UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A study of scour below Ruskin Dam spillway using a non-cohesive bed hydraulic model Galvagno, Giampiero 1998

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1998-0156.pdf [ 13.88MB ]
JSON: 831-1.0050300.json
JSON-LD: 831-1.0050300-ld.json
RDF/XML (Pretty): 831-1.0050300-rdf.xml
RDF/JSON: 831-1.0050300-rdf.json
Turtle: 831-1.0050300-turtle.txt
N-Triples: 831-1.0050300-rdf-ntriples.txt
Original Record: 831-1.0050300-source.json
Full Text

Full Text

A STUDY OF SCOUR BELOW RUSKIN DAM SPILLWAY USING A NON-COHESIVE BED HYDRAULIC MODEL by Giampiero Galvagno B.A. (Philosophy), The University of British Columbia, 1990 B.A.Sc. (Civil Engineering), The University of British Columbia, 1995 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE in THE FACULTY OF GRADUATE STUDIES CIVIL ENGINEERING We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA April, 1998 © Giampiero Galvagno, 1998 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of The University of British Columbia Vancouver, Canada DE-6 (2/88) Abstract Hydraulic models, using either cohesive or non-cohesive beds, have traditionally been employed to gauge scour below a spillway. Empirical equations, based on world-wide prototype and model experience, have been developed for estimating scour depth depth. In this study both the hydraulic modelling and analytical approach were used to assess the maximum potential scour in the rock lined channel below Ruskin Dam. Ruskin Dam, built in 1928, has an overflow spillway with seven bays discharging into a rock-lined channel. The spillway has a complex arrangement of concrete benches lining the sides and a large concrete bridge pier in directly downstream of the spillway. A non-cohesive bed, hydraulic model was employed to understand how the degree of scour is affected by the presence of the bridge pier and redirection of the spillway discharge by the concrete benches, and to locate areas of maximum scour intensity. Five scour equations were used to determine theoretical scour depths in both the prototype and model. Data on scour depths gathered from the model study was compared to the theoretical values calculated from the equations. From the model study it was shown that if the bridge pier was removed from the spillway channel the point of maximum scour moves downstream and the intensity is reduced. The location of the point of maximum scour was also influenced by gate openings. High scour intensities were recorded for gate openings where flow was deflected to one side of the channel and localized by the pier armature. The results of the scour depth comparisons using theoretical and measured model depths showed that 2 of the 5 equations reasonably represented the model scour. This result was generalized to the scour calculations using prototype data and an estimate of scour depth was made for the rock lined spillway channel below Ruskin Dam. ii Table of Contents Abstract ii List of Tables v List of Figures vi Acknowledgments vii CHAPTER 1. INTRODUCTION.. 1 CHAPTER 2. LITERATURE REVIEW 5 2.1 BEHAVIOUR OF A FALLING JET IN THE PLUNGE POOL 5 2.2 TURBULENT ENERGY DISSIPATION IN THE PLUNGE POOL : 7 2.3 MODELLING SPILLWAY SCOUR 11 CHAPTER 3. THE RUSKIN PROJECT 13 3.1 T H E GENERAL ARRANGEMENT OF THE SITE 13 3.2 RUSKIN D A M SPILLWAY 15 3.3 GEOLOGY OF THE SITE 16 3.4 SPILLWAY OPERATION 17 3.5 FLOOD ROUTING 17 3.6 OVERVIEW OF OBSERVED SCOUR TO DATE 18 3.7 PREVIOUS SCOUR STUDY TO PREDICT M A X I M U M SCOUR DEPTH 19 3.7.1 Scour Study Approach 20 3.7.2 Results of Scour Calculations 22 3.7.3 Conclusion of Analytical Scour Study 26 CHAPTER 4. RUSKIN DAM SPILLWAY MODEL AND STUDY APPROACH 35 4.1 DESCRIPTION OF THE MODEL 37 4.1.1 Operation of the Model 41 4.1.2 Calibration of the Model 43 4.1.3 Results Of The Model Test : 45 4.2 OBSERVATIONS OF FLOW PATTERNS AND THE RESULTING SCOUR 47 4.2.1 Return Currents 47 4.2.2 The Roller in the Plunge Pool 48 4.2.3 The Rooster Tail Effect 49 4.2.4 The Effect of Different Spill Cases the Water Level 50 4.3 RESULTS OF SCOUR DEPTH MEASUREMENTS 51 4.4 COMPARISON OF PREDICTED AND MEASURED DEPTHS 54 4.5 APPLICATION OF MODEL RESULTS TO ANALYTICAL STUDY 58 4.6 SUMMARY OF RESULTS 59 iii CHAPTER 5. CONCLUSIONS AND RECOMMENDATIONS WORKS CITED & EXTENDED BIBLIOGRAPHY iv List of Tables Table 3.1: Pertinent Elevations 17 Table 3.2: Spillway Flood Discharge Annual Excedence Probabilities 18 Table 3.3: Summary of Calculated Maximum Scour Depths 25 Table 4.1: Summary of Model Tests 46 Table 4.2: Results of Measured and Calculated Scour Depths 81 v LIST OF FIGURES Figure 3.1: General Arrangement of the Ruskin Project 27 Figure 3.2: Ruskin Spillway and Concrete Benches 28 Figure 3.3: Extension of Concrete Benches 29 Figure 3.4: Concrete Bridge Pier in the Prototype Spillway 30 Figure 3.5: Detail of 1935/36 Survey of Spillway Channel 31 Figure 3.6: 1989 Sounding Survey of Spillway Channel Showing Orientation of Lines 32 Figure 3.7: 1989 Sounding Survey of Spillway Channel 33 Figure 3.8: Detail of Historical Scour in Side Channel in the Region of the Bridge Pier 34 Figure 4.1: Plot of Measured Maximum Scour Depths 47 Figure 4.2: Scour Depth Calculate Using Damle's, Veronese, and Calibrated Mason's Eq. 56 Figure 4.3: Scour Depth Calculate Using Coleman's and Mason's Eq 57 Figure 4.4: Rating Curve for Prototype Spillway 62 Figure 4.5: Ruskin Dam Spillway Hydraulic Model - General Layout 63 Figure 4.6: Construction Detail of Spillway Face 64 Figure 4.7: View of Hydraulic Model 65 Figure 4.8: Prototype Spillway Discharge from Gate 66 Figure 4.9: Prototype Spillway Discharge Through Gates 6 & 7 67 Figure 4.10: Prototype Spillway Discharge Through Gates 1 & 2 68 Figure 4.11: Plan View of Model Spillway Benches and Apron 69 Figure 4.12: Plan View of Model Spillway Discharge Through Gates 4, 5, & 6 69 Figure 4.13: Model Spillway Discharge, Gates 2 to 6 Open 70 Figure 4.14: Model Spill Discharge, Gates 2 to 7 Open 70 Figure 4.15: Test #1 Scour Patterns (Bays 4, 5, & 6; Q=800 m3/s) 71 Figure 4.16: Test #2 Scour Patterns (Bays 4, 5, & 6; Q=980 m3/s) 71 Figure 4.17: Test #3 Scour Patterns (Bays 4, 5, & 6; Q=650 m3/s) 72 Figure 4.18: Test #4 Scour Patterns (Bays 3, 4, & 5; Q=800 m3/s) 72 Figure 4.19: Test #5 Scour Patterns (Bays 3, 4, 5, & 6; Q=1400 rrrVs) 73 Figure 4.20: Test #6 Scour Patterns (Bays 3, 4, 5, 6 & 7; Q=1950 m3/s) 73 Figure 4.21: Test #7 Scour Patterns (Bays 2, 3, 4, 5, 6 & 7; Q=2400 m3/s) ....74 Figure 4.22: Test #8 Scour Patterns (Bays 1, 2, 3, 4, 5, 6 & 7; Q=2700 m3/s) 74 Figure 4.23: Test #9 Scour Patterns (Bays 1,2, 3, 4, 5, 6 & 7, pier removed; Q=2950 m3/s) 75 Figure 4.24: Test #10 Scour Patterns (Bays 2, 3, 4, 5, 6 & 7, pier removed; Q=2600 m3/s) 75 Figure 4.25: Test #11 Scour Patterns (Bays 2, 3, 4, 5 & 6, pier removed; Q=2400 rrfVs) 76 Figure 4.26: Test #12 Scour Patterns (Bays 2, 3, 4, 5 & 6; Q=2300 m3/s) 76 Figure 4.27: Test #13 Scour Patterns (Bays 3, 4, 5 & 6; Q=1850 m3/s) 77 Figure 4.28: Test #14 Scour Patterns (Bays 2, 3, 4, 5 & 6; Q=1950 rrfVs) 77 Figure 4.29: Test #15 Scour Patterns (Bays 1, 2, 3, 6 & 7; Q=2000 m3/s) 78 Figure 4.30: Test #16 Scour Patterns (Bays 4, 5 & 6; Q=1350 m3/s) 78 Figure 4.31: Test #17 Scour Patterns (Bays 2, 3, 4, 5, 6 & 7; Q=2750 m3/s) 79 Figure 4.32: Test #18 Scour Patterns (Bays 1, 2, 3, 4, 5, 6 & 7; Q=3600 m3/s) 79 Figure 4.33: Test #19 Scour Patterns (Bays 1, 2, 3, 4, 5, 6 & 7; Q=4200 m3/s) 80 Figure 4.34: Test #20 Scour Patterns (Bays 1, 2, 3, 4, 5, 6 & 7, pier removed; Q=4100 m3/s) 80 vi Acknowledgments I would like to acknowledge the contribution and support of B C Hydro and its Partnership Program without which this thesis could not have been possible. I would especially like to thank those individuals at B C Hydro who contributed directly and indirectly to the outcome of this work: to Dave Cattanach for his patience and trust in me; to Warren Bell whose valuable feedback gave shape to much of the analysis; to John Taylor whose encouragement and ideas set this thesis in motion; to Kathy Groves for her generosity of time and excellent guidance during my work terms at B C Hydro; to my colleagues in the program, and lunch budsCharissa Dharmatessia and Milton Siu; and to the many other people who in the short time I was at B C Hydro I had the privilege to work with. I would also like to thank Dr. Dennis Russell, Dr. Greg Lawrence, and Dr. Michael Quick at U B C who together showed me the magic of Hydrotechnical Engineering and inspired me to pursue the discipline as a career. Most importantly I would like to acknowledge the crucial role thatKerrin Spurr played in the evolution of this thesis: it was his perspective and experience with plunge pool scour that helped me to define the hydraulic model and the scope of this thesis. I would also like to thank all those people who, collectively, have helped me to arrive at this milestone: to Susan Rundle for the love and companionship she has given me; to my family for all their support through the years; to Dennis & Velma, M & M , Eric Berglund, John Blackburn, Greg Garland, and Rob & Suzanne Lionello for their friendship; to the people of Kaleden for giving me a sense of identity; to John Dixon for introducing me to the ideas of history; to Bob Dylan just 'cause your music's so fine; and to F. Doestoyevsky, G.G. Marquez, A. Camus, S. Clements, J. Conrad, B. Bettleheim, B. Russell, Descartes, and many others, for leaving behind such amazing literature and glimpses of truth to guide us in our search for meaning. Chapter 1 Introduction The prediction of scour depth and extent below spillways has been the subject of many papers. The very complex nature of the fluid dynamics in a plunge pool has made any meaningful analytical analysis nearly impossible. Instead empirical equations, based on world-wide prototype and model experience have been developed to estimate scour depth. Even the best of these equations, though, can overestimate scour depth by as much as 100% (Mason, 1985). Another limitation is that the location of the maximum scour depth is not predicted by the scour depth equations. Hydraulic modelling of the spillway and bed is a second approach which produces more detailed and site specific information on the extent and location of scour. For many spillways the location of the point of maximum scour can be predicted analytically. For a flip-bucket or ski jump energy dissipator the maximum scour w i l l occur at the point of jet impact ( I C O L D , 1987). The trajectory of the jet issuing from these types of spillways can be predicted given basic Newtonian physics and design curves (Mason, 1993). For other stilling basins, the point of maximum scour wi l l occur in the vicinity of the hydraulic jump or roller (Henderson, pp.221-227). There are spillway types, though, where the point of jet impact or location of the roller changes for different spill conditions. The spillway at Ruskin dam exemplifies a energy dissipator where the location of maximum scour cannot be predicted analytically with any reasonable degree of certainty. Ruskin Dam is a concrete gravity dam with an overflow spillway with seven bays discharging into a rock-lined spillway channel. A n estimate of maximum potential scour at the site is hampered by the presence of a complex arrangement of concrete benches lining the 1 spillway, and a large concrete bridge pier located immediately downstream of the spillway. Both the concrete benches and bridge pier interact with the spillway discharge to different degrees, depending on which of the seven discharge gates are open. The result of the interaction of the discharge with these boundary conditions is a highly complex flow pattern in the plunge pool, and elements of both a ski-jump and roller-type energy dissipator. When the central bays are open a submerged jump or roller develops at the foot of the spillway, above a short, horizontal concrete apron. During high spill events, when the outside gates are utilized, discharge is deflected by concrete benches lining both sides of the spillway resulting in one or more detached jets entering the channel. At maximum discharge, when all seven gates are open, both the roller and detached jets are dominant features of the flow pattern. In an analysis of maximum scour depth using empirically derived equations, there is a lowered level of confidence in the results due to the uncertainty in the way the spill discharge. interacts with these structures, and the influence of the geology. An additional level of understanding is required which can only be achieved by modelling the spillway discharge and the resulting scour. Information from a scour model study can be used in conjunction with results of scour depth calculations to achieve greater confidence in the understanding of the location and extent of scour under different spill conditions. Scale models have been used for more than 100 years to study river hydraulics, beginning with Osborn Reynolds' modelling of River Mersey in 1885 (Henderson, p.498). The refinement of techniques in this area has led to models for understanding scour in rock beds below spillways. In these scour models the rock bed of a site can be modeled by gravel, using a specific design gradation. Gravel in regions of high scour intensity along the non-cohesive beds is transported downstream. In past comparisons between non-cohesive bed scour models and their rock bed prototypes, the extent and location of the resulting scour hole 2 shows good agreement. The shape of the scour hole is similar except for the side slopes which, due to the granular material of the bed, w i l l have flatter side slopes. Unfortunately, the scour depth in non-cohesive scale bed models has been shown to be exaggerated and information of this kind from the model must be used cautiously. In the model study described in this thesis, information on the extent of scour below the Ruskin dam spillway was gauged using a classical non-cohesive bed approach. Scour conditions were studied for two possible scenarios: the first was for a P M F event where the spill discharge approaches 4500 cms; the second was for the same spill discharge magnitude but with the assumption that the concrete bridge pier becomes undermined and washed out. The removal of the bridge pier changes the hydraulic characteristics of the channel in an important way. Taking into account the concrete pier, the cross-sectional area of the channel is smallest at this point and was thought to represent a channel control at high discharges. The result of a control would be an increase in water depth upstream of the constriction at high discharges. The sudden removal of this hydraulic control, as in the case of the pier being washed out during ah extreme flood event, may intensify scour by reducing the water level in the plunge pool. A 1:50 scale model of Ruskin Dam was constructed in the U B C Hydraulics laboratory to assess scour in the plunge pool. A non-cohesive bed model was employed to locate regions of maximum scour. The location and extent of scour in the plunge pool was documented for various discharge magnitudes and gate operation spill scenarios. Information from the model study was combined with the results of an analytical study using 5 scour depth equations to arrive at a reasonable estimate of the maximum scour depth for Ruskin Dam. A review of various articles pertaining to scour is made in Chapter 2. A description of the site, including a summary of relevant data for the study is presented in Chapter 3. The 3 results of a scour analysis which makes use of various equations to estimate maximum scour potential is documented and included in Chapter 3. A description and the results of the model study are given in Chapter 4 . Chapter 5 summarizes the conclusions and recommendations of the study. To facilitate describing the site and the results of the scour studies, a standard convention of reference is used. The channel is always referenced from the perspective of someone looking downstream from the dam; thus, the right and left banks are those viewed looking downstream from the dam crest. 