Complementary Methods for Determining the Sedimentation and Flushing in a Reservoir

Castillo, L. G.; Carrillo, J. M.; Álvarez, M. A.

doi:10.1061/(ASCE)HY.1943-7900.0001050

Open access

Case Studies

Jun 8, 2015

Complementary Methods for Determining the Sedimentation and Flushing in a Reservoir

Authors: L. G. Castillo [email protected], J. M. Carrillo [email protected], and M. A. Álvarez [email protected]Author Affiliations

Publication: Journal of Hydraulic Engineering

Volume 141, Issue 11

https://doi.org/10.1061/(ASCE)HY.1943-7900.0001050

PDF

Abstract

This study analyzes the changes expected in the Paute River in Ecuador as a result of the future construction of the Paute-Cardenillo Dam (owned by Celec Ep-Hidropaute). The project must remain in use throughout its projected useful life. For this reason, the operational rules at the reservoir need to consider sedimentation effects. Four complementary methods are used to study the sediment transport and flushing. Empirical formulas and one-dimentional (1D) simulations are used to estimate sedimentation in the reservoir. Two-dimensional simulations allow the analysis of a 72-h flushing operation in the reservoir. Three-dimensional simulations show the detail of the sediment transport through bottom outlets, where the effect of increasing the roughness due to the sediment transport in the bottom outlets is considered. The results demonstrate the utility of crossing different methods to achieve adequate resolution in the calculation of sedimentation and flushing operation in reservoirs.

Main Characteristics of the Project

The Dam is in the Paute River basin in Ecuador, 23.128 km downstream from the Amaluza Dam. The area to analyze drains is approximately

275 {km}^{2}

of the surface area, and the average slope of the river reach is

0.05 m / m

.

The Paute-Cardenillo is a double-curvature arch dam with a maximum height of 135 m to the foundations. Considering as the reference the Universal Transverse Mercator (UTM) system with the projection PSAD 56 (Provisional South American Datum 1956) corresponding to area 17 of the Southern Hemisphere, the top level is 926 m above mean sea level (MASL), whereas the normal maximum water level is 924 MASL. The reservoir has a length of 2.98 km along the thalweg of the underlying river channel when the water level is at 924 MASL.

The estimated total bed load was

1.75 {Mm}^{3} / year

and the maximum volume of the reservoir

12.33 {Mm}^{3}

. In order to prevent the deposition of sediments into the reservoir, the owner proposes periodic discharges of bottom outlets or flushing. These operations should be able to remove the sediments, avoiding the advance of the delta from the tail of the reservoir.

The river bed material is characterized by independently considering the finer and coarser materials. The points count method allowed us to characterize the coarsest materials (

> 75 mm

) of the Paute River (Fig. 1). Fig. 2 shows the field work in a selected area of

6.47 \times 5.70 m^{2}

carried out to obtain the most representative sieve curve (Fig. 3). A mesh of 0.50 m is used for the processing of images. Thus, 156 subsectors (Consorcio PCA 2012) are identified.

Fig. 1. Paute river in the Cardenillo Dam; transversal section and principal parameters for $Q = 540 m^{3} / s$

Fig. 2. Field work for obtaining the sieve curve; mesh size of $0.50 \times 0.50 m$

Fig. 3. Sieve curve of the Paute-Cardenillo river near to the site of the dam

According to the feasibility study carried out by Consorcio PCA (2012), the minimum flow evacuated by the bottom outlet to achieve an efficient flushing should be at least twice the annual mean flow (

Q_{m a} = 136.3 m^{3} / s

, corresponding to the subbasin between Amaluza Dam and Paute-Cardenillo Dam). Based on the results of this study, the owner adopted a conservatively high flow of

408.9 m^{3} / s

(

3 Q_{m a}

).

Flushing Considerations

Great amounts of sediments have accumulated in numerous reservoirs, thus reducing their performance. The problem most commonly occurs in small-and medium-sized reservoirs, where the contributing rivers have high sediment transport. The decreased reservoir volume shortens the lifetime of the reservoir and thereby reduces the economic value of the project.

Based on field and laboratory data, some authors have proposed several empirical formulas to estimate sediment outflow discharge that may be related to flow discharge, energy slope, or channel width. Under appropriate conditions, researchers consider that hydraulic flushing can remove fine sediments as well as coarse sediments. Other authors state that hydraulic flushing methods are not very effective because they are limited to the amount of sediments close to the outlet, even with considerable water discharges. It is therefore necessary to investigate the effectiveness of flushing operations.

The effect of the flushing varies greatly in different reservoirs. There are no general rules that can be extrapolated from one reservoir to another; however, Shen (1999) highlighted some ideas to consider:

•

Water level in reservoir should be drawn down to improve the efficiency of flushing;

•

Flushing sediment is more efficient in narrow reservoirs than wide reservoirs;

•

For wide reservoirs, a distinct flushing channel is formed and retrogressive erosion occurs mainly inside this flushing channel; and

•

The width of the flushing channel was found to be a coefficient of about 11 to 12 times the square root of the bankfull discharge inside the flushing channel. This has been found both in real conditions using data from four different reservoirs (Atkinson 1996) and in experimental laboratory conditions (Lai and Shen 1996; Janssen 1999).

