An eVTOL battery pack temperature change prediction method

Abstract

The application discloses an eVTOL battery pack temperature change prediction method, comprising the following steps: constructing an eVTOL battery pack temperature change prediction model; setting a trigger condition for eVTOL battery pack temperature change prediction; screening a prediction stage, collecting battery operation data of N operation cycles in the prediction stage; constructing a linear regression formula of temperature change amount based on the battery pack temperature change prediction model, and iteratively updating a regression vector and a parameter vector of the linear regression formula by using a recursive least square method with a forgetting factor, and estimating the parameter vector to output a convergent parameter vector; predicting a temperature change rate of an eVTOL battery pack in a future prediction stage to obtain maximum, minimum and average temperatures of the eVTOL battery pack in the future prediction stage. The application can dynamically update model parameters based on actual operation processes on a simple model embedded device, and estimate battery temperature changes in a whole life cycle.

Description

Technical Field

[0001] This invention relates to the field of eVTOL battery pack temperature analysis, and specifically to a method for predicting temperature changes in eVTOL battery packs. Background Technology Currently, there are three main methods for predicting battery temperature: 1. Thermoelectric model: Establish an equivalent circuit model of the battery's electro-thermal coupling to estimate the temperature rise (e.g., patent CN202510117570.7: a modeling method for an electrochemical-thermal coupling model of a lithium battery). This algorithm requires high computing power and cannot be run on a single-chip microcomputer. Furthermore, the parameters will drift as the battery ages, making it impossible to achieve accurate prediction in the later stages of battery use.

[0002] 2. Finite element thermal simulation model: Although it has high accuracy, it is mainly completed on high-performance workstations and is used for simulation in the design stage. The calculation is time-consuming and it cannot be run online in real time.

[0003] 3. Estimation based on historical data: The model is trained using historical temperature rise data, and then the model is used to estimate the current temperature rise. For example, patent CN202511502989.0: A method and system for thermal management of new energy vehicle batteries based on big data mentions an algorithm based on historical data trends and dynamic threshold hierarchical intervention. However, the operating conditions of historical data will always differ from the current operating conditions. This prediction method, which is based on data only and not on principles, can only rely on more correction conditions and will still have more estimation errors.

[0004] Deficiencies of existing technology: The predictive model is too complex and cannot run in real time on the MCU of the embedded BMS or has fixed parameters. It cannot adapt to parameter drift due to battery aging. The algorithm, which is entirely based on data features, does not consider physical factors and relies too much on correction.

[0005] The operating conditions of eVTOL (electric vertical takeoff and landing aircraft) may fluctuate drastically during takeoff and landing, but these phases are usually short in duration. During longer phases such as hovering, cruise, and charging, the operating conditions are relatively stable, providing a basis for temperature prediction applications.

[0006] The temperature variation of eVTOL batteries is influenced by a variety of factors, including the battery's internal resistance (different SOCs and different aging stages of the battery pack affect the heat generation and dissipation of the battery system), the temperature difference with the environment, and the structure of the battery system itself. Since the battery pack is largely exposed to the air, flight speed also directly affects the heat exchange rate. These factors are difficult to model accurately, making it extremely difficult to accurately estimate the battery temperature variation under steady-state operating conditions. Summary of the Invention

[0007] To address the aforementioned shortcomings of existing technologies, this invention provides a method for predicting temperature changes in eVTOL battery packs.

[0008] To achieve the above-mentioned objectives, the technical solution adopted by this invention is as follows: A method for predicting temperature changes in an eVTOL battery pack is provided, comprising the following steps: S1: Construct a battery pack temperature change prediction model for eVTOL; S2: Set the trigger conditions for eVTOL battery pack temperature change prediction based on the battery pack temperature collected in real time by the battery thermal management system. S3: The period between the last time the trigger condition was met and the current time the trigger condition is met is taken as the prediction phase. Battery operation data for N operating cycles are collected during the prediction phase, and the temperature difference-speed term is calculated. SOC-SOH-Current Square Term TMS power item and constant term ; S4: Calculate the temperature change during the prediction stage. Based on the battery pack temperature change prediction model, construct a linear regression formula for the temperature change. Use the recursive least squares method with forgetting factor to iteratively update the regression vector and parameter vector of the linear regression formula, estimate the parameter vector, and output the converged parameter vector. S5: Substitute the model coefficients in the converged parameter vector into the battery pack temperature change prediction model to predict the eVTOL battery pack temperature change rate in the future prediction stage, and obtain the maximum, minimum and average temperature of the eVTOL battery pack in the future prediction stage.

