Connected vehicle accident battery cloud evaluation method based on inner and outer field coupling twins

Abstract

This invention discloses a cloud-based assessment method for battery accidents in connected vehicles based on an internal-external field coupled twin. First, a transient accident data window is defined. The vehicle-to-everything (V2X) cloud platform extracts sensor data within this window to construct a multi-field incomplete time-series dataset of external impact and internal cell voltage. Second, an internal-external field coupled twin model is constructed. Damping and RC polarization attenuation algorithms are used to deduce and repair blind spots in the accident data, obtaining an impact-voltage twin enhancement vector. Then, through a multi-source torque collaborative dynamics unit and a voltage entropy spatiotemporal outlier mining unit, the mechanical clamping force decay value and the deviation of the cell thermodynamic anomaly are calculated, constructing a full-dimensional failure degree matrix. Normalized weighted calculations generate a quantitative assessment vector, outputting a risk location report. This invention can achieve accurate quantification and graded early warning of battery pack accident failure states under extreme accident conditions, providing important technical support for battery safety emergency response and management strategy optimization.

Description

Technical Field

[0001] This invention relates to the field of battery pack safety prediction technology for intelligent connected vehicles, and specifically to a battery cloud assessment method for connected vehicles in accident based on an internal and external field coupling twin. Background Technology

[0002] The new energy intelligent connected vehicle industry is booming. As a core energy storage component, the safety of power batteries is directly related to the operational reliability of vehicles and the safety of passengers. In real-world road traffic environments, vehicles inevitably face sudden accidents such as collisions and scraping. The severe mechanical impacts from these accidents not only cause mechanical structural damage such as deformation of the battery pack casing and loosening of bolts, but also often induce the degradation of the internal electrochemical performance of the battery, leading to catastrophic consequences such as thermal runaway or even fire and explosion.

[0003] However, existing assessment methods have significant limitations: traditional battery management systems (BMS) focus on threshold monitoring of voltage and temperature, making it difficult to detect minute structural failures; while high-precision finite element simulations involve enormous computational demands, making it difficult to meet the real-time requirements of accident sites. Furthermore, severe collisions often damage sensors or disrupt communication, creating data "blind spots." Existing data-driven methods are highly dependent on data integrity; once signals are lost, they become ineffective, and they lack joint quantitative analysis of mechanical, electrochemical, and other multi-factor characteristics.

[0004] Therefore, how to leverage the powerful computing capabilities of the cloud in an intelligent connected environment to construct a digital twin assessment method that integrates real-time monitoring and deep mechanical-electrochemical dual characteristics, thereby achieving full-dimensional quantification of the battery pack risk of accident vehicles from structure to thermal runaway, and thus improving passenger safety and emergency response efficiency, has become a critical technical bottleneck that urgently needs to be overcome in the field of new energy vehicle safety emergency management. Summary of the Invention

[0005] This invention overcomes the shortcomings of existing technologies by proposing a cloud-based battery assessment method for connected vehicles based on an internal and external field coupled twin. This method aims to overcome the limitations of existing technologies, such as incomplete assessment of single factors after a vehicle accident, limited vehicle-side computing power preventing complex real-time calculations, and susceptibility to sensor damage leading to data interruptions. Therefore, it improves the ability to assess the overall state of the battery pack during emergencies, enhancing passenger safety and accident decision-making efficiency, and providing a reference for emergency response.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] A cloud-based battery assessment method for connected vehicle accidents, based on an internal / external field coupling twin, involves the following steps performed using computer equipment.

[0008] Step 1: The cloud server receives a vehicle accident signal, defines a transient accident data window, extracts sensor data within this window, and constructs a multi-field incomplete time-series dataset of external impact and internal single-cell voltage. The aforementioned multi-field incomplete time-series dataset specifically includes the external mechanical feature matrix of the battery. With internal electrochemical property matrix ;

[0009] Step 2: Construct an internal and external field coupled twin model, combining the external shock with the internal single-cell voltage multi-field incomplete time-series dataset. Input the model and perform inference transformation to obtain the full-dimensional failure degree matrix of the battery pack. ;

[0010] Step 3: For the matrix Normalization and weighted calculations are performed to generate a battery failure quantification evaluation vector. The analysis of mechanical structure integrity loss and electrochemical performance degradation will generate quantitative failure assessment results and risk location reports, which will be submitted to the cloud server for output.

[0011] Further, step 1 includes the following steps:

[0012] Step 1.1: The cloud server receives the collision message frame sequence reported by the vehicle via V2I communication. The moment the last collision message frame is received is recorded as the moment the accident occurred. ;

[0013] Step 1.2: Based on the time of the accident anchor point Define a full diagnostic time window covering both the early warning and later stages of an accident. The definition is as follows:

[0014] (1)

[0015] (2)

[0016] in The preset duration of historical data review before the accident. This is the preset duration of continuous monitoring after an accident. If the sensor continues to upload data after the accident, the complete sequence is extracted; if the sensor is damaged due to the accident and data is interrupted, the valid historical sequence before the interruption is extracted.

[0017] Step 1.3: Retrieve from the historical storage queue of the cloud database. Transverse vibration amplitude sequence acquired by vibration monitoring sensor within window With transverse vibration frequency sequence Using this as the mechanical modal features of the multi-field incomplete time-series dataset of external shock and internal single-cell voltage, an external mechanical feature matrix of the battery is constructed. for:

[0018] (3)

[0019] in, .

[0020] Step 1.4: Retrieve from the historical storage queue of the cloud database. Battery voltage data is used as the electrochemical modal characteristics of a multi-field incomplete time-series dataset of external shocks and internal single-cell voltages to construct an internal electrochemical characteristic matrix of the battery. for:

[0021] (4)

[0022] in, , Indicates the first The single cell in the first Historical voltage values ​​at each sampling time.

