Change search
ReferencesLink to record
Permanent link

Direct link
Automatic shape optimization of hydropower flows: the draft tube
2002 (English)Independent thesis Advanced level (professional degree), 20 credits / 30 HE creditsStudent thesis
Abstract [en]

This work illustrates an automatic shape optimization technique for hydropower flows. In the present energy market, the demands of rehabilitation and modernization of old constructions are increasing. Among the renewable energy sources hydropower offers the greatest potential for further improvements. This can be realized by small alterations to the geometry of the waterways. The draft tube is an important part of a hydropower plant: this is especially true for low-head plants where the losses in it often account for about half of the overall losses. Hence, there is great potential for improvements. The draft tube is often shaped as a curved diffuser connecting the runner to the outlet. A way to redesign an existing draft tube is presented which uses an automated optimization based on Computational Fluid Dynamics (CFD) calculations. By this method a design can be predicted in terms of optimize a predefined objective function, here the pressure recovery factor. The design parameters used are all based on spline-functions and they are set under certain restrictions. One is that the new design should fit into the old one. The optimization is performed with the Response Surface Method (RSM) implemented on the commercial code iSIGHT6.0. The RSM, which is a global optimization technique, were chosen since it has shown to give excellent results with relatively few numbers of iterations. The flow simulations are carried out with the commercial code, CFX4, with the standard k-epsilon turbulence model and with boundary conditions based on detailed experimental data. Since the method is fully automatic, great efforts must be spent on the geometrical parameterization and mesh generation. It is also obvious that ways to reduce the overall time for a CFD-based optimization must be found. One way to do this that is evaluated here is to run most of the optimization on a course grid and then refine the grid to get the final result. It must be noticed that draft tube flow predictions are challenging since the flow in these tubes are characterized by many flow phenomena, e.g. separation, swirling flow, streamline curvature, unsteadiness and turbulence among others. It has been demonstrated that it is possible to do fully automatic design optimization of draft tube flows in a reasonable time and with standard tools (iSIGHT, CFX4, Sunblade 1000). The resulting pressure recovery factor increased with over 4% but will possibly increase even further with an extension of the design space. As a suggestion an economy (cost-function) should be considered in the future optimization process. And as well as financial considerations should be made, geometrical and structural considerations should also be taken into account. Hence, in a future work the objective function will be redefined in order to consider these demands. The most time-consuming part in this thesis work was to build the geometry and mesh: great effort should be spent on that in the future.

Place, publisher, year, edition, pages
Keyword [en]
Technology, optimization, optimisation, hydropower, draft tube, RSM, CFD, fluid mechanics
Keyword [sv]
URN: urn:nbn:se:ltu:diva-56168ISRN: LTU-EX--02/085--SELocal ID: cf512163-5ed2-4628-8410-0b6bf6a515ceOAI: diva2:1029555
Subject / course
Student thesis, at least 30 credits
Educational program
Engineering Physics, master's level
Validerat; 20101217 (root)Available from: 2016-10-04 Created: 2016-10-04Bibliographically approved

Open Access in DiVA

No full text

Search outside of DiVA

GoogleGoogle Scholar

Total: 12 hits
ReferencesLink to record
Permanent link

Direct link