Industrial plant alarm system forms the core element of almost all modern operator interfaces used to automatically monitor plant conditions and alert plant operators to any significant changes that require diagnosis and/or countermeasures. In this paper, we propose a method for quantitatively evaluating the diagnostic and timely characteristics of alarm system that uses a two-layer cause-effect model to measure three rates used as indices: effective, recall, and timeliness rates. The effective and recall rates are used to evaluate the diagnostic abilities of the alarm system in identifying root causes of assumed malfunctions. The timeliness rate is used to evaluate the plant alarm system's ability to generate diagnostic alarms quickly enough for operators to respond in a timely manner and correct the problem. The case study demonstrated the feasibility of the proposed method.
|ジャーナル||Computer Aided Chemical Engineering|
|出版ステータス||出版済み - 2011|
All Science Journal Classification (ASJC) codes
- コンピュータ サイエンスの応用