4 Chapter 2 Literature Review Scouring of the rock bed downstream of a spillway is caused by the water pressure or total water load at a particular point. The three components which make up the total water load are the mean dynamic pressures caused by the impact of the jet on the rock surface; the truly mean dynamic time dependent water pressures (pressure fluctuations); and the Reynolds shear stresses (Otto, 1986). The point of maximum scour depth in a plunge pool typically coincides with the point of momentum impact of the jet ( ICOLD, 1987). The influence of the shear stresses along the boundaries, as well as the presence of pressure fluctuations wi l l contribute to the depth, size and shape of the scour hole. Much research has been carried out with the aim of further understanding the impact behaviour of water jets on underwater surfaces (Doddiah, 1953; Cola, 1965; Beltaos, 1977; Aderibigbe, 1996). A significant body of work also exists which is directed at trying to quantify by empirical equations the depth of scour given particular discharge and/or geological parameters; (Damle, 1966; Mason, 1984, 1985, 1988; Spurr, 1985; Akhmedov, 1988; Yildiz, 1994). In recent work, a greater emphasis has been placed on the importance of pressure fluctuations in analyzing the scouring process (Ervine etal, 1997; Annandale, 1995) 2.1 Behaviour of a Falling Jet in the Plunge Pool There is difficulty in applying research on underwater jets to the understanding of plunging turbulent jets, as Ervine and Falvey (1987) point out, because plunging jets are a more complex phenomenon. Turbulent eddies, expressed as waves, on the surface of a plunging jet entrain air. When the jet enters the plunge pool it is composed of an expanding, aerated outer core and a decaying solid, inner core of water. Because of the surface waves on 5 the jet and the irregular outer region, the plunging jet which enters the pool is not clearly defined. Plunging aerated jets will produce surface waves in the pool rather than well-defined penetrating shear layers, as with submerged jets (ibid.). There are three important features of a falling jet. The first is its tendency to spread laterally during its fall; the second is its ability to become increasingly distorted; and the third, is its propensity to break-up if the plunging length is sufficiently long. All three of these features affect the degree to which the jet entrains air and the impact pressures in the plunge pool, and thus, are important for understanding plunge pool scour. In general, disintegrated, aerated jets produce smaller mean pressures than intact, non-aerated jets (Ervine et. al., 1987). The three features can be understood in terms of turbulence that occurs within the jet and at the water-air interface (Falvey, 1988). The falling jet begins to spread laterally as air is entrained and droplets of water are produced. The falling jet becomes increasingly distorted because the eddies, which form riffles on the surface of the jet, become increasingly larger in period and amplitude. As aeration progresses, the outer region increases laterally with a proportional decrease in the width of the inner core. Eventually, given a sufficient fall-length, the inner core decays completely, after which the jet begins to break-up (ibid.) Turbulence is the most significant air-entraining mechanism of a plunging jet (McKeogh and Elsaw,1980; Ervine et. al., 1987). Turbulence is a factor which can vary significantly between outlet arrangements (Mason, 1989). One of the features of a falling jet is that the velocity gradient within the jet becomes negligible, except near the surface, within a few jet diameters of the outlet. Since the degree of turbulence intensity depends on the velocity gradient, there is little turbulence generated within the core of the falling jet; rather, 6 the potential for turbulence occurs at the outer edges of the jet where the velocity gradient is generally much steeper (Falvey, 1988). Ervine (1976) related the volumetric air to water ratio ((3) in the plunge pool to the jet fall height (H), the jet thickness (t) at impact, the jet impact velocity (v), and the minimum jet velocity (ve) require to entrain air. Ervine empirically determined the entrainment velocity (ve) to be 1.1 m/s. Mason (1989) used this result to refine an empirical equation he developed earlier (1985) to estimate the depth of scour from free jets below dams and flip buckets. In more recent work, Ervine (1987) derived an expression for the minimum jet velocity required for the onset of aeration in terms of the turbulence intensity; this is given as V = 0.275/(u7U) = 0.275/Tu, where u' is the root mean square velocity fluctuation and U is the mean jet velocity. For a typical turbulence intensity (Tu) of 4-5% a jet velocity of 5.5-5.7 m/s is required for the onset of aeration (ibid.). 2.2 Turbulent Energy Dissipation in the Plunge Pool Pressure fluctuations occur within the plunge pool as a result of the turbulent diffusion of energy and are an important part of the scour process (Annandale, 1995). Momentary pressure bursts can cause hydrofracturing or the 'jacking' of large pieces of rock along the pool lining through fissures in the rock or concrete/rock interface (ibid.). While the work in this area is incomplete, it does suggest a parameter by which scour can be gauged in a rock lined plunge pool. The presence of pulsating pressures in stilling basins, below hydraulic jumps, is well researched. Fiorotto and Rinaldo (1992) show the distribution of pressure fluctuations is well correlated to the area below an hydraulic jump, in the region of highest energy dissipation. Other authors have emphasized the importance of fluctuating pressures in the erosion process 7 (Akhmedov, 1968, 1988; Kobus, 1988; Spurr, 1985). Bowers (1988), in his analysis of the 1961 Karnafuli Project spillway damage, concluded that fluctuating pressures in the hydraulic jump initiated uplift and failure of the chute slab. In a study of pressure fluctuations at the base of a free overfall, Robinson et. al. found that average dynamic pressure values could increase to above static pressure as the backwater was decreased, with the instantaneous pressure fluctuations showing a higher variation at the point of maximum average pressure (1989). The maximum average pressure occurs at a point where the nappe impacts the basin floor; the maximum stress occurs just downstream of the impact point, presumably as the flow accelerates away from the impact (ibid.). The frequency and amplitude of pressure fluctuations for a model of the Morrow Point Dam, with a free detached jet issuing from the spillway and an impact energy absorbing pool, were measured and analyzed (King, 1967); these results were compared to pressure fluctuation data from a hydraulic jump stilling basin modeled after Crystal Dam. The pressure fluctuation data from the floor of each basin showed different characteristics. Pressure data from the Morrow Point pool show that amplitudes generally increase with decreasing frequency. At a model frequency of about 2 Hz the effective prototype pressure head was calculated to be 57.7 ft with a static head of 57.6 ft. The contribution of the dynamic pressure head is negligible for this frequency; however, larger fluctuations with frequencies below the range of the analyzer are likely present (ibid.). Pressure fluctuations measured in the stilling basin of the Crystal Dam model showed a different frequency spectrum. The largest pressures occurred at isolated peaks in the model around 3 Hz. In both setups, the models were scaled according to the Froude law, on the assumption that the lower frequencies are due primarily to the surface waves, surges, and large scale eddies in which gravity forces dominate (ibid.). This is supported by Vasiliev and 8 Bukryev (1967) who found that oscillations less than 50 Hz were scaleable according to the Froude Law. Lesleighter (1971) reviewed the results of pressure measurements in the stilling basins of three structures; these were: along the floor of a modeled trapezoidal hydraulic jump stilling basin; the dividing walls of a prototype hollow-jet valve stilling basin; and the vertical sidewall of a modeled roller-bucket type spillway (1971). Typical energy spectra show a predominance of energy in the lower frequency bands between 0.2 and 5 Hz, and, in some cases, with large pressure contributions at well defined frequencies (ibid.). Measurements were made of pressure fluctuations generated by turbulent flows in the fluid between two counter-rotating disks (Abry, Fauve et al, 1994). The turbulence was shown to display a random occurrence of strong pressure drops which correspond to vorticity concentrations. The pressure probability distribution function (PDF) was non-symmetrical with a roughly exponential tail toward low pressures and parabolically with a negative concavity toward the high pressure side. In an earlier experiment no qualitative differences were discerned between measurements taken using a 1 cm diameter pressure transducer and one with a diameter of 3 mm (Fauve et al, 1993). Kobus compared erosion from pulsating jets with that of steady jets (1979). The results indicate a greater degree of erosion from pulsating jets over steady jets; this he surmised occurs when the pulsation frequency of the jet resonates with turbulent elements (eddies), thus prolonging the duration of their existence (ibid.). Kirsten and Annandale have based their work on the assumption that the process of erosion is caused by fluctuating differential pressures between the fissures in the rock and the surrounding pool, and show that the rate at which energy is dissipated is a good indicator of the magnitude of such differential pressure fluctuations (Annandale & Kirsten, U.S.B.R., 1993, p.57). Annandale has recently proposed an erodibility classification index (1995). Here 9 the rate of energy dissipation, and thus pressure fluctuations, is used in conjunction with a geological indexing system to allow the determination of the onset of erosion. The significance seems to be that for a higher degree of variability in pressure fluctuations there exists a condition where extreme differential pressures could occur; this suggests that erosion is determined by the magnitude of the pressure fluctuation or resulting force, and with a higher variability there is a greater likelihood that the threshold differential pressure or force necessary for erosion will occur. Annandale's work provides a potentially useful framework to define the onset of erosion given some measure of turbulent pressure fluctuations in a scour hole. If pressure fluctuations could be correlated to energy dissipation then a further understanding of how pressure fluctuations vary with depth and flow could be used to predict scour depth at a particular site The vibrations that are induced by turbulent flow is an additional parameter which could be used to study scour. Work has been done to understand the frequency response of various structures to turbulent wind gusts. As well, vibrations have also been studied in mechanical systems where turbulent fluid can induce instabilities and result in failures. In these cases turbulent fluctuations of the flow are transmitted to the structure by fluctuating pressures on the surface of the structure; this, in turn, induces vibration. The phenomenon of spilling water inducing seismic waves in the rock was observed by the author at the Bennett Dam seismic recorder station. During a large spill event in 1996 at both Bennett Dam and Peace Canyon Dam a significant increase in seismic activity was recorded at the nearby Bennett Dam seismic station. Although it is difficult to ascertain whether the induced vibrations were the result of water spilling over the dam or turbulent pressure fluctuations in the plunge pool, the literature suggests that the latter is possible. 