KAEiOG - Konferencja Algorytmy Ewolucyjne i Optymalizacja Globalna

XIII Krajowa Konferencja
Warszawa, 21 - 22 września 2011

 
Archiwum / Archives  
 
Piotr Bilski, Jacek Wojciechowski


An Automated Algorithm of the Fuzzy Logic Generation for the Diagnostics of the Analog Systems


Modern diagnostic techniques and methods resort to advanced concepts and algorithms. The concerns of the diagnostic procedures are not only the fault detection, but also detection of the failure in the system under test (SUT), determining which element of the SUT is the source of the fault, and what is the actual value of that element. The aims make the whole procedure complex and difficult for the industrial applications. The knowledge about the SUT’s state must be drawn based on its responses for the excitation signals. Although the detection is relatively easy, the identification and localization face certain obstacles. Firstly, in every object there might be more than one source of the fault. Multiple faults are relatively rare, but make the approach more complicated. Engineers take into account that in practical applications the measurement contains not only useful signals, but noise as well, which should be minimized by the de-noising process. The latter still leaves room for uncertainties. The last issue are the tolerance margins of the system elements, which result in responses different from nominal, although still regarded as correct. For the on-line diagnostic procedure, interaction with the user should be minimal and the automation of the procedure is important. In the ideal case the latter should cover not only the diagnostic process, but also the creation of the object oriented diagnostic system. A generic universal architecture of the diagnostic system was introduced in [1]. It consists of five functional modules, responsible for data acquisition, de-noising, making decision about the SUT state, etc. This paper introduces the method for automation of the decision making procedure. Although the mentioned architecture is to be application independent and so are the modules, the examination of the decision making algorithm will be presented, using the electronic circuit. The paper contains description of the SUT on which the algorithm was tested, the detailed presentation of the algorithm and experimental results. The algorithm is described in two parts: the decision tree generation and fuzzy logic design. These contain the plan of the approach, presentation of the structure storing data from learning experiments, algorithm for the tree construction and the way to project the tree structure into fuzzy logic module.

Inne artukuły z wybranej konferencji.
 
 
 
[Strona główna] [Tematyka] [Ważne terminy] [Komitet Naukowy] [Harmonogram] [Goście specjalni] [Dla autorów] [Dla uczestników] [Dla recenzentów] [Lokalizacja] [Organizatorzy] [Kontakt] [Sponsorzy] [Plakat konferencji] [Galeria] [Archiwum] 

autor: Krzysztof Adamski