Data repository for ‘Predicting battery end of life from solar off-grid system field data using machine learning’

Aitio A, Howey D

This dataset was generated by 1027 lead-acid batteries from BBOXX Ltd., each with nominal voltage 12 V (internally comprising 6 cells in series), nominal capacity 20 Ah, and attached to a 50 Wp photovoltaic panel. These systems are used for lighting, phone charging and small appliances, and are located across sub-Saharan Africa. Each battery was in use for 400-760 days. A full explanation is given in the associated paper which may be found at https://arxiv.org/abs/2107.13856 .

Keywords:

Gaussian process

,

rural electrification

,

battery

,

machine learning