[0023] Step 1.5: Assemble the external mechanical features of the battery. With internal electrochemical properties Synchronous alignment and integration are performed along the time dimension to construct a multi-field incomplete time-series dataset of external shocks and internal individual voltages to characterize the transient characteristics of accidents. :

[0024] (5)

[0025] Further, step 2 includes the following steps:

[0026] Step 2.1: Constructing the data vector reconstruction unit of the internal and external field coupled twin model: This involves data repair for periods where sensors were unable to upload data due to accidents, generating coverage data. Impact-voltage twin enhancement vector throughout the entire accident cycle:

[0027] Step 2.1.1: The model automatically parses the imported multi-field incomplete time-series dataset of external shocks and internal individual voltages. The last moment of location data interruption ;

[0028] Step 2.1.2: Addressing the blind spot period after sensor failure. At the time of extraction Lateral vibration amplitude The initial amplitude is given by time. transverse vibration frequency The vibration frequency, combined with the preset structural damping ratio The transverse vibration amplitude sequence during the failure period was generated by deduction. :

[0029] (6)

[0030] Based on the physical characteristics of damped free vibration, a transverse vibration frequency sequence is set for the failure period. Maintain a constant dominant frequency at the impact end:

[0031] (7)

[0032] Step 2.1.3 at the extraction time Battery cell voltage The initial voltage is combined with the preset theoretical steady-state recovery voltage. With cell polarization time constant The voltage recovery sequence during the failure period is generated by deduction. :

[0033] (8)

[0034] Step 2.1.4: The model will use the matrix... and The historical data sequences stored in the database are respectively compared with those generated by the deduction. Sequence and The sequences are spliced ​​together along the time dimension to reconstruct a cover. Impact-voltage twin enhancement vector throughout the entire accident diagnosis cycle This serves as the input for subsequent feature recombination calculations.

[0035] Step 2.2: Construct a multi-source torque cooperative dynamics calculation unit for the coupled internal and external field twin model: For the vector... Perform feature reorganization to estimate clamping force decay value :

[0036] Step 2.2.1: The model analyzes the input impulse-voltage twin enhancement vector. traversal Each sampling time in The instantaneous motion state of the bolt head is updated using the phase accumulation method, and the cumulative phase at the current moment is updated using an iterative formula. :

[0037] (9)

[0038] in, It is a transverse vibration frequency sequence. to Set the sampling time step to , The accumulated phase from the previous sampling time, the initial phase Set to 0;

[0039] Step 2.2.2: Based on the transverse vibration amplitude sequence The amplitude of the lateral vibration at the current moment With accumulated phase Calculate the end-face friction shear force at that moment. :

[0040] (10)

[0041] In the formula: Indicates the elastic modulus of the bolt material; This represents the moment of inertia on the cross-section of the bolt; Indicates the effective bending length of the bolt. Indicates the bolt bending factor. The bending stiffness coefficient of the bolt head in the battery box installation scenario.

[0042] Step 2.2.3: Use numerical iteration to find the parameters in the following formula. , Indicates the translational velocity of the end face and end face rotation speed The ratio:

[0043] (11)

[0044] In the formula: and These represent the minimum and maximum end-face contact radii, respectively. Indicates the coefficient of friction of the end face; The average pressure at the end face is expressed by the following formula:

[0045] (12)

[0046] In the formula, The bolt clamping force at the current moment. Indicates the initial preload. This represents the cumulative decrease in clamping force.

[0047] Step 2.2.4, from the already obtained Substitute into the following formula to calculate the end face torque. :

[0048] (13)

[0049] In the formula, This represents the circumferential integral angle in a polar coordinate system centered on the bolt axis.

[0050] Step 2.2.5: Calculate the thread friction tangential force according to the following formula. :

[0051] (14)

[0052] In the formula: ; and These represent the minor diameter and major diameter of the thread, respectively. Represents the thread flank angle; Indicates the effective bending length of the bolt; Indicates the pitch.

[0053] Step 2.2.6: Calculate the unknown parameters using the numerical iteration method according to the following formula. , Indicates the translational speed of the thread. and thread rotation speed The ratio:

[0054] (15)

[0055] In the formula: and These represent the thread flank angle and the thread lead angle, respectively. This represents the coefficient of friction of the thread; where The average pressure on the threaded contact surface is expressed by the following formula:

[0056] (16)

[0057] in This represents the thread contact area.

[0058] Step 2.2.7: Calculate the thread torque according to the following formula. :

[0059] (17)

[0060] Step 2.2.8: Calculate the pitch torque according to the following formula. :

[0061] (18)

[0062] In the formula: This indicates the bolt clamping force.

[0063] Step 2.2.9: Construct a system containing pitch torque. Thread torque and end face torque The transient dynamic equilibrium equations, traversing the sampling times Calculate the instantaneous angular acceleration of the bolt. :

[0064] (19)

[0065] In the formula: This represents the moment of inertia of the bolt (calculated based on the moment of inertia of a cylinder). , (For bolt quality).

[0066] Step 2.2.10: Update the current time using the discrete-time integration method. angular velocity of rotation Cumulative angular displacement :

[0067] (20)

[0068] (twenty one)

[0069] Step 2.2.11: Calculate the time corresponding to the sampling point. Degree of decrease in clamping force :

[0070] (twenty two)

[0071] In the formula: For battery housing rigidity; This refers to the bolt stiffness.

