Validation of the impact of architectural flaws in six machine risk estimation tools

摘要

To address the hazards inherent in industrial machinery, machine designers and users must conduct risk assessments and use risk reduction measures. Machine risk estimation plays a crucial role in choosing and prioritizing risk reduction methods (e.g., level of performance required for the safety-related control system). A large number of machine risk estimation tools exist, and each tool has its own specific parameters and architecture. Flaws in a tool may bias risk estimation and lead to the adoption of inappropriate or insufficient risk reduction methods. An earlier study identified potential flaws in risk estimation tool parameters and architecture and proposed construction rules. In this paper, potential flaws in the architecture of six tools are tested by 25 machine safety experts. Four scenarios involving industrial machines and representing different risk levels were used for that purpose. The experimentation served to validate the potential flaws which were the impact of (i) a non-uniform distribution of risk levels, (ii) greater relative weight given to one parameter, (iii) discontinuity in risk levels and (iv) an overly sensitive risk matrix. Construction rules for machine risk estimation tools that should help improve inter- and intra-user repeatability, making the tools more reliable and robust, are proposed. The recommendations can potentially guide users of risk estimation tools when choosing, designing or using a tool. The results of this study will also help improve national and international standards in machinery risk assessment.