Several authors indicate that the efficiency of the hydraulic flushing depends on the relationship between the storage volume of the reservoir and the annual amount of incoming runoff. Annandale (1987) indicates that the flushing is effective if this ratio is less than 0.02, whereas Basson and Rooseboom (1996) raised this threshold to 0.05. The Cardenillo Reservoir ratio is about 0.003. Hence, an effective flushing process can be expected.

Methodology

The sediment transport is divided into the following: wash load (very fine material transported in suspension) and total bed transport (bed sediment transported plus suspension bed sediment transport, depending on the sediment size and flow velocity). The main properties of sediment are particle size, shape, density, fall velocity, porosity, and concentration, whereas the main properties of sediment transport are the interrelated hydraulic variables such as discharge, bed slope, flow area, velocity, turbulence, and hydraulic roughness.

In this paper, four complementary methodologies have been used to analyze the flushing process and the time required between flushing operations.

Some formulas have been applied for estimating the sediment transport capacity. The mean section and slope have been considered in the analyzed river reach. These values are an upper envelope of the sediment transport obtained at the site of the dam when the sedimentation and erosion processes are considered.

Taking into account the entire river reach, the time required for the sediment to reach the bottom outlet level of the future Cardenillo Dam was analyzed. Different water levels in the reservoir were studied. Sedimentation in the reservoir is a long process that may require simulations of a long period of time. In this particular case, the reservoir sedimentation (shape of the delta of sediments) has been simulated with a one-dimensional (1D) model due to the large length of the river reach (23.128 km).

The flushing of the sediments involves several complex processes. The water flow field is three-dimensional (3D), with recirculation zones and secondary flows. The erosion and sediment transport processes are also complex. Sediments are transported as bed load and suspended load. The bed form may increase the roughness. As the water level is drawn down during the flushing process, there will be time-dependent changes in water level, bed level and the other processes described previously. The geometry in the Cardenillo Reservoir is complex, so the decision was made to analyze the flushing operation by two-dimensional (2D) numerical models. The details of the flushing process through the bottom outlets were also analyzed with 3D numerical models.

Estimate of the Manning Resistance Coefficient

Following the methodology applied in Castillo (2004), four aspects were checked to determine the hydraulic characteristics of the flow: macro roughness, bed form resistance, hyperconcentrated flow, and bed armoring phenomenon.

The resistance coefficients due to the grain influence were estimated by means of 10 formulas: Strickler (1923), Limerinos (1970), Jarret (1984), Bathurst (1985), van Rijn (1987), Fuentes and Aguirre-Pe (1990), García-Flores (1996), Grant (1997), Aguirre-Pe et al. (2000), and Bathurst (2002). For each analyzed flow, calculations were carried out by the adjustment between the hydraulic characteristics of the river reach (mean section and slope) and the mean roughness coefficients and, finally, the different sediment transport formulas were considered.

To estimate the grain roughness, only those formulas whose values fall in the range of the mean value– one standard deviation of the Manning resistance coefficient of all formulas at each specific flow rate were considered. The formulas that fall out of this range are discarded. The process was repeated twice. Fig. 4 shows the final result. It can be observed that the mean grain roughness is between 0.045 and 0.038.

Fig. 4. Manning resistance coefficients of the grain sediment

A macro roughness behavior was identified in all the flows analyzed. This leads to a significant increase in the various Manning coefficients. Therefore, the total flow resistance can be estimated by means of (Palt 2001; Rickenmann 2001; Wang et al. 2013)

n_{tot} = 10 S^{0.36} \sqrt{n_{s}^{2} + n_{b}^{2}}

(1)

where

S

is the bed slope,

n_{s} = 0.01

the Manning resistance coefficient of prismatic glass flume, and

n_{b}

the resistance caused by protrusions on the river bed, obtained as follows:

n_{b} = α [{(1 - B)}^{β} - 1]

(2)

where

B

is the blockage factor (relation between area of protrusions and total area of the cross section). The coefficients

α

and

β

must be evaluated by experimental and field data with different protrusions, including sediment of various sizes and vegetation. According to Wang et al. (2013), these coefficients are

α = B^{- 0.8}

and

β = - 0.15 B

. Table 1 shows the total flow resistance obtained for different flows.

Table 1. Total Flow Resistance for Different Flows

$Q (m^{3} / s)$	$h$ (m)	$B$	$α$	$β$	$n_{tot} (m^{1 / 2} / s)$
136.30	1.64	0.306	2.581	$- 0.0458$	0.152
400.00	2.69	0.261	2.932	$- 0.0391$	0.123
540.00	3.03	0.248	3.052	$- 0.0372$	0.116
820.00	3.54	0.217	3.391	$- 0.0326$	0.099
1,180.00	4.15	0.205	3.555	$- 0.0307$	0.092
2,340.00	5.46	0.170	4.119	$- 0.0256$	0.075
3,790.00	6.46	0.147	4.636	$- 0.0220$	0.065
5,520.00	7.29	0.132	5.063	$- 0.0198$	0.059

Considering the water levels measured in the field for the flows 136, 540, and

820 m^{3} / s

(Consorcio PCA 2012), the flood plain and the main channel resistance coefficients have been obtained by means of the calibration of the 1D program HEC-RAS (U.S. Army Corps of Engineers 2010). After obtaining these coefficients, the values corresponding to higher flows were calculated.

