TY - JOUR N2 - The paper is aimed at presenting a study of the main limitations and problems influencing the robustness of diagnostic algorithms used in diagnostics of complex chemical processes and to present the selected exemplary solutions of how to increase it. The five major problems were identified in the study. They are associated with: uncertainties of fault detection and reasoning, changes of the diagnosed process structure, delays of fault symptoms formation and multiple faults. A brief description and exemplary solutions allowing increase of the robustness of diagnostic algorithms were given. Proposed methods were selected keeping in mind applicability for the on-line monitoring and diagnostics of complex chemical processes. L1 - http://journals.pan.pl/Content/103646/PDF/3_Koscielny_CPE-39(2)-2018-3.pdf L2 - http://journals.pan.pl/Content/103646 PY - 2018 IS - No 2 KW - complex systems KW - diagnostic system KW - alarm systems KW - diagnosis robustness KW - diagnostic inference A1 - Kościelny, Jan Maciej A1 - Syfert, Michał PB - Polish Academy of Sciences Committee of Chemical and Process Engineering VL - vol. 39 DA - 2018.07.31 T1 - Robust diagnostics of complex chemical processes: Main problems and possible solutions UR - http://journals.pan.pl/dlibra/publication/edition/103646 T2 - Chemical and Process Engineering ER -