10 2.3 Modelling Spillway Scour Free surface flows associated with most spillways are good candidates for scale modelling. The flow processes in these specialized open channels can be assumed to be governed primarily by gravity and only slightly affected by viscosity. These conditions allow small-scale hydraulic structures to be built which are based on the Froude model law. The modelling of energy dissipation, though, requires that turbulence, an important component of energy dissipation, be represented. Turbulence is characterized by the range of Reynold's number because it results from an interaction of viscous and inertial reactions (Kobus, p.76). A plunging jet can, in general, be characterized as a system of large eddies. The main flow generates these eddies, the length scale of which are of the order of the main dimensions of the flow field. The large eddies are inertial, with the viscosity effects being negligible, and contain a large amount of the energy entering the pool. "Due to the decay of these large eddies, energy, is transferred to smaller and smaller eddies, until their size is finally so small that dissipation due to viscosity takes place" (ibid.). The pressure fluctuations which are associated with these large eddies can be scaled according to the Froude law; however, the pressures due to the finer scale turbulence cannot accurately be scaled in the same way because of the distortion of the Reynold's number. Lopardo (1988) found that macroturbulent pressure fluctuation amplitudes and frequencies for a hydraulic jump basin were able to be accurately scaled according to the Froude law for energy spectra lower than 20 Hz; he suggested for accuracy that the jump Reynold's Number be at least 100,000 and the scale factor be not less than 1:50. Physical models, employing a non-cohesive or granular bed are important tools for approximating the shape and size of scour holes. The implied assumption of the non-cohesive bed approach is that the fracturing process of the rock bed has already taken place (Otto, 11 1986). The broken-up rock, therefore, is represented by a moveable bed made up of uncohesive material. The greatest difficulty for such models is the proper representation of the non-cohesive bed material (ICOLD, 1987). A comparison of model and prototype data for the Tarbela dam spillway showed that the ability of the current in the prototype to transport scoured sediment out and away from the plunge pool could not be reproduced in the model (Ahmad, 1988). Despite the use of ever finer sediment in the model, a sediment bar continued to form at the downstream edge of the pool, something that was not observed in the prototype (ibid.). This difference between model and prototype may be explained by the ability in the prototype scour hole to grind material through ballmill action until it is able to be transported out of the pool. While modelling spillway flows can give a "fairly good" indication of scour pattern any quantitative interpretation of scour should be used very cautiously (ibid.). In a comparison of a 1:60 model and prototype of Pandoh Dam spillway, the modeled spillway scour was overestimated (Dhillon, 1982). Past experience and research into scour has shown that a Froude model can represent the important components of the flow which result in scouring. The greater difficulty in modelling scour is in the representation of the strength of rock at the model scale. In a non-cohesive bed model, the practical difficulties of modelling the rock are circumvented by making the implicit assumption that the rock is already broken up (Otto, 1986). Scour in non-cohesive bed models becomes more a function of the momentum impact and transport capacity (Reynold's stresses) of the modeled flow. Since the Reynold's stresses become distorted at a Froude scale, the model represents the momentum impact of the prototype most accurately. While this imposes a limitation on the results of a depth analysis, it still provides useful information on the progression and extent of scour. 12 Chapter 3 The Ruskin Project This chapter gives the context in which the scour investigation has been carried out. The arrangement of the study site is described; as well, specific features of the Ruskin site and historical information relevant to the study are detailed. In some of the descriptions that follow, important features of the site have been defined and are used in the following way. The plunge pool is taken to be the portion of the channel which extends from the spillway face to the end of the concrete apron and benches (Figure 3.1). The spillway channel is defined to begin immediately downstream of the concrete apron and benches. The side channel is defined as the short, narrow channel between the bridge pier and right abutment (Figure 3.1). In addition, the standard convention of referencing the spillway plunge pool and channel from the perspective of someone looking downstream of the dam crest is used. 3.1 The General Arrangement of the Site The Ruskin Project is a hydro-electric facility situated on the Stave River, located about 65 km east of Vancouver, North of the Fraser River between Haney and Mission. The facility is part of the larger Stave Falls-Alouette-Ruskin Project; this includes, in addition to the Ruskin Project, the Alouette Project and Stave Falls Project. The Stave Falls-Alouette-Ruskin Project includes four dams, and at full flow, produces 166,100 kW of electricity. The Alouette Project is the upper most portion of the overall project and is comprised of Alouette Dam, Alouette Lake and a single 8000 kW generating facility on the shore of Stave Lake. The Alouette Project was built between 1926 and 1928 with the dam replaced in 13 1984 by a new earth-fill structure. Alouette Lake is a 19.2 km long reservoir, containing 210 million cubic metres of water. The elevation difference between Alouette Lake and Stave Lake is 23 metres. The generating facility is fed by a 1070 m long, 4.6 m diameter tunnel at the North end of Alouette Lake. A low level outlet at Alouette Dam, located on the South end of Alouette Lake, provides minimum flows to be maintained in the natural Alouette River course. The Stave Project is comprised of Stave Lake, Stave Falls Dam, Blind Slough Dam, and the Stave Falls hydroelectric plant. The Stave Project was built between 1908 and 1925. Stave Lake is the second storage reservoir in the system and is fed by tributaries to the Lake, as well as water passed through the tunnel from Alouette Lake. Stave Lake has a storage volume of 580 million cubic metres. Blind Slough Dam and Stave Falls Dam are concrete structures located adjacent to each other, separated by a rock bluff. Blind Slough Dam is at the head of what used to be a natural flood channel to the Stave River. The spillway has 14 bays which discharge into a channel below Stave Falls Dam. The five turbines at the Stave Falls plant produce 52,500 kW of power and, currently, have a combined discharge capacity of 225 m3/s . A major powerplant extension, expected to be complete by 1999, will increase the discharge capacity from the turbines to 280 m3/s. * The Ruskin Project, completed in 1928, is comprised of Ruskin Dam, Hayward Lake, and the Ruskin hydroelectric plant. Hayward Lake is 5.6 km long and is impounded by Ruskin Dam. Hayward Lake itself has a negligible contributing watershed; thus, its water supply comes almost entirely from water discharged from the Stave Falls plant or over Blind Slough Dam (BC Hydro, SOO 4P-47). The plant has a power capacity of 105,600 kW, on 39 m of static head, and can discharge a maximum of 340 m3/s through the turbines. Ruskin 14 Dam is a concrete structure with seven spillway bays which allow water to be discharged into Hay ward River (Figure 3.1). The Stave Falls-Alouette-Ruskin Project, at the time of its completion, was the cornerstone of BC's power supply. With the addition during the 1960's and 1970's of much larger complexes in the Columbia and Peace River basins, the Stave Falls-Alouette-Ruskin Project has become less important in terms of supplying base power; however, because of its close proximity to the largest power demand centre in Vancouver the project remains important for meeting peak load demands. 3.2 Ruskin Dam Spillway The Ruskin spillway consists of 7 overflow bays, each of which has a width of 10.06 m, with the flow controlled by 7.92 m high radial gates. The gates are numbered in sequence from 1 to 7 starting from the right side of the dam, looking downstream. The gates are actuated by electrically driven hoisting machinery located underneath the bridge deck and can be controlled remotely from the powerhouse or dam. Concrete benches and training walls line both sides of the downstream banks of the spillway (Figure 3.2). In the original construction, a protective concrete apron was built extending 18.3 m out from the toe of the spillway. In 1937/38, the apron and some of the lower benches on the right side were extended an additional 7.62 m (Figure 3.3) after bedrock at the downstream end was discovered to be eroded and the apron undermined (Lazenby, p.26). The dimensions and orientation of the benches appear random but follow the contours of the natural channel. In addition to providing erosion protection to the banks, the benches redirect discharge into the main channel, especially when the outside gates are utilized. The significance of this redirection of flow becomes important in understanding the progression of 15 scour in the plunge pool. Reported observations of spilling from the outside gates in the prototype.are rare since they are used only for very large discharges. A characterization of the redirection of discharge, gained from observations during testing of the hydraulic model, is given in Chapter 4. A bridge crosses the spillway channel at a point approximately 35 m downstream of the end of the spillway apron. A bridge pier is situated in the channel and offset to the right from the centreline of the channel (Figure 3.4). The bridge pier is a large concrete structure with a width of approximately 17 m. A concrete slab at an elevation of 5.54 m (GSC) extends between the concrete pier and the right abutment. 3.3 Geology of the Site The bedrock at the Ruskin site is characterized as irregularly grained diorite (BCH, H1837). Diorite is a hard, dark grey, igneous rock composed mostly of feldspar and dark (mafic) minerals, consisting mostly of hornblende, with up to 10% quartz. The rock in the channel bed, just downstream of the spillway, is made up of moderately fractured hornblende diorite with minor faulting. Much of the rock is closely jointed and contains numerous shear and fracture zones (ibid.). No distinct sets of joints have been mapped but some general patterns have been observed. In addition to random jointing and orientation, steeply dipping joints greater than 60° occur most often parallel to the Stave River and dip in northwesterly to southeasterly direction. Flatter dipping joints of 30° to 60° were commonly found striking more or less perpendicular to the river in an upstream or downstream direction. 16 3.4 Spillway Operation The operation order for Ruskin Dam stipulates the opening sequence for the spillway gates. The preferred gate sequence is 5, 4, 6, 3, 2, 7, 1 (S.O.O. 4P-47, p.6). Discharge is established first in the central bays (5 and 4) where the flow enters directly into the plunge pool without being redirected by the benches. As more capacity is required, additional gates are opened in the prescribed sequence. Pertinent elevations related to the operation of the dam and spillway are summarized in Table 3.1. Table 3.1: Pertinent Elevations Elevation GSC (m) COMMENTS 45.73 Elevation of roadway (crest of dam) 44.76 Expected maximum reservoir elevation while routing PMF 42.91 Normal maximum operating level. Top of closed gates 34.98 Spillway crest elevation 24.30 Elevation of top concrete bench on right bank 10 Expected elevation of tailwater during PMF 9.10 Elevation of top bench on left bank 2.4 Approximate tailwater elevation during normal operation -10.50 Elevation of concrete apron at base of spillway -12.3 Minimum plunge pool elevation 3.5 Flood Routing Since there is no significant local inflow into Hayward Lake, and its storage capacity is small, any major flood routed through Ruskin Dam will be equivalent in magnitude to the routed flow from the Stave Falls Project upstream (BCFf, Report H1862, p. 6-2). Recently, the annual excedence probabilities for peak 6-hourly spillway discharges were assessed for the new Stave Falls Project, and thus, can be generalized to discharges at Ruskin; these are 17 summarized in Table 3.2 (BCH, Report # MEP182, p.2). The spillway discharge at Ruskin associated with the peak 6 hour probable maximum flood (PMF) is taken to be 4242 m3/s. The annual average peak daily spill discharge for the period 1961-1991 is 558 m3/s (ibid.). The peak discharge of the spillway, before overtopping of the roadway, is taken to be 4500 m3/s (BCH, MEP188). The maximum estimated spillway discharge to-date occurred in January, 1961 and was estimated to have a magnitude of 2340 m /s (BCH, Report HI862, p.6-4). Table 3.2: Spillway Flood Discharge Annual Excedence Probabilities Annual Excedence Probability Peak 6-Hourly Spillway Discharges Hayward Lake Elevation (m3/s) (m GSC) 0.005 0.001 0.0001 PMF 2555 2830 3440 4242 42.9 42.9 43.2 44.8 3.6 Overview of Observed Scour to Date The scour resulting from historical discharges is shown on two plunge pool contour maps (Figure 3.5, 3.6, 3.7). The first and more detailed of the two plunge pool surveys was done in 1935 and 1936; the second sounding survey was carried out in 1989. In addition to the surveys, information about scour in the plunge pool has been gathered over the years from underwater inspections by divers. During the 1935 and 1936 surveys erosion downstream of the spillway apron was discovered. In 1937 the concrete apron and some of the benches were extended 7.62 m (25 ft) downstream. A contour drawing was produced which details the extent of scour in 1935. In 1975, the right side of the bridge pier was damaged, just upstream of the concrete slab. An underwater inspection showed local scouring and undermining of the rock at the 18 base of the concrete bridge pier. Figure 3.8 details the extent of scouring of the concrete and rock, and the undermining of the right side of the concrete pier. The protective concrete of the bridge pier was subsequently repaired and the voids filled. An underwater inspection in 1983 did not reveal any serious erosion at the toe of the spillway (BCH, H1862). The 1989 survey was not extensive enough to produce a contour map to compare to that from the 1935/36 survey. Rather, in the 1989 survey, various cross-sectional and longitudinal profiles were produced of the spillway channel. The T/R Midpoint cross-section of Figure 3.7 was used to establish the current minimum bed elevation (-12.3 m GSC) and the maximum historical scour depth, which was used in the scour calculations (Section 3.7.1). The bed elevations and shape of the downstream channel cross-section of the hydraulic model, described in Chapter 4, was taken from the T/R Powerhouse cross-section. The large scour hole reported in Figure 3.8 would not have appeared in the 1989 T/R Bridge cross-section (Figure 3.7) because it was filled in shortly after its discovery in 1975. To appreciate the magnitude of the 1975 scour hole, its maximum depth profile, taken from Figure 3.8, was sketched into the T/R Bridge cross-section (Figure 3.7). The maximum depth profile is estimated to have occurred approximately 4.5 m upstream of the line of sight used for the T/R Bridge cross-section (ibid.). 