Morris & Fan. Reservoir Sedimentation Handbook

Подождите немного. Документ загружается.

ENGINEERING FEATURES OF DAMS AND RESERVOIRS 3.11

3.4 OUTLETS AND GATES

Outlets to control the release of water from a reservoir may be located at high,

intermediate, and low levels. Water quality characteristics such as dissolved oxygen,

suspended solids, and temperature vary as a function of depth, and a multiple-level intake

structure allows the selective withdrawal of water from different depths to better control

the quality of released water. Bottom outlets may be used to drain the reservoir and

release sediments. Inadequate outlet capacity is one of the principal causes of dam

failure, and most dams provide overflow spillways to safely discharge extreme floods.

Few dams are designed as nonoverflow structures with the entire flood flow designed to

pass through low-level gates. An ungated overflow spillway will begin discharging as

soon as the water overtops the spillway crest, whereas water normally stands above the

spillway crest in a gated spillway and both the discharge and pool level are controlled by

gate operation. The design flood criteria differ from one country to another, and also

depend on the downstream hazard and type of dam, but spillway capacity is typically

based on flood events having return intervals in the range of the 1,000 to 10,000 years

(Berga, 1995).

Gates controlling the release of water may be classified according to their direction of

opening as bottom-opening or underflow gates, or as top-opening or overspill gates. They

can also be classified by mechanical configuration (e.g., radial, vertical lift, flap), and

placement level (e.g., crest, bottom). Several basic types of gate mechanisms are

illustrated in Fig. 3.9. The bottom-opening radial or sector gate, also called Tainter gate

after its inventor, inscribes a sector of a circle and may be installed either as a crest gate

or as a bottom gate. They are normally operated by a chain hoist mechanism, although

hydraulic actuators are used in high-head, low-level installations. They are sometimes

provided with a counterweight. Vertical lift gates have a leaf which travels along vertical

slots. They may be constructed as slide gates, or in larger sizes may be equipped with

rollers to reduce friction, and may be operated by a hydraulic actuator or by a hoist and

wire rope (Fig. 24.5).

FIGURE 3.9 Types of gates frequently used in dams

ENGINEERING FEATURES OF DAMS AND RESERVOIRS 3.12

Overspill gates are unsuited for sediment passage for two reasons: (1) when partially

open they discharge from the water surface where sediment concentration is lowest, and

(2) bed material can become trapped in the gate recess. However, they are well-suited for

passing floating debris and ice. Some radial gates may be equipped with overspill flaps to

provide the sediment-passing characteristics of a bottom-opening gate, while still

retaining overspill capacity for release of floating debris (Figs. 3.3 and 3.13). Both radial

and bottom-opening flap gates are well suited for the closure of a sediment discharge

conduit because neither requires guide slots or mechanical equipment in the channel that

conveys the sediment. Gate seals for the flap gate are flush with the conduit, and the seal

on the radial gate faces downstream, where it is not subject to scour by sediment and

stones. Vertical lift gates require a guide slot along either wall to the base of the gated

opening, and there is the possibility of stones lodging in the slot. The slots used in

vertical sluices can also produce cavitation problems at high flow velocities. For all types

of gates passing coarse sediment, the grade on the apron in the immediate vicinity of the

gate should be designed so that stones are not deposited in the area of the seal during

decreasing flow, thus preventing complete gate closure (Bouvard, 1988).

Low-level outlets are normally constructed with two gates in series, the downstream

gate being used for normal operations and the upstream gate serving as an emergency

gate which can be closed to allow repair of the downstream gate. Because the flushing of

coarse sediment through bottom outlets can erode gate seals, separate gates for flow

modulation and for watertight closure should be considered, as described in the Gebidem

case study.

3.5 HYDROPOWER PLANTS

3.5.1 Nomenclature

A hydropower facility is schematically shown in Fig. 3.5. Trashracks consist of steel bars

that strain debris from water entering the intake. Shallow trashracks are generally set at

an incline and use vertical bars so that they can be cleaned by a rake. The penstock is the

pressure conduit connecting the intake to a powerhouse. In some facilities water is

conveyed to a point above the powerhouse through a canal or low-pressure tunnel on a

low gradient, which feeds into a short, steep, high-pressure penstock. The powerhouse

contains the hydraulic turbines, generators, hoists, and control room. Flow exiting

reaction-type turbines enters the draft tube, which is designed to decelerate flow velocity

and recover kinetic energy, creating a negative pressure on the downstream side of the

turbine runner. The draft tube discharges to the tailrace, which conducts flow back into

the river. Headwater and tailwater elevations respectively refer to water surface in front

of the intake and in the tailrace at the exit of the draft tube.

