Proposal of data pre-processing for purpose of analysis in accordance with the concept industry 4.0

Abstract

The paper deals with a process of lowering amount of errors on automated screwing mechanism. Research focuses on analysis and individual evaluation of problematic robot screwing heads. The anticipated result is finding a problematic point of the whole process, what should lead to lowering the amount of errors in the process, shortening errors duration, and to improving effectiveness of predictive maintenance. Many errors occur in the screwing process due to not meeting the criteria of key parameters – adherence pressure, axial torque, screwing depth. There are additional parameters that may affect the process, like quality and chemical composition of material, its thickness, placing of the screw, etc. The paper focuses on identification of useful data sources, joining and pre-processing the downloaded data, and performing basic analysis thus preparing for analysis via DM methods. We will use the methods of KDD and Big Data, with respect to Industry 4.0 and CRISP-DM methodology.