3.7 Previous Scour Study to Predict Maximum Scour Depth An analysis to estimate the maximum scour depth in the spillway channel at Ruskin Dam was made by the author during the final work term at BC Hydro and was supervised by Dr. Warren Bell, Principal Engineer (BCH, W.W. Bell). The scour study was undertaken to assess the risk of dam failure due to the upstream progression of scour and undermining of the 19 dam foundation during the probable maximum flood (PMF). An outline of the approach and conclusions of the scour analysis are provided here. 3.7.1 Scour Study Approach Equations designed to predict scour from free-falling jets were used to estimate the ultimate scour depth below the Ruskin Dam spillway. Four of the five equations used here to predict scour are of the form, D = Kq vHwh x/g yd z where D is the depth below the tailwater surface elevation K is a constant, q is the discharge per unit width, H is the gross head drop between the forebay and tailwater; \ h is the depth downstream of the plunge pool; g is the acceleration due to gravity; and d is the characteristic particle size. The fifth equation, in addition to q and H, considers the angle (0) at which the jet impacts the plunge pool. The first equation was developed by Mason for prototype scour (Mason, 1989) and includes the effect of air entrainment on scour. The formula is based on scour caused by jets issuing from flip buckets, pressure gates, and overspills and incorporates research into the entrainment of air in water (Ervine, 1976, 1987). The constant, K, is taken to be 3.27, and the exponents are v = 0.6, w = 0.05, x = 0.15, y = 0.3, and z =0.1. The second method, offered by Mason in a discussion of Spurr's article (1985), involves calibrating Mason's equation to a known scour event, and then applying the 20 calculated parameter, K, to the PMF conditions (Mason, 1986). The existing scour hole was assumed here to have been formed by the largest spill event (2340 m3/s in January, 1961). The constant, K, can be solved from the equation using data from the Jan./61 flood, the known scour hole depth, and using exponent values recommended by Mason (v = 0.5, w = 0.5, x = 0.15, y = 0.3, z = 0.1). Using the calibrated value for K, the scour hole depth for the PMF discharge can then be calculated from the same equation. The third equation uses as exponents v = 0.5, w = 0.5, x = 0, y = 0, and z = 0 (Hager, 1995). The equation was developed by Damle in 1966, and with a K-value of 0.362, gives the upper bound value of prototype scour in feet for a specified flow. The fourth equation is the Veronese equation and is endorsed by the Bureau of Reclamation (1977). The Veronese equation uses a constant (K) of 1.32 and exponents v = 0.54 and w = 0.225. The Veronese calculates scour depth, in feet, below the tailwater surface elevation. The fifth equation was developed by Coleman and is a modification of the Veronese equation (Hager, 1995). Coleman's equation takes into account the angle that the jet impinges on the plunge pool and gives the depth of scour below the tailwater depth, in metres, as Z = 1.9H°' 2 2 5qO 5 4sin0. The tailwater depth is defined as the depth downstream of the scour hole; this is a different reference elevation than the other four equations. The elevation of the tailwater bed was taken to be -6.0 m GSC for the calculation of scour depth in Coleman's equation. The peak spillway discharge used in the four equations to estimate maximum scour depth was taken to be 4500 m3/s. The corresponding forebay and tailwater elevations for the maximum spill discharge was estimated to be 46 m GSC and 10 m GSC, respectively. The current minimum bed elevation was taken to be -12.3 m GSC. 21 3.7.2 Results of Scour Calculations All the equations used in the calculations were derived from prototype scour data associated with free falling jets. The crest shape of Ruskin dam follows U.S. Army Corps Engineers design for free overflow (BC Hydro, Report HI862). Thus, flow over the crest of Ruskin dam can be assumed to approximate nappe flow, with a roughly a free-falling jet down the spillway. The impact angle of the jet in the plunge pool is assumed to follow that of the spillway face. However, the concrete benches and apron deflect the flow and make analysis of the true parameters for the equations more difficult to assess. The geology is taken into account in Mason's equation by a single parameter, the median partical size. In the other equations the geology is not considered The concrete apron at the toe of the spillway deflects the flow into the central unlined portion of the spillway channel. Although the free-surface jet strikes the central portion of the plunge pool at a steep angle (approximately 56°), it issues from the end of the apron as a submerged jet and strikes the rock at a much flatter angle (Figure 3.2). The concrete benches, as well, contribute to the complexity of the scour problem by deflecting the free-surface flow into the concrete lined plunge pool area and the unlined downstream portions of the spillway channel. Discharges from the outside gates, particularly #1 and #7, are deflected by angled walls which effectively flip the bulk of the water into the unlined central channel, and create a rooster tail effect. The combined effect of the concrete benches and apron on a spillway discharge with the seven gates operating simultaneously is a combination of multiple, highly turbulent free-falling jets impacting in the central channel, and an underwater horizontal jet issuing at the level of the apron. Although the spillway discharge is not strictly speaking a free jet discharge, the results of the scour equations are still relevant. All the equations were derived from a large database 22 of either model or prototype data, and, in Mason's case, both prototype and model data. The results of the equation are meant to represent the worst case spill condition. The spill over Ruskin Dam can be thought of as a single free-falling jet with the concrete benches and apron deflecting the flow, but also providing some degree of energy dissipation. There is some uncertainty with respect to the value of unit discharge (q) since flow may tend to become concentrated in the central, lined portion of the plunge pool. In order to take into account of this uncertainty, two scenarios were considered. The two cases can be seen to represent a range of upper and lower bound values of unit discharge. In addition, a consideration of the results of four scour equations allows further the development of a range of values for maximum scour depth. The two possible spill scenarios considered give an upper and lower bound for the value of q, discharge per unit length, a parameter used in all four equations. In scenario # 1 , all the PMF discharge is assumed to be concentrated across the width of the plunge pool, represented by the width of the backwater area at the spillway face during a non-spill condition. In scenario # 2 , the discharge is assumed to be spread evenly over the width of the spillway; this scenario allows for the possibility of scour to be caused by the deflection of the discharge by the benches. In both cases, the effects of the benches and apron on the dissipation of energy are neglected. The actual value of q that contributes to scour is uncertain but should be accounted for by the range of values. More importantly, the equations give no, information about where the highest scour intensity will occur. The precise way in which the discharge is distributed and contributes to scour cannot be determined in any analytical way. The manner in which the flow is deflected by the benches, and thus, the momentum impact in the spillway channel ultimately affect the intensity as well as the location of scour along the bed. No documentation could be found 23 which describes the flow patterns in the plunge pool during past high magnitude spills. An hydraulic model of the spillway and benches is ideally suited to locating the highest scour intensity because it can represent the location of the highest momentum impact of the jet and take into account the effects of secondary currents. There is a significant difference between the scour depth calculated from Mason's first equation and that from the calibrated version. Mason's equation is a refinement of an earlier equation which was based on 26 sets of prototype and 47 sets of model scour data, and is considered an upper bound for prototype and model scour (Mason, 1989). Mason's equation predicts the maximum scour level to go to -37.1 m GSC given conditions expected during a PMF and Scenario #1; this is 24.8 m deeper than the current plunge pool elevation. The calibration method, however, is prone to underestimation; this is because using information from the largest spill event on record and assumes that the spill duration was adequate for the maximum or equilibrium depth to be reached. If the full scour depth was not reached during the 1961 event then the value derived for K, and the maximum depth calculated for the PMF conditions, would be underestimated. Coleman's equation takes into account the angle that the jet impinges the plunge pool. The maximum angle that of jet impingement is 56.2°, the angle of the spillway face. This value would clearly overestimate the channel scour depth in Coleman's equation in central spillway channel because the deflection of the underwater jet by the spillway apron. The concrete apron deflects the discharge entering the plunge which results in an underwater jet entering the channel at a much flatter angle. However, discharge that is flipped by the outside concrete benches into the spillway channel can be assumed to represent a water jet impacting the channel at a steep angle. In the absence of any data on the angle of impact by the deflected discharge the value of 56.2° was used. 24 The calculations for maximum scour were based on a tailwater elevation of 10 m GSC and a minimum bed elevation of -12.3 m GSC. A comparison is made of the five calculated values in Table 3.3. Table 3.3: SUMMARY OF CALCULATED MAXIMUM SCOUR DEPTHS Predicted Additional Minimum Scour to Equation Used Elevation of Current Comments Scour Hole Minimum Bed Elevation (all calculations based on a discharge of 4500 m3/s) (m GSC) (m) Scenario Scenario Scenario Scenario #1 #2 #1 #2 Mason's -37.1 -26.8 24.8 14.5 Considered an upper bound for scour depth & includes effects of air entrainment. Calibrated Version -17.7 -18.7 5.4 6.4 Prone to underestimation of of Mason's scour depth. Damle's -24.2 -18.3 11.9 6.0 Equation developed as an upper bound of prototype scour. Veronese -33.8 -25.7 21.5 13.4 Modified Veronese -42.4 -35.7 17.0 12.9 Result based on the angle of (Coleman's Eq.) jet entry. The results of this study show great variability in the value of predictive depth of maximum scour (5.4 - 24.8 m of additional scour). There is additional uncertainty with respect to the prediction of the location and extent of scouring within the channel. The complex hydraulics that are created by the interaction of the spilled water deflected by the concrete benches makes the analytical prediction of scour development and the influence of secondary currents difficult. The presence of the bridge pier and concrete slab extending from the right abutment and pier introduces an obstruction which may affect the scour development. The constriction 25 in the cross-sectional area of the channel due to the presence of a bridge pier and concrete slab could, at very high discharges, increase the water level in the plunge pool area. The depth of the plunge pool is an important variable in the scour equations considered above. A higher water level in the plunge pool will reduce the gross head of the spilled water and, theoretically, will reduce the maximum scour depth. Although this reduction of scour depth is desirable, a failure of the bridge pier would remove the channel control and result in a sudden increase in gross head and potentially a sudden acceleration of scouring in the plunge pool. Given this scenario, the risk of scouring in the plunge pool becomes dependent on the risk of failure of the bridge pier. The hydraulic model study was initiated, in part, to test whether there is any significant backwater in the plunge pool area during high discharges as a result of the channel constriction and to assess the effect on scour due to changes in the downstream hydraulic conditions. A description of the model test and the results are given in Chapter 4. 3.7.3 Conclusion of Analytical Scour Study The additional scour calculated by the five equations gives a range of depth between 5.4m and 24.8m. The worst case scour was predicted using Mason's equation for scour depth; from this equation, as much as 24.8 metres of additional scour could be caused by a PMF event. This figure represents an extreme value, in part because it is calculated using an equation which is prone to overestimation; but also because it takes into account an upper bound estimate for unit discharge. As well, the lower estimates of scour depth from the calibration method is prone to underestimation. Neglecting the upper and lower bound values as an overestimation and underestimation, respectively, the range of additional scour is calculated to be between 6.0m and 21.5m. 26 27 / *3 28 29 Figure 3.4: Concrete Bridge Pier in the Prototype Spillway Channel (Photo taken from a point above Gate 7) 3 0 32 33 34 Chapter 4 Ruskin Dam Spillway Model and Study Approach In this chapter, the testing and results of a non-cohesive bed, physical model study are described. The primary purpose of the hydraulic model was to supplement the analytical study, undertaken to predict maximum scour below Ruskin spillway, as described in Chapter 3. This chapter is divided into 4 main sections. Included in the first section are a description of the hydraulic model and summary of test results. In Section 4.2, observations of spill flow patterns from the model are discussed and used to characterize the location and depth of the resulting scour. In Section 4.3 the observed scouring in the model for various test cases is presented. The same five equations used in Chapter 3 to calculate prototype scour depth were used to calculate scour depth in the model. The predicted and model observed values of scour are presented in Section 4.4. In Section 4.5, data from the model scour study is applied to the analytical study to refine the prototype scour estimate. A summary of results is presented in Section 4.6. The use of equations to predict the depth of scour assumes that the location of the maximum scour is known. For most spillways, where the impact location of a water jet or hydraulic jump is clearly defined, the maximum depth of scour is obvious. Information about the location of maximum scour is important for assessing dam stability in the