Search

CN-121997004-A - Network-connected vehicle accident battery cloud assessment method based on internal and external field coupling twin body

CN121997004ACN 121997004 ACN121997004 ACN 121997004ACN-121997004-A

Abstract

The invention discloses a network-connected vehicle accident battery cloud assessment method based on an internal and external field coupling twin body, which comprises the steps of firstly defining a transient accident data window, extracting sensor data in the window by a vehicle network cloud platform, constructing a multi-field incomplete time sequence data set of external impact and internal single voltage, secondly constructing an internal and external field coupling twin body model, deducing and repairing accident data blind areas by using a damping and RC polarization attenuation algorithm to obtain an impact-voltage twin enhancement vector, calculating a mechanical clamping force declining value and a battery cell thermodynamic anomaly deviation degree by a multi-source moment cooperative dynamics unit and a voltage entropy time-space outlier mining unit, constructing a full-dimension failure degree matrix, carrying out normalized weighted calculation to generate a quantized assessment vector, and outputting a risk positioning report. According to the invention, under the extreme accident working condition, accurate quantification and grading early warning can be realized on the accident failure state of the battery pack, and an important technical guarantee is provided for the emergency treatment and management strategy optimization of the battery safety.

Inventors

  • WANG YUEFEI
  • Yi Zhejin
  • ZHAO RUIHUA
  • YUAN YICHEN
  • Zhou Yaman
  • ZHAO JINGLONG

Assignees

  • 合肥工业大学

Dates

Publication Date
20260508
Application Date
20260409

Claims (8)

