CN-121601844-B - Battery thermal runaway identification method based on pressure mutation

CN121601844BCN 121601844 BCN121601844 BCN 121601844BCN-121601844-B

Abstract

The invention relates to the technical field of battery thermal runaway identification, and particularly discloses a battery thermal runaway identification method based on pressure mutation, which comprises the following steps of S1, collecting a pressure vibration signal, and judging whether an abnormal event occurs in a battery pack based on the pressure vibration signal; S2, synchronously acquiring original waveform data of all pressure sensors in an event time window to form a signal data set when an abnormal event occurs, S3, conducting mode separation is conducted on the complete signal data set to obtain a first conducting mode signal subset and a second conducting mode signal subset, S4, a first space coordinate solution set and a second space coordinate solution set are calculated based on the first conducting mode signal subset and the second conducting mode signal subset respectively, and S5, cross verification is conducted on the first space coordinate solution set and the second space coordinate solution set to determine a thermal runaway physical source point. The invention improves the safety and controllability of the battery system in the whole operation process.

Inventors

YUAN JIE
ZHANG RAN
CHEN XIAORONG
ZHAO XU
WAN HAO
ZHANG ZHENTING
DING YUWEN
HU WEI
FU YONG

Assignees

广州科通达信息科技有限公司

Dates

Publication Date: 20260508
Application Date: 20260128

Claims (5)

1. The battery thermal runaway identification method based on the pressure mutation is characterized by comprising the following steps of: S1, arranging a plurality of pressure sensors in a battery pack to acquire pressure vibration signals, and judging whether an abnormal event occurs in the battery pack or not based on the pressure vibration signals; S2, synchronously collecting original waveform data of all pressure sensors in an event time window when an abnormal event occurs, and forming a signal data set; s3, conducting mode separation is carried out on the complete signal data set, and a first conducting mode signal subset and a second conducting mode signal subset are obtained; S4, calculating a first space coordinate solution set and a second space coordinate solution set based on the first conduction mode signal subset and the second conduction mode signal subset respectively; S5, performing cross verification on the first space coordinate solution set and the second space coordinate solution set to determine a thermal runaway physical source point; conduction mode separation includes: Extracting signals of each pressure sensor in multiple dimensions from a signal data set, and combining the signals of all the sensors to form a space-time signal matrix; processing the space-time signal matrix by using a modal decomposition method to obtain a plurality of global conduction mode components, wherein the modal decomposition method comprises multi-scale step modal decomposition; Calculating, for each global conduction mode component, its signal propagation characteristics between different pressure sensors; matching the signal propagation characteristics with preset solid structure conduction characteristics and non-solid medium conduction characteristics; combining global conduction mode components matched as solid structure conduction features and non-solid medium conduction features into a first conduction mode signal subset and a second conduction mode signal subset respectively; calculating the first set of spatial coordinate solutions includes: extracting a first peak moment when a signal reaches each pressure sensor based on the first conduction mode signal subset; Calculating the difference value of the wave crest moments among all non-repeated pressure sensor pairs based on the first wave crest moment of all the pressure sensors, and sequencing the difference values of the wave crest moments to obtain a first sequence; traversing each node except for a sensor node in a network topology model of a battery pack solid structure as an assumed source point, wherein the sensor nodes in the network topology model correspond to the mounting positions of the pressure sensors one by one, and the edges represent continuous solid force transmission paths and are associated with theoretical wave velocity values; for each assumed source point, calculating theoretical propagation time of vibration waves from the assumed source point to each sensor node based on the network topology model and theoretical wave velocity values of each edge; calculating the difference value of the theoretical propagation time between all the non-repeated pressure sensor pairs, and sequencing the difference value of the theoretical propagation time to obtain a second sequence; And calculating the root mean square error of the second sequence corresponding to each hypothesized source point and the first sequence, screening out hypothesized source points with the root mean square error smaller than a preset error tolerance, and forming the first space coordinate solution set by three-dimensional space coordinates of all the screened hypothesized source points.
2. The method for identifying thermal runaway of a battery based on abrupt pressure change according to claim 1, the method is characterized in that the step of collecting the signal data set comprises the following steps: Continuously calculating the instantaneous energy of the signal output by each pressure sensor, wherein the instantaneous energy is the integral of the square of the signal amplitude in a short time window; when the duration of the instant energy continuously exceeding the preset background energy threshold reaches the preset duration, extending the moment reaching the preset duration forwards for a preset duration to obtain the starting moment of the event; setting a starting point of an event time window as an event starting moment, wherein the length of the event time window is preset manually; and in an event time window, synchronously acquiring and storing the original waveform data of all the pressure sensors to form a signal data set.
3. The method of claim 2, wherein calculating the second subset of conduction mode signals comprises: Extracting a first peak moment when a signal reaches each pressure sensor based on the second subset of conduction mode signals; calculating the difference value of the wave crest moments among all non-repeated pressure sensor pairs based on the first wave crest moment of all the pressure sensors, and sequencing the difference values of the wave crest moments to obtain a third sequence; Setting the average propagation velocity v of pressure vibration in a non-solid medium inside the battery pack; for any two different pressure sensors i and j, the coordinates of the pressure sensors i and j are (xi, yi, zi) and (xj, yj, zj), respectively, the following equation is established: ; wherein, (x, y, z) represents an event source point coordinate to be solved, and Δtij represents a difference value of peak moments corresponding to the pressure sensors i and j; All equations form an equation set, a numerical optimization algorithm is adopted to solve the equation set, one or more event source point coordinates are found, and the total residual error generated after the event source point coordinates are substituted into all the equations is minimum; all event source point coordinates constitute a second set of spatial coordinate solutions.
4. A method of identifying thermal runaway in a battery based on abrupt pressure change according to claim 3, wherein determining the physical source point of thermal runaway comprises: Calculating Euclidean distances between all coordinate point pairs in the first space coordinate solution set and the second space coordinate solution set, and marking the corresponding coordinate point pair as a target pair if the Euclidean distances are smaller than a preset space distance threshold; If the proportion of the number of the target pairs to the total number of the coordinate pairs exceeds a preset threshold, calculating geometric center points of two coordinate points in the target pairs, and selecting the geometric center point with the highest occurrence frequency as a thermal runaway physical source point.
5. The method of claim 4, wherein calculating a first set of spatial coordinate solutions further comprises: After each time of thermal runaway physical source point determination, storing the coordinates of the thermal runaway physical source point corresponding to the abnormal event and the first sequence; When the number of the stored abnormal events reaches the preset number, a single thermal runaway physical source point coordinate of the stored abnormal events is taken as a vibration starting point, and a network topology model is combined to obtain a corresponding second sequence, and the second sequence is recorded as a check sequence; obtaining errors between the check sequences and the corresponding first sequences, recording the errors as check errors, taking the minimized total check errors as an optimization problem, and adopting a gradient descent method to adjust theoretical wave velocity values of each side in the network topology model; and replacing the original theoretical wave velocity values of each side in the network topology model with the adjusted theoretical wave velocity values.

