Inversin R. Allen Micro-hydropower Sourcebook

is placed vertically on top of the stake and the level of

the stream is read directly off the scale. Numerous

variations of this method can be used.

Although specific details in the construction of a weir

can vary, it is important that the critical features

described above be incorporated. When constructing a

temporary weir, the task of sealing around the weir can

be simplified by laying a sheet of plastic on the

upstream side of the weir and using sand, gravel, or

rocks to hold it in place (Fig. 2.15).

The formulas shown in Fig. 2.14 or the curves shown in

Fig. 2.16 yield t!ie discharge through each of the three

types of notches described. Note that for rectangular

and Cipoletti weirs, the discharge is given per unit

effective weir length. Actual discharge is obtained by

multiplying this discharge by the effective weir length-

by “L - 0.2 h” for the rectangular weir or by “L” for the

Cipoletti weir.

In reality, a much wider range of designs for sharp-

crested weirs is available, but then the formulas for

expressing discharge must include the effects of

decreasing end contractions, velocity of approach, and

submergence. In this section, the simple forms of the

formulas are presented. For these to be valid, it is

necessary to adhere to the limitations noted in

Fig. 2.14. If the notch is too wide or deep, contractions

Fig. 2.15 Three separate rectangular boards are used to

improvise a tempomry gaging weir. WIthout readily

avail-

able soil

in the vicinity,

preventing leakage

crowd such

a

weir on bedrock wollld be difficult,

but

a sheet of plastic

become suppressed and the formulas which are given

facilitates that task

0.8

0.6

2

.

N

El

-3 0.4

0.2

C

For discharge through:

I. rectangular weir - read “q” on left-hand

scale and multiply by (L - 0.2 h)

2. Cipoletti weir - read “q” on left-hand

scale and multiply by L

3. triangular weir

- read “Q” directly on

right-hand scale

30 40

h (here in cm)

50

60

Fig. 2.16.

Curves

for

determining

flow through a weir.

18 Measuring head and discharge

underestimate the discharge. If the velocity of

approach is too high, the formulas again underestimate

the discharge. And if there: is insufficient room for the

water to fall freely downstream of the weir, the weir

becomes submerged and the formulas presented overes-

timate the discharge.

Sa.IMilution method

There are two approaches to determining discharge by

measuring the dilution of a salt released into a stream.

With the first approach, a concentrated salt solution is

released into a stream all at once and the change in

concentration in the stream is measured over time as

the water flows past a point downstream. With the

second approach, a salt solution is released into the

stream at a known rate and the salt concentration is

measured downstream.

The first approach requires that a known mass “M” of

salt-about 0.3 kg for each 0.1 m3/s of streamflow-

which has been completely dissolved in a bucket of

water be quickly dumped into the stream to be gaged.

At a point 50 m or more downstream-far enough for

the salt to be uniformly dispersed throughout the stream

cross-section-electrical conductivity readings are

made. This may cover a period of several minutes. A

curve of conductivity (ohm-l) vs. time (s) is then plotted

(Fig.

2.17). The shaded area under the curve is calcu-

lated (oh

-T

-1 s). Using a conversion factor “k” (kg/m3

per ohm

) between concentration and conductivity, the

discharge in the stream is then found using the following

equation:

Qb3/s) =

M (kg)

area (ohm-‘,) x k (kg/m3/ohm-‘)

(2.11)

A brief derivation of this equation is as follows: At any

point downstream, the volume of water passing through

the stream cross-section during an increment of time

” dt” simply equals Q At (m3). If, from the conductivity

measurement “p” and the conversion factor “k”, the salt

concentration in that volume of water is determined to

be C (kg/m3) = c k, the mass of salt passing this point

during this increment of time is CQ At (kg). The total

salt in the stream, which is already known, is then the

sum of the mass of salt passing during each increment

of time or M = E CQ At. Since the discharge “Q” is

assumed to be constant during the sampling period

(which it essentially always is), then M = Q C C At, or

Q= -!!i- =

M

(2.12)

CC At

5

C dt

Because the factor EC At equals k E 1 At, the deno-

minator in this expression is simply the product of the

conversion factor and the shaded area under the conduc-

tivity curve (Fig. 2.17).

