New energy automobile power battery temperature adjusting method, system and equipment

Abstract

The invention relates to the technical field of battery temperature control, and discloses a new energy automobile power battery temperature adjusting method, system and equipment, and the method comprises the following steps: synchronously collecting vehicle operation data, traffic information, charging pile state and battery aging parameters, and carrying out timestamp alignment and abnormal data filtering; converting the road test data into a vehicle coordinate system, predicting an aging trend in combination with LSTM, and generating a cross-domain feature vector through feature coding and tensor fusion; and generating a basic temperature control strategy based on reinforcement learning, predicting SOH deviation in combination with digital twinning, dynamically calculating an aging compensation coefficient, and synthesizing an anti-aging enhancement strategy. According to the invention, the maximum temperature is controlled within a safety threshold value by constructing a full-link cooperative temperature control system, so that the effectiveness of a thermal management strategy in a full life cycle is ensured.

Description

Technical Field

[0001] This invention relates to the field of battery temperature control, and more specifically, to a method, system, and device for regulating the temperature of a power battery for new energy vehicles. Background Technology

[0002] With the rapid development of intelligent connected vehicles, the efficiency of the power battery thermal management system (BTMS) has become a core factor restricting the performance and safety of new energy vehicles.

[0003] The degradation of the state of health (SOH) of existing automotive power batteries leads to asymptotic mismatch between the twin and the physical battery. Existing systems rely on closed-loop control using local vehicle sensors, which fails to integrate global information such as road conditions, charging station status, and weather in a vehicle-road cooperative environment. The failure of the aging compensation coefficient leads to the risk of overcooling, and the misjudgment rate of thermal runaway prediction has increased to 6.7%. Electrolyte crystallization has been observed in actual road tests. Summary of the Invention

[0004] This invention provides a method, system, and device for regulating the temperature of a power battery in a new energy vehicle, solving the technical problems in related technologies.

[0005] This invention provides a method for regulating the temperature of a power battery in a new energy vehicle, comprising the following steps:

[0006] The S100 synchronously collects vehicle operation data, traffic information, charging pile status and battery aging parameters, and performs timestamp alignment and abnormal data filtering.

[0007] S200 transforms road test data into the vehicle coordinate system, combines LSTM to predict aging trends, and generates cross-domain feature vectors through feature encoding and tensor fusion.

[0008] S300 is based on reinforcement learning to generate basic temperature control strategies, combined with digital twin prediction of SOH deviation, dynamic calculation of aging compensation coefficient, and synthesis of anti-aging enhancement strategies.

[0009] The S400 uses a current / heat dissipation component analysis strategy to quantify the sensitivity of the aging rate to temperature control and applies temperature regulation with physical constraints to achieve precise control within the safety boundary.

[0010] S500 calculates the temperature residual between the digital twin and the physical entity, evaluates the stability of the strategy, triggers online calibration of model parameters and reconstruction of the heat transfer network, and forms a closed-loop optimization mechanism.

[0011] Furthermore, vehicle operating data includes vehicle power battery temperature, load current, and SOC device data;

[0012] Traffic information includes traffic density and road gradient;

[0013] The charging station status includes charging power and remaining charging time;

[0014] Battery aging parameters include the current capacity of the vehicle's power battery and the resistance of each individual cell.

[0015] Furthermore, the formula for calculating the aging trend using LSTM is as follows: ; ;

[0016] in Represents the aging feature vector. Indicates the length of the time window. This represents the encoder, whose hidden layer dimension is 64, where Indicates the first A valid battery health status. Indicates the timestamp after compensation. Indicates the amount of delay compensation; ;

[0017] in Represents the aging trend prediction vector. Represents the prediction weight matrix. , Indicates the prediction bias term. ∈ .

[0018] Furthermore, the formula for calculating the cross-domain feature vector is as follows: ; ;

[0019] in This indicates vehicle operation data. Indicates the status of the charging station. This represents the transformed geographic coordinates. This represents the cumulative sum of element-wise multiplication. Represents the fused feature vector. This indicates a splicing operation. This indicates a transformation from WGS84 to the vehicle's local coordinate system. Indicates the first One road test data; ;

[0020] in This represents the final fused feature vector. Indicates the weights of the fully connected layer. , This represents the ELU activation function.

