An alternative approach of Risk Analysis for Multivariable Alarm System

Abstract

Alarm system of a plant works as a safeguard to prevent hazards and unexpected accidents. But too many false alarms may make the operators indifferent to critical alarms that may lead to serious mishaps. In this study, the risk based multivariable alarm system proposed by [1] has been restudied and an alternative approach of quantifying the system failure probability is proposed. Since in all the modern chemical plants there are adequate safety measures and the safety is ensured by installing different instruments, and the failure data of these instruments are available in [2], so the risk of the system can be quantified by applying statistical methods as a combination of several instruments If the risks and hazards associated with a system can be properly identified, the hazards can be correlated with the failure of the different parts of the instrumentation system and the probabilities of failure can be estimated [3]. Then it is necessary to know the consequence factor of a hazardous situation since risk depends on both the failure probability and the consequence probability. According to the severity of failure probabilities of the instruments of the physical system, the alarms can be indexed and all alarms of a single system can be consolidated into a single alarm, which will notify the operator about the importance of the situation. Since risks can also be properly indexed, it is easier for the operator to take necessary steps as soon as he gets the alarm. Thus the purpose of the alarm system will be fully satisfied = to keep the process operation safe and sound. In this study, failure probabilities were calculated for a simple model process namely a tank heating system.

DOI: http://dx.doi.org/10.3329/jce.v26i1.10189

JCE 2011; 26(1): 75-79