Investigation of nickel-metal hydride battery sorting based on charging thermal behavior

In this study, the sorting of nickel-metal hydride batteries is investigated based on their charging thermal behavior. A self-organization map (SOM) model affiliated to artificial neural network is constructed to conduct the sorting work. The sorting principle is described in detail to support the model. A batch of batteries is charged in various rates to collect training data closely related to battery thermal behavior. It is indicated that the model can master the regulation of sorting well after training. As a result, the batteries are classified by the SOM model into three categories of high heat generation battery, middle heat generation battery, and low heat generation battery, which corresponds well with the training result. The model thus allows the batteries in the same category to be selected for the consistency in thermal behavior as well as discharge performance. (C) 2012 Elsevier B.V. All rights reserved.