theses_web.jpg

Thesis for the Degree of Master of Science

 

Year 2013
Anna Katriina Kähö

Improving Performance of Flotation-deinking by MPC Tuning


The ever tightening competition in the markets forces paper mills to occasional market-related shutdowns and the industry seeks for technological solutions to adapt the production to the shutdowns. This thesis proposes to adapt the model predictive control (MFC) of the flotation-deinking process to reduced production. The aim is to evaluate process performance at different production levels in order to improve the performance in reduced production through MPC adaptation. The research was carried out in a recycled fibre (RCF) plant, which uses MFC as a multivariable control strategy.

The literature review provided insight to the functionality of MFC, and to the RCF process and its multivariable control. MFC implementations in RCF plants were also reviewed.

The experimental study had both analytical and practical goals. Regarding the former, process performance evaluation methods were developed: performance was evaluated using technical and economic measures based on the dynamic and steady-state optimization objectives of the MFC. Measures were used in the performance comparison between different production levels. This resulted in the discovery that performance degrade when the production level decreases as a result of lost ash-yield. The MPC-related causes behind the performance degradation were also analysed: model accuracy was studied through the variation of control, and tuning through the behaviour of manipulated variables. A conclusion was reached that the models were inaccurate when the operating point differs significantly from the operating point of full production. This degraded the contror7e^ponse."It was" further concluded, that the tuning obtained in full production limits higher ash-yield in reduced production. The analysis results were used to improve the process performance in practice.

This thesis recommends two ways to adapt MFC to reduced production: modelling at different production levels and tuning the MPC to improve ash-yield. The MPC tuning trialed at plant resulted in improved ash-yield and significantly improved economic process performance at all production levels. Finally, this-thesis suggests means to further adapt MPC to reduced production and to further develop the MPC structure.

Thesis electronical version can be downloaded from here


This info last modified 22 Nov 2017 by Jukka Kortela