event of an extreme discharge event. At the Ruskin site the discharge over the spillway is not uniform. It is not clear from studying the layout of the spillway where the location of the maximum momentum impact and maximum depth of scour would occur for a high discharge event. The concrete benches and training walls along the sides of the spillway redirect the flow into the downstream channel. The precise way in which the flow from the seven bays interact and 35 impact the channel is not well understood, given historical discharges, and cannot be predicted analytically with any high degree of confidence. In addition, the presence of a concrete bridge pier immediately downstream of the apron and a concrete slab between the right abutment and pier, represents a constriction in the cross-section of the channel which may affect the upstream hydraulics for extreme discharge events and influence the degree of scouring. The hydraulic model, theoretically, takes into account the effects of the complex boundary conditions at the site and indicates, the location and relative severity of scouring for varying discharge magnitudes and open-gate sequences. A non-cohesive or coarse sand bed is used to model the bedrock in the spillway channel. In areas of high scour intensity, the sand particles are transported downstream resulting in scour hole. A comparison of the depth of these scour holes, then, gives a relative measure of the worst case scour for the different spill scenarios. Scaling the depth of the scour hole from the model bed to the prototype bed tends to overestimate scour depth. The information from the above analytical scour study, though, can be used in conjunction with the model study to gauge the depth and location of maximum scour. Scour depth data, measured from the model, can be compared to values calculated using the scour equations of the analytical study in Chapter 3. The equations were developed to predict model scour, in addition to prototype scour. The ability of the scour equations to give a reasonable upper bound estimate for the model scour is a test of their application to the prototype. The stacking of information in this way, that is, using the results of the hydraulic model in addition to calculations using prototype data, provides a new level of understanding of the scour processes and facilitates a more accurate assessment of the risk of dam foundation erosion due to scouring in the spillway channel. 36 In the descriptions of the model study which follow some conventions are defined here. The side channel is defined as the short, narrow channel between the bridge pier and right abutment (Figure 3.1). 4.1 Description of the Model The hydraulic model of Ruskin spillway used in this study was constructed at an undistorted scale of 1:50 (Figure 4.5). The model dimensions were chosen in accordance with the Froude criterion for dynamic similarity. The model was located in a large wave pond in the UBC Hydraulics Laboratory (Figure 4.7). The basin has concrete walls with a height of 0.6 m and a free surface outlet with an invert 0.2 m above the floor. Water pumped into the head pond flows over the spillway and down the 1.72 m wide channel, discharging into the basin and back to the pumps. Two gates at the end of the 2.2 m long channel control the backwater and define the downstream bed level. The top of the apron was 0.533 m above the concrete floor of the basin. The as-built dimensions of the model spillway and dam were measured and found to be within 5% of the scaled prototype as-built dimensions. At a 1:50 Froude scale the length, discharge, velocity and time parameters have the following model to prototype scales (Henderson, p.491): Length 1:50 Discharge 1:17,700 Velocity 1:7.07 Time 1:7.07 Water was supplied to the model by a 0.17 m3/s and 0.11 m3/s pump. Two separate pipes (250 mm and 200 mm) were connected to the head pond from the laboratory supply lines. Inside the head pond, two baffled pipes, with perforated holes directed toward the back 37 corner of the pond, distributed the flow across the width of the pond. The dimensions of the head pond were 1.87 m wide, by 2.44 m long, and with a height of 1.8 m. The head pond was built as large as practically possible to ensure a smooth transition of flow over the dam. At the larger design flows, the two baffles working together performed well with only minor surface disturbances observed in the head pond. The radial gates on the dam were constructed from 0.30 m (12 in) diameter PVC pipe. To simplify the testing, the spillway gates were not designed to be partially opened. The gates were set in place by counter-acting turnbuckles which held the gates in the closed position. Each spill scenario, then, is based on the gates being either fully open or closed with a predetermined discharge magnitude established prior to testing. The pre-determined discharge was established by satisfying the head requirement in the pond. The head-discharge relationship for the model was derived from the spillway curves from the prototype spillway (Figure 4.4). A v-notch weir was used to verify the accuracy of the approach of scaling down the stage-discharge curves for the spillway. The v-notch weir was installed at the outlet to the basin and measured flows were compared against the flow calculated from the scaled down rating curves. The measured discharges agreed to the rating curve values to within 7% error for flows of 0.014 m3/s and 0.021 m3/s; that is, assuming that the flow measured using the v-notch weir is accurate, the use of the scaled rating curves underrepresented the flow by about 7%. Comparisons using higher flows could not be carried put because of the shallow nature of the basin. The invert of the outlet and the top of the walls were 0.2m and 0.6m above the floor of the basin. A discharge greater than about 0.028 m3/s with the v-notch weir in place would have over-topped the walls. The level of accuracy indicated by the v-notch weir was 38 considered adequate for justifying the approach of using scaled rating curves as a measure of flow in the model. The head above the spillway crest in the head pond, which was used in conjunction with the scaled stage-discharge curves to measure flow (Figure 4.4), was measured using a single water level indicator attached to the outside of the head pond. The open-ended manometer which was used as the water level indicator consisted of a clear plastic tube inserted into a hole drilled through the wall of the head pond approximately 1.4 metres upstream of the crest; the other end was connected to a fixed vertical tube and ruler mounted on the outside of the wall. The elevation of the spillway crest was measured using a surveyor's level and marked on the ruler. The head above the spillway was measured as the difference in head between the water level reading and the elevation of the spillway crest. The location of the manometer was chosen in order to balance the influences on the water level readings caused by drawdown effects near the spillway crest and turbulence effects in the region of the baffles at the rear of the stilling pond. The error in calculating discharges using measured values of head at the manometer would be greatest for the maximum flow where drawdown effects would be greatest. In fact, the calculated maximum discharge in model using measured head values and the scaled rating curves was 0.237 m3/s; the combined discharge of the pumps is rated as 0.280 m3/s. The difference, an error of 15%, is not considered significant since it is the combined result of losses in the supply pipes, as well as error in the water level readings due to the drawdown effect. Since two separate lines and baffle pipes were used to supply water to the model, it is not unreasonable to expect some reduced discharge capacity associated with frictional losses in the pipes and bends (3), and exit losses in the two baffles. The prototype discharge which corresponds to the maximum 39 model discharge is considered in this study to adequately represent the magnitude of 6-hour PMF spill discharge (Table 3.1). The spillway and benches were constructed of wood and painted with several coats of melamine paint to achieve a smooth, hard surface (Figure 4.6, 4.7). The smoothness of the melamine paint was considered adequate for scaling the roughness of the prototype spillway (Henderson, p.498). The spillway benches were attached by bolts to large interconnected concrete blocks, the top surface of which served as the spillway apron (Figure 4.6). The spillway itself was also attached by bolts to the concrete blocks. The concrete blocks each measured 0.533 m in height, with varying lengths and widths. The 5 concrete blocks were cast separately and bolted together by two 19 mm diameter steel bars which ran the width of the blocks. The total weight of the blocks was approximately 1400 kg and thus provided a measure of stability to counter the hydrostatic forces, as well as increased stiffness to reduce vibrations induced by water flowing over the spillway and benches. The downstream channel walls of the model were constructed of 19mm plywood. Concrete blocks and large stones were used to model the contours of the banks. The bridge pier was built up from formed concrete blocks having a cross-section scaled down from that of the prototype pier. Two gates at the downstream end of the model were installed to control the water and bed levels. A variable gate at the end of the channel was used to control the tailwater level. A second fixed gate, 0.6 m upstream of the water level gate, was installed to keep the downstream bed level constant. The shape and elevation of the bed level gate was generalized from the downstream cross-section taken from Figure 3.7. The space between the bed level gate and water level gate provided a sump for trapping bed material transported down the channel. 40 A commercial grade of silica abrasive with a dso of 3 mm was used as sediment in the non-cohesive bed. The size of the bed material was initially determined from practical considerations given the absence of any data on sediment in the prototype spillway channel. The 3 mm grain size of the sediment was considered a good initial approximation given the modelling experience of the Pandoh Dam spillway which was of similar scale and discharge (Dhillon, 1982). The sediment was initially assumed to be slightly undersized. This would have allowed the gradation to be increased in size by mixing in a larger grain size, if it were required. However, the performance of the 3 mm silica abrasive was found to be adequate. Good scour profiles were achieved for a range of flows without excessive scouring at the maximum discharge. Scour data was taken from the model in the form of photographs of contours of the drained bed. After each run, the contours of the bed were mapped using fluorescent string. The contour intervals of the string at the model scale were set at 0.02 m and referenced to the top of the spillway apron which was arbitrarily set at 900 mm. The contour elevations were measured down from a tracked aluminum beam, with the bottom face set a distance 0.745 m above the apron. The beam could be positioned anywhere along the length of the channel to allow measurement of the entire bed profile. Nails were used to temporarily mark individual measurements which then aided in positioning the string along the contours of the bed. 4.1.1 Operation of the Model The testing procedure was identical for all the tests carried out on the model. The gravel was first contoured to approximate the bed of prototype spillway channel bed. Next the channel and head pond were filled and the model allowed to run at low discharges until the basin was filled to capacity. Both supply lines were then opened until a predetermined 41 discharge, measured by the upstream head over the spillway crest, was reached. The model was allowed to run for 60 minutes. At the end of the run time the water was drained from the basin and mapping of the bed contours was carried out using fluorescent string. The time needed for a single test run, including preparation and photographing the bed contours, was approximately 6 hours. The 400mm supply line was used to slowly fill the head pond. While the pond filled, the downstream channel was also filled using two supply lines. The first supply line was a 50mm hose connected to a valve at the bottom of the head pond; the second supply line was a fire hose connected to an external source. Both hoses discharged into the sediment sump, between the bed level gate and water level gate; this ensured that the channel would be back-filled with water with no disturbances to the coarse sand bed. The rate of rise in the head pond was timed such that when water level reached the crest of the spillway, the channel was near the top of the downstream water level gate. A small amount of flow was allowed to spill over the dam until the channel and basin were completely filled. The extra time taken to prepare the model for testing ensured that the bed material would be undisturbed by water spilling directly onto the bed. Once the model was filled with water, both supply lines to the head pond were opened together until the required water level, corresponding to the desired discharge, was reached. The model was allowed to run for a duration of 1 hour. Given that the scaling of time for a 1:50 Froude model is 1:7.07, the 1 hour run time represents, theoretically, a prototype spill duration of 7.07 hrs. The 1 hour run time was determined during the calibration phase of the testing to ensure the development of stable bed forms. 