[0072] Step 2.3: Constructing the spatiotemporal outlier mining algorithm unit for the voltage entropy value of the internal and external field coupled twin model: For the time series vector... Perform feature recombination to calculate the thermodynamic deviation of the battery cell. :

[0073] Step 2.3.1, targeting from Time's up For each voltage sampling point in the voltage data of each individual cell at any given time, define the discretization interval of the voltage state space:

[0074] (twenty three)

[0075] in For the number of intervals, and From respectively Time's up Minimum and maximum values ​​of individual cell voltage data at any given time;

[0076] Step 2.3.2, Statistics of the first Frequency distribution of single cell voltage data falling within various discrete intervals Calculate its probability distribution vector , No. The voltage of the No. 1 cell falls within the range of the No. 1 cell. The probability of each discrete interval is ,in The calculation formula is as follows:

[0077] (twenty four)

[0078] Step 2.3.3: Calculate the full diagnostic time window according to the following formula. Inner Shannon entropy of a single cell :

[0079] (25)

[0080] In the formula: The number of discrete regions. The process involves iterating through all individual cells, performing voltage state space discretization interval division, frequency distribution statistics, probability distribution vector calculation, and Shannon entropy solving, to generate the Shannon entropy feature vector for all individual cells within the full diagnostic time window. ;

[0081] Step 2.3.4: Calculate the Shannon entropy eigenvectors of all individual cells. Mapping to a state space, calculating the target single cell Corresponding locally reachable density :

[0082] (26)

[0083] In the formula: For target single cell battery The Distance neighborhood, representing the distance to the target single cell. The Shannon entropy value closest to A collection of other individual cells, The number of individual cells in the neighborhood. For single cell batteries With single cell The entropy value and Euclidean distance between them For single cell batteries The Proximity distance, i.e., distance from a single cell Shannon entropy value Closest Euclidean distance;

[0084] Step 2.3.5: Calculate the target unit based on local reachability density. Local outlier :

[0085] (27)

[0086] Step 2.4: Constructing a quantitative evaluation unit for the coupled internal and external field twin model: This involves calculating the clamping force decay value. Combining cell thermodynamic deviation Construct a full-dimensional failure degree matrix for battery packs. :

[0087] Step 2.4.1: The model calculates the time-varying trajectory of the decaying clamping force. Perform an extreme value search to extract the maximum mechanical loosening force throughout the entire diagnostic cycle. Simultaneously, the local outlier sequence of all individual cells throughout the entire cycle is scanned to extract the maximum abnormal deviation of the thermodynamic state. ;

[0088] Step 2.4.2: Extract the... and Vectorized encapsulation is performed to construct a full-dimensional failure degree matrix representing the final damage state of the battery pack under accident conditions. :

[0089] (28)

[0090] Furthermore, step 3 includes the following steps:

[0091] Step 3.1: Analyze the full-dimensional failure degree matrix of the battery pack. Extract the maximum mechanical loosening force Deviation from maximum anomaly Introduce a preset mechanical failure limit threshold. thermal runaway safety threshold The normalized index of mechanical risk was calculated using the linear ratio method. Electrochemical risk normalization index :

[0092] (29)

[0093] (30)

[0094] If the calculation result is greater than 1, its value is truncated and locked to 1;

[0095] Step 3.2: Introduce the preset feature weight coefficient vector. ,in The normalized index , A point-to-point multiplication operation is performed with the feature weight coefficients to construct a battery failure quantification evaluation vector. :

[0096] (31)

[0097] In the formula, The weighted mechanical failure assessment components, This is the weighted electrochemical failure assessment component.

[0098] Step 3.3: Define the finite set of discrete failure states of the battery pack. :

[0099] (32)

[0100] in, For safety reasons, It is a single mechanical loosening state. This is a single precursor state to thermal runaway. This represents a critical state of structural-electrochemical concurrent cascade.

[0101] Step 3.4: Based on the battery failure quantification evaluation vector Components in and The following decision logic is executed:

[0102] like and Determined as The output is "The system structure is intact and the electrochemical state is stable"; if and Determined as Output a "Level 1 Mechanical Connection Failure Warning" and mark the battery box mounting point where the connection has become loose; if and Determined as Output "Early warning of Level 2 thermal runaway" and mark the cell numbers of cells exhibiting outliers; if and Determined as It outputs a "critical alarm for the tertiary structure-electrochemical cascade".

[0103] In another aspect, the present invention also discloses a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the steps of the method described above.

[0104] In another aspect, the present invention also discloses a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor performs the steps of the method described above.

[0105] As described above, the battery cloud assessment method for connected vehicle accidents based on an internal-external field coupled twin of the present invention locks the vehicle impact sensor data and battery voltage time-series data before the accident through a preset window, establishes the external mechanical feature matrix and internal electrochemical characteristic matrix of the battery at the moment of failure; constructs an internal-external field coupled twin model for vector reconstruction and feature recombination to obtain a full-dimensional failure degree matrix of the battery pack; then, it calculates the battery failure quantitative assessment vector using normalization processing and feature weight coefficients, automatically analyzes the loss of mechanical structural integrity and the degree of electrochemical performance degradation, and finally outputs the quantitative failure assessment results and risk location report through a cloud server. This invention can accurately assess the battery pack accident failure state and quantify the degree of loss in an intelligent connected environment, providing important technical support for battery safety emergency response and management strategy optimization.

[0106] Compared with existing technologies, the beneficial effects of the present invention are as follows:

[0107] 1. This invention constructs a data repair and reconstruction mechanism based on physical mechanisms, effectively solving the blind spot problem caused by sensor damage or data packet loss due to accidental impacts. By constructing a data vector reconstruction unit and utilizing various physical parameters, the interrupted impact-voltage time series data is dynamically deduced and repaired, ensuring that the evaluation model can still obtain complete and continuous input data under extreme accident conditions.

