Semi-empirical assessment of plunge pool scour: two-dimensional application of Annandale's Erodibility Index Method on four dams in British Columbia, Canada, A

Abstract

Rock scour downstream of dam foundations and spillways has become a significant dam safety concern in recent years. As design flood estimates increase and older infrastructure is expected to pass larger amounts of water, downstream river beds and plunge pools are subjected to progressively greater stream power from rapidly flowing water. A need exists to quantify the erosive capacity of the flowing water and the erodibility of earth materials to evaluate potential scour in these susceptible areas. Annandale's Erodibility Index Method, widely considered a state of the art scour prediction method, offers an approach to quantify scour depth by comparing the erosive capacity of flowing water and the ability of rock to resist it. This study assesses the accuracy of Annandale's Erodibility Index Method for estimating rock scour depth in plunge pools. The success by which the method may be implemented is dependent on the accuracy of methods to quantify the rate of energy dissipation of plunging jets (applied stream power) and the ability to estimate the capacity of rock to resist the power of the flowing water. The stream power of plunging jets is quantified by making use of published research, while the ability of rock to resist scour is computed using a geo-mechanical index, known as the Erodibility Index. The Erodibility Index that was used to estimate the scour resistance of the various stratigraphic layers downstream of four BC Hydro dam spillways located in British Columbia, Canada relies on in-situ rock parameters consisting of UCS strength values, RQD values, joint spacing, aperture, alteration, roughness, and orientation. The jet stream power was calculated using continuous daily discharge records and spillway geometries at each of the dams, and published research on stream power quantification. The spillway types included one long spillway chute with a free overfall and a number of flip bucket-type energy dissipaters. Comparison between the numerically generated scour profiles and a series of plunge pool surveys at each of the dams provided a means of determining accuracy. Scour depths and the distances between the end of the spillways and the points of maximum scour were matched. The study revealed that correlations between calculated and observed scour profiles improved with the quality of geologic information and with the certainty by which the stream power of jets and their decay could be quantified. The geologic information at two of the dams, Revelstoke and Seven Mile Dams, was incomplete and resulted in a generalized characterization of the scour resistance of the plunge pool rock. At these dams it was not possible to spatially characterize changes in scour resistance of the rock in the plunge pool. The geologic information at Peace Canyon and W.A.C. Bennett Dams was more informative and allowed quantification of the spatial distribution of plunge pool scour resistance in each case. The research further identified that jet theory associated with flip buckets provides good estimates of stream power and its decay, but that hydraulic theory used to quantify the energy dissipation in long spillway chutes may be incomplete at this point in time. The combined inadequacy of geologic data and the insufficiency of hydraulic theory related to long spillway chutes resulted in the comparison between surveyed and calculated scour depths at Revelstoke Dam being particularly poor. Contrasting this result, it was found that better understanding of flip bucket hydraulics and more informative geologic information resulted in very good correlations between calculated and surveyed scour profiles at Peace Canyon and W.A.C. Bennett Dams. The comparison provided more pleasing results at Seven Mile Dam than at Revelstoke Dam, in spite of the relatively poor geologic information in the plunge pool area of Seven Mile Dam. The improved correlation at Seven Mile Dam may be attributed to the fact that it has a flip bucket spillway, which resulted in better quantification of jet stream power and its decay. In summary the analysis results indicate a strong correlation between surveyed and modeled plunge pool depth and invert location at dam sites where both the hydraulics and geology are well understood. It is concluded from this research that Annandale's Erodibility Index Method is an accurate method for estimating plunge pool depth when geologic information is available and spillway and plunge pool hydraulics are well-understood. A lack of information or understanding of the spatial distribution of materials generally results in less accurate predictions. Although this research project was not directed towards determining the rate of scour of rock, a good correlation was found between cumulative energy (the product of stream power and duration) and scour depth. This preliminary result provides encouragement for future research into using the Erodibility Index Method to quantify the rate of scour in rock. The analysis would benefit from more robust rock parameter datasets that would allow for the inclusion of a Monte Carlo simulation within the model. Additionally, a larger dataset of dams including those with various spillway structure types and geologic environments would be valuable moving forward.