Bioinspired uncertainty network for state of health estimation of lithium-ion batteries based on partial charging data

Lithium-ion batteries (LIBs) are extensively utilized in modern systems. The state of health (SOH) of LIBs directly influences the reliability and efficiency of systems. However, in practical applications, the partial charging and discharging processes pose significant challenges for SOH assessment. Existing methods commonly address this issue by extracting features from segment data within predefined voltage ranges, which lack adaptability across varying operating conditions and may overlook vital information related to battery degradation. Moreover, most of the data-driven methods mainly provide point estimates without considering the inherent uncertainty in battery SOH assessment, which increases risks in subsequent decision making. To overcome these limitations, this paper proposes a novel SOH assessment method which incorporates incremental capacity (IC) curves recovery for partial charging scenarios and utilizes a bioinspired uncertainty network. Complete IC curves are recovered based on functional principal analysis (FPCA) with partial charging data to reveal detailed degradation information of LIBs. Then, a bioinspired uncertainty network is employed to provide SOH assessment results. Experimental results show that the proposed method can effectively recover complete IC curves and provide more accurate and reliable SOH results than other methods.