Venla Aleta Kuuluvainen
Testing of a fault detection and diagnosis toolkit for a board machine
Due to tightening global competition companies strive to produce their products efficiently. This leads to tighter quality requirements and higher production goals, which can be achieved by an efficient manufacturing execution system that utilizes advanced control applications. These advanced control applications require progressive fault detection and diagnosis (FDD) systems. Therefore, the area of fault detection and diagnosis has received much attention in the last decades. FDD methods have been developed and studied as individual methods as well as combined or enhanced methods. Recent research indicates that the combined and enhanced methods perform better compared to individual methods. In this thesis, the relevant literature is reviewed and a case example of combining two valve stiction detection methods is given. Also, the validation result of a SOM for thickness sensor fouling is presented, and a MATLAB based toolkit is tested. The results of the toolkit testing are compared to the validation results and the toolkit was found to be working correctly. In the future, the toolkit can be used for both educational and research purposes. Two tutorials for using the toolkit are presented in the appendices to serve as a basis for the educational use of the toolkit.
This info last modified 19 Jan 2018 by Jukka Kortela