Sustainable optimization based process control

Purpose and goal

The ultimate aim of this project is to facilitate the use of optimization based feedback control in Swedish process industry, in particular in paper mills, by developing a platform for Economical Model Predictive Control (EMPC). Integration of optimization and feedback control represents the next major step in process automation and opens up new possibilities for optimizing economical objectives, but also increased flexibility. In this first part of the project, the objectives have been to develop the core of such a platform and to explore what benefits can be achieved through EMPC.

Expected results and effects

A mathematical model of a paper machine has been implemented and validated by comparison with process data. This model is used as substitute for a real paper machine for control algorithm development. EMPC also uses a process model, and we have developed a suitable model for some of the key relations in the process, and a procedure to identify unknown parameters using process data. One specific EMPC scenario has been studied in a simplified setting, indicating promising results. Soft-sensing techniques have been studied to extend the range of quality variables in the EMPC.

Planned approach and implementation

A major effort has been knowledge transfer between the industrial partners and the academic partner. This has been achieved through a mini course on paper machines, a number of joint project meetings with technical presentations, weekly meetings with academic and industrial representatives, internal reports, sharing of internal documentation, and extensive on-site visits/experimental campaigns in a paper mill. An infrastructure has been set-up for efficient data collection and information sharing. Software tools to facilitate algorithm development have been implemented.