[0021] Furthermore, the specific steps in S300 are as follows:

[0022] S310, Basic Policy Reasoning: Based on the output of S250 Generate a baseline control strategy;

[0023] The calculation formula for the baseline control strategy is as follows: ; ;

[0024] in Indicates the baseline control strategy, This indicates reinforcement learning. Represents the state vector. , Indicates policy network parameters, This indicates a flattening operation;

[0025] S320, Digital Twin Prediction: Obtain the SOH prediction value of the battery digital model;

[0026] S330, Aging Deviation Calculation: Calculates the SOH deviation between the digital model and the physical system;

[0027] S340, Dynamic compensation coefficient generation: Calculate adaptive compensation weights;

[0028] S350, dual-mode strategy synthesis: superimposed compensation strategies.

[0029] Furthermore, the formula for calculating the SOH deviation is as follows: ;

[0030] in The actual measured value representing the real-time health status of the battery. This indicates the calculation of the L2 norm. Indicates SOH deviation; ;

[0031] in Indicates the length of the historical encoded feature window. This indicates a fully connected three-layer network. This represents the predicted SOH value.

[0032] Furthermore, the calculation formula for the dual-mode strategy synthesis is as follows: ; ;

[0033] in Indicates the maximum allowable control output. This represents a multimodal pre-trained neural network. This indicates the control quantity of the compensation strategy. Indicates adaptive compensation weights. This represents the synthesized dual-mode strategy; ;

[0034] in This represents the slope coefficient of the Sigmoid function. =0.8.

[0035] Furthermore, the specific steps in S400 are as follows:

[0036] S410, Aging Rate Quantization: Calculates the derivative of the battery health status output;

[0037] S420, Policy Component Analysis: The combined dual-mode policy after decomposition output is the control component;

[0038] S430, Sensitivity Field Construction: Establishing the Response Relationship between Aging Rate and Temperature Control Parameters;

[0039] S440, Compensation Calculation: Generate Temperature Adjustment Amount;

[0040] S450, Temperature Status Update: Applying constrained temperature regulation.

[0041] This invention also proposes a temperature regulation system for a new energy vehicle power battery, used to perform the steps in the aforementioned method for regulating the temperature of a new energy vehicle power battery, including:

[0042] Data acquisition module: Enhanced data acquisition, synchronously collecting vehicle operation data, traffic information, battery aging parameters, and charging pile status from multiple sources;

[0043] Spatiotemporal calibration module: Converts road test data to the vehicle coordinate system and combines it with an LSTM model to predict battery aging trends;

[0044] Dual-mode strategy generation module: Generates basic temperature control strategies based on reinforcement learning, and combines them with a digital twin model to predict SOH deviation;

[0045] Dynamic compensation execution module: Analyzes the temperature control strategy into current and heat dissipation components, and quantifies the sensitivity of aging rate to temperature control;

[0046] Evolutionary Verification and Update Module: Calculates the temperature residual between the digital twin and the physical entity to evaluate the stability of the strategy.

[0047] The present invention also proposes a temperature regulation device for a new energy vehicle power battery, comprising one or more processors for controlling the operation of a computing device; and a memory for storing data and program instructions used by the one or more processors, wherein the one or more processors are configured to execute instructions stored in the memory for performing the steps in the aforementioned method for regulating the temperature of a new energy vehicle power battery.

[0048] The beneficial effects of this invention are as follows:

[0049] This invention constructs a full-link collaborative temperature control system. Based on a multi-source data fusion framework with spatiotemporal calibration, it improves the accuracy of cross-domain information association. The dual-mode dynamic compensation strategy enables the temperature difference control accuracy under extreme operating conditions to reach ±0.5℃, reducing energy consumption by 29%. The self-evolutionary verification mechanism reduces the model drift rate to 0.3% / thousand kilometers, keeping the maximum temperature within the safe threshold, reducing the peak temperature difference by 63%, and reducing the annual SOH decay rate by 26.9%, ensuring the effectiveness of the thermal management strategy throughout the entire life cycle. Attached Figure Description

[0050] Figure 1 This is a flowchart of a method for regulating the temperature of a power battery for new energy vehicles proposed in this invention;

[0051] Figure 2 This is a structural block diagram of a new energy vehicle power battery temperature regulation system proposed in this invention.

