can be fully observed, and the core temperature and surface temperature of each battery in the battery pack can be accurately estimated; as can be seen from equation (9), since 0B is determined by H, its rank is related to the number and installation position of temperature sensors; by calculating the rank of the observability matrix 0B corresponding to different combinations of the number and installation position of temperature sensors, the combination with the least number of temperature sensors is selected when the rank of the observability matrix 0B is equal to , to determine the number and installation position of temperature sensors; S103. Data acquisition, obtaining the current and temperature sensor data of the battery at the installation position of the temperature sensor, and recording the environmental temperature; S104. Establishing an extended random discrete state space model; the extended random discrete state space model is: （10） （11） （12） （13） where k is the discrete time point, generally the smaller the discrete time interval, the more accurate the discrete model; is the parameter vector composed of 5 battery parameters to be identified; A d is the extended matrix of system matrix A, where is the unit matrix, and 0 is the zero matrix, and the right lower corner of the zero matrix identifies the number of rows and columns; B d is the extended matrix of input matrix B; H a is the extended matrix of output matrix H; represents the extended state vector at the kth discrete time point, which is composed of state vector and parameter vector ; represents the output vector at the kth discrete time point; is the system noise at the kth discrete time point, representing the imperfection of the battery equivalent circuit model and the random fluctuation of the internal electrochemical reaction; is the measurement noise at the kth discrete time point, representing the measurement error and sampling quantization error of each temperature sensor; the system noise and the measurement noise are assumed to be zero-mean white Gaussian noise, and their covariance is and , and are adjustable hyperparameters; S105. Through the parameter-state joint estimation algorithm, the core temperature and parameters of each battery of the battery pack are jointly estimated.

2. The battery pack core temperature and thermal parameter joint estimation method based on sparse surface temperature measurement according to claim 1, characterized in that, In step S101, the mathematical model is: where, Ticore is the core temperature of the i-th cell, Tisurface is the surface temperature of the i-th cell, Tambient is the ambient temperature, Cg is the heat capacity of the gel inside the cell, Rcore-surface is the thermal resistance between the core and the surface, Cshell is the heat capacity of the cell shell, Rsurface-air is the thermal resistance between the surface and the air, Rcell-cell is the thermal resistance between the cells, Rcell is the internal resistance of the cell, I is the current flowing through each cell, the cells are connected in series so the current through each cell is equal, Qi is the energy received by the i-th cell from the adjacent cells, Qcore is the heat generated by the core of the cell, N is the number of cells; equation (1) can be converted to a continuous time state space equation: wherein, wherein, is the system matrix, which is composed of the thermal parameters of the battery, describing how the current temperature state of the system evolves to the next time temperature state, A1, A2, A3, A4, A5 are elements in the A matrix, also composed of the thermal parameters of the battery; is the input matrix, which is composed of the physical parameters of the battery, describing how the external input influences the change of the system state; is the state vector, which is composed of the temperature and of each battery; is the input vector, which is composed of the external excitation, the external excitation being the current and the ambient temperature; in formula (4), and the number of .

3. The battery pack core temperature and thermal parameter joint estimation method based on sparse surface temperature measurement according to claim 2, characterized in that, In step S102, the observation equation is expressed as: wherein y is an array of rows of temperature measurements made by the actual temperature sensor, wherein the number of rows is 1, and wherein the number of columns is N, wherein N is the total number of temperature sensors installed, and wherein the number of the battery in which the temperature sensor is installed is then ; is an output matrix, determined by the location of the temperature sensor as well; is a row column matrix; The i-th row of the matrix is the number of the temperature sensor installed on the surface of the battery numbered ; The i-th row is configured as follows: the element in the j-th column of the i-th row is , when the column index , ; for other column indexes , .

4. The battery pack core temperature and thermal parameter joint estimation method based on sparse surface temperature measurement according to claim 3, characterized in that, In step S105, the to-be-estimated quantities are first initialized: An extended state vector is first defined comprising temperature state estimates and thermal parameter estimates which has the same internal composition as is mathematically defined as （14） wherein, and Ticore,i and Tisurface,i are the core temperature and surface temperature of the i-th battery estimated by the present method, respectively, is the battery thermal parameter estimated by the present method; at initialization, , is a set value.

5. The battery pack core temperature and thermal parameter joint estimation method based on sparse surface temperature measurement according to claim 4, characterized in that, In step S105, after the to-be-estimated quantities are initialized, the core temperature and parameters are updated in a loop iteration, and iteration termination condition judgment and estimated quantity output are performed; The updating process of the core temperature and thermal parameters needs to execute formulas (15) to (18) in a loop iteration, formula (15) is Jacobian matrix calculation, formula (16) is gain matrix calculation, formula (17) is covariance matrix updating, and formula (18) is state and thermal parameter updating, the state is temperature: （15） （16） （17） （18） wherein, is the Jacobian matrix at the kth discrete time instant, which is an intermediate computational product necessary to achieve the battery temperature and thermal parameter estimation; is the gain matrix at the k+1th discrete time instant, which is also an intermediate computational product necessary to achieve the battery temperature and thermal parameter estimation; is the covariance matrix of the estimation error at the kth discrete time instant, which is not only an intermediate computational product necessary to achieve the battery temperature and thermal parameter estimation, but also can be used to determine whether the algorithm converges; is the temperature data measured by the battery temperature sensor at the k+1th discrete time instant.

6. The battery pack core temperature and thermal parameter joint estimation method based on sparse surface temperature measurement according to claim 5, characterized in that, The iteration termination condition judgment in step S105 and the estimated value output process are: when the absolute value of the residual error between the estimated battery temperature and the battery temperature measured by the corresponding temperature sensor is less than ε for consecutive D times of iteration, or the trace of the matrix is less than ε for consecutive D times of iteration, it is determined that the algorithm converges, the iteration is stopped, and the estimated battery core temperature and the thermal parameter are extracted from .

Images