锂电池热失控早期典型气体精准检测方法

Accurate typical gas detection of lithium battery in early thermal runaway period

锂电池储能是消纳新能源发电,达成国家“双碳”目标的重要途径,是地面储能及工商业储能的核心装备,锂电池热失控消防安全是储能大规模应用的前提保障。锂电池热失控早期典型气体检测是储能消防预警的主要手段,但是在混合气体场景下各气体传感器数据存在交叉干扰,导致检测失准而造成预警延迟或误报警,进而导致消防隐患。针对上述问题,围绕典型气体H2和CO的浓度精准检测,提出一种混合气体场景下传感器数据的解耦方法。通过建立各传感器在不同单一气体环境下的响应模型,建立不同气体对传感器的交叉耦合关系;进而推导混合气体场景下各传感器信号与气体组分与浓度的构成关系,建立方程组得出各气体的精确浓度数据,实现各传感器数据解耦。最后,搭建H_(2)和CO混合气体场景,用于模拟不同化学体系、不同SOC的锂电池热失控早期气体环境,进行实验测试,结果显示在0~1000 mL/m^(3)浓度范围内的检测误差小于50 mL/m^(3),检测精度最大提升了15%,验证了本文所提方法的有效性。

Lithium battery energy storage plays a crucial role in harnessing new energy for power generation and achieving the national"dual-carbon"objectives.This technology is indispensable for ground-based,industrial,and commercial energy-storage applications.A critical aspect of its widespread adoption is ensuring the safety of lithium batteries against thermal runaway and fire hazards.To this end,early detection of gases typically released during the thermal runaway phase of a lithium battery is essential for fire warning in energy storage.However,accurately identifying these gases poses challenges in environments with mixed gases owing to cross-interference from data collected by various gas sensors.Such interferences often lead to inaccurate detections,delayed or false alarms,and potential fire hazards.To address these issues,we propose a decoupling method designed to enhance the precision of detecting typical gases,namely H_(2)and CO concentrations,in mixed gas environments.This method involves establishing response models for each sensor in different single-gas environments and understanding the cross-coupling relationships between different gases and sensors.By deriving the relationship between the sensor signals and the gas components and concentrations in mixed-gas scenarios,we can establish an equation system.This system is crucial for obtaining precise concentration data for each gas,thus facilitating the decoupling of sensor data.In experimental testing with a mixed gas scenario of H2 and CO,which was built to simulate the early gas environment of lithium battery thermal runaway across different chemical systems and different states of charge.The results demonstrated a detection error of less than 50 mL/m^(3)within the concentration range of 0—1000 mL/m^(3).The maximum improvement in detection accuracy reached 15%,validating the effectiveness of the proposed method outlined in this study.