New energy automobile battery thermal management method and system based on big data

Abstract

The invention provides a new energy automobile battery thermal management method and system based on big data. The method comprises the steps that real-time data flow and a historical temperature change curve in a battery pack are collected in real time through a distributed sensor array; inputting a pre-trained time sequence prediction model, outputting a predicted temperature change curve and calculating a real-time temperature rise slope; obtaining an environment comprehensive compensation amount according to the environment temperature and the battery health state, and subtracting the environment comprehensive compensation amount from the basic safety threshold value to obtain a dynamic safety threshold value; determining a dynamic temperature compensation amount through a preset slope grading mechanism, and subtracting the dynamic temperature compensation amount from the dynamic safety threshold to obtain an advanced intervention temperature point; and when the temperature of the battery pack reaches the advanced intervention temperature point, a graded cooling system is started, and cooling power grades are dynamically switched according to the growth interval where the real-time temperature rise slope is located. The battery temperature is accurately controlled, and the energy consumption is remarkably reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer, in particular to a new energy vehicle battery thermal management method and system based on big data. BACKGROUND

[0002] With the rapid development of new energy vehicle industry, as the core energy storage component, the thermal safety management of battery is directly related to the vehicle range, battery life and driving safety; under high temperature environment or large rate charging and discharging conditions, irreversible side reaction chain reaction will be triggered in the battery, including accelerated decomposition of solid electrolyte interface film, electrolyte oxidation and positive material lattice collapse, these reactions not only cause battery capacity to drop and internal resistance to increase, but also may trigger thermal runaway and cause safety accidents; especially in the summer in southern China, desert areas and fast charging scenarios, battery pack temperature management faces severe challenges, and a more intelligent thermal management mechanism needs to be built to cope with complex conditions.

[0003] The current mainstream battery thermal management system generally uses a fixed temperature threshold to trigger the cooling strategy, and when the battery temperature reaches the preset threshold (usually 45-50℃), the cooling device is started. This responsive mechanism has the following defects: first, due to the thermal inertia of the battery and the delay of the cooling system, the temperature will continue to rise for several minutes after reaching the threshold, causing the battery to be exposed to high temperature environment above the safety threshold for a long time, causing irreversible electrochemical damage; second, the fixed threshold does not consider the influence of dynamic factors such as environmental temperature, battery aging degree and driving load, for example, in a 40℃ environment, the actual safety threshold needs to be lowered to below 42℃, but the existing system cannot be adjusted adaptively, resulting in protection failure; third, the cooling strategy is single, and full load refrigeration is used regardless of the temperature rise rate, resulting in low energy efficiency. SUMMARY

[0004] In order to solve the problems existing in the prior art, the present application provides a new energy vehicle battery thermal management method and system based on big data, which solves the irreversible damage problem of new energy vehicle battery caused by fixed temperature threshold triggering lag under high temperature environment, realizes accurate control of battery temperature and significantly reduces energy consumption.

[0005] In the first aspect, the present application provides a new energy vehicle battery thermal management method based on big data, comprising: The temperature distribution data of each single battery cell in the battery pack, the charging and discharging current and voltage data, the environmental temperature and humidity data and the vehicle driving condition data are collected in real time by a distributed sensor array to obtain a real-time data stream, and based on the current spatiotemporal characteristic parameters, the historical temperature change curve under similar scenarios is retrieved from the cloud historical database; input the real-time data stream and the historical temperature change curve into a pre-trained time series prediction model, output a predicted temperature change curve within a future set time window through multi-source data fusion processing, and calculate a real-time temperature rise slope based on the predicted temperature change curve; According to the negative compensation effect of the ambient temperature on the basic safety threshold and the nonlinear attenuation effect of the battery health state on the basic safety threshold, an environmental comprehensive compensation quantity is obtained, the environmental comprehensive compensation quantity monotonically increases with the increase of the ambient temperature and accelerates the growth with the aggravation of the battery aging degree, the basic safety threshold is subtracted by the environmental comprehensive compensation quantity, and a dynamic safety threshold is obtained; Based on the real-time temperature rise slope, a dynamic temperature compensation quantity is determined through a preset slope grading mechanism, the dynamic temperature compensation quantity increases in steps with the increase of the slope, the dynamic safety threshold is subtracted by the dynamic temperature compensation quantity, and an early intervention temperature point is obtained. When the battery pack temperature reaches the early intervention temperature point, the hierarchical cooling system is started, and the cooling power level is dynamically switched according to the growth interval of the real-time temperature rise slope, wherein the cooling intensity is positively correlated with the slope.

[0006] In a second aspect, the application further provides a new energy vehicle battery thermal management system based on big data, which is applied to the new energy vehicle battery thermal management method based on big data in the first aspect; the new energy vehicle battery thermal management system based on big data comprises: A data acquisition and retrieval module acquires temperature distribution data of each single battery cell in the battery pack, charge and discharge current and voltage data, ambient temperature and humidity data, and vehicle driving condition data in real time through a distributed sensor array, obtains a real-time data stream, and retrieves a historical temperature change curve under a similar scene from a cloud historical database based on current spatiotemporal characteristic parameters; A temperature prediction and slope generation module inputs the real-time data stream and the historical temperature change curve into a pre-trained time series prediction model, outputs a predicted temperature change curve within a future set time window through multi-source data fusion processing, and calculates a real-time temperature rise slope based on the predicted temperature change curve; A dynamic safety threshold generation module obtains an environmental comprehensive compensation quantity according to the negative compensation effect of the ambient temperature on the basic safety threshold and the nonlinear attenuation effect of the battery health state on the basic safety threshold, the environmental comprehensive compensation quantity monotonically increases with the increase of the ambient temperature and accelerates the growth with the aggravation of the battery aging degree, the basic safety threshold is subtracted by the environmental comprehensive compensation quantity, and a dynamic safety threshold is obtained; An early intervention point decision module determines a dynamic temperature compensation quantity based on the real-time temperature rise slope through a preset slope grading mechanism, the dynamic temperature compensation quantity increases in steps with the increase of the slope, the dynamic safety threshold is subtracted by the dynamic temperature compensation quantity, and an early intervention temperature point is obtained. The staged cooling execution module starts the staged cooling system when the battery pack temperature reaches the early intervention temperature point, and dynamically switches the cooling power level according to the growth interval of the real-time temperature rising slope, wherein the cooling intensity is positively correlated with the slope.

