通过动态深度学习实现离子电池的现实故障检测
Jingzhao Zhang1,2, Yanan Wang3, Benben Jiang4
1IIIS Tsinghua University, Beijing, China.
Nature communications
|September 23, 2023
概括
这项研究引入了一个深度学习框架,用于检测电动汽车离子电池中的异常. 该模型通过分析真实世界的数据来提高安全性和降低成本,优于现有的方法.
科学领域:
- 电池技术 电池技术
- 人工智能的人工智能
- 电动汽车系统 电动汽车系统
背景情况:
- 准确的离子电池 (LiB) 安全评估对于防止故障,实现部署和支持低碳经济至关重要.
- 由于复杂的故障模式和有限的大规模数据集,目前的AI异常检测方法缺乏针对现实的电池条件的定制和验证.
研究的目的:
- 开发和验证一个现实的深度学习框架,用于电动汽车 (EV) 异常检测.
- 通过结合真实世界的数据和定制的算法来解决现有的故障检测模型的局限性.
主要方法:
- 开发了一个深度学习框架,其中包括一个专门为动态系统设计的动态自动编码器.
- 使用社会和财务因素配置自动编码器.
- 在一个大型数据集上测试了检测算法,其中包括来自347辆电动汽车的69万多个LiB充电片段.
主要成果:
- 与包括其他深度学习方法在内的最先进的故障检测模型相比,开发的模型表现出更高的性能.
- 该框架成功地降低了预期的直接电动汽车电池故障和检查成本.
- 在一个大规模的真实数据集上验证了异常检测算法的有效性.
结论:
- 深度学习具有显著的潜力,可以提高电动汽车中的LiB安全性.
- 将社会和财务信息集成到深度学习模型设计中,对于提高电池安全性和性能至关重要.
- 该研究为更可靠的LiB异常检测提供了经过验证的框架.
相关概念视频
03:12Batteries and Fuel Cells
27.6K
A battery is a galvanic cell that is used as a source of electrical power for specific applications. Modern batteries exist in a multitude of forms to accommodate various applications, from tiny button batteries such as those that power wristwatches to the very large batteries used to supply backup energy to municipal power grids. Some batteries are designed for single-use applications and cannot be recharged (primary cells), while others are based on conveniently reversible cell reactions that...
27.6K
01:26Linear Approximation in Frequency Domain
106
Linear systems are characterized by two main properties: superposition and homogeneity. Superposition allows the response to multiple inputs to be the sum of the responses to each individual input. Homogeneity ensures that scaling an input by a scalar results in the response being scaled by the same scalar.
In contrast, nonlinear systems do not inherently possess these properties. However, for small deviations around an operating point, a nonlinear system can often be approximated as linear....
In contrast, nonlinear systems do not inherently possess these properties. However, for small deviations around an operating point, a nonlinear system can often be approximated as linear....
106