A more detailed description of this method based on

v = average flow velocity in channel (m/s)

work by Andrew Brown, lTlS Micro-Hydro Engineer, is

r = hydraulic radius = A/P

found in “Stream Flow Measurement by Salt Dilution

A = cross-sectional area of flow (m2)

120

I=7

I

100 -

conductivity due

n

+“‘I-

to salt introduced

background co:ductivity

0

30 60

90

120

Time Is)

150 180

Fig. 2.17. A typical curve

of conducffvity vs.

time at T

point downstream from

where a

known quantity

of salt

has

been released.

.

Gauging” (33). This reference also includes a discussion

of meter calibration and temperature correction.

The second approach requires that a salt solution with a

high concentration “Cl” (kg/m3) be gradually discharged

into a stream at a known rate “q”, This disperses trans-

versely throughout the stream as the water continues

downstream. At a point downstream where mix ng is

complete, the salt concentration “C2” is measured by

measuring conductivity. The discharge in the stream is

then

Q

= q(C1/C2)

(2.13)

This assumes that the natural salt concentration in the

stream has been subtracted to get “C2n and that “Q” is

much larger than “q”.

If either of these assumptions is

incorrect, the more precise form of this equation must

be used (24).

Slopearea method

This method uses an open channel flow equation, such as

Manning’s equation which is frequently used to deter-

mine discharges in power canals (see Determining canal

dimdons and slope, p. 96). In metric units, this

equation is expressed as:

,2/3,1/2

v=

n

where

(2.14)

Measuring head and discharge

19

Straight bank, full stage, no pools

clean

with some weeds and stones

Winding, some pools and shoals

clean

clean but at lower stages

with weeds and stones

lower stages, stony sections

Sluggish, weedy or with deep pools

Very weedy and sluggish

n (roughness coefficient)

0 .020

.040 .060

.080 ,100

,120 .140

.16(1

-

I

1

-

Qg. 2.18. Roughness coefficient for natural channels.

P = wetted perimeter of this area (m)

s = slope of the stream’s surface

n = roughness coefficient (see Fig. 2.18)

roughness, and size of the streambed over a significant

distance, the channel of the stream itself serves as the

control. In any case, the stage measurement site should

be selected so that, for a given change in discharge, as

large a change in stage will result. A broad control sec-

tion should be avoided, because a large change in dis-

charge will cause only a smali fluctuation in stage.

To use this method, a straight channel 50-300 m long

with reasonably uniform slope and cross-section should

be selected. Its bed and banks should be permanent and

the slope should be steep enough to be measured without

a large percentage error. To derive the hydraulic

radius, the mean cross-sectional area and wetted peri-

meter must be determined. With irregular channels,

this may have to be done at several points and the mean

value used to determine hydraulic radius. The slope of

the stream’s surface should be measured by dividing the

difference in the water surface elevation at the two

ends of the channel section being considered by its

length. The roughness of the bed and banks must also be

estimated (Fig. 2.18). Manning’s equation gives the

average flow velocity associated with that cross-sec-

tion. Discharge is then simply the product of this velo-

city with the mean cross-sectional area. Because cor-

rect determination selection of the roughness coeffi-

cient is difficult, the value derived for discharge by

using this equation is only approximate.

Stag-charge method

The validity of the stage-discharge method rests on the

axiom that the discharge for any given stage or depth of

water will remain unchanged (a) as long as the stream-

flow is steady and (b) there is an effective control sec-

tion below the stage-gaging site where the streambed

and banks are permanent. Because the control section

is equivalent to a notch in a weir used with the weir

method, this is analogous to saying that the stage-dis-

charge relation for a weir will remain unchanged as long

as the notch in the weir remains unchanged. Even if

there are changes in the bed or banks near where head is

measured, this will have no effect on the relation bet-

ween head and discharge through the notch as long as

the changes do not extend to the control section (the

notch). For example, if debris is lodged at the control

section (or around the notch of a weir), sediment or

boulders are deposited there, or erosion causes the con-

trol section to change in shape, then readings will no

longer remain valid because performance of the control

section (or notch) will have changed.

As mentioned in Weir metbod(p. 16), one method of

measuring discharge is to construct a notched weir and

The gage itself must be placed so that it is accessible at

all stages and its datum remains constant. it should be

measure the depth of water, or stage, behind the weir.

