Data-driven fault diagnosis and thermal runaway warning for battery packs using real-world vehicle data

Battery fault diagnosis is essential to ensure the safe and reliable operation of electric vehicles. Early detection of battery faults can reduce battery incidents and property losses. However, early warning of battery thermal runaway is still a challenging task. This paper proposes a novel data-driven method for lithium-ion battery pack fault diagnosis and thermal runaway warning based on state representation methodology. The normalized battery voltages are used to achieve accurate identification of battery early faults. The proposed method calculates the real-time state of each cell to characterize the internal characteristics of the battery cell, and the state changes are recorded to achieve battery fault diagnosis. The fault detection time is compared with the alarm time of real vehicles to verify the effectiveness of the proposed method. The real-world operation data of four electric vehicles with different specifications are used to verify the feasibility, robustness, and reliability of the proposed method. The results show that the method can achieve not only the accurate identification of the faulty cells and accurate determination of the voltage fault type but also the early detection of faults and early warning of thermal runaway. (c) 2021 Elsevier Ltd. All rights reserved.

Automated summary

This paper introduces a novel data-driven method for lithium-ion battery pack fault diagnosis and thermal runaway warning, achieving accurate identification of battery early faults and early warning of thermal runaway using normalized battery voltages. The proposed method is validated with real vehicle operation data to demonstrate its effectiveness.