Improving quality of statistical process control by dealing with non-normal data in automotive industry

Abstract

[EN] Study deals with an analysis of data to the effect that it improves the quality of staƟsƟcal tools in processes of assembly of automobile seats. Normal distribuƟon of variables is one of inevitable condiƟons for the analysis, examinaƟon, and improvement of the manufacturing processes (f. e.: manufacturing process capability) although, there are constantly more approaches to non‐normal data handling. An appropriate probability distribuƟon of measured data is firstly tested by the goodness of fit of empirical distribuƟon with theoreƟcal normal distribuƟon on the basis of hypothesis tesƟng using programme StatGraphics Centurion XV.II. Data are collected from the assembly process of 1st row automobile seats for each characterisƟc of quality (Safety RegulaƟon ‐S/R) individually. Study closely processes the measured data of an airbag´s assembly and it aims to accomplish the normal distributed data and apply it the staƟsƟcal process control. Results of the contribuƟon conclude in a statement of rejecƟon of the null hypothesis (measured variables do not follow the normal distribuƟon) therefore it is necessary to begin to work on data transformaƟon supported by Minitab15. Even this approach does not reach a normal distributed data and so should be proposed a procedure that leads to the quality output of whole staƟsƟcal control of manufacturing processes