Description

Battery thermal runaway identification method based on pressure mutation Technical Field The invention relates to the technical field of thermal runaway identification, in particular to a battery thermal runaway identification method based on pressure mutation. Background During operation of a power battery system, thermal runaway of the battery is often accompanied by complex and dramatic physical changes, the early stages of which often appear as local structural abrupt changes in stress and rapid disturbances in the pressure environment inside the battery pack. Because the internal structure of the battery pack is highly integrated, the space is closed, the propagation path is complex, and pressure abrupt signals caused by thermal runaway show significantly different propagation characteristics in different media and structures, the related signals are highly coupled in time, space and morphology. In the prior art, the monitoring of the thermal runaway of the battery is based on the slow variable information such as temperature, voltage or gas concentration, and the like, so that the physical initial position of the sudden thermal runaway event is difficult to timely and accurately reflect, and the method based on single pressure signal or simple threshold judgment is difficult to reliably identify and position the thermal runaway event under the multi-source interference and complex propagation environment. Especially, under the condition that a solid structure and a non-solid medium coexist in the battery pack, pressure abrupt signals often propagate along different paths at the same time, so that signal arrival characteristics are mixed, source point directions are not clear, and therefore large uncertainty exists in identification of a thermal runaway physical source point. Therefore, how to reliably identify the thermal runaway event of the battery and accurately judge the physical occurrence position of the thermal runaway event of the battery under the complex propagation condition by utilizing the transient information carried by the abrupt change of the internal pressure of the battery pack without damaging the structure of the battery pack becomes a key technical problem to be solved in the field of safety monitoring of power batteries. Disclosure of Invention The invention aims to provide a battery thermal runaway identification method based on pressure mutation, which solves the technical problems. The aim of the invention can be achieved by the following technical scheme: a method for identifying thermal runaway of a battery based on abrupt pressure change, comprising the steps of: S1, arranging a plurality of pressure sensors in a battery pack to acquire pressure vibration signals, and judging whether an abnormal event occurs in the battery pack or not based on the pressure vibration signals; S2, synchronously collecting original waveform data of all pressure sensors in an event time window when an abnormal event occurs, and forming a signal data set; s3, conducting mode separation is carried out on the complete signal data set, and a first conducting mode signal subset and a second conducting mode signal subset are obtained; S4, calculating a first space coordinate solution set and a second space coordinate solution set based on the first conduction mode signal subset and the second conduction mode signal subset respectively; S5, performing cross verification on the first space coordinate solution set and the second space coordinate solution set to determine a thermal runaway physical source point. As a further aspect of the invention, collecting the signal data set comprises: Continuously calculating the instantaneous energy of the signal output by each pressure sensor, wherein the instantaneous energy is the integral of the square of the signal amplitude in a short time window; when the duration of the instant energy continuously exceeding the preset background energy threshold reaches the preset duration, extending the moment reaching the preset duration forwards for a preset duration to obtain the starting moment of the event; setting a starting point of an event time window as an event starting moment, wherein the length of the event time window is preset manually; and in an event time window, synchronously acquiring and storing the original waveform data of all the pressure sensors to form a signal data set. As a further aspect of the invention, the conduction mode separation comprises: Extracting signals of each pressure sensor in multiple dimensions from a signal data set, and combining the signals of all the sensors to form a space-time signal matrix; processing the space-time signal matrix by using a modal decomposition method to obtain a plurality of global conduction mode components, wherein the modal decomposition method comprises multi-scale step modal decomposition; Calculating, for each global conduction mode component, its signal propagation characteristics between differe