Department of Chemical Engineering, NTNU

Department of Engineering Cybernetics, NTNU

1999-04-27

The field of process systems engineering and process cybernetics has a strong history at the university in Trondheim, going back to the activities of Jens Balchen (at Control Engineering) and Odd Andreas Asbjørnsen (at Chemical Engineering) in the 1960's. More recently, especially during the last ten years, there has been a renewed activity in this area in Trondheim, and in 1994 the University and SINTEF established "Process Systems Engineering" in Trondheim (PROST) as a strong point center. The group has since then established itself as one of the leading in the world in the area of process systems engineering. The activities are described in more detail below.

The group has received substantial economic contributions from industry, but nevertheless most of the Ph.D. students, who form the basis for the research, have been financed by various programs and projects at the national research council (NFR). However, it is now very clear that, unless the present proposal for a strategic program is funded, there will be a significant drop in the output in candidates from the group; probably from a level of more than 5 Ph.D. candidates per year, to less than half. This is undesirable, especially since the candidates from the program have a unique training, mixing process knowledge and information technology, that make them ideal candidates for building the industry of the future.

At the Department of Chemical Engineering there is a strong activity within the field of process systems engineering. The main focus is on process control and optimization, process modeling and simulation, and on process design and synthesis. The research has been supervised by professors Sigurd Skogestad, Terje Hertzberg and Kristian Lien. Some key figures for the last 5 year period (1994-98) is:

• 16 dr.ing. students have graduated.
• 50 papers have been published in journals, and 80 papers have been published in conferences with review.
• A strong point center (PROST) has been initiated and led by the group.
• A prestigious conference (ESCAPE'7-PSE97) with 400 participants from 34 countries was arranged in Trondheim in June 1997.
• An international publisher has published a text book by Skogestad and Postlethwaite, on Multivariable Feedback Control (Wiley, 1996). The book came in as a close second for the prize for the best text book in control published during the last three years at the IFAC World Congress in 1999.
• In 1998 the Norwegian research council published a review of chemistry research, including chemical engineering, at Norwegian universities (the Weitkamp report). The following is said about the Process Systems Engineering Group: This is a particularly strong group. It enjoys international standing and has many international links. Generally, the research maintains an excellent standard.
• Since 1994 the group has participated actively in several European programs and networks, including the human capital and mobility program on "chemical process control", the CAPE.NET network on process systems engineering, the European Science Foundation network on control of complex systems (COSY), and the CAPE.OPEN and CAPE.GLOBAL projects on standards for process systems engineering. The group has also participated actively in numerous Nordic projects.
• There has been a strong interaction with other departments at the university, especially the process cybernetics group at the Department of Engineering Cybernetics through the strong point center and the INPRO-program.

Process cybernetics is an active research group at the Department of Engineering Cybernetics. The main focus is development of systems technology, meaning modeling, simulation, control and optimization, for the process industries. Professor Bjarne A. Foss and professor em. Jens G. Balchen have supervised the research. During the two latter years assoc. professor Tor A. Johansen and professor Morten Hovd have joined the department contributing actively to the area. Some key figures for the last 5 year period (1994-98) is:

• 14 dr.ing. students have graduated.
• 33 paper have been published in journals or as book chapters, and 39 papers have been published in conferences with review.
• Two international workshops on multi-model techniques (35 participants from 11 countries) and process plant operation (31 participants from 8 countries), respectively, have been arranged in June 1997 and September 1997.
• An international publisher (Taylor & Francis, 1996) has published a book edited by Dr. R. Murray-Smith and assoc. professor T. A. Johansen on modeling and control.
• A modeling technology developed in doctorate studies at our department (1996 and 1997) has been commercialized. The product development has been financed by the Norwegian process industries. SINTEF has been a partner in this activity in which software engineering plays a vital part.
• Since 1994 the group has participated actively in the European Science Foundation network on control of complex systems (COSY).
• There has been a strong interaction with other departments, especially the process systems engineering group at the Department of Chemical Engineering through the strong point center and the INPRO-program.

