Battery pack short circuit early warning and protection method based on multi-sensor fusion

Abstract

The invention discloses a battery pack short circuit early warning and protection method based on multi-sensor fusion, and the method comprises the steps: S1, collecting the multi-dimensional sensor data, including temperature distribution data, voltage fluctuation data, gas component data, deformation data and current ripple data, of a battery pack in real time; s2, performing space-time alignment and feature extraction on the multi-source sensor data, and constructing a fusion feature vector; s3, based on the fusion feature vector, performing short circuit type identification and risk level evaluation through a two-stage short circuit risk judgment model; and S4, triggering a grading response mechanism according to the risk grade, wherein the grading response mechanism comprises an early warning signal, current limiting, fuse action or battery cluster isolation. By fusing five-dimensional data of temperature, voltage, gas, deformation and current, a two-stage short circuit risk judgment model is constructed, and accurate classification and risk grading evaluation of short circuit types are realized. A four-level response mechanism is triggered based on an evaluation result, and protection is strengthened step by step from early warning to physical isolation.

Description

Technical Field

[0001] The present invention relates to the technical field of battery safety management technology, and specifically to a battery pack short-circuit early warning and protection method based on multi-sensor fusion. Background Art

[0002] Battery safety and performance are inseparable from the key technology of the battery management system (BMS). Like a shrewd guardian, the BMS ensures stable battery operation in a variety of complex environments through real-time monitoring, intelligent regulation, and multiple protections. The BMS utilizes high-precision sensors to track key battery indicators such as voltage, current, and temperature in real time, while accurately calculating the battery's remaining capacity (SOC) and state of health (SOH). Upon detecting an abnormal cell voltage, such as overcharge or over-discharge, the BMS quickly activates its protection mechanism, shutting off the circuit to prevent battery damage. Furthermore, the BMS can predict battery aging trends, issuing early warnings and providing a scientific basis for battery maintenance.

[0003] However, current battery short-circuit protection mainly relies on voltage and temperature detection, which has several defects: such as high false alarm rate: the voltage change in the early stage of micro-short circuit is weak and cannot be detected by traditional BMS; response delay: the fuse needs to wait for the current to rise to the threshold before it can be activated, during which thermal runaway has been triggered; no type distinction: progressive short circuits such as dendrite growth cannot be identified. Summary of the Invention

[0004] (1) Technical problems solved

[0005] In response to the shortcomings of the existing technology, the present invention provides a battery pack short-circuit warning and protection method based on multi-sensor fusion. Through multi-source heterogeneous data fusion and a hierarchical response mechanism, it solves the problems of early, accurate and rapid battery short-circuit detection.

[0006] (2) Technical solution

[0007] To achieve the above objectives, the present invention provides the following technical solution: a battery pack short circuit warning and protection method based on multi-sensor fusion, comprising the following steps:

[0008] Step S1: real-time collection of multi-dimensional sensor data of the battery pack, including temperature distribution data, voltage fluctuation data, gas composition data, deformation data, and current ripple data;

[0009] Step S2: Perform spatiotemporal alignment and feature extraction on multi-source sensor data to construct a fusion feature vector;

[0010] Step S3: Based on the fused feature vector, the short circuit type is identified and the risk level is assessed using a two-level short circuit risk determination model;

[0011] Step S4: triggering a graded response mechanism according to the risk level, including early warning signal, current limiting, fuse operation or battery cluster isolation.

[0012] As a preferred solution, the sensor combination in step S1 includes: a distributed thermocouple array for monitoring the temperature gradient of the cell, a high-frequency voltage sampling module for capturing μs-level voltage drops, a MEMS hydrogen sensor for detecting electrolyte decomposition gas, a piezoelectric film deformation sensor attached to the battery casing, and a Hall current sensor for measuring current harmonic components.