[0009] Furthermore, the battery pack temperature change prediction model is based on the eVTOL battery pack temperature change rate. express: ; in, The current temperature of the battery pack. For ambient temperature, For the flight speed of eVTOL, This represents the total current of the eVTOL battery pack. The battery is in its state of charge. For battery health status, These are the correction functions for the battery state of charge and the battery state of health, respectively. For the output power of the battery thermal management system, This serves as a status indicator for the battery thermal management system. The coefficients of the model to be identified are... This is a compensation term for unmodeled factors.

[0010] Furthermore, the battery state-of-charge correction function To describe the nonlinear characteristics of the battery's internal resistance as a function of its state of charge (SOC), a piecewise linear form is used, specifically as follows: .

[0011] The battery health status correction function This describes the trend of battery internal resistance increasing with the state of health (SOH) as the battery ages, using a linear form, specifically: .

[0012] Furthermore, the triggering condition for eVTOL battery pack temperature change prediction is as follows: like If the triggering condition is met, the eVTOL battery pack temperature change prediction will begin, and step S3 will be executed. Otherwise, if the triggering conditions are not met, the battery thermal management system will continue to monitor the battery pack temperature. ; The initial temperature threshold for triggering temperature change prediction.

[0013] Furthermore, the temperature difference-velocity term The calculation method is as follows: Where i is the cycle number within the prediction phase; The difference between the battery pack temperature and the ambient temperature during the i-th operating cycle. Let eVTOL be the flight speed during the i-th operating cycle; SOC-SOH-Current Square Term The calculation method is as follows: ;in, These are the correction functions for the battery state of charge and battery health state in the i-th operating cycle, respectively. The total current of the eVTOL battery pack during the i-th operating cycle; TMS power item The calculation method is as follows: ;in, , These are the output power and operating status flag of the battery thermal management system during the i-th operating cycle, respectively. constant term =1.

[0014] Further, step S4 includes: S41: Based on the initial temperature threshold of the prediction phase and the end temperature threshold Calculate the temperature change during the prediction phase. And calculate the duration of the prediction phase. , The runtime of the running cycle; S42: Based on the battery pack temperature change prediction model, the temperature difference-rate term is used. SOC-SOH-Current Square Term TMS power item and constant term Constructing temperature change The linear regression formula; ; in, This represents the error term in linear regression. S43: Extract the regression vector for the current prediction stage from the linear regression formula. and parameter vector ; , ; S44: Based on regression vectors and parameter vector Write the linear regression formula in the general regression form; ; in, This represents the temperature change during the current forecast phase. S45: Read the laboratory-calibrated initial parameter vector from the NVM memory of the battery thermal management system. , These are the initial model coefficients; and an initial covariance matrix is ​​constructed. , It is a unit vector. It is a tiny positive value; S46: Iterative updates are performed using recursive least squares with a forgetting factor, based on the regression vector of the current prediction stage. and parameter vector Calculate the gain vector, estimate the parameter vector, and output the converged parameter vector. ; Gain vector in the current prediction phase for: ; in, This is the covariance matrix from the previous prediction stage. Forgetting factor; Based on the gain vector in the prediction phase The parameter vector is updated to obtain the estimated parameter vector. ; ; in, This is the parameter vector estimated in the previous prediction stage; Update the covariance matrix during the iterative recursion process: ; The convergence condition for iterative recursive updates is: Or until the set number of iterations is reached.

[0015] The beneficial effects of this invention are as follows: This invention simplifies the multi-factor coupled battery temperature change model into a linear equation related to the internal and external temperature difference, the square root of the reciprocal of the flight speed, the square of the current (including corrections for State of Charge (SOC) and State of Health (SOH)), thermal management power, and thermal management operating status through engineering simplification. After initial parameter calibration in the laboratory, the recursive least squares method with a forgetting factor is used online to identify and dynamically adjust the model coefficients, thereby achieving prediction of battery pack temperature changes. This approach balances the requirements of the aviation field for real-time computation, model interpretability, and long-term prediction accuracy.

