A proposed model to account human factors in safety-critical systems

This article is fostered by some major ferry accidents which occurred during severe conditions. The aim of this paper is to propose a model for predicting the risk of total loss of a vessel, by considering human factors (HF) as the active parts of a safety-critical system (SCS). HF is regarded as a barrier function in the SCS. The safety instrumented system (SIS) framework is adopted to represent HF as safety functions (SF) on a vessel. Environmental conditions are modelled as hazard systems, while the crew are modelled as SF, i.e. sensors, logic solvers and actuators. The HF relations to the SF are discussed, variables are defined and explained, and examples are presented. How the crew control the vessel, both in normal condition and in harsh conditions are analyzed for the case studies. The helmsman and the master are defined as detectors which work as parallel system. The master is also defined as a solver for the system. In the bridge operations the helmsman who steers the rudder and the master who regulates the engine telegraph are defined as the actuators. While in the bridge-to-engine room operations, the engineers or technicians are defined as the actuating elements. All the actuators are assumed to work as a serial system. Mathematical formulations based on IEC 61508 are presented to calculate both the probability of failure on demand of the system (PFDSYS) and the probability of failure per hour (PFHSYS). They are identified as the probability of capsized in normal condition (low demand system) and in hard weather (high demand system) respectively. A specific model of PFD based on system reliability perspective by considering human factors as the main component is proposed. Further validation of the model should be done, for instance by experiments in a ship simulator.