Adoption of a Bayesian Belief Network for the System Safety Assessment of Remotely Piloted Aircraft Systems. (October 2019)

• Record Type: Journal Article
• Authors: Washington, Achim; Clothier, Reece; Neogi, Natasha; Silva, Jose; Hayhurst, Kelly; Williams, Brendan
• Is Part Of: Safety science. Volume 118(2019)
• Page Start: 654
• Page End: 673
• Publication Date: 2019-10
• Subjects: Remotely Piloted Aircraft Systems -- Bayesian Belief Networks -- System Safety Regulations -- Failure rate -- Uncertainty; RPAS Remotely Piloted Aircraft Systems (RPAS) -- SSA System Safety Assessment -- BBN Bayesian Belief Networks -- CONOPs Concept of Operations -- EASA European Aviation Safety Agency -- SSR System Safety Regulations -- COTS Commercial Off The Shelf -- SSPR System Safety Performance Requirement -- CA Compliance Assessment -- CF Compliance Findings -- FHA Functional Hazard Assessment -- ETA Event Tree Analysis -- FTA Fault Tree Analysis -- APFH Average Probability of Failure per Flight Hour -- FPOs Failure Probability Objectives -- FHA Functional Hazard Assessment -- ATSB Australian Transport Safety Bureau -- RPS Remote Pilot Station -- RPA Remotely Piloted Aircraft -- UDS Unpremeditated Descent Scenario -- LOC Loss of Control -- CFIT Controlled Flight into Terrain -- DOJC Dropped or Jettisoned Components (DOJC) -- LOSS Loss of Safe Separation -- RP Remote Pilot -- NPT Node Probability Table -- KE Kinetic Energy -- FAA Federal Aviation Administration -- ADF Australian Defence Force -- NSA NATO Standardization Agency -- JARUS Joint Authorities for Rulemaking of Unmanned Systems; Industrial accidents -- Periodicals; Accident Prevention -- Periodicals; Safety -- Periodicals; Travail -- Accidents -- Périodiques; 363.11
• DOI: 10.1016/j.ssci.2019.04.040
• Languages: English
• ISSNs: 0925-7535

Abstract

Highlights: Provides a template for high level classification of functions and failures. Uses BBNs to capture uncertainty in the assessed compliance scenario. Uses BBNs to extend compliance scenarios to multiple assessments. Removes the requirement for assessing single credible (often worst-case) scenarios. Applies BBNs within an aviation SSR "Part 1309" system safety context. Provides a case study application of unmanned aircraft systems. Abstract: There can be significant uncertainty as to the safety of novel or complex aviation systems, such as Remotely Piloted Aircraft Systems (RPAS). Current aviation safety assessment and compliance processes do not adequately account for uncertainty. The aim of this research is to support more objective, transparent, systematic and consistent regulatory outcomes in relation to the safety assessment of such systems. The objective of this work is to provide a systematic means of accounting for the various uncertainties inherent to any System Safety Assessment (SSA) process. The paper first defines the system safety compliance process and its modification to account for uncertainty. The SSA process, its various outputs, and associated uncertainties are defined and then applied to a generic RPAS. A Bayesian Belief Network (BBN) is adopted that facilitates a more comprehensive treatment of the uncertainty in each of the outputs of a typical SSA process. A case study of a generic RPAS is used to illustrate the features of the new approach. The adoption of the Proposed SSA approach would allow for the high uncertainty associated with the safety assessment of novel or complex aviation systems, such as RPAS, to be taken into consideration. Such an approach would enable the risk-based regulation of the sector.