Wang, Guanyu;
(2022)
Using Artificial Intelligence to Predict the Discharge Performance of Cathode Materials for Lithium-ion Batteries Applications.
Doctoral thesis (Ph.D), UCL (University College London).
Preview |
Text
Guanyu_Wang_Thesis.pdf - Submitted Version Download (51MB) | Preview |
Abstract
A comprehensive understanding of the composition-structure-property relationships for doped cathode materials used in lithium-ion batteries remains lacking which delays the progress of developing new cathode materials. This thesis proposes that machine learning (ML) techniques can be used to predict the discharge capacities of the cathode materials whilst revealing these underlying relationships. To achieve this, the data for three different doped cathodes are curated from the publications, namely, the doped spinel cathode, LiMxMn2−xO4, the M-doped nickel- cobalt-manganese layered cathode, LiNixCoyMnzM1−x−y−zO2, and the carbon -coated and doped olivine cathode, C/LiM1M2PO4 (M1, M2 denote different metal ions). Several linear and non-linear ML models are trained with the data and compared for the power of predicting initial and higher cycle discharge capacity. Gradient boosting models have shown the best prediction power for predicting the initial and 20th cycle end discharge capacity of 102 doped spinel cathode and the initial and 50th cycle discharge capacity of 168 doped nickel-cobalt-manganese layered cathodes. For the doped spinel cathode, higher discharge capacities at both cycles can be achieved through increasing the material formula mass, reducing the crystal lattice constant and using dopants with smaller electronegativity. For the doped layered cathodes, it is revealed that the higher lithium content, lower formula molar mass, small doping content and doped with low electronegativity dopant are more likely to possess greater capacities at both cycles. Bayesian ridge regression and gradient boosting model are shown to have the highest prediction power over the initial and the 20th cycle discharge capacity of carbon-coated and doped olivine cathode. In addition, the olivine systems with lower dopant content, higher base-metal content and smaller unit cells are shown to be more likely to possess higher capacities at both cycles. Finally, future research directions are presented including the suggestion of involving other new input variables and using principal component analysis and feature selection algorithms to use to improve the model performance.
| Type: | Thesis (Doctoral) |
|---|---|
| Qualification: | Ph.D |
| Title: | Using Artificial Intelligence to Predict the Discharge Performance of Cathode Materials for Lithium-ion Batteries Applications |
| Open access status: | An open access version is available from UCL Discovery |
| Language: | English |
| Additional information: | Copyright © The Author 2022. Original content in this thesis is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) Licence (https://creativecommons.org/licenses/by-nc/4.0/). Any third-party copyright material present remains the property of its respective owner(s) and is licensed under its existing terms. Access may initially be restricted at the author’s request. |
| UCL classification: | UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of Maths and Physical Sciences UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of Maths and Physical Sciences > MAPS Faculty Office > Institute for Materials Discovery UCL > Provost and Vice Provost Offices > UCL BEAMS UCL UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of Maths and Physical Sciences > MAPS Faculty Office |
| URI: | https://discovery.ucl.ac.uk/id/eprint/10146392 |
Archive Staff Only
 |
View Item |