The strategic plan for the department defines process cybernetics as a strategic technology both in terms of education and research. A consequence of this is the fact that the department has hired two faculty members within this area lately: Assoc. professor Tor A. Johansen in 1997 and professor Morten Hovd in 1998. Further, the Department is increasing the throughput of siv.ing. students from 50-60 per year to 100 per year, and approximately 1/3 of these students will specialize in process cybernetics. This increase is based on guidelines issued by the Ministry of Education.

The technologies within process systems engineering and process cybernetics are quite similar in the sense that they address the same application domain: the process industries. Further, modeling and control are essential technologies in both areas. The main difference appears in the fact that process knowledge is somewhat stronger emphasized in process systems engineering while information technology is more strongly emphasized in process cybernetics. Through this the two areas form an excellent, complementary technology basis covering methodology from basic process knowledge to applications.

The groups have been cooperating for several years. This is a valuable asset, as all the principle actors of the proposed program know each other.

This proposal is based on the fact that there is a continuing and growing need for highly educated people within the area of process systems engineering and process cybernetics. This can be documented by the following facts:

• There has been a high demand for the doctorate students educated within the area. Below we list completed doctorate research projects since 1994 in addition to the first or present employer of the candidate.

• Knut W. Mathisen, Integrated design and control of heat exchanger networks, 1994. (Norsk Hydro research Center)

• Erik A. Wolff, Studies on control of integrated plants, 1994. (ABB Miljø AS)

• Eva Sørensen, Studies on optimal operation and control of batch distillation columns, 1994 (Imperial College, London)

• H. Petter Lundström, Studies on robust multivariable control of distillation columns, 1994. (Aitos AS)

• John C. Morud, Dynamics and control of integrated plants with reactors, 1996. (Norske Skog ASA)

• Ying Zhao Studies on modeling and control of continuous biotechnical processes, 1996. (Process industry; control group, Canada)

• Atle C. Christiansen Studies on optimal design and operation of integrated distillation arrangements, 1998. (Nansen-instituttet)

• Kjetil Havre. Studies on controllability analysis and control structure design, Feb. 1998. (ABB Technology AS)

• Bjørn Glemmestad, Optimal operation of integrated processes. Study on heat recovery systems, Telemark Institute of Technology, 1997 (Supervisor: Truls Gundersen) (Norsk Hydro ASA)

• Håvard I. Moe, Dynamic process simulation - studies on modeling and index reduction, 1995 (Statoil)
• Sverre Støren, Dynamic process optimization, studies on local thermodynamic models and sensitivity estimation, 1997 (Statoil)
• Hanne Kvamsdal, Studies on modelling, 1995 (Aker Engineering AS)

• Steinar Hauan, On the behavior of reactive distillation systems, 1999 (Ass. professor Carnegie Mellon University, Pittsburgh),
• Eirik Bjerve Sund, Synthesis and optimization of chemical reactor systems, 1998 (Fantoft prosess AS)
• Torbjørn Pettersen, Design of membrane-based separation processes, 1994 (Aitos AS)
• Tore Omtveit, Studies on conceptual design of chemical reactors, 1994 (Nycomed Imaging AS)

• Jo Simensen, Learning Support in Process Plant Operation, 1998 (Nycomed Imaging AS)
• Olav Slupphaug, On Robust Constrained Nonlinear Control and Hybrid Control: BMI- and MPC-based State-Feedback Schemes, 1998 (ABB Industri AS and NTNU)
• Kjell Støle-Hansen, Studies of some phenomena in control engineering projects - with application to precipitation processes, 1998 (Falconbridge Nikkelverk AS)
• Tormod Drengstig, On process model representation and AlF3 dynamics of aluminum electrolysis cells, 1997 (Hydro Aluminium AS)
• Stein Wasbø, Ferromanganese Furnace Modelling Using Object-Oriented Principles, 1996 (Elkem ASA)
• Finn Are Michelsen, A Dynamic Mechanistic Model and Model-based Analysis of a Continuous Kamyr Digester, 1995 (Started his own company)
• Ole Bjørn Gjøsæter, Structures for Multivariable Robust Process Control, 1995 (Steinco AS)
• Tor Arne Johansen, Operating Regime Based Process Modeling and Identification, 1995 (SINTEF)
• Are Mjaavatten, Topology-based Diagnosis for Chemical Process Plants, 1994 (Norsk Hydro AS)
• Jan Aage Walseth, Diagnostic Reasoning in Continuous Systems, 1994 (Aker Engineering AS)