[0013] As a preferred solution, the feature extraction in step S2 includes:

[0014] Temperature characteristics: maximum temperature difference between cells ΔT_max, temperature rise rate dT / dt;

[0015] Voltage characteristics: voltage standard deviation σ_V, adjacent cell voltage difference ΔV;

[0016] Gas characteristics: hydrogen concentration change rate d[H2] / dt;

[0017] Deformation characteristics: shell expansion ε and expansion acceleration d 2 ε / dt 2 ;

[0018] Current characteristics: high-frequency ripple energy ratio E_ripple.

[0019] As a preferred solution, the two-level short circuit risk determination model in step S3 includes:

[0020] First-level model: A random forest-based short circuit type classifier that distinguishes external short circuits, internal micro short circuits, and dendrite penetration short circuits;

[0021] Second-level model: LSTM-based risk level prediction network, outputting risk level R∈[0,4].

[0022] As a preferred solution, step S3 further includes performing voltage balance verification, the determination conditions of which are:

[0023] When R≥0.6, start the battery pack terminal voltage balance detection; calculate the single cell voltage standard deviation σ_V, and detect the maximum voltage deviation ΔV_max;

[0024] If σ_V>0.05×rated voltage or ΔV_max>3%×rated voltage, it is determined that the voltage is abnormal due to imbalance, and the short-circuit protection action is suppressed;

[0025] Otherwise go to step S4.

[0026] As a preferred solution, the risk level determination logic in step S3 is:

[0027] R=1 (low risk): trigger cloud warning and local sound and light alarm;

[0028] R=2 (medium risk): start-up active current limiting to 50% of rated current;

[0029] R=3 (high risk): cut off the main circuit and trigger the fuse;

[0030] R=4 (critical): Activate the battery cluster level isolation device.

[0031] As a preferred solution, the short-circuit warning and protection method further includes a fault tracing step: constructing a battery pack topology relationship model based on a graph neural network (GNN) to locate the position of the short-circuit starting cell.

[0032] As a preferred solution, it includes a multi-sensor acquisition module, a data fusion processing unit, a hierarchical execution mechanism and an isolation control unit.

[0033] (3) Beneficial effects

[0034] Compared with the existing technology, the present invention provides a battery pack short-circuit warning and protection method based on multi-sensor fusion, which has the following beneficial effects: By integrating five-dimensional data of temperature, voltage, gas, deformation, and current, the present invention constructs a two-level short-circuit risk determination model, realizing accurate classification of short-circuit types and risk grading assessment. Based on the assessment results, a four-level response mechanism is triggered, gradually strengthening protection from warning to physical isolation. At the same time, the introduction of fault location and dynamic threshold adjustment technology significantly improves the accuracy of short-circuit identification and the real-time response. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 Schematic diagram of the process of the present invention. DETAILED DESCRIPTION

[0036] In order to better understand the purpose, structure and function of the present invention, the following is a further detailed description of a battery pack short circuit warning and protection method based on multi-sensor fusion of the present invention in conjunction with the accompanying drawings.

[0037] See also Figure 1 The present invention provides a battery pack short circuit warning and protection method based on multi-sensor fusion, comprising the following steps:

[0038] Step S1: real-time collection of multi-dimensional sensor data of the battery pack, including temperature distribution data, voltage fluctuation data, gas composition data, deformation data, and current ripple data;

[0039] Step S2: Perform spatiotemporal alignment and feature extraction on multi-source sensor data to construct a fusion feature vector;

[0040] Step S3: Based on the fused feature vector, the short circuit type is identified and the risk level is assessed using a two-level short circuit risk determination model;

[0041] Step S4: triggering a graded response mechanism according to the risk level, including early warning signal, current limiting, fuse operation or battery cluster isolation.

[0042] Specifically, the sensor assembly in step S1 of the present invention includes: a distributed thermocouple array for monitoring cell temperature gradients, a high-frequency voltage sampling module for capturing μs-level voltage drops, a MEMS hydrogen sensor for detecting electrolyte decomposition gases, a piezoelectric film deformation sensor attached to the battery casing, and a Hall effect current sensor for measuring current harmonics. Taking a 280Ah lithium-ion battery module as an example, the installation method is as follows: thermocouples: one each on the positive, negative, and side surfaces of each cell; piezoelectric film: attached to the center of the casing; hydrogen sensor: mounted centered on the top of the module.

