Search

In this article, the previously developed single block fully coupled algorithm [1,2] for solving three-dimensional incompressible turbulent flows is extended to resolve transient flows in multiple rotating reference frames using the arbitrary Lagrange–Euler (ALE) formulation. Details on the discretization of ALE terms along with a recently developed extension to the conservative and fully implicit treatment of multi-block interfaces into three-dimensional space are presented. To account for turbulence, the kω − SST turbulence model in ALE formulation is solved using Navier–Stokes equations. This multi-block transient coupled algorithm is embedded within the OpenFOAM® Computational Fluid Dynamics (CFD) library, and its performance evaluated in a real case involving a turbulent flow field in a swirl generator by comparing numerical predictions with experimental measurements.

In this article, the fully coupled block algorithm for the solution of three-dimensional incompressible turbulent flows presented in a companion article [1] is extended for use with multiple reference frames and multiple mesh blocks. The implicit block coupling is applied to the extra rotational terms, and to the multiblock interfaces. Furthermore, implementation details on the linearization of cyclic and other boundary conditions are detailed. These modifications allow the coupled solver to retain its improved performance and robustness in addition to mesh size scalability while solving turbomachinery-type applications. The performance and mesh size scalability of the coupled solver is compared to that of a segregated pressure based solver [2] using three industrial-size test cases.

An experimental method is described to trace scalar S/sub 21/-measurements with a precision network analyzer to the basic quantities frequency and power. The achieved measurement uncertainty is 0.11 dB for two standard deviations (95.5% confidence level) at a dynamic range up to 70 dB. Precision reference dipoles have been constructed for use as electric field-strength standards in the frequency range of 30-1000 MHz. The transmission loss between such pairs was calculated by using the numerical electromagnetics code and measured both on an open-area test site and in a fully anechoic chamber. Deviations between computations and measurement of the transmission loss were less than 0.3 dB for any measurement and less than 0.25 dB for the mean values. The expanded uncertainty in the antenna factor of one precision dipole is 0.15 dB for two standard deviations.

There are two established methods for antenna calibration in the frequency range of 30-1000 MHz: The standard site method assumes ideal, calculable wave propagation conditions on an open-area test site with large metal groundplane. The reference antenna method uses a tuned half-wave dipole as a standard with calculable voltage at the feedpoint when exposed to a certain electric field-strength. The aim of our work was to minimize and precisely evaluate uncertainties involved in antenna calibrations using the two methods. In order to establish an accredited antenna calibration laboratory we took the following steps to obtain precision primary standards: We constructed an open-area test site that should exhibit wave-propagation conditions as ideal as possible. We constructed a set of precision, tuned, half-wave dipoles. We calculated site attenuation on an ideal open-area test site for our primary standard dipoles using the NEC-code. We finally measured site attenuation at 10 m distance in horizontal and vertical polarization. In the comparison of measured and calculated data we achieved an agreement of /spl plusmn/0.1 dB between 30 and 80 MHz, /spl plusmn/0.3 dB between 80 and 400 MHz, and /spl plusmn/0.5 dB between 400 and 1000 MHz for horizontal polarization. For vertical polarization agreement was within /spl plusmn/1 dB. We carefully analyzed the residual uncertainty and could dearly identify standing waves on the site by probing the field distribution along longitudinal and transversal axes. This influence can be taken into account by careful selection of antenna positions for the calibrations.

An efficient hydraulic optimization procedure, suitable for industrial use, requires an advanced optimization tool (EASY software), a fast solver (block coupled CFD) and a flexible geometry generation tool. EASY optimization software is a PCA-driven metamodel-assisted Evolutionary Algorithm (MAEA (PCA)) that can be used in both single- (SOO) and multiobjective optimization (MOO) problems. In MAEAs, low cost surrogate evaluation models are used to screen out non-promising individuals during the evolution and exclude them from the expensive, problem specific evaluation, here the solution of Navier-Stokes equations. For additional reduction of the optimization CPU cost, the PCA technique is used to identify dependences among the design variables and to exploit them in order to efficiently drive the application of the evolution operators. To further enhance the hydraulic optimization procedure, a very robust and fast Navier-Stokes solver has been developed. This incompressible CFD solver employs a pressure-based block-coupled approach, solving the governing equations simultaneously. This method, apart from being robust and fast, also provides a big gain in terms of computational cost. In order to optimize the geometry of hydraulic machines, an automatic geometry and mesh generation tool is necessary. The geometry generation tool used in this work is entirely based on b-spline curves and surfaces. In what follows, the components of the tool chain are outlined in some detail and the optimization results of hydraulic machine components are shown in order to demonstrate the performance of the presented optimization procedure.