Detail publikace

Application of Neural Networks for Hot-Air System Control

Originální název

Application of Neural Networks for Hot-Air System Control

Anglický název

Application of Neural Networks for Hot-Air System Control

Jazyk

en

Originální abstrakt

Application of an adaptive semi-inversion neural controller for a laboratory hot-air system model is described. The system can be used to control two parameters – the airflow and the temperature inside the tunnel. The hot-air system displays negative effects commonly occurring in industrial applications – different static amplifications at different operating points, large offset increasing with time, dead zone and noise. The used semi-inversion neural controller is based on an inversion controller, but is capable of solving problems such as oscillating control action, noise sensitivity and ill-estimated parameters in the initial phase of control or adjustment.

Anglický abstrakt

Application of an adaptive semi-inversion neural controller for a laboratory hot-air system model is described. The system can be used to control two parameters – the airflow and the temperature inside the tunnel. The hot-air system displays negative effects commonly occurring in industrial applications – different static amplifications at different operating points, large offset increasing with time, dead zone and noise. The used semi-inversion neural controller is based on an inversion controller, but is capable of solving problems such as oscillating control action, noise sensitivity and ill-estimated parameters in the initial phase of control or adjustment.

BibTex


@article{BUT41840,
  author="Petr {Pivoňka} and Václav {Veleba}",
  title="Application of Neural Networks for Hot-Air System Control",
  annote="Application of an adaptive semi-inversion neural controller for a laboratory hot-air system model is described. The system can be used to control two parameters – the airflow and the temperature inside the tunnel. The hot-air system displays negative effects commonly occurring in industrial applications – different static amplifications   at different operating points, large offset increasing with time, dead zone and noise. The used semi-inversion neural controller is based on an inversion controller, but is capable of solving problems such as oscillating control action, noise sensitivity and ill-estimated parameters in the initial phase of control or adjustment.",
  chapter="41840",
  journal="WSEAS Transactions on Systems",
  number="2",
  volume="3",
  year="2004",
  month="march",
  pages="757",
  type="journal article - other"
}