[0007] The new energy vehicle battery thermal management method based on big data provided by the embodiment of the application can accurately generate a future temperature change curve and a dynamic safety threshold by collecting environmental parameters, driving conditions and battery state data in real time, combining a historical database to train a time series prediction model, and calculating an early intervention temperature point based on a real-time slope of the prediction curve. The staged cooling is started before the temperature reaches the safety threshold, the temperature overshoot phenomenon is eliminated, the cooling intensity is dynamically adjusted according to the slope change, the battery temperature is always kept below the dynamic threshold, and the energy consumption is significantly reduced. The scheme realizes the technical leap from passive response to active prevention, and achieves a breakthrough in high-temperature environment adaptability, battery life protection and energy utilization efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0008] Figure 1 is a flowchart of a new energy vehicle battery thermal management method based on big data provided by the embodiment of the application; Figure 2 is a structural diagram of a new energy vehicle battery thermal management method based on big data provided by the embodiment of the application. DETAILED DESCRIPTION

[0009] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of the application.

[0010] In the description of the application, the terms "first", "second" are used only for the purpose of description, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined with "first", "second" can explicitly or implicitly include one or more features. In the description of the application, the meaning of "multiple" is two or more, unless otherwise specifically limited.

[0011] In the description of the present application, the term "for example" is used to indicate "serving as an example, instance, or illustration." Any embodiment described as "for example" in this application is not necessarily to be construed as preferred or advantageous over other embodiments. The following description is presented to enable any person skilled in the art to practice the present application as claimed. In the following description, details are set forth in order to provide a thorough understanding of the present application. It will be apparent to one ordinarily skilled in the art that the present application can be practiced without using these specific details. In other instances, well-known structures and processes have not been elaborated in order to avoid unnecessary detail, which might obscure the description of the present application. Accordingly, the present application is not intended to be limited to the embodiments shown, but is to be accorded the widest scope consistent with the principles and features disclosed herein.

[0012] Reference Figure 1 , Figure 1 is a flowchart of a new energy vehicle battery thermal management method based on big data provided by the present application, a new energy vehicle battery thermal management method based on big data, comprising: Step 10, real-time acquisition of temperature distribution data of each single battery cell in the battery pack, charge and discharge current and voltage data, environmental temperature and humidity data, and vehicle driving condition data by a distributed sensor array, obtaining a real-time data stream, based on current spatiotemporal characteristic parameters, retrieving a historical temperature change curve under a similar scene from a cloud historical database; Step 20, inputting the real-time data stream and the historical temperature change curve into a pre-trained time series prediction model, outputting a predicted temperature change curve within a future set time window through multi-source data fusion processing, and calculating a real-time temperature rising slope based on the predicted temperature change curve; Step 30, according to the negative compensation effect of environmental temperature on the basic safety threshold and the nonlinear attenuation effect of battery health state on the basic safety threshold, obtaining an environmental comprehensive compensation amount, the environmental comprehensive compensation amount monotonically increases with the increase of environmental temperature, and accelerates the growth with the aggravation of battery aging degree, subtracting the environmental comprehensive compensation amount from the basic safety threshold to obtain a dynamic safety threshold; Step 40, based on the real-time temperature rising slope, determining a dynamic temperature compensation amount through a pre-set slope grading mechanism, the dynamic temperature compensation amount increases in steps with the increase of the slope, subtracting the dynamic temperature compensation amount from the dynamic safety threshold to obtain an early intervention temperature point; Step 50, when the battery pack temperature reaches the early intervention temperature point, starting a hierarchical cooling system, and dynamically switching the cooling power level according to the growth interval of the real-time temperature rising slope, wherein the cooling intensity is positively correlated with the slope.

[0013] Specifically, in step 10, data collection and historical retrieval are implemented through a distributed multi-source sensing network, which includes: At least three groups of temperature sensor arrays are deployed inside the battery pack, each group containing K-type thermocouples arranged in the battery module gap and thin film temperature sensors attached to the surface of the single cell, to collect temperature distribution data at a sampling rate of 10 Hz; A closed-loop Hall current sensor and a high-precision voltage acquisition circuit are installed in the charging and discharging circuit to synchronously acquire current and voltage waveforms at a frequency of 1 kHz; A multi-in-one environmental monitoring unit is integrated on the roof to obtain real-time environmental temperature, humidity, pressure, and light intensity data; Through the vehicle CAN bus interface, broadcast messages of vehicle speed, battery state of health (SOH), and battery state of charge (SOC) are read at a period of 500 ms; When the vehicle starts, the historical retrieval mechanism is automatically triggered, combining the current timestamp, GPS geographic coordinates, and environmental parameters into a 128-dimensional feature vector, which is uploaded to the cloud historical database through the Internet of Vehicles module. The improved Euclidean distance algorithm is used to match similar scenarios, and the three highest matching historical temperature change curve original data are returned.

[0014] In a specific embodiment, the step 10 process is executed when a certain electric vehicle is driving on a highway in the summer afternoon: The temperature sensor array detects that the middle module temperature in the battery pack reaches 42.3℃, and the edge module is 38.7℃, generating a temperature distribution thermal map; The current sensor captures a continuous discharge current of 150A, and the voltage fluctuation range is 350-365V; The roof environmental unit measures the external air temperature of 38℃ and humidity of 65%; The CAN bus analysis shows that the vehicle speed is 110 km / h, SOH=88%, and SOC=45%; The system automatically constructs a feature vector and uploads it to the cloud through the 5G network; The cloud database retrieves data for the same period and section in 2020-2025, matches three similar curves: ① historical curve with environmental temperature of 37℃ at the same speed on July 12, 2024; ② curve under the same SOC on August 3, 2023; ③ extreme temperature curve on the historical high-temperature warning date, and downloads the original time-temperature data stream of the three curves to the vehicle cache.

