Home / All / Industry News /

Researchers use artificial intelligence algorithm to evaluate the health status of battery

Researchers use artificial intelligence algorithm to evaluate the health status of battery

Dec 29,2017

Foreign media reported, an R&D team of the University of Science and Technology of China proposed to use artificial 

intelligence algorithm to the health status of electric car battery pack. This is the new evaluating method for 

the battery health. The team published an article on the Journal of Power Supply to introduce their research work.


In the application of electric car, it is an important thing to evaluate the health status of battery pack 

accurately.Because through this way can obtain the dynamic response of battery pack and improve its safe 

reliability.However,the discharge and charge performance of battery and the working environment of battery pack are 

different, and these make evaluating the health status of battery pack being difficult.


The health status of battery pack is defined as the change of battery pack’s max power storage by researchers. And 

it contains all the information of cell, like battery capacity, the relevance between the state of charge and the open-

circuit voltage, and the inconsistency of battery.


In order to estimate the health status of battery pack, the team adopted particle swarm optimization. Base on 

the experimental result, they used particle filter to predict the battery’s state of charge and open-circuit voltage, to 

avoid noise and drift current in the terminal voltage test. They also adopted recursive least square method to improve 

battery capacity.


According to the experimental result,this test method can predict the status of battery in the actual operation, and with 

high accuracy.


Contact: Acacia Huang

Email: sales03@liliangbattery.com

Skype: liliangbattery03

Website: www.liliangbattery.com


Please send your message to us
*Email
Name
Tel
*Title
*Content
Upload
  • Only supports .rar/.zip/.jpg/.png/.gif/.doc/.xls/.pdf, maximum 20MB.
Address