3 Testing

Testing design: In the overcharge test, three batteries with different chemical properties (LFP, NMC, and NCA) were used for comparative behavior analysis. The reason for choosing these batteries is that LFP has a mild overcharge reaction, NMC electrode has stronger reactivity as a cathode material, and NCA oxide releases oxygen and causes thermal runaway. The selection of batteries is based on the main criteria, which is that the batteries should have a CID. Before the experiment, samples of each battery type were opened and inspected.

Testing device: The testing device includes a power circuit and a measurement circuit. The measurement circuit includes a high-voltage measurement module, current clamp, temperature sensor, and data acquisition equipment. The power circuit consists of a voltage source, a load contactor, and a battery. The overcharge abuse test was conducted in Explosion-proof Chamber, and high-definition cameras and infrared cameras were used to record the events.

Testing process: The testing is conducted according to the FreedomCAR testing specification, but at the normal operating temperature of the battery. The testing equipment is charged to twice the rated voltage, and data collection stops after 30 minutes, regardless of the battery’s reaction state. The reaction of the battery was evaluated using the EUCAR hazard level, dividing its behavior into eight hazard levels. Three color levels were defined to represent the safe behavior of the battery, and binary logistic regression analysis was conducted.

Test parameters: Conduct ten tests on each battery at voltage levels of 120V, 400V, and 800V, as most electric vehicles are within these voltage ranges. According to the manufacturer’s battery data sheet, the current level of each battery was selected, with NCA and NMC batteries set to 4A and LFP batteries set to 1.5A. The battery is charged until the CID interrupts the charging flow or the test is terminated, with each test lasting for 30 minutes.

Data analysis: Binary logistic regression is used for evaluation based on binary expressions of “safe” or “unsafe”. The statistical evaluation of the test includes discrete (descriptive) and analytical (inferential) parts. The test can be described using three variables: chemical properties (discrete categorical variables), voltage (continuous ratio scaling variables), and test results (binary 0-1 variables, safe and unsafe).

4 Results

Classification of test results: In order to provide an overview of the raw data, three categories with hazard levels 3-5 have been defined for the test series.

The behavior of correctly triggering CID: The first test result category summarizes the data on the correct behavior of CID (hazard level 3). All tested batteries, after being overcharged for 10 minutes, had internal air pressure sufficient to open CID, causing battery exhaust (current drop, voltage increase).

CID triggered incorrect behavior: The second category summarizes CID triggered incorrect behavior, in which the CID partially interrupts current flow, resulting in strong smoke and temperature rise, and is classified as unsafe condition hazard level 4 (yellow unsafe behavior).

Behavior triggered by CID errors: The last category includes data triggered by CID errors, where CID can only briefly or completely separate current and voltage, and therefore cannot prevent battery overcharging, ultimately leading to battery combustion or explosion, classified as an unsafe condition of hazard level 5 or higher (red unsafe behavior).

5 Discussion

Limitations of testing standards: According to FreedomCAR’s battery testing standards, it is difficult to push the battery to the safe limit, that is, when overcharged at twice the rated voltage, the battery will not be pushed to extreme limits and will not exhibit dangerous behavior. Within this voltage range (2-5V), CID can correctly separate the positive and negative poles without igniting the battery. However, the testing standards do not reflect the actual use of lithium batteries. In the energy storage market, there are higher interconnected series switching systems with voltages up to 800V.

The performance of batteries with different chemical properties: Considering the results of the 120V test series, NMC and NCA chemical batteries exhibited the first critical battery behavior, while LFP chemical batteries were relatively safe and did not experience ignition or fire with a hazard level of 5 or higher. In the 400V test, the critical conditions of NMC and NCA chemistry batteries doubled compared to the 120V test, but LFP batteries can still be considered non critical. In the 800V test, the performance of NMC and NCA batteries was almost the same, in the ignition stage, while LFP batteries showed the first key behavior compared to the 120V and 400V test series.

Reasons for unsafe behavior: For all batteries classified as “unsafe”, the energy supply cannot be stopped, that is, the charging current cannot be interrupted, which may be due to the arc generated when CID is triggered, causing the charging current to continue flowing, resulting in a small contact point between the anode and cathode, leading to high current density. In addition, the distance between the two contacts created when CID is triggered is very short, which also increases the breakdown voltage and may cause arcing.

6 Conclusion

Shortcomings of current standards: Based on the results of all test series, it can be concluded that the current standards for testing battery safety in battery systems are insufficient. In the battery system of cylindrical batteries connected in series, the disconnection of CID under high system voltage may lead to the formation of critical arcs, resulting in battery combustion or explosion.

Therefore, if the batteries are connected in series in the battery system, battery testing at twice the rated voltage is not important for the safe behavior of the batteries, and the current standards must be revised. It is recommended that the testing conducted at the battery level should at least reach the maximum voltage level of the battery system planned for installation and operation.

Consideration for CID application: It has been found that overcharging the battery with very high voltage increases the potential for danger. Therefore, when a large number of batteries with CID are used in series in the battery system, their application should be reconsidered, as triggering CID may lead to catastrophic battery failure. The alternative solution to this problem is to design a CID battery that can withstand such high voltage.