Abstract
The growth in electric vehicles market share is one of the main actions taken to fight greenhouse gases emissions, but it also brings new environmental challenges to the table. Because of the high costs connected to the extractions of Lithium and other battery raw materials, the environmental risks posed by battery disposal and the intermittent nature of electricity production from renewable sources, batteries which are not suitable anymore for traction use can turn from waste to a resource, through their recycling for stationary applications. This paper presents a stand-alone device for the diagnostics of battery modules with the following applications: a) it can be used by EV owners for predictive maintenance; b) it enables end-of-line quality control in manufacturing plants after the production of battery cells; c) it can be used to test batteries exhausted for traction use, matching batteries with similar quality for stationary storage units based on batteries in second life. The application of the proposed device can have noticeable impacts on battery manufacturing procedures, accelerating the transition towards EV-based mobility. The diagnostics is performed through Artificial Intelligence trained with experimental characterization of battery cells. This paper in particular presents the hardware selected for a 48 V 25 Ah battery, but the general architecture of the device and the methodology of the training procedure can be extended to any battery size.