Lithium ion batteries pose safety risks in both civil aviation and road transportation, as well as in transportation and household appliances. In order to find solutions to these problems, the impact of mechanical vibration on the performance and thermal runaway of lithium-ion batteries under low voltage conditions was studied. Under low voltage conditions, mechanical vibration causes an increase in the temperature of thermal runaway of lithium-ion batteries and a significant change in the popping time of safety valves.
Low frequency and large amplitude mechanical vibrations accelerate the process of battery short circuit, while high-frequency vibrations increase the ignition temperature and exhaust volume of lithium batteries. For 18650 lithium-ion batteries, the temperature at which the safety valve pops out is between 140~150 ℃. When the safety valve pops out, the gas release intensity is relatively high, and the gas signal can be a key variable for lithium battery safety warning. After low-voltage vibration, the discharge capacity of lithium-ion batteries is smaller than that of the original battery, and the discharge speed is faster than that of the original battery
1 Type of lithium-ion battery failure
As the “heart” of a car, lithium-ion batteries determine the distance traveled by the car. The lithium-ion battery system consists of four major components: battery modules (single cell series parallel), battery management system (BMS), thermal management system, and electrical and mechanical components. The main function of BM is to monitor the battery voltage, current, and temperature in real time through shell sensors, actuators, main control chips, etc. However, due to the internal aging and decay of each individual battery in the battery pack, as well as the corrosion of the circuit during the battery grouping process, improper operation during use can lead to abnormal external factors such as overcharging and discharging. The combined effect of these internal and external factors leads to internal and external failures of the battery. The internal faults of lithium-ion batteries can be classified into types such as overcharging, over discharging, internal short circuit, and thermal runaway.
Internal faults are caused by the malfunction of the BMS and the sensor itself, which prevents the sensor from working properly. On the other hand, they are caused by electrochemical reactions and internal short circuits in the battery’s internal structure, resulting in lithium dendrite phenomenon. However, the danger of external failures in lithium-ion batteries is usually greater than internal failures. External failures can trigger a chain reaction of internal failures, ultimately leading to uncontrolled heating. Sensor failures are often the most easily overlooked, but this can lead to serious consequences.
BMS relies on sensors to achieve functions such as balanced control, fault diagnosis, and State of Charge (SOC) estimation. If the sensor experiences deviations, drift, or stops working, it cannot obtain real-time data and make accurate and reasonable judgments about the current state of the battery. This may not only reduce performance but also cause major safety accidents. However, due to the high concealment of sensors, diagnosis is difficult, which is also the focus and difficulty of current sensor research.
BMS manages battery packs composed of hundreds or even thousands of individual batteries. Due to the similar characteristics of battery cell failures, sensor failures, and connector failures, as well as the fact that many faults are essentially small and highly concealed, it is difficult to quickly identify them. Therefore, BMS occasionally produces misdiagnosis and misoperation. It is crucial for car safety to quickly detect and accurately diagnose multiple battery failures. The process of battery fault diagnosis can be roughly divided into four aspects: fault detection, fault classification, fault localization, and fault isolation. Data processing of batteries plays a foundational role in battery fault diagnosis, and the effectiveness of denoising can be effectively verified by incorporating mathematical morphological filtering methods.
2 Problems with diagnostic methods for voltage measurement
(1) In practical applications, battery management systems can only measure the terminal voltage of each individual battery in the battery pack. To make the voltage measurement value include the voltage at the individual terminals and the voltage on the connectors, additional measurement lines need to be added, which undoubtedly increases the complexity of the equipment. If, at the beginning of the design of the battery management system, the collected voltage includes the terminal voltage and the voltage on the connector, it will cause the battery management system to be unable to accurately obtain the terminal voltage of the battery, and thus unable to effectively control and manage the charging and discharging of the battery, which is prone to overcharging and discharging faults
(2) The voltage changes caused by changes in battery internal resistance are similar to the voltage changes during connection faults Both are on the same order of magnitude, so diagnostic methods based on voltage signals also face the challenge of distinguishing between connection faults and faults with increased internal resistance of the battery.
(3) When a slight connection looseness occurs within the battery pack, the increase in contact resistance is very small. If the working current of the battery pack is small, such a connection fault will not cause significant changes in the voltage signal. Therefore, the voltage signal based diagnostic method mentioned above may not be able to detect early slight connection looseness.
3 The impact of vibration
Conduct atmospheric pressure charging and discharging tests on batteries subjected to low voltage vibration treatment, to demonstrate the changes in the charging and discharging performance of lithium batteries after aviation transportation. The results showed that after vibration frequency treatment at 60 Hz and 80 Hz in a low-voltage environment, the charging effect was lower than that of the original battery, and the 60 Hz treatment had the most significant impact on the capacity of lithium batteries;
The other vibration frequencies all accelerate the charging speed, and the variation is significant at 120 Hz. During the discharge process, the lithium-ion battery curve with 120 Hz vibration treatment has the largest deviation, while the curve with 200 Hz has the smallest deviation. The remaining deviations range from small to large, including 60 Hz, 80 Hz, 180 Hz, 160 Hz, 140 Hz, and 100 Hz.
It was found that there was no positive correlation between vibration frequency and discharge situation, which may indicate that the damage to the structure caused by vibration is related to the responsiveness of the battery’s own structure to frequency and amplitude. Some frequencies and amplitudes of vibration have greater damage to the structure of 18650 lithium batteries.
From the overall comparison of the charging and discharging curves, it can be found that the discharge capacity of all batteries that have been vibrated is smaller than that of the original battery, and the discharge speed is faster than that of the original battery. This can well indicate that the capacity and discharge performance of the battery will be damaged by mechanical vibration.
In fact, after being subjected to mechanical vibration, the contact degree between the positive and negative electrodes in the electrolyte of lithium-ion batteries will change, resulting in a change in the contact area between the electrode and the electrolyte solution, which hinders the transfer of lithium ions to the negative electrode during charging and discharging, and thus reduces the amount of lithium embedded in the negative electrode.
4 Conclusion
This article proposes a fault diagnosis method for lithium-ion battery pack connections based on mechanical vibration signals and obtains the following conclusions. Using different piezoelectric ceramic sheets to generate vibration excitation, measure vibration response, and extract time-frequency features from the response signal can effectively achieve the classification of single point and multi-point connection fault modes. Future work will optimize the layout of piezoelectric ceramic sheets, collect vibration signals from the vehicle environment, and make this method suitable for connection fault diagnosis of lithium-ion battery packs in vehicles