[0043] The feature extraction in step S2 includes:

[0044] Temperature characteristics: maximum temperature difference between cells ΔT_max, temperature rise rate dT / dt;

[0045] Voltage characteristics: voltage standard deviation σ_V, adjacent cell voltage difference ΔV;

[0046] Gas characteristics: hydrogen concentration change rate d[H2] / dt;

[0047] Deformation characteristics: shell expansion ε and expansion acceleration d 2 ε / dt 2 ;

[0048] Current characteristics: high-frequency ripple energy ratio E_ripple.

[0049] The short-circuit detection logic is as follows: micro-short characteristics: ΔV>50mV and d[H2] / dt>5ppm / s; dendrite short: E_ripple>0.1 and ε>0.5% for 10s; external short: dT / dt>5°C / s and σ_V>0.2V. A dynamic risk threshold library is established to adaptively adjust the detection threshold based on battery SOC, aging, and ambient temperature.

[0050] Furthermore, the two-level short circuit risk determination model in step S3 of the present invention includes:

[0051] First-level model: A random forest-based short circuit type classifier that distinguishes external short circuits, internal micro short circuits, and dendrite penetration short circuits;

[0052] Second-level model: LSTM-based risk level prediction network, outputting risk level R∈[0,4].

[0053] The risk level determination logic is as follows:

[0054] R=1 (low risk): trigger cloud warning and local sound and light alarm;

[0055] R=2 (medium risk): start-up active current limiting to 50% of rated current;

[0056] R=3 (high risk): cut off the main circuit and trigger the fuse;

[0057] R=4 (critical): Activate the battery cluster level isolation device.

[0058] To more safely protect the battery pack, the present invention also proposes a fault tracing step: constructing a battery pack topology relationship model based on a graph neural network (GNN) to locate the short-circuit starting cell position. A protection mechanism is also provided, including: a fast fuse based on shape memory alloy (SMA) with a response time of <10ms; a modular battery cluster isolation unit that achieves physical isolation through a magnetic latching relay. Furthermore, the present invention also includes performing voltage balance verification, and its judgment conditions are:

[0059] When R≥0.6, start the battery pack terminal voltage balance detection; calculate the single cell voltage standard deviation σ_V, and detect the maximum voltage deviation ΔV_max;

[0060] If σ_V>0.05×rated voltage or ΔV_max>3%×rated voltage, it is determined that the voltage is abnormal due to imbalance, and the short-circuit protection action is suppressed;

[0061] Otherwise go to step S4.

[0062] Furthermore, the present invention provides a battery pack short circuit warning and protection method based on multi-sensor fusion, which is based on a battery pack management system and includes a multi-sensor acquisition module, a data fusion processing unit, a hierarchical execution mechanism and an isolation control unit.

[0063] Among them, the multi-sensor acquisition module is used to capture data and sense the changes in the physical state of the battery in real time; the data fusion processing unit converts multi-source data into decision-making features and makes decisions in real time, completing the short-circuit risk assessment within 5ms; the hierarchical actuator matches the protection strength according to the risk level to prevent false operation; the isolation control unit uses physical isolation to block the spread of thermal runaway and suppress the re-ignition of the isolation area. Specifically, the sensor data is transmitted via SPI / I 2 C is aggregated to the processing unit, and after time and space alignment, a fusion feature vector is generated. Then, the FPGA runs the classification model to the ARM to calculate the risk level and output the execution instruction. Finally, the fuse status / isolation resistance value is returned to the processing unit to close the loop and verify the effectiveness of the action.

[0064] The technical solution of the present invention solves the problems of early detection, accuracy and rapidity of battery short circuit detection through multi-source heterogeneous data fusion and hierarchical response mechanism, and is particularly suitable for the safety protection of high energy density battery systems.