[0052] In the diagram: 101, Data Acquisition Module; 102, Spatiotemporal Calibration Module; 103, Dual-Mode Strategy Generation Module; 104, Dynamic Compensation Execution Module; 105, Evolutionary Verification and Update Module. Detailed Implementation

[0053] The subject matter described herein will now be discussed with reference to exemplary embodiments. It should be understood that these embodiments are discussed only to enable those skilled in the art to better understand and implement the subject matter described herein, and changes may be made to the function and arrangement of the elements discussed without departing from the scope of this specification. Various processes or components may be omitted, substituted, or added as needed in the examples. Furthermore, features described in some examples may be combined in other examples.

[0054] like Figure 1 As shown, a method for regulating the temperature of a power battery in a new energy vehicle includes the following steps:

[0055] The S100 synchronously collects vehicle operation data, traffic information, charging pile status and battery aging parameters, and performs timestamp alignment and abnormal data filtering.

[0056] In one embodiment of the present invention, the following steps are specifically included:

[0057] S110, Multi-source data synchronous acquisition: Parallel acquisition of four types of data sources and timestamp; ;

[0058] in This indicates vehicle operation data. Indicates the temperature of the car's power battery. Indicates the load current. Represents SOC device data; ;

[0059] Indicates traffic information, Indicates traffic density. Indicates the road slope; ;

[0060] in Indicates the status of the charging station. Indicates charging power. Indicates the remaining charging time; ;

[0061] in Indicates battery aging parameters. Indicates the current capacity of the car's power battery. Indicates the resistance of a single unit;

[0062] S120 and SOH parameter calculation: Health indicators are calculated based on raw BMS data;

[0063] Calculation formula: ;

[0064] in Indicates the real-time health status of the battery. Indicates the battery's nominal capacity; ;

[0065] in Indicates the internal resistance growth rate. This indicates the reference value for the internal resistance of the new battery; ; ;

[0066] in Indicates the capacity decay rate. Indicates the cycle life design capacity. Indicates the timestamp after compensation. Indicates the amount of delay compensation; ;

[0067] in Indicates the battery's health status;

[0068] S130, Time Alignment Processing: Unifying the Time Base for Multi-Source Data;

[0069] Alignment method: ; ;

[0070] in, This indicates the battery health status after interpolation alignment. Indicates the sensor acquisition time. Indicates the system acquisition time. Indicates the frequency of sample acquisition;

[0071] S140, Abnormal data filtering: Sliding window filtering is used;

[0072] Filtering formula: ;

[0073] in This represents the t-th valid battery health state. Indicates the radius of the time window. =5, This represents the battery health status after the kth interpolation. Indicates Gaussian weights, , =1.5;

[0074] Final output standard dataset: ;

[0075] S200 transforms road test data into the vehicle coordinate system, combines LSTM to predict aging trends, and generates cross-domain feature vectors through feature encoding and tensor fusion.

[0076] In one embodiment of the present invention, the following steps are specifically included:

[0077] S210, Spatiotemporal reference alignment: unifying the spatiotemporal coordinate system of multi-source data;

[0078] Alignment formula: ;

[0079] in This represents the transformed geographic coordinates. This indicates a transformation from WGS84 to the vehicle's local coordinate system. Indicates the first One road test data;

[0080] S220, Aging Feature Encoding: Extracting the output of S140 Temporal characteristics;

[0081] Encoding formula: ;

[0082] in Represents the aging feature vector. Indicates the length of the time window. This indicates the encoder, whose hidden layer dimension is 64;

[0083] S230, Trend Prediction Modeling: Generating Aging Trend Prediction Vectors;

[0084] Prediction formula: ;

[0085] in Represents the aging trend prediction vector. Represents the prediction weight matrix. , Indicates the prediction bias term. ∈ ;

[0086] S240, Tensor Fusion Computation: Performs cross-modal feature fusion;