[0108] 2. This invention establishes a dual-factor coupled evaluation system of "external mechanical and internal electrochemical" factors, overcoming the limitations of traditional single-physical factor monitoring and improving the comprehensiveness of fault identification. It analyzes the correlation between mechanical structure loosening and electrochemical performance degradation, enabling more accurate identification of hidden composite faults caused by impact.

[0109] 3. This invention achieves quantitative classification and precise location of battery failure risks, providing an intuitive basis for graded rescue and decision-making after an accident. By constructing a full-dimensional failure degree matrix for the battery pack and introducing dual thresholds for mechanical failure and thermal runaway safety, the efficiency and targeting of emergency response are significantly improved. Attached Figure Description

[0110] Figure 1 This is a flowchart of the present invention;

[0111] Figure 2 This is a detailed logic block diagram for constructing the internal and external field coupled twin model in this invention;

[0112] Figure 3 The test result diagram was constructed to verify the present invention. Detailed Implementation

[0113] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments.

[0114] In this embodiment, refer to Figure 1 and Figure 2 This paper discloses a cloud-based battery assessment method for connected vehicle accidents based on an internal-external field coupled twin. It primarily utilizes cloud-based digital twins and physical mechanism deduction, deeply integrating external mechanical impact characteristics and internal electrochemical properties of the battery for analysis, to enhance the assessment capability of battery pack composite faults. Specifically, the method first uses a cloud server to lock impact and voltage monitoring data before and after the accident, establishing an initial feature matrix. Then, using the constructed internal-external field coupled twin model, it performs dynamic deduction and repair on sensor data that may have been interrupted due to impact damage, and simultaneously reconstructs features to inversely calculate the decay value of mechanical clamping force and the deviation of cell thermodynamics. Finally, through normalization and weighted calculation of multi-dimensional features, a quantified battery failure assessment vector is generated, automatically outputting a graded early warning report including mechanical integrity loss and electrochemical performance degradation.

[0115] The specific steps include:

[0116] Step 1: The cloud server receives a vehicle accident signal, defines a transient accident data window, extracts sensor data within this window, and constructs a multi-field incomplete time-series dataset of external impact and internal single-cell voltage. The aforementioned multi-field incomplete time-series dataset specifically includes the external mechanical feature matrix of the battery. With internal electrochemical property matrix ,like Figure 1 As shown;

[0117] Step 1.1: The cloud server receives the collision message frame sequence reported by the vehicle via V2I communication. The moment the last collision message frame is received is recorded as the moment the accident occurred. ;

[0118] In this embodiment, the collision message frame sequence Specifically, the data packets originate from the vehicle uploaded via Ethernet. Each frame of data includes: a frame header identifier, a UTC timestamp, instantaneous vibration sensor values, and an airbag deployment status bit. The cloud server monitors when the "airbag deployment status bit" changes from 0 to 1 or when the impact acceleration exceeds a preset threshold of 3g to pinpoint the anchor point of the accident. .

[0119] Step 1.2: Based on the time of the accident anchor point Define a full diagnostic time window covering both the early warning and later stages of an accident. The definition is as follows:

[0120] (1)

[0121] (2)

[0122] in The preset duration of historical data review before the accident. This is the preset duration of continuous monitoring after an accident. If the sensor continues to upload data after the accident, the complete sequence is extracted; if the sensor is damaged due to the accident and data is interrupted, the valid historical sequence before the interruption is extracted.

[0123] In this embodiment, , If the sensor continues to upload data after the accident, the complete sequence is extracted; if the sensor is damaged due to the accident and the data is interrupted, the valid historical sequence before the interruption is extracted as the input to the data vector reconstruction unit in the subsequent model.

[0124] Step 1.3: Retrieve from the historical storage queue of the cloud database. Transverse vibration amplitude sequence acquired by vibration monitoring sensor within window With transverse vibration frequency sequence Using this as the mechanical modal features of the multi-field incomplete time-series dataset of external shock and internal single-cell voltage, an external mechanical feature matrix of the battery is constructed. for:

[0125] (3)

[0126] in, .

[0127] Step 1.4: Retrieve from the historical storage queue of the cloud database. Battery voltage data is used as the electrochemical modal characteristics of a multi-field incomplete time-series dataset of external shocks and internal single-cell voltages to construct an internal electrochemical characteristic matrix of the battery. for:

[0128] (4)

[0129] in, , Indicates the first The single cell in the first Historical voltage values ​​at each sampling time.

[0130] Step 1.5: Assemble the external mechanical features of the battery. With internal electrochemical properties Synchronous alignment and integration are performed along the time dimension to construct a multi-field incomplete time-series dataset of external shocks and internal individual voltages to characterize the transient characteristics of accidents. :

[0131] (5)

[0132] Step 2: Construct an internal and external field coupled twin model, combining the external shock with the internal single-cell voltage multi-field incomplete time-series dataset. Input the model and perform inference transformation to obtain the full-dimensional failure degree matrix of the battery pack. ,like Figure 2 As shown;

[0133] Step 2.1: Constructing the data vector reconstruction unit of the internal and external field coupled twin model: This involves data repair for periods where sensors were unable to upload data due to accidents, generating coverage data. Impact-voltage twin enhancement vector throughout the entire accident cycle:

[0134] Step 2.1.1: The model automatically parses the imported multi-field incomplete time-series dataset of external shocks and internal individual voltages. The last moment of location data interruption ;

[0135] Step 2.1.2: Addressing the blind spot period after sensor failure. At the time of extraction Lateral vibration amplitude The initial amplitude is given by time. transverse vibration frequency The vibration frequency, combined with the preset structural damping ratio The transverse vibration amplitude sequence during the failure period was generated by deduction. :

[0136] (6)

[0137] In this embodiment, the preset structural damping ratio .

