Thermal management control method and device for new energy automobile battery

Abstract

The invention discloses a thermal management control method and device for a new energy automobile battery, and relates to the field of battery thermal management control, the device comprises an entropy change module, a fault module, a decision module, a phase change module and a scheduling module, entropy change intensity and nuclear heat flow field analysis is carried out on collected current and voltage signals to obtain an aging thermal topology, and the aging thermal topology is subjected to thermal management control. The method comprises the following steps: constructing an anisotropic diffusion equation based on aging thermal topology in combination with a fault diffusion tensor, discretizing a sub-strategy set for a safety margin field, constructing a cooling Nash game, analyzing through a payment matrix and a replicated dynamic equation, screening a stable strategy, mapping, and constructing an implicit phase change criterion model for the surface temperature of a battery to obtain a phase change cumulant; the phase change cumulant latent heat is analyzed and then cooperatively arbitrated with the cooling membership degree to obtain arbitration cooling capacity, the water pump flow is obtained through limiting operation based on the arbitration cooling capacity and the temperature deviation, then cubic root mapping is conducted on the water pump flow and the arbitration cooling capacity to obtain the fan rotating speed, and accurate control over battery thermal management is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of battery thermal management control, and in particular to a thermal management control method and device for a new energy vehicle battery. Background Art

[0002] Existing thermal management technology for new energy vehicle batteries has obvious deficiencies in analyzing battery aging. It is difficult to deeply analyze the changes in thermal distribution inside the battery through electrical signals, cannot accurately locate local overheating areas, and cannot effectively identify potential faults such as lithium plating and internal short circuits. This leads to inaccurate assessments of battery health status and difficulty in detecting thermal runaway risks in advance.

[0003] In terms of cooling control strategies, traditional technologies lack differentiated management of the safety status of battery cells. The formulation of cooling strategies is not intelligent enough, and it is impossible to dynamically adjust the cooling distribution according to the actual needs of the battery. It is also difficult to find the optimal balance between energy consumption and cooling effect, resulting in energy waste and affecting battery performance and service life due to untimely cooling.

[0004] In order to solve the above-mentioned defects, a technical solution is now provided. Summary of the Invention

[0005] In order to solve the technical problems raised by the above background technology, the present invention is proposed. Embodiments of the present invention provide a thermal management control method and device for a new energy vehicle battery.

[0006] The purpose of the present invention can be achieved through the following technical solutions: In a first aspect, the present invention provides a thermal management control method for a new energy vehicle battery, comprising the following steps: Step S100: performing entropy change intensity, relaxation thermal spectrum, and nuclear thermal flow field analysis on the collected current and voltage signals to obtain aging thermal topology; Step S200: constructing an anisotropic diffusion equation based on the aging thermal topology and the fault diffusion tensor, solving it to obtain the risk temperature field, and then scanning and mapping it to obtain the safety margin field; Step S300: Discretize the safety margin field into a set of strategies, construct a cooling Nash game, analyze the payoff matrix and the replication dynamic equation, screen the stable strategies, and then map them to obtain the cooling membership; Step S400: constructing an implicit phase change criterion model for the battery surface temperature to obtain the phase change accumulation, analyzing the latent heat of the phase change accumulation and arbitrating it with the cooling membership to obtain the arbitration cooling capacity; Step S500: Based on the arbitration cooling capacity and the deviation between the measured temperature and the target temperature, a water pump flow rate is obtained through a limiting operation, and then a cube root mapping is performed on the water pump flow rate and the arbitration cooling capacity to obtain a fan speed.

[0007] Furthermore, the aging thermal topology analysis steps are as follows: Perform Fourier transform on the relaxation thermal spectrum, extract the inherent frequency variable in the Fourier transform, and obtain the heat source fluctuation angular frequency. Based on the relaxation time constant, calculate the harmonic mean to obtain the comprehensive time scale of the dominant relaxation process. Perform numerical conversion on the comprehensive time scale of the dominant relaxation process to obtain the smoothing characteristic bandwidth coefficient. Dynamic collaborative calculation and parameter coupling configuration of the heat source fluctuation angular frequency and smoothing characteristic bandwidth coefficient are performed to obtain the frequency domain filter kernel function parameter set. Frequency domain fusion analysis is performed on the frequency domain filter kernel function parameter set and the relaxation thermal spectrum. The nuclear thermal flow field is obtained by applying the convolution kernel operator in the frequency domain space to perform point multiplication filtering and space-time domain inversion transformation. Perform Laplace transform on the nuclear thermal flow field to obtain complex frequency variables. Establish an embedded capacity attenuation model for battery discharge and perform multiplication coupling between the temperature compensation function and the Coulomb integral to obtain the health status value. Based on the linear multiplication of the health status value and the aging sensitivity coefficient, the aging time constant is obtained. The degradation transfer analysis of the complex frequency variable and the aging time constant is carried out through the first-order inertia link model to obtain the degradation transfer function. The Laplace transform result of the nuclear thermal flow field is multiplied by the degradation transfer function in the frequency domain, and then reconstructed through the inverse Laplace transform to obtain the aging thermal topology.

[0008] Furthermore, the relaxation thermogram analysis steps are as follows: By analyzing the voltage-to-current change rate through the real-time acquisition of current and voltage signals, the dynamic equivalent resistance is obtained. The entropy change intensity is obtained by coupling the temperature gradient and entropy change factor of the dynamic equivalent resistance. The relaxation characteristic coefficient is obtained by performing standardized electrochemical impedance spectroscopy calibration analysis on the battery; The entropy change intensity is subjected to Prony spectrum decomposition algorithm and numerical iterative separation to obtain the relaxation amplitude coefficient; The relaxation time constant is obtained by performing algebraic division operation on the relaxation characteristic coefficient and the absolute value of the entropy change intensity. The relaxation amplitude coefficient and the relaxation time constant are superimposed on each other through an exponential decay model to obtain the relaxation thermal spectrum.

