Refining Fault Trees for Accurate Risk Assessment of Unmanned Aerial Systems (UAS)
Leung, Tszhim.
2017
-
Abstract: Unmanned aerial systems (UAS) are unlike most of the aircraft that currently operate in the National Airspace because they lack human pilots. Existing risk assessment methods, specifically fault trees, are not well-suited for unmanned aircraft as: (1) they cannot accurately model all the risks of the automated software in UAS and (2) are overconservative when representing general UAS fai... read morelure events. This thesis presents two refinements to those methods to improve accuracy and reduce overconservatism. The first recognizes and demonstrates that the MAJORITY gate, an existing logic gate, is an improvement upon the traditional AND gate model by additionally accounting for false alarm scenarios present in Fault Detection, Isolation, and Recovery (FDIR) software. A case study involving the application of FDIR to mitigate UAS control-surface failure suggests that the MAJORITY gate both sufficiently conservative as well as accurate in assessing overall system risk. The second, dubbed the Consequence Severity Level (CSL) analysis, replaces binary analysis with a multi-level analysis that directly maps failure events into aviation definitions for consequence severity. An unmanned aircraft example demonstrates how accounting for severity levels can relax certain design requirements by multiple orders of magnitude.
Thesis (M.S.)--Tufts University, 2017.
Submitted to the Dept. of Mechanical Engineering.
Advisor: Jason Rife.
Committee: Pratap Misra, and Kye Taylor.
Keyword: Mechanical engineering.read less - ID:
- mc87q2783
- Component ID:
- tufts:20663
- To Cite:
- DCA Citation Guide EndNote
