Projects / Programmes
Računalniško podprto vodenje in optimiranje zveznih in šaržnih procesov. (Študij kombinacij načrtovanih metod in vključevanje pristopov (Slovene)
| Code |
Science |
Field |
Subfield |
| 2.06.00 |
Engineering sciences and technologies |
Systems and cybernetics |
|
| Code |
Science |
Field |
| T125 |
Technological sciences |
Automation, robotics, control engineering |
| P176 |
Natural sciences and mathematics |
Artificial intelligence |
System theory, Control theory, Process modelling, Identification, Simulation, Adaptive Systems, Multivariable systems, Intelligent Control, Supervisory Systems, Sequential Control, Predictive control, Industrial processes, Biomedical processes, Education aspects, Multimedia
Researchers (14)
Organisations (2)
Abstract
Project deals with the problems connected with three basic hypotheses. The first one tends to support the statement that combination of control design methods with included expert knowledge increases the efficacy of control design procedure as well as the simplicity of the multivariable control design methods completed with the corresponding expert system, the multivariable predictive controllers designed bz the aid of individual channel design approach and iterative procedure combining identification and control design steps are investigated. Second hypothesis indicates the possibility that fuzzy control systems and process identification with neural nets improve the properties and usefulness of the adaptive controllers. The latter base on the identification of the inverse fuzzy model given in the relational matrix form. Neural networks on the other hand exhibit very good results in identification of nonlinear processes, what is confirmed also by the trials on laboratory plants. The obtained knowledge enables the application on concrete industrial batch process controlled bz the aid of flexible recipes.
Fuzzy-neural model was used also in the model based control scheme. Third hypothesis deals with the fact that the development of the environment enabling the simulation and direct application of the designed control algorithm on the target computer as well as the use of object oriented tools for modelling and simulation significantly simplify the transfer of theoretical results in practise. The main idea is to enable the implementation of selftuning, adaptive, fuzzy and other complex control algorithms on the low cost computer controllers and programmable logic controllers. In this sense the corresponding adaptations in the MATLAB - SIMULINK environment are proposed. Object oriented modelling and simulation tools completed with problem oriented libraries contribute to the simplification of model development procedure. Problem domains are mainly control technology and biomedicine.
