The discharge coefficient - experimental measurement of a dependence on density contrast.

At sufficiently high Reynolds numbers, the discharge coefficient Cd associated with a sharplycontracting flow through a square-edged opening is, in general, taken to be constant. Theeffect of buoyancy forces, due, for example, to temperature differences between the air oneither side of the opening, is ignored.We hypothesise that the buoyancy force may result in a significant reduction in the dischargecoefficient associated with the flow through a square-edged opening in a horizontal surface.We test this hypothesis by deducing Cd from laboratory measurement of the two-layerstratification produced by a localised buoyancy source in an enclosure which ventilatespassively in a displacement mode. Experiments were performed at small scale in a water tankusing saline solutions to generate buoyancy forces.In the absence of a buoyancy contrast, the discharging flow contracts due to inertial effectsgiving Cd ? 0.6. In the presence of a buoyancy contrast, our results show that the dischargethrough a horizontal vent takes the form of a rising plume-like flow, which may contractfurther giving a smaller value of Cd. The plume-like discharge is characterised by conditionsat the discharge opening, which we represent by a single dimensionless parameter ?d - thedischarge parameter. Our results demonstrate that once a critical value of ?d is exceeded, Cdexhibits a strong dependence on the density contrast and rapidly decreases with increasing ?d.The implications of a dependence of Cd on density contrast are for potentially serious errorsin the prediction of airflow rates if a constant value of Cd is assumed.