Fig. 5 shows the Manning coefficients for the different flow rates. The different values for the flood plain, main channel, and global section can be observed. Furthermore, the sensitivity analysis with respect to a variation of

\pm 0.03

in the mean blockage of the mean section can be seen. Finally, the grain resistance of the global section is included.

Fig. 5. Manning resistance coefficients in the main channel and floodplain

Estimation of Sediment Transport

Fourteen formulations of sediment transport capacity were used: Meyer-Peter and Müller (1948), Brown (1950), Einstein and Barbarrosa (1952), Colby (1964), Engelund and Hansen (1967), Yang (1984), Yang (1996), Parker et al. (1982), Smart and Jaeggi (1983), Mizuyama and Shimohigashi (1985), van Rijn (1987), Bathurst et al. (1987), Ackers and White (1973), Aguirre-Pe et al. (2000), and Yang (2005). From these, the formulas selected fell within a range of the mean value—one standard deviation of the sediment transport capacity of all formulas at each specific flow rate. The formulas that fall out of this range were discarded. The process was repeated twice.

Fig. 6 shows the sediment transport capacity and the result of the numerical simulations at the site of the dam. The mean capacity varies from

4 t / s

for the flow

345 m^{3} / s

to

130 t / s

for the flow

5,730 m^{3} / s

. The Yang (1984) results tend to be higher than the Meyer-Peter and Müller (1948) and the Meyer-Peter and Müller corrected by Wong and Parker (2006).

Fig. 6. Sediment transport capacity (reach mean values) and net sediment transport (sedimentation and erosion processes simulation of all reach) in the site of the dam

The sedimentation and erosion process of the complete river reach (23.128 km) was simulated. The formulas applied were those of Yang (1984) and Meyer-Peter and Müller (1948) corrected by Wong and Parker (2006). The Yang results are lower than their corresponding sediment transport capacity. This difference indicates that the deposition process is bigger than the erosion process. It can be observed that the results with Meyer-Peter and Muller corrected are lower than the Yang values. This is in agreement with the fact that Meyer-Peter and Müller only considers the bed load. The total bed load (bed load plus suspension bed load) obtained with Yang varies from

0.4 t / s

for the flow

345 m^{3} / s

to

4.3 t / s

for the flow

5,730 m^{3} / s

.

Reservoir Sedimentation

The bed level change

Z_{b}

was calculated from the overall mass balance equation for the bed load sediment or Exner (1919, 1925) equation

(1 - p) \frac{\partial Z_{b}}{\partial t} + \frac{\partial Q_{b x}}{\partial x} + \frac{\partial Q_{b y}}{\partial y} = D - E

(3)

where

p

is the porosity of the sediment that forms the bed layer, while

Q_{b x}

and

Q_{b y}

are the two solid flow components. The difference

D - E

is the balance between the bed load and suspended load.

The time required for the sediment level to reach the height of the bottom outlets (elevation 827 MASL) operating at reservoir levels was analyzed. Simulations were carried out with the 1D HEC-RAS v4.1 (U.S. Army Corps of Engineers 2010) program.

Fig. 7 shows the initial and final ground profile between the Amaluza and Paute-Cardillo dams. The input flows are the annual mean flow (

Q_{m a} = 136.3 m^{3} / s

) equally distributed in the first 23.128 km and the incorporation of the annual mean discharge flow of the Sopladora hydroelectric power plant (

Q_{m a_sop} = 209.0 m^{3} / s

). Sopladora discharge came from two dams located upstream of the analyzed river reach. Hence, the inflow condition into the reservoir is a constant flow of

345.3 m^{3} / s

.

Fig. 7. Initial and final ground profile and input flows

The simulations were carried out using the complete sieve curve of the river reach. Fig. 7 also shows the results of erosion and sedimentation areas computed using the Yang (1984) formula and the Meyer-Peter and Müller (1948) formula, corrected by Wong and Parker’s (2006), formulas. Due to the fact that the differences with the initial channel invert elevation are not easy to distinguish with the scale length used, the deposition and erosion areas have been increased by a factor of 10. This change in the calculated values is only valid to enable the reader to visualize the location of the erosion-deposition areas.

The corrected Meyer-Peter and Müller (1948) formula shows some numerical oscillations in the 26 km analyzed. The overall estimates of inflowing load at the upper limit of the reservoir seem reasonable when compared with empirical methods and to the Yang (1984) simulation. For that reason, the numerical oscillations in the model were subsequently ignored.

The suspended sediment concentration calculated at the inlet section of the reservoir was

0.258 kg / m^{3}

. This value takes into account the sediment transport of the upstream river and the annual average of sediment concentration from the Sopladora hydroelectric power plant.

Furthermore, the effect of different return period flood flows in the sedimentation process and in the sediment transport was analyzed. The unitary hydrograph pattern has a duration of 45 h. The correspondent hydrographs have been added to the annual mean flow.