[0087] Fusion formula: ;

[0088] in This represents the cumulative sum of element-wise multiplication. Represents the fused feature vector. Indicates a splicing operation;

[0089] S250, Feature dimensionality reduction output: Generates the final fused feature vector;

[0090] Dimensionality reduction formula: ;

[0091] in This represents the final fused feature vector. Indicates the weights of the fully connected layer. , This represents the ELU activation function;

[0092] S300 is based on reinforcement learning to generate basic temperature control strategies, combined with digital twin prediction of SOH deviation, dynamic calculation of aging compensation coefficient, and synthesis of anti-aging enhancement strategies.

[0093] In one embodiment of the present invention, the following steps are specifically included:

[0094] S310, Basic Policy Reasoning: Based on the output of S250 Generate a baseline control strategy;

[0095] The calculation formula is as follows: ; ;

[0096] in Indicates the baseline control strategy, This indicates reinforcement learning. Represents the state vector. , Indicates policy network parameters, This indicates a flattening operation;

[0097] S320, Digital Twin Prediction: Obtain the SOH prediction value of the digital model of the battery;

[0098] Prediction formula: ;

[0099] in Indicates the length of the historical encoded feature window. This indicates a fully connected three-layer network. This indicates the predicted SOH value;

[0100] S330, Aging Deviation Calculation: Calculates the SOH deviation between the digital model and the physical system;

[0101] Deviation formula: ;

[0102] in The actual measured value representing the real-time health status of the battery. This indicates the calculation of the L2 norm. Indicates SOH deviation;

[0103] S340, Dynamic compensation coefficient generation: Calculate adaptive compensation weights;

[0104] Adaptive formula: ;

[0105] in This represents the slope coefficient of the Sigmoid function. =0.8, Indicates the adaptive compensation weight;

[0106] S350, Dual-mode strategy synthesis: superimposed compensation strategies;

[0107] Synthetic formula: ; ;

[0108] in Indicates the maximum allowable control output. This represents a multimodal pre-trained neural network. This indicates the control quantity of the compensation strategy. This represents the synthesized dual-mode strategy;

[0109] The S400 uses a current / heat dissipation component analysis strategy to quantify the sensitivity of the aging rate to temperature control and applies temperature regulation with physical constraints to achieve precise control within the safety boundary.

[0110] In one embodiment of the present invention, the following steps are specifically included:

[0111] S410, Aging Rate Quantization: Calculates the battery health status output by S140. differential;

[0112] Differential formula: ;

[0113] in Indicates the control period. This indicates the effective internal resistance growth rate. Indicates the effective capacity decay rate;

[0114] S420, Policy Component Analysis: Decomposing the output of S350 To control the components;

[0115] Analysis formula: ;

[0116] in This represents the current coefficient, corresponding to the first dimension of the S310 policy network output. This represents the heat dissipation coefficient, corresponding to the second dimension of the policy network;

[0117] S430, Sensitivity Field Construction: Establishing the Response Relationship between Aging Rate and Temperature Control Parameters;

[0118] Sensitivity model: ;

[0119] in Represents the sensitivity weight matrix. , Indicates the current / heat dissipation sensitivity coefficient;

[0120] S440, Compensation Calculation: Generate Temperature Adjustment Amount;

[0121] Compensation formula: ;

[0122] in Indicates the amount of temperature adjustment. Indicates the temperature rise at maximum current. Indicates the maximum heat dissipation temperature drop. Indicates the current sensitivity coefficient. Indicates the heat dissipation sensitivity coefficient;

[0123] S450, Temperature Status Update: Applying constrained temperature regulation;

[0124] Updated formula: ;

[0125] in Indicates the minimum operating temperature. Indicates the maximum permissible temperature. Represents a saturation model. This represents the current operating temperature at time t;

[0126] S500 calculates the temperature residual between the digital twin and the physical entity, evaluates the stability of the strategy, triggers online calibration of model parameters and reconstruction of the heat transfer network, and forms a closed-loop optimization mechanism.

