Grandison, Angus ORCID: 0000-0002-9714-1605 , Deere, Steven ORCID: 0000-0001-5171-2014 , Lawrence, Peter ORCID: 0000-0002-0269-0231 and Galea, Edwin Richard ORCID: 0000-0002-0001-6665 (2017) The use of confidence intervals to determine convergence of the total evacuation time for stochastic evacuation models. Ocean Engineering, 146. pp. 234-245. ISSN 0029-8018 (doi:https://doi.org/10.1016/j.oceaneng.2017.09.047)

Preview

PDF (Author Accepted Manuscript) 17742 GRANDISON_Total_Evacuation_Time_for_Stochastic_Evacuation_Models_2017.pdf - Accepted Version Download (1MB) | Preview

Abstract

International guidelines (IMO MSC.Circ 1533) specify that evacuation models used to certify evacuation performance of passenger ships must demonstrate that the calculated representative evacuation time, the sample 95th percentile time τS, is lower than a prescribed Pass/Fail Criterion Time (PFCT). In this paper a Confidence Interval Convergence Test (CICT) method is presented that minimises the computational burden required to demonstrate that a model design has passed/failed by calculating a CI for the population 95th percentile time, τP, rather than simply relying on τS determined from an arbitrary sample of 500 simulations as specified in the current guidelines. The CICT has comparable pass/fail accuracy to using 500 simulations whilst significantly reducing the number of simulations required when the PFCT is far from the τP. In addition, the proposed method has superior accuracy to the convergent method described in the IMO guidelines. Furthermore, the methodology described in the guidelines fails to identify situations where there may be uncertainty in the pass/fail status due to proximity of τP to PFCT. The CICT identifies these situations and provides a means of resolving the uncertainty. The CICT can be applied to any stochastic evacuation model to determine parameter convergence.

Item Type:	Article
Uncontrolled Keywords:	95th percentile; Confidence intervals; Evacuation modelling; Convergence; Maritime safety; IMO
Subjects:	Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Faculty / School / Research Centre / Research Group:	Faculty of Engineering & Science > Centre for Numerical Modelling & Process Analysis (CNMPA) Faculty of Engineering & Science > Centre for Numerical Modelling & Process Analysis (CNMPA) > Fire Safety Engineering Group (FSEG) Faculty of Engineering & Science > School of Computing & Mathematical Sciences (CMS) Faculty of Engineering & Science
Last Modified:	04 Mar 2022 13:07
URI:	http://gala.gre.ac.uk/id/eprint/17742

Actions (login required)

View Item

Downloads