Table 2 shows the total volume of sediments deposited in the reservoir based on numerical simulation when the sediment deposition reaches the bottom outlet level (827 MASL). The two transport equations were examined, as well as various water levels in the reservoir based on future operations at the dam, e.g., 860 MASL is the water level considered to operate the bottom outlets, 920 MASL the pondered average water level, and 924 MASL the normal upper storage level (Fig. 8).

Table 2. Time Required and Volume of Sediment to Reach the Bottom Outlets

Reservoir elevation (MASL)	Yang (1984)		Meyer-Peter and Müller (1948)
Reservoir elevation (MASL)	Required time (years)	Sediment volume ( ${hm}^{3}$ )	Required time (years)	Sediment volume ( ${hm}^{3}$ )
860	0.35	0.65	0.32	1.47
892	5.10	2.77	2.58	3.86
918	12.90	6.07	8.80	7.34
920	13.60	6.33	9.50	7.64
924	14.80	6.62	10.90	8.97

Fig. 8. Scheme of the dam, water levels and delta of sediment into the reservoir in the initial condition of the flushing

According to the results, the volume of sediment in the reservoir rises as the water level in the reservoir increases, and requires a longer duration to reach the bottom outlet elevation. The Meyer-Peter and Müller (1948) expression and the level of the reservoir located at 860 MASL obtained the biggest sedimentation volume (

1.47 {hm}^{3}

) and the smallest time to reach the level of the bottom outlets (3 months and 27 days), whose invert elevation is located at 827 MASL.

Flushing Simulation

Two-Dimensional Simulation

With the 1D model, the flushing during the initial emptying of the reservoir (HEC-RAS simulates the sediment transport by using quasi-unsteady flow data) could not be simulated. The regressive erosion of the delta of sediments does not clearly appear during the flushing operation time (72 h according to the future operations at the dam). Furthermore, 1D modeling cannot show the spatial patterns of 2D erosion in the reservoir.

The flushing process was analyzed using the Iber v1.9 (Iberaula 2013) 2D program. Iber uses an unstructured mesh and finite volume scheme. The hydrodynamic module solves shallow water equations (2D Saint-Venant equations). The

k - ε

turbulence model was considered. The sediment transport module solves the transport equations by the Meyer-Peter and Müller (1948) expression and the evolution of the bottom elevation is calculated by sediment mass balance.

The evolution of the flushing over a continuous period of 72 h was studied, according to the operational rules at the Paute-Cardenillo Dam. The initial condition of the sedimentation profile (the lower level of the bottom outlet) was

1.47 {hm}^{3}

of sediment. The suspended sediment concentration at the inlet section was

0.258 kg / m^{3}

. In accordance with the future operations at the dam, the initial water level at the reservoir was 860 MASL. The flushing process was observed in the reservoir during the operation of the bottom outlets. Fig. 9 shows the evolution of the profiles of the sediment during the flushing operation.

Fig. 9. Evolution of the level of sediments during a flushing period of 72 h with the 2D model

After a flushing period of 72 h, the sediment volume removed from the bottom was

1.77 {hm}^{3}

. This value is greater than the initial sediment value (

1.47 {hm}^{3}

).

Fig. 10 shows the transversal profiles of the reservoir bottom before and after the flushing operation. The mean width of the flushing channel is over 220 m, which is about 11 times the square root of the flushing discharge (Shen 1999).

Fig. 10. Level of sediments before and after a flushing period of 72 h with the 2D model

Three-Dimensional Simulation

Two-dimensional simulation might not properly simulate the instabilities of the delta of sediments that could block the bottom outlets nor the flow pressurized in the bottom outlets that occurs in the first steps of the flushing. A 3D simulation could clarify the doubts during the first steps of the flushing operation.

The computational fluid dynamics (CFD) program FLOW-3D v11.0 (Flow Sciences 2011) was used, which solves the Navier-Stokes equations discretized by finite differences. It incorporates various turbulence models, a sediment transport model and an empirical model bed erosion (Mastbergen and Von den Berg 2003; Brethour and Burnham 2010), together with a method for calculating the free surface of the fluid (Hirt and Nichols 1981). The bed load transport was calculated using the Meyer-Peter and Müller (1948) expression. The closure of the Navier-Stokes equations was calculated with the Re-Normalisation Group (RNG) k-epsilon model turbulence model (Yakhot and Smith 1992).

Due to the high-capacity equipment requirements and long simulation times to calculate the flushing of all the reservoir during 72 h, the results are focused in the first 5,000 s. the initial conditions of the delta of sediments was obtained from the HEC-RAS program. The inlet boundary was considered 150 m upstream of the dam, considering that the water level was 860 MASL in the initial condition. The total number of cells used is 432,090. The mesh is made of hexahedral elements of 1 m length scale near the dam and 2 m in the reservoir.

Due to the high concentration of sediments that pass through the bottom outlets, it is recommended to consider the variation of the roughness in the bottom outlets. The formula proposed by Nalluri and Kithsiri (1992) was used to estimate the hyperconcentrated flow resistance coefficient on rigid bed

λ s s = 0.851 λ_{0}^{0.86} C_{v}^{0.04} D_{g r}^{0.03}

(4)

where

λ_{s s}

is the Darcy-Weisbach’s resistance factor on rigid bed with sediment transport,

λ_{0}

the Darcy-Weisbach’s resistance factor on rigid bed with clear water,

C_{v}

the volumetric sediment concentration, and

D_{g r}

the grain size nondimensional factor.