3.5.2 Types of Hydraulic Turbines

Hydrostatic head is converted into mechanical energy by hydraulic turbines, which can

be classified into three types: impulse (Pelton wheel), Francis reaction, and propeller

reaction. The propeller-type turbine may have blades of either fixed or variable pitch, the

Kaplan turbine being an example of the latter. Basic turbine configurations are shown in

Fig. 3.10, and a Pelton wheel is shown in Fig. 3.11. All turbines consist of two elements,

a static guide (or nozzle in the case of the impulse turbine), which converts the static head

partially or wholly into velocity, and a revolving part, called the runner.

ENGINEERING FEATURES OF DAMS AND RESERVOIRS 3.13

FIGURE 3.10 Types of hydropower turbines: (a) Pelton wheel; (b) propeller turbine;

FIGURE 3.11 Pelton wheel impulse turbine with casing removed (G. Morris).

e nozzle in the impulse turbine converts the entire static head into velocity head, and

atmospheric pressure surrounds both the jet issuing from the nozzle and the bucket wheel.

Impulse turbines are generally used for installations with static head exceeding 150 m.

The guide case for reaction-type turbines converts only a portion of the head into

ENGINEERING FEATURES OF DAMS AND RESERVOIRS 3.14

velocity, producing an overpressure between the guide case and the runner. The flow

continues to accelerate as it passes through the throat of the runner, and the draft tube

generates a negative pressure on the downstream side of the runner which increases

energy transfer. Reaction-type turbines are generally used at heads of less than 300 m,

and the propeller-type turbine is more suited for high-volume, low-head applications.

3.5.3 Energy Relations

Energy is the ability to do work, computed as force times distance. The unit of measure

for energy is the joule (J) which is equivalent to a force of 1 newton (N) acting over a

distance of 1 meter. In the English system the unit of energy is the foot-pound (ft•lb), a

force of 1 pound acting over a distance of 1 foot (1 J = 0.7376 ft•lb). The potential energy

in a volume of water is the product of weight and static head. Power is the rate of work,

or work done per unit of time and may be expressed in watts (W), the unit equivalent to

energy expenditure at the rate of 1 J/s. In hydroelectric practice kilowatts or horsepower

are used to express power, and the kilowatt-hour (kWh) is used to express energy. These

may be computed as follows:

1 kWh = 3.6 × 10

J = 2.655 × 10

ft•lb

1 hp = 745.7 W = 550 ft•lb/s

The potential energy available in 1 m

of water (1000 kg) at 1 m head is 9800 J = 7228

ft•lb = 2.722x 10

-3

kWh. Power output may be related to flow as follows:

Power (kW) = 9.8 × Q (m

/s) × head (m)

Power (hp) = Q (gal/min) × head (ft)/3,960

Hydraulic turbines can operate at efficiencies exceeding 90 percent based on head loss

across the turbine. One of the important effects of sediment on hydropower stations is the

erosion of turbine runners, which causes substantial declines in efficiency and increasing

maintenance.

3.5.4 Sediment Impact on Tailwater

The gross head at a hydropower plant is the difference between headwater and tailwater

elevation when the plant is not operating. The net head or effective head is the total net

height of the water column effective on the turbine runner when it is generating power. It

represents the gross head less all friction and turbulent losses before and after the turbine,

and the increase in tailwater level caused by the turbine discharge plus any sediment

accumulation in the river in the vicinity of the power house. Sedimentation problems

affecting tailwater elevation are exemplified by the 3.4-Mm

Ralston Afterbay Reservoir

on the American River in California.