[0127] In one embodiment of the present invention, the following steps are specifically included:

[0128] S510, Temperature Residual Window Calculation: Collect the temperature differences between the most recent N digital twins and the physical system;

[0129] The calculation formula is as follows: ;

[0130] in Indicates the length of the sliding window. Represents the temperature of a digital twin. This represents the actual value of the i-th temperature. This indicates the predicted latency of the digital model. This represents the temperature residual value. Indicates L1 norm calculation;

[0131] S520, Strategy Performance Verification: Generate strategy stability indicators;

[0132] Verification function: ;

[0133] in Indicates the prevention of zero constant, The Frobenius norm represents the policy change.

[0134] S530, drift detection triggered: when When the temperature exceeds 0.8℃, the update flag and historical data cache are triggered.

[0135] The calculation formula is as follows: ;

[0136] in This indicates the parameter update backtracking window. This indicates that the status flag has been updated. This represents the validated feature dataset;

[0137] S540, EKF_SOH Online Calibration: Update Health Status Estimator Parameters;

[0138] Calibration process: ; ;

[0139] in Indicates the rated temperature rise. This represents the noise covariance of the original EKF process. This represents the validated features in the SOH dataset. This represents the median. This represents the observation noise covariance matrix of the original EKF process;

[0140] S550, Thermal Network Reconfiguration: Update the heat transfer coefficient matrix;

[0141] Reconstructing the formula: ;

[0142] in This represents a matrix concatenation operation. Indicates the heat dissipation coefficient. This represents the updated and reconstructed heat transfer coefficient matrix;

[0143] like Figure 2 As shown, a temperature regulation system for a new energy vehicle power battery includes the following modules:

[0144] Data acquisition module 101: Enhanced data acquisition, synchronously collecting vehicle operation data (such as temperature, current, SOC), traffic information, and charging pile status from multiple sources;

[0145] Spatiotemporal calibration module 102: converts road test data to the vehicle coordinate system and combines it with an LSTM model to predict battery aging trends;

[0146] Dual-mode strategy generation module 103: Generates basic temperature control strategies based on reinforcement learning, and combines digital twin models to predict SOH (State of Health) deviation;

[0147] Dynamic compensation execution module 104: Analyzes the temperature control strategy into current and heat dissipation components, and quantifies the sensitivity of aging rate to temperature control;

[0148] Evolutionary Verification and Update Module 105: Calculates the temperature residual between the digital twin and the physical entity to evaluate the stability of the strategy;

[0149] Through the above modules, the following application scenarios are addressed:

[0150] Urban transportation: Electric vehicles driving in cities can adjust battery temperature in real time to adapt to traffic flow and changes in the external environment;

[0151] Fast charging station: Dynamically adjusts battery temperature during charging to ensure the safety and efficiency of fast charging;

[0152] Extreme climate conditions: In high or low temperature environments, the system can automatically adjust the battery temperature to ensure battery performance and safety;

[0153] Through the design of the above modules, the power battery temperature regulation system of new energy vehicles can achieve intelligent and dynamic management, improving battery efficiency and safety.

[0154] Here is an example of an application scenario involving fast charging in urban areas with high temperatures and traffic congestion:

[0155] Scenario problem description:

[0156] Electric vehicles face two major challenges in urban traffic during the summer heat (ambient temperature 40℃):

[0157] Fast charging thermal runaway risk: When fast charging from 20% to 80% SOC, the battery temperature rise rate exceeds 3℃ / min;

[0158] Congested roads reduce heat dissipation efficiency: Low-speed driving limits the pump speed of the liquid cooling system, and the battery pack temperature difference reaches 8°C.

[0159] This module-based collaborative solution for the temperature regulation system of power batteries in new energy vehicles:

[0160] 1. Data Acquisition Module:

[0161] Real-time monitoring of individual battery cell temperature (up to 52℃), charging current (300A), and SOC (65%).

