Papers by Keyword: RANS

Abstract: Oscillating water column (OWC) wave energy converters can be integrated in harbor protection structures, such as vertical, rubble mound and piled breakwaters. The interaction between the incident wave and the structure, in which the OWC device is integrated, is significantly different, since the structure of the vertical breakwater is impermeable, while that of the rubble mound breakwater is porous. The performance of the OWC device for the three configurations is analyzed for a range of wave periods from 6 to 12 s and a wave height of 1 m. The OWC device integrated into the vertical breakwater shows the best performance (maximum mean pneumatic power of 70 kW), and the mean pneumatic power is globally 3 % higher than that of the OWC device integrated into the rubble mound breakwater (maximum mean pneumatic power of 67.4 kW). The performance of the OWC device integrated into the piled breakwater shows a similar trend to the OWC device integrated into the vertical breakwater for wave periods lower than 9 s, but it has a significant loss of performance for higher wave periods.

Abstract: This paper presents the application of the OpenFOAM^® software package, using the IHFoam/OlaFlow solver, in the simulation of the water waves and air flow in an OWC-WEC located on a breakwater, the Mutriku power station in the north of Spain. The numerical code solves the transient Reynolds averaged Navier-Stokes (RANS) equations, using the Volume of Fluid (VoF) technique to identify the free surface. The standard k-ε turbulence model was used evaluate the Reynold stresses. Two geometries were considered: one with the pneumatic chamber completely open to the atmosphere; and other with the chamber connected to the turbine duct. The solutions obtained by OpenFOAM^® are compared with those obtained by the commercial code Fluent.

Abstract: Process waste heat recovery has been gaining attention in last decades in order to use resourcesefficiently. Even though the mechanisms of heat transfer in static systems such as cross flow heatexchangers are well understood, the situation is totally different for the heat transfer mechanisms insystems consisting of parallel, rotating discs. The situation becomes even more complex when sealingelements prevent fluid from following the natural path in rotational direction. In this work ReynoldsAveraged Navier-Stokes (RANS) simulations using the k-omega-SST turbulence model were performedfor various disc radii, rotational and cross flow velocities as well as disc spacings. Simulations wereexecuted for rotational Reynolds numbers Re! ranging from 94,094 to 435,623 and cross flow Reynoldsnumbers ReU between 3,993 and 36,052. In this paper we show the correlation of resultingflow structures, temperature distribution and resulting heat transfer from multiple rotating discs. Resultsfrom simulations of several different geometrical setups of the discs for various operating pointsshow a maximum of convective heat transfer at a given point that is dependent on the cross flowvelocity, disc diameter, disc spacing and rotational velocity. An increase in air cross flow velocity aswell as larger disc spacing show positive effects on the convective heat transfer in contrast to higherrotational velocity and larger discs. Even though the simulations show some limitations of classicalmeans to increase heat transfer from rotating discs, the influence of sealing elements and resultingflow structures in the gap have not been investigated. It is very likely that the thermal performancecan be further increased by optimizing shape and arrangement of these elements which is subject tofurther research.

Abstract: Reduction of drag and flow resistance in systems containing moving fluids is a prominenttool to increase energy efficiency. Besides active flow control – such as moving surfaces or boundarylayer suction – passive techniques such as surface patterning by means of dimples are promising sinceno additional energy consumer is introduced into the system. Even though the effect of drag reductiondue to dimples has often been observed, the physical principals responsible for this effect are not yetunderstood. Most of the research concerning dimples and drag reduction published so far has beencarried out experimentally, not many numerical investigations on this topic have been done. The mainreason for this is that the tiny, transient flow structures generated in direct vicinity of dimples can noteasily be resolved in simulations. Even in case of time dependent numerical investigations it is notclear, whether and with which method of sub-grid scale modeling Large Eddy Simulations are capableof modeling these structures sufficiently. In this work we investigated different surfaces with dimpledepth to diameter ratios h/D reaching from 0.01 to 0.1 in channels of height H = 0.417D at Reynoldsnumbers ReD 5 830 and ReD 11 650 using steady state simulations with a k-omega-SST turbulencemodel. Drag reductions were observed for all setups h/D < 0.08 compared to the smooth channel.The best results were obtained with dimple depths of 4-5 % of diameter showing a slight dependenceof Re which is in good agreement with literature. As the experimental investigation of the flow overdimpled surfaces is limited in spacial and temporal resolution we could demonstrate that numericalinvestigations give the possibility to overcome this drawback. However the solution of simulationsstrongly depends on numerous factors such as the discretization scheme, the numerical models andthe grid used to obtain results which might be a reason for slightly varying results of such simulationsfound in literature.