Since closure of Ralston Dam in 1966, sediment has accumulated at an average rate

of 43,200 m

/yr, and about 30 percent of the reservoir capacity had been sedimented by

1990. The general layout of this system and the pattern of sediment accumulation is

illustrated in Fig. 3.12. The 79,200-kW Ralston power plant discharges 2.6 km upstream

of the dam in an area affected by delta deposition, and bed aggradation raises the tail-

water elevation. When tailwater reaches 358.7 m, backsplash within the impulse turbine

discharge chamber causes vibration and reduces turbine efficiency because of increased

drag. Peak load diminishes as tailwater rises to 360.0 m, at which point the unit must be

shut down. At the Oxbow intake located about 360 m up stream of the dam, sediment

ENGINEERING FEATURES OF DAMS AND RESERVOIRS 3.15

ENGINEERING FEATURES OF DAMS AND RESERVOIRS 3.16

TABLE 3.3 Sediment Excavation at Ralston Afterbay Reservoir

Year Excavation volume, m

Base year for costs Excavation cost, 1990 $/m

1984 9,990 1990 7.32

1985 9,200 1990 5.20

1986 34,400 1990 15.47

1986 61,200 1990 4.82

1989 24,900 1990 8.04

1994 58,900 1994 23.77

Source: Plumas County (1990); Jones (1996).

entrainment during high flows overloads the strainers in the cooling water supply for the

generators. The amount of power lost due to shutdowns at Ralston and Oxbow is

variable, with larger amounts being lost during wet years.

Sediment accumulation to date has been controlled by excavation (Table 3.3), and

excavation volume is expected to increase with time as sedimentation continues. Large

amounts of coarse sediment were transported by the 1986 flood, causing 2.4 m of bed

aggradation along the river in the vicinity of the powerhouse discharge and requiring

larger excavation volumes. Excavation costs may be expected to increase over time as

nearby disposal sites become filled up. The extremely high unit cost for sediment

excavation from the delta area in 1994 reflected the cost of a 5-km haul with an elevation

gain of 365 m to reach the disposal site, where the material was used to construct a dry

dam. Regulatory agencies have not allowed placement of excavated material back into

the stream channel below the dam, despite more than 2 m of channel degradation at the

gage station located 2.5 km below the dam since dam closure. The owner is presently

examining alternatives to route sediment through the reservoir to reduce or eliminate

reliance on mechanical removal (Jones, 1996).

3.6 ABRASION AND CAVITATION

3.6.1 Abrasion of Turbines

Sediment entrained in water can severely abrade hydraulic machinery of all types,

causing efficiency declines in turbines and pumps and also affecting valve and gate seals.

In extreme cases sediment can render hydraulic machinery useless in a matter of weeks

(Bouvard, 1992). Grain sizes over 0.1 mm should be excluded for heads exceeding 50 m,

and for heads exceeding 200 m even silts should be excluded (Raudkivi, 1993). Angular

quartz particles are particularly abrasive and are generally abundant in glacial melt

streams, posing a particular problem for hydraulic equipment in this environment. Francis

turbines are highly sensitive to wear, often starting in the labyrinth where it causes higher

leakage and requires dismantling for repair. In Pelton wheels the abrasion is focused on

the needle and nozzle tip, both of which can be easily replaced, a factor that favors their

use in installations when abrasion is a possibility. However, abrasion increases rapidly as

a function of head, especially above 400 m, and high-head Pelton wheels can be abraded

by 0.05-mm quartz in suspension (Bouvard, 1992).

3.6.2 Abrasion of Concrete Structures

Sediment will abrade spillways, aprons, and outlets that pass coarse bed material load,

which may range from sand to boulders. Conventional concrete has limited abrasion

ENGINEERING FEATURES OF DAMS AND RESERVOIRS 3.17

resistance and may be faced with a variety of abrasion-resistant materials such as dressed

stone, steel (at least 2 cm thick, not including anchors), timber, fiber-reinforced concrete,

or concrete with a coarse granite aggregate. When dressed stone is used, each block

should be shaped like a molar tooth, consisting of "roots" 0.5 to 0.7 m long capped with

squared vertical faces at least 15 cm deep which will be exposed to wear. The blocks

should be laid with the smallest possible joints, and jointing should be arranged

perpendicular to the flow to prevent the development of preferential paths for erosion.

Dressed dense granite is highly resistant to both abrasion and shock, but costly. It may be

considered in areas where facing replacement would be extremely costly, even if it could

not be afforded elsewhere in the project. Oak or larch timbers 10 to 15 cm thick have

been extensively used in the French Alps to line outlet works on streams transporting

gravels and pebbles, including structures discharging as much as 1000 m

/s during floods.