[0009] Furthermore, the safety margin field analysis steps are as follows: Based on the fault fingerprint vector, a self-organizing fault mapping algorithm is performed to obtain a fault type code; Based on the fault type code, a preset material characteristic template is called to obtain the fault diffusion tensor; Anisotropic diffusion equations are constructed based on the fault diffusion tensor and aging thermal topology, and the risk temperature field is obtained by solving the equations using the finite element method. Dynamic spatial boundary scanning and critical distance mapping are performed on the risk temperature field to obtain the safety margin field.

[0010] Furthermore, the fault fingerprint vector analysis steps are as follows: The discharge current is superimposed with a slightly triangular wave current to obtain a slightly disturbed charge and discharge current. The slightly disturbed charge and discharge current and voltage response signal are subjected to second-order mixed partial derivatives and temperature-normalized integration to obtain the differential conductivity entropy. The instantaneous heat flux density peak value of the aging thermal topology is analyzed to obtain the maximum heat flux density. The heat flux gradient field of the aging thermal topology is analyzed to obtain the heat flux gradient norm. Based on the maximum heat flux density, heat flux gradient norm and differential conductivity entropy, multi-physical field feature vectorization synthesis is performed to obtain the fault fingerprint vector.

[0011] Furthermore, the cooling membership analysis steps are as follows: The thermal safety benefit function and the safety margin field are subjected to global strategy traversal processing to dynamically generate the benefit value of each unit under each combination. The unit sequence is used as the row and the cooling strategy combination arrangement is used as the column. The benefit value calculation results of the unit in all strategy combinations are integrated into the game payoff matrix. Based on the game payoff matrix, a replication dynamic equation is constructed to analyze the stability of the dynamic changes in the strategy proportions, and the steady-state distribution of the relative benefit evolution of the strategy proportions is obtained. The dominant strategy screening analysis is performed on the steady-state distribution of the relative benefit evolution of the strategy proportions, and a global convergent evolutionary stable strategy is obtained. Based on the global convergence evolutionary stability strategy, fuzzy membership function mapping is performed to obtain the cooling membership.

[0012] Furthermore, the thermal safety benefit function analysis steps are as follows: The safety margin field is discretized into N units of safety margin values, where N corresponds to the total number of battery module units evaluated in the safety margin field. A safety margin interval is preset. If the unit safety margin value is in different safety margin intervals and corresponds to high cooling capacity, medium cooling capacity, and low cooling capacity, respectively, the strategy set is obtained through analysis.

[0013] Based on the safety margin field and strategy set, spatial security situation mapping is performed to construct a cooling Nash game; In the cooling Nash game framework, the safety distance and average safety state of the battery cells are analyzed based on the safety margin field, and the thermal safety benefit function is constructed in combination with the selection cost of the cooling strategy.

[0014] Furthermore, the arbitration cooling capacity analysis steps are as follows: An implicit phase change criterion model is constructed by measuring the battery surface temperature, and a step-excitation piecewise integration operation is performed to obtain the phase change accumulation. The dynamic phase change latent heat flow is obtained by quantifying the differential heat absorption rate, scaling the physical properties and performing nonlinear correction on the phase change cumulant. The dynamic phase change latent heat flux, cooling membership and maximum heat flux density are collaboratively compared and the cooling demand is arbitrated to obtain the arbitration cooling capacity.

[0015] Furthermore, the fan speed analysis steps are as follows: The temperature range is divided into normal, slightly hot, and overheated discrete intervals. For each sensor and each discrete interval, the probability of associating the sensor's measured temperature with the interval median is calculated to obtain the support of each sensor for each interval. Define the weight value when no failure rate occurs. Obtain the weight of each temperature sensor through the mapping relationship of actual historical failure rate statistical data. Based on the weight of each temperature sensor and the support of a single sensor for each interval, use the DS rule to fuse the support of all sensors for each interval to obtain the fusion probability. Based on the fusion probability, the credible interval is determined by the arg max function. The indicator function is used to obtain only the sensor temperature values ​​that fall within the credible interval. The weighted average is performed according to the sensor weight to obtain the credible temperature. The water pump current and fan speed are obtained and multiplied to amplify the composite fault to obtain the health attenuation index. The health attenuation index is nonlinearly compressed to obtain the compressed health attenuation index. The compressed health attenuation index is linearly reverse mapped to obtain the actuator health. The temperature rise deviation is calculated based on the trusted temperature. The water pump flow is obtained through a restricted operation on the arbitration cooling capacity, actuator health, and temperature rise deviation. The fan speed is obtained by performing a nonlinear cube root mapping on the water pump flow, arbitration cooling capacity, and temperature rise deviation.

[0016] In a second aspect, the present invention provides a thermal management control device for a new energy vehicle battery, comprising an entropy change module, a fault module, a decision module, a phase change module, and a scheduling module; The entropy change module analyzes the entropy change intensity, relaxation thermal spectrum, and nuclear thermal flow field of the collected current and voltage signals to obtain the aging thermal topology; The fault module constructs an anisotropic diffusion equation based on the aging thermal topology and the fault diffusion tensor, solves it to obtain the risk temperature field, and then scans and maps it to obtain the safety margin field; The decision module discretizes the safety margin field into strategy sets, constructs the cooling Nash game, analyzes the payoff matrix and the replication dynamic equation, selects the stable strategy and then maps it to obtain the cooling membership. The phase change module constructs an implicit phase change criterion model for the battery surface temperature to obtain the phase change accumulation. After analyzing the latent heat of the phase change accumulation, it coordinates arbitration with the cooling membership to obtain the arbitration cooling capacity. The scheduling module obtains the pump flow rate through limiting calculation based on the arbitration cooling capacity, the deviation between the measured temperature and the target temperature, and then performs cube root mapping of the pump flow rate and the arbitration cooling capacity to obtain the fan speed.

