SUBSTANTIATION OF THE CHOICE OF METHODS OF NON-DESTRUCTIVE TESTING OF ELEMENTS OF ENERGY EQUIPMENT USING A FUZZY LOGIC APPARATUS

• Main Authors: Ihor Hryhorenko, Svіtlana Hryhorenko, Mykola Ovcharenko
• Format: Article
• Language: English
• Published: National Technical University "Kharkiv Polytechnic Institute" 2020-10-01
• Series: Сучасні інформаційні системи
• Subjects: control object; defect; non-destructive testing; fuzzy logic; heuristic analyzer; fuzzification
• Online Access: http://ais.khpi.edu.ua/article/view/213339

The paper illustrates the solution of the problem of choosing methods of quality control of manufacturing parts and assemblies of power equipment using a fuzzy logic device. The main methods of non-destructive testing for the detection of surface and internal defects are considered, as well as the main indicators of quality of metal products. The types of metal defects and welded joints are inspected. The description of the equipment and means of control for detection of defects is executed. The sequence and methods of quality control by ultrasonic, capillary and magnetic powder methods of control are described in detail. The results of quality control of parts during production and during their operation are obtained. The analysis of the revealed defects is carried out. An example of using an integrated approach to control is given. The obtained results of control of the percentage of coincidence of detection of defects on the product are analyzed. Comprehensive quality control was performed by visual, ultrasonic, capillary and magnetic powder methods of non-destructive testing to determine the percentage of coincidences of defects. By creating a heuristic analyzer based on the interface of the fuzzy logic system Fuzzy Logic Toolbox of the Matlab program, an example of determining a combination of non-destructive testing methods for quality control of a steam turbine bearing liner is considered. Computer simulation according to the Mamdani algorithm is carried out, which consists of fazzification with determination of ranges of change of input values for each example, task of distribution functions for each input parameter; calculation of rules based on the adequacy of the model; defuzzification with the transition from linguistic terms to quantitative assessment and graphical construction of the response surface. The simulation made it possible to determine the optimal combination of non-destructive testing methods, which provides the highest quality of defect detection in the steam turbine bearing liner.