Abstract
This paper proposed an interval prediction strategy for lithium-ion battery remaining useful life (RUL) based on fuzzy information granulation and linguistic description to solve the limitations on current numerical prediction strategies. Firstly, the fuzzy information granulation is introduced to process the time series of battery capacity degradation and then the original numerical level data is treated as granular level, which is the basic of achieving interval prediction. Secondly, in order to solve the problem of fluctuation information loss caused by fuzzy granulation when processing battery degradation data, this paper creatively introduced a linguistic description method to attach semantic label for each granule to represent the fluctuation characteristics. Then, combined with the least square support vector machine, the granules with linguistic labels are used for modeling, by which the fluctuation characteristics of degradation sequence are considered while implementing the RUL interval prediction. Finally, four groups of NASA battery aging data were used for the experiment, and the interval prediction evaluation criterion P was introduced to evaluate the RUL interval prediction performance. Compared with the model without linguistic description, the P% of the model with linguistic description is improved by 32% on average.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call