Mechanical Engineering - Research Publications
Permanent URI for this collection
Search Results
1 - 10 of 12
-
ItemTurbulence intensity similarity formulations for wall-bounded flows(CIMNE - International Center for Numical Methods in Engineering, 2004)
-
ItemSimilarity predictions based on the attached eddy hypothesis in turbulent boundary layers( 1995)The paper presents a similarity formulation for the streamwise component of the fluctuating velocity u1 in a turbulent boundary layer based on the attached eddy model of wall turbulence being developed at the University of Melbourne by Perry and co-workers. The consequences of this formulation for increasing Reynolds numbers is tested against recent high Reynolds number data. The model is based on the assumption that there exist eddies of different length scales in a turbulent boundary layer and the probability density function (p.d.f.) of the eddy length scale distribution follows an inverse power law for eddies in the turbulent wall region. Such a distribution is necessary to obtain a log-law of the mean velocity: the spectral scaling laws provide indirect evidence for this. In this paper the results from a template matching technique will be presented which gives further support for the proposition of an inverse power law p.d.f. of attached eddy length scales.
-
ItemSimilarity laws and attached eddy shapes in turbulent boundary layers( 1995)This paper presents a similarity formulation for the streamwise component of the fluctuating velocity u1 in a turbulent boundary layer based on the attached eddy model of wall turbulence being developed at the University of Melbourne by Perry and co-workers. The consequences of this formulation for increasing Reynolds numbers is tested against recent high Reynolds number data. The model is based on the assumption that there exist eddies of different length scales in a turbulent boundary layer and the probability density function (p.d.f.) of the eddy length scale distribution follows a -1 power law for eddies in the turbulent wall region. Such a distribution is necessary to obtain the log-law of the mean velocity: the spectral scaling laws provide indirect evidence of this. In this paper the results from a template matching technique will be presented which gives further support for the proposition of a -1 power law p.d.f. of attached eddy length scales.Using space time correlation coefficients further details can be obtained regarding eddy shape. The simple II-shaped representative eddy needs to be modified to give reasonable correlation coefficients.
-
ItemA wall-wake model for turbulent boundary layers with pressure gradients( 1995)The attached eddy hypothesis is considered here for boundary layers with arbitrary streamwise pressure gradients. It is found that in order to obtain the correct quantitative results for all components of the Reynolds stresses, two basic types of eddy structure geometries are required. The first type, called type-A, is interpreted to give a "wall structure" and the second referred to as type-B gives a "wake structure". This is in analogy with the conventional mean velocity formulation of Coles where the velocity is decomposed into a law of the wall and a law of the wake. If the above mean velocity formulation is accepted then in principle, once the eddy geometries are fixed for the two eddy types, all Reynolds stresses and associated spectra contributed from the attached eddies can be computed without any further empirical constants.
-
ItemApplication of the wavelet transform in turbulence( 1997)Traditionally, Fourier transforms have been used to elicit the scale-based behaviour of the turbulent motion and one speaks synonymously of its wavenumber components with scales (large scales are associated with small wavenumbers and vice-versa). Although, this approach is theoretically correct, many workers have questioned its appropriateness on the grounds that a Fourier mode represents a wave like disturbance which is global in the physical domain, whereas an eddy is a disturbance with finite spatial extent. Consequently, a more appropriate scheme should involve a local decomposition of the velocity field which is more reminiscent of eddy like phenomena. In this paper we have explored the feasibility of the wavelet transform as an analyzing tool in deducing the turbulence spectrum.
-
ItemApplication of the attached eddy hypothesis for the evolution of turbulent boundary layers( 1997)The wall-wake attached eddy model of Perry & Marusic [1] is incorporated in a new approach to the classic closure problem for turbulent boundary layers recently proposed by Perry, Marusic & Jones [2]. This involves using the well known mean-flow scaling laws such as Prandtl's law of the wall and the law of the wake of Coles together with the mean continuity and the mean momentum differential and integral equations. The important parameters governing the flow in the general non-equilibrium case are identified and are used for establishing a framework for closure.
