Performance Assessment and Systematic Design of Industrial Alarm Systems

• Author / Creator

  Adnan, Naseeb A

• In process plants, alarms are configured to notify operators of any abnormalities or faults. However, in practice a majority of raised alarms are false or nuisance and create problems for operators as they face an increasing number of alarms to handle. In the past, many catastrophic incidents happened only because of poor performance of alarm systems. Therefore, a dependable and efficient alarm system is needed to ensure plant safety and uninterrupted operation. To assess performance of a plant's alarm system, it is necessary to evaluate various performances indices. Motivated by this, this thesis develops quantitative relationships among commonly used alarm attributes and performance indices.

  Alarm design techniques, like, deadbands and delay-timers can significantly reduce false and nuisance alarms. However, these techniques introduce some delay in raising the alarm (detection delay). In this thesis, detection delays are calculated using Markov processes for deadbands and delay-timers. A design procedure is then proposed that compromises between detection delay, false alarm rate (Type I error) and missed alarm rate (Type II error) for an optimal configuration. Inclusion of these indices in alarm design makes the system more reliable and effective in nature.

  Filtering is another widely used alarm design technique in industries. Two most commonly used filter types in industry are the moving average filter and the exponentially weighted moving average filter. However, the effect of filter parameters on the alarm detection delay is not well known. We investigated this relationship for the moving average filter and proposed a method to design filter order.

  We proposed a generalized delay-timer framework where instead of consecutive n samples in the conventional case, n1 out of n consecutive samples (n1 <= n) are considered to raise an alarm. For the generalized delay-timer, three important performance indices, namely, the false alarm rate (FAR), the missed alarm rate (MAR) and the detection delay (EDD) are calculated using Markov processes. Also the performance and sensitivity of generalized delay-timers are compared with conventional delay-timers. All theoretical development and alarm design technique proposed in this thesis are validated through simulations and several case studies are conducted to illustrate industrial applications.

• Subjects / Keywords

  • Delay-timer
  • Filtering
  • Alarm Systems
  • Deadband
  • Alarm Management
  • Detection Delay
  • Fault detection

• Graduation date

  Fall 2013

• Type of Item

  Thesis

• Degree

  Doctor of Philosophy

• DOI

  https://doi.org/10.7939/R3930P582

• License

  This thesis is made available by the University of Alberta Libraries with permission of the copyright owner solely for non-commercial purposes. This thesis, or any portion thereof, may not otherwise be copied or reproduced without the written consent of the copyright owner, except to the extent permitted by Canadian copyright law.