Lithium-ion battery capacity fading dynamics modelling for formulation optimization: A stochastic approach to accelerate the design process. (15th September 2017)

• Record Type: Journal Article
• Authors: Tao, Laifa; Cheng, Yujie; Lu, Chen; Su, Yuzhuan; Chong, Jin; Jin, Haizu; Lin, Yongshou; Noktehdan, Azadeh
• Is Part Of: Applied energy. Volume 202(2017)
• Page Start: 138
• Page End: 152
• Publication Date: 2017-09-15
• Subjects: Lithium ion battery -- Capacity fading dynamics modelling -- Various formulations -- Accelerating battery design process -- Five-state nonhomogeneous Markov chain model; Power (Mechanics) -- Periodicals; Energy conservation -- Periodicals; Energy conversion -- Periodicals; 621.042
• DOI: 10.1016/j.apenergy.2017.04.027
• Languages: English
• ISSNs: 0306-2619

Abstract

Highlights: The model is linked to known physicochemical degradation processes and material properties. Aging dynamics of various battery formulations can be understood by the proposed model. Large number of experiments will be reduced to accelerate the battery design process. This approach can describe batteries under various operating conditions. The proposed model is simple and easily implemented. Abstract: A five-state nonhomogeneous Markov chain model, which is an effective and promising way to accelerate the Li-ion battery design process by investigating the capacity fading dynamics of different formulations during the battery design phase, is reported. The parameters of this model are linked to known physicochemical degradation dynamics and material properties. Herein, the states and behaviors of the active materials in Li-ion batteries are modelled. To verify the efficiency of the proposed model, a dataset from approximately 3 years of cycling capacity fading experiments of various formulations using several different materials provided by Contemporary Amperex Technology Limited (CATL), as well as a NASA dataset, are employed. The capabilities of the proposed model for different amounts (50%, 70%, and 90%) of available experimental capacity data are tested and analyzed to assist with the final design determination for manufacturers. The average relative errors of life cycling prediction acquired from these tests are less than 2.4%, 0.8%, and 0.3%, even when only 50%, 70%, and 90% of the data, respectively, is available for different anode materials, electrolyte materials, and individual batteries. Furthermore, the variance is 0.518% when only 50% of the data are available; i.e., one can save at least 50% of the total experimental time and cost with an accuracy greater than 97% in the design phase, which demonstrates an effective and promising way to accelerate the Li-ion battery design process. The qualitative and quantitative analyses conducted in this study suggest that the proposed model provides an accurate, robust, and simple way to accelerate the Li-ion battery design process for battery manufacturers, thereby enabling rapid market capture.