[0016] During eVTOL operation, the battery cells generate a large amount of heat, and the battery itself operates in a complex environment. Throughout its life cycle, as the equipment and insulation materials age, the temperature characteristics of the cells will also change. Therefore, temperature changes are usually difficult to estimate. This invention can dynamically update model parameters based on the actual operation process on a simple model embedded device, thereby achieving the estimation of battery temperature changes throughout its entire life cycle. Attached Figure Description

[0017] Figure 1 This is a flowchart of a method for predicting temperature changes in eVTOL battery packs. Detailed Implementation

[0018] The specific embodiments of the present invention are described below to enable those skilled in the art to understand the present invention. However, it should be understood that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, various changes are obvious as long as they are within the spirit and scope of the present invention as defined and determined by the appended claims. All inventions utilizing the concept of the present invention are protected.

[0019] like Figure 1 As shown, a method for predicting temperature changes in an eVTOL battery pack includes the following steps: S1: Construct a battery pack temperature change prediction model for eVTOL; The battery pack temperature change prediction model in this embodiment is based on the eVTOL battery pack temperature change rate. express: ; in, This represents the current temperature of the battery pack (°C). The ambient temperature is (°C). The flight speed of the eVTOL is (m / s). The total current (A) of the eVTOL battery pack. This represents the battery's state of charge (%). Battery health status (%) These are the correction functions for the battery state of charge and the battery state of health, respectively. For the output power of the battery thermal management system (TMS), For the operating status indicators of the battery thermal management system (dimensionless) The value can be 0 or 1, where 0 indicates the battery thermal management system is in operation, and 1 indicates the battery thermal management system is enabled. The coefficients of the model to be identified are... These are unmodeled factors.

[0020] The first term on the right side of the battery pack temperature change prediction model For convective heat transfer, a < 0, indicating that the battery dissipates heat to the environment at a rate of [missing value]. The larger the value, the stronger the convective heat transfer; This indicates that the battery temperature is lower than the ambient heat dissipation. This indicates that the battery temperature equals the heat dissipated from the environment.

[0021] Second item For Joule heating, Based on the basic heat production coefficient, The modulation effects of battery state of charge and battery health state on battery internal resistance are described respectively. Third item Indicates the intervention items of the thermal management system, parameters The value depends on the mode of the battery thermal management system. When the battery thermal management system is in cooling mode... Heating mode Other modes ; Fourth item This is a compensation term for unmodeled factors.

[0022] The battery state-of-charge correction function in this embodiment To describe the nonlinear characteristics of the battery's internal resistance as a function of its state of charge (SOC), a piecewise linear form is used, specifically as follows: .

[0023] The battery health status correction function in this embodiment This describes the trend of battery internal resistance increasing with the state of health (SOH) as the battery ages, using a linear form, specifically: .

[0024] S2: Based on the battery pack temperature collected in real time by the battery thermal management system. Set the trigger conditions for eVTOL battery pack temperature change prediction; like If the triggering condition is met, the eVTOL battery pack temperature change prediction will begin, and step S3 will be executed. Otherwise, if the triggering conditions are not met, the battery thermal management system will continue to monitor the battery pack temperature. ; The initial temperature threshold for triggering temperature change prediction; This embodiment applies to the temperature prediction component of the battery management system for eVTOL (electric vertical takeoff and landing aircraft). It also needs to determine whether to activate the temperature prediction program based on the current flight phase of the eVTOL. Predicted activation of stable operating conditions: Currently in the hovering phase, cruising phase, or ground charging phase; Prohibited conditions: During the transition period between takeoff and landing phases and flight phases.

[0025] S3: The period between the last time the trigger condition was met and the current time the trigger condition is met is taken as the prediction phase. Battery operation data for N operating cycles are collected during the prediction phase, and the temperature difference-speed term is calculated. SOC-SOH-Current Square Term TMS power item and constant term .

[0026] In this embodiment, the temperature difference-velocity term The calculation method is as follows: Where i is the cycle number within the prediction phase; The difference between the battery pack temperature and the ambient temperature during the i-th operating cycle. Let eVTOL be the flight speed during the i-th operating cycle; SOC-SOH-Current Square Term The calculation method is as follows: ;in, These are the correction functions for the battery state of charge and battery health state in the i-th operating cycle, respectively. The total current of the eVTOL battery pack during the i-th operating cycle; TMS power item The calculation method is as follows: ;in, , These are the output power and operating status flag of the battery thermal management system during the i-th operating cycle, respectively. constant term =1.

[0027] S4: Calculate the temperature change during the prediction stage. Based on the battery pack temperature change prediction model, construct a linear regression formula for the temperature change. Use the recursive least squares method with a forgetting factor to iteratively update the regression vector and parameter vector of the linear regression formula, estimate the parameter vector, and output the converged parameter vector.

