A digital twin based reliability assessment of Superbuck converters

Abstract

Addressing the challenge of assessing the reliability of "Long-life, high-reliability" DC/DC analog converters in submarine observation networks and space power systems, which have limited internal observation points and insufficient valid information, and initial assessment rendered unreliable with the long-term operation of the converter. This paper proposes a novel method based on digital twin technology.The method eliminates the need for measuring internal variables, and only requires the measurement of the converter's ripple peak-to-peak. A digital Twin(DT) is constructed based on the underlying mechanism, and the parameters of the operating converter are identified through the collection of ripple peak-to-peak, which are then injected into the digital twin for dynamic tracking of the physical entity. By obtaining internal current and voltage of the DT and utilizing the Markov model, the reliability of the converter can be assessed.This method offers real-time reliability assessment without the need for additional measurement points. To validate the effectiveness of the proposed approach, a Superbuck converter is used as an example to demonstrate the high fidelity of the DT through isomorphic single-sample repeatability and isomorphic multiple-sample consistency experiments.

Advanced digital twin-enabled fault diagnosis framework for unmanned vehicle systems

“…Wright and Davidson [29] discussed the key characteristics of digital twins and used metrology and metrology data to support the use of digital twins. Xia et al [30] applied digital twins to power systems and proposed a reliability evaluation method for Superbuck converters.…”

AI-enabled industrial equipment monitoring, diagnosis and health management

“…The model, integrating just-in-time learning (JITL) with Gaussian process regression (GPR) in a JITL GPR, accurately forecasted flow rate, coefficient for dynamic head, and shaft power, showcasing enhanced reliability in predicting performance parameters for centrifugal pumps without the need for sensors. Xia et al [37] developed a digital twin based on ripple peak-to-peak measurements to evaluate the reliability of DC/DC analog converters in submarine observation networks and space power systems, overcoming challenges related to internal observation points and inadequate information. Wang et al [38] proposed a similarity-based probabilistic approach for system-level remaining useful life (RUL) estimation, emphasizing the capture of effective degradation features, integration of multiple nonlinear sensor data, and management of multiple source uncertainties.…”