Abstract
Industrial experience shows that the fire and explosion hazards associated with metal and light metal powders are still frequently underestimated, mainly due to the reliance on database values of explosion parameters that are not representative of real process conditions, and an incomplete understanding of the specific behavior of metal powders under real process conditions. In particular, recent studies on aluminium dust have shown that explosion parameters such as Pmax and KSt are subject to considerable variability, while several influencing factors are still not fully understood. This paper proposes a structured discussion of practical experience gained in the protection against metal powder explosions, focusing on recurring critical issues observed in industrial plants, powder extraction and separation systems, and related protection concepts. Key aspects analyzed include the limitations of database values for explosion parameters, the influence of particle size distribution, oxidation processes, and dust aging, as well as specific issues related to hydrogen formation and the high combustion temperatures typical of light metal dust. The analysis is based on evidence from industrial practice and recent scientific contributions relating to the explosion characteristics of aluminum dust, as well as established methodologies for explosion risk assessment and ATEX classification. Without resorting to specific case studies, the work systematizes lessons learned, identifies open questions, and highlights critical factors that must be considered in order to develop reliable explosion protection concepts. The considerations presented are intended to support designers and safety managers in developing more robust protection strategies that are more in line with actual process conditions, contributing to a more realistic assessment of the explosion parameters of metal dust in industrial practice. This manuscript is presented as a perspective based on industrial experience, systematizing lessons learned rather than reporting new experimental data.
