This title appears in the Scientific Report :
2021
Please use the identifier:
http://hdl.handle.net/2128/26371 in citations.
Please use the identifier: http://dx.doi.org/10.1016/j.jpowsour.2020.228863 in citations.
Online capacity estimation of lithium-ion batteries with deep long short-term memory networks
Online capacity estimation of lithium-ion batteries with deep long short-term memory networks
There is an increasing demand for modern diagnostic systems for batteries under real-world operation, specificallyfor the estimation of their state of health, for example, via their remaining capacity. The online estimationof the capacity of a cell is challenging due to the dynamic nature of cell ag...
Saved in:
Personal Name(s): | Li, Weihan (Corresponding author) |
---|---|
Sengupta, Neil / Dechent, Philipp / Howey, David / Annaswamy, Anuradha / Sauer, Dirk Uwe | |
Contributing Institute: |
Helmholtz-Institut Münster Ionenleiter für Energiespeicher; IEK-12 |
Published in: | Journal of power sources, 482 (2021) S. 228863 - |
Imprint: |
New York, NY [u.a.]
Elsevier
2021
|
DOI: |
10.1016/j.jpowsour.2020.228863 |
Document Type: |
Journal Article |
Research Program: |
Methods and Concepts for Material Development |
Link: |
Restricted OpenAccess |
Publikationsportal JuSER |
Please use the identifier: http://dx.doi.org/10.1016/j.jpowsour.2020.228863 in citations.
There is an increasing demand for modern diagnostic systems for batteries under real-world operation, specificallyfor the estimation of their state of health, for example, via their remaining capacity. The online estimationof the capacity of a cell is challenging due to the dynamic nature of cell aging and the limited variety of inputsavailable from a cell under operation. The scope of this work is the development of a data-driven capacityestimation model for cells under real-world working conditions with recurrent neural networks having longshort-term memory capability. Voltage-time sensor data from the partial constant current phase charging curve isused as input, reflecting input availability in the real world. The network achieves a best-case mean absolutepercentage error of 0.76% and is extremely robust while handling input noise. It also has the ability to handlevariations in the length of the input time series and can generate a viable estimation even with an incompletecollection of input due to sensor errors. The model validation with several scenarios is done in a local embeddeddevice, highlighting the use case of such models in future battery management systems. |