Publication detail

Optimization of Stator Channels Positions Using Neural Network Approximators

SIKORA, M. ANČÍK, Z.

Czech title

Optimalizace pozic statorových radiálních kanálů s využitím aproximátorů na bázi neuronových sítí

English title

Optimization of Stator Channels Positions Using Neural Network Approximators

Type

conference paper

Language

en

Original abstract

This paper deals with an improvement of synchronous generator cooling. Finding of the best positions of stator radial channels is a way to cooling improvement. An optimization method for finding of the best channels positions is presented here. The main goal of optimization is to achieve uniform temperatures field of winding along slot length, without overheated locations. The Computational Fluid Dynamic (CFD) is used for estimating of stator temperatures. Input parameters for CFD models are computed by optimization algorithm which is able to predict stator temperatures behavior. This algorithm is based on neural networks approximators. There is a feedback between CFD models and optimization algorithm.

Czech abstract

Tento příspěvek se zabývá zlepšením chlazení synchronních generátorů. Cestou ke zlepšení je nalezení optimálních pozic statorových radiálních kanálů. Je zde prezentována metoda optimalizace těchto pozic. Hlavním cílem optimalizace je docílení rovnoměrného teplotního pole podél celého vinutí, bez přehřátých míst. Pro výpočet teplot statoru je použito CFD metod. Vstupní parametry pro CFD modely jsou počítány optimalizačním algoritmem schopným predikovat chování teplot. Tento algoritmus je založen na aproximátorech pomocí neuronových sítí a je zde zpětná vazba mezi CFD modely a tímto algoritmem.

English abstract

This paper deals with an improvement of synchronous generator cooling. Finding of the best positions of stator radial channels is a way to cooling improvement. An optimization method for finding of the best channels positions is presented here. The main goal of optimization is to achieve uniform temperatures field of winding along slot length, without overheated locations. The Computational Fluid Dynamic (CFD) is used for estimating of stator temperatures. Input parameters for CFD models are computed by optimization algorithm which is able to predict stator temperatures behavior. This algorithm is based on neural networks approximators. There is a feedback between CFD models and optimization algorithm.

Keywords in Czech

CFD model, generátor, aproximátor pomocí neuronové sítě, optimalizace

Keywords in English

CFD model, generator, neural network aproximator, optimization

RIV year

2011

Released

21.09.2011

Publisher

Springer

Location

Berlin

ISBN

978-3-642-23243-5

Book

Mechatronics Recent Technological and Scientific Advances

Edition number

1

Pages from–to

365–373

Pages count

9

BIBTEX


@inproceedings{BUT73882,
  author="Michal {Sikora} and Zdeněk {Ančík},
  title="Optimization of Stator Channels Positions Using Neural Network Approximators",
  booktitle="Mechatronics Recent Technological and Scientific Advances",
  year="2011",
  month="September",
  pages="365--373",
  publisher="Springer",
  address="Berlin",
  isbn="978-3-642-23243-5"
}