[0162] Simultaneously acquire roadside weather station data (humidity 70%) and traffic density 3 kilometers ahead (85 vehicles / km);

[0163] Monitor battery aging parameters: =92%, =+18%;

[0164] Key function: To build a real-time dataset containing 16-dimensional feature vectors and complete multi-source alignment within 5ms;

[0165] 2. Spatiotemporal calibration module:

[0166] LSTM predicts aging rate over the next 10 minutes: =0.3%;

[0167] Map the location (latitude and longitude) of the charging pile to the vehicle coordinate system and calculate the pre-cooling time window before arrival;

[0168] Generate fused feature vectors: temperature gradient, traffic flow entropy, and aging acceleration;

[0169] The output predicts that the temperature will exceed the safety threshold (55℃) at the end of the charging process.

[0170] 3. Dual-mode strategy generation module:

[0171] Basic strategy: Reinforcement learning outputs current reduction requests (from 300A to 250A);

[0172] Compensation strategy: Increase the liquid cooling pump speed compensation based on the SOH deviation (92% vs. 95% predicted by the digital model);

[0173] Final strategy: Current coefficient = 0.83, heat dissipation coefficient = 1.2;

[0174] While ensuring that the charging time increases by ≤8%, the temperature rise rate is controlled to be ≤2℃ / min;

[0175] 4. Dynamic Compensation Execution Module:

[0176] The parsing strategy is as follows:

[0177] Adjustment of communication protocol for charging piles (GB / T 27930-2015);

[0178] The liquid cooling system was overclocked to 120% of its rated power.

[0179] Sensitivity compensation: targeting Abnormal, add an extra 2°C of cooling margin;

[0180] The maximum temperature is controlled at 49℃, and the temperature difference of the module is ≤3℃.

[0181] 5. Evolution Verification and Update Module:

[0182] Temperature deviation detected in digital twin =0.6℃ (<0.8℃ threshold);

[0183] Dynamically update the heat transfer coefficient matrix: ← ×1.15;

[0184] Optimize policy network weights: a policy stability metric An increase of 37%;

[0185] This enables the temperature regulation system of the power battery of new energy vehicles to adapt to local climate characteristics, with a false alarm rate of less than 0.1% for thermal runaway warning throughout the year.

[0186] The present invention also discloses a temperature regulation device for a power battery of a new energy vehicle, comprising: one or more processors for controlling the operation of a computing device; and a memory for storing data and program instructions used by the one or more processors, wherein the one or more processors are configured to execute instructions stored in the memory so that: the temperature regulation system for the power battery of the new energy vehicle can achieve intelligent and dynamic management, thereby improving the battery's efficiency and safety.

[0187] The embodiments of the present invention have been described above, but the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention, all of which are within the protection scope of the present invention.

Claims

1. A method for regulating the temperature of a power battery in a new energy vehicle, characterized in that, Includes the following steps: The S100 synchronously collects vehicle operation data, traffic information, charging pile status and battery aging parameters, and performs timestamp alignment and abnormal data filtering. S200 transforms road test data into the vehicle coordinate system, combines LSTM to predict aging trends, and generates cross-domain feature vectors through feature encoding and tensor fusion. S300 is based on reinforcement learning to generate basic temperature control strategies, combined with digital twin prediction of SOH deviation, dynamic calculation of aging compensation coefficient, and synthesis of anti-aging enhancement strategies. The S400 uses a current / heat dissipation component analysis strategy to quantify the sensitivity of the aging rate to temperature control and applies temperature regulation with physical constraints to achieve precise control within the safety boundary. S500 calculates the temperature residual between the digital twin and the physical entity, evaluates the stability of the strategy, triggers online calibration of model parameters and reconstruction of the heat transfer network, and forms a closed-loop optimization mechanism.

2. The method for regulating the temperature of a power battery for a new energy vehicle according to claim 1, characterized in that, Vehicle operating data includes vehicle power battery temperature, load current, and SOC device data; Traffic information includes traffic density and road gradient; The charging station status includes charging power and remaining charging time; Battery aging parameters include the current capacity of the vehicle's power battery and the resistance of each individual cell.

3. The method for regulating the temperature of a power battery for a new energy vehicle according to claim 2, characterized in that, The formula for calculating aging trends using LSTM is as follows: ； ； in Represents the aging feature vector. Indicates the length of the time window. This represents the encoder, whose hidden layer dimension is 64, where Indicates the first A valid battery health status. Indicates the timestamp after compensation. Indicates the amount of delay compensation; ； in Represents the aging trend prediction vector. Represents the prediction weight matrix. , Indicates the prediction bias term. ∈ .

