CN-122017643-A - Online sensing system for full life cycle security situation of lithium battery box

CN122017643ACN 122017643 ACN122017643 ACN 122017643ACN-122017643-A

Abstract

The invention discloses an on-line sensing system for a full life cycle safety situation of a lithium battery box, which relates to the technical field of lithium battery safety monitoring and comprises a signal excitation and acquisition unit and a data processing and control unit, wherein the signal excitation and acquisition unit is arranged in the battery box; the signal excitation and acquisition unit is configured to apply a small alternating current excitation signal to the battery cell and acquire a voltage-current response thereof. According to the online perception system for the full life cycle security situation of the lithium battery box, the defect of insufficient adaptability of a single model in the full life cycle of the battery is effectively overcome by constructing a collaborative perception architecture with weighted fusion of double-engine state estimation and dynamic credibility. By introducing a feedback verification closed-loop mechanism, the online secondary verification and self-correction of the internal state estimation result are realized, so that the accuracy and reliability of early thermal runaway risk early warning are greatly improved, and the occurrence of false alarm and false alarm is remarkably reduced.

Inventors

HE PENG
YANG YANG
WANG XIAOHU
LIU YUTAO

Assignees

扬州中远海运重工有限公司

Dates

Publication Date: 20260512
Application Date: 20260129

Claims (10)

1. An on-line sensing system for a full life cycle safety situation of a lithium battery box is characterized by comprising a signal excitation and collection unit and a data processing and control unit, wherein the signal excitation and collection unit is arranged in the battery box and is electrically connected with the signal excitation and collection unit, the signal excitation and collection unit is configured to apply a micro alternating current excitation signal to a battery cell and collect voltage and current responses of the battery cell, and the data processing and control unit comprises: A dual engine state estimation module having a simplified physical engine configured to extract internal state parameters of the battery based on the voltage-current response data and a data-driven engine configured to learn and map internal state of health of the battery based on a history of battery operation and real-time external electrical parameter data in parallel; The dynamic credibility weighting fusion device is connected to the double-engine state estimation module and is configured to receive the preliminary estimation values of the same internal state parameters of the simplified physical engine and the data driving engine, allocate proper weights for the preliminary estimation values of the two engines according to the current working condition and the service life stage of the battery, perform first fusion calculation and output a primary fusion state quantity; The safety situation scoring and predicting model is connected to the dynamic credibility weighted fusion device and is configured to receive the primary fusion state quantity, comprehensively calculate the current safety situation scoring of the battery and simultaneously predict the battery state at the future moment in a short term based on the current state; The feedback verification closed loop module is connected between the safety situation score and the prediction model and the dynamic credibility weighting fusion device, and is configured to continuously compare a short-term prediction value made by the safety situation score and the prediction model with an actual measurement value subsequently acquired by the signal excitation and acquisition unit, and transmit the generated prediction deviation information as a feedback signal to the dynamic credibility weighting fusion device so as to drive the dynamic credibility weighting fusion device to adjust the credibility of the dynamic credibility weighting fusion device to the two engine output values in the dual-engine state estimation module; the system forms an enhanced closed loop of sensing, fusing, predicting, verifying and correcting through the feedback verification closed loop module, and finally outputs a security situation score which tends to be stable after multiple iterations for early warning.
2. The system for online sensing of full life cycle security situations in a lithium battery case as recited in claim 1, wherein the simplified physical engine is a reduced electrochemical-thermal coupling model focused on online extraction of critical internal state parameters of internal resistance and lithium ion diffusion coefficient of the battery.
3. The system of claim 1 wherein the data driven engine is a lightweight deep learning network model with inputs for battery voltage, current and temperature data streams and outputs as estimates of battery internal health.
4. The on-line perception system for full life cycle safety situation of lithium battery box according to claim 1, wherein the basis for weight distribution by the dynamic credibility weighting fusion device comprises working condition information of standing, charging or discharging of the battery and battery life initial, middle or later stage information divided according to cycle times.
5. The on-line perception system for the full life cycle safety situation of the lithium battery box according to claim 1, wherein the feedback verification closed loop module feeds the prediction deviation information back to the dynamic credibility weighting fusion device to realize on-line self-adaptive adjustment of the credibility of the output result of the double-engine state estimation module, so that the secondary verification and correction of the primary fusion state quantity are completed.
6. The system for online sensing of full life cycle safety situation of lithium battery box according to claim 1, wherein the frequency components of the micro alternating current excitation signals applied by the signal excitation and acquisition unit are adaptively selected according to the current health state and temperature condition of the battery.
7. The system of claim 1, wherein the safety profile score and the safety profile score calculated by the predictive model are a continuous number ranging from zero to one, the number being associated with a thermal runaway risk level of the battery.
8. The system for online sensing of full life cycle security posture of a lithium battery box of claim 1, wherein said data driven engine is capable of receiving an estimate of internal state parameters provided by said reduced physical engine as its auxiliary input features for enhancing accuracy of its mapping relationship.
9. The system for online perception of full life cycle safety situation of lithium battery box according to claim 1, wherein final output of the system is safety situation score and internal state quantity which are output by the safety situation scoring and predicting model and tend to be stable after the enhanced closed loop is iterated for a plurality of times.
10. The system for online sensing of full life cycle safety situation of lithium battery box according to any one of claims 1 to 9, further comprising a cloud data platform communicatively connected to the data processing and control unit, wherein the cloud data platform is configured to perform statistical analysis on operation modes and failure characteristics of the battery population in the full life cycle, and synchronize the analysis result downstream to a local safety situation scoring and prediction model of the individual battery, so as to implement continuous evolution of the model.

