Reliability Assessment Method for DC/DC Converters of AgO-Al Batteries

Silver oxide–aluminum (AgO-Al) batteries are widely used as power sources for underwater vehicles, where direct current/direct current (DC/DC) converters play a critical role in supplying low-voltage loads. Accurate reliability assessment of DC/DC converters is essential for ensuring successful missions. Existing studies rarely consider the constraints between key performance parameters, as well as the load-dependent characteristics of measurement errors, limiting their applicability to DC/DC converters. To address these issues, a probabilistic model describing the relationship between operating load and measurement error is proposed, with particular emphasis on the load-dependent stochastic characteristics of measurement errors. Based on this relationship, load-dependent stochastic models for both the power-on and power-off voltages are developed by jointly considering unit-to-unit variability and random measurement errors. The model parameters are estimated by maximum likelihood estimation. A failure criterion incorporating the operational constraints between the power-on and power-off voltages is established according to the practical requirements of DC/DC converters. A reliability assessment method considering the dependence between the power-on and power-off voltages is provided according to the practical requirements. Based on test data, the results show that the Kolmogorov–Smirnov statistic of the proposed method is at least 86% lower than the compared methods, indicating a more accurate reliability assessment.