Controlling industrial processes in pulp&paper and bioenergy by MPC with greybox models

Important results from the project

Accomplishment of aim and objectives: 1. The Greybox-MPC approach was demonstrated on two industrial processes: cardboard manufacture and bioethanol production. Real time sensors were used to monitor the processes. Physical models have been used in the MPC. 2. The possibility of using real-time sensors in combination with the Greybox-MPC approach was verified on two continuous industrial processes. 3. The Greybox-MPC approach applied to the two studied processes has potential to result in substantial cost savings for the companies making implementation a viable alternative

Expected long term effects

Essential results and expected effects: By using step responses and real process data the Greybox-MPC approach has been shown feasible. Feedback control with the Greybox-MPC has been demonstrated at laboratory scale. The expected cost savings have been credibly shown. Two scientific communications have been drafted for publication of the findings.

Approach and implementation

The original intentions of the project have been realised: Object-oriented software (Wolfram Mathcore) has been used for design of generic and transferable greybox-MPC solutions that can be implemented on a variety of industrial processes. The Greybox-MPC approach has been developed in an industrial plant environment (Iggesunds Bruk) and in close cooperation with a large scale bioprocess plant for bioethanol (Lantmännen Agroetanol) New sensors have been tested in MPC setups at laboratory scale and have led to implementation at the plant (Lantmännen Agroetanol)