Wood is both elastic and smooth and has a service life on the order of 10 to 15 years

(Bouvard, 1992). Railroad rails placed parallel to the flow line and interfilled with

concrete have been used in a variety of projects, although the concrete between the rails

can be eroded by gravel. An example of abrasion protection at a small project is shown in

Fig. 3.13.

The Compagnie Nationale du Rhone (CNR) operates 18 barrages on the Rhone River

between Switzerland and the Mediterranean Sea. Abrasion is a significant problem at

these sites, which pass sands and gravels through gates and across aprons. Dams on the

Rhone have used epoxy-resin concrete covered with a 2-cm-thick layer of pitch epoxy

resin corundum on aprons and metal linings on the lower 2 or 2.5 m of piers (Bouvard,

1992). Pierrer and Fruchart (1995) reported on shock and abrasion tests performed by

CNR on various substances including Abraroc, a proprietary concrete-based material

formulated for use in abrasive environments (Table 3.4). In these tests shock resistance

was measured as the volume of material eroded from a sample by a 1-kg metal ball 7 cm

in diameter falling from 1 m, at 15 strikes per minute, for a period of 3 h. Abrasion

resistance was measured by jetting water entraining 99 percent pure silica sand, 1.0 to 2.2

mm in diameter, at a 45

angle against a submerged sample.

3.6.3 Cavitation

Areas of negative pressure can occur in high-velocity flows, causing water to separate

into liquid and gas phases, forming bubbles that subsequently collapse. The dynamic

energy released by cavity collapse is highly erosive and cavitation can cause serious

damage to hydraulic equipment and structures. Cavitation noise can be used as an

indicator of potential cavitation damage. Raudkivi (1993) states that suspended sediments

cause cavitation damage and noise to increase compared to clear water for sediment

concentrations up to about 100 g/L, with the maximum effect at about 25 g/L of sediment

concentration. At sediment concentrations above 100 g/L, cavitation is reduced compared

to clear water.

Liu (1987) reported on the effect of sediment concentration on cavitation, using a

cylinder cavitation test and silts from the Yellow River with a d

diameter of 0.014 mm.

In the experimental setup the length of the cavitation cavity downstream of the cylinder

could be observed and noise level was measured with a transducer. The difference

between the cavitation number K in clear and sediment-laden water, for constant values

of cavity length or noise level, was interpreted as the effect of sediment on cavitation.

Compared to clear water, cavitation was initiated more readily for suspended sediment

concentrations up to about 15 g/L, and cavitation noise was increased for sediment

concentrations up to about 7 g/L. Cavitation was reportedly less than in clear water at

higher sediment concentrations (Fig. 3.14).

ENGINEERING FEATURES OF DAMS AND RESERVOIRS 3.18

FIGURE 3.13 The small La Peuffeyre hydroelectric diversion dam, Bex. Switzerland, is

located on a steep gravel-bed stream. The spillway is fitted with a radial gate with an

overspill bascule for passing floating debris. The spillway is armored with granite blocks

and the apron below with railroad rails. Note the erosion of the unprotected concrete

between the rails (G. Morris)

TABLE 3.4 Shock and Abrasion Resistance of Selected Materials

Material Shock resistance, cm

Abrasion, cm

Steel

— 0.02-0.04

Granite <100 0.35-0.80*

Abraroc 175 0.4

Very resistant concrete < 150 <1

Resistant concrete 150-250 <2

Common concrete 250-400 2-3

Mediocre concrete >400 >4

* Depending on origin

Source: Perrier and Fruchart (1995).

ENGINEERING FEATURES OF DAMS AND RESERVOIRS 3.19

FIGURE 3.14 Effect of sediment concentration on cavitation number (after Liu, 1987).

3.7 RESERVOIR BACKWATER AND FLOOD ROUTING

During floods the water surface profile above a dam will be elevated above the level-pool

elevation at full reservoir level. The backwater curve corresponding to any flow condition

can be computed from the stage-discharge characteristics of outlets, as regulated by gate

operation; geometry and hydraulic characteristics along the impounded reach; and the

unsteady discharge characteristics of the inflowing hydrograph. Backwater profiles can

be computed by using either steady-flow models such as HEC-2 (U.S. Army, 1990) or

unsteady flow models such as UNET (U.S. Army, 1992) and DAMBRK (Fread, 1988).