The stage-discharge method is similar to the weir

referred to bench marks entirely removed from the gage

so that, if disturbed, it may be replaced to the same

method except that, rather than a notch, some physical

feature of the stream downstream of the gage controls

datum. The gage should not be situated where it is

the relation between stage and discharge. When this

exposed to possible damage from debris nor where silt

controlling feature is situated in a short reach of the

deposited by the normal current can cut off the connec-

stream, a “control section” is said to exist. This might

tion between it and the free water level of the flowing

be a stretch of rapids where there is at least a moderate

water. It is not necessary, however, that the site also

fall or a culvert with a cascading discharge. If the

be suitable for measuring discharge in order to calibrate

stage-discharge relationship is governed by the slope,

the gage. This can be done upstream or downstream

from the gaging site, if the flow is steady and no water

20 Measuring head and discharge

enters or leaves the stream in the reach between where

the discharge is measured to calibrate the gage and

where the gage itself is located.

The most common and least expensive gage is a staff

gage, a vertical or inclined scale rigidly and perma-

nently secured in a stream, either attached to solid

rock, a bridge pier, or other structure built at the river’s

edge. If no suitable site is available, it may be attached

to a masonry pier built for that purpose (Fig. 2.19).

Water-stage recorders which cperate continually are

also available but involve substantially increased cost

(Fig. 2.20).

Fig.

2.19. A

staff

gage mounted

on

a masonry pier.

To calibrate a staff or other type of gage, hoth the

stage and corresponding discharge must be measured for

each of several stages spanning the range of stages that

will to be encountered during the life of the gaging sta-

tion. The velocity-area method is commonly used to

measure discharge. This is done over several months,

because the discharge at any stage can be determined

accurately only when the stream itself attains that

stage. A rating curve--a graph of stage vs. discharge--

can then be prepared to determine the discharge corre-

sponding to any stage read off the gage (Fig. 2.21). A

complete rating curve requires that both minimum and

maximum stream discharges be gaged, as well as several

discharges in between,

A permanent control section is essential to obtain valid

results using this method. Therefore, it may be neces-

sary to verify the stage-discharge relationship periodi-

cally, after a monsoon season for example, to determine

whether erosion or sediment deposits have changed the

control section since the last calibration.

MEASUREMENT OF HEAD

Measuring the available head is often seen as a task for

a surveyor, but for all schemes, much quicker and less

costly methods can be used for the preliminary deter-

mination of head. For micro-hydropower schemes,

Fig. 2.20. A staff gage and a water-stage recorder

on the

Ruvyironra River near Kibimbu, Burundi.

0.8

0.6

3

& 0.4

rd

iz

0.2

0

0.2

0.4 0.6

Discharge (m3/s)

Fig.

2.21. A stage-discharge

rating curve.

these methods often can also be used for final deter-

mination of gross head, although the apprzte method

must then be selected, especially for sites with very low

heads or small penstock gradients. Because the power

available from a given turbine is proportional to H3i2,

the percentage error in the power output due to incor-

Measuring head and discharge 21

rect head measurement is about 1.5 times the percent-

age error in this measurement.

Of the methods described ‘below, using a level is the

most straightforward and requires the minimum invest-

ment. A carpenter’s level can be used, although a less

expensive line level or a more expensive Locke hand

level can also be used. For a typical micro-hydropower

site, accuracy is generally within *5%; accuracy is

usually be?t* r than this figure in steeper terrain and

poorer in gr~ctually sloping terrain. One person can use

this method with a carpenter’s or Locke level, but two

are preferable if a line level is used.

Using a clinometer or Abney leve! (Fig. 2.22) also

requires a measuring tape, unless the clinometer is sim-

ply used as a level as in the first method described.

Two persons are usually required and accuracy is of the

same order as with a level. It involves making, record-

ing, and manipulating both linear and angular measure-

ments. However, rather than moving laboriously in

2-20 m increments up a slope, the length of the incre-

ments is limited only by the length of the tape and the

nature of the terrain.

If a rangefinder is used with a

clinometer, greater increments are possible. With a

rangefinder and clinomcter, it may even be possible to

measure gross head in a single step, provided the site

for the intake to the penstock is visible from the power-

house site. However, in this case, error will probably be

greater.

Fig. 2.22.

A clinometer

or

Abney level used to

measure

angles of elevation.

Using water pressure to determine head is not generally

used, primarily because it has not been considered by

most and because it requires a pressure gage, which is

usually not found around the home. This method

requires at least a 10 m length of preferably clear plas-

tic tubing and a pressure gage. The error associated

with this method depends on the accuracy of the gage

used. A single person can use this method, although two

would save time.

Using an altimeter is probably the most expensive,

although one of the easiest, methods of measuring head.