4. The method for regulating the temperature of a power battery for a new energy vehicle according to claim 3, characterized in that, The formula for calculating cross-domain feature vectors is as follows: ； ； in This indicates vehicle operation data. Indicates the status of the charging station. This represents the transformed geographic coordinates. This represents the cumulative sum of element-wise multiplication. Represents the fused feature vector. This indicates a splicing operation. This indicates a transformation from WGS84 to the vehicle's local coordinate system. Indicates the first One road test data; ； in This represents the final fused feature vector. Indicates the weights of the fully connected layer. , This represents the ELU activation function.

5. The method for regulating the temperature of a power battery for a new energy vehicle according to claim 4, characterized in that, The specific steps in S300 are as follows: S310, Basic Policy Reasoning: Based on the output of S250 Generate a baseline control strategy; The calculation formula for the baseline control strategy is as follows: ； ； in Indicates the baseline control strategy, This indicates reinforcement learning. Represents the state vector. , Indicates policy network parameters, This indicates a flattening operation; S320, Digital Twin Prediction: Obtain the SOH prediction value of the battery digital model; S330, Aging Deviation Calculation: Calculates the SOH deviation between the digital model and the physical system; S340, Dynamic compensation coefficient generation: Calculate adaptive compensation weights; S350, dual-mode strategy synthesis: superimposed compensation strategies.

6. The method for regulating the temperature of a power battery for a new energy vehicle according to claim 5, characterized in that, The formula for calculating SOH deviation is as follows: ； in The actual measured value representing the real-time health status of the battery. This indicates the calculation of the L2 norm. Indicates SOH deviation; ； in Indicates the length of the historical encoded feature window. This indicates a fully connected three-layer network. This represents the predicted SOH value.

7. The method for regulating the temperature of a power battery for a new energy vehicle according to claim 6, characterized in that, The calculation formula for dual-mode strategy synthesis is as follows: ； ； in Indicates the maximum allowable control output. This represents a multimodal pre-trained neural network. This indicates the control quantity of the compensation strategy. Indicates adaptive compensation weights. This represents the synthesized dual-mode strategy; ； in This represents the slope coefficient of the Sigmoid function. =0.

8.

8. The method for regulating the temperature of a power battery for a new energy vehicle according to claim 7, characterized in that, The specific steps in S400 are as follows: S410, Aging Rate Quantization: Calculates the derivative of the battery health status output; S420, Policy Component Analysis: The combined dual-mode policy after decomposition output is the control component; S430, Sensitivity Field Construction: Establishing the Response Relationship between Aging Rate and Temperature Control Parameters; S440, Compensation Calculation: Generate Temperature Adjustment Amount; S450, Temperature Status Update: Applying constrained temperature regulation.

9. A temperature regulation system for a power battery in a new energy vehicle, characterized in that, The steps in the method for regulating the temperature of a power battery for a new energy vehicle as described in any one of claims 1-8 include: Data acquisition module: Enhanced data acquisition, synchronously collecting vehicle operation data, traffic information, battery aging parameters, and charging pile status from multiple sources; Spatiotemporal calibration module: Converts road test data to the vehicle coordinate system and combines it with an LSTM model to predict battery aging trends; Dual-mode strategy generation module: Generates basic temperature control strategies based on reinforcement learning, and combines them with a digital twin model to predict SOH deviation; Dynamic compensation execution module: Analyzes the temperature control strategy into current and heat dissipation components, and quantifies the sensitivity of aging rate to temperature control; Evolutionary Verification and Update Module: Calculates the temperature residual between the digital twin and the physical entity to evaluate the stability of the strategy.

10. A temperature regulation device for a power battery in a new energy vehicle, characterized in that, It includes one or more processors for controlling the operation of a computing device; and a memory for storing data and program instructions used by the one or more processors, wherein the one or more processors are configured to execute instructions stored in the memory for performing steps in a method for regulating the temperature of a power battery for a new energy vehicle as described in any one of claims 1-8.

Images