The bottom outlets are four rectangular ducts (

5.00 \times 6.80 m

). The slope corresponding to the stretch under study is

S = 0.001

. The

C_{v}

estimated is 0.04. Wan and Wang (1994) consider that the energy supplied by the solid phase for a volume unit and a distance unit downstream (in a nondimensional way) is

\frac{E_{d}}{γ} = C_{v} S = 0.00014

(5)

This value is considerably lower than the limit between the hyperconcentrated flow and mudflow (0.004). The coefficient of the kinematic viscosity of water with sediments concentration (

v_{s}

) was estimated using the following formula (Graf 1984):

\frac{ν_{s}}{ν} = 1 + K_{e} C_{v} + K_{2} C_{v}^{2}

(6)

where

ν

is the coefficient of kinematic viscosity of clear water (for

T = 15 ° C

,

ν ≅ 1.14 \times 10^{- 6} m^{2} / s

),

K_{e}

the Einstein viscosity constant (

≅ 2.5

), and

K_{2}

the particles interaction coefficient (

≅ 2

). Therefore,

ν_{s} = 1.26 \times 10^{- 6} m^{2} / s

, a value 10.5% higher than

ν

. The nondimensional size of the grain corresponding to

D_{50} = 0.150 m

is

D_{g r} = D_{50} {[\frac{(s - 1) g}{ν_{s}^{2}}]}^{1 / 3} = 3255

(7)

Ducts are covered by iron, so the absolute roughness

ε = 5 \times 10^{- 5} m

. With the Coolebrok-White formula,

λ_{0} = 8.32 \times 10^{- 3} (n_{0} = 0.0148)

has been obtained. Using the Nalluri and Kithsiri formula,

λ_{s f} = 1.55 \times 10^{- 2} (n_{s f} = 0.0203)

. This value corresponds to an absolute roughness

ε_{s f} = 2.37 \times 10^{- 3} m

. This value was used in the 3D simulations.

Fig. 11 shows the total discharge flow of the four bottom outlets for the first 5,000 s of simulation. Different roughness was considered in the ducts depending if clear water and water with sediments is being transported. Outlets worked in a pressured and unsteady regime at the initial emptying of the reservoir, reaching a discharge near

2,700 m^{3} / s

in both cases. After reaching the steady regime (around 500 s of simulation), there was a free surface flow and the discharged flow was as expected (

408.90 m^{3} / s

during the flushing operation). With the smooth roughness, the amount of solid removed during the flushing process is higher than with the high roughness. This is due to the occlusion of the bottom outlets in the first steps of the simulation. The simulation with high roughness tends to occlude the bottom outlets, not only in the first steps, but also in the rest of the simulation (there are intermittent occlusion and nonocclusion processes). This shows the great complexity and variability of the hydraulic behavior during the flushing process. After a determined time, simulations seem to reach a constant rate of sediments removed.

Fig. 11. Liquid and solid flows considering smooth and high roughness during the flushing operation simulated with the 3D program

Table 3 shows the mean liquid flow through the bottom outlets of the first 5,000 s of simulation. Bottom outlets in the middle tend to have the higher discharge values. This is mainly due to the nonalignment of the bottom outlets with respect to the main channel (Fig. 12).

Table 3. Mean Flow Water per Each Bottom Outlet after 5,000 s of Simulation

Surface roughness	Liquid flow though bottom outlet ( $m^{3} / s$ )
Surface roughness	1	2	3	4	Total
Smooth roughness ( $n = 0.0148$ )	88.60	104.96	128.74	86.59	408.90
High roughness ( $n = 0.0203$ )	94.34	101.61	122.00	90.95	408.90

Fig. 12. Velocity vectors thought the bottom outlets

Fig. 13 shows the volume of sediment removed and the transient sediment transport during the first hours of operation. As in the 2D simulation, there is significant sediment transport at the beginning of the simulation. There is a maximum of

130 m^{3} / s

of sediments rate near the second hour for the high roughness 3D simulation. Later, the sediment transport rate tends to decrease to values similar to the 2D result. The total volume of sediment calculated by FLOW-3D is higher than with the Iber program, since the simulations of the flushing phenomenon are very different in the first hours.

Fig. 13. Comparison of the flushing operation simulated with 2D and 3D simulations

The 2D simulations considered that all the sediments (

1.47 {hm}^{3}

) may be removed in 60 h. Considering that the rate of sediments would continue during the rest of the flushing operation, the 3D simulations with the high roughness would require 54 h to remove all the sediments. Hence, the Iber simulations err on the side of caution.

Conclusions

In this paper, the complex phenomenon of sedimentation and flushing was analyzed by using four interrelated methodologies: empirical formulations, 1D simulations, 2D simulations, and 3D simulations.

Sediment transport capacity constitutes an upper envelope of the erosion and sedimentation processes. Calculating this allowed for the estimation of the coefficients of resistance as a function of the grain influence. There is a high increment of the roughness due to the blockage effect. Those values were applied in the numerical simulations.

