Radiation Assurance of CubeSat Payloads Using Bayesian Networks and Fault Models

• Published in: 2018 Annual Reliability and Maintainability Symposium (RAMS) pp. 1 - 5
• Main Authors: Austin, Rebekah A., Mahadevan, Nagabhushan, Witulski, Arthur F., Evans, John
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01-01-2018
• Subjects: Bayes methods; Bayesian Networks; COTS; CubeSat; Fault Models; Goal Structuring Notation; Load modeling; Radiation Hardness Assurance; Random access memory; Reliability; Sensitivity analysis; Switches
• ISSN: 2577-0993
• DOI: 10.1109/RAM.2018.8463043

Commercial off-the-shelf (COTS) electronics are used in satellites that span the size, cost, and acceptable risk spectrum. At one extreme the CubeSat, a platform originally created for university-based spacecraft design, is often constructed only from COTS electronics. At the other extreme, even high-budget NASA spacecraft use COTS electronics in non-mission critical circuits, especially when no radiation-hardened alternatives exist or the radiation-hardened parts do not meet the system requirements. There is significant uncertainty in the radiation test results of COTS electronics; hence additional means of assessing risk and radiation hardness are needed. In this work we construct a fault-propagation model of a CubeSat payload in a SysML platform, and then develop a Bayesian network from the fault model of the system. The sensitivity of the system to variations in component radiation response can be evaluated quantitatively, within the limits of the assumptions and available data. This sensitivity study then informs the structure and content of the overall radiation-reliability assurance case.