Statistisk processtyrning och mineralanalyser : En fallstudie vid LKAB:s kulsinterverk

• Main Author: Kettunen, Johan
• Format: Bachelor Thesis
• Language: Swedish
• Published: 2015
• Subjects: Technology; Teknik; spc; sps; ARIMA; ARMA; MSA; RnR; Kiruna
• Online Access: http://urn.kb.se/resolve?urn=urn:nbn:se:ltu:diva-43292

LKAB is a mining company in Sweden, which among other things, is running an iron mine and several ore enrichment plants in Kiruna. The main product produced in Kiruna is iron pellets: finely grinded iron ore and additives sintered together. The ore enrichment process separates iron ore from contaminating materials. Enriched ore is thereafter mixed with additives which combined gives the product the desired quality. The mix between enriched ore and additives is critical to quality and a consistent product quality is important for every competitive company.Product quality is defined by several parameters, one of which is chemical composition. Several measurement systems are used within the enrichment process. Data from these systems are used for monitoring and controlling the process, where of additive/iron ore ratio is one adjustable factor. LKAB is dependent on these measurement systems. LKAB is therefore interested in knowing whether these measurement systems are reliable and how data from these measurement systems is best put to use.This study addresses the reliability of two manual and one automated measurement system at LKAB’s pelletization plant KK2 in Kiruna. The study also investigate how data from the measurement systems can be used together with statistical process control.Some conclusions drawn from the study are:- The measurements system error is large in comparison to the internal quality requirements at LKAB but well within the requirements of the ISO standard 3082.- There is no proven bias between the manual and automatic measurement system controlling enriched iron ore.- The bias between pre and post pelletization is significant.- Grinders and x-ray equipment used in the measurement systems needimprovement.- Data from the automated measurement system is auto correlated andcurrent control charts should be replaced or combined with suggestedARIMA-charts.- The additive mixing process is not in statistical balance and need to bestabilized in order to get the best results out of the SPC:system. ...