García, A.C.B ; Leme, L.A.P. ; Pinto, F and Sanchez-Pi, N.
IEA-AIE - The 25th International Conference on Industrial, Engineering & Other Applications of Applied Intelligent Systems, 2012, Dalian. Modern Advances in Intelligent Systems and Tools Studies in Computational Intelligence Volume 431, 2012, pp 9-18
Publication year: 2012


In this paper, we present an alarm management system focused on guiding offshore platform operators’ attention to the essential information that calls for immediate action during emergency situations. Due to the imminent associated nayat chapter 4danger involved in the petroleum operation domain, only well trained workers are allowed to operate in offshore oil process plants. Although their vast experience, human errors may happen during emergency situations as a result of the overwhelmed amount of information generated by a great deal of triggered alarms. Alarm devices have become very cheap leading petroleum equipment manufacturers to overuse them transferring safety responsibility to operators. Not rarely, accident reports cite poor operators’ understanding of the actual plant status due to too many active alarms. A petroleum process plant can be understood as a system composed of a set of equipments interacting with each other to transform and conduct safely a fluid. Each equipment has its own set of rules and safety devices (alarms). The system is subjected to external, non-predictable, effects coming from nature. Hence, the petroleum process plant system can be represented as a set of agents with rules for acting, reacting and interacting with each other. Each equipment is represented as an agent. This AI multi-agent based approach is the basis of our alarm management system for assisting operators to make sense of alarm avalanche scenarios. Our model was implemented using stored procedure statements, installed into the automation circuit of a actual offshore petroleum platform and we are currently collecting results. During initial tests we identified unexpected benefits concerning verification of the process plant automation procedure.