The river reach (23.128 km) sediment transport and the reservoir sedimentation simulations were carried out with a 1D program. Two- and three-dimensional programs need higher-capacity equipment and longer simulation times. The flushing operation was simulated with two-dimensional (Iber) and three-dimensional (FLOW-3D) programs. For 72 h of flushing simulation, the Iber program required nearly 24 h (Intel Core i7 CPU, 3.40 GHz processor, 16 GB RAM and 8 cores, Santa Clara, California).

The FLOW-3D program, using the same equipment, would require more than 1,200 h (50 days) to solve the complete reservoir. Hence, only the three-dimensional simulations were used to analyze the behavior of the flow in the first steps of the flushing near the dam. The increase of the roughness in the ducts drives to a reduction of the amount of sediment removed in the reservoir.

Suspended fine sediments in the reservoir can result in certain cohesion of the deposited sediments, which might influence the flushing procedure. Carrying out a flushing operation every 4 months, the cohesion effect in increasing the shear stress will be avoided.

Designers must take into account the high degrees of uncertainty inherent in sediment transport (numerical modeling and empirical formulae). Sensitivity analysis must be performed to prove the models are robust to various inputs and not only to one single scenario. Some parameters need to be considered to reduce this uncertainty in numerical modeling, including the following: turbulence parameters; estimated volumetric sediment concentration; and sensitivity to roughness, among others. In the near future, a reduced physical model will be built that will allow for checking and calibrating the main parameters of the numerical model.

The results demonstrated the suitability of crossing different methodologies to achieve an adequate resolution of complex phenomena such as flushing operations. Thus, numerical simulations of differing degrees of complexity were used to complement the classical formulations and allow a better understanding of the physical phenomena.

Notation

The following symbols are used in this paper:

$B$: blockage factor;
$C_{v}$: volumetric sediment concentration;
$D$: characteristic diameter;
$D_{g r}$: grain size non-dimensional factor;
$E_{d}$: energy provided by the solid fraction;
$K_{2}$: particles interaction coefficient;
$K_{e}$: Einstein viscosity constant;
$n_{0}$: resistance coefficient in the outlet with clear water;
$n_{b}$: resistance coefficient caused by protrusions on river bed;
$n_{c}$: resistance coefficient in the main channel;
$n_{p}$: resistance coefficient in the flood plain;
$n_{s}$: resistance coefficient of prismatic glass flume;
$n_{s f}$: resistance coefficient in the outlet with sediment transport;
$n_{tot}$: total flow resistance coefficient in the river;
$p$: porosity of the sediment that forms the bed layer;
$Q$: flow;
$Q_{b x}$ , $Q_{b y}$: solid flow components in $x$ and $y$ directions;
$Q_{m a}$: annual mean flow;
$S$: slope;
$T$: water temperature;
$Z_{b}$: bed level;
$α$ , $β$: coefficients of the Manning resistance coefficient;
$β$: coefficient of dimensionless bed load transport rate;
$γ$: density;
$ε$: absolute roughness in the outlet with clear water;
$ε_{s f}$: absolute roughness in the outlet with sediment transport;
$λ_{0}$: Darcy-Weisbach’s resistance factor on rigid bed with clear water;
$λ_{s s}$: Darcy-Weisbach’s resistance factor on rigid bed with sediment transport;
$ν$: coefficient of kinematic viscosity of clear water; and
$ν_{s}$: coefficient of kinematic viscosity of water with sediment concentration.

Acknowledgments

We are grateful to CELEC EP-Hidropaute and the Consorcio POYRY-Caminosca Asociados for the data provided and to Ministerio de Economía y Competitividad and Fondo Europeo de Desarrollo Regional (FEDER) for the financial aid received through the project (BIA2011-28756-C03-02). The authors are also grateful to the editors and the anonymous reviewers for their constructive comments that helped us to improve the quality and clarity of the paper.

References

Ackers, P., and White, W. R. (1973). “Sediment transport: New approach and analysis.” J. Hydraul. Div., 99(11), 2041–2060.

Crossref

Google Scholar

Aguirre-Pe, J., Olivero, M. I., and Moncada, A. T. (2000). “Transporte de sedimentos en cauces de alta montaña.” Revista Ingeniería del Agua, 7(4), 353–365 (in Spanish).

Crossref

Google Scholar

Annandale, G. W. (1987). Reservoir sedimentation, Elsevier Science, Amsterdam, Netherlands.

Google Scholar

Atkinson, E. (1996). “The feasibility of flushing sediment from reservoirs.”, HR Wallingford, Oxfordshire, U.K.

Google Scholar

Basson, G. R., and Rooseboom, A. (1996). “Sediment pass-through operations in reservoirs.” Proc., Int. Conf. on Reservoir Sedimentation, Fort Collins, 1107–1130.

Google Scholar

Bathurst, J. C. (1985). “Flow resistance estimation in mountain rivers.” J. Hydraul. Eng., 625–643.

Crossref

Google Scholar

Bathurst, J. C. (2002). “At-a-site variation and minimum flow resistance for mountain rivers.” J. Hydrol., 269(1–2), 11–26.

