Authors: J. Groeneveld, C. Sweeney, C. Mannheim, C. Simonsen, S. Fry, and K. Moen
Waterpower XV in Chattanooga, Tennessee, July 23-26, 2007
Abstract
Dissolved gas at supersaturated concentrations has been identified as a water quality issue
affecting aquatic organisms. The principal source of supersaturation at dams is gas transfer in
the highly aerated flood discharges from spillways. Avista Utilities has determined that when
operating, the spillway of the Cabinet Gorge Dam produces high concentrations of dissolved
gas in the downstream river. Avista is therefore looking to reconstruct the original river bypass
tunnels to reduce the use of the spillway when the hydraulic capacity of the powerhouse is
exceeded. The bypass tunnel design is intended to reduce downstream dissolved gas
concentration by avoiding use of the spillway and air entrainment in the tunnel discharge.
Avista chose to use physical and numerical modeling efforts to address the risks and issues
associated with this highly complex project. ENSR constructed a 1:50 scale physical model to
define the hydraulic performance of an initial tunnel design concept, support design
development of the tunnels, and to confirm the hydraulic performance of the final concept. Hatch Energy used Flow-3D, a commercially available computational fluid dynamics (CFD)
computer program, to evaluate design alternatives prior to testing final concepts in the physical
model.
In this paper, a comparison of intake pressures measured in the physical model and computed
in the CFD model is presented. The location and magnitude of low pressures in the intake are of
particular importance to avoid cavitation and to verify the actual discharge capacity of the
tunnel. The CFD model was used to identify low pressure areas in the intake. Piezometric taps
were installed in the physical model at the corresponding locations and connected to calibrated
pressure transducers. The physical model and CFD model results showed good agreement of
minimum pressure locations and reasonably good agreement of average pressure magnitudes,
but not the dynamic fluctuations of pressure. Determination of the tunnel discharge capacity
required careful interpretation of both sets of model results due to the potential for cavitation to
occur in the tunnel intakes.