[0138] Based on the physical characteristics of damped free vibration, a transverse vibration frequency sequence is set for the failure period. Maintain a constant dominant frequency at the impact end:

[0139] (7)

[0140] Step 2.1.3 at the extraction time Battery cell voltage The initial voltage is combined with the preset theoretical steady-state recovery voltage. With cell polarization time constant The voltage recovery sequence during the failure period is generated by deduction. :

[0141] (8)

[0142] In this embodiment, the preset cell polarization time constant Preset theoretical steady-state recovery voltage The moment the accident occurred The open-circuit voltage (OCV) corresponding to the battery's state of charge (SOC).

[0143] Step 2.1.4: The model will use the matrix... and The historical data sequences stored in the database are respectively compared with those generated by the deduction. Sequence and The sequences are spliced ​​together along the time dimension to reconstruct a cover. Impact-voltage twin enhancement vector throughout the entire accident diagnosis cycle This serves as the input for subsequent feature recombination calculations.

[0144] Step 2.2: Construct a multi-source torque cooperative dynamics calculation unit for the coupled internal and external field twin model: For the vector... Perform feature reorganization to estimate clamping force decay value :

[0145] Step 2.2.1: The model analyzes the input impulse-voltage twin enhancement vector. traversal Each sampling time in The instantaneous motion state of the bolt head is updated using the phase accumulation method, and the cumulative phase at the current moment is updated using an iterative formula. :

[0146] (9)

[0147] in, It is a transverse vibration frequency sequence. to Set the sampling time step to , The accumulated phase from the previous sampling time, the initial phase Set to 0; in this embodiment, .

[0148] Step 2.2.2: Based on the transverse vibration amplitude sequence The amplitude of the lateral vibration at the current moment With accumulated phase Calculate the end-face friction shear force at that moment. :

[0149] (10)

[0150] In the formula: Indicates the elastic modulus of the bolt material; This represents the moment of inertia on the cross-section of the bolt; Indicates the effective bending length of the bolt. Indicates the bolt bending factor. This refers to the bolt head bending stiffness coefficient in the battery enclosure mounting scenario. In this embodiment, , , , .

[0151] Step 2.2.3: Use numerical iteration to find the parameters in the following formula. , Indicates the translational velocity of the end face and end face rotation speed The ratio:

[0152] (11)

[0153] In the formula: and These represent the minimum and maximum end-face contact radii, respectively. Indicates the coefficient of friction of the end face; The average pressure at the end face is expressed by the following formula:

[0154] (12)

[0155] In the formula, The bolt clamping force at the current moment. Indicates the initial preload. This represents the cumulative decrease in clamping force.

[0156] In this embodiment, , , , .

[0157] Step 2.2.4, from the already obtained Substitute into the following formula to calculate the end face torque. :

[0158] (13)

[0159] In the formula, This represents the circumferential integral angle in a polar coordinate system centered on the bolt axis.

[0160] Step 2.2.5: Calculate the thread friction tangential force according to the following formula. :

[0161] (14)

[0162] In the formula: ; and These represent the minor diameter and major diameter of the thread, respectively. Represents the thread flank angle; Indicates the effective bending length of the bolt; Indicates the pitch. In this embodiment, , .

[0163] Step 2.2.6: Calculate the unknown parameters using the numerical iteration method according to the following formula. , Indicates the translational speed of the thread. and thread rotation speed The ratio:

[0164] (15)

[0165] In the formula: and These represent the thread flank angle and the thread lead angle, respectively. This represents the coefficient of friction of the thread; where The average pressure on the threaded contact surface is expressed by the following formula:

[0166] (16)

[0167] in This represents the threaded contact area. In this embodiment, , , .

[0168] Step 2.2.7: Calculate the thread torque according to the following formula. :

[0169] (17)

[0170] Step 2.2.8: Calculate the pitch torque according to the following formula. :

[0171] (18)

[0172] In the formula: This indicates the current bolt clamping force.

[0173] Step 2.2.9: Construct a system containing pitch torque. Thread torque and end face torque The transient dynamic equilibrium equations, traversing the sampling times Calculate the instantaneous angular acceleration of the bolt. :

[0174] (19)

[0175] In the formula: This represents the moment of inertia of the bolt (calculated based on the moment of inertia of a cylinder). , (For bolt mass). In this embodiment, .

[0176] Step 2.2.10: Update the current time using the discrete-time integration method. angular velocity of rotation Cumulative angular displacement :

[0177] (20)

[0178] (twenty one)

[0179] Step 2.2.11: Calculate the time corresponding to the sampling point. Degree of decrease in clamping force :

[0180] (twenty two)

[0181] In the formula: For battery housing rigidity; This refers to the bolt stiffness. In this embodiment, , .

[0182] Step 2.3: Constructing the spatiotemporal outlier mining algorithm unit for the voltage entropy value of the internal and external field coupled twin model: For the time series vector... Perform feature recombination to calculate the thermodynamic deviation of the battery cell. :

[0183] Step 2.3.1, targeting from Time's up For each voltage sampling point in the voltage data of each individual cell at any given time, define the discretization interval of the voltage state space:

[0184] (twenty three)

[0185] in For the number of intervals, and From respectively Time's up The minimum and maximum values ​​of the individual cell voltage data at any given time; in this embodiment... .

