Methods for the analysis of two dimensional measurement data of paper web
Ptak, Piotr (2006)
Diplomityö
Ptak, Piotr
2006
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi-fe20061595
https://urn.fi/URN:NBN:fi-fe20061595
Tiivistelmä
This work is devoted to the problem of reconstructing the basis weight structure at paper web with black{box techniques. The data that is analyzed
comes from a real paper machine and is collected by an o®-line scanner.
The principal mathematical tool used in this work is Autoregressive Moving Average (ARMA) modelling. When coupled with the Discrete Fourier
Transform (DFT), it gives a very flexible and interesting tool for analyzing
properties of the paper web. Both ARMA and DFT are independently used
to represent the given signal in a simplified version of our algorithm, but the
final goal is to combine the two together. Ljung-Box Q-statistic lack-of-fit
test combined with the Root Mean Squared Error coefficient gives a tool to
separate significant signals from noise.
comes from a real paper machine and is collected by an o®-line scanner.
The principal mathematical tool used in this work is Autoregressive Moving Average (ARMA) modelling. When coupled with the Discrete Fourier
Transform (DFT), it gives a very flexible and interesting tool for analyzing
properties of the paper web. Both ARMA and DFT are independently used
to represent the given signal in a simplified version of our algorithm, but the
final goal is to combine the two together. Ljung-Box Q-statistic lack-of-fit
test combined with the Root Mean Squared Error coefficient gives a tool to
separate significant signals from noise.