Year 2009
Erkki Salo

Long-Term data acquisition from printing press - data processing and methods of analysis

Modern automation systems and process instruments have enabled a great amount of online measurements to be collected from a process. The correct analysis of the collected data leads to an improved understanding of the process which may be used to improve the process efficiency and product quality. The aim of the literature part of this thesis was the familiarization with various analysis methods and with the theory of data pretreatment. The objective of the experimental part was to create a data collection system for an operational printing machine as well as develop computer programs to support research within the project in question.

The theoretical part of this thesis investigates multivariate analysis methods such as the principal component analysis and the least squares method, as well as data pretreatment techniques necessary for the aforementioned methods. In addition to these, the control charts of statistical process control are also reviewed.

In the experimental part a data collection system was designed and implemented for Technical Research Centre of Finland. The large-scale and long-term data collection began in October 2008 and was completed in June 2009. Additionally several programs were written to facilitate the inspection, combination and processing of the data in the databases. Both the data collection system and the assisting programs were realized within schedule and they were observed to perform according to their design.

This info last modified 27 Apr 2024 - contact webmaster