[0186] Step 2.3.2, Statistics of the first Frequency distribution of single cell voltage data falling within various discrete intervals Calculate its probability distribution vector , No. The voltage of the No. 1 cell falls within the range of the No. 1 cell. The probability of each discrete interval is ,in The calculation formula is as follows:

[0187] (twenty four)

[0188] Step 2.3.3: Calculate the full diagnostic time window according to the following formula. Inner Shannon entropy of a single cell :

[0189] (25)

[0190] In the formula: The number of discrete regions. The process involves iterating through all individual cells, performing voltage state space discretization interval division, frequency distribution statistics, probability distribution vector calculation, and Shannon entropy solving, to generate the Shannon entropy feature vector for all individual cells within the full diagnostic time window. ;

[0191] Step 2.3.4: Calculate the Shannon entropy eigenvectors of all individual cells. Mapping to a state space, calculating the target single cell Corresponding locally reachable density :

[0192] (26)

[0193] In the formula: For target single cell battery The Distance neighborhood, representing the distance to the target single cell. The Shannon entropy value closest to A collection of other individual cells, The number of individual cells in the neighborhood. For single cell batteries With single cell The entropy value and Euclidean distance between them For single cell batteries The Proximity distance, i.e., distance from a single cell Shannon entropy value Approximate Euclidean distance; in this embodiment, it is set .

[0194] Step 2.3.5: Calculate the target unit based on local reachability density. Local outlier :

[0195] (27)

[0196] Step 2.4: Constructing a quantitative evaluation unit for the coupled internal and external field twin model: This involves calculating the clamping force decay value. Combining cell thermodynamic deviation Construct a full-dimensional failure degree matrix for battery packs. :

[0197] Step 2.4.1: The model calculates the time-varying trajectory of the decaying clamping force. Perform an extreme value search to extract the maximum mechanical loosening force throughout the entire diagnostic cycle. Simultaneously, the local outlier sequence of all individual cells throughout the entire cycle is scanned to extract the maximum abnormal deviation of the thermodynamic state. ;

[0198] Step 2.4.2: Extract the... and Vectorized encapsulation is performed to construct a full-dimensional failure degree matrix representing the final damage state of the battery pack under accident conditions. :

[0199] (28)

[0200] Step 3: For the matrix Normalization and weighted calculations are performed to generate a battery failure quantification evaluation vector. The analysis of mechanical structure integrity loss and electrochemical performance degradation will generate quantitative failure assessment results and risk location reports, which will be submitted to the cloud server for output.

[0201] Step 3.1: Analyze the full-dimensional failure degree matrix of the battery pack. Extract the maximum mechanical loosening force Deviation from maximum anomaly Introduce a preset mechanical failure limit threshold. thermal runaway safety threshold The normalized index of mechanical risk was calculated using the linear ratio method. Electrochemical risk normalization index :

[0202] (29)

[0203] (30)

[0204] If the calculation result is greater than 1, its value is truncated and locked to 1; in this embodiment, ,

[0205] Step 3.2: Introduce the preset feature weight coefficient vector. ,in The normalized index , A point-to-point multiplication operation is performed with the feature weight coefficients to construct a battery failure quantification evaluation vector. :

[0206] (31)

[0207] In the formula, The weighted mechanical failure assessment components, This represents the weighted electrochemical failure assessment component. In this embodiment, , .

[0208] Step 3.3: Define the finite set of discrete failure states of the battery pack. :

[0209] (32)

[0210] in, For safety reasons, It is a single mechanical loosening state. This is a single precursor state to thermal runaway. This represents a critical state of structural-electrochemical concurrent cascade.

[0211] Step 3.4: Based on the battery failure quantification evaluation vector Components in and The following decision logic is executed:

[0212] like and Determined as The output is "The system structure is intact and the electrochemical state is stable"; if and Determined as Output a "Level 1 Mechanical Connection Failure Warning" and mark the battery box mounting point where the connection has become loose; if and Determined as Output "Early warning of Level 2 thermal runaway" and mark the cell numbers of cells exhibiting outliers; if and Determined as It outputs a "critical alarm for the tertiary structure-electrochemical cascade".

[0213] To verify the effectiveness of the proposed cloud-based battery assessment method for connected vehicle accidents based on internal and external field coupling twins in practical engineering, this embodiment utilizes a cloud server testing platform for scenario verification. The test data originates from an offline vehicle-to-everything (V2X) dataset of chassis bottoming-out accidents involving a certain type of new energy vehicle equipped with a ternary lithium-ion battery pack.

[0214] Set the full diagnostic time window duration for the cloud system to be [duration]. The system calibrates the anchor point time of the accident occurrence as follows: (i.e., extracting the contents before the accident) (Historical cache sequence). To address the damage to the physical communication link caused by the accident, the data is centrally configured with the vehicle-side voltage sensor at... Data transmission interruptions occurred intermittently. After experiencing a brief, severe impact and aftershocks, the vehicle... Completely halted. The preset model feature weight coefficients are set to... , Mechanical failure limit threshold thermal runaway safety threshold .

[0215] The mechanical risk normalization index was calculated. Electrochemical risk normalization index (Because it is greater than 1, the system truncates it and locks it as 1.0), the result is as follows Figure 3 As shown, Figure 3 The A-parameter derivation completes the continuous voltage time sequence after the disconnection time. Figure 3By mining data from 100 individual cells, the system identified a thermodynamic anomaly in cell number 42. Figure 3 The C-analysis examined the loosening of bolts in the vehicle battery pack throughout the entire diagnostic time window. Based on the test results, the cloud server test platform should output a Level 2 early warning for thermal runaway.

[0216] In another aspect, the present invention also discloses a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the steps of the method described above.

