ABSTRACT
In the industry 4.0 era, large numbers of sensors are employed in industrial plants for safe production. However, the sensors may generate numerous alarms which cannot be effectively interpreted and managed by employees. Thus, industrial plants need a tool to help employees respond to alarms in a proper and timely way. Some works have attempted to isolate faults of an alarm based on occurrence likelihood of the faults. However, if employees only focus on the most likely fault, the real fault with severe consequences can be ignored, which easily causes accidents. The current work proposes a risk-based assessment system of faults to overcome the drawback. Moreover, some research has prioritized alarms to facilitate safety management. However, alarms cannot convey information of faults in a straightforward way, and thus employees have to interpret alarms based on their knowledge and experience before responding. The current work identifies and prioritizes faults, and also pre-stores checking and fixing measures of the faults in the management system; thus, it can effectively guide practical safety management in a straightforward way. The framework of fault management and proposed methods are explained in this paper.
