Applying process mining techniques to improve emergency response planning for chemical spills

Abstract

To evaluate an implementation process for emergency planning and to assess its effectiveness, a technical method based on process mining is utilized for the first time in hazardous chemical accidents from the perspectives of workflow and organizational structure. The proposed method investigates: (1) how the emergency-response tasks of a target emergency plan are effectively executed; and (2) how the emergency-response actors coordinate and cooperate with each other. For this purpose, a fuzzy mining algorithm was applied to reconstruct the drill process for analysing workflow. Social network mining techniques, namely density, reciprocity, node degree, centrality, and handover of work metrics, were utilized to explore the organizational structure among the actors. By means of emergency-response drills, a pre-defined emergency plan for chlorine release was analyzed, for which mining results were obtained. Regarding the workflow, four deviations were detected from the target emergency plan, which suggest ways of improving the plan rules. Regarding the organizational structure, the most influential actors A1, A5, A10, whose actions directly determined the plan implementation and outcome. The proposed method can facilitate the improvement of existing processes and organizational structures for chemical accident emergency response planning.