[0015] In the specific implementation process, not only data collection and retrieval are performed, but also the following matters need to be paid attention to, including: Temperature sensors undergo zero-point calibration every quarter, and current sensors employ dynamic calibration compensation to ensure that data errors are less than industry standard tolerances. When GPS signal is lost, switch to base station positioning and inertial navigation fusion mode. When environmental parameters are abnormal, activate vehicle-mounted weather radar to supplement data and ensure data integrity. In the historical data filtering rules, historical curves with differences in battery health status greater than 15% or deviations in charge / discharge rate exceeding 0.5C are excluded, and data of the same model battery pack are selected first. Set up a safety redundancy mechanism. If cloud communication is interrupted, call up the historical data of the most recent 30 days stored locally. When there are fewer than two available data points, use the simulation curve based on the electrochemical model as a substitute. To protect privacy, geolocation information is rasterized and blurred before transmission, reducing accuracy to 1km. 1km grid, timestamps blurred to 30-minute intervals.

[0016] Specifically, in step 20, temperature prediction and slope calculation are achieved through an embedded LSTM model, and the real-time data stream output from step 10 is combined with three historical temperature curves. , , Input to a pre-trained multi-head attention LSTM network; real-time data stream includes a temperature distribution matrix. Current ,Voltage Ambient temperature Speed ;in The first similar scene temperature-time curve (such as the temperature change curve over time for the scene of "high-speed driving in the summer afternoon"). The second similar scene temperature-time curve (and) (The scene type is the same, but the collection time or working conditions are slightly different). The third similar scene temperature-time curve (used to improve the redundancy of data fusion and prediction accuracy). I Battery charging and discharging current (unit: A, positive value during charging, negative value during discharging, reflecting the energy flow intensity of the battery). V: Battery cell voltage (unit: V, a basic parameter reflecting the current state of charge and health of the battery). Ambient temperature (unit: °C, a key external parameter affecting battery heat dissipation efficiency, used for environmental compensation calculation of dynamic threshold). The model normalizes the input data into a 12-dimensional feature vector through a feature fusion layer wherein Tavg is the average temperature of the battery pack, Tstd is the temperature standard deviation, Tdev is the deviation of the historical curve from the current temperature; the LSTM unit is calculated by time step expansion, including: forget gate:

[0017] generates a value between 0 and 1, determining the proportion of historical information retained (for example, when the ambient temperature changes suddenly, the weight of the historical temperature rise trend is reduced); input gate:

[0018] controls the degree of adoption of new input information (such as when a sudden large current discharge occurs, the weight of the current data is increased); candidate state:

[0019] generates the latent state at the current time, representing the short-term thermal dynamic characteristics (such as the acceleration of instantaneous temperature rise); cell state update:

[0020] fuses long-term regularity and short-term characteristics to form updated knowledge memory (for example, combining the battery aging baseline with the temperature rise pattern under new working conditions); output gate:

[0021] filters the information that needs to be output at the current time (such as filtering the effective temperature rise signal after filtering sensor noise); hidden layer output:

[0022] generates the final prediction basis, which is passed to the next time step or the fully connected layer output temperature prediction value; wherein : hidden state of the previous time step, carrying the evolution law of the historical temperature sequence (such as the temperature rise trend of the battery in the previous minute); : feature vector of the input time series prediction model at time t (core parameter set integrating real-time data and scenario-based historical data); : cell state of the previous time step, as a long-term memory carrier to store the essential laws of battery thermal characteristics (such as the heat capacity characteristics of battery materials); : weight matrices of forget gate, input gate, candidate state, and output gate, obtained through training and learning, used to extract the contribution weight of different input features (such as the influence coefficient of current on temperature rise); : Bias vector corresponding to the above gating, used to adjust the threshold of the activation function, enhancing the model fitting ability; : Sigmoid activation function, compressing the input to the interval (0, 1), realizing the gating effect; : Hyperbolic tangent activation function, normalizing the input to the interval (-1, 1), stabilizing the gradient flow; : Hadamard product (element-wise multiplication), used for local modulation of the gating signal and state vector; T : Battery pack average temperature (unit: ℃, reflecting the overall temperature level of the battery pack); : Temperature standard deviation (unit: ℃, reflecting the temperature dispersion of each monomer in the battery pack, used to identify local temperature difference); S: Vehicle speed (unit: km / h, reflecting the vehicle driving condition, directly related to battery discharge power and heat generation); : First historical curve : Deviation value from current temperature (unit: ℃, used to correct the prediction deviation of real-time data); : Second historical curve : Deviation value from current temperature (unit: ℃, same as , improving redundancy); : Third historical curve : Deviation value from current temperature (unit: ℃, same as , improving redundancy); : Finally output the future 10-minute temperature sequence through the full connection layer , and calculate the real-time slope by first-order difference , the model iteratively updates the prediction result on the vehicle GPU at a frequency of 5Hz.

[0023] In a specific embodiment, the step 20 process is executed with a certain electric SUV climbing at a speed of 100 km / h at an ambient temperature of 40℃, and the input data includes: : Temperature distribution matrix (center module 46.2℃ / edge module 41.3℃); discharge current 182A; voltage fluctuation 352-367V; three historical curves (2022 same section curve , same model high temperature curve , SOH=85% aging curve ); : Feature vector .

[0024] The LSTM network computation process is as follows: Time step Forgotten Gate 87% of the historical state has been preserved; Candidate state The calculated temperature rise trend index is +0.35; Output gate Generate the first prediction point ; After 120 consecutive iterations (10-minute window), the predicted curve is output: it reaches a peak of 47.8℃ in the 3rd minute and drops to 45.2℃ in the 8th minute. The slope calculation module detects this. hour This triggers the high-growth phase indicator.