[0017] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention obtains the aging thermal topology by analyzing the entropy change intensity, relaxation thermal spectrum, and nuclear thermal flow field of the collected current and voltage signals. Based on the aging thermal topology and the fault diffusion tensor, an anisotropic diffusion equation is constructed, and the risk temperature field is obtained by solution. The safety margin field is then scanned and mapped to obtain the safety margin field. The safety margin field is discretized into a strategy set, and a cooling Nash game is constructed. After analysis of the payoff matrix and the replicated dynamic equation, the stable strategy is screened and mapped to obtain the cooling membership. An implicit phase change criterion model is constructed for the battery surface temperature to obtain the phase change accumulation. After analyzing the latent heat of the phase change accumulation, it is coordinated with the cooling membership to obtain the arbitration cooling capacity. The electrical signal can be used to deeply analyze the heat distribution changes inside the battery, accurately locate the local overheating area, and effectively identify potential faults such as lithium plating and internal short circuit. The battery health status is accurately assessed and the risk of thermal runaway can be discovered in advance.

[0018] 2. The present invention obtains the water pump flow through limiting calculation based on the arbitration cooling capacity, the deviation between the measured temperature and the target temperature, and then performs cube root mapping on the water pump flow and the arbitration cooling capacity to obtain the fan speed. In terms of cooling control strategy, it can manage the safety status of battery cells in a differentiated manner. The formulation of cooling strategy can be dynamic and intelligent, and can dynamically adjust the cooling capacity distribution according to the actual needs of the battery. It can also find the best balance between energy consumption and cooling effect, reduce the occurrence of energy waste, and reduce the situation where battery performance and service life are affected by untimely cooling. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. The following drawings are not intentionally scaled to the actual size, and the focus is on illustrating the main purpose of the present invention.

[0020] Figure 1 is a flow chart of the method of the present invention; Figure 2 is a block diagram of the device of the present invention; Figure 3 This is a flow chart of step S200 of the present invention. DETAILED DESCRIPTION

[0021] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts also fall within the scope of protection of the present invention.

[0022] like Figure 1As shown, a thermal management control method for a new energy vehicle battery includes the following steps: Step S100: performing entropy change intensity, relaxation thermal spectrum, and nuclear thermal flow field analysis on the collected current and voltage signals to obtain aging thermal topology; Step S200: constructing an anisotropic diffusion equation based on the aging thermal topology and the fault diffusion tensor, solving it to obtain the risk temperature field, and then scanning and mapping it to obtain the safety margin field; Step S300: Discretize the safety margin field into a set of strategies, construct a cooling Nash game, analyze the payoff matrix and the replication dynamic equation, screen the stable strategies, and then map them to obtain the cooling membership; Step S400: constructing an implicit phase change criterion model for the battery surface temperature to obtain the phase change accumulation, analyzing the latent heat of the phase change accumulation and arbitrating it with the cooling membership to obtain the arbitration cooling capacity; Step S500: Based on the arbitration cooling capacity and the deviation between the measured temperature and the target temperature, a water pump flow rate is obtained through a limiting operation, and then the water pump flow rate and the arbitration cooling capacity are mapped to a cube root to obtain a fan speed.

[0023] Specifically, the detailed analysis of step S100 is as follows: By analyzing the voltage-to-current change rate through the real-time acquisition of current and voltage signals, the dynamic equivalent resistance is obtained. The entropy change intensity is obtained by coupling the temperature gradient and entropy change factor of the dynamic equivalent resistance. Specifically, the dynamic equivalent resistance ,in Indicates the conditional symbol, T represents the absolute temperature of the battery system, measured in a constant temperature environment, They represent the real-time voltage measured by the current sensor and the real-time current measured by the voltage sensor respectively. The entropy change intensity is obtained by coupling analysis of the temperature gradient and the entropy change factor of the dynamic equivalent resistance. , It represents the temperature coefficient of resistance, reflecting the temperature drift characteristics of the material resistance; The relaxation characteristic coefficient Ck is obtained by performing standardized electrochemical impedance spectroscopy calibration analysis on the battery; The entropy change intensity is subjected to Prony spectrum decomposition algorithm and numerical iterative separation to obtain the relaxation amplitude coefficient Ak.

[0024] Specifically, the battery is calibrated with a standardized electrochemical impedance spectrum, multi-frequency sinusoidal current scanning excitation is applied to the battery system under a constant temperature environment, and the voltage dynamic response data in the high-frequency to low-frequency range is synchronously collected. The measured data is fitted to generate a Nyquist curve map, and the equivalent circuit parameters corresponding to the three characteristic trough areas on the curve are identified. The characteristic parameter combinations of different relaxation processes such as charge transfer, ion diffusion, and interface polarization are extracted; the conversion of key physical quantities is completed based on the relationship between the frequency position and time constant of the relaxation peak, and finally the relaxation characteristic coefficient Ck (k=1, 2, 3) is obtained, where k is a subscript index representing three independent electrochemical relaxation processes, and its value is regularly updated and calibrated with the aging status of the battery.

[0025] Based on the entropy change intensity signal Se(t) collected during the real-time operation of the battery, a dynamic thermal effect waveform segment lasting for a certain period of time is intercepted as input during the constant current charge and discharge stage. The certain period of time is specifically 5 minutes. The Prony spectral decomposition algorithm is used to fit the time domain signal with a third-order exponential decay model. A matrix equation with a specific structure is constructed and the characteristic roots are solved through least squares optimization. The amplitude components corresponding to the three dominant relaxation processes in the signal are separated through numerical iteration, and finally the relaxation amplitude coefficient Ak (k=1, 2, 3) is obtained. This coefficient is recalculated and updated every minute according to the battery working status.

[0026] Perform algebraic division on the relaxation characteristic coefficient and the absolute value of the entropy change intensity to obtain the relaxation time constant , the relaxation amplitude coefficient and the relaxation time constant are superimposed by the exponential decay model to obtain the relaxation thermal spectrum. The specific exponential decay model includes: ,in represents the exponential decay function, e represents the natural constant, Represents the relaxation thermal spectrum, which is a dynamic three-dimensional thermal distribution map that changes with time.

[0027] The relaxation thermal spectrum is Fourier transformed, and the inherent frequency variables in the Fourier transform are extracted to obtain the heat source fluctuation angular frequency. The comprehensive time scale of the dominant relaxation process is obtained through harmonic average calculation based on the relaxation time constant. The comprehensive time scale of the dominant relaxation process is numerically converted to obtain the smoothing characteristic bandwidth coefficient.