• Emil H. Edwin, Modeling, modelbased control, and optimization of thermal cracking and ethylene production, 1994 (Statoil)
• Hilde Christine Meisingset, Modeling and model-based control of a rotary hearth calciner, 1995 (Bayley Norge AS)
• Helge Didriksen, Modeling and identification for control of an industrial copolymer reactor, 1996 (Danish company)
• Øyvind Gundersen, Modelling of structure and properties of soft carbons with application to carbon anode baking, 1999 (SINTEF)

• As documented earlier the Department of Engineering Cybernetics is increasing the throughput of siv.ing. students significantly within the area of process cybernetics.

• Several programs are now in their closing stages.

• The INPRO-program, in which the participating departments have 4 candidates; two of these have graduated and the two others are in the closing stages.
• In the strategic program ORD at the Department of Engineering Cybernetics two of the candidates are linked to the process cybernetics area; one of these have graduated and the other will finish in 1999.
• The 5 year funding of NOK 1.0million/year from NTNU/SINTEF for the strong point center PROST terminates in the summer of 1999. It should be mentioned that industrial support of about NOK 0.5 million/yr has been secured to continue the university part of this activity.

• The process industries is the largest industry in Norway. It has a throughput of NOK 100 billion, of which NOK 80 billion is export revenues. Its interest in process systems engineering and process cybernetics has been very profound as documented by the results presented above. Further, feedback from the industry suggest a continuing and increasing need for the knowledge developed through our strategic research.

• Process and offshore industry companies in Norway all have improved production as an important part of their strategy for maintaining and strengthening global competitiveness. Process systems engineering and process cybernetics technologies is recognized as playing a crucial role to accomplish this goal. Some examples of this are: Statoil has defined model predictive control (MPC) as a key technology, and Norsk Hydro ASA and Elkem ASA has defined model-based methods as key technologies. These are all technologies at the heart of the area for this proposal.

• The engineering and vendor companies are moving from being vendor companies to solution providers for the process and offshore industries. This is quite clear from studying large provider companies like ABB Norway and Kongsberg Simrad, and in particular their hiring of personnel. Further, there is a flourishing industry of small innovative companies (examples of these are Prediktor A.S. in Fredrikstad and Dynamica A.S. in Trondheim) in which the research and personnel within this area is highly in demand.

We have identified four primary areas for research. These areas are selected by the the research interests of the participating professors, and industrial needs.

1. Optimizing and self-optimizing control (4 candidates)

Various optimizing control strategies are now routinely used in parts of the process industry, as an important and profitable technology to improve plant operation. However, limiting factors are the availability of models for some industries and in general the availability of trained people at the Ph.D. level. In addition, the theory of both optimization and modelling needs further development .

The use optimization and the use of equivalent feedback structures to achieve optimal control is an important research area. Both departments have contributed significantly and are well ranked internationally for their research in this area. Some selected references are given in the appendix.

Principal supervisors: Professors Sigurd Skogestad and Bjarne A. Foss

2. Control performance monitoring (3 candidates).

In a prosess plant the nummer of control loops is very large – many plants have several thousand loops. It is common knowledge that in most plants there are many control loops that function poorly. Considering the large number of control loops, it is not reasonable to expect plant operators to be able to monitor the performance of many control loops. Furthermore, the typical operator has limited knowledge of control engineering. Hence, even if an operator is able to assess whether control performance is acceptable, the typical operator will not be able to determine whether significantly better control performance can be achieved nor diagnose the cause for poor control performance.