However, the user must be aware of how to handle this

device properly and of what factors, beyond change in

elevation, can affect the readings (temperature,

changes in atmospheric conditions, and, to a lesser

extent, relative humidity). Accuracy obtainable

depends on whether a pocket or surveying altimeter

(Fig. 2.23) is used. When care is used with a surveying

altimeter, maximum

errors can

be as small as 1.0 m.

Fig. 2.23. A

surveying

altimeter.

Using a level

Although there are several variations for using a level

to measure head, all are based on the same principle.

The approach described below is probably the most

straightforward and quickest, involves the fewest mani-

pulations of numbers, and requires only a carpenter’s

level.

Basically, one begins at a point at the bottom of the

hill, at point X in Fig. 2.24 which represents the pro-

posed powerhouse location. A carpenter’s level is set on

a stake of known length which rests on this point. A

horizontal line is sighted along the upper edge of a level

to a point XI on the ground. Therefore, the difference

in elevation or head from X to Xl is “h”, equal to the

length of the stake plus the width of the level. The

stake and level are then placed at XI to sight to the

next point, X .

4

This procedure is continued untii point

Y, the level o the water at the location for the pro-

22 Measuring head and discharge

Fig. 2.24.

Using a carpenter’s level for determining gross head “Hg” between two points, X and Y.

posed intake to the penstock. The final reading “hf”

It should be noted that,

using

this method, two tasks

might be less than the full distance “h”. The totalhead

is then

Hg

= nh + hf (2.15)

where n is the number of intermediary points between X

and Y. In preparing the stake, the distance “h” selected

should enable convenient sighting along the level from a

standing position slightly below the stake.

This method requires that head be measured from the

point of lowest elevation and that measurement pro-

gress upward along the slope. Surveyors with transits

can progress downhill but then a longer stadia or level-

ing rod would be required. If going from the powerhouse

to intake locations requires going downhill for a portion

of

the distance (Fig. 2.251, the method described

requires that segments be measured separately from the

lowest to highest points. The corresponding changes in

elevation are then added

or

subtracted as appropriate to

derive the gross head “Hgn.

Flg. 2.25. Determintng the gross head over undulating ter-

min.

have to be performed simultaneously-viewing along the

level to some point X.

while at the same time viewing

the bubble to ensure t h

at the carpenter’s level is exactly

horizontal. Clearly this can be performed by two per-

sons. If only one person is available, he can perform

both tasks using a small

mirror

oriented at 45’ to the

axis

of

the level; this permits the bubble to be viewed

from the same position from which the aighting to the

next point Xi is made.

If

a tripod is available to which

the level can be secured, the level can be mounted hori-

zontally first and then the sightings can be made.

The error in measurement of gross head using this

method depends on the average gradient of the slope.

The error is positioning the level horizontally may aver-

age itself out over the entire distance. The error in

sighting along the level might not, depending on the

user. With a little care, on a fairly steep slope of 1:2

(vertical to horizontal) or about 30°, error should be less

than 2%4%. For a fairly level grade

of

l:lO,

error

increases considerably, to possibly lo%- 20%.

Note that for the determination of head, horizontal dis-

teances are of no concern. In addition, the distance

along the ground from X to Y is of importance only

later, when the length and diameter of the penstock

pipe to be used need to be specified.

Using a clinometer

-

This method requires a clinometer or Abney level tv

measure vertical angles. The view through the eyepiece

of this device is split so that while the instrument can

be aimed at the next station up (or down) the hill, an

attached graduated arc can be oriented horizontally by

means of a spirit level which is also visible through the

eyepiece.

The angle of elevation “0” in the direction

being sighted can then be determined. Though clinom-

Measuring head and discharge

23

H

!3

Fig. 2.26. Using

a clInometer to determ!ne gross head

between

two points, X

and

Y.

eters are available with or without magnification,

sighting is often easier through one without.

In using a clinometer, a measuring tape is stretched

between two stakes of equal length to determine the

distance “L” between the two stakes (Fig. 2.26). A cli-

nometer resting on one stake is used to measure the

angle of elevation (or depression ) “0” to the top of the

next stake. A trigonometric function, sine, is then used

to convert these two readings to the difference in ele-

vation of the two points “h”. The gross head at the site

is then determined by completing a table such as that

shown

in Fig. 2.25.

