Analysis of control systems for the application of data mining methods

In high speed continuous production process the machine control system consists many controllers of different type and different level of complexity. Paper making is an example of such a high-speed continuous process. Can be listed PID, MPC, fuzzy controllers and a few others that are in use in different parts of the paper machine at different stages of production. The machine-directional (MD) and cross-directional (CD) control system includes loops that influence the paper properties in the cross and machine direction. However, PID regulators only look at a small part of the technology, and only account for one or at most cases two process parameters. MD and CD control reacts to deviations from the parameters determined by the technological requirements of the paper. However, these systems do not have the ability to analyse a wider range of the machine or to cover the whole machine area with its scope. Minimalization the time of production of the wrong product means multifarious profits for the production plant. The ideal situation would be to eliminate the wrong product at all. Finding new ways of detecting or prediction the emergence of defect areas and minimizing the time of adjusting the system to nominal conditions of a paper making process is the aim for research in this paper. The paper is an attempt to compile and analyse currently used control loops and the possibility of modifying them to improve the quality of the final product.