[0028] Specifically, based on the three relaxation time constants, the comprehensive time scale of the dominant relaxation process is obtained by solving the harmonic mean of these time constants. The comprehensive time scale of the dominant relaxation process is then substituted into the preset relationship of twice the pi multiple to complete the numerical conversion, and the smoothing characteristic bandwidth coefficient is obtained. This coefficient is dynamically updated according to the thermal relaxation characteristics.

[0029] The heat source fluctuation angular frequency and smoothing characteristic bandwidth coefficient are dynamically coordinated and parameter coupled to obtain the frequency domain filter kernel function parameter set. The frequency domain filter kernel function parameter set and the relaxation thermal spectrum are subjected to frequency domain fusion analysis, and the nuclear thermal flow field is obtained by applying the convolution kernel operator in the frequency domain space to perform point multiplication filtering and space-time domain inversion transformation.

[0030] Specifically, the frequency domain filter kernel function parameter set , where ω represents the angular frequency of heat source fluctuation, σ represents the smoothing characteristic bandwidth coefficient, represents the set cutoff frequency, sinc(*) represents the Singer function, and the nuclear thermal flow field ,in represents the nuclear thermal flow field that characterizes the spatial distribution of the intrinsic heat source, represents the inverse Fourier transform operator, F represents the Fourier transform operator, Represents point multiplication operation and frequency domain filtering processing.

[0031] The nuclear thermal flow field is Laplace transformed to obtain complex frequency variables. An embedded capacity attenuation model is established for battery discharge to multiply and couple the temperature compensation function with the Coulomb integral to obtain the health status value.

[0032] Specifically, Laplace transform is performed on the nuclear thermal flow field to obtain a complex frequency variable, where the real part represents the heat source decay rate and the imaginary part maps the thermal fluctuation frequency. By establishing an embedded capacity decay model for battery discharge and multiplying the temperature compensation function with the Coulomb integral, the health status value is obtained. The specific embedded capacity decay model includes: , where SH represents the health status value, Indicates the initial nominal capacity, the battery factory value, Indicates the discharge current, the current sensor samples in real time, Represents the temperature compensation factor. A three-dimensional temperature-capacity compensation coefficient matrix is ​​established through calibration experiments. The compensation factor is obtained by looking up the table and matching it according to the real-time temperature value of the battery module NTC sensor.

[0033] Based on the linear multiplication of the health status value and the aging sensitivity coefficient, the aging time constant is obtained. The degradation transfer analysis of the complex frequency variable and the aging time constant is carried out through the first-order inertia link model to obtain the degradation transfer function. The Laplace transform result of the nuclear thermal flow field is multiplied by the degradation transfer function in the frequency domain, and then reconstructed through the inverse Laplace transform to obtain the aging thermal topology.

[0034] The specific aging sensitivity coefficient is obtained by synchronously monitoring the associated data set of capacity attenuation rate and thermal response delay characteristics in the battery aging experiment, and fitting the capacity attenuation percentage and response time extension by the least squares method, where the aging time constant is , where k represents the aging sensitivity coefficient, and the first-order inertia link model includes ,in represents the degradation transfer function, s represents the complex frequency variable, and the aging thermal topology , where L represents the Laplace transform operator, represents the inverse Laplace transform operator, represents a frequency domain multiplication operation, where x represents the spatial position.

[0035] Specifically, a Laplace transform is performed on the nuclear thermal flow field to obtain a complex frequency variable. The aging time constant is then calculated using the health state value and aging sensitivity coefficient. A degradation transfer analysis is then performed using a first-order inertial link model to obtain the degradation transfer function. The Laplace transform of the nuclear thermal flow field and the degradation transfer function are then multiplied in the frequency domain. Finally, an inverse Laplace transform is performed to reconstruct the aging thermal topology. This method can identify the maximum heat flux density and heat flux gradient norm in the battery, localize the overheating focus, and provide data for generating a fault fingerprint vector, enabling subsequent determination of the battery fault type.

[0036] like Figure 3 Specifically, the specific analysis of step S200 is as follows: A slight triangular wave current is superimposed on the discharge current to obtain a slightly disturbed charge and discharge current. , where ΔI is the triangular wave perturbation amplitude, tri(*) is the triangular wave function, and f is the frequency. The second-order mixed partial derivative and temperature normalized integral are performed on the slightly perturbed charge and discharge current and voltage response signals to obtain the differential conductivity entropy , where U is the battery charge and discharge current under slight disturbance Real-time dynamic voltage generated under excitation.

[0037] The instantaneous heat flux peak value analysis of the aging thermal topology was performed to obtain the maximum heat flux density , It means scanning the aging thermal topology, locating the maximum instantaneous heat flux density, capturing the local overheating focus, analyzing the heat flux gradient field of the aging thermal topology, and obtaining the heat flux gradient norm. ,in Represents the norm, and performs multi-physics field feature vectorization synthesis based on the maximum heat flux density, heat flux gradient norm and differential conductivity entropy to obtain the fault fingerprint vector .

[0038] Based on the fault fingerprint vector, the self-organizing fault mapping algorithm is converted to obtain the fault type code. The specific self-organizing fault mapping algorithm includes: ,in represents the minimization operation, optimizing the objective function, Represents the cluster center vector, which has the same dimension as Vf. represents the fault fingerprint vector of the i-th sample, the fault characteristics of different battery cells, Represents double summation, traversing all samples and cluster centers, Represents the square of the Euclidean distance, the distance from the sample to the cluster center, represents the neighborhood function, σ represents the neighborhood radius coefficient, which controls the clustering range, and Cf represents the fault type code, specifically 0=normal, 1=lithium deposition, and 2=internal short circuit.