[0028] Step S4 specifically includes: S41: Based on the initial temperature threshold of the prediction phase and the end temperature threshold Calculate the temperature change during the prediction phase. And calculate the duration of the prediction phase. , The runtime of the running cycle; S42: Based on the battery pack temperature change prediction model, the temperature difference-rate term is used. SOC-SOH-Current Square Term TMS power item and constant term Constructing temperature change The linear regression formula; ; in, This represents the error term in linear regression. S43: Extract the regression vector for the current prediction stage from the linear regression formula. and parameter vector ; , ; S44: Based on regression vectors and parameter vector Write the linear regression formula in the general regression form; ; in, This represents the temperature change during the current forecast phase. S45: Read the laboratory-calibrated initial parameter vector from the NVM memory of the battery thermal management system. , These are the initial model coefficients; and an initial covariance matrix is ​​constructed. , It is a unit vector. For a small positive value, this embodiment takes... ; S46: Iterative updates are performed using recursive least squares with a forgetting factor, based on the regression vector of the current prediction stage. and parameter vector Calculate the gain vector, estimate the parameter vector, and output the converged parameter vector. ; Gain vector in the current prediction phase for: ; in, This is the covariance matrix from the previous prediction stage. As the forgetting factor, this embodiment takes ; Based on the gain vector in the prediction phase The parameter vector is updated to obtain the estimated parameter vector. ; ; in, This is the parameter vector estimated in the previous prediction stage; Update the covariance matrix during the iterative recursion process: ; The convergence condition for iterative recursive updates is: Or until the set number of iterations is reached.

[0029] S5: Converge the parameter vector Model coefficients Substituting the data into the battery pack temperature change prediction model, the rate of temperature change of the eVTOL battery pack in the future prediction stage is predicted, and the maximum, minimum and average temperatures of the eVTOL battery pack in the future prediction stage are obtained.

[0030] By multiplying the predicted eVTOL battery pack temperature change rate by different times within the future prediction period, the eVTOL battery pack temperature at any time within the future prediction period can be obtained, and then the maximum, minimum and average temperatures can be selected.

[0031] During initial operation, the system initializes model parameters based on laboratory calibration data for prediction. In later use, the BMS collects data on the internal and external temperatures of the eVTOL battery pack, real-time current, flight speed, SOC / SOH calculated by other components, and the operating status and thermal management power of the thermal management system. It uses a simplified linear temperature change model to predict battery temperature changes in real time and continuously corrects the model coefficients using a recursive least squares method with a forgetting factor.

[0032] In this embodiment, several engineering projects used in actual applications are simplified as follows: 1. In the short term, the heat capacity of the battery pack in the eVTOL battery pack is fixed, and the equivalent heat capacity of the battery pack as a whole is regarded as constant in the short term.

[0033] 2. The battery's internal resistance is considered fixed for a short period, but it is strongly correlated with SOC (State of Charge) and SOH (State of Health, reflecting the degree of performance degradation relative to the factory settings). The battery's internal resistance R... int Influenced by both SOC and SOH, the internal resistance is high at both ends of the SOC and low in the middle; as the number of usage cycles increases, the SOH decreases, and the overall internal resistance increases. Therefore, the rate of temperature change caused by current-generated heat is proportional to the square of the current, and the proportionality coefficient is modulated by SOC and SOH.

[0034] 3. Ignoring the heat of electrochemical reaction (entropy change heat), Joule heating dominates during charging and discharging, while the contribution of heat of electrochemical reaction is relatively small.

[0035] 4. The battery pack is largely exposed to the outside, and it exchanges heat with the outside world through convection. Therefore, the temperature change is positively correlated with the internal and external temperature difference and the square root of the speed.

[0036] 5. The output power of the thermal management system is positively correlated with the cooling and heating capacity, and is linearly correlated with the temperature change of the battery pack when the heat capacity is fixed.