Description

Online sensing system for full life cycle security situation of lithium battery box Technical Field The invention relates to the technical field of lithium battery safety monitoring, in particular to an online sensing system for a full life cycle safety situation of a lithium battery box. Background In the field of safety monitoring of lithium battery boxes, the prior art mainly relies on online monitoring of external electrical parameters such as voltage, current, surface temperature and the like, and safety early warning is carried out based on a preset fixed threshold value. Such a solution is imperative in the face of complex evolution of the full life cycle of the battery. Particularly, the battery has a slight internal state change in an early use stage, and the internal critical parameters such as internal resistance and interface film thickness can undergo nonlinear evolution along with the increase of the cycle number. The existing system generally lacks real-time and self-adaptive sensing capability on internal states, and a static model adopted by the system cannot accurately capture tiny characteristic changes in the early stage of battery aging, so that early prediction on serious faults such as thermal runaway and the like is extremely unreliable. The method is characterized in that false alarms frequently occur in the early stage of the service life of the battery, and real risks are easily missed in the later stage of aging. While some improvements attempt to introduce more sensors or complex algorithms, they often either require extensive historical data for offline training, or are computationally overly burdened to enable online continuous updates, failing to fundamentally address the core difficulty of early risk accurate sensing throughout the battery's entire cycle from a brand-new state to the end of aging. Therefore, an online sensing system for the full life cycle safety situation of a lithium battery box aims to solve the practical problem that the prior art cannot realize real-time accurate prediction of early thermal runaway risk in the full life cycle of a battery. Disclosure of Invention The invention aims to provide an on-line sensing system for the full life cycle safety situation of a lithium battery box, so as to solve the problems in the background technology. In order to solve the technical problems, the invention provides the technical scheme that the on-line sensing system for the full life cycle safety situation of the lithium battery box comprises a signal excitation and acquisition unit and a data processing and control unit, wherein the signal excitation and acquisition unit is arranged in the battery box and is electrically connected with the signal excitation and acquisition unit, the signal excitation and acquisition unit is configured to apply a micro alternating current excitation signal to a battery cell and acquire the voltage current response of the battery cell, and the data processing and control unit comprises: a dual engine state estimation module having a simplified physical engine configured to quickly extract internal state parameters of the battery based on the voltage-current response data and a data-driven engine configured to learn and map internal state of health of the battery based on a history of battery operation and real-time external electrical parameter data in parallel; The dynamic credibility weighting fusion device is connected to the double-engine state estimation module and is configured to receive the preliminary estimation values of the same internal state parameters of the simplified physical engine and the data driving engine, allocate proper weights for the preliminary estimation values of the two engines according to the current working condition and the service life stage of the battery, perform first fusion calculation and output a primary fusion state quantity; The safety situation scoring and predicting model is connected to the dynamic credibility weighted fusion device and is configured to receive the primary fusion state quantity, comprehensively calculate the current safety situation scoring of the battery and simultaneously predict the battery state at the future moment in a short term based on the current state; The feedback verification closed loop module is connected between the safety situation score and the prediction model and the dynamic credibility weighting fusion device, and is configured to continuously compare a short-term prediction value made by the safety situation score and the prediction model with an actual measurement value subsequently acquired by the signal excitation and acquisition unit, and transmit the generated prediction deviation information as a feedback signal to the dynamic credibility weighting fusion device so as to drive the dynamic credibility weighting fusion device to adjust the credibility of the dynamic credibility weighting fusion device to the two engine output values in the dual