[0217] In another aspect, the present invention also discloses a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor performs the steps of the method described above.

[0218] In another embodiment provided in this application, a computer program product containing instructions is also provided, which, when run on a computer, causes the computer to execute any of the connected vehicle accident battery cloud assessment methods based on internal and external field coupling twins in the above embodiments.

[0219] It is understood that the systems, devices, and storage media provided in the embodiments of the present invention correspond to the methods provided in the embodiments of the present invention, and the explanations, examples, and beneficial effects of the relevant content can be referred to the corresponding parts of the above methods.

[0220] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially as a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid state disk (SSD)).

[0221] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0222] The various embodiments in this specification are described in a related manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0223] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A cloud-based assessment method for battery accidents in connected vehicles based on internal and external field coupling twins, characterized in that, Includes the following steps: Step 1: The cloud server receives a vehicle accident signal, defines a transient accident data window, extracts sensor data within the window, and constructs a multi-field incomplete time-series dataset of external impact and internal single-cell voltage. The multi-field incomplete time-series dataset specifically includes the battery's external mechanical feature matrix and internal electrochemical characteristic matrix. Step 2: Construct an internal and external field coupled twin model. Input the multi-field incomplete time series dataset of the external impact and the internal single-cell voltage into the internal and external field coupled twin model for inference transformation to obtain the battery pack full-dimensional failure degree matrix. Step 3: Normalize and weight the full-dimensional failure degree matrix of the battery pack to generate a quantitative assessment vector for battery failure, analyze the loss of mechanical structural integrity and the degree of electrochemical performance degradation, form a quantitative failure assessment result and risk positioning report, and submit it to the cloud server for output. Step 2 specifically includes, Step 2.1: Construct a data vector reconstruction unit for the coupled internal and external field twin model, perform data repair for periods when sensors are unable to upload data due to accidents, and generate a full diagnostic time window. Impact-voltage twin enhancement vector throughout the entire accident cycle; Step 2.2: Construct a multi-source torque co-dynamic calculation unit for the coupled internal and external field twin model, perform feature recombination on the impact-voltage twin enhancement vector, and calculate the clamping force decay value. ; Step 2.3: Construct a spatiotemporal outlier mining algorithm unit for the voltage entropy value of the internal and external field coupled twin model, perform feature recombination on the impact-voltage twin enhancement vector, and calculate the thermodynamic deviation of the battery cell. ; Step 2.4: Construct a quantitative evaluation unit for the coupled internal and external field twin model, and input the clamping force decay value. Combining cell thermodynamic deviation Construct a full-dimensional failure degree matrix for the battery pack; Step 2.3 includes, Step 2.3.1: For the voltage sampling points of each individual battery cell within the full diagnostic time window, obtain the minimum and maximum values ​​of the voltage data, and combine them with the preset number of intervals to divide the voltage state space into discretized intervals. Step 2.3.2, Statistics of the first The frequency distribution of the voltage data of the individual cells falls in each discrete interval, and the probability distribution vector is generated by calculating the frequency ratio of each interval. Step 2.3.3: Based on the probability of each interval in the probability distribution vector, calculate the Shannon entropy of the target single cell within the full diagnostic time window; traverse all single cells and perform the calculation process to generate a Shannon entropy feature vector covering all single cells within the full diagnostic time window; Step 2.3.4: Map the Shannon entropy eigenvectors of all individual cells to the state space, and calculate the target individual cell. The entropy Euclidean distance between the cell and the other individual cells is used to determine the closest one. The first group of individual cells constitutes the first Distance to the neighborhood, and based on this, the target single cell battery can be estimated. The corresponding locally reachable density; Step 2.3.5: In summary, the target single cell... The local achievable density of the target single cell and its neighboring cells is calculated. Local outlier This is the thermodynamic deviation of the battery cell.

2. The connected vehicle accident battery cloud assessment method based on internal and external field coupling twins according to claim 1, characterized in that: Step 1 includes, Step 1.1: The cloud server receives the sequence of collision message frames reported by the vehicle through V2I communication, and records the time when the last collision message frame is received as the time of the accident. Step 1.2: Based on the time of the accident, define a full diagnostic time window covering the accident precursor period and the accident evolution period. If the sensor continues to upload data after the accident, the complete data sequence is extracted; if the sensor is damaged due to the accident and the data is interrupted, the valid historical sequence before the interruption is extracted. Step 1.3: Extract the full diagnostic time window from the historical storage queue of the cloud database. The transverse vibration amplitude sequence and transverse vibration frequency sequence obtained by the internal vibration monitoring sensor are used to construct a set of external mechanical features of the battery; Step 1.4: Extract the full diagnostic time window from the historical storage queue of the cloud database. The voltage data of each individual cell within the battery are used to construct a set of internal electrochemical characteristics of the battery that includes time-series data of all individual cells; Step 1.5: Synchronously align and integrate the set of external mechanical features and the set of internal electrochemical characteristics of the battery in the time dimension to construct a multi-field incomplete time-series dataset of external shock and internal single-cell voltage for characterizing the transient characteristics of an accident.

