Turbulence of vertical round buoyant jets in a cross flow M. Ben Meftah & A. Petrillo Water Engineering and Chemistry Department, Technical University of Bari, Italy P.A. Davies Civil Engineering Department, University of Dundee, United Kingdom D. Malcangio & M. Mossa Civil and Environmental Engineering Department, Technical University of Bari, Italy ABSTRACT: The paper looks at a vertical round buoyant jet issued in a cross flow and investigates the effect of turbulence in the cross flow upon the structure of the jet. This general problem of the influence of ambient turbulence has been investigated for the case of cross flow turbulence associated with flow over a rough bot- tom boundary. The results of laboratory investigations of this problem are described and it is shown that the influence of the roughness of the bottom surface is manifested by significant modifications to the buoyant jet trajectory (as compared to the smooth boundary reference case). Preliminary attempts to quantify and param- eterize these modifications are presented. 1 INTRODUCTION One crucial problem in hydraulic engineering concerns the dilution of turbulent thermal discharges in a cross flow. The effects of cross flow turbulence (generated, as in this case, by the presence of a rough bottom boundary) upon the structure of the discharge flow is an important aspect of this prob- lem. The study of a heated water discharge in the form of a turbulent buoyant jet into a channel with a flat bottom and with a cross current is well estab- lished in the literature (e.g. Lee, 1984; Il Won Seo et al., 2001; Young Do Kim et al., 2002). The flow dy- namics of the thermal discharge in the near field are governed purely by the momentum source repre- sented by the high velocity injection. In the far field, instead, mixing is governed by ambient currents and stratification. Moreover, processes at the air-water interface, such as heat exchange towards the atmos- phere and wind stress, can further affect the heat dis- tribution. As regards the ambient turbulence, the visualization study of Grass et al. (1991) revealed that powerful vortical structures with a general horseshoe-type configuration occur over smooth sur- faces as well as in the turbulent boundary layer near rough walls and that they are similarly linked to bursting events in both the smooth and rough bed cases. The nature of coherent structures and their processes of development over rough surfaces, therefore, appear to be similar to those originating over smooth walls. Furthermore, Grass et al. (1991) and Grass & Mansour-Tehrani (1996) observed that a remarkable feature of the rough wall flow is its ap- parent ability to order itself very rapidly in a small vertical distance above the tops of the roughness elements. Their measurements indicated that for fully rough wall conditions, the near-wall turbulence structures are directly proportional to the bed rough- ness size for geometrically similar roughness ele- ments and the packing used in their tests. In order to clarify the dependence of the buoyant jet upon the background turbulence and to improve knowledge of the physical processes that determine its dilution, an experimental study was conducted. Localized background turbulence was generated within the ambient flow by means of two different corrugated plastic surfaces of known "wavelength" λ and "amplitude" ε, fixed at the bottom of a rectangu- lar channel. During this experimental work, the characteristics of the turbulence field produced in the surrounding flow by particular combinations of λ and ε were analyzed, with the aim of finding a typi- cal characteristic length scale of the turbulent eddies associated with flow over the rough boundary and its effect upon the structure of the buoyant jet. Since the study consisted of a three-dimensional buoyant jet interacting with a three-dimensional "patch" of shear turbulence, measurements of velocity and tempera- ture fields were assessed at different vertical levels to obtain the spatial structure of the temperature and the velocity profile of the cross flow. Such an ap- proach required a long time for each run, especially because relevant differences were observed at a dis- tance of mms, and accurate and appropriate tech- niques to collect data, as described below were nec- essary. For each configuration, the objectives of the ex- periments were (i) to compare the mixing of a buoy-