Derivation of Diagnostic Models Based on Formalized Process Knowledge

详细信息    查看全文

• 作者：Esteban Arroyo^* ; ^{e.arroyo@hsu-hh.de" class="auth_mail" title="E-mail the corresponding author} ; Denis Schulze^* ; Lars Christiansen^* ; Alexander Fay^* ; ^{alexander.fay@hsu-hh.de" class="auth_mail" title="E-mail the corresponding author} ; Nina F. Thornhill^** ; ^{thornhill@imperial.ac.uk" class="auth_mail" title="E-mail the corresponding author}
• 关键词：Process ; knowledge representation ; model ; fuzzy inference ; diagnosis ; information exchange
• 刊名：IFAC-PapersOnLine
• 出版年：2014
• 出版时间：2014
• 年：2014
• 卷：47
• 期：3
• 页码：3456-3464
• 全文大小：1177 K

文摘

Industrial systems are vulnerable to faults. Early and accurate detection and diagnosis in production systems can minimize down-time, increase the safety of the plant operation, and reduce manufacturing costs. Knowledge- and model-based approaches to automated fault detection and diagnosis have been demonstrated to be suitable for fault cause analysis within a broad range of industrial processes and research case studies. However, the implementation of these methods demands a complex and error-prone development phase, especially due to the extensive efforts required during the derivation of models and their respective validation. In an effort to reduce such modeling complexity, this paper presents a structured causal modeling approach to supporting the derivation of diagnostic models based on formalized process knowledge. The method described herein exploits the Formalized Process Description Guideline VDI/VDE 3682 to establish causal relations among key-process variables, develops an extension of the Signed Digraph model combined with the use of fuzzy set theory to allow more accurate causality descriptions, and proposes a representation of the resulting diagnostic model in CAEX/AutomationML targeting dynamic data access, portability, and seamless information exchange.