[0025] In the specific implementation process, the following should also be noted: To achieve lightweight model constraints, the LSTM hidden layer dimension is ≤128, and the floating-point operation volume is <15G FLOPS, meeting the computing power limitations of automotive chips. Real-time performance is guaranteed, with a single prediction period of ≤200ms. If the timeout occurs, cached data will be activated and an alarm will be triggered. Perform confidence verification when the extreme point of the predicted curve is reached. With current temperature deviation When the battery safety tolerance is set to 30%, a degradation strategy is initiated, which includes: like Using weighted average ; like Switch to electrochemical model .

[0026] Perform slope filtering to adjust the original slope. Apply Kalman filter The process noise covariance The observation noise covariance R=0.1 is used to prevent abrupt changes in the slope.

[0027] in: : Represents the filtered slope estimate at the current moment, which is the actual temperature rise rate after noise suppression and is used for cooling strategy decision-making; : Represents the filtered slope estimate of the previous moment, which is a smooth record of the historical temperature rise trend and reflects the continuous influence of battery thermal inertia; : the original slope observation value at the current time, which is the instantaneous slope obtained by fitting the predicted temperature curve by the least square method; : state transition matrix (scalar form), which represents the time correlation coefficient of the slope and reflects the inertia degree of the temperature rise trend; : Kalman gain (scalar form), which represents the trust weight of the observation value and balances the contribution proportion of the historical estimate and the current observation.

[0028] Specifically, in step 30, a dynamic safety threshold is generated by a three-stage linkage mechanism, which is specifically divided into steps 301, ambient temperature compensation amount calculation; step 302, battery aging compensation amount calculation; and step 303, dynamic safety threshold synthesis.

[0029] Further, the step 301 collects the ambient temperature in real time through a high-precision environment sensor (collecting rate ), and inputs the segmented linear compensation engine to perform a three-order judgment, which includes: performing a reference temperature judgment, calling the best working temperature median value in the battery material database (such as NMC battery 25℃), when , it is judged that the temperature has no thermal risk, and the ambient temperature compensation amount is output; performing a regular temperature rise compensation judgment, when entering the first high temperature interval ), based on the unit temperature rise side reaction acceleration rate (0.6-0.8℃ / ℃ calibrated by isothermal calorimeter), the is calculated, where is the material characteristic coefficient; performing a high-temperature strengthening compensation judgment, when (the side reaction mutation point, such as 40℃) enters the second high temperature interval, according to the exponential growth characteristic in the Arrhenius equation , a strengthening factor is used to calculate .

[0030] The step 302 obtains the state of health SOH (update period ≤24h) through the BMS, inputs the SOH into the double-threshold attenuation model to perform a hierarchical operation, including: when SOH≥ (the capacity first accelerated attenuation critical point, such as 90%) the battery health state is higher than the first health threshold, it is judged that the battery is in a safe service life and belongs to the health state partition, and the battery aging compensation amount is output; When (For example, 80%≤SOH<90%) the battery health state falls between the first health threshold and the second health threshold, it is determined that the battery is in the linear decay area, and the linear compensation is calculated according to the heat generation increment (0.3-0.5W / %) caused by each percentage capacity decay (calibrated by accelerated aging experiment) , wherein is the aging sensitivity coefficient; When SOH (Thermal stability collapse critical point, for example, 80%) the battery health state is lower than the second health threshold, it is determined that the battery is in the exponential decay area, and the exponential parameter is fitted based on the reaction activation energy measured by the aging battery adiabatic calorimeter (ARC) , and the calculation is , wherein is the baseline decay amount.

[0031] The step 303 performs four-step operations of bidirectional compensation operation, boundary constraint, data reliability verification, and real-time early warning by inputting the , , into the synthesis engine.

[0032] , wherein : environmental temperature compensation amount (unit: ℃, calculated based on , the higher the environmental temperature, the greater the compensation amount, which is used to lower the dynamic safety threshold); : battery aging compensation amount (unit: ℃, calculated based on the battery health state SOH, the lower the SOH, the exponentially increasing compensation amount, which is used to adapt to the thermal stability decline characteristics of the aging battery); : dynamic safety threshold (unit: ℃, core control threshold, the calculation formula is , which is dynamically adjusted according to the environment and the battery state); : basic safety threshold (unit: ℃, safety temperature upper limit of the battery in a brand-new state and standard environmental temperature (25℃), which is a fixed value, such as 52℃).

[0033] Specifically, the bidirectional compensation operation is to perform arithmetic superposition of the dynamic safety threshold ; boundary constraint is performed, if the result is lower than the system allowed lower limit (the freezing point of the electrolyte + 10℃ safety margin), the output is forced ; data reliability verification is performed, when the environmental or aging data source is abnormal, the historical mean safety mode is switched to (the mean value of the same scene threshold in the past 30 days is called); real-time early warning is performed when the total compensation amount ​At this time, the high-temperature aging risk warning of the vehicle-mounted system is activated.

[0034] A dynamic safety threshold is generated through a three-stage linkage mechanism to quantify the erosion effect of environmental temperature rise on the thermal stability of the battery, solve the problem of fixed threshold protection failure under high-temperature working conditions, capture the nonlinear thermal risk caused by battery aging, and solve the defect of insufficient protection of fixed threshold for aged batteries. A dynamic safety boundary driven by both environment and aging is established to achieve the essential protection of battery thermal safety.

[0035] In a specific embodiment, the full process of step 30 is performed by taking a certain electric vehicle running in a tropical rainforest area under high load as an example.

[0036] During the implementation of the environmental compensation in step 301, the environmental temperature is collected in real time by the roof environmental monitoring unit = 43.5°C (sliding average per minute), and the reference temperature in the system parameter library is called synchronously = 25°C (NMC811 battery optimal working average), high-temperature critical point = 40°C (DSC measured side reaction mutation point), conventional compensation coefficient = 0.65 (calibration basis: side reaction rate increment 1.8 times / °C in 35-40°C temperature rise zone); since 43.5°C > 40°C, the high-temperature intensive compensation mode is triggered, and the intensive factor is called = 1.75 (according to Arrhenius equation , wherein the activation energy = 85 kJ / mol, and the gas constant); the environmental temperature compensation amount is calculated by executing the intensive compensation formula .