42 4.1.2 Calibration of the Model Prior to testing, several calibration tests were performed to confirm the similarity of the scour profile to the prototype channel bed, as well as to assess the time needed for the formation of a stable bed. Unfortunately, the calibration of the model, due to a lack of detailed prototype historical flow and scour data, was more a subjective process to corroborate model-prototype similtude. The approach of the model calibration was to confirm that at relatively low flows similar bed forms of prototype channel could be reproduced in the model. For the calibration process, the only detailed information of prototype scour, taken from a 1935/36 (Figure 3.5) contour map of the channel bed, was compared to model scour. The limitation of this approach is that the surveyed information for the contour map was taken in 1935 and 1936, only 11 years after the dam went into operation. The 1989 sounding survey, consisting of 2 channel cross-sections and 1 longitudinal profile, does not adequately represent the prototype bed forms and could not be used. Records of spillway discharge for the period between 1924 and 1935 were not available from either Ruskin or Stave Falls for this study, and thus, some assumptions about the spill magnitudes and gate operation needed to be made. It was assumed that the spillway gates were operated in the same way as they are today and the peak spill event for the period up to 1935 utilized the first 3 gates in the operating sequence (gates 4, 5, and 6); that is, the peak discharge was less than about 1200 m3/s for the period up to 1935. A discharge of 1200 m3/s was chosen as an upper bound because it is more than twice the average magnitude spill event for the 1961-1991 period (BCH, Report# MEP182, p.2) and has a fairly low likelihood of having been exceeded given the small time frame. A discharge greater than 1200 m /s requires a fourth gate and changes 43 the scour profile significantly. Using these assumptions, the general bed form of the prototype, using Figure 3.5 as a guide, can be produced in the model. The resulting scour on the non-cohesive bed as a result of various discharge magnitudes from bays 4, 5 and 6 were compared to the 1935 survey. The scour profile from the model with gate 4,5, and 6 open, represented by Figure 4.15, Figure 4.16, Figure 4.17, and Figure 4.30, all show elements of similarity with the 1935 survey (Figure 3.5). The model tests having a smaller magnitude discharge (Figure 4.15 and Figure 4.17) give a better representation of the prototype conditions than the other two tests (Figure 4.16, Figure 4.30). The indentation at the base of the benches on the left side, caused by the return current is represented well in Figure 4.15 and Figure 4.17; these runs also represent the general shape of the scoured main channel, between the pier and left abutment. The difference between Tests 1 and Test 3 is the high scour which shows up along the face of the pier. The difference in scour intensity along the boundary of the bridge pier suggests that equilibrium conditions are not represented in the prototype data; this difference may also be the effect of introducing a fixed boundary, the bridge pier, in the model. The presence of the concrete pier in the model introduces a boundary which does not model accurately the prototype conditions. The prototype pier is fixed to the bedrock channel and thus can be undercut if exposed to scour intensities at the base. The model pier represents a discontinuity, that is, a long, effectively infinite, column of unscourable material. Insofar as the model test is designed to locate areas of high scour intensities, it is assumed that this dissimilarity in model-prototype behaviour will not affect the results significantly. The areas of high scour intensity are initially set up by the boundary conditions which are represented, in part, by the pier and banks. The scouring will over time change the boundaries in the 44 prototype channel and the magnitude of scour intensity along the bed; but this, it is assumed, will not alter significantly the location of the scouring. The decision to use a 1 hour spill duration used in the model was reached after assessing the resulting scour of several calibration runs. The objective of these calibration runs was to choose a run time beyond which the scour pattern did not change significantly. The testing was done for a model spill of approximately 0.11 m3/s (1900 m3/s prototype). After successive trials, a one hour run time was selected as one which ensured that an equilibrium bed condition had been reached; this run time was verified for discharges of 0.17 m3/s (3000 m3/s prototype) and 0.23 m3/s (4100 m3/s prototype). It was found that for all the discharges tested, the equilibrium bed condition was reached in less than 1 hour; that is, the bed did not change in any appreciable way for tests having a spill duration of greater than 1 hour. For the 0.17 m3/s and 0:23 m3/s model flows, the equilibrium bed condition was reached in less that 1/2 hour; 4.1.3 Results Of The Model Test In total, 20 tests were run with varying discharge magnitudes and gate opening configurations. A summary of discharges and the gate openings for these tests is given in Table 4.1. The elevation of the bottom of the scour holes are listed in Table 4.1 and are referenced to the prototype GSC (Geological Survey of Canada) datum. As well, the additional depth of scour, referenced to the assumed prototype bed elevation (-12.3 m GSC), in prototype metres, is tabulated; thus, a positive depth value indicates the additional scour which could be expected if the model results are used as an indicator. In order to make comparisons between the prototype and model more meaningful in the descriptions which follow, all model discharges are referenced to the equivalent prototype discharge. 45 Table 4.1: SUMMARY OF MODEL TESTS Test# Gates Open Model Equivalent Elevation Additional 1 Referenced to prototype scale Discharge Prototype Discharge of Max. Depth Scouring1 [m3/s] [m3/s]' [m GSC] 1 Location 1 4,5,6 0.045 800 -13.3 1.0 -left U/S base of bridge pier 2 4,5,6 0.056 982 n/a n/a n/a 3 4,5,6 0.037 661 -12.3 0 -D/S of apron 4 3,4,5 0.045 800 -15.3 3.0 -left U/S base of bridge pier 5 3,4,5,6 0.079 1400 -19.3 7.0 -right U/S base of bridge pier 6 3,4,5,6,7 0.111 1960 -19.3 5.0 -left U/S base of bridge pier 7 2,3,4,5,6,7 0.137 2430 -25.3 13 -base of slab 8 1,2,3,4,5,6,7 0.155 2740 -22.3 10 -left U/S base of bridge pier 9 1,2,3,4,5,6,7 (pier removed) 0.167 2950 -15.3 3.0 -D/S base of left benches 10 2,3,4,5,6,7 (pier removed) 0.149 2630 -19.3 7.0 -right bridge abutment 11 2,3,4,5,6 (pier removed) 0.135 2390 -17.3 5.0 -D/S base of left benches 12 2,3,4,5,6 0.129 2280 -20.3 8.0 -base of slab 13 3,4,5,6 0.105 1860 -18.3 6.0 -left U/S base of bridge pier 14 2,3,4,5,6 0.110 1940 -19.3 7.0 -base of slab 15 1,2,3, 6,7 0.113 2000 -21.3 9.0 -left U/S base of bridge pier 16 4,5,6 0.076 1340 -19.3 7.0 -left U/S base of bridge pier 17 2,3,4,5,6,7 0.155 , 2740 -26.3 14 -base of slab 18 1,2,3,4,5,6,7 0.204 3600 -25.3 13 -left U/S base of bridge pier 19 1,2,3,4,5,6,7 0.237 4190 -28.3 16 -left U/S base of bridge pier 20 1,2,3,4,5,6,7 (pier removed) 0.233 4120 -19.3 7.0 -D/S centre of channel The results of the depth measurements tabulated in Table 4.1 are presented in 4.1 as a graph of discharge versus depth. Each of the data points in Figure 4.1 represents the elevation, at the prototype scale, of the deepest area of the measured model scour. The 'Test #' id from Table 4.1 is labeled next to each of the measured data points. 46 Figure 4.1: Plot of Measured Maximum Scour Depths (Integer next to data point refers to Test#) Sca led Model Discharge (m/s) 4.2 Observations of Flow Patterns and the Resulting Scour 4.2.1 Return Currents Return currents were observed in the model and are responsible for the scour along the downstream end of the benches. The indentation seen in Figure 3.5 to the left of the apron, at the base of the benches, is similar to that produced in the model at lower flows with only the central gates open (Figure 4.15, Figure 4.16, Figure 4.17). In the model, this local scouring was the result of a return current which started at the left bridge abutment and travels upstream along the bank; near the benches, the current becomes drawn into strong underwater jet issuing from the central plunge pool area. At higher discharges, with only the 47 open, the scouring in the model at the base of the benches becomes more severe (Figure 4.30). Similar return currents in the lower discharge tests cause scouring at the base of the benches on the right side; however, the scouring is not as pronounced as in the higher discharge tests. When only the central spillway bays 4, 5, 6 are used, the resulting discharge produces a return current in the model which travels upstream, in relation to the spillway discharge, through the channel between the pier and right abutment (Figure 4.15). All the discharge from the 3 bays is deflected into the main channel between the pier and left abutment by the concrete benches. The resulting upstream return current is produced by a back eddy in the main channel which curls around the bridge pier and produces flow in the upstream direction. The upstream current was seen for all four tests incorporating bays 4, 5, and 6 with discharges ranging in magnitude between 650 m3/s and 1340 m3/s. When the outside gates are opened, in addition to the central gates, more flow is directed through the side channel and the upstream current disappears (Figure 4.19). 4.2.2 The Roller in the Plunge Pool A prominent feature of the flow pattern in the plunge pool area for all tests is the roller produced when gates 4 and 5 are open. The roller produces a strong, flat underwater jet which enters the spillway channel at the elevation of the spillway apron. The benches just above the top of the apron are angled such that most of the underwater jet is directed into the main channel, between the pier and left bridge abutment (Figure 3.5). Strong vortices are shed at the downstream edges and corners of the benches in the model as the jet issues from the plunge pool area. Since the jet is nearly horizontal in Orientation when it enters the spillway channel it does not produce any deep, local scouring of the bed, except where it interacts with the left side of the pier. The effect of the underwater jet on the shape of the spillway channel 48 in plan can be seen in the prototype contour map of the bed (Figure 3.5). The orientation of channel just downstream of the apron has an angle in plan that coincides with the angle of the lower benches. The overall shape of the main spillway channel, in plan, is determined by the underwater jet issuing from the base of the roller and the lower benches which, in effect, deflect the underwater jet away from the bridge pier. 4.2.3 The Rooster Tail Effect A rooster tail effect is caused when the spillway discharge from the end gates is deflected by the angled concrete benches and training walls and is essentially flipped into the central channel (Figure 4.13, Figure 4.14). The jet of water that is flipped by the outside boundaries of the spillway is extremely turbulent and fans across the trajectory like a rooster's tail. This effect is especially prevalent with discharge through bays 1 and 7, but occurs both in the model and prototype, to a lesser extent, with discharge through gate 6. The same phenomenon can be observed for spillway discharges from gate lin the prototype (Figure 4.10). Figure 4.9 shows the effect of the deflection in the prototype by the lower benches for a relatively low discharge. At higher flows the rooster tail effect in both the model and prototype is much more pronounced. It is precisely this deflection of the water by the benches which introduced the greatest uncertainty in trying to analytically predict the development of scour in the spillway channel. The model shows good agreement in the way the flow patterns are established; and thus, the scoured bed patterns in the model should take into account the complex flow patterns set up by the concrete benches. An interesting effect of the interaction of rooster tails, or deflected water jets, can be observed when all the gates are open. When all seven gates are open water is deflected by the spillway boundaries on both sides creating two free surface jets which impact the spillway 49 channel. The bulk of the momentum impact from these two rooster tails, associated with the deflection of flow through bays 1 and 7, impact the spillway channel at about the same point, directly downstream of the bridge pier. That is, the two jets originate from opposite sides of the spillway yet impact at the same point directly upstream of the bridge pier. The result of this flow condition can be seen in the scoured bed profile as two distinct depressions directly in front of the bridge pier (Figure 4.22, Figure 4.32, Figure 4.33). 4.2.4 The Effect of Different Spill Cases the Water Level The backwater effect which was thought to occur in the prototype plunge pool area during very high discharge magnitude scenarios was not observed in model. To test whether the bridge pier and concrete slab created enough of a constriction to increase the water level in the plunge pool, a comparison of the water level in the plunge pool was measured for similar discharge magnitudes with and without the bridge pier and concrete slab in-place. In Tests 8 and 9, the model was run with and without the pier in-place for discharge magnitudes of 2740 m3/s and 2950 m3/s; in Tests 19 and 20, the model was run with and without the pier in-place for discharge magnitudes of 4190 m /s and 4120 m /s. The difference in water level for both these tests was less than 0.02 m (1.0 m prototype) and was considered to be insignificant given the inherent difficulties of measuring the water level precisely. The water level in the model was measured directly from a beam laid across the channel just downstream of left spillway benches. The turbulence in the channel flow and the fluctuating water level during spilling made taking a precise measurement of the water difficult. A water level reading, the average of 5 measurements taken every 2 minutes, showed an increase of 0.02 m (1.0 m prototype) between Tests 19 and 20. In the prototype, a 1.0 metre difference might be 50 significant; however, given the imprecision inherent in the model measurements, it suggests there is little or no backwater effect in the plunge pool during high discharge events. A significant local increase in flow and water level was observed upstream of the side channel, between the right abutment and bridge pier with the highest increase observed when most of the discharge was distributed over bays 2 to 7 (Figure 4.31). A local increase in water level upstream of the side channel was observed when there is greater propensity for the flow to be concentrated on the right side of the spillway channel. The increase in flow is caused by discharge from bay 6 and 7 being deflected into the spillway channel, directly in front of the pier; this deflects much of the flow from bays 2 and 3, and some from bay 4, through the right side channel. With gate 1 open, the local increase in flow through the side channel diminishes significantly. Flow from bay 1 is deflected by the training wall, along the upper portion of the spillway, causing it to be deflected towards the centre of the channel. Some of the deflected flow interacts with the water issuing from bays 2 and 3, which in turn becomes deflected into the central channel. The result is that more of the spillway discharge through bays 2 and 3 leaves through the main channel when spill from gate 1 is balanced with spill from gate 7. When gates 2 to 7 are open a large part of the flow becomes funneled through the narrow opening between the bridge pier and right abutment and causes the water level to increase locally. The deflection of flow from gate 2 by spill from gate 1 can be observed in the prototype at low spill magnitudes in Figure 4.10. 