Using a premsure gage

This method for measuring head relies on the fact that

for every meter vertically below the free surface of

water, water pressure increases by 9.84 kPa (1.43 psi),

Independent of the shape of whatever is confining the

water.

Consequently, the distance in elevation “h” (m)

between the free water surface and the pressure gage

can be expressed as

h=&

(2.16)

.

where “p” is the pressure reading (kPa). On the other

hand, a pressure gage directly calibrated in meters

could be used.

A device to measure head in this fashion requires a long

length of flexible tubing, one end of which is connected

to an appropriately sized and correctly calibrated pres-

sure gage. The tubing is then filled with water. The

tubing must be la.rge enough to allow all air bubbles to

be expelled easily, because they could introduce errors.

To help verify that no air remains in the tubing, the tub-

ing should be made of clear plastic.

24 Measuring head and discharge

To apply this method, the distance between the pro-

posed penstock inlet and powerhouse is measured in

increments approximately equal to the length

of

the

tubing (Fig. 2.27). The pressure reading is noted for

each increment. The sum of all these readings, con-

verted to meters, will equal the

gross

head.

plastic tubing

filled with water

Fig. 2.27

A

pressure

gage can be

used to measure the

heat

bet

ween two points.

Although this method has been used in the field, it is not

always practical. If the area to be measured is wooded

or covered with brush, maneuvering a long length of

tubing is difficult. In addition, gages with different

full-scale readings are required to measure slopes with

small and large gradients. It is also more difficult to

verify the calibration of pressure gages.

U&g an altimeter

For an increase in elevation of 100 m, atmospheric pres-

sure decreases by approximately 9 mm of mercury.

Therefore, if one were to measure the pressure at each

of two locations with a barometer, the difference in

pressure could then be used to derive the difference in

elevation between those two points. An altimeter used

to measure changes in elevation is essentially a barom-

eter calibrated in meters.

Although an altimeter carries a scale indicating eleva-

tion, this scale, unfortunately, is not absolute. If peri-

odic elevation readings are taken during the day from an

altimeter whose position remained unchanged, the read-

ings will change because of changing weather, tempera-

ture, and humidity.

Ambient temperature is a major factor affecting altim-

eter readings. If the temperature increases at the loca-

tion of the altimeter, atmospheric pressure also

increases. Because of this increase in pressure, the

altimeter indicates an elevation less than the actual

value. Altimeter manufacturers provide simple tables

or graphs for determining temperature corrections

which must be applied to altimeter readings. A rule of

thumb is that a 10 ‘C increase in temperature results in

a 4% apparent decrease in elevation. Therefore, in this

case, the actual elevation is obtained by increasing the

elevation obtained from the altimeter by 4% (102).

In certain parts of the tropics, the variations in pressure

and, therefore, in altimeter readings during the day are

very regular and remain nearly the same day after day

over fairly long periods. A typical daily variation is

shown in Fig. 2.28. In this case, altimeter readings

change as much as 7 m/hr, and this may introduce a

major error in head measurements if not accounted for.

-I

0 3 6

9 Noon 3 6 9 12

Time of Day

Fig.

2.28.

The

daily variation

in

pressure

recorded

at a

meteorological office

in

Kakete, Kenya (30).

Another factor affecting altimeter readings is the

change in pressure accompanying a change in weather.

These changes can be gradual or sudden. Tropical

storms and depressions can cause variations of 15 m or

more within half an hour. These storms can also be very

local, with considerable variation in pressure over

several kilometers (30).

When high humidity accompanies high temperature,

relative humidity also affects altimeter readings. The

altimeter manufacturer specifies corrections for this as

well.

If only a single altimeter is used, the difference in ele-

vation between two points can be determined by reading

the altimeter and thermometer first at one point and

then the other. The true difference in elevation is

obtained by applying a temperature correction to each

elevation reading and then subtracting the two cor-

rected altimeter

readings.

If the temperature remains

the same, the temperature correction can be applied

directly to the observed difference in elevation to

obtain the actual difference. Because changing atmos-

pheric conditions affect altimeter readings, it is possi-

ble to determine whether a significant change in these

conditions has occurred during the measurement period

by taking a reading back at the first point and compar-

ing that with the original reading.

Because changing atmospheric conditions do affect

altimeter readings, delay between readings at different

points should be minimal. Readings during windy or

inclement weather should be avoided, because these are

indications that atmospheric conditions vary rapidly and

widely. The best results are obtained two to four hours

after sunrise

or

before sunset.

