摘要：To perform advanced model based control it is important to
know what is fed into a system such as a waste or biomass fired
boiler or a pulp digester. In this paper, we present correlations
between the lignin content of different types of wood chips and
their Near-infrared (NIR) spectra. The Principal Component
Regression (PCR) method is used for deriving the correlation,
as well as selecting certain wave lengths. Analysis is made
including different parts of the spectra in the wave length range
700 – 2500 nm. The model is then used as input to an Open
Modelica pulp digester model to tune the reactivity constant of
the dissolution of lignin. The lignin content of wood-chips is
determined on-line through the NIR measurement at the feed
to the digester. Simulations are carried out to determine the
content of residual lignin on fibers at the exit (continuous
digester) or at the end of a cook (batch digester). By comparing
the deviation between predicted values and actual measured
values the reactivity constant of the lignin is determined. The
regression can be made to the NIR spectrum aside of the lignin
content as such. The original content of lignin together with
reactivity may then be used for optimized on-line control of
the digester. It can also be used for diagnostic purposes with
regard to process issues like hang-ups or channeling, as well
as possible sensor faults and data reconciliation.