4.3 Results of Scour Depth Measurements The maximum scour in the model reached a level 0.32 m (16 m prototype) below the equivalent prototype bed elevation, for the maximum discharge of 4190 m3/s. The point of maximum scour was located at the left upstream face of the pier (Figure 4.33). 51 The second deepest scour hole was located at the base of the concrete slab between the right abutment and bridge pier and was more a function of gate sequence rather than spill magnitude (Figure 4.31). This test incorporated gates 2 to 7 and had a combined discharge of 2740 m3/s; the resulting scour hole in the model had an invert 0.28 m (14 m prototype scale) below the equivalent bed elevation. The same gate sequence was tested under a slightly lower spill discharge, resulting in a similarly deep hole (Figure 4.21). In test 7 the spill discharge was 2400 m3/s and produced a scour hole with an invert 0.26 m (13 m prototype scale) below the equivalent prototype bed elevation at the same location. In test 8 a similar discharge magnitude (2700 m3/s) was run with all gates open and resulted in a scour hole 0.20 m (10 m prototype scale) below the prototype bed elevation; this is 0.06 m (3 m prototype) shallower than in test 7 (Figure 4.22). In addition, the location of deepest scour was different for test 7 and test 8. The differences in the associated bed profile between spilling with gates 1 to 7 open and with gates 2 to 7 open is important because it clearly shows the significant change that can be made to the scour profile by making seemingly small changes to the gate openings. The two deepest scour depths are associated with the maximum spill discharge spread over the seven bays, and with a lesser spill magnitude from bays 2 to 7; yet, the location of these two scour holes is quite different. There are two factors which probably led to the spill scenario incorporating bays 2 to 7 having a high scour intensity in the side channel. The first factor is the presence of the concrete slab which extends between the right bridge abutment and pier. The slab confines the scour hole and forces the current downward which intensifies the scouring process. When the slab is removed there is a reduced scour depth for a similar discharge magnitude (Figure 4.25). However, the point of deepest scour remains in the vicinity of the right bridge 52 abutment. The second, more important factor, is the unbalanced flow which results from spilling water from bays 2 to 7, bays 3 to 7 or bays 3 to 6. An unbalanced flow condition results when flow is deflected or flipped from one side of the channel to the other. The most severe unbalanced flow condition results from spilling from bays 2 to 7, but occurs to a lesser extent when spilling from bays 3 to 7 or bays 3 to 6. The influence of the 2 to 7 open gate sequence on scour can be understood by considering individual flow patterns of discharge from the 6 bays. Discharge from bays 2 and 3, in the absence of spill from bay 1, is not deflected sideways and flows over the benches and into the channel directly upstream of the slab. The flow from 2 and 3 impacts the surface of the channel as a free jet discharge. Figure 4.8 shows how the flow from only bay 2 in the prototype travels over the benches, directly into the side channel; in this photo, the flow from bay 2 impacts the channel between the pier and benches as a semi-detached jet. The confinement by the concrete slab and bridge pier intensifies the plunging action of the free jet discharge and the scouring. Further aggravating the scour potential is the effect of the flipping of flow from the opposite side of the spillway. In the model, the discharge from bay 7, and, to a lesser extent, that from bay 6 was deflected by the angled benches into the spillway channel directly in front of the pier. In the photograph of the spill through bay 2 some of the water can be seen to flow into the main channel (Figure 4.8). Flipped discharge from the opposite side of the spillway in the model would impact just upstream of the bridge pier, thus forcing all of the flow from gates 2 and 3 through the side channel and contributing to the plunging action of the jet. The concentration of flow through the side channel is evident from the high water level which was observed locally in the side channel for the spill scenario incorporating spill from bays 2 to 7 (Figure 53 4.14). Similar but less severe unbalanced flow scenarios occur with spill scenarios where gates 3 to 7 and gates 3 to 6 are open. The effect of the unbalanced flow condition can be seen in the contour photo of test 10, where the same spill scenario was run in the absence of the pier and slab (Figure 4.24). A scour hole around the right bridge abutment, in the area where the flow is concentrated is clearly evident and suggests the presence of high scour processes. The high scouring observed in the side channel as a result of the unbalanced flow conditions is particularly interesting in light of the undermining of the pier which was discovered in 1975 (Figure 3.8). The scouring and undermining was located in the side channel, at the same location where the highest scour depths were observed in the model. Although there is no detailed information on spill discharges and gate operating sequences in the time prior to the discovery of the scour in the prototype, the model results do suggest that this area is prone to scouring. Although the scour for Test 6 was not as severe as that in Test 7 or Test 17, considering the lower magnitude of spillway discharge, it does suggest a higher scour intensity. This gate sequence, it should be noted, is not part of BC Hydro's operating order for the spillway. 4.4 Comparison of Predicted and Measured Depths Calculated scour depths, using the five equations of Chapter 3, were compared with the measured values from the model. Since all the test data was plotted, a wide variety of flow scenarios were represented. The calculated values of depth using each of the five equations were derived using input parameters from each of the tests; these are tabulated in Table 4.2. Mason's equation was calibrated to the maximum depth (7.0 m) for a spillway 54 discharge of 2280 m3/s in Test 12. Both the measured and calculated values were scaled upwards and referenced to the GSC datum of the prototype. In each of the plots of measured and calculated values, the 'Test #' id from Table 4.1 is labeled next to each of the measured data points. The results of the five scour depth equations show a high degree of variability. Damle's equation severely underestimates all the measured scour depth values. The Veronese and Mason's Calibrated Equation overestimate scour for low to moderate magnitude spill discharges but underestimate scour for large discharges (Figure 4.2). Mason's and Coleman's Equation tend to overestimate scour (Figure 4.3). Coleman's Equation tended to overestimate the scour depth by a higher degree than Mason's Equation. 55 Figure 4.2: Scour Depth Calculate Using Damle's, Veronese, and Calibrated Mason's Eq. -5 Assumed Minimum Bed Elevation of Prototype A 9 A 6 • • V 1 A 20 A Measured Model Depth x Damle's Equation X Veronese Equation • Calibrated Version of Mason's Equation Trendline of Depth From Veronese Eq. Trendline of Depth From Damle's Eq. Trendline of Measured Depth Trendline From Mason's Calibrated Eq. A 17 A 19 500 1000 1500 2000 2500 3000 3500 4000 4500 S c a l e d M o d e l D i s c h a r g e ( m / s ) 56 Figure 4.3: Scour Depth Calculate Using Coleman's and Mason's Eq -5 -10 1 o IB <3 -15 E H. a 3 o u co -20 E D E o c •2 -25 ts > UJ -30 + -35 A Measured Model Depth • Mason's Equation + Coleman's Equation Trendline of Depth From Mason's Eq. Trendline of Measured Depth Trendline of Depth From Coleman's Equation Assumed Minimum Bed Elevation of Prototype A 1 A S V A 9 \ ^ A 6 A 11 13 16 > N * 14 A A 5 N ^ A 12 A 10 A 20 X \ ° \ A 15 ' A 8 • * \ + + -m D $ 17 - . • • + it L T \ A 19 + X \ 500 1000 1500 2000 2500 3000 3500 4000 4500 Scaled Model Discharge (m/s) Both Coleman's equation and Mason's equation, in general, overestimate the depth of scour for the model for all the test scenarios. However, for Test 7 where gates 2 to 7 are open, Mason's equation underestimates slightly the amount of scour. In Test 17, for the same gate sequence, both Coleman's and Mason's equation overestimate the scour for the same open gate sequence. Applying the five scour equations to the model data suggests that only Mason's or Coleman's equation consistently over-represent the maximum scour depth in the model. 57 Mason's equation tends to ;over-represent the maximum scour to a lesser degree than Coleman's equation which suggests it is a more;accurate tool for gauging the maximum scour depth in the model and perhaps the prototype. 4.5 Application of Model Results to Analytical Study The data gathered from the model adds to the understanding of scour below Ruskin dam spillway. The depth data can be used to refine the results of the analytical study and obtain a better estimate of prototype scour. In the analytical scour study a range of scour depths between 5.4 m and 24.8 m was calculated from the five equations, assuming different spill scenarios. The two spill scenarios considered in the analytical study represented an upper and lower bound value for q, the unit discharge. In Scenario #1, all flow was assumed to be funneled by the concrete benches into a narrow width, represented by the width of the plunge pool. Observations from the model testing indicated that this spill scenario does take place. At maximum discharge the flow is deflected by the concrete benches but not into a narrow slot. Thus, all of the scour depth calculations using the model data, thus, were made using the lower bound, or Scenario #2 type conditions, where the unit discharge is measured across the width of the gate openings. If the scour depth calculations using Scenario#l type conditions are neglected from the analytical study in Chapter, the range of predicted scour depth by the five equations is 6.0 - 14.5 m. It is clear from the model results that Damle's equation grossly under-represents all scour depths. The calibrated version of Mason's equation and the Veronese equation under-represent the maximum scour depths. The results of the analytical scour study from Chapter 3 show that, of the five equations, the lower bound scour depths calculated from the prototype conditions are from Damle's and the calibrated version of Mason's equation; the model results 58 indicate that these two values underestimate the maximum scour. Given the model results, the maximum prototype scour depth calculated from Damle's equation and the calibrated version of Mason's in Chapter 3 can be discounted as underestimationing the scour. However, there is an inconsistency in the relative depth of maximum scour calculated using the Veronese equation between the prototype and model conditions if taken together with the results of Coleman's equation. Coleman's equation and Mason's equation are better predictors of scour depth in the model because they consistently over-represented all of the measured model scour depths. The scour depth calculations from the model data also show that Coleman's equation overestimates the scour depth by a greater amount than Mason's equation. In the scour study from Chapter 3, the depth calculated from Coleman's equation is less than that of Mason's equation; as well, the depth of maximum prototype scour calculated from the Veronese equation is greater than that calculated using Coleman's equation. These results are somewhat different than the values calculated from the model conditions. The difference in the relative magnitudes of scour for model and prototype conditions by Coleman's equation and the Veronese equation is likely the effect of scaling. If the lower bound values of prototype scour calculated in Chapter 3 are discounted as underestimates of prototype scour, and only Scenario#2 type conditions are considered, the range of scour predicted is between 12.9 and 14.5 m. The range takes into account the uncertainty due to the scaling of Coleman's equation and the Veronese equation. 4.6 Summary of Results The maximum scouring in the model to -28.3 m GSC represents an additional scouring at the prototype scale to 16 m below the assumed current prototype minimum bed elevation. 59 Applying results of measured scour depth from the model to the prototype is likely conservative given that, in general, scour depths are exaggerated in non-cohesive bed models. A refinement of the range of scour calculated in Chapter 3 using the results of measured and predicted values of scour from the model indicates that between 12.9 and 14.5 metres of additional scour can be expected as a result of a PMF event. The model results also indicate that the maximum scour depth will occur in the main channel at the base of the concrete bridge pier. The location of the scouring is of no significant risk to the dam foundation since it is likely that the pier itself will be undermined and washed out before the maximum depth is reached. In the absence of the bridge pier, the point of maximum scour depth moves downstream and away from the dam foundation. Although the calculation of maximum scour in the analytical study was based on a maximum flow scenario, it did not address the influence of gate sequence on scour. The model study indicates that an open gate sequence favouring bays 2-7 results in high local flows in the side channel and deep scouring along the right channel bank. With this type of unbalanced flow scenario there is a risk of erosion of the channel bank. Even with the bridge pier washed out propensity for scour along the right bank remains high. In comparison to a PMF discharge scenario, the unbalanced flow scenario is of greater concern. The probability of a discharge magnitude which produces the unbalanced flow is higher than a PMF event. As well, potential erosion to the right bank as a consequence of the unbalanced flow is less desirable than the scouring around the bridge pier and central channel associated with a PMF event. The model testing indicates that scouring in the side channel could be reduced by changing the opening sequence of gates in order to produce a more balanced flow scenario. The maximum scour depth in Test 8, with all seven gates open, was 4 metres less than in Test 60 17, with gates 2 to 7 open, for the same discharge magnitude. Currently, the operating order for Ruskin dam spillway stipulates that gate 7 be opened before gate 1. If gate 7 were opened concurrently with gate 1 a more balanced flow scenario would be produced. For this sequence a greater part of the discharged water is deflected into the main channel, as illustrated in the model. In the model tests, a spill scenario with all seven gates open results in the deepest scouring occuring the central spillway channel. With gates 2 to 7 open, the point of deepest scour occurs along the right embankment. Changing the operating order to favour the balanced discharge scenario may avoid scouring along the right channel bank for moderate to high spill magnitudes. Prior to undertaking a procedural change of the gate operations, a simple and relatively inexpensive test could be undertaken to confirm the results of the model study. Dyanmic pressure measurements along the side-channel bed could be compared for a 2 to 7 and 1 to 7 open gate sequence to test the propensity of scour in the side channel. The gates would only need to be partially opened (1 to 2 metres) to establish the flow patterns seen in the model. 