[0037] During the implementation of the aging compensation in step 302, the state of health SOH = 77% is obtained by the BMS (based on the capacity attenuation model calculation), and the health threshold in the parameter library is called = 90% (capacity retention rate inflection point after 500 cycles), = 80% (ARC test heat generation rate mutation point), reference attenuation amount = 2.3 (fitting self-aging battery out-of-control experiment), exponential parameter = 0.55 (calculated according to the enthalpy change of SEI film regeneration reaction , and then normalized); since SOH = 77% < 80%, the attenuation area is determined, and it is determined to enter the exponential attenuation area; the battery aging compensation amount is calculated, and the exponential formula is executed .

[0038] During the implementation of the dynamic synthesis in step 303, the basic safety threshold in the electrochemical database is called = 52℃ (NMC811 initial exothermic temperature 62℃ minus 10℃ safety margin), receiving the output of step 301 and the output of step 302 = 11.98℃; performing a bidirectional compensation operation, executing a synthesis formula to calculate a dynamic safety threshold ; performing a boundary constraint processing, since the synthesis result 18.98℃ is higher than the system lower limit , directly outputting a final threshold; performing a risk warning, an environmental comprehensive compensation amount , activating a high-temperature aging red warning icon of a vehicle-mounted instrument panel.

[0039] In the specific implementation process, attention also needs to be paid to parameter calibration specifications, setting a real-time guarantee mechanism, and fault tolerance design.

[0040] The parameter calibration specifications include The differential scanning calorimetry (DSC) method must be used to measure the side reaction rate mutation point, with an error tolerance 1.5℃; The in-situ measurement of SEI film regeneration reaction enthalpy change needs to be based on a three-electrode system fitting, meeting (R is the gas constant); It is required to be at least 10℃ higher than the freezing point of the electrolyte, and verified by low-temperature differential thermal analysis (DTA).

[0041] The setting of the real-time guarantee mechanism includes synchronization of the environmental compensation period and temperature sampling (≤1 second), and binding of the aging compensation with the BMS health report event; when the fluctuation is >5℃ per minute, starting the sliding window filtering (N is adjusted according to the climate type).

[0042] The fault tolerance design includes switching to V2X vehicle-road cooperation weather data when the environmental sensor fails, using the open circuit voltage (OCV)-capacity mapping model for estimation when the SOH is abnormal, and cross-verification by a three-module redundant processor for synthesis operation, with the output of a majority decision result.

[0043] Specifically, in step 40, an early intervention temperature point is generated by a four-order decision chain, which is specifically divided into step 401, real-time slope extraction; step 402, slope interval determination; step 403, dynamic compensation amount generation; and step 404, intervention point synthesis.

[0044] Further, the step 401 receives a predicted temperature change curve in a future time window by a temperature prediction module , inputs the predicted temperature change curve into a sliding window analysis engine for window cutting, linear fitting, and dynamic optimization.

[0045] The window is captured at the current time. Starting from the beginning, extract a fixed duration. curve segment ( (Based on the battery thermal time constant setting) The linear fitting uses the least squares method to fit the temperature-time relationship within the window. Output the real-time temperature rise slope ; The dynamic optimization involves automatically expanding the window to 1.5 when the prediction confidence level p < 0.9. Improve the stability of the fit.

[0046] In step 402, the real-time temperature rise slope k is input into a three-level classifier to perform interval mapping. when Activate low-growth intervals ( (Based on the first abrupt change point of the side reaction rate of battery materials) when Time-activated growth interval ( (Calibrated based on the rate of temperature rise at the onset of electrolyte boiling). when Activate the high-growth zone in a timely manner; in Real-time temperature rise slope (unit: ℃ / min, reflecting the rate at which the battery temperature rises, and is the core basis for graded compensation and cooling control); Critical slope in the low / medium growth range (unit: °C / min, determined based on the first abrupt change in the side reaction rate of the battery material, e.g., 1 °C / min, slope). (In the low growth range). Critical slope in the medium / high growth range (unit: °C / min, calibrated based on the rate of temperature rise at the electrolyte boiling point, e.g., 2 °C / min, slope). (for high growth range) Dynamic temperature compensation (unit: °C, based on) The growth range is determined; the steeper the slope, the greater the compensation amount, used to calculate the early intervention temperature point and reserve a safety margin. The , The inflection point of reaction kinetics was determined by differential scanning calorimetry (DSC).

[0047] Step 403 calls the compensation mapping rule based on the interval determination result, and outputs when it is in a low growth interval. When in the medium growth range, output When output is in a high-growth range ; Calculate the rate of change of slope within the window ,when Time (vehicle rapid acceleration threshold), plus buffer compensation amount Final compensation amount .

[0048] Step 404 receives the dynamic security threshold. and compensation amount Execute the imposed constraints And apply constraints ( (for system protection margin).

[0049] In one specific embodiment, step 40 is executed as an example of an electric vehicle overtaking at high speed.

[0050] In step 401, during the real-time slope extraction process, the predicted curve and the temperature sequence for the next 5 minutes are input. (Time step 1 minute); Configure the sliding window settings. =3min (capturing three points: 38.9, 41.5, and 44.8℃); perform least squares fitting and solve the equation. We get k = 2.97℃ / min.

[0051] Step 402, slope interval determination, involves calling calibration parameters. ℃ / min (transition point of side reaction in graphite negative electrode) =2.0℃ / min (Electrolyte boiling point); because k=2.97> The assessment results indicate that the economy has entered a high-growth phase.

[0052] Step 403, during the dynamic compensation generation process, involves basic compensation and the activation of high-growth intervals. =8℃; Perform acceleration detection and calculate the slope change of the preceding window. Perform buffer stacking and activate. ℃, to obtain ℃.