  1. 1. The network-connected vehicle accident battery cloud assessment method based on the internal and external field coupling twin body is characterized by comprising the following steps of: Step 1, a cloud server receives an accident signal of a vehicle, defines a transient accident data window, extracts sensor data in the window, and constructs a multi-field incomplete time sequence data set of external impact and internal single voltage, wherein the multi-field incomplete time sequence data set specifically comprises a battery external mechanical characteristic matrix and an internal electrochemical characteristic matrix; step 2, constructing an internal and external field coupling twin model, inputting the multi-field incomplete time sequence data set of the external impact and the internal single voltage into the internal and external field coupling twin model for reasoning transformation, and obtaining a full-dimension failure degree matrix of the battery pack; And step 3, carrying out normalization and weighted calculation on the full-dimension failure degree matrix of the battery pack to generate a battery failure quantitative evaluation vector, analyzing the mechanical structure integrity loss and the electrochemical performance attenuation degree to form a quantitative failure evaluation result and a risk positioning report, and submitting the quantitative failure evaluation result and the risk positioning report to a cloud server for output.
  2. 2. The network-connected vehicle accident battery cloud assessment method based on the internal and external field coupling twin body according to claim 1, wherein the step 1 comprises, Step 1.1, a cloud server receives a collision message frame sequence reported by a vehicle through V2I communication, and records the moment of receiving the last collision message frame as the accident occurrence moment; Step 1.2, defining a full diagnosis time window covering the accident precursor period and the accident evolution period based on the accident occurrence time If the sensor continuously uploads data after an accident, extracting a complete data sequence; if the sensor is damaged due to accident, the data is interrupted, and then the effective history sequence before interruption is extracted; Step 1.3, extracting the full diagnosis time window from a history storage queue of a cloud database A transverse vibration amplitude sequence and a transverse vibration frequency sequence acquired by the internal vibration monitoring sensor are used for constructing a battery external mechanical feature set; step 1.4, extracting the full diagnosis time window from a history storage queue of a cloud database The voltage data of each single battery in the battery is used for constructing a battery internal electrochemical characteristic set containing all single time sequence data; And 1.5, synchronously aligning and integrating the external mechanical characteristic set and the internal electrochemical characteristic set of the battery in a time dimension to construct a multi-field incomplete time sequence data set for representing external impact and internal single voltage of the transient characteristics of accidents.
  3. 3. The network-connected vehicle accident battery cloud assessment method based on the internal and external field coupling twin body according to claim 2, wherein the step 2 specifically comprises, Step 2.1, constructing a data vector reconstruction unit of an internal and external field coupling twin body model, repairing data in a failure period when a sensor cannot upload data due to an accident, and generating a coverage full diagnosis time window Impact-voltage twinning enhancement vector of accident full period; 2.2, constructing a multisource moment cooperative dynamics calculation unit of an internal and external field coupling twin model, carrying out characteristic recombination on the impact-voltage twin enhancement vector, and calculating a clamping force decay value ; Step 2.3, constructing a voltage entropy value space-time outlier mining algorithm unit of an internal and external field coupling twin model, carrying out characteristic recombination on the impact-voltage twin enhancement vector, and calculating the thermodynamic deviation of the battery cell ; Step 2.4, constructing a quantitative evaluation unit of the internal and external field coupling twin body model, and reducing the clamping force decay value Combined cell thermodynamic bias And constructing a full-dimension failure degree matrix of the battery pack.
  4. 4. The network-connected vehicle accident battery cloud assessment method based on the internal and external field coupling twin body according to claim 3, wherein step 2.1 comprises, Step 2.1.1, analyzing the imported multi-field incomplete time sequence data set of the external impact and the internal single voltage by the model, and positioning the final moment of data interruption ; Step 2.1.2, aiming at the dead zone period after the sensor fails By the time of extraction Transverse vibration amplitude of (2) For initial amplitude, in time of day Transverse vibration frequency of (2) For vibration frequency, combined with preset structural damping ratio Deducing a sequence of transverse vibration amplitudes that generates a failure period And transverse vibration frequency sequence : (6) Setting a transverse vibration frequency sequence of a failure period according to damping free vibration physical characteristics Maintained at a constant dominant frequency of striking the end, i.e ; Step 2.1.3, at the time of extraction Is a battery cell voltage of (a) For initial voltage, combined with preset theoretical steady-state recovery voltage And cell polarization time constant Deducing a voltage recovery sequence that generates a failure period : (8) Step 2.1.4, the model respectively compares the historical data sequences stored in the external mechanical characteristic matrix and the internal electrochemical characteristic matrix of the battery with the transverse vibration amplitude sequences of the failure period generated by deduction Voltage recovery sequence for failure period Splicing the sequences in the time dimension, and reconstructing to generate a coverage full diagnosis time window The impulse-voltage twin enhancement vector of the accident full diagnosis period is used as the input of the subsequent characteristic recombination calculation.
