Abstract
Fish passing through hydroturbines are subjected to a variety of conditions that can injure or kill them. Hydropower owners and operators need to understand injury mechanisms to identify designs and operational criteria to increase the survival of fish passing through dams. Of the various mechanisms believed to be responsible for injuries to fish during turbine passage, strike by turbine runner blades is one of the most predominant.
This study is the initial stage of further investigation into the dynamics of injury to fish during passage through a turbine runner. As part of the study, Pacific Northwest National Laboratory (PNNL) estimated the probability of blade strike, and associated injury, as a function of fish length and turbine operating geometry at two adjacent turbines in Powerhouse 1 of Bonneville Dam. Units 5 and 6 had identical intakes, stay vanes, wicket gates, and draft tubes, but Unit 6 had a new runner and curved discharge ring to minimize gaps between the runner hub and blades and between the blade tips and discharge ring. We used a mathematical model to predict blade strike associated with two Kaplan turbines and compared results with empirical data from biological tests conducted in 1999 and 2000. Blade-strike models take into consideration the geometry of the turbine blades and discharges as well as fish length, orientation, and distribution along the runner.
Previously, we integrated historical and recent test results for turbine passage in conjunction with published theory for turbine runner blade strike to examine the relationship among fish size, turbine operation, and injury to fish during turbine passage. Use of Monte Carlo techniques allowed consideration of factors such as fish orientation at entry to the runner, distribution of fish along runner blades, and important sources of error and variability into the strike probability estimation. Blade strike was considered for analysis because it is an obvious source of injury, and injury mechanisms have been mathematically and empirically studied.
The first phase of this study included a sensitivity analysis to consider the effects of difference in geometry and operations between families of turbines on the strike probability response surface. The analysis revealed that the orientation of fish relative to the leading edge of a runner blade and the location that fish pass along the blade between the hub and blade tip are critical uncertainties in blade-strike models.
Blade-strike models predicted that injury increases with decreasing discharge and with increasing fish-passage radius. Over a range of discharges, the average prediction of injury from blade strike was two to five times higher than average empirical estimates of visible injury from shear and mechanical devices. However, differences in relative blade velocities may explain why predictions of injury were within 2-3% of empirical 48-h mortality estimates for fish passing at mid-blade and tip locations but were 6-8% higher than 48-h mortality estimates for fish passing near the hub. Mortal injuries are more likely to result when relative blade velocities are high like those at mid-blade and tip locations. Empirical estimates of mortality may be better metrics for comparison to predicted injury rates than other injury measures for fish passing at mid-blade and blade-tip locations.