Open AccessDiagnostics of low-capacity solar power station equipment with 2- and 3-valued logic
Author(s) -
Stanisław Duer,
Paweł Wrzesień,
Radosław Duer,
Dariusz Bernatowicz
Publication year - 2018
Publication title -
biuletyn wojskowej akademii technicznej
Language(s) - English
Resource type - Journals
ISSN - 1234-5865
DOI - 10.5604/01.3001.0012.6613
Subject(s) - power station , set (abstract data type) , diagnostic test , reliability engineering , medical diagnosis , computer science , power (physics) , artificial neural network , solar power , divergence (linguistics) , inference , engineering , computer engineering , data mining , artificial intelligence , electrical engineering , physics , medicine , emergency medicine , linguistics , philosophy , pathology , quantum mechanics , programming language
The paper outlines research issues relating to 2- and 3-valued logic diagnoses developedwith the diagnostic system (DIA G 2) for the equipment installed at a low-capacity solar power station.The presentation is facilitated with an overview and technical description of the functional anddiagnostic model of the low-power solar power station. A model of the low-power solar power station(the tested facility, a.k.a. the test object) was developed, from which a set of basic elements and a setof diagnostic outputs were determined and developed by the number of functional elements j of j.The work also provides a short description of the smart diagnostic system (DIA G 2) used for the testsshown herein. (DIA G 2) is a proprietary work. The diagnostic program of (DIA G 2) operates by comparinga set of actual diagnostic output vectors to their master vectors. The output of the comparisonare elementary divergence metrics of the diagnostic output vectors determined by a neural network.The elementary divergence metrics include differential distance metrics which serve as the inputsfor (DIA G 2) to deduct the state (condition) of the basic elements of the tested facility.Keywords: technical diagnostics, diagnostic inference, multiple-valued logic, artificial intelligence.