[0053] Step 404: During the implementation of intervention point synthesis, input the dynamic safety threshold. ,calculate The result was higher than Perform boundary checks and output directly.

[0054] In the specific implementation process, attention should also be paid to ensuring the robustness of slope calculation, the standardization of interval boundary calibration, the dynamic adjustment mechanism of compensation amount, and the design of fault tolerance.

[0055] The slope calculation robustness guarantee includes: Perform window adaptation, when the prediction curve confidence p < 0.85, the window length is dynamically extended according to Dynamic expansion; Perform outlier removal by filtering outliers using Tukey's rule (IQR is the interquartile range); Perform fitting verification, requiring linear regression decision coefficient , otherwise switch to polynomial fitting.

[0056] The interval boundary calibration specification includes: Determine the negative electrode SEI film decomposition reaction activation energy by cyclic voltammetry (CV) , calculate (R gas constant, T temperature) to Calibration; According to the temperature rise rate of the 5% weight loss point of the electrolyte thermogravimetric analysis (TGA), the error tolerance is 0.2 ℃ / min, to Calibration; Online calibration by rechecking the validity of the boundary value every quarter on the real vehicle platform.

[0057] The compensation amount dynamic adjustment mechanism includes: , , The initial value is set according to the battery thermal runaway time constant, and the running is adaptively adjusted according to the formula The learning rate is , and the recent prediction error mean is to perform basic compensation; Perform buffer compensation, positively related to the absolute value of acceleration, satisfying ; Perform synthesis constraint, the total compensation amount shall not exceed 40% of the dynamic safety threshold.

[0058] The fault tolerance design includes: Predictive failure degradation, when the fitting is for 3 consecutive times, switch to a simplified model based on current integration ( is the battery resistance coefficient); If , perform boundary protection, forcibly raise to and limit power operation; Store key parameters such as slope k, acceleration , compensation amount , save period​ A safety audit is conducted every 10 years.

[0059] Specifically, in step 50, the graded cooling dynamic regulation is executed through a four-level joint control mechanism, which is divided into step 501, intervention point triggering; step 502, cooling mode decision; step 503, operation status maintenance; and step 504, energy saving degradation determination.

[0060] Furthermore, step 501 involves real-time monitoring of the battery pack's highest temperature. When detected At that time, a start command is sent to the cooling system, with a command transmission delay. 50ms; Synchronously read the real-time slope of the predicted temperature change curve .

[0061] Step 502 executes a cooling power level classification strategy based on the real-time slope k of the predicted temperature change curve: like (Low growth range), enters low power mode, activates fan cooling array, power consumption... 300W; like (Medium growth range), enters medium power consumption mode, activates the semiconductor cooling chip assembly, power consumption... 1.5kW; like (High growth range) Entering high power consumption mode, activating the liquid cooling circulation pump and semiconductor refrigeration linkage, power consumption... 3.5kW.

[0062] After entering the high-power mode in step 503, when the slope drops to the medium-growth range, the high-power operation is maintained for a fixed duration. The fixed duration The protection duration is set according to the battery's thermal inertia, and the system switches to medium power consumption mode after the countdown ends.

[0063] Step 504 involves continuous monitoring of the temperature difference. ,when (Safety margin) and actual duration of continuous stability During the (stable duration) period, if the current mode is high / medium power consumption, it will gradually downgrade to low power consumption mode.

[0064] In one specific embodiment, step 50 is performed as an example of an electric sports car being driven on a racetrack.

[0065] During the implementation of the intervention point in step 501, the real-time temperature Reaching the intervention point Immediately send a cooling start command.

[0066] In the step 502, the cooling mode decision implementation process, read the real-time slope = 3.1 ℃ / min, because (calibration value), trigger high-power mode, liquid cooling pump full-speed operation, semiconductor refrigeration piece full load.

[0067] In the step 503, the running state maintenance implementation process, 3 minutes after the slope to = 1.8 ℃ / min (into the middle growth interval), start fixed time = 45 seconds countdown, maintain high-power operation, after the countdown switch to medium-power mode, turn off the liquid cooling pump, semiconductor refrigeration to 70% power.

[0068] In the step 504, the energy-saving degradation determination implementation process, when the temperature difference and last 120 seconds, switch to low-power mode, only keep the fan running.

[0069] In the specific implementation process, also need to pay attention to the mode switching safety guarantee, delay time dynamic calibration, energy-saving degradation robustness design and fault tolerance mechanism.

[0070] The mode switching safety guarantee: Perform power-on constraints, when the slope pre-start the liquid cooling pump, eliminate hydraulic delay; Perform degradation speed limit, medium and low-power mode switching need to be 30 seconds slope power down ( ), prevent temperature rebound; Ensure communication redundancy, control instructions are transmitted through CAN bus and Ethernet dual-channel, packet loss rate is less than 0.001%.

[0071] The delay time dynamic calibration includes: Perform thermal inertia modeling, protection time ( specific heat capacity, m is the mass, allowed temperature rise, effective refrigeration power of the cooling system); Perform adaptive adjustment, when the battery is aging (SOH < 85%), the protection time extended by 20%; ambient temperature > 40℃, the protection time shortened by 15%; Perform boundary protection, the protection time maximum value does not exceed 30% of the battery thermal runaway time constant.

[0072] The energy-saving degradation robustness design includes: Perform safety margin calibration, increase with SOH decrease Stable duration constraint is performed, According to the ambient temperature setting (120 seconds at 25℃, shortened to 60 seconds at 40℃); Abnormal lock is performed, if the slope breaks twice within 10 minutes after degradation , the current mode is forced to lock for 30 minutes.

[0073] The fault tolerance mechanism comprises: When the prediction module fails, the slope degradation is invalid, and the simplified model based on current is switched to (I is the current); A three-way temperature sensor voting mechanism is adopted to ensure temperature monitoring redundancy, and the median is taken when the deviation is greater than 1℃; When the single cell temperature difference is greater than 5℃ or the voltage drop is greater than 10%, an emergency fuse is triggered, and the highest cooling is directly started regardless of the mode.