These hydraulic models use fixed streambed geometry. However, reservoir geometry is

altered by sediment accumulation, and delta deposits can infill the upper reaches of the

reservoir and also cause the streambed to aggrade upstream of the normal pool level. The

subsequent growth of brush or trees on delta deposits, which increases hydraulic

roughness, can substantially increase the backwater profile as compared to initial non-

sedimented conditions. The modification of in-channel geometry and roughness due to

sedimentation and vegetation should be considered in backwater computations.

When unsteady flood flows are routed through reservoirs, the temporary storage of

floodwater will significantly reduce the flood peak at the dam. Sedimentation that

reduces flood storage will not only increase backwater levels above the dam due to rising

bed level, but it will also increase the peak discharge at the dam. In this manner,

extensive sedimentation may cause the spillway design capacity to be exceeded.

3.8 RESERVOIR YIELD

Reservoir yield varies as a function of available storage volume in the conservation pool

and will decline over time as this volume is lost to sedimentation. The impact of

sedimentation management alternatives on reservoir yield should be quantified before

selecting a sediment management strategy. McMahon and Mein (1986) summarize yield

ENGINEERING FEATURES OF DAMS AND RESERVOIRS 3.20

analysis technique emphasizing rivers and individual reservoirs, while Basson et. al.

(1994) describe techniques developed and used in a complex 50-reservoir system.

McCuen (1993) provides an overview of a number of statistical techniques for hydrologic

analysis and includes diskettes with programs and sample datasets. Time series modeling

is reviewed by Salas et al. (1980).

3.8.1 Storage-Yield Relationship

Reservoir inflow is stochastic, and yield is expressed as the withdrawal rate that can be

sustained at a stated probability level without emptying the storage pool. Thus, a

municipal supply reservoir designed to deliver 1 m

/s of water at the 99 percent level of

reliability may deliver a lesser flow rate on 1 percent of the days (36 days per decade).

Failure occurs when the desired yield cannot be sustained for more than 1 percent of the

time over the long run. In the simplest case, the reservoir is drafted at a constant rate until

empty, at which point the rate of withdrawal will be limited to the rate of inflow. Firm

yield is the maximum uninterruptable amount of water that can be continuously

withdrawn from a reservoir for a given (e.g., historical) inflow series and subject to a

given set of operational rules and a fixed storage volume.

However, supply reservoirs may not be drafted at a constant rate until empty; water

rationing may be imposed when the pool level drops to a certain point. Thus, the total

conservation pool may be divided into an upper active pool and a lower inactive pool

which functions as the reserve pool, with water rationing being triggered once the active

pool is depletes.

The yield at zero storage is the minimum streamflow at the stated exceedance level

(the unregulated streamflow equaled or exceeded 99 percent of the time in the municipal

example). Reservoir yield increases as a function of storage volume for any stated level

of reliability. For a given hydrologic sequence, a unique relationship exists between

storage volume, yield, and service reliability. Yield increases as a function of storage, but

additional increments in storage volume generate less yield than the previous storage

increments producing storage-yield relationships having the asymptotic pattern shown in

Fig. 3.1. Because the slope of yield curve steepens as storage volume declines, if

conservation storage declines at a constant rate, the rate of yield decline will accelerate.

3.8.2 Gould's Gamma Method for Estimating Yield

Hydrologic datasets are often short at existing or proposed dam sites, and it is both con-

venient and necessary to use approximate methods to estimate yield as a function of con-

servation storage. Even when apparently adequate data are available, it is useful to use an

independent method to check yield estimates for reasonableness. McMahon and Mein

(1986) present Gould's gamma method for making a preliminary estimate of the constant

rate of withdrawal sustainable from a single reservoir, as a function of probability of

failure, given information on the mean (x¯ )and standard deviation (σ) of

annual

streamflow discharge. This method incorporates the assumption that the sequence of

annual flows is gamma-distributed, but for the sake of computational simplicity it use the

normal distribution and a correction factor to approximate the gamma distribution.

The equation for determining the storage volume required for each combination of

yield and level of reliability is:



 {Z

/[4(1 D)] d}C

whe

 = storage volume expressed as hydrologic size (C/I ratio)

= coefficient of variation of annual flows (C

=  / x¯ )

= sta

ndardized normal variate at p%

D = withdrawal rate as ratio of the mean annual discharge rate

d = correction factor ( Table 3.5)

Morris &amp; Fan. Reservoir Sedimentation Handbook

Подождите немного. Документ загружается.

Morris & Fan. Reservoir Sedimentation Handbook