[0039] Based on the fault type code, the preset material characteristic template is called to obtain the fault diffusion tensor. Specifically, when Cf=0, the fault diffusion tensor is , is an isotropic template, where d0 represents the base thermal diffusivity. When Cf=1, the fault diffusion tensor is , is a radially anisotropic template, Thermal diffusivity in the parallel direction, represents the thermal diffusivity in the vertical direction. When Cf=2, the fault diffusion tensor is , is a high thermal conductivity template, dx, dy are the thermal diffusivities in the short-circuit path direction, all parameters of the fault diffusion tensor (d0, , d⊥, dx, dy) are all pre-stored through experimental calibration, and lithium metal is deposited on the negative electrode of the battery using the low-temperature, high-current cycling method (corresponding to fault type 1 = lithium deposition). An infrared thermal imager is aimed at the battery to obtain the vertical diffusion rate, and metal particles are implanted inside the battery to induce dendrite short circuit (corresponding to fault type 2 = internal short circuit). The infrared thermal imager tracks the short circuit point and obtains the thermal diffusivity in the direction of the short circuit path. The thermal diffusivity of a normal and healthy battery is directly measured through an infrared thermal imaging experiment to obtain a benchmark thermal diffusivity. The benchmark value under normal conditions is directly used to obtain the parallel thermal diffusivity.

[0040] Based on the fault diffusion tensor and aging thermal topology, an anisotropic diffusion equation is constructed and solved by the finite element method to obtain the risk temperature field. , where T represents the temperature field, the spatially distributed temperature, represents the divergence operator, flux density calculation, represents the fault diffusion tensor, Represents the gradient operator, and the equation is solved by the finite element method to obtain the risk temperature field .

[0041] Perform dynamic spatial boundary scanning and critical distance mapping on the risk temperature field to obtain the safety margin field ,in Represents minimization operation, looking for the minimum distance that meets the conditions, x represents the spatial position vector, represents the Euclidean norm, Indicates a conditional separator, if the condition...is met, represents the safe temperature threshold and the thermal stability limit of the material represents the safety margin and design redundancy.

[0042] Specifically, an anisotropic diffusion equation is constructed based on the fault diffusion tensor and the aging heat topology, and the risk temperature field is obtained by solving it with the finite element method. Then, dynamic spatial boundary scanning and critical distance mapping are performed on the risk temperature field to obtain the safety margin field. This can classify battery cells into strategy sets of high cooling capacity, medium cooling capacity, and low cooling capacity according to the safety margin values, providing a basis for constructing a cooling Nash game, and enabling subsequent formulation of cooling strategies according to the safety conditions of each battery cell.

[0043] Specifically, the detailed analysis of step S300 is as follows: Discretize based on the safety margin field into safety margin values of N cells , where N corresponds to the total number of battery module cells evaluated in the safety margin field. Preset safety margin intervals XT1, XT2, and XT3, and XT1 < XT2 < XT3. If the cell safety margin value is within the safety margin interval XT1, it corresponds to high cooling capacity (urgent risk). If the cell safety margin value is within the safety margin interval XT2, it corresponds to medium cooling capacity (medium risk). If the cell safety margin value is within the safety margin interval XT3, it corresponds to low cooling capacity (slight risk), obtaining the strategy set .

[0044] Construct a cooling Nash game based on the safety margin field and the strategy set through spatial safety situation mapping. Specifically , where G is the cooling Nash game, and the mathematical representation of the Nash game is the tuple symbol, defining a multi-dimensional structure. N is the number of players, corresponding to the total number of battery module cells is the set of thermal safety benefit functions, and the thermal safety benefit function of player i, corresponding to the thermal safety benefit function of battery module cell i is the strategy of player i, corresponding to the strategy of battery module cell i , is the strategies of other players, the strategy combination of all players except i.

[0045] Under the framework of the cooling Nash game, analyze the safety distance and average safety status of battery cells based on the safety margin field, and combine with the selection cost of cooling strategies to construct the thermal safety benefit function , where represents the safety margin of cell i is the average safety margin of the battery safety margin field is the strategy cost of cell i. Preset: high cooling capacity = 3, medium cooling capacity = 2, low cooling capacity = 1 is the weight coefficient.

[0046] Perform global strategy traversal on the thermal safety benefit function and safety margin field, and dynamically generate the benefit value of each unit under each combination ,in Represents each strategy combination, where each unit has 3 strategies and N units have 3 N The unit sequence is used as a row and the cooling strategy combination is arranged as a column. The calculation results of the unit's benefit value in all strategy combinations are integrated into the game payment matrix. , which represents the benefit value of player i under strategy combination j.

[0047] Based on the game payoff matrix, a replication dynamic equation is constructed to analyze the stability of the dynamic changes in strategy proportions, and the steady-state distribution of the relative benefit evolution of strategy proportions is obtained. The dominant strategy screening analysis is performed on the steady-state distribution of the relative benefit evolution of strategy proportions, and a global convergent evolutionary stable strategy is obtained. The replication dynamic equation includes: ,in represents the proportion of strategy k, , represents the average benefit of strategy k, , represents all elements in set k The average value of represents the overall average benefit, , It represents the steady-state distribution of the relative benefit evolution of the strategy proportion, and the dominant strategy screening analysis: if the benefit of k > the average benefit → more units turn to strategy k, if the benefit of k < the average benefit → units abandon strategy k, when the proportion of all strategies no longer changes , the system reaches a stable state, at which point the strategy with the highest return takes over the entire system (accounting for 100%) and becomes the globally convergent evolutionary stable strategy a*. .

[0048] Based on the global convergence evolutionary stability strategy, the fuzzy membership function mapping is performed to obtain the cooling membership. The specific fuzzy membership function is: , Indicates cooling membership.

[0049] Specifically, a replication dynamic equation is constructed based on the game payoff matrix. The stability of the dynamic changes in strategy proportions is analyzed to obtain the steady-state distribution of the relative benefits of the strategy proportions. From this distribution, globally convergent and stable strategies are selected. Fuzzy membership functions are then applied to these strategies to obtain the cooling membership. The cooling membership can be used together with the dynamic phase change latent heat flow and maximum heat flux density to arbitrate the cooling demand and determine the final cooling demand, providing a basis for subsequent calculations of water pump flow and fan speed.

