https://www.mdu.se/

In this paper the digester house process variables

are first studied to identify the process variations in

different time perspectives. The focus is on the

perspective of controlling the process on-line for

the next 24 hours. With a mathematical model first

the cook and the flows associated with the digester

house are modelled. From this different scenarios

with different “faults” are implemented in the

model and compared to the process data. From this

the performance can be predicted for different

conditions. Key numbers are determined from this

and these are used together with suitable objective

functions to make optimal production planning for

model based control. The set points are sent to the

control system. By determining the reactivity of

different qualities of wood chips by measuring NIR

and measuring final kappa, a tool for feed forward

control is developed.