Drive system monitoring: Requirements and suggestions: MODUS-Project Case Study ODT2

A starting point for the case study of a paper machine monitoring is presented along with a requirement analysis. The monitoring is carried out by analyzing drive control measurements and making inferences about the state of the machine based on the measurements and their abnormal changes. The limitations to usefulness of the measurements posed by the properties of the measurement equipment are presented. The demands of different target groups of the monitoring tool and the essential states of the monitoring of the process are also presented in the requirement analysis. Moreover, some attention is paid to the important properties of the paper making process requiring especially good adaptivity and generality from the monitoring method. The rest of the report is devoted to presenting the monitoring process based on feature extraction from the measurements, feature selection, and inference machine. Additionally theoretical background and properties of some common feature extraction methods, including descriptive statistics, trend analysis, principal component analysis, Fourier analysis, and wavelet analysis, are presented.