61 62 63 Figure 4.6: Construction Detail of Spillway Face 6»4 Figure 4.12: Plan View of Model Spillway Discharge Through Gates 4, 5, & 6 (Q ~ 150 m3/s) &9 Figure 4.13: Model Spillway Discharge, Gates 2 to 6 Open Showing Direction of Flow Into Spillway Channel (Q = 2280 m3/s) Figure 4.14: Model Spill Discharge, Gates 2 to 7 Open Showing Direction of Flow Redirected by Benches Below Gates 6 & 7 and High Local Flows Through Side Channel (Q = 2740 m 3/s; Pond E l . = 43.8 m GSC) Figure 4.16: Test #2 Scour Patterns (Bays 4, 5, & 6; Q = 980 m3/s; Pond E l . = 41.2 m GSC) Figure 4.17: Test #3 Scour Patterns (Bays 4, 5, & 6; Q = 650 m3/s; Pond E l . = 39.7 m G S C ) Figure 4.18: Test #4 Scour Patterns (Bays 3, 4, & 5; Q = 800 m3/s; Pond E l . = 40.5 m G S C ) 1-2. Figure 4.21: Test #7 Scour Patterns (Bays 2, 3, 4, 5, 6 & 7; Q = 2400 trrVs; Pond E l . = 42.1 m GSC) Figure 4.22: Test #8 Scour Patterns (Bays 1, 2, 3, 4, 5, 6 & 7; Q = 2700 m3/s- Pond El 42.0 m G S C ) Figure 4.23: Test #9 Scour Patterns (Bays 1, 2, 3, 4, 5, 6 & 7, pier removed; Q = 2950 m3/s; Pond El. = 42.4 in GSC) Figure 4.24: Test #10 Scour Patterns (Bays 2, 3, 4, 5, 6 & 7, pier removed; Q = 2600 m3/s; Pond El. =42.5 m GSC) IS Figure 4.25: Test #11 Scour Patterns (Bays 2, 3, 4, 5 & 6, pier removed; Q = 2400 m/s : Pond E l . = 43.1 m GSC) Figure 4.27: Test #13 Scour Patterns (Bays 3, 4, 5 & 6; Q = 1850 m3/s; Pond El . = 42.9 mGSC) Figure 4.28: Test #14 Scour Patterns (Bays 2, 3, 4, 5 & 6; Q = 1950 m'Vs; Pond El. = 42.0 m GSC) 7 ? Figure 4.29: Test #15 Scour Patterns (Bays 1, 2, 3, 6 & 7; Q = 2000 m3/s; Pond El. = 42.1 mGSC) Figure 4.30: Test #16 Scour Patterns (Bays 4, 5 & 6; Q = 1350 mVs; Pond El. = 42 7 m GSC) 7 8 -M Figure 4.31: Test #17 Scour Patterns (Bays 2, 3, 4, 5, 6 & 7; Q = 2750 m3/s; Pond El . = 42.8 m GSC) Figure 4.32: Test #18 Scour Patterns (Bays 1, 2, 3, 4, 5, 6 & 7; Q = 3600 m3/s; Pond El. = 43.5 m GSC) 7 9 Figure 4.33: Test #19 Scour Patterns (Bays 1, 2, 3, 4, 5, 6 & 7; Q = 4200 m3/s; Pond E l . = 44.4 m GSC) j , .2 § s 11-- § I-a-8 I h s iff c o V 8 i •2 c ro .-E n <u 3 — o -LU -2 5 81 £ red ept 3 Q <n k_ ro Me COI CO "o . o i f S a •»-• fll g o •o — n a) o 2 o 3 g h - 111 ca u a. <u O to to. | in § C O C O ^ N . o CM" g) "= 'a. - f 6 O CO a S o ra co •5 " c E .2 tn 1 1 •I -° '2 e £ co E <2 O co CO o C3 O g —: C <D O C "D 7 ui o II ™ E c 11 £ 2 « s O. <D N « o "> s ».= 03 CO £ = O .5 OI O Q. S O CO X) C M C O IT) CO z o Z> CO CO < Chapter 5 Conclusions and Recommendations The results of the scour study indicate that the maximum scour depth will occur at the base of the bridge pier for a PMF type scenario. The combined results of the model and analytical scour study suggest the maximum depth of scour for a PMF flood scenario to be in the range of 12.9 to 14.5 m below thecurrent minimum bed elevation. The removal of the pier, in an event where it might be undermined and washed out in an extreme spill, presents no particular danger with respect to increased scour intensities in the main channel. The modelling results indicate that if the bridge pier and slab are washed out the point of maximum scour intensity moves downstream, away from the dam foundation. In addition, the maximum scour depth decreases for the same peak discharge after bridge pier and slab are removed. The model tests are inconclusive with respect to whether there is a backwater effect in the main channel as a result of extreme discharges over the spillway. The high turbulence in the spillway channel made an accurate water level reading difficult and inherently inaccurate. A significant local increase in flow through the side channel, between the bridge pier and right abutment, was observed in the model when gates 2 to 7 were open. The increase in flow through the side channel occurs as the result of an unbalanced flow for this particular gate opening. The local flows in the side channel cause relatively deep scouring, compared to the maximum scour depth. The model testing indicates that a spill discharge evenly distributed over seven bays produces less scouring for the same magnitude spill distributed over bays 2 to 7. In light of this, it is recommended that modification of the operating order for Ruskin spillway be considered to avoid any unbalanced flow condition which results in 82 scouring in the side channel. An operating procedure which requires gates 1 and 7 to be opened concurrently rather than consecutively is preferred. Opening gates 1 to 7 concurrently redistributes the flow and moves the point of highest scour intensity into the main channel, away from the channel banks. Prior to modifying the operating order, a study of the correlation between model and prototype could be undertaken to confirm the validity of the conclusion of the modelling results. Pressure tranducers could be attached to the channel rock bed at 2 or 3 points, including one on the side-channel bed to acquire pressure data as an indicator of scour propensity for different gate openings. A comparison of the prototype pressure data for two spill scenarios, one with gates 2 to 7 open and the second with all 7 gates open should indicate higher average and maximum dynamic pressure fluctuations in the side channel for the 2 to 7 open gate sequence. This spill test could be undertaken for partial but equal gate openings for each of the scenarios. The spill duration would only need to be long enough to establish flow patterns in the channel and to collect statistically representative data. This test, and the data acquired, would provide a useful insight into scour and would affirm the accuracy of the hydraulic model. The approach of this scour study was to integrate information from equations which predict scour depth with information taken from a physical hydraulic model. The equations gave an estimated range of scour depth but gave no information on where the highest scour might be located. In addition, there was uncertainty with respect to how well the equations could represent the scour given the complex flow patterns which are generated by the interaction of flow down the spillway with the concrete benches and bridge pier. The scour study has shown that Mason's and Coleman's equation reasonably represented the maximum scour depth in the model, whereas Damle's equation, the calibrated version of Mason's 83 equation and the Veronese equation tended to underestimate the maximum scour. Eliminating the values of prototype scour calculated by Damle's equation and the calibrated version of Mason's equation gave a more refined and more certain range of maximum scour depth. In addition, the model study was able to capture conditions of high scour intensity associated with a less-than-maximum discharge; this is something which the equations could not have been expected to predict given the interaction of the flow with the concrete benches and the presence of the concrete slab in the side channel. 84 Works Cited and Extended Bibliography Abry P., Fauve S. et al (1994). "Analysis of Pressure Fluctuations in Swirling Turbulent Flows", Journal De Physique IL May, pp725-733. Ahmad, Mushtaq (1988). "Tarbela dam Project: Model-Proto Comparison of Scour In Plunge Pool" in Model Prototype Correlations of Hydraulic Structures. ASAE, New York, New York: 1988. Akhmedov, T.Kh. (1968). "Local Erosion of Fissured Rock at the Downstream End of Spillways", Hydrotechnical Construction. No.9, pp. 821-824. Akhmedov, T.Kh. (1988). "Calculation of the Depth of Scour in Rock Downstream of a Spillway", Water Power and Dam Construction. Vol.40, No. 12, December, pp. 25-27. Annandale, G.W. (1995). "Erodibility", Journal of Hydraulic Research. Vol.33, No.4, pp. 471-493. Bellin, A & Fiorotto, V. (1995). "Direct Dynamic Force Measurement of Slabs in Spillway", Journal of Hydraulic Engineering. Vol. 121, No. 10, October, pp. 686-693. Annandale, G.W. (1995). "Erodibility", Journal of Hydraulic Research. Vol.33, No.4, pp. 471-493. BC Hydro, "Report No. MEP182: Summary of Information For Advisory Board Meeting No.l", March, 1996. BC Hydro, "S.0.0 4P-47: System Operating Order for Ruskin Project", January 27, 1997. BC Hydro, "Report H1837: Ruskin Dam 1984-85 Geotechnical Investigations", June 1985. BC Hydro, "Report H1862: Ruskin Project Civil Inspection Report (CIR)", 1985. BC Hydro, "Report MEP188: Ruskin Dam Comprehensive Inspection and Review, 1995", March 1996. Bell, W.W., "Ultimate Scour Depth For the Ruskin Dam Plunge Pool". Memorandum to Alec De Rham, 4 July, 1996. Bellin, A & Fiorotto, V. (1995). "Direct Dynamic Force Measurement of Slabs in Spillway", Journal of Hydraulic Engineering. Vol. 121, No. 10, October, pp. 686-693. Blake, William K. (1986). Mechanics of Flow-Induced Sound and Vibration. Academic Press Inc., Toronto: 1986. Blevins, Robert D. (1977). Flow Induced Vibration. Von Nostrand Reinhold Company, Toronto: 1977. 85 Bowers, Edward C. & Toso, Joel (1988). "Karnafuli Project, Model Studies of Spillway Damage", Journal of Hydraulic Engineering, Vol.114, No.5, May. Burgi, Phillip H. ed. Model-Prototype Correlation of Hydraulic Structures, ASAE, New York, New York: 1988. Cola, Raffaele (1965). "Energy Dissipation of a High-Velocity Vertical Jet Entering a Basin", Proc. IAHR Conference No. XL Leningrad, USSR, Paper 1.52, Vol. 1. Dhillon, G.S. et al. (1982). "Behaviour of Pandoh Dam Spillway: A Field-Cum-Model Study", Proceedings from International Conference on the Hydraulic Modelling of Civil Engineering Structures, Conventry, England: pp. 539-552. Doddiah, D., et al (1953). "Scour From Jets", 5th Congress of the IAHR, Minneapolis, Minn., pp. 161-169. Elder, R.A. (1961). "Model-Prototype Turbulence Scaling", Proceedings of the 9th IAHR Convention, Dubrovnik,Yogoslavia: pp.24-31. Ervine, D.A. (1976), "The Entrainment of Air in Water", International Water Power and Dam Construction, 28(12), pp. 27-30. Ervine, D.A. & Falvey, H.T. (1987). "Behaviour of Turbulent Water Jets in the Atmosphere and In Plunge Pools", Proc. Instn Civil Engineers, Part 2, 83March, pp. 295-314. Ervine, D.A., Falvey, H.T. (1997). & Withers W. "Pressure Fluctuations on Plunge Pool Floors", Journal of Hydraulic Research, Vol.35, 1997, No.2, pp. 257-279. Falvey, Henry T. and Ervine, Alan (1988). "Aeration in Jets and High Velocity Flows", in Model-Prototype Correlation of Hydraulic Structures, ASAE, New York, New York: 1988. Fauve, S., et al (1993). "Pressure Fluctuations in Swirling Turbulent Flows', Journal De Physique II March, pp. 271-278. Fiorotto V. & Rinaldo, A. (1992). "Turbulent Pessure Fluctuations Under Hydraulic Jumps", Journal of Hydraulic Research, Vol. 30, No. 4, pp. 499-520. Fiorotto V. & Rinaldo, A. (1992). "Fluctuating Uplift and Lining Design in Spillway Stilling Basins", Journal of Hydraulic Engineering, Vol.118, No.4, April, pp. 578-597. ICOLD Bulletin 58 (1987) "Spillways for Dams," pp. 75-93. King,' D.L. (1967). "Analysis of Random Pressure Fluctuations in Stilling Basins", Proceedings of the 12th Conference of IAHR, Fort Collins, Colorado, USA, Vol.2: pp.210-217. 86 Kobus, H., Leister, P. & Westrich, B. (1979). "Flow Fields and Scouring Effects of Steady and Pulsating Jets Impinging on a Movable Bed", Journal of Hydraulic Research. Vol.17, No. 2, pp. 175-192. Lazenby, F.A. "Report on Design and Construction Features of Ruskin Dam", May 1, 1940. Lesleighter, Eric J. (1971) "Pressure Fluctuations in Large Energy Dissipation Structures". Proceedings of the Fourth Australasian Conference on Hydraulics and Fluid Mechanics. Melbourne, Australia: pp. 434-442. Lopardo, Raul A. (1988). "Stilling Basin Pressure Fluctuations", in Model-Prototype Correlations of Hydraulic Structures, ASAE, New York, New York: 1988. Martin, H.M., Wagner, W.E. (1961). "Experience in Turbulence in Hydraulic Structures", Proceedings of the 9th IAHR Convention, Dubrovnik, Yugoslavia: pp. 153-155. Mason, P.J. (1993). "Practical Guidelines for the Design of Flip Buckets and Plunge Pools", Water Power and Dam Construction, Sept/Oct, pp.40-45. Mason, P.J. & Arumugam, K. (1985). "Free Jet Scour Below Dams and Flip Buckets", Journal of Hydraulic Engineering, Vol. 11, No. 2, Feb., pp. 220-235. Mason, P.J. (1986). Reply to Kerrin Spurr, Water Power and Dam Construction, Feb. Mason, P.J. (1989). "Effects of Air Entrainment on Plunge Pool Scour", Journal of Hydraulic Engineering, Vol.115, No.3, March, pp. 385-399. Mason, P.J. "Erosion of Plunge Pools Downstream of Dams Due to the Action of Free-Trajectory Jets", Proc. Instn Civil Engineers, Part I, May, 1984, pp.523-537. Otto, B. (1986). "Study of Scour Potential at Burdekin Falls Dam Due to Rock Stresses". Queensland Water Resources Commission. Robinson, Kerry M. (1989). "Hydraulic Stresses on an Overfall Boundary", Transactions of the ASAE, Vol. 32,(4): July-August, 1989, pp.1269-1274. Spurr, K.J.W. (1985). "Energy Approach to Estimating Scour Downstream of a Large Dam", Water Power and Dam Construction, July, pp. 81-89. U.S. Bureau of Reclamation, Design of Small Dams, Oxford & IBH Publishing Co., New Delhi: 1977 (p. 410). Vasiliev, O.F. and Bukreyev, V.I. "Statistical Characteristics of Pressure Fluctuations in the Region of Hydraulic Jump", Proceedings 12th Congress IAHR, Fort Collins, Colorado, Vol. 2, p. 1-8. 87 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items