Towards zero-defect manufacturing with a combined online - offline fuzzy-nets approach in wire electrical discharge machining

Current practice of wire electrical discharge machining (W-EDM) is not defect free and still a number of parts produced by WEDM processes are rejected due to different defects. The main defects related to surface quality of W-EDMed parts include occurrence of surface lines, surface roughness, and recast layer (also called white layer). In this paper we characterize the different types of defects and for each of them we propose the appropriate approach to handle it. To achieve zero-defect manufacturing in W-EDM, one can intervene before starting machining operations through selecting the best combination of values of process parameters to meet required levels of performance for part quality characteristics, and during machining operations through monitoring machining conditions to predict any potential defect and take appropriate actions to prevent the effective occurrence of defects if and when predicted. This paper addresses both issues and a combined online - offline fuzzy-nets approach is proposed. It is noteworthy, that for the first time in the published literature, an approach is presented to prevent the occurrence of surface lines in W-EDM; this is accomplished by our experimentally validated system by: (i) high frequency sensing and processing of the current signal in the discharge zone; (ii) the development of a new algorithm for the detection and calculation of the duration of a series of consecutive short circuits; and (iii) the proactive adjustment of the pulse–off time through an adaptive fuzzy nets system.