Crossref

Google Scholar

Bathurst, J. C., Graf, H., and Cao, H. H. (1987). “Bed load discharge equations for steep mountains rivers.” Chapter 15, Sediment transport in gravel bed rivers, Wiley, New York, 453–491.

Google Scholar

Brethour, J., and Burnham, J. (2010). “Modeling sediment erosion and deposition with the FLOW-3D sedimentation and scour model.”, Flow Science, Santa Fe, NM, 1–22.

Google Scholar

Brown, C. B. (1950). “Sediment transportation.” Engineering hydraulics, H. Rouse, ed., Wiley, New York, 769–857.

Google Scholar

Castillo, L. (2004). “Estimation of sediment transport and dominant flow in a hyperconcentrated flow.” Proc., 6th Int. Conf. on Hydroscience and Engineering, Univ. of Mississippi, Oxford, MS.

Google Scholar

Colby, B. R. (1964). “Practical computations of bed-material discharge.” J. Hydraul. Div., 90(2), 217–246.

Crossref

Google Scholar

Consorcio PCA. (2012). “FASE B: Informe de factibilidad. Anexo 2.” Meteorología, Hidrología y Sedimentología, Quito, Ecuador (in Spanish).

Google Scholar

Einstein, H. A., and Barbarossa, N. L. (1952). “River channels roughness.” Trans. ASCE, 117(2), 1121–1132.

Crossref

Google Scholar

Engelund, F., and Hansen, E. (1967). A monograph on sediment transport in alluvial streams, Teknisk Forlag, Copenhagen, Denmark.

Google Scholar

Exner, F. M. (1919). “Zur theorie der flussmäander.” Sitzber. Akad. Wiss., 128, 10 (in German).

Google Scholar

Exner, F. M. (1925). “Über die wechselwirkung zwischen wasser und geschiebe in flüssen.” Sitzber. Akad, Wiss., Wien, Vol. 134 (in German).

Google Scholar

Flow Sciences. (2011). FLOW-3D users manual version 10.0, Santa Fe, NM.

Google Scholar

Fuentes, R., and Aguirre-Pe, J. (1990). “Resistance to flow in steep rough streams.” J. Hydraul. Eng., 1374–1387.

Crossref

Google Scholar

García Flores, M. (1996). “Resistencia al flujo en ríos de montaña.” Proc., XVII IAHR, Guayaquil, Ecuador (in Spanish).

Google Scholar

Graf, W. H. (1984). Hydraulics of sediment transport, Water Resources Publications, LLC, CO.

Google Scholar

Grant, G. E. (1997). “Critical flow constrains flow hydraulics in mobile-bed streams: A new hypothesis.” Water Resour. Res., 33(2), 349–358.

Crossref

Google Scholar

Hirt, C. W., and Nichols, B. D. (1981). “Volume of fluid (VOF) method for the dynamics of free boundaries.” J. Comput. Phys., 39(1), 201–225.

Crossref

Google Scholar

Iberaula. (2013). “Iber: Hydraulic reference manual.” 〈www.iberaula.es〉 (Mar. 20, 2014).

Google Scholar

Janssen, R. (1999). “An experimental investigation of flushing channel formation during reservoir drawdown.” Ph.D. thesis, Univ. of California, Berkeley, Berkeley, CA.

Google Scholar

Jarrett, R. D. (1984). “Hydraulics of high-gradient streams.” J. Hydraul. Eng., 1519–1539.

Crossref

Google Scholar

Lai, J. S., and Shen, H. W. (1996). “Flushing sediment through reservoirs.” J. Hydraul. Res., 34(2), 237–255.

Crossref

Google Scholar

Limerinos, J. T. (1970). “Determination of the manning coefficient for measured bed roughness in natural channels.”, U.S. Geological Survey, Washington, DC.

Google Scholar

Mastbergen, D. R., and Von den Berg, J. H. (2003). “Breaching in fine sands and the generation of sustained turbidity currents in submarine canyons.” Sedimentology, 50(4), 625–637.

Crossref

Google Scholar

Meyer-Peter, E., and Müller, R. (1948). “Formulations of the bed-load transport.” Proc., Int. Conf. II IAHR, Stockholm, Netherlands, 39–64.

Google Scholar

Mizuyama, T., and Shimohigashi, H. (1985). “Influence of fine sediment concentrations on sediment transport rates.” Japan Civ. Eng. J., 27(1), 46–49 (in Japanese).

Google Scholar

Nalluri, C., and Kithsiri, M. M. A. U. (1992). “Extended data on sediment transport in rigid bed rectangular channels.” J. Hydraul. Res., 30(6), 851–856.

Crossref

Google Scholar

Palt, S. (2001). “Sedimenttransporte im Himalaya-Karakorum und ihre bedeutung für wasserkraftanlagen.” Ph.D. thesis, Mitteilungen des Institutes für Wasserwirtschaft und Kulturtechnik der Universität Karlsruhe (TH), Karlsruhe, Germany (in German).

Google Scholar

Parker, G., Klingeman, P. C., and McLean, D. C. (1982). “Bedload and size distribution in paved, gravelbed streams.” J. Hydraul. Div., 108(4), 544–571.

