Part 3: Experimental verification, Experimental setup
Cracks (or cuts, scratches)
In the manufacturing process of batteries, potential issues such as cracks may arise from improper welding or instances of compression and impact during usage. After sealing or formation, it is difficult to obtain spatial resolved data of the battery from the outside. Utilizing magnetic field distribution maps with cracks allows for an intuitive assessment of the crack’s condition, facilitating fault detection and precise crack localization.
Figure 10 shows cracks in different orientations in the lithium-ion battery. The first scenario involves a crack parallel to the X-axis located at x = 2.5 cm, y = 5 cm, with a depth of 65 μm. As evident from the current density distribution diagram, the presence of a surface current from negative to positive in the xy plane causes the current to “circumvent” the crack in the battery, resulting in higher current density at both ends of the crack.
Figure 10
The current distribution of the battery when there are cracks in different directions: (a) along the x-axis; (b) along the y-axis; (c) At a 45° angle to the x-axis.
For encapsulated batteries, obtaining current density distribution images from the exterior is challenging; however, magnetic field information can be easily detected through magnetic sensors. Regarding the magnetic field distribution of the battery, results similar to current distribution can be observed.
As shown in Fig. 11, the magnetic field images around a battery with a crack are displayed. Apart from significantly elevated magnetic field values at the tabs, there are also abnormally high magnetic fields at both ends of the crack. Additionally, there are conspicuous gaps in the magnetic field image at the crack’s site. Through the magnetic field distribution images, the locations and quantities of cracks in the battery can be easily detected. For the charging condition of 1C, the abnormally high magnetic field values are approximately 8 × 10-8 T.
From Fig. 11a, it can be noted that the abnormal magnetic field values at the two ends of the crack are different. The abnormal magnetic field values are stronger at x = 0.02 m than at x = 0.03 m. This is because x = 0.02 m is closer to the tabs of the battery, resulting in a relatively larger current flowing through this region. In addition, in Fig. 11a, the current and magnetic field at both ends of the crack are greater than in Fig. 11b and c. This is because the direction of the crack in Fig. 11a (in the x-direction) is perpendicular to the flow direction of the current in the current collector (in the y-direction). Therefore, the current needs to “circumvent” the incision at both ends, leading to two positions with abnormally increased magnetic fields.
Figure 11
Magnetic field distribution of the battery when there are cracks in different directions: (a) along the x-axis; (b) along the y-axis; (c) At a 45° angle to the x-axis.
Experimental verification
In practical applications, the nominal voltage and rated capacity of individual batteries are often insufficient to meet the demands of high-power usage. Thus, single batteries are commonly combined to form battery packs. During the configuration process of these packs, a prevalent approach involves initially parallel connecting individual batteries to create smaller battery pack units. Subsequently, these smaller units are arranged in series to construct larger battery packs. This hierarchical configuration method is designed to enhance the overall stability of the battery pack during operation.
Parallel operation effectively equalizes voltage and current discrepancies among individual batteries, thereby enhancing the performance consistency of the battery pack. Meanwhile, series operation facilitates voltage superposition to cater to scenarios requiring higher energy demands. Consequently, this parallel-then-series configuration method holds significant application value in battery system design.
However, over time and with repeated cycles, the performance consistency of batteries within the pack tends to deteriorate. Research indicates that variations in capacity or internal resistance directly contribute to uneven current distribution in parallel battery packs, accelerating battery aging and diminishing overall pack performance. Moreover, low-capacity batteries may induce abnormally high currents, generating excessive heat and posing potential safety hazards.
The power banks represent a ubiquitous application of battery packs in daily life. Typically, these banks comprise single batteries arranged in series, parallel, or series–parallel combinations. A dedicated circuit board governs and manages the charging and discharging processes of the batteries, ensuring each operates within a safe range. In commercial power banks, the closure of the entire battery pack due to a single battery malfunction often impedes prompt troubleshooting of the faulty unit.
Based on the background, to further validate the efficacy of the model developed in this study for practical applications, this section will conduct detailed measurement procedures on an 18650-battery pack. However, acquiring faulty batteries for experi-mentation poses significant challenges. Manual disassembly or fault analysis, such as through nail penetration test, not only demands professional instrumentation and equipment but also entails inherent safety risks. Particularly with commercial power banks, unauthorized dismantling without proper equipment greatly heightens the risk of severe safety incidents, including fire and explosions.
Considering these factors, the experimental design of this section adopts an innovative approach: within a four-cell compartment, batteries will be selectively removed from different locations to simulate single-cell failures within the battery pack. Furthermore, when the battery has a fault such as short circuit or particle rupture, it will produce a weak magnetic field change. This change, while not easy to detect, can be detected with sensitive magnetic field sensors.
For example, fluxgate can detect magnetic field changes of ~ nT magnitude; and the SERF atomic magnetometer can detect magnetic field changes of the order of 10fT. This method allows us to effectively replicate the fault conditions of the battery pack in real-world usage scenarios while prioritizing safety. Subsequently, precise measurements will be conducted on both the intact battery pack and the simulated fault state to comprehensively assess the model’s performance in practical settings.
Experimental setup
When the battery is in operation, its magnetic field undergoes fluctuations during the charging and discharging processes. A magnetic sensor is employed to capture the magnetic field variations induced by the battery. This section utilizes scanning techniques to gather magnetic field data. Specifically, a fluxgate probe (with a measuring range of ± 70 μT and frequency domain noise less than 5pT/Hz1/2@1 Hz) is mounted onto a linear module sliding table, which moves around the battery during its operation to capture the magnetic field distribution within a defined plane.
For data collection, the fluxgate is fixed onto the holder, and the linear translation stage is connected to the bracket, enabling movement and scanning. The spatial position of the probe can be adjusted and calibrated manually. The stepper motor and controller integrated with the linear module enable precise control over its motion, facilitating the fluxgate’s scanning measurements.
The reference coordinate system utilized during measurement is defined by the three-axis measurement coordinate system of the fluxgate, as illustrated in Fig. 12. (a). The blue arrow denotes the three-axis component measured by the fluxgate. When assessing the magnetic field distribution surrounding the battery, the y–z plane is designated as the scanning plane, and the fluxgate gathers magnetic field data based on its preset parameters.
The scanning step length is 12 mm, with a stabilization time of 3 s, and each data point’s collection duration is set to 3 s. Following the stabilization period, the software of the three-axis fluxgate samples and stores the magnetic field data. The magnetic field value of each point is determined by averaging the data collected within 3 s. This process is iterated across the designated scanning plane, with the resultant magnetic field values correlated with their corresponding coordinates to generate a magnetic field distribution map within the scanning plane.
Figure 12
Experimental setup: (a) Arrangement of measurement setup. (b) Diagram of the power bank used in the experiment.
The diagram illustrating the power bank utilized in the experiment is depicted in Fig. 12b. It can be placed inside 1–4 batteries, and the connection between the batteries is in parallel. The internal working battery of the battery box is a commercial 18650 battery with a rated capacity of 2.5 Ah.
Throughout the experiment, the battery testing system (Neware, BTS-20V10A) was initially employed to regulate the charging and discharging states of the battery, followed by the utilization of the magnetic field acquisition module for scanning and data collection. During the measurement process, the distance maintained between the fluxgate probe and the surface of the power bank was 4 cm.