  5. 5. The network-connected vehicle accident battery cloud assessment method based on the internal and external field coupling twin body according to claim 3, wherein step 2.2 comprises, Step 2.2.1, analyzing the input impact-voltage twin enhancement vector by the model, and traversing each sampling moment in the impact-voltage twin enhancement vector Updating the instantaneous motion state of the bolt head by adopting a phase accumulation method; Step 2.2.2 based on the transverse vibration amplitude sequence Transverse vibration amplitude at the present moment in the (b) And cumulative phase Calculating the end face friction shear force at the moment : (10) Wherein: Representing the elastic modulus of the bolt material; Representing the moment of inertia in the cross section of the bolt; Indicating the effective bending length of the bolt, The bending factor of the bolt is indicated, The bending rigidity coefficient of the bolt head in the installation scene of the battery box body; step 2.2.3, based on the frictional shear force Minimum end face contact radius Maximum end contact radius Coefficient of friction of end face Average pressure at end face Inverse parameter determination by numerical iteration ; Step 2.2.4, from the determined Substituting the following to calculate the end face torque : (13) In the formula, Representing a circumferential integral angle in a polar coordinate system centered on the bolt axis; Step 2.2.5 calculating the thread friction tangential force according to the following formula : (14) Wherein: ; And Respectively representing a small thread diameter and a large thread diameter; Representing thread flank angles; indicating the effective bending length of the bolt; Representing the pitch; Step 2.2.6, tangential force based on said thread friction Thread flank angle Lead angle of thread Coefficient of thread friction Average pressure determined by bolt clamping force and thread contact area The unknown parameters are reversely calculated by a numerical iteration method ; Step 2.2.7, calculating the thread torque according to : (17) Step 2.2.8, calculating the Pitch Torque according to : (18) The bolt clamping force at the current moment is calculated; Step 2.2.9, build up comprising pitch torque Screw torque End face torque Traversing the sampling moment of transient dynamic balance equation Calculating the instantaneous rotational acceleration of the bolt : (19) Wherein: Representing the rotational inertia of the bolt; Step 2.2.10, based on the instantaneous cornering angular acceleration, using a discrete time integration method Updating the current time Is fixed at a rotational angular velocity of (2) Thereby deriving cumulative rotational displacement Calculating the corresponding moment of the sampling point Degree of decrease in clamping force : (22) Wherein: is the rigidity of the battery box body; Is the bolt stiffness.
  6. 6. The network-connected vehicle accident battery cloud assessment method based on the internal and external field coupling twin body according to claim 3, wherein step 2.3 comprises, Step 2.3.1, acquiring the minimum value and the maximum value of voltage data aiming at the voltage sampling points of all the single batteries in the full diagnosis time window, and dividing a voltage state space into discretization intervals by combining the preset interval number; step 2.3.2, statistics of the first The frequency distribution of the voltage data of the single battery falls in each discrete interval, and a probability distribution vector is generated by calculating the frequency duty ratio of each interval; Step 2.3.3, solving shannon entropy of the target single battery in the full diagnosis time window based on each interval probability in the probability distribution vector; Step 2.3.4, mapping shannon entropy feature vectors of all the single batteries into a state space, and calculating a target single battery Entropy Euclidean distance between rest single battery to determine nearest The first unit battery is formed by assembling Distance neighborhood and calculate the target single battery accordingly Corresponding local reachable densities; step 2.3.5, synthesizing the target single battery And calculating the local reachable density of other single batteries in the neighborhood to obtain the target single battery Local outlier factors of (a) As a thermodynamic deviation of the cell.
  7. 7. The network-connected vehicle accident battery cloud assessment method based on the internal and external field coupling twin body according to claim 3, wherein step 2.4 comprises, Step 2.4.1, modeling the time-varying track of the clamping force decay generated by deduction Searching extremum, extracting maximum mechanical release force in whole diagnosis period Scanning local outlier factor sequences of all single batteries in the whole period at the same time, and extracting the maximum abnormal deviation degree of the thermodynamic state ; Step 2.4.2 maximum mechanical release force to extract Degree of deviation from maximum abnormality And vectorizing packaging is carried out, and a full-dimension failure degree matrix representing the final damage state of the battery pack under the accident working condition is constructed.
  8. 8. The network-connected vehicle accident battery cloud assessment method based on the internal and external field coupling twin body according to claim 1, wherein the step 3 comprises the following steps: Step 3.1, analyzing the full-dimension failure degree matrix of the battery pack, and extracting the maximum mechanical release force Degree of deviation from maximum abnormality Introducing a preset mechanical failure limit threshold Safety threshold for thermal runaway Calculating mechanical risk normalization indexes by using linear ratio method Normalized index to electrochemical risk If the calculation result is greater than 1, the value is truncated and locked to be 1; step 3.2, introducing a preset characteristic weight coefficient vector Wherein Normalizing the mechanical risk index Normalized index of electrochemical risk Performing point-to-point multiplication operation with the characteristic weight coefficient to construct a battery failure quantization evaluation vector: (31) in the formula, To evaluate the components for weighted mechanical failure, Evaluating the components for weighted electrochemical failure; step 3.3 defining a set of finite discrete failure states of the battery pack including a safe state, a single mechanical release state, a single thermal runaway precursor state, and a structure-electrochemical concurrency cascade critical state ; Step 3.4, quantifying the components in the evaluation vector based on the battery failure And (3) with Battery pack limited discrete failure state set The following decision logic is executed: If it is And is also provided with Output "the system is complete and the electrochemical state is stable", if And is also provided with Outputting a first-level mechanical connection failure early warning, marking the mounting point of the battery box body with looseness, if And is also provided with Outputting early warning of secondary thermal runaway, and marking the number of the monomer battery with outlier abnormality, if And is also provided with Output a "tertiary structure-electrochemical cascade critical alarm".