3. The connected vehicle accident battery cloud assessment method based on internal and external field coupling twins according to claim 2, characterized in that: Step 2.1 includes, Step 2.1.1: Analyze the imported incomplete time-series dataset of external shocks and internal individual voltages to pinpoint the last moment of data interruption. ; Step 2.1.2: Addressing the blind spot period after sensor failure. At the time of extraction Lateral vibration amplitude The initial amplitude is given by time. transverse vibration frequency The vibration frequency, combined with the preset structural damping ratio The transverse vibration amplitude sequence during the failure period was generated by deduction. With transverse vibration frequency sequence : (6) Based on the physical characteristics of damped free vibration, a transverse vibration frequency sequence is set for the failure period. Maintaining a constant dominant frequency at the impact end, i.e. ; Step 2.1.3, based on the extraction time Battery cell voltage The initial voltage is combined with the preset theoretical steady-state recovery voltage. With cell polarization time constant The voltage recovery sequence during the failure period is generated by deduction. : (8) Step 2.1.4: The model compares the historical data sequences stored in the battery's external mechanical characteristic matrix and internal electrochemical characteristic matrix with the transverse vibration amplitude sequences generated during the failure period, respectively. and voltage recovery sequence during the failure period The sequences are spliced ​​together along the time dimension to reconstruct a time window covering the entire diagnosis. The impact-voltage twin enhancement vector throughout the entire accident diagnosis cycle is used as input for subsequent feature reconstruction calculations.

4. The battery cloud assessment method for connected vehicle accidents based on internal and external field coupling twins according to claim 3, characterized in that: Step 2.2 includes, Step 2.2.1: Analyze the input impulse-voltage twin enhancement vector of the model, and traverse each sampling time in the impulse-voltage twin enhancement vector. The instantaneous motion state of the bolt head is updated using the phase accumulation method; Step 2.2.2: Based on the transverse vibration amplitude sequence The amplitude of the lateral vibration at the current moment With accumulated phase Calculate the end-face friction shear force at that moment. : (10) In the formula: Indicates the elastic modulus of the bolt material; This represents the moment of inertia on the cross-section of the bolt; Indicates the effective bending length of the bolt. Indicates the bolt bending factor. The bending stiffness coefficient of bolt heads in the battery box installation scenario; Step 2.2.3, based on the aforementioned frictional shear force Minimum contact radius of end face Maximum end face contact radius end face friction coefficient Average pressure at the end face The parameters are obtained by inverse numerical iteration. ; Step 2.2.4, from the already obtained Substitute into the following formula to calculate the end face torque. : (13) In the formula, This represents the circumferential integral angle in polar coordinates centered on the bolt axis. Step 2.2.5: Calculate the thread friction tangential force according to the following formula. : (14) In the formula: ; and These represent the minor diameter and major diameter of the thread, respectively. Represents the thread flank angle; Indicates the effective bending length of the bolt; Indicates the pitch; Step 2.2.6: Based on the aforementioned thread friction tangential force Thread flank angle Thread lead angle Thread friction coefficient The average pressure determined by the bolt clamping force and the thread contact area The unknown parameters are obtained by numerical iteration. ; Step 2.2.7: Calculate the thread torque according to the following formula. : (17) Step 2.2.8: Calculate the pitch torque according to the following formula. : (18) The calculated bolt clamping force at the current moment; Step 2.2.9: Construct a system containing pitch torque. Thread torque and end face torque The transient dynamic equilibrium equations, traversing the sampling times Calculate the instantaneous angular acceleration of the bolt. : (19) In the formula: Indicates the moment of inertia of the bolt; Step 2.2.10: Using the discrete-time integration method, based on the instantaneous angular acceleration... Update the current time angular velocity of rotation Thus, the cumulative rotational angular displacement is derived. Calculate the time corresponding to the sampling point Degree of decrease in clamping force : (22) In the formula: For battery housing rigidity; This refers to the bolt stiffness.

5. The battery cloud assessment method for connected vehicle accidents based on internal and external field coupling twins according to claim 3, characterized in that: Step 2.4 includes, Step 2.4.1: Model analyzes the time-varying trajectory of the decaying clamping force generated during the simulation. Perform an extreme value search to extract the maximum mechanical loosening force throughout the entire diagnostic cycle. Simultaneously, the local outlier sequence of all individual cells throughout the entire cycle is scanned to extract the maximum abnormal deviation of the thermodynamic state. ; Step 2.4.2: Extract the maximum mechanical loosening force. Deviation from maximum anomaly Vectorized encapsulation is performed to construct a full-dimensional failure degree matrix representing the final damage state of the battery pack under accident conditions.

6. The battery cloud assessment method for connected vehicle accidents based on internal and external field coupling twins according to claim 1, characterized in that: Step 3 includes the following steps: Step 3.1: Analyze the full-dimensional failure matrix of the battery pack and extract the maximum mechanical loosening force. Deviation from maximum anomaly Introduce a preset mechanical failure limit threshold. thermal runaway safety threshold The normalized index of mechanical risk was calculated using the linear ratio method. Electrochemical risk normalization index If the calculation result is greater than 1, its value is truncated and locked to 1; Step 3.2: Introduce the preset feature weight coefficient vector. ,in The mechanical risk normalization index is used to... Electrochemical risk normalization index Perform point-to-point multiplication with the feature weight coefficients to construct a battery failure quantification evaluation vector: (31) In the formula, The weighted mechanical failure assessment components, The weighted electrochemical failure assessment components; Step 3.3: Define a finite discrete set of battery pack failure states, including a safe state, a single mechanical loosening state, a single thermal runaway precursor state, and a structural-electrochemical concurrent cascade critical state. ; Step 3.4: Based on the components in the battery failure quantification evaluation vector and 1. A finite set of discrete failure states of a battery pack The following decision logic is executed: like and The output is "The system structure is intact and the electrochemical state is stable"; if and Output a "Level 1 Mechanical Connection Failure Warning" and mark the battery box mounting point where the connection has become loose; if and Output "Early warning of Level 2 thermal runaway" and mark the cell numbers of cells exhibiting outliers; if and It outputs a "critical alarm for the tertiary structure-electrochemical cascade".

Images