A single hazard-based matrix method for data-driven risk assessment in aviation

Abstract

The reason we investigate a flight accident is simple. We try to get information about system failure and its cause, share lessons learned and plan successful prevention to future threat to promote aviation safety. This process is called risk management. Risk assessment has a central role in making a right decision in risk management. As it is well known, however, risk assessment is not adequately operated or not performed at all in many case even if risk management is implemented. There could be several reasons for this. These include too flexible guideline, inconsistent definitions of related concepts, feasibility of existing method in aviation environment and data availability, just to name a few. In this condition, implementation risk assessment has been faced with two main problems. There has been lack of consensus on 1) the object (what it is to be assessed) and 2) feasible method (how to do it) of risk assessment.In this study, we propose the safety-event-data-driven methodology to assess risk of hazard in aviation safety in quantitative way. We construct the entity of events that involve two attributes of cause and consequence. By matrix transpose and multiplication, event table transforms into single hazard-based matrix. As a result, we get total number of occurrences for each hazard and their number of occurrences by consequence. Based on the result, the risk associated to hazard is evaluated in terms of likelihood of occurrence multiplied by severity of consequence. Then, the risk of hazard is ranked for prioritization. We performed a case study with National Transportation Safety Board (NTSB) accident/incident to demonstrate the applicability of our methodology. Our methodology is very simple, fast and flexible. The result is coherent, useful and replicable, making it understandable by non-experts.