The article presents an example of analysis of the influence of selected parameters deriving from data acquisition in foundries on the occurrence of Gas porosity defects (detected by Visual testing) in castings of ductile cast iron. The possibilities as well as related effectiveness of prediction of this kind of defects were assessed. The need to rationally limit the number of possible parameters affecting this kind of porosity was indicated. Authors also benefited from expert group's expertise in evaluating possible causes associated with the creation of the aforementioned defect. A ranking of these parameters was created and their impact on the occurrence of the defect was determined. The classic statistical tools were used. The possibility of unexpected links between parameters in case of uncritical use of these typical statistical tools was indicated. It was emphasized also that the acquisition realized in production conditions must be subject to a specific procedure ordering chronology and frequency of data measurements as well improving the casting quality control. Failure to meet these conditions will significantly affect the difficulties in implementing and correcting analysis results, from which INput/OUTput data is expected to be the basis for modelling for quality control.
A significant development of the foundry industry contributes to the creation of high reliability and operational strength castings so that they meet specific standards in accordance with customers’ needs. This technology, however, is inseparably connected with casting defects in finished products. Cast products are subject to various defects which are considered acceptable or not, which is conditioned by the alloy chemical composition and strength characteristics, that is, generally – qualities to be agreed between the foundry and the customer. It is the latter that led the authors to research on designing a tool enabling the most reliable possible assessment of the emerging casting defects, which after proper consultations can be repaired and the casting – sold. The paper presents an original tool named the Open Atlas of Defects (OAD), developed for the last few years to support the evaluation of cast iron defects using Non-Destructive Testing (NDT) casting defects analysis tools (DCC card – Demerit Control Chart, Pareto-Lorenz analysis and ABC analysis). The OAD tool structure was presented as an integral part of the original system module for acquisition and data mining (A&DM) in conjunction with the possibilities of using selected tools for defect analysis support on the example of cast iron casting.