Skip to main content
European Commission logo print header
Content archived on 2022-12-23

Adaptive and Predictive Control with Physical Constraints

Exploitable results

Aerothermal System is a portable physical model of a thermal distributed parameter system (DSP). It can be applied in the research and testing of advanced control algorithms for DPSs before their implementation on real industrial processes. The system also supports practical education in courses of automatic control systems analysis and design. Key features include the ability to simulate multi-Input, multi-output (MIMO) systems and one class of thermal DSPs. A number of steps must be undertaken before a control algorithm can be successfully applied to a real environment. The main benefit of the laboratory system can be expected in obtaining first practical experiences with created control algorithms and in shortening a test phase before their implementation to a real industrial process.
Water Power Plant is a portable physical model of a pumped storage hydro-plant. It can be applied in the research and testing of advanced control algorithms before their implementation on real industrial processes. The system is also suitable for supporting practical education in courses of automatic control systems analysis and design. The priority of the system is its ability to simulate single-input, single-output (SISO) and multi-input, multi-output (MIMO) processes with a variable time delay. There are a number of steps that must be undertaken before a control algorithm can be successfully applied to a real environment. The main benefit of the laboratory system can be expected in obtaining first practical experiences with created control algorithms and in shortening a test phase before their implementation to an industrial process. The main market for the result is universities and research laboratories.
The PREDCON package is in the form of a MATLAB toolbox and is based on new methods of multiobjective optimization. These methods employ two different approaches to solving multiobjective problems: optimality principles and dialogue with designer. All these techniques are based on Rykov's methods of deformed configurations and dialogue method of mirror reflections. The design problems are generally multiobjective because several conflicting design aims need to be achieved simultaneously. There has been a renewed interest in multiobjective control system design using parameter search techniques. Several multiobjective computer-aided control system design CACSD packages are available; successful applications have also been reported.
PREDHEAT is a family of model-based predictive control (MBPC) algorithms intended for robust predictive and adaptive control of various industrial processes. The algorithms are easy to implement, using standard commercially available software or special purpose software environments like PREDICTOR. The main usage of the algorithms is in the design of advanced controllers for industrial processes. PREDHEAT has enabled a stable prediction of the process output using a properly chosen model of the controlled process and minimization of given cost function with or without constraints. On-line determination of process parameters is a further key element of the MBPC algorithms. The family of algorithms contains: conventional generalized predictive control (GPC) algorithm, stabilizing constrained receding horizon predictive control (CRHPC) algorithm and continuous-time spline-based predictive control (SBPC) algorithm. The algorithms have been extensively tested in laboratory conditions and currently are implemented in a standard industrial heat exchange station for control of outlet water temperature of heat exchangers. Benefits can be expected in increasing quality of control and cost-effectiveness of the industrial technologies. The controllers can adapt to changes in the process dynamics and the disturbances.

Searching for OpenAIRE data...

There was an error trying to search data from OpenAIRE

No results available