-
ItemAnalysis of vortex packets and Reynolds stress in a turbulent boundary layer( 2003)Sets of stereo PIV measurements were obtained instreamwise-spanwise planes of a turbulent boundary layerwith Re¿ = 1060. Two-point spatial correlations obtainedfrom fields in the logarithmic region revealed that bothstreamwise-streamwise (Ruu) and streamwise-wall-normal(Ruw) correlations were significant for streamwise displacementsof more than 1500 wall units. Zero crossing datafor the streamwise fluctuating component u revealed thatstreamwise strips between zero crossings of 1500 wall units orlonger occurred more frequently for negative u than positiveu. This result suggested that the long streamwise correlationsin Ruu are dominated by slower streamwise structures.Additional analysis of Rww correlations suggestedthat the long slow-moving streamwise structures containdiscrete zones of strong upwash over extended streamwisedistances as might occur within packets of angled hairpinvortices. At a wall-normal location outside of the log region(z=± = 0:5), the correlations were shorter in the streamwisedirection and broader in the spanwise direction. Allof the correlation results are consistent with earlier studies(Ganapathisubramani et al, 2003) in which a feature detectionalgorithm identified packets of hairpins in the log regionbut not in the outer region.
-
ItemEvidence of the -1-law in a high Reynolds number turbulent boundary layer( 2004)Dimensional analysis leads to a prediction of a -1-power-lawfor the streamwise velocity spectrum in a turbulent boundarylayer. This law can be derived from overlap arguments or fromphysical arguments based on the attached eddy hypothesis ofTownsend (1976). Some recent experiments have questionedthe existence of this power-law region in wall-bounded ows.In this paper experimental spectra are presented which supportthe existence of the -1-law in a high Reynolds number boundarylayer, measured in the large boundary layer facility in theWalterBasset laboratory at the University of Melbourne. The paperpresents the experimental results and discusses the theoreticaland experimental issues involved in examining the existence ofthe -1-law and reasons why it has proved so elusive.
-
ItemNew evolution equations for turbulent boundary layers in arbitrary pressure gradients( 1997)A new approach at looking at the classic closure problem for turbulent boundary layers is presented. This involves using the well known mean-flow scaling laws such as Prandtl's law of the wall and Coles' law of the wake together with the mean momentum integral and differential equations. The important parameters governing the flow in the general non-equilibrium case are identified and are used to formulate the closure hypothesis. Once the mean flow field has been determined, relevant turbulence quantities can be computed using a coherent structure eddy model based on the attached eddy hypothesis.
-
ItemDual-plane PIV investigation of structural features in a turbulent boundary layer( 2004)Simultaneous dual-plane PIV experiments were performed instreamwise-spanwise planes in the log region of a turbulentboundary layer at a moderate Reynolds number (Ret » 1100).The acquired datasets were used to resolve all 9 velocity gradientsfrom which the complete vorticity vector and other invariantquantities like 3-D swirl strength were computed. Thesederived quantities were used to analyze and interpret the structuralcharacteristics and features of the boundary layer. Resultsof the vorticity vector and the 2-D swirl strength from the twoneighbouring planes indicate the existence of hairpin shapedvortices inclined downstream along the streamwise direction.These vortices envelop low speed zones and generate Reynoldsshear stress that enhances turbulence production. Plots of full3D swirl strength indicate the existence of additional vorticalstructures in the middle of the low speed zones that may representheads of smaller eddies intersecting the measurementplane. This concept is in accordance with the hierarchy of structuresize in a hairpin packet proposed by Adrian et al.[2]. Computationof inclination angles of individual eddies using the vorticityvector suggests that most cores are inclined at 25± to thestreamwise-spanwise plane with a resulting projected eddy inclinationof 32±.