Crossref

Google Scholar

Rickenmann, D. (2001). “Comparison of bed load transport in torrents and gravel bed streams.” Water Resour. Res., 37(12), 3295–3305.

Crossref

Google Scholar

Shen, H. W. (1999). “Flushing sediment through reservoirs.” J. Hydraul. Res., 37(6), 743–757.

Crossref

Google Scholar

Smart, G. M., and Jaeggi, M. (1983). “Sediment transport on steep slopes.”, Versuchsanstalt für Wasserbau, Hydrologie und Glaziologie, ETH, Zürich, Switzerland.

Google Scholar

Strickler, A. (1923). “Beiträge zur frage der geschwindigkeitsformel und der rauhigkeitszahlen für ströme, Kanäle und geschlossene leitungen.” Sekretariat d. eidg. Amtes f. Wasserwirtschaft, Bern (in German).

Google Scholar

U.S. Army Corps of Engineers. (2010). HEC-RAS 4.1 hydraulic reference manual, Davis, CA.

Google Scholar

Van Rijn, L. C. (1987). “Mathematical modelling of morphological processes in the case of suspended sediment transport.” Ph.D. thesis, Delfy Univ. of Technology, Delft, Netherlands.

Google Scholar

Wan, Z., and Wang, Z. (1994). Hyperconcentrated flow, CRC Press, Rotterdam, Netherlands.

Google Scholar

Wang, D., Cheng, J., and Fu, X. (2013). “Quantification of flow resistance in mountain rivers based on resistance partitioning.” Proc., Int. Conf. 35th IAHR World Congress, Tsinghua University Press, Beiging.

Google Scholar

Wong, M., and Parker, G. (2006). “Reanalysis and correction of bed-load relation of Meyer-Peter and Müller using their own database.” J. Hydraul. Eng., 1159–1168.

Crossref

Google Scholar

Yakhot, V., and Smith, L. M. (1992). “The renormalisation group, the ε-expansion and derivation of turbulence models.” J. Sci. Comput., 7(1), 35–61.

Crossref

Google Scholar

Yang, C. T. (1984). “Unit stream power equation for gravel.” J. Hydraul. Div., 1783–1797.

Crossref

Google Scholar

Yang, C. T. (1996). Sediment transport: Theory and practice, McGraw-Hill, New York.

Google Scholar

Yang, S. Q. (2005). “Prediction of total bed material discharge.” J. Hydraul. Res., 43(1), 12–22.

Crossref

Google Scholar

Information & Authors

Information

Published In

Journal of Hydraulic Engineering

Volume 141 • Issue 11 • November 2015

Copyright

This work is made available under the terms of the Creative Commons Attribution 4.0 International license, http://creativecommons.org/licenses/by/4.0/.

History

Received: Oct 8, 2014

Accepted: Apr 6, 2015

Published online: Jun 8, 2015

Published in print: Nov 1, 2015

Discussion open until: Nov 8, 2015

Authors

Affiliations

L. G. Castillo [email protected]

Titular Professor, Hidr@m Group, Dept. of Civil Engineering, Universidad Politécnica de Cartagena, Paseo Alfonso XIII, 52, 30203 Cartagena, Spain (corresponding author). E-mail: [email protected]

View all articles by this author

J. M. Carrillo [email protected]

Lecturer, Hidr@m Group, Dept. of Civil Engineering, Universidad Politécnica de Cartagena, Paseo Alfonso XIII, 52, 30203 Cartagena, Spain. E-mail: [email protected]

View all articles by this author

M. A. Álvarez [email protected]

Associate Professor, GEAMA Group, Centro de Innovación en Edificación e Enxeæería Civil, Universidade da Coruña, Campus de Elviña, 15071 A Coruña, Spain. E-mail: [email protected]

View all articles by this author

Metrics & Citations

Metrics

Citations

Download citation

If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.

Complementary Methods for Determining the Sedimentation and Flushing in a Reservoir

Abstract

Main Characteristics of the Project

Flushing Considerations

Methodology

Estimate of the Manning Resistance Coefficient

Estimation of Sediment Transport

Reservoir Sedimentation

Flushing Simulation

Two-Dimensional Simulation

Three-Dimensional Simulation

Conclusions

Notation

Acknowledgments

References

Information & Authors

Information

Published In

Copyright

History

Authors

Affiliations

Metrics & Citations

Metrics

Citations

Download citation

Cited by

View Options

Media

Figures

Other

Tables

PREVIOUS ARTICLE

NEXT ARTICLE

Verify Phone

Congrats!

Abstract

Main Characteristics of the Project

Flushing Considerations

Methodology

Estimate of the Manning Resistance Coefficient

Estimation of Sediment Transport

Reservoir Sedimentation

Flushing Simulation

Two-Dimensional Simulation

Three-Dimensional Simulation

Conclusions

Notation

Acknowledgments

References

Information

Published In

Copyright

History

Authors

Affiliations

Metrics

Citations

Download citation

Cited by

Figures

Other

Share

Copy the content Link

Share with email

Share

Request Username

Create a new account

Change Password

Password Changed Successfully

Verify Phone

Congrats!