[0050] Specifically, the detailed analysis of step S400 is as follows: By measuring the battery surface temperature, an implicit phase change criterion model is constructed, and a step-excitation piecewise integral operation is performed to obtain the phase change accumulation. The specific implicit phase change criterion model includes: ,in Indicates the battery surface temperature measured by the infrared sensor. Represents the phase change temperature of PCM (phase change material), which is pre-calibrated by thermal analysis experiments of PCM (such as differential scanning calorimetry DSC) and used as a fixed input in the model. represents a step function, represents the cumulative amount of phase change.

[0051] The dynamic phase change latent heat flow is obtained by quantifying the differential heat absorption rate, scaling the physical properties and performing nonlinear correction on the phase change cumulant. ,in represents the PCM density, The latent heat of phase change is obtained by measuring the peak value of the net heat absorbed by the PCM in the phase change temperature range through differential scanning calorimetry (DSC) experiments. It is obtained by integrating the endothermic peak area of ​​the DSC curve and dividing it by the sample mass. Tanh represents the hyperbolic tangent function. represents the phase transition activation threshold.

[0052] The dynamic phase change latent heat flux, cooling membership and maximum heat flux density are collaboratively compared and the cooling demand is arbitrated to obtain the arbitration cooling capacity, which is specifically: ,in represents arbitration cooling capacity, final cooling demand, α represents latent heat sufficiency coefficient, It represents the latent heat conversion coefficient, and IF-THEN-ELSE is a conditional statement.

[0053] The battery surface temperature is measured using an infrared sensor. An implicit phase change criterion model is used to perform a step-excitation piecewise integral operation to obtain the phase change cumulative value. This phase change cumulative value is then quantified using the differential heat absorption rate and physical property scaling. The phase change state is then nonlinearly corrected to obtain the dynamic phase change latent heat flow. The dynamic phase change latent heat flow, cooling membership, and maximum heat flux density are then compared and compared together. Conditional statement arbitration is used to determine the arbitration cooling capacity. This arbitration cooling capacity is a key parameter for subsequent calculations of water pump flow and fan speed, determining the required cooling capacity of the cooling system.

[0054] Specifically, the detailed analysis of step S500 is as follows: The temperature range is divided into normal, slightly hot, and overheated discrete intervals. Then, for each sensor and each discrete interval, the probability of associating the sensor's measured temperature with the interval median is calculated to obtain the support of each sensor for each interval.

[0055] The weight is defined as 1 when the failure rate is 0%. The weight of each temperature sensor is obtained through the mapping relationship of actual historical failure rate statistical data (specifically, the weight decreases by 0.4 for every 10% increase in the failure rate. If the failure rate is greater than 25%, the weight is 0). Based on the weight of each temperature sensor and the support of a single sensor for each interval, the support of all sensors for each interval is fused using the DS rule to obtain the fusion probability.

[0056] Based on the fusion probability, the credible interval is determined through the argmax function. Through the indicator function, only the sensor temperature values ​​that fall within the credible interval are obtained, and the weighted average is performed according to the sensor weight to obtain the credible temperature.

[0057] Specifically, the discrete intervals of normal, slightly hot, and overheat are A1: normal (20-45℃), A2: slightly hot (45-60℃), and A3: overheat (>60℃). The support degree of each sensor for each interval is ,in k pairs of discrete intervals for sensors (such as normal, slightly hot, overheated) support, is the real-time temperature value measured by sensor k, Temperature range proposition The median of the interval (e.g., the normal interval is 20-45°C, where the median is the middle temperature of the interval), σ is the Gaussian distribution parameter, which is used to control the degree of influence of the deviation between the temperature value and the median of the interval when allocating the probability, and determines the "width" of the membership function. i and j are indexes used to traverse the temperature interval proposition, such as i and j represent different temperature intervals (normal, slightly hot, overheated), which facilitates the calculation of the fusion and conflict of different intervals, and the fusion probability. ,in Represents the weight of sensor k, the weight of each temperature sensor obtained by the above analysis, Represents the multiplication operator, which performs multiplication operations on the items corresponding to all sensors k, and is used to calculate the joint probability of multi-sensor evidence, etc. It represents the sum of the items corresponding to different temperature interval proposition combinations (i and j are different), which is used to handle conflicting evidence and other calculations, and the credible interval , where argmax represents the independent variable when the function is at its maximum, and the interval that maximizes the fusion probability is taken, and the credible temperature ,in Represents the indicator function. When the measured temperature of sensor k falls within the credible interval A*, the function value is 1; otherwise, it is 0. It is used to filter the sensor data involved in the credible temperature calculation. Represents the credible temperature, which is the temperature obtained by weighted average of the sensor temperature values ​​falling within the credible interval A* according to the sensor weight, as the credible temperature result for the final judgment.

[0058] Get the water pump current and fan speed to amplify the composite fault in the form of product to obtain the health attenuation index, perform nonlinear compression on the health attenuation index to obtain the compressed health attenuation index, perform linear reverse mapping on the compressed health attenuation index to obtain the actuator health, specifically, the health attenuation index , where λ represents the health decay index, capturing the anomalies of current and speed at the same time, Indicates the actual current of the water pump, measured by the current sensor, Indicates the rated current of the pump. Indicates the actual fan speed. Indicates the fan set speed and health decay index after compression , min means comparing the size of 1 and λ and taking the smaller value. The health of the actuator is h=1- , mapping relationship, where 0 indicates complete failure and 1 indicates complete health.

[0059] Based on the trusted temperature, the temperature rise deviation is calculated. The arbitration cooling capacity, actuator health and temperature rise deviation are used to obtain the water pump flow through the restriction operation. The water pump flow, arbitration cooling capacity and temperature rise deviation are mapped nonlinearly to obtain the fan speed. Specifically, the temperature rise deviation ,in Indicates the set target temperature and water pump flow , where k represents the temperature rise compensation coefficient, and the fan speed ,in Indicates the maximum speed set for the fan. represents the specific heat capacity of the coolant, Indicates coolant density.

