Publication detail

Predictive Controller with Neural Network Classificator

DOBIAS, M. PIVONKA, P.

Original Title

Predictive Controller with Neural Network Classificator

Czech Title

Prediktivní regulátor s neuronovým klasifikátorem

English Title

Predictive Controller with Neural Network Classificator

Type

journal article

Language

en

Original Abstract

Predictive controllers are nowadays a necessary part of industrial processes. The main reason for application of this kind of controllers is their specific attributes which provide many advantages in this area. In this paper we will design multi-model predictive controller and make a comparison with standard predictive controller in closed loop which contains plant with uncertain parameters. Also we will be interested in problems which can be caused by noise in the closed loop. In the last part of the article we apply both predictive controllers in control circuit which contains the physical model.

Czech abstract

Prediktivní regulátory jsou v dnešní době nevyhnutnou součástí průmyslových provesů. Hlavním důvodem jejich aplikace jsou specifické vlastosti těchto regulátorů, které přinášejí mnoho výhod. V článku je navržen prediktivní regulátor, který pracuje se třemi modely a je porovnán se standardním prediktivním regulátorem. Dále se článek zabývá vlyvem šumu v regulační smyčce. V poslední části článku jsou prediktivní regulátory porovnány při řízeni v regulačním obvodu obsahujícím fyzikální model.

English abstract

Predictive controllers are nowadays a necessary part of industrial processes. The main reason for application of this kind of controllers is their specific attributes which provide many advantages in this area. In this paper we will design multi-model predictive controller and make a comparison with standard predictive controller in closed loop which contains plant with uncertain parameters. Also we will be interested in problems which can be caused by noise in the closed loop. In the last part of the article we apply both predictive controllers in control circuit which contains the physical model.

Keywords

predictive control, artificial neural network, plant with uncertainty, classification

RIV year

2011

Released

26.05.2011

Publisher

Technical University Sofia, branch Plovdiv

Location

Plovdiv

Pages from

259

Pages to

263

Pages count

5

BibTex


@article{BUT72641,
  author="Michal {Dobias} and Petr {Pivoňka}",
  title="Predictive Controller with Neural Network Classificator",
  annote="Predictive controllers are nowadays a necessary part of industrial processes. The main reason for application of this kind of controllers is their specific attributes which provide many advantages in this area. In this paper we will design multi-model predictive controller and make a comparison with standard predictive controller in closed loop which contains plant with uncertain parameters. Also we will be interested in problems which can be caused by noise in the closed loop. In the last part of the article we apply both predictive controllers in control circuit which contains the physical model.",
  address="Technical University Sofia, branch Plovdiv",
  chapter="72641",
  institution="Technical University Sofia, branch Plovdiv",
  journal="TechSys 2011 Internationa Conference of Engieering, Technologies and Systems",
  number="1",
  volume="16",
  year="2011",
  month="may",
  pages="259--263",
  publisher="Technical University Sofia, branch Plovdiv",
  type="journal article"
}