Research in the area of control performance monitoring was intitiated by a paper published by T. Harris in 1989. A recent review paper by Harris et al. identifies several unresolved areas where further research is necessary.

Principal supervisors: Professors Morten Hovd and Sigurd Skogestad

3. Process modeling (3 candidates)

Process models are used in engineering activities ranging from process synthesis and design, operation planning, process control and monitoring. The increasing demands on product quality, safety, environmental regulations and competitive markets require continuous improvement of chemical process models in minimal time and at minimal cost. Nevertheless, the effort of setting up a detailed mathematical model for a chemical process remains still high because of the large variety of chemical process units and physical phenomena as well as increasing requirements on the sophistication of models. In order to overcome these limitations a systematization of modeling as well as the development of computer-aided modeling tools is required. Both departments have contributed significantly and are well ranked internationally for their research in modeling and model methodology. Some selected references are given in the appendix.

Principal supervisors: Professors Bjarne A. Foss and Terje Hertzberg

4. Control structure selection (2 candidates)

Control structure design (2 candidates).

Academic research up to now has been focused mainly on improving the control algorith, and much less effort has been on defining the control problem itself, including selection of controlled variables, manipulated variables and measurements. The latter is known as the control structure design probem, and although its importance is widely recognized, there has been little research in this area, probably because no one has been able to formulate the problem clearly. Therefore, control structure design in industrial plants is based on experience and heuristics. However, recent research in this field, for example at the Department of Chemical Engineering at NTNU, give a promise to change this.

The research will focus on further developing the systematic methods for control structure design and applying them to some industrially important cases.

Principal supervisors: Professors Morten Hovd and Sigurd Skogestad

---------------

Assoc. professor Tor A. Johansen and adjunct professor Kristian Lien will participate actively in the program.

This research program shall contribute by developing innovative concepts and results within an inter-departmental research environment. The research will be focussed towards an application area of vital impoartnace for Norway; the process industries. The research will be based on in-depth knowlegde in the following three components: (i) process knowledge, (ii) control theory, and (iii) software engineering.

The main goal of the program is to graduate 12 dr.ing. candidates within the field of process systems engineering and process cybernetics; 6 candidates from each of the participating departments. 4 candidates will start in 2000; thereafter 4 candidates will start in 2001 and 2002, respectively. The total cost is NOK 16.2million. This is based on NOK 450.000 per candidate per year. This includes funding for travel, research stays at foreign universities and experiments. The budget is shown in Table 1.

The research will be structured and run according to the following guidelines:

• The candidates will work in groups of at least two, preferably across departmental boundaries.
• The international dimension will be strong in that we will strengthen our links with selected universities. These are, among others, ETH in Zurich, Imperial College in London, RWTH in Aachen and Delft University, and in addition the major technical universities in the Nordic countries. The goal is to coordinate our research with similar strategic programs at these universities.
• There will be a tight coordination with application-oriented projects at the participating department and their cooperating SINTEF-departments.
• We will continue the cooperation developed within the PROST-program, especially the information link to industry.

The program will have a joint leadership. It will be jointly led by department head, professor Sigurd Skogestad from the Department of Chemical Engineering, and department head, professor Bjarne A. Foss from the Department of Engineering Cybernetics. Further professor Morten Hovd and assoc. professor Tor A. Johansen from the Department of Engineering Cybernetics and professor Terje Hertzberg and adjounct professor Kristian Lien from the Department of Chemical Engineering will participate actively in the program.

We plan to appoint a board consisting of all actively participating professors, and members from the Federation of Norwegian Process and Manufacturing Industries (PIL), a process industry company, a vendor company and the Norwegian Research Council.

Curriculum vitae of the principal investigators as well as selected publications can be found in the appendix.

The appendices include

1. 10 selected publications highlighting publications within the research areas to be investigated in this program.
2. CVs for the principal investigators