[0060] Specifically, the temperature rise deviation is calculated based on the trusted temperature. The arbitration cooling capacity, actuator health, and temperature rise deviation are then combined to determine the water pump flow rate through a constraint operation. A nonlinear cube root mapping is then performed on the water pump flow rate, arbitration cooling capacity, and temperature rise deviation to determine the fan speed. These water pump flow rate and fan speed directly control the cooling system's operation, adjusting the coolant flow rate and fan speed to maintain the battery temperature within a reasonable range, ensuring battery performance and lifespan.

[0061] like Figure 2 As shown, another embodiment, a thermal management control device for a new energy vehicle battery includes an entropy change module, a fault module, a decision module, a phase change module, and a scheduling module.

[0062] The entropy change module analyzes the entropy change intensity, relaxation thermal spectrum, and nuclear thermal flow field of the collected current and voltage signals to obtain the aging thermal topology; The fault module constructs an anisotropic diffusion equation based on the aging thermal topology and the fault diffusion tensor, solves it to obtain the risk temperature field, and then scans and maps it to obtain the safety margin field; The decision module discretizes the safety margin field into strategy sets, constructs the cooling Nash game, analyzes the payoff matrix and the replication dynamic equation, and then maps the stable strategy to obtain the cooling membership. The phase change module constructs an implicit phase change criterion model for the battery surface temperature to obtain the phase change accumulation. After analyzing the latent heat of the phase change accumulation, it coordinates arbitration with the cooling membership to obtain the arbitration cooling capacity. The scheduling module obtains the pump flow rate through limiting calculation based on the arbitration cooling capacity, the deviation between the measured temperature and the target temperature, and then performs cube root mapping of the pump flow rate and the arbitration cooling capacity to obtain the fan speed.

[0063] The above is an illustration of the present invention and should not be considered as limiting thereof. Although several exemplary embodiments of the present invention have been described, it will be readily understood by those skilled in the art that many modifications may be made to the exemplary embodiments without departing from the novel teachings and advantages of the present invention. Therefore, all such modifications are intended to be included within the scope of the present invention as defined by the claims. It should be understood that the above is an illustration of the present invention and should not be considered as being limited to the specific embodiments disclosed, and modifications to the disclosed embodiments and other embodiments are intended to be included within the scope of the appended claims. The present invention is defined by the claims and their equivalents.

Claims

1. A thermal management control method for a new energy vehicle battery, characterized in that: The following steps are involved: Step S100: performing entropy change intensity, relaxation thermal spectrum, and nuclear thermal flow field analysis on the collected current and voltage signals to obtain aging thermal topology; Step S200: constructing an anisotropic diffusion equation based on the aging thermal topology and the fault diffusion tensor, solving it to obtain the risk temperature field, and then scanning and mapping it to obtain the safety margin field; Step S300: Discretize the safety margin field into a set of strategies, construct a cooling Nash game, analyze the payoff matrix and the replication dynamic equation, screen the stable strategies, and then map them to obtain the cooling membership; Step S400: constructing an implicit phase change criterion model for the battery surface temperature to obtain the phase change accumulation, analyzing the latent heat of the phase change accumulation and arbitrating it with the cooling membership to obtain the arbitration cooling capacity; Step S500: Based on the arbitration cooling capacity and the deviation between the measured temperature and the target temperature, a water pump flow rate is obtained through a limiting operation, and then a cube root mapping is performed on the water pump flow rate and the arbitration cooling capacity to obtain a fan speed.

2. The thermal management control method for a new energy vehicle battery according to claim 1, characterized in that: The steps of the aging thermal topology analysis are as follows: Perform Fourier transform on the relaxation thermal spectrum, extract the inherent frequency variable in the Fourier transform, and obtain the heat source fluctuation angular frequency. Based on the relaxation time constant, calculate the harmonic mean to obtain the comprehensive time scale of the dominant relaxation process. Perform numerical conversion on the comprehensive time scale of the dominant relaxation process to obtain the smoothing characteristic bandwidth coefficient. Dynamically coordinate the calculation and parameter coupling configuration of the heat source fluctuation angular frequency and smoothing characteristic bandwidth coefficient to obtain the frequency domain filter kernel function parameter set; The frequency domain filter kernel function parameter set and the relaxation thermal spectrum are analyzed in the frequency domain. The nuclear thermal flow field is obtained by applying the convolution kernel operator in the frequency domain to perform point multiplication filtering and space-time inversion transformation. Perform Laplace transform on the nuclear thermal flow field to obtain complex frequency variables. Establish an embedded capacity attenuation model for battery discharge and perform multiplication coupling between the temperature compensation function and the Coulomb integral to obtain the health status value. Based on the linear multiplication of the health status value and the aging sensitivity coefficient, the aging time constant is obtained. The degradation transfer analysis of the complex frequency variable and the aging time constant is carried out through the first-order inertia link model to obtain the degradation transfer function. The Laplace transform result of the nuclear thermal flow field is multiplied by the degradation transfer function in the frequency domain, and then reconstructed through the inverse Laplace transform to obtain the aging thermal topology.

3. The thermal management control method for a new energy vehicle battery according to claim 2, characterized in that: The relaxation thermogram analysis steps are as follows: By analyzing the voltage-to-current change rate through the real-time acquisition of current and voltage signals, the dynamic equivalent resistance is obtained. The entropy change intensity is obtained by coupling the temperature gradient and entropy change factor of the dynamic equivalent resistance. The relaxation characteristic coefficient is obtained by performing standardized electrochemical impedance spectroscopy calibration analysis on the battery; The entropy change intensity is subjected to Prony spectrum decomposition algorithm and numerical iterative separation to obtain the relaxation amplitude coefficient; The relaxation time constant is obtained by performing algebraic division operation on the relaxation characteristic coefficient and the absolute value of the entropy change intensity. The relaxation amplitude coefficient and the relaxation time constant are superimposed on each other through an exponential decay model to obtain the relaxation thermal spectrum.

