Electrochemical batteries, especially of lithium-ion type, have been primarily adopted to feed electric vehicles thanks to their superior overall performance. However, they suffer from some limitations such as lower efficiency at peaking power demand. In some applications, they are combined with supercapacitors, who can deliver high power at the expense of lower energy aptitudes. This combination of power sources constitutes a very attractive hybrid energy storage system for electric vehicles. There are numerous topologies and control schemes for such systems and standard drive cycles are usually used to validate such systems. In this paper, a machine learning method is proposed to manage the energy demand from the Li-ion battery and the supercapacitor with the objective of maximizing the efficiency of these devices. Initial simulations and experimental testing show promising results.
Paper
Full text
Hybrid Vehicle Energy Management Using Deep Learning
Semantic Scholar · Engineering · 2019
Abstract
Electrochemical batteries, especially of lithium-ion type, have been primarily adopted to feed electric vehicles thanks to their superior overall performance. However, they suffer from some limitations such as lower efficiency at peaking power demand. In some applications, they are combined with supercapacitors, who can deliver high power at the expense of lower energy aptitudes. This combination of power sources constitutes a very attractive hybrid energy storage system for electric vehicles. There are numerous topologies and control schemes for such systems and standard drive cycles are usually used to validate such systems. In this paper, a machine learning method is proposed to manage the energy demand from the Li-ion battery and the supercapacitor with the objective of maximizing the efficiency of these devices. Initial simulations and experimental testing show promising results.