Variationsreducering vid mätning av centrumsegring : Ett Sex Sigma projekt vid SSAB

• Main Author: Modig, Jens
• Format: Bachelor Thesis
• Language: Swedish
• Published: 2016
• Subjects: Social Behaviour Law; Samhälls-; beteendevetenskap; juridik; Sex Sigma; centrumsegring; SSAB; försöksplanering; Luleå
• Online Access: http://urn.kb.se/resolve?urn=urn:nbn:se:ltu:diva-59125

A common problem during production of steel by continuous casting is centreline segregation. Centreline segregation is an accumulation of unwanted substances in thecentre of the end product, called slabs. Centreline segregation occurs because different alloys have higher solubility with liquid steel than solid steel. The solidification of steel is done from the outside and in, which makes the alloys move towards the centre where they eventually stick.At SSAB’s production facility in Luleå the centreline segregation is measured in a sampling process, where samples from the end product are cut out and etched in acid bath. After the acid bath the sample is scanned to specify the centreline segregation. Previous studies have concluded that the sampling process has high variance, which creates uncertainties for the specification of the centreline segregation. The scanner and other parts of the sampling process have been excluded during previous studies.The purpose of this project was to identify influencing factors for the acid, determine which of these factors are active for the variation and the settings of the factors to generate low variation. Since experiments and quantitative data collection were needed the method used by this thesis was Six Sigma by the DMAIC-method.A project team was appointed containing experts of both the process and the chemical properties of the acid. During discussions with the team three influencing factors where identified; purity of the acid, time and temperature. These factors where then examined by design of experiments (DoE).During the DoE each sample where etched two times at each setting and a difference between these two was calculated. The sample size was kept at seven. The average of the seven samples and the variance between these samples then became the response variablesfor the experiment. Two DoE were conducted where the purpose of the first was toeliminate the purity of the acid factor to save time. Since it could not be excluded the second experiment continued where the ...