Energy and AI · 2021 · 77 citations · 55 references
EngineeringManufacturing EngineeringMaterials EngineeringData MiningMachine LearningManufacturing ProcessesBattery MaterialsElectrode MaterialsIndustrial ElectrochemistryManufacturing Process PlanningManufacturing SimulationProcess EngineeringBattery ManufacturingBattery ModelingLithium-ion Battery MaterialsMachine Learning For Production SystemsManufacturingManufacturing Process OptimizationProcess ParametersProcess Parameters Optimization
Li-ion battery is one of the key players in energy storage technology empowering electrified and clean transportation systems. However, it is still associated with high costs due to the expensive material as well as high fluctuations of the manufacturing process. Complicated production processes involving mechanical, chemical, and electrical operations makes the predictability of the manufacturing process a challenge, hence the process is optimised through trial and error rather systematic simulation. To establish an in-depth understanding of the interconnected processes and manufacturing parameters, this paper combines data-mining techniques and real production to offer a method for the systematic analysis, understanding and improving the Li-ion battery electrode manufacturing chain. The novelty of this research is that unlike most of the existing research that are focused on cathode manufacturing only, it covers both of the cathode and anode case studies. Furthermore, it is based on real manufacturing data, proposes a systematic design of experiment method for generating high quality and representative data, and leverages the artificial intelligence techniques to identify the dependencies in between the manufacturing parameters and the key quality factors of the electrode. Through this study, machine learning models are developed to quantify the predictability of electrode and cell properties given the coating process control parameters. Moreover, the manufacturing parameters are ranked and their contribution to the electrode and cell characteristics are quantified by models. The systematic data acquisition approach as well as the quantified interdependencies are expected to assist the manufacturer when moving towards an improved battery production chain.
55
Greedy function approximation: A gradient boosting machine.
Jerome H. Friedman · The Annals of Statistics · 2001 · 29K citations · Full text
Jerome H. Friedman · Computational Statistics & Data Analysis · 2002 · 7.1K citations
The future cost of electrical energy storage based on experience rates
Oliver Schmidt, Adam Hawkes, Ajay Gambhir et al. · Nature Energy · 2017 · 1.1K citations
Engineering, Electrical Engineering, Industrial Engineering +15
Current status and challenges for automotive battery production technologies
Arno Kwade, Wolfgang Haselrieder, Ruben Leithoff et al. · Nature Energy · 2018 · 1.1K citations