Abstract:Turbines are critical parts in hydropower facilities, and the cross-flow turbine is one of the widely applied turbine designs in small-and micro-hydro facilities. Cross-flow turbines are relatively simple, flexible and less expensive, compared to other conventional hydro-turbines. However, the power generation efficiency of cross-flow turbines is not yet well optimized compared to conventional hydro-turbines. In this article, a Computational Fluid Dynamics (CFD)-driven design optimization approach is applied to one of the critical parts of the turbine, the valve. The valve controls the fluid flow, as well as determines the velocity and pressure magnitudes of the fluid jet leaving the nozzle region in the turbine. The Non-Uniform Rational B-Spline (NURBS) function is employed to generate construction points for the valve profile curve. Control points from the function that are highly sensitive to the output power are selected as optimization parameters, leading to the generation of construction points. Metamodel-assisted and metaheuristic optimization tools are used in the optimization. Optimized turbine designs from both optimization methods outperformed the original design with regard to performance of the turbine. Moreover, the metamodel-assisted optimization approach reduced the computational cost, compared to its counterpart.

-Fluid-structure interaction problems occur in a wide variety of science and engineering fields. In solving such problems, two domains of different characteristics shall be modeled and analyzed through separate kinematic equations. The complexity in the computation of such kind of problems arise mainly from the modeling and analysis of the problem at the interface. Thus, the interfacing condition is also treated as an independent part. A range of computational techniques based on spatial and temporal discretization schemes have been proposed and being employed in solving diverse FSI problems. Monolithic and partitioned approaches are the two broad classes of FSI numerical analysis approaches under the temporal discretization scheme. The monolithic approach requires development of complex mathematical model to represent the whole FSI domain collectively. The partitioned approach, unlike the monolithic approach, make good use of the existing developed and advanced tools with segregate analysis. Nevertheless, the latter requires an additional external code for coupling at the interface. Based on the behavior of the coupling conditions at the interface, FSI computational methods are also categorized under either one-way or two-way coupled approaches. Moreover, the formulation scheme chosen to describe mesh motions, spatial discretization and more other approaches determine classification schemes of FSI solution techniques. In line with the broader classifications based on temporal discretization and coupling schemes, simulation results on a benchmark problem employing selected computational approaches have been analyzed and discussed. Results from a strongly coupled two-way partitioned approach, from an open source code, is compared against one-way coupled partitioned approach utilizing ANSYS code (academic research mechanical). The strongly coupled portioned approach shows a more realistic result than the one-way coupled system.

No abstract

No abstract