Enhancing safety management through occurrence reporting : modelling and predictive analytics

Abstract

This dissertation explores the integration of Six Sigma methodologies within a Safety Management System (SMS) applied within an aviation base maintenance facility. The ultimate goal was to improve existing safety models and analysis methods by integrating the predictive rationale; into a safety causation model and as an additional analysis approach as a means of aircraft layover preparation. The study begins with a comprehensive literature review of existing safety causation models and methodologies, highlighting their limitations and strengths. Building on this foundation, the research introduces a novel complex-emergent safety causation model designed to address non-linearity, complexity, and emergent properties inherent in aviation safety management. This model is specifically tailored to capture the dynamic interactions and provide a more robust framework for understanding safety incidents and accidents in a Maintenance, Repair, and Overhaul (MRO) environment. To address the dynamic rationale of this model, a Six Sigma Define-Measure-Analyse-Design-Optimise-Verify (DMADOV) methodology is followed to systematically investigate and enhance safety occurrence reporting within an MRO organisation. The methodology integrates advanced data analytics, including Machine Learning (ML) and Natural Language Processing (NLP), to develop predictive models and address the defined problem. The findings demonstrate significant improvements in the accuracy and reliability of safety predictions, thereby enabling more effective preventive measures. The study concludes that incorporating Six Sigma principles within an SMS framework not only optimises safety performance but also supports a proactive and predictive safety rationale in the aviation industry.