A novel safety analysis method for marine cyber-physical systems

Student thesis: Doctoral Thesis

Abstract

Cyber-Physical Systems (CPSs) represent a systems category expected to enhance the safety and improve the efficiency of maritime operations. However, new challenges due to the CPSs complexity are also anticipated leading to an unpredictable system behaviour, thus jeopardising safety.This thesis aims at developing a novel safety analysis method and system for enhancing the safety of the marine CPSs considering both their design and operation. Based on a comprehensive literature review, the safety-related properties for CPSs are identified. Then, the different hazard identification methods are analysed on their effectiveness to identify scenarios linked to the CPSs safety related properties.As the existing literature demonstrates, the existing hazard identification methods such as Fault Tree Analysis (FTA), Failure Modes and Effects Analysis and System-Theoretic Process Analysis (STPA) applications to the CPSs have been criticised for not capturing either the software-intensive character of CPSs or not allowing for quantitative safety analysis.To address these limitations, a novel Combinatorial Approach for Safety Analysis (CASA) is developed by integrating STPA, Events Sequence Identification (ESI) method and FTA. The method initiates with STPA, then employs ESI using input from STPA to identify the different scenarios and develops a Fault Tree based on ESI results. This Fault Tree is populated with STPA results, further refined, and enriched with the FTA results. The final Fault Tree can be used for estimation of the top-event failure rate and frequency, importance measures estimation and uncertainty analysis.The novel method is applied for estimating the failure rate and importance measures estimation of two types of marine CPSs: exhaust gas open-loop scrubber system and a reference cruise ship Diesel-Electric Propulsion (DEP). Failure rate for 12 DEP system alternatives blackout is also estimated. The derived results for the scrubber system and DEP system demonstrate that the developed Fault Tree is much richer than for the previous studies. Moreover, it is demonstrated that the increase of the DEP system reliability/availability does not always result in DEP system blackout frequency reduction, as other system parameters have significant influence on blackout.Based on the CASA method results for the DEP reference system, a novel automated blackout monitoring concept for the DEP system is proposed. This concept is used to estimate the blackout probability variation in time in a virtual environment for the reference DEP system by integrating a number of measured system parameters, historical data and the CASA developed Fault Tree by providing a functional alarm to the crew and allowing better system monitoring and control.The novel CASA method is expected to support the system safety analysis and enhancement during the system design, whilst the proposed blackout monitoring concept is expected to enhance the safety of the DEP system operations.
Date of Award19 May 2020
Original languageEnglish
Awarding Institution
  • University Of Strathclyde
SponsorsUniversity of Strathclyde
SupervisorGerasimos Theotokatos (Supervisor) & Evangelos Boulougouris (Supervisor)

Cite this

'