Dobór urządzeń do zgrzewania oporowego punktowego z wykorzystaniem analizy skupień

Abstract

Dobór urządzeń do zgrzewania oporowego punktowego z wykorzystaniem analizy skupień Selection of resistance spot-welding machines with the use of cluster analysis Streszczenie W pracy omówiono zagadnienie doboru urządzeń do zgrzewania oporowego punktowego, stwierdzając na podstawie przeglądu literatury, że nie istnieje obiektywna metoda wspomagająca decyzję zakupu takich zgrzewarek. Do rozwiązania problemu zaproponowano zastosowanie metod analizy skupień. Zgromadzono i opracowano dane diagnostyczne opisujące 35 zgrzewarek oporowych punktowych, a następnie przeprowadzono ich klasyfikacje metodami Warda i k-średnich. Na podstawie porównania wyników wykazano, że analiza skupień może być stosowana jako wstępna metoda wspomagania decyzji zakupu urządzeń do zgrzewania oporowego punktowego.

Application of Multivariate Analysis Methods in Welding Engineering

“…It should be emphasized that there is a factually justified possibility of the complimentary application of various analyses enabling the more precise investigation of issues. Examples of the above-named approach include the application of various cluster analysis techniques [27][28][29][30], cluster analysis and regression analysis [1], principal component analysis and the design of experiment (Taguchi method) [33], principal component analysis and neural network analysis [36] and other variants [36,37].…”

“…In spite of the high potential usability of these methods, their use in cases of welding-related issues is very rare. Figure 2 presents the results of the principal components analysis in relation to stationary spot resistance welding machines based on data presented in publication [29]. In turn J.S.…”

“…In welding engineering, the cluster analysis was used when clustering MAG and TIG welding machines, resistance welding machines, rutile electrodes, covered electrodes, fluxes for automatic submerged arc welding and welding imperfections [27][28][29][30][31][32]. In the above-named cases, good results were obtained using the hierarchical Ward method followed by the verification and supplementation of its results using the non-hierarchical k-means method.…”

“…The results of the linear ordering of stationary spot resistance welding machines based in data presented in publication [29] are presented in Figure 3. The worst device, in terms of assumed diagnostic features, is the welding machine designated as D2, whereas the best device is designated as A8.…”

“…The most popular graphic presentation methods include columnar, star, radar and profile charts as well as Chernoff faces. Figure 4 presents exemplary charts created in relation to data presented in article [29], concerned with the clustering of spot welding machines. A disadvantage of the above-named technique is the necessity of performing the subjective visual assessment of similarities between graphically presented objects (which could be difficult in cases of many slightly varying objects).…”