Machine Learning Assessment of Battery State-of-Health
摘要
The topic of LiIon battery State-of-Health assessment has become extremely prevalent within scientific research in the past 15 years due to the push towards a more sustainable industry. Within this work, we have developed and presented a Machine Learning method for evaluating the State-of-Health of batteries based on curve approximation coefficients. Additionally, we have proposed a method for converting the captured battery charging and discharging transient data into spectrograms, which can be used as data entries for the training of a Convolutional Neural Network image classifier.