If daily pressure variations are regular and therefore

fairly predictable, this fact can be used to improve the

accuracy

of

elevation measurements made using a single

altimeter. These variations can be determined by a

series of observations taken over the course of a day at

some point in the center of the area

to

be studied and

plotted against time. If the time when altimeter read-

ings are made in the field is noted, these readings can

be corrected by removing the effects of the regular

daily changes in pressure (or elevation) during the day.

If two altimeters are available, keeping one in a fixed

location or “base” permits changes in pressure caused by

changing atmospheric conditions to be recorded inde-

pendently from those caused by changes in elevation.

Altimeter and thermometer readings are made at 5-10

minute intervals at the base. A “roving” altimeter, pre-

viously compared with the base instrument, is taken to

other points to determine the difference in elevation

between these locations and the base station. When the

roving altimeter is read, time, temperature, and altime-

ter readings are recorded simultaneously. Elevation

readings from the roving altimeter are then corrected

for changes in temperature and atmospheric conditions.

If a single altimeter is used without correction, errors

can amount to 3-30 m or more. With corrections made

for regular daily pressure variations, this error can be

reduced considerably. If two altimeters are used, errors

can be reduced to about *l m per measurement.

Instrument temperature can also affect the altitude

reading, but altimeters are often compensated for this

effect and corrections need not be applied unless the

temperature changes by more than about 10 ‘C.

Measuring head and discharge 25

III. STREAMFLOW cIIARAcTERIsTIcs AND

DESIGN FLOW

INTRODUCTION

In planning a micro-hydropower scheme, the total

flow

required to generate the desired power with the head

available at a specific site can be derived easily using

the power equation [Eq. (4.2)]. One or more turbines

then convert the power available in this flow into

mechanical power. In preparing turbine specifications,

it is necessary to know the flow each turbine will

accommodate during normal opezn. This figure is

known as the “design flow” for that turbine.

If the flow required for power generation is always less

than the annual minimum flow in the stream to be

tapped, determining the design flow for the turbine(s) is

straightforward. This is briefly covered in SCHEMES

USING LESS THAN A.NNUAL MINIMUM STREAMFLOW

(pa 27).

However, power generation at some sites may some-

times require flows greater than the minimum flows

found in the stream. In these cases, it is necessary to

know actual streamflows over the year, especially dur-

ing periods of low flows. The larger part

of

this chap-

ter, SCHEMES USING GREATER THAN ANNUAL

MINIMUM

STREAlWLOW

(p. 28), addresses this issue.

This section begins by describing procedures for gather-

ing streamflow data at a gaged site and processing it

into a hydrograph and flow-duration curve. This discus-

sion is included to familiarize the reader with some of

the basic concepts involved.

Unfortunately, regular gagings have been carried out at

very few potential micro-hydropower sites, and even in

these cases, data is sparse. Furthermore, there may not

be sufficient time and manpower to gather the addi-

tional data before implementing a micro-hydropower

scheme. Therefore this section also explains several

approaches for estimating the mean annual flow, mini-

mum flow, and flow-duration curve for an ungaged site.

These assume that streamflows originate from

rainfall

and not from melting snows. Depending on the quality

of the data available, these estimates may be very

rough, but they still provide a basis for decision-making.

This section continues by describing how the power and

annual energy potential for a run-of-river scheme can

be determined from a flow-duration curve. It concludes

by presenting

several

simple turbine configurations

intended to reduce equipment cost and sophistication

and by reviewing the implications of these approaches

on the Bower and energy potential of a run-of-river

plant.

SCHEMES USING LE!X THAN ANNUAL MINIAAUM

STREAMFLOW

The simplest micro-hydropower scheme is a run-of-river

installation requiring a flow that is always available

from the stream. In this case, a dam would be included

only to increase available head, if necessary; it would

not be needed to create a reservoir for storing water. If

a single turbine

were

used, its design flow would simply

equal the required flow. Because a precise knowledge

of the streamflow variation is unnecessary, there is no

need to collect data over an extended period. However,

an estimate of flood flows might be useful in order (a)

to design an adequate dam, weir, or intake structure

that can withstand these flows and (b) to place the

powerhouse floor sufficiently above flood stage.

On occasion, the possibility of using two turbines rather

than one might be considered, for example, when a plant

serves

a remote hospital and some power must be

assured all the times. Using two turbines increases the

probability that some power will always be available,

because if one turbine breaks down, the other will still

be available to run essential services.