[0074] The embodiment of the application realizes a fundamental breakthrough in battery thermal safety through a dynamic thermal management system. Its core advantage is to eliminate the temperature hysteresis risk of the traditional fixed threshold mechanism, and to change passive response to active defense. Based on the dual dynamic compensation mechanism of ambient temperature and battery aging state, a precise adaptive safety boundary is established for different working conditions, effectively solving the protection failure problem caused by high temperature environment and battery decay. Through the predictive intervention point and the slope response type hierarchical cooling strategy, the optimal intensity of heat dissipation measures is started before the temperature reaches the critical point, not only completely avoiding the exposure of the battery to dangerous high temperature environment, significantly inhibiting the erosion of irreversible side reactions on the battery life, but also realizing the intelligent matching of cooling energy consumption. Combined with the online calibration of the prediction model and the multi-level fault tolerance design, the system maintains high precision decision and robust operation throughout the life cycle, greatly reduces the maintenance demand, and finally realizes a leap in four dimensions of zero overshoot temperature control, battery life extension, energy efficiency improvement and system reliability, providing an all-weather, all-scenario intrinsic safety solution for new energy vehicles.

[0075] Further, the new energy vehicle battery thermal management system based on big data provided by the application is described below. The new energy vehicle battery thermal management system based on big data described below can be correspondingly referred to the new energy vehicle battery thermal management method based on big data described above. Optionally, referring to Figure 2 , Figure 2 is a structural schematic diagram of the new energy vehicle battery thermal management system based on big data provided by the application. The new energy vehicle battery thermal management system based on big data comprises.

[0076] ​The data acquisition and retrieval module 210 acquires the temperature distribution data of each single battery cell in the battery pack, the charging and discharging current and voltage data, the environmental temperature and humidity data, and the vehicle driving condition data in real time through a distributed sensor array, obtains a real-time data stream, retrieves a historical temperature change curve under a similar scenario from a cloud historical database based on current spatiotemporal characteristic parameters, and retrieves the historical temperature change curve under the similar scenario from the cloud historical database based on the current spatiotemporal characteristic parameters; The temperature prediction and slope generation module 220 inputs the real-time data stream and the historical temperature change curve into a pre-trained time series prediction model, processes multiple source data fusion, outputs a predicted temperature change curve in a future set time window, and calculates a real-time temperature rising slope based on the predicted temperature change curve. The dynamic safety threshold generation module 230 obtains an environmental comprehensive compensation amount according to the negative compensation effect of the environmental temperature on the basic safety threshold and the nonlinear attenuation effect of the battery health state on the basic safety threshold, the environmental comprehensive compensation amount monotonically increases with the increase of the environmental temperature, and the environmental comprehensive compensation amount accelerates growth with the aggravation of the battery aging degree, subtracts the environmental comprehensive compensation amount from the basic safety threshold to obtain a dynamic safety threshold. The early intervention point decision module 240 determines a dynamic temperature compensation amount based on the real-time temperature rising slope through a preset slope grading mechanism, the dynamic temperature compensation amount increases in steps with the increase of the slope, subtracts the dynamic temperature compensation amount from the dynamic safety threshold to obtain an early intervention temperature point. The hierarchical cooling execution module 250 starts the hierarchical cooling system when the battery pack temperature reaches the early intervention temperature point, and dynamically switches the cooling power level according to the growth interval of the real-time temperature rising slope, wherein the cooling intensity is positively correlated with the slope.

[0077] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be realized by means of software and a necessary general hardware platform, and of course, can also be realized by hardware. Based on such understanding, the above technical solutions can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as a ROM / RAM, a magnetic disk, an optical disk, etc., and includes a plurality of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute the method described in each embodiment or some parts of the embodiment.

[0078] It should be pointed out finally that the above embodiments are only used to illustrate the technical solutions of the present application, but not to limit the same; and although the present application has been described in detail with reference to the foregoing embodiments, it should be appreciated by those skilled in the art that the technical solutions recorded in the foregoing embodiments can be modified, or some technical features thereof can be replaced equivalently; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for thermal management of new energy vehicle batteries based on big data, characterized in that, include: The distributed sensor array collects real-time data on temperature distribution, charging and discharging current and voltage, ambient temperature and humidity, and vehicle driving conditions of each individual cell in the battery pack to obtain a real-time data stream. Based on the current spatiotemporal characteristic parameters, the historical temperature change curves under similar scenarios are retrieved from the historical database in the cloud. The real-time data stream and historical temperature change curves are input into a pre-trained time series prediction model. Through multi-source data fusion processing, the predicted temperature change curve within a set future time window is output. The real-time temperature rise slope is calculated based on the predicted temperature change curve. Based on the negative compensation effect of ambient temperature on the basic safety threshold and the nonlinear decay effect of battery health status on the basic safety threshold, the comprehensive environmental compensation amount is obtained. The comprehensive environmental compensation amount increases monotonically with the increase of ambient temperature and accelerates with the increase of battery aging. The comprehensive environmental compensation amount is subtracted from the basic safety threshold to obtain the dynamic safety threshold. Based on the real-time temperature rise slope, a dynamic temperature compensation amount is determined through a preset slope grading mechanism. The critical slope of the slope grading mechanism is calibrated according to the temperature rise rate at the first abrupt change point of the side reaction rate of the battery material and the starting point of electrolyte boiling, and is divided into three growth intervals: low, medium, and high. The dynamic temperature compensation amount increases stepwise with the increase of the slope. The low growth interval corresponds to the first fixed compensation amount, the medium growth interval corresponds to the second fixed compensation amount which is greater than the first fixed amount, and the high growth interval corresponds to the third fixed compensation amount which is greater than the second fixed amount. The dynamic temperature compensation amount is subtracted from the dynamic safety threshold to obtain the early intervention temperature point. When the battery pack temperature reaches the pre-intervention temperature point, the graded cooling system is activated, and the cooling power level is dynamically switched according to the growth range of the real-time temperature rise slope, wherein the cooling intensity is positively correlated with the slope.