Claims

1. A method for predicting temperature changes in an eVTOL battery pack, characterized in that, Includes the following steps: S1: Construct a battery pack temperature change prediction model for eVTOL; S2: Set the trigger conditions for eVTOL battery pack temperature change prediction based on the battery pack temperature collected in real time by the battery thermal management system. S3: The period between the last time the trigger condition was met and the current time the trigger condition is met is taken as the prediction phase. Battery operation data for N operating cycles are collected during the prediction phase, and the temperature difference-speed term is calculated. SOC-SOH-Current Square Term TMS power item and constant term ; S4: Calculate the temperature change during the prediction stage. Based on the battery pack temperature change prediction model, construct a linear regression formula for the temperature change. Use the recursive least squares method with forgetting factor to iteratively update the regression vector and parameter vector of the linear regression formula, estimate the parameter vector, and output the converged parameter vector. S5: Substitute the model coefficients in the converged parameter vector into the battery pack temperature change prediction model to predict the eVTOL battery pack temperature change rate in the future prediction stage, and obtain the maximum, minimum and average temperature of the eVTOL battery pack in the future prediction stage.

2. The method for predicting temperature changes in an eVTOL battery pack according to claim 1, characterized in that, The battery pack temperature change prediction model is based on the eVTOL battery pack temperature change rate. express: ； in, The current temperature of the battery pack. For ambient temperature, For the flight speed of eVTOL, This represents the total current of the eVTOL battery pack. The battery is in its state of charge. For battery health status, These are the correction functions for the battery state of charge and the battery state of health, respectively. For the output power of the battery thermal management system, This serves as a status indicator for the battery thermal management system. The coefficients of the model to be identified are... This is a compensation term for unmodeled factors.

3. The method for predicting temperature changes in an eVTOL battery pack according to claim 2, characterized in that, The battery state of charge correction function To describe the nonlinear characteristics of the battery's internal resistance as a function of its state of charge (SOC), a piecewise linear form is used, specifically as follows: ； The battery health status correction function This describes the trend of battery internal resistance increasing with the state of health (SOH) as the battery ages, using a linear form, specifically: 。 4. The method for predicting temperature changes in an eVTOL battery pack according to claim 1, characterized in that, The triggering condition for predicting the temperature change of the eVTOL battery pack is as follows: like If the triggering condition is met, the eVTOL battery pack temperature change prediction will begin, and step S3 will be executed. Otherwise, if the triggering conditions are not met, the battery thermal management system will continue to monitor the battery pack temperature. ; The initial temperature threshold for triggering temperature change prediction.

5. The method for predicting temperature changes in an eVTOL battery pack according to claim 1, characterized in that, The temperature difference-rate term The calculation method is as follows: Where i is the cycle number within the prediction phase; The difference between the battery pack temperature and the ambient temperature during the i-th operating cycle. Let eVTOL be the flight speed during the i-th operating cycle; SOC-SOH-Current Square Term The calculation method is as follows: ;in, These are the correction functions for the battery state of charge and battery health state in the i-th operating cycle, respectively. The total current of the eVTOL battery pack during the i-th operating cycle; TMS power item The calculation method is as follows: ;in, , These represent the output power and operating status flag of the battery thermal management system during the i-th operating cycle; constant term =1.

6. The method for predicting temperature changes in an eVTOL battery pack according to claim 1, characterized in that, Step S4 includes: S41: Based on the initial temperature threshold of the prediction phase and the end temperature threshold Calculate the temperature change during the prediction phase. And calculate the duration of the prediction phase. , The runtime of the running cycle; S42: Based on the battery pack temperature change prediction model, the temperature difference-rate term is used. SOC-SOH-Current Square Term TMS power item and constant term Constructing temperature change The linear regression formula; ； in, This represents the error term in linear regression. S43: Extract the regression vector for the current prediction stage from the linear regression formula. and parameter vector ; ， ； S44: Based on regression vectors and parameter vector Write the linear regression formula in the general regression form; ； in, This represents the temperature change during the current forecast phase. S45: Read the laboratory-calibrated initial parameter vector from the NVM memory of the battery thermal management system. , These are the initial model coefficients; and an initial covariance matrix is ​​constructed. , It is a unit vector. It is a tiny positive value; S46: Iterative updates are performed using recursive least squares with a forgetting factor, based on the regression vector of the current prediction stage. and parameter vector Calculate the gain vector, estimate the parameter vector, and output the converged parameter vector. ; Gain vector in the current prediction phase for: ； in, This is the covariance matrix from the previous prediction stage. Forgetting factor; Based on the gain vector in the prediction phase The parameter vector is updated to obtain the estimated parameter vector. ; ； in, This is the parameter vector estimated in the previous prediction stage; Update the covariance matrix during the iterative recursion process: ； The convergence condition for iterative recursive updates is: Or until the set number of iterations is reached.

Images