When two turbines are used for this purpose, the design

flow for the first turbine would be selected to enable it

to generate at least the minimum power necessary to

meet the the end

users’

critical needs. The design flow

for the second would then be selected so that the sum of

the two design flows equals the full design flow-the

amount required to generate the maximum power

desired. With two small turbines, using turbines of

equal capacity has several advantages: cost can proba-

bly be reduced, one set of spare parts can be used to

service either unit, or if parts are temporarily lacking,

one unit can be cannibalized to service the other.

Although using more than one turbine has advantages,

doing so would generally be more expensive-two 40 kW

turbines, for example, would usually cost more than a

single 80 kW turbine. In addition, if each turbine drives

its own generator, appropriate governing devices must

be incorporated to ensure that the two units can be, and

remain, synchronized; otherwise, two separate power

distribution networks must be used.

Streamflow characteristics and design flow

27

gaging site and because of the variation of rainfall

over

the catchment, flow variations tend to average out over

larger catchments (Fig. 3.1). Streams with small drain-

age areas may therefore require more frequent gagings.

rdx.EMES USING GREATER THAN ANNDAL MINIMUM

STREAMFLOW

If the peak power output of a powerplant requires a flow

greater than the minimum streamflow, more detailed

streamflow information may be needed to specify design

flows. This information is usually reduced to graphical

forms known as hvdrographs and flow-duration curves,

which can then be used:

l

to ascertain the percentage of the year a particular

level of power cannot be generated hecause of insuf-

ficient flow and during which months this occurs;

l

to determine the required storage capacity of a

reservoir, if one is needed to meet expected power

and energy demands;

l

to select the turbine configuration that will generate

power of minimum cost; and

l

to predict the average annual energy that can be

generated.

Generally, the larger and costlier the scheme, the more

critical is a full knowledge of the flow patterns of the

stream to be used.

Gaged sites

Data collection

A gaged site here refers to

one

where streamflow data

have been gathered regularly over a period of time. The

frequency of these gagings represents a trade-off

between the accuracy of the conclusions drawn from

this data and the cost and effort involved in gathering

it. Decreasing this frequency saves time but also tends

to decrease the accuracy with which recorded flows

represent actual flows. For example, peaks in stream-

flow caused by heavy downpours are relatively brief and

likely to be overlooked; therefore, large flows tend to

be underestimated, and this can have a disastrous

impazt on civil structures which have not been properly

sited or designed. On the other hand, low flows vary

much less in magnitude and are of longer duration than

peak flows; therefore, they are more likely to be

recorded. However, if there is too much time between

gagings, low flows may also pass unnoticed, and there

may be unexpectedly insufficient streamflow during the

dry season to generate the power which is expected by,

and possibly essential to, the end user,

Although measurements taken at six-hour intervals give

a more precise picture of the actual flow than daily or

weekly gagings, there is a point beyond which the

increased accuracy of the data does not warrant the

increased effort required to gather and process it. To

determine a gaging frequency which yields adequate

data with the minimum effort, it is necessary to have a

general knowledge of the flow characteristics of the

area’s streams. If streamflow variations are gradual,

weekly gagings may be adequate. However, for small

catchment areas, which are common to micro-hydro-

power schemes, flow variations are often more pro-

nounced. Because of the time it takes for flows from

various points within a large catchment to reach the

28 Streamflow characteristics and design flow

drainage area

=56kmL

drainage area

= 1800 km2

0 location of

gaging stations

QA (m3/S)

A

QB (m3/s)

1

ig. 3.1. Comparison of dally flow Wriations of streams

with

small and large cotchments

in the same

geogmphical

region

during

the

dry season

The gaging stations are

about 20 km apart (15).

The regularity of gaging is as important as its fre-

quency. This may be daily, every other day, or every

_

Monday, for example, but not at times which bias parti-

cular flows, such as only when the weather is nice, the

sun is shining, and readings can be taken in comfort.

Rainfall, and thus streamflow patterns, occur on a

yeariy cycle. Therefore, it is also necessary that data

gathered cover yearly increments of time-one, two, or

five years, for example-and not two months

or

half a

year,

If data can be gathered only for one year, it would be

advisable to have a longer-term record of rainfall over

or near the catchment area to help ascertain whether

that year was significantly wetter or drier thaa the

Inversin R. Allen Micro-hydropower Sourcebook

Подождите немного. Документ загружается.