Search

The six papers included in this special issue represent a selection of extended contributions presented at the 17th International Conference on Hybrid Artificial Intelligent Systems, HAIS 2022 held in Salamanca, Spain, September 6th-8th, 2022, and organized by the BISITE group at the University of Salamanca. The International Conference on Hybrid Artificial Intelligence Systems (HAIS 2022) has become a unique, established, and broad interdisciplinary forum for researchers and practitioners who are involved in developing and applying symbolic and sub-symbolic techniques aimed at the construction of highly robust and reliable problem-solving techniques to present the most relevant achievements in this field. HAIS Series of Conferences provides an interesting opportunity to present and discuss the latest theoretical advances and real-world applications in this multidisciplinary research field. [ABSTRACT FROM AUTHOR]

• An autonomous root-cause fault diagnosis framework for industrial processes. • There is no need for a priori process knowledge and expert intervention. • A novel multi-dimensional joint xD-Markov machine for probability analysis. • A fast symbolic dynamic based normalized direct transfer entropy (SDNDTE). • Real-time application of SDNDTE for root-cause diagnosis in industrial processes. • Definition of Immediate and Source intermediate variable required for Direct TE. • Autonomous Algorithms for determining IIVs, SIVs and root-cause of the process fault. This paper proposes a general framework for autonomous root-cause fault diagnosis in a complex process. In this framework, as a prerequisite step after conducting fault detection, the potential root-cause candidates are selected using a contribution score-based method (e.g. accumulative rate contribution scores). Then a fully automated procedure is proposed to determine the root-cause(s) of the detected fault amongst potential candidates without a priori knowledge of the base-line model or intervention of an expert. To locate the root-cause variable(s), firstly symbolic dynamic-based normalized transfer entropy (SDNTE) is proposed to generate an initial causal graph of root-cause candidates. Then symbolic dynamic-based normalized direct transfer entropy (SDNDTE) is proposed and utilized for pruning the initial graph (i.e. discard indirect and spurious causal edges). To this aim, explicit definitions of immediate intermediate variables (IIV) and source intermediate variables (SIV) are given and systematic algorithms are developed to find them efficiently. At last, a topological approach is proposed to locate the root-cause variables according to the pruned causal graph. The proposed generalized SDNDTE approach overcomes the computational complexity in the calculation of normalized direct transfer entropy (NDTE) using conventional multi-dimensional kernel probability density functions (PDFs) fitting, which enables real-time root-cause fault diagnosis. To demonstrate the effectiveness and applicability, the proposed autonomous scheme is tested on a numerical example at first, and its computational complexity is analysed and compared to other existing works. Finally, it is validated on the Tennessee Eastman process (TEP) benchmark model. [ABSTRACT FROM AUTHOR]