[0065] It will be understood that the present invention is described by way of certain embodiments, and it will be appreciated by those skilled in the art that various changes or equivalent substitutions may be made to these features and embodiments without departing from the spirit and scope of the present invention. In addition, under the guidance of the present invention, these features and embodiments may be modified to suit specific circumstances and materials without departing from the spirit and scope of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of this application are intended to be protected by the present invention.

Claims

1. A battery pack short circuit warning and protection method based on multi-sensor fusion, characterized in that: The steps include: Step S1: real-time collection of multi-dimensional sensor data of the battery pack, including temperature distribution data, voltage fluctuation data, gas composition data, deformation data, and current ripple data; Step S2: Perform spatiotemporal alignment and feature extraction on multi-source sensor data to construct a fusion feature vector; Step S3: Based on the fused feature vector, the short circuit type is identified and the risk level is assessed using a two-level short circuit risk determination model; Step S4: triggering a graded response mechanism according to the risk level, including early warning signal, current limiting, fuse operation or battery cluster isolation.

2. The battery pack short circuit warning and protection method based on multi-sensor fusion according to claim 1 is characterized in that: The sensor combination in step S1 includes: a distributed thermocouple array for monitoring the temperature gradient of the cell, a high-frequency voltage sampling module for capturing μs-level voltage drops, a MEMS hydrogen sensor for detecting electrolyte decomposition gas, a piezoelectric film deformation sensor attached to the battery casing, and a Hall current sensor for measuring current harmonic components.

3. The battery pack short circuit warning and protection method based on multi-sensor fusion according to claim 1 is characterized in that: The feature extraction in step S2 includes: Temperature characteristics: maximum temperature difference between cells ΔT_max, temperature rise rate dT / dt; Voltage characteristics: voltage standard deviation σ_V, adjacent cell voltage difference ΔV; Gas characteristics: hydrogen concentration change rate d[H2] / dt; Deformation characteristics: shell expansion ε and expansion acceleration d 2 ε / dt 2 ; Current characteristics: high-frequency ripple energy ratio E_ripple.

4. The battery pack short circuit warning and protection method based on multi-sensor fusion according to claim 1, characterized in that: The two-level short circuit risk determination model in step S3 includes: First-level model: A random forest-based short circuit type classifier that distinguishes external short circuits, internal micro short circuits, and dendrite penetration short circuits; Second-level model: LSTM-based risk level prediction network, outputting risk level R∈[0,4].

5. The battery pack short circuit warning and protection method based on multi-sensor fusion according to claim 1, characterized in that: The step S3 further includes performing voltage balance verification, the determination conditions of which are: When R≥0.6, start the battery pack terminal voltage balance detection; calculate the single cell voltage standard deviation σ_V, and detect the maximum voltage deviation ΔV_max; If σ_V>0.05×rated voltage or ΔV_max>3%×rated voltage, it is determined that the voltage is abnormal due to imbalance, and the short-circuit protection action is suppressed; Otherwise go to step S4.

6. The battery pack short circuit warning and protection method based on multi-sensor fusion according to claim 4, characterized in that: The risk level determination logic in step S3 is: R=1 (low risk): trigger cloud warning and local sound and light alarm; R=2 (medium risk): start-up active current limiting to 50% of rated current; R=3 (high risk): cut off the main circuit and trigger the fuse; R=4 (critical): Activate the battery cluster level isolation device.

7. The battery pack short circuit warning and protection method based on multi-sensor fusion according to claim 1, characterized in that: The short-circuit warning and protection method also includes a fault tracing step: building a battery pack topology relationship model based on a graph neural network (GNN) to locate the position of the short-circuit starting cell.

8. The battery pack short circuit warning and protection method based on multi-sensor fusion according to claim 1 is based on a battery pack management system and includes a multi-sensor acquisition module, a data fusion processing unit, a hierarchical execution mechanism and an isolation control unit.

Images