2. The method for thermal management of new energy vehicle batteries based on big data according to claim 1, characterized in that, The step of establishing a piecewise linear compensation model based on the negative compensation effect of ambient temperature on the basic safety threshold includes: Compensation is not activated when the ambient temperature is lower than or equal to the reference temperature. When the ambient temperature is in the first high temperature range, a first-level compensation amount is generated that is proportional to the temperature rise. When the ambient temperature enters the second high temperature range, a second-level compensation amount greater than the first-level compensation amount is generated. The reference temperature is set based on the battery's optimal operating temperature range, and the boundary values ​​between the first high-temperature range and the second high-temperature range are calibrated according to the critical ambient temperature at which the battery's side reaction rate changes abruptly.

3. The method for thermal management of new energy vehicle batteries based on big data according to claim 1, characterized in that, The nonlinear decay effect of the battery health state on the basic safety threshold is used to construct a compensation growth model associated with the battery health state, including: When the battery health status is higher than the first health threshold, aging compensation is not activated; When the battery health status drops to between the first health threshold and the second health threshold, the compensation amount increases linearly as the health status decreases. When the battery health status is below the second health threshold, the compensation amount increases exponentially as the health status decreases. The first health threshold is set based on the critical point at which the battery capacity first accelerates its decline, and the second health threshold is calibrated based on the critical point at which the battery thermal stability deteriorates significantly.

4. The method for thermal management of new energy vehicle batteries based on big data according to claim 1, characterized in that, The determination of dynamic temperature compensation through a preset slope grading mechanism includes: In the low growth range, the absolute value of the slope is less than the first critical slope, which corresponds to the normal temperature rise stage of the battery. The first critical slope is calibrated based on the temperature change rate of the first sudden change in the side reaction rate of the battery material. In the medium growth range, the absolute value of the slope is between the first critical slope and the second critical slope, which corresponds to the accelerated stage of the side reaction chain. The second critical slope is calibrated based on the temperature rise rate at the boiling point of the electrolyte. In the high-growth range, the absolute value of the slope is greater than the second critical slope, corresponding to the stage of thermal runaway risk.

5. The method for thermal management of new energy vehicle batteries based on big data according to claim 4, characterized in that, The mapping relationship between the slope grading mechanism and the dynamic temperature compensation amount includes: The low-growth interval maps to the dynamic temperature compensation amount of the first fixed value. The dynamic temperature compensation amount of the second fixed value is mapped to the dynamic temperature compensation amount that is greater than the first fixed value in the medium growth interval. The high-growth interval is mapped to a third fixed value of dynamic temperature compensation that is greater than the second fixed value. When the rate of change of slope is detected to exceed the preset acceleration threshold, a buffer dynamic temperature compensation amount is added on the basis of the corresponding fixed value of dynamic temperature compensation amount. The acceleration threshold is calibrated according to the battery temperature rise characteristics under the rapid acceleration condition of the vehicle.

6. The method for thermal management of new energy vehicle batteries based on big data according to claim 1, characterized in that, The method of dynamically switching the cooling power level based on the growth range of the real-time temperature rise slope includes: The highest cooling power will be activated immediately upon entering a high-growth phase. After exiting the high-growth range, maintain the highest cooling power operation for the set protection duration before downgrading; Once the system enters a low-growth zone and remains stable for longer than the set protection time, the high-power cooling device will be shut down. When the difference between the real-time temperature and the dynamic safety threshold exceeds the safety margin, cooling power degradation is triggered to optimize energy consumption. The protection duration is set based on the battery thermal inertia time constant, and the safety margin is determined based on the system temperature control accuracy requirements.

7. The method for thermal management of new energy vehicle batteries based on big data according to claim 1, characterized in that, The method also includes a safety redundancy control mechanism. When the temperature difference between individual cells exceeds the safety tolerance, the battery internal resistance mutation rate exceeds the critical risk threshold, or the ambient temperature exceeds the system high temperature warning line and the battery health status is lower than the aging risk threshold, the maximum cooling power is directly activated. The safety tolerance, critical risk threshold and high temperature warning line are all calibrated based on battery thermal failure boundary experimental data.

8. A big data-based new energy vehicle battery thermal management system implementing the method of any one of claims 1-7, characterized in that, Includes the following modules: The data acquisition and retrieval module collects real-time data on temperature distribution, charging and discharging current and voltage, ambient temperature and humidity, and vehicle driving conditions of each individual cell in the battery pack through a distributed sensor array. It obtains a real-time data stream and retrieves historical temperature change curves for similar scenarios from the cloud historical database based on the current spatiotemporal characteristic parameters. The temperature prediction and slope generation module inputs the real-time data stream and historical temperature change curves into a pre-trained time series prediction model. Through multi-source data fusion processing, it outputs the predicted temperature change curve within a set future time window and calculates the real-time temperature rise slope based on the predicted temperature change curve. The dynamic safety threshold generation module obtains the comprehensive environmental compensation amount based on the negative compensation effect of ambient temperature on the basic safety threshold and the nonlinear decay effect of battery health status on the basic safety threshold. The comprehensive environmental compensation amount increases monotonically with the increase of ambient temperature and accelerates with the increase of battery aging. The dynamic safety threshold is obtained by subtracting the comprehensive environmental compensation amount from the basic safety threshold. The early intervention point decision module determines the dynamic temperature compensation amount based on the real-time temperature rise slope through a preset slope grading mechanism. The dynamic temperature compensation amount increases stepwise as the slope increases. The early intervention temperature point is obtained by subtracting the dynamic temperature compensation amount from the dynamic safety threshold. The graded cooling execution module activates the graded cooling system when the battery pack temperature reaches the pre-intervention temperature point, and dynamically switches the cooling power level according to the growth range of the real-time temperature rise slope, wherein the cooling intensity is positively correlated with the slope.

Images