Description

Network-connected vehicle accident battery cloud assessment method based on internal and external field coupling twin body Technical Field The invention relates to the technical field of battery pack safety prediction of intelligent network-connected vehicles, in particular to a network-connected vehicle accident battery cloud assessment method based on an internal-external field coupling twin body. Background The intelligent network-connected automobile industry of new energy is vigorously developed, and the safety of the intelligent network-connected automobile is directly related to the running reliability of a vehicle and the life safety of passengers by taking a power battery as a core energy storage component. In an actual road traffic environment, vehicles inevitably face sudden accidents such as collision and bottom scraping, and severe mechanical impact generated by the accidents not only can cause mechanical structure damages such as deformation of a battery pack box body and loosening of bolts, but also can induce the attenuation of electrochemical performance in a battery, so that catastrophic consequences such as thermal runaway and even fire explosion are caused. However, the conventional evaluation means have obvious limitations that the conventional Battery Management System (BMS) focuses on threshold monitoring of voltage and temperature, and is difficult to sense tiny failures in the structural level, and the high-precision finite element simulation calculation amount is huge, so that the real-time requirement of an accident scene is difficult to meet. Meanwhile, severe collision often causes damage or communication interruption of a sensor to form a data 'blind area', the existing data driving method extremely depends on data integrity, cannot work effectively once a signal is lost, and lacks joint quantitative analysis on multi-factor characteristics such as machinery, electrochemistry and the like. Therefore, how to construct a digital twin evaluation method with real-time monitoring and deep fusion of mechanical-electrochemical dual characteristics by utilizing strong computing capacity of a cloud end in an intelligent networking environment, and realize full-dimension quantification of the risk from structure to thermal runaway of an accident vehicle battery pack, so that the safety of passengers and the emergency treatment efficiency of sudden accidents are improved, and the digital twin evaluation method becomes a technical bottleneck which is urgently needed to break through in the current new energy automobile safety emergency management field. Disclosure of Invention The invention provides a network-connected vehicle accident battery cloud evaluation method based on an internal-external field coupling twin body, which overcomes the defects that in the prior art, single factor evaluation is incomplete after a vehicle accident, calculation force at the vehicle end is limited and cannot be operated for complex and real-time calculation, and meanwhile, a sensor is easy to damage and data is interrupted. Therefore, when an emergency accident occurs, the evaluation capability of the overall state of the current battery pack is improved, the safety of passengers and the accident decision efficiency are improved, and a reference basis is provided for emergency treatment. In order to achieve the above purpose, the present invention adopts the following technical scheme: a network-connected vehicle accident battery cloud evaluation method based on an internal and external field coupling twin body comprises the following steps of executing by computer equipment, Step 1, a cloud server receives a vehicle accident signal, defines a transient accident data window, extracts sensor data in the window, and constructs a multi-field incomplete time sequence data set of external impact and internal single voltageThe multi-field incomplete time sequence data set specifically comprises a battery external mechanical characteristic matrixAnd internal electrochemical characteristic matrix; Step 2, constructing an internal and external field coupling twin body model, and integrating the external impact and internal single voltage into a multi-field incomplete time sequence data setInputting the model to perform reasoning transformation to obtain a full-dimension failure degree matrix of the battery pack; Step 3, matrix alignmentNormalization and weighting calculation are carried out, and a battery failure quantitative evaluation vector is generatedAnd analyzing the mechanical structural integrity loss and the electrochemical performance attenuation degree to form a quantized failure evaluation result and a risk positioning report, and submitting the quantized failure evaluation result and the risk positioning report to a cloud server for output. Further, step 1 includes the steps of: Step 1.1, a cloud server receives a collision message frame sequence reported by a vehicle through V2I communication The mo