Abstract: The motion of the free-surface inside a surface-piercing vertical cylindrical tube can be seen as a simplified approximation of an oscillating-water-column ocean-wave energy-converter (OWC-OWEC). In the present work the IHFOAM code, which solves the Reynolds-averaged Navier-Stokes (RANS) equations by a finite-volume method using the volume-of-fluid (VoF) technique, was used to simulate the action of a regular wave on the free-surface inside a vertical cylinder open to the atmosphere. In this paper the results obtained by IHFOAM are compared with other numerical results and with experimental data, showing a good correspondence between these results. In this way the IHFOAM code was verified (by comparison with other numerical results) and validated (by comparison with experimental data).

Abstract: Conventional ventilation systems with heat recovery used for building aeration exhibit characteristic disadvantages arising from their operating principle such as noise generation from bladed ventilators or remarkable pressure losses generated by heat exchangers. A novel concept that combines ventilators and heat exchanger in one compact friction ventilator that rotates in two separated ducts producing two opposed airflows and transferring thermal energy from the higher temperature airflow to the lower temperature level can overcome the mentioned shortcomings. In order to demonstrate the feasibility of a friction ventilator to operate as ventilation system with heat recovery computational fluid dynamics were used to analyze the resulting pressure jump and volume flow for different geometrical setups. An extensive grid dependency study for a defined operating point that represents the typical use has been carried out in order to improve the numerical results. Furthermore, the results were compared to experimental data whenever possible.

Abstract: The article deals with results of the implementation of the k-k_L-ω turbulence model for compressible transitional flow into OpenFOAM. This model was firstly proposed by Walters and Leylek [2] and utilizes the approach of the laminar kinetic energy in order to predict the transition between laminar and turbulent flows. The performance of the implemented model has been tested for the case of flow over a flat plate and the flow through VKI and SE 1050 turbine cascades. The properties of the implementation of the model for compressible flow simulations into OpenFOAM are discussed.

Abstract: For the needs of high-performance steam turbines producer the data of a blade section measurement have been analyzed in detail using an experimental and numerical approach. The blade section is used on prismatic blades in high and medium pressure steam turbine parts. The linear blade cascade was tested at four pitch/chord ratios at two different stagger angles. The blade cascade was tested under two levels of Reynolds number in the range of output izentropic Mach numbers from 0.4 to 0.9.The inlet of the test section was measured pitch-wise by five-hole probe to determine the inlet flow angle. The free stream turbulence of inlet flow was determined at 2.5% what is very close to the operating conditions on first high pressure stages. Two-dimensional flow field at the center of the blades was traversed pitch-wise downstream the cascade by means of a five-hole needle pressure probe to find out the overall integral characteristics. The blade loading was measured throughout surface pressure taps at the blade center. An in-house code based on a system of Favre-averaged Navier-Stokes equation closed by non-linear two-equation EARSM k-ω turbulence model was adopted for the predictions. The code utilizes an algebraic model of bypass transition valid for both attached and separated flows taking into account the effect of free-stream turbulence and pressure gradient. Results are presented by integral characteristic in means of kinetic energy loss coefficient and velocity or pressure distribution in the blade wakes or on the blade surface. In this article, the effect of investigated criteria and comparison of experimental and numerical approach are presented and discussed.

Abstract: The paper presents RANS numerical simulations carried out to calibrate the Pitot tube velocity measurements in order to extend the accurate measurement range to high velocities, up to 250 m/s. Three calibration methods are proposed, able to reduce the rms error of the velocity measurement from an initial value of 6.5 m/s, to 3.11 m/s, 2.33 m/s, and 0.08 m/s, respectively. As the accuracy of the calibration method increases, the complexity and the portability of the method for other flows and other Pitot tube dimensions is found to decrease.

Abstract: Film cooling has been extensively used to provide thermal protection for the external surfaces of gas turbine components. For the past 40 years, numerous number of film cooling hole designs and arrangements have been introduced. Due to broad designs and arrangements of film cooling, numerical investigation has been utilized to provide initial insight on the aerodynamics and thermal performance of the new film cooling designs or arrangements. The present work focuses on the numerical investigation of RANS and URANS analyses on a flat plate film cooling. The investigation aims to provide comparison between various turbulent models available for the Reynolds Average Navier Stokes (RANS) analyses and extended to unsteady Reynolds Average Navier Stokes (URANS). The numerical investigations make used of ANSYS CFX ver. 14 and were carried out at Reynolds Number, Re = 7,000 based on the hole diameter at blowing ratio, BR = 0.5. The results of the RANS analyses show significant influence of the turbulent models on the predicted aerodynamics and thermal performance of the film cooling. The result of URANS indicates limitation of RANS analyses to provide details on the eddied and vortices formation in film cooling flow structure.