4. The thermal management control method for a new energy vehicle battery according to claim 1, characterized in that: The safety margin field analysis steps are as follows: Based on the fault fingerprint vector, a self-organizing fault mapping algorithm is performed to obtain a fault type code; Based on the fault type code, a preset material characteristic template is called to obtain the fault diffusion tensor; Anisotropic diffusion equations are constructed based on the fault diffusion tensor and aging thermal topology, and the risk temperature field is obtained by solving the equations using the finite element method. Dynamic spatial boundary scanning and critical distance mapping are performed on the risk temperature field to obtain the safety margin field.

5. The thermal management control method for a new energy vehicle battery according to claim 4, characterized in that: The fault fingerprint vector analysis steps are as follows: The discharge current is superimposed with a slightly triangular wave current to obtain a slightly disturbed charge and discharge current. The slightly disturbed charge and discharge current and voltage response signal are subjected to second-order mixed partial derivatives and temperature-normalized integration to obtain the differential conductivity entropy. The instantaneous heat flux density peak value of the aging thermal topology is analyzed to obtain the maximum heat flux density. The heat flux gradient field of the aging thermal topology is analyzed to obtain the heat flux gradient norm. Based on the maximum heat flux density, heat flux gradient norm and differential conductivity entropy, multi-physical field feature vectorization synthesis is performed to obtain the fault fingerprint vector.

6. The thermal management control method for a new energy vehicle battery according to claim 1, characterized in that: The cooling membership analysis steps are as follows: The thermal safety benefit function and the safety margin field are subjected to global strategy traversal processing to dynamically generate the benefit value of each unit under each combination. The unit sequence is used as the row and the cooling strategy combination arrangement is used as the column. The benefit value calculation results of the unit in all strategy combinations are integrated into the game payoff matrix. Based on the game payoff matrix, a replication dynamic equation is constructed to analyze the stability of the dynamic changes in the strategy proportions, and the steady-state distribution of the relative benefit evolution of the strategy proportions is obtained. The dominant strategy screening analysis is performed on the steady-state distribution of the relative benefit evolution of the strategy proportions, and a global convergent evolutionary stable strategy is obtained. Based on the global convergence evolutionary stability strategy, fuzzy membership function mapping is performed to obtain the cooling membership.

7. The thermal management control method for a new energy vehicle battery according to claim 6, characterized in that: The thermal safety benefit function analysis steps are as follows: The safety margin field is discretized into N units of safety margin values, where N corresponds to the total number of battery module units evaluated in the safety margin field. Safety margin intervals are preset. If the unit safety margin values ​​are in different safety margin intervals and correspond to high cooling capacity, medium cooling capacity, and low cooling capacity, respectively, a strategy set is obtained through analysis. Based on the safety margin field and strategy set, spatial security situation mapping is performed to construct a cooling Nash game; In the cooling Nash game framework, the safety distance and average safety state of the battery cells are analyzed based on the safety margin field, and the thermal safety benefit function is constructed in combination with the selection cost of the cooling strategy.

8. The thermal management control method for a new energy vehicle battery according to claim 1, characterized in that: The arbitration cooling capacity analysis steps are as follows: An implicit phase change criterion model is constructed by measuring the battery surface temperature, and a step-excitation piecewise integration operation is performed to obtain the phase change accumulation. The dynamic phase change latent heat flow is obtained by quantifying the differential heat absorption rate, scaling the physical properties and performing nonlinear correction on the phase change cumulant. The dynamic phase change latent heat flux, cooling membership and maximum heat flux density are collaboratively compared and the cooling demand is arbitrated to obtain the arbitration cooling capacity.

9. The thermal management control method for a new energy vehicle battery according to claim 1, characterized in that: The fan speed analysis steps are as follows: The temperature range is divided into normal, slightly hot, and overheated discrete intervals. For each sensor and each discrete interval, the probability of associating the sensor's measured temperature with the interval median is calculated to obtain the support of each sensor for each interval. Define the weight value when no failure rate occurs. Obtain the weight of each temperature sensor through the mapping relationship of actual historical failure rate statistical data. Based on the weight of each temperature sensor and the support of a single sensor for each interval, use the DS rule to fuse the support of all sensors for each interval to obtain the fusion probability. Based on the fusion probability, the credible interval is determined by the argmax function. The indicator function is used to obtain only the sensor temperature values ​​that fall within the credible interval. The weighted average is performed according to the sensor weight to obtain the credible temperature. The water pump current and fan speed are obtained and multiplied to amplify the composite fault to obtain the health attenuation index. The health attenuation index is nonlinearly compressed to obtain the compressed health attenuation index. The compressed health attenuation index is linearly reverse mapped to obtain the actuator health. The temperature rise deviation is calculated based on the trusted temperature. The water pump flow is obtained through a restricted operation on the arbitration cooling capacity, actuator health, and temperature rise deviation. The fan speed is obtained by performing a nonlinear cube root mapping on the water pump flow, arbitration cooling capacity, and temperature rise deviation.

10. A thermal management control device for a new energy vehicle battery, characterized in that A thermal management control method for a new energy vehicle battery according to any one of claims 1 to 9 is implemented, comprising: an entropy change module, a fault module, a decision module, a phase change module, and a scheduling module; The entropy change module analyzes the entropy change intensity, relaxation thermal spectrum, and nuclear thermal flow field of the collected current and voltage signals to obtain the aging thermal topology; The fault module constructs an anisotropic diffusion equation based on the aging thermal topology and the fault diffusion tensor, solves it to obtain the risk temperature field, and then scans and maps it to obtain the safety margin field; The decision module discretizes the safety margin field into strategy sets, constructs the cooling Nash game, analyzes the payoff matrix and the replication dynamic equation, selects the stable strategy and then maps it to obtain the cooling membership. The phase change module constructs an implicit phase change criterion model for the battery surface temperature to obtain the phase change accumulation. After analyzing the latent heat of the phase change accumulation, it coordinates arbitration with the cooling membership to obtain the arbitration cooling capacity. The scheduling module obtains the pump flow rate through limiting calculation based on the arbitration cooling capacity, the deviation between the measured temperature and the target temperature, and then performs cube root mapping of the pump flow rate and the arbitration cooling capacity to obtain the fan speed.

Images