Page - 833 - in Book of Full Papers - Symposium Hydro Engineering

effect of those parameters with small-scale model tests. The employed scale model exceeds most previous model studies in terms of energy head HE and water discharge Qw. Nevertheless, scale and model effects still need to be addressed carefully. Detailed prototype data are crucial for the validation and upscaling of model results. Therefore, this project additionally includes air demand measurements at three bottom and one middle outlet(s) in southern Switzerland. 2. SCALE MODEL TESTS 2.1. MODEL SETUP AND INSTRUMENTATION A physical bottom outlet model was built at the Laboratory of Hydraulics, Hydrology and Glaciology (VAW) at ETH Zurich (Fig. 1a). Two pumps provide a maximum energy head at the gate of HE = 30 m w.c. at a water discharge of Qw ≈ 600 l/s. The rectangular sharp-crested gate is 0.2 m wide, 0.25 m high and the downstream tunnel extends to a height of ht = 0.3 m (Fig. 1b). The horizontal BO tunnel has a maximum length of L = 20.6 m which can be varied due to detachable elements. The aeration chamber is connected to a circular air vent of diameter d = 0.15 m which can be throttled with an orifice plate to vary the loss coefficient ζ of the whole aeration system. A similar air vent is located at the tunnel end to measure the airflow into or out of the tunnel. All channel walls consist of PVC or acrylic glass with a hydraulic roughness of k = 0.003 mm. Qw is measured with an inductive flow meter with an accuracy of ±0.5% of the measurement value (MV) and ±0.4 l/s absolute error. The air flow velocity in the air vent Ua,o is measured with a thermal anemometer while the air flow velocity at the tunnel end Ua,u is measured with a bidirectional vane anemometer. The thermal anemometer has an accuracy of ±2.5% of MV and the vane anemometer has an accuracy of ±1.5% of MV ±0.2 m/s. From the measured velocity, the air discharges Qa,o and Qa,u are computed assuming a logarithmic velocity profile. The cross section above the air-water mixture is blocked with a gate at the tunnel end. Consequently the air has to be supplied through the second air vent which allows for a precise measurement of Qa,u. A total of 40 relative pressure sensors with a measurement range of ±100 mbar and an accuracy of ±1 mbar are installed at the invert and the soffit along the tunnel centerline to measure both the water (subscript w) and air pressures (subscript a), respectively. The following parameters were varied in the model tests: The relative gate opening a/amax was increased from 0.1 to 1 in steps of 0.1. For each a/amax, HE was varied from 5 to 30 m w.c. in steps of 5 m w.c. and six ζ-values were tested, ζ = 0.7, 2.7, 9, 19, 28 and 57. All parameter combinations were tested for tunnel lengths of L = 20.6, 12.6 and 6.6 m. Combinations of HE > 20 m w.c. and ζ > 10 could not be measured due to strong flow pulsations. Large relative gate openings of a/amax > 0.8 led to a drowning of the aeration chamber and were thus excluded from the measurements. 833