A methodology to assess human-automated system conflict from safety perspective. (September 2022)

• Record Type: Journal Article
• Authors: Wen, He; Amin, Md. Tanjin; Khan, Faisal; Ahmed, Salim; Imtiaz, Syed; Pistikopoulos, Stratos
• Is Part Of: Computers & chemical engineering. Volume 165(2022)
• Publication Date: 2022-09
• Subjects: Conflict -- Fault diagnosis -- Failure analysis -- Probabilistic risk analysis -- Human-automated system -- Digitalization; Chemical engineering -- Data processing -- Periodicals; 660.0285
• DOI: 10.1016/j.compchemeng.2022.107939
• Languages: English
• ISSNs: 0098-1354

Abstract

Highlights: Introduces an evolutionary framework that shows how a conflict can lead to an accident. A methodology and associated models to assess and manage conflict risk is presented. The methodology and models are explained using a two-phase separator. The study concludes that human-AI conflict in automation and digitalization should be emphasized. Human-centered design is vital to avoid catastrophic accidents due to conflicts in human-AI systems. Abstract: Most automated systems are equipped with shared control systems with two intelligent decision-makers: humans and automated machines. The contradictory observations, states, goals, and actions may result in a conflict between these two decision-makers. The definitions, cause(s), and path(s) of such a conflict from a process safety perspective have not been explored and assessed. This work introduces an evolutionary framework that shows how a conflict can lead to an accident. A methodology and associated models to assess and manage conflict risk are also presented. The methodology and models are explained using a two-phase separator. The results suggest that there are conflicts associated with faults that may lead to failure. A sensor fault can trigger observation conflict, which may lead to action conflict. The study concludes that human-automated system conflict in automation and digitalization should be emphasized. Human-centered design is vital to avoid catastrophic accidents due to conflicts in human-automated systems.