Abstract
Continually managing occupational risks is the cornerstone of any prevention program. ISO 12100:2010 defines risk in the field of safety of machinery as the combination of the probability of occurrence of harm and the severity of that harm. The means available to help with risk identification present the hazards isolated from one another. Moreover, estimation of probability is a recurrent problem. To overcome these issues, this paper proposes a method using logical analysis of data to generate patterns from belt-conveyorrelated accident investigation reports. The patterns represent accident scenarios involving combinations of hazards and risk factors. The probabilities of the patterns are estimated to establish a hierarchy of prevention measures that will be part of a prevention action plan. Updating data from accident reports impacts risk estimation, thus entailing adjustment of the prevention action plan.