A novel cuckoo search particle filtering strategy for the remaining useful life prediction of the lithium‐ion batteries in hybrid electric vehicle. (8th September 2022)

• Record Type: Journal Article
• Authors: Pu, Ren; Wang, Shunli; Chen, Xianpei; Huang, Junhan; He, Mingfang; Cao, Wen
• Is Part Of: International journal of energy research. Volume 46:Number 15(2022)
• Page Start: 21703
• Page End: 21715
• Publication Date: 2022-09-08
• Subjects: battery management system -- capacity degradation -- cuckoo search optimization particle filtering -- exponential growth model -- maximum likelihood region -- remaining useful life; Power resources -- Periodicals; Power (Mechanics) -- Periodicals; Power resources -- Research -- Periodicals; 621.042
• DOI: 10.1002/er.8712
• Languages: English
• ISSNs: 0363-907X

Abstract

Summary: The remaining useful life (RUL) is a core parameter of the battery management system. To realize accurately predict the RUL, the paper takes the National Aeronautics and Space Administration battery test data set as the research object, and a battery capacity degradation model based on an exponential growth model is built to characterize the battery aging process. A novel cuckoo search optimization particle filtering algorithm is proposed for the RUL prediction by transferring the particles in the prior distribution region of the particle filtering algorithm to the maximum likelihood region. The initial cycle numbers are set differently, the capacity decay process of the four groups of batteries can be predicted completely, and the RUL of the batteries can be obtained. Compared with the commonly used particle filtering and unscented particle filtering algorithms, the results show that the proposed method has obvious advantages in the relative error of prediction and resampling rate under the four datasets. For the B0005 dataset, the relative errors are 4.8%, 3.2%, and 0.8%, respectively under 30, 50, and 70 cycles, corresponding to the particle filtering algorithm errors are 10.3%, 7.9%, and 4.8%, and the unscented particle filter algorithm errors are 6.3%, 4.8%, and 2.4%, respectively. In addition, the method for RUL prediction present in the paper has a high confidence level, and low resampling rate, which plays an important role in promoting the further application of lithium‐ion batteries. Highlights: A novel cuckoo search optimization particle filtering algorithm is proposed. Different battery aging cycle number is distinguished to predict. Multiple battery RUL prediction is realized. RUL prediction has high prediction accuracy and high confidence.