Synthetic data for LFP/LMO, NMC/LMO, and NCM/NCA blended electrodes vs. Graphite

Description

Three synthetic datasets, each consisting of more than 260,000 unique Voltage vs. Capacity curves, covering all possible degradations for three exemplary cells with blended electrodes are provided here. More details can be found in "Electrode blending simulations Using the Mechanistic Modeling Approach" (10.3390/batteries10050159). These datasets were calculated using the mechanistic modeling approach. See “Big data training data for artificial intelligence-based Li-ion diagnosis and prognosis“ (Journal of Power Sources, Volume 479, 15 December 2020, 228806) and "Analysis of Synthetic Voltage vs. Capacity Datasets for Big Data Li-ion Diagnosis and Prognosis" (Energies 2021, 14, 2371 ) for more details on the methodology. The analysis of these particular datasets is provided in "Electrode blending simulations Using the Mechanistic Modeling Approach" published in Batteries (10.3390/batteries10050159). This work investigated the behavior of electrodes with blends composed of materials with a completely separated electrochemical response (LFP and LMO), mostly over-lapping response (NMC and NCA), and an only partially overlapping response (NMC and LMO). For each blend, a set of 5450 different degradation paths were defined combining every combination of loss of lithium inventory (LLI) and LAMs on NE as well as both components of the PE in 1/30th increments. For each path, 31 combinations of LAMPE1/LAMPE2 were considered to account for loss on either component of the PE. Finally, 49 simulations at different levels of degradation (from 0 to 50% with a periodically increasing time step) were then performed for each path for a total of 267,050 simulations. 4 Variables are included, see attached readme.m file for details and example how to use.

Institutions

Categories

Funding

Licence