Projects / Programmes source: ARIS

Poenostavljeni eksplicitni prediktivni regulator (Slovene)

Research activity

Code Science Field Subfield
2.06.03  Engineering sciences and technologies  Systems and cybernetics  Methods and tools for design and implementation of control systems 

Code Science Field
2.02  Engineering and Technology  Electrical engineering, Electronic engineering, Information engineering 
Evaluation (rules)
source: COBISS
Researchers (5)
no. Code Name and surname Research area Role Period No. of publicationsNo. of publications
1.  16161  PhD Samo Gerkšič  Systems and cybernetics  Researcher  2009 - 2012  132 
2.  10598  PhD Juš Kocijan  Systems and cybernetics  Researcher  2009 - 2012  447 
3.  25655  PhD Boštjan Pregelj  Systems and cybernetics  Researcher  2009 - 2012  123 
4.  02830  PhD Stanislav Strmčnik  Systems and cybernetics  Head  2009 - 2012  488 
5.  12342  PhD Damir Vrančić  Systems and cybernetics  Researcher  2009 - 2012  338 
Organisations (1)
no. Code Research organisation City Registration number No. of publicationsNo. of publications
1.  0106  Jožef Stefan Institute  Ljubljana  5051606000  89,990 
Significance for science
The research topics are of extreme importance for the development of control theory, as a serious opportunity of a breakthrough of advanced control algorithms to the industrial practice is encountered. In the recent scientific literature on automatic control, conventional (non-explicit) MPC very frequently appears as a method of choice among advanced control algorithms, with which more efficient control is achieved but industrial application is not economically feasible due to high implementation costs. A robustification and a decrease of the off-line computational demand of eMPC is currently of key importance for a wide industrial success of this technology in a wide application field of control at the physical level for processes that benefit from advanced handling of constraints. This would open possibilities for further research of more demanding theoretically advanced eMPC approaches with guaranteed robust satisfaction of constraints subject to a given model uncertainty, that would be also suited for safety-critical applications. To the academic research environments, industrial application of advanced control algorithms brings important feedback regarding practically relevant properties of control algorithms and valuable pilot-plant operation experience, and steers the development of theory to practically important directions. Knowledge on the transfer of theory into practice are an important competitive advantage in scientific environments, which also contribute to the theoretical development of the field of anthropocentric systems. The project allowed continued work in the field that was open within the successfull international project CONNECT (6th Framework Programme of European Community) and an applicative upgrade of its results. The published improvements of the eMPC methods are a world-class achievement, because we have pointed out the problem of poor numerical conditioning of eMPC computations and have achieved a considerable improvement of numerical robustness in demanding practical examples. The contribution of the new method of partition reduction by sparse placement of constraints is important because it produces controller partitions with much less regions without a considerable deterioration of regulation efficiency and with better numerical conditioning of computation; opposed to other known methods, this approach does not require neither the prior computation of the non-reduced partition nor the conversion of polyhedra to the vertex description. This facilitates a widening of the application field from univariable to smaller multivariable applications, which is important because the advantages of MPC's constraint-handling abilities are most prominent in multivariable control applications. It is also important that the treatment of output constraints as soft constraints has become computationally feasible with the improved algorithms. Based on these algorithms we have successfully carried out the presented pilot applications. Among them, we wish to point out the challenging application of vertical stabilisation of tokamak plasma on a model of the ITER reactor, where a lower thermal load of in-vessel ohmic coils was achieved by applying a soft output constraint.
Significance for the country
Predictive control algorithms are general-purpose methods for control of dynamic systems, which may be used in very different industries where problems of control in the presence of constraints are encountered. Typical examples are cases where a controlled variable must not exceed a certain physical or security constraint, while it is economically beneficial to operate as close as possible to this limit; in such cases, important savings may be achieved using MPC. Predictive controllers also facilitate enforcement of lower-priority economic targets whenever redundant degrees of freedom are available; they also enable on-line reconfiguration of the control system in cases of actuator or sensor failures. Implementation of the algorithm on a PLC opens up the application field of the process industries where the PLC is the standard automation building block, mostly in chemical and pharmaceutical industries. Due to the possibility of very short sampling times, eMPC may also be used for control of mechanical, pneumatic and hydraulic systems, in robotics, etc. Implementability in microcontrollers and in FPGA chips means that this technology may be used in mass-produced items, such as air-conditioning systems, cars, valve positioners, etc. The expected advantages are in savings of energy and raw materials and in faster or more suitable dynamic response of the controlled processes. Benefits from project results are expected for the co-financer INEA d.o.o. and their custormers. INEA d.o.o. company, which acts in the roles of co-financer and the end-user of the project results, expects the following advantages: 1. Implementation capability for advanced control applications on processes with constraints using simplified eMPC. The results of the pilot applications will allow recognition of suitable processes and marketing of control solutions. 2. The blocks EMPC and KALMAN are available as additional modules of a general-purpose package for PLC-based closed-loop control "IDR BLOK", which will enhance the list of functionality and the market position. 3. The block PERFMON containing the algorithm for controller supervision and evaluation of control efficiency is also available as an additional function block in IDR BLOK. It can also be used with different types of controllers. 4. The gained know-how will be used in the activities of the company in the field of industrial automation services, and will represent a competitive advantage in control of demanding processes. 5. The algorithms in the new function blocks are an opportunity for marketing of advanced technology to automation equipment vendors (Mitsubishi Electric).
Most important scientific results Annual report 2009, 2010, 2011, final report, complete report on dLib.si
Most important socioeconomically and culturally relevant results Annual report 2010, 2011, final report, complete report on dLib.si
Views history