Publication detail

Development of high specific speed Francis turbine for low head HPP

KRAUSOVÁ, H. OBROVSKÝ, J. ŠPIDLA, J. ZOUHAR, J.

Original Title

Development of high specific speed Francis turbine for low head HPP

English Title

Development of high specific speed Francis turbine for low head HPP

Type

journal article - other

Language

en

Original Abstract

Nowadays we can commonly encounter with revitalizations of an original HPPs which were earlier fitted with Francis turbines. They were often placed to the locations with low head and higher discharge which means high specific speed (ns > 400). Generally it is quite complex to design Francis turbines for such high specific speed. These very old turbines usually have lower efficiency due to the earlier limited possibilities of hydraulic design. An exchange of a water turbine with another type can be quite expensive and therefore it can be more suitable to change only an old runner for a new one. In this article the design process of high specific speed turbine ns = 430 is described. Optimization was done as the full-automatic cycle and was based on a simplex optimization method as well as on a genetic algorithm. For the parameterization of the runner blade, the BladeGen software was used and CFD (Computational Fluid Dynamics) analysis was run in Ansys CFX v.14 software. The final shape of the runner blade was reached after computing about 1000 variants which lasted about 250 computational hours.

English abstract

Nowadays we can commonly encounter with revitalizations of an original HPPs which were earlier fitted with Francis turbines. They were often placed to the locations with low head and higher discharge which means high specific speed (ns > 400). Generally it is quite complex to design Francis turbines for such high specific speed. These very old turbines usually have lower efficiency due to the earlier limited possibilities of hydraulic design. An exchange of a water turbine with another type can be quite expensive and therefore it can be more suitable to change only an old runner for a new one. In this article the design process of high specific speed turbine ns = 430 is described. Optimization was done as the full-automatic cycle and was based on a simplex optimization method as well as on a genetic algorithm. For the parameterization of the runner blade, the BladeGen software was used and CFD (Computational Fluid Dynamics) analysis was run in Ansys CFX v.14 software. The final shape of the runner blade was reached after computing about 1000 variants which lasted about 250 computational hours.

Keywords

Francis turbine, high specific speed, CFD, optimization method, objective function

RIV year

2013

Released

06.12.2013

Publisher

Association for Engineering Mechanics

Location

Praha

Pages from

139

Pages to

148

Pages count

10

Documents

BibTex


@article{BUT100654,
  author="Hana {Krausová} and Jiří {Obrovský} and Jiří {Špidla} and Josef {Zouhar}",
  title="Development of high specific speed Francis turbine for low head HPP",
  annote="Nowadays we can commonly encounter with revitalizations of an original HPPs which were earlier fitted with Francis turbines. They were often placed to the locations with low head and higher discharge which means high specific speed (ns > 400). Generally it is quite complex to design Francis turbines for such high specific speed. These very old turbines usually have lower efficiency due to the earlier limited possibilities of hydraulic design. An exchange of a water turbine with another type can be quite expensive and therefore it can be more suitable to change only an old runner for a new one. In this article the design process of high specific speed turbine ns = 430 is described. Optimization was done as the full-automatic cycle and was based on a simplex optimization method as well as on a genetic algorithm. For the parameterization of the runner blade, the BladeGen software was used and CFD (Computational Fluid Dynamics) analysis was run in Ansys CFX v.14 software. The final shape of the runner blade was reached after computing about 1000 variants which lasted about 250 computational hours.",
  address="Association for Engineering Mechanics",
  chapter="100654",
  institution="Association for Engineering Mechanics",
  number="2",
  volume="20",
  year="2013",
  month="december",
  pages="139--148",
  publisher="Association for Engineering Mechanics",
  type="journal article - other"
}