Assignment 4

From Latent Variable Methods in Engineering
Revision as of 17:37, 30 October 2011 by Kevin Dunn (talk | contribs)
Jump to navigation Jump to search

<rst> <rst-options: 'toc' = False/>

Please provide **short** answers (no more than 5 pages, including all your plots) to the questions below. Here is some `background reading for this assignment < http://literature.connectmv.com/item/124>`._

The aim is to build a soft-sensor model for Kappa number, which is measured at the outlet of the Kamyr digester. This creates problems with feedback control to keep :math:`y` on target. Long time delays exist between variables that affect :math:`y`, and those columns have been time-shifted already to align the data.

We already examined in class the important variables that affect :math:`y` = Kappa number. Now your task is to build as good a soft-sensor model as possible using `the extended data set <http://latent.connectmv.com/images/0/06/Kamyr-full.csv>`_.

Using the data from a bleach kraft mill in Alberta, build a soft sensor model using observations of the first 2000 hours (about a month's worth of data). Then test your model's prediction ability on data from hours 2000 to 3000, and hours 3000 to 4000.

For each of these prediction periods, show time-series plots of

  • SPE,
  • :math:`T^2`,
  • and plot observed *and* predicted on a time-series plot (not the usual scatter plot)
  • Also calculate RMSEP for the prediction periods.

How do you rate the soft-sensor's performance? </rst>