|
|
RESERVOIR SEDIMENTATION EFFECTS ON
HYDROPOWER GENERATION - A CASE STUDY
Newton
de Oliveira Carvalho
ELETROBRÁS, Centrais Elétricas
Brasileiras S.A.
Rua da Quitanda, 196 - 22o
andar
20091-000 - Rio de Janeiro, RJ -
Brasil
Phone (021) 514-5430 - Fax (021)
516-4462
e.mail:
newtonc@eletrobras.gov.br
ABSTRACT
The analysis of the existing
hydroelectric power plants in Brazil shows that, in many cases, the
sedimentological studies didn?t reach the adequate extension or weren?t even
made. This procedure can be partially explained by the fact that 80% of the
53000 MW installed in hydroelectrics by 1991, including the total capacity of
Itaipu, are concentrated in almost 20 power plants, the reservoirs volumes of
which are bigger than 1000 hm3. With the perspective of building
power plants with smaller reservoirs, the sedimentation problems must increase.
In this paper the case of the Mascarenhas power plant, in the Doce river, ES,
operating since 1974, is being analysed. The project of sedimentological
studies of the reservoir was not included. It?s shown that the problems which
occur nowadays could have been forecasted during the project study, if data and
technology available on that occasion had been used. The results show the
importance of making sedimentological studies before the dam construction.
Keywords: Sedimentology, Reservoir sedimentation, Hydrographic basin,
Hydropower generation, Useful lifetime the reservoir, Sedimentometric network
INTRODUCTION
Brazil is a huge country
presenting a surface area of 8 512 000 km2 with a population of
about 150x106 inhabitants. Its surface is traversed by many rivers,
including several important hydrographic basins. The Amazon, the biggest
hydrographic basin in the world, covers more than half of the surface of
Brazil, two thirds of its total catchment area. Most of the Paraná basin is
located also in the country. The Paraguay basin (1 095 000 km2) has
almost half of the area on brazilian territory. The Tocantins (757
000 km2) and the São Francisco (634 000 km2) basins are
located entirely in the country. Other big basins, tributaries to the Amazon,
are also entirely located in the country, as the Tapajós (490 000 km2),
and the Xingu (504 300 km2). The total average water discharge of
the brazilian rivers, which represents 8% of the sweet water available in the
world, is of 257 790 m3/s [3]. One of the main uses of the brazilian
rivers is for power generation, this representing 95% of the total generation
in the country. The total hydropower installed by 1996 exceeds 53 000 MW, for a
possible capacity of about 260 000 MW. In order to accompany the industrial
development in the country, the electric sector is planning to install several
hydropower plants in the next 15 years. Many reservoirs have been formed by the
dams constructed for hydropower plants. There are more than 300 reservoirs,
including those for small plants. However most of the hydraulic generation has
been done by less than 20 hydropower plants with big reservoirs, each one of
volumes higher than 1 000 hm3. The evaluated sedimentation time of
these reservoirs shows a useful life time much longer than the economical
useful lifetime of the plant. Probably due to this, the sedimentological
studies on reservoir formation haven?t got enough importance. Another aspect of
the problem is that until few years ago the rivers in the country presented a
relative small sediment load, with an average value of less than 250 ppm. However,
Brazil has several medium and small hidropower plants, presenting sedimentation
in their reservoirs, which had the generation capacity reduced or completely
interrupted [2]. For instance, the hydropower Funil, at the Contas river, had
its operation interrupted from January/1992 to March/1993, presenting a loss of
US$ 1 200 000. To recover its generation capacity it was necessary to dredge
the sediment deposit near the intake at expenses of US$ 220 000. No previous
references about sedimentological studies had been done for the reservoir. Although
the dam was the bottom unloading type it was never in operation [1]. Another
example is the hydropower plant Mascarenhas, on the Doce river, the dam
construction of which was finished in 1974. Surprisingly in 1979 the reservoir
was almost totally sedimentated. Due to this, generation is only possible nowadays
by dredging every day near the intake. In the project, no references of
sedimentological studies about the formation of the reservoir had been
mentioned. This paper shows that the present problems at Mascarenhas could have
been antecipated if a sedimentological study had been done before the formation
of the reservoir, using data from other basins nearby.
A CASE STUDY
Characteristics
of the Doce river basin
The catchment area
of the Doce river (85 028 km2) is one of the most prolific sediment
yielding in Brazil. The intense rains and the high slopes in the basin are the
main natural causes for the high soil erosion in the region. As the main
antropic causes there can be mentioned the mineral iron exploitation of the
surface soil and the disforestation for getting wood coal for siderurgic
purposes. By travelling in the region it is possible to see several eroded
areas, presenting soils completely unprotected and slopes slippers along the
roads [1]. The reservoirs formed by the basin and also the channels of the
rivers are sedimentated. Every hydraulic construction in the river presents
problems due to the high sediment load. Along all the water course of the Doce
river and its tributaries the cities and properties have been seriously reached
by the floods because the sedimentated channel cannot support any longer the
volume of water. Almost all the sedimentated reservoirs in the rivers of the
basin were formed for hydroelectric power plant as Brecha at the Piranga river,
Peti at the Santa Bárbara river, Piracicaba and Sá Carvalho at the Piracicaba
river, Salto Grande Complex at the Santo Antônio river, Bretas at the Suaçuí
Pequeno river, Dona Rita at theTanque river, also Mascarenhas at the Doce
river, and others. In spite of the ocurrence of so many problems, the studies
didn?t make any reference to sediment transport nor recommendations for
sediment control. The dam doesn?t have any structure to atenuate the
sedimentation effects, as, for example, an unloading bottom [1].
The
sedimentation problem
The hydropower
plant Mascarenhas, of the ESCELSA, hydroelectric company of the Espírito Santo
State, which was constructed on the low Doce river in order to generate 120 MW,
has been operating since 1974. The dam has a height of 33 m and the location
has a catchment area of 74 300 km2, corresponding to 88% of the
total area of the basin. The water capacity of the reservoir at its maximum
level is of 42,1x106 m3 and the dead volume is only of
8x106 m3. At Mascarenhas the river has an average of 927
m3/s of water discharge [1]. In 1979, after the big flood of
February, 9th, with a water discharge of 14 000 m3/s, the reservoir
was almost completely sedimentated. The sediment discharge in that wet year was
of 30x106 t/year, more than half the capacity of the reservoir,
considering a unit weight of 1,2 t/m3. But the reservoir was already
sedimentated in the five previous years of high sediment load as computed
later. The mud and sand transported by the river during the flood reached the
intake easily and flowed through the intake to the turbines and equipments,
causing operation difficulties for the power generation. The height of the
sediment deposit near the dam achieved 27 meters, remaining a little more of
3,0m of water depth.
Present
evaluation of the useful lifetime of the reservoir
Mascarenhas? small reservoir
of 7 km length, presenting a lengthened shape, has a small trap efficiency of
sediment. This permits the suspended sediment load to be discharged through the
spillway and the intakes of water. So the planners of the dam had not worried
about the reservoir sedimentation. They also forgot the fact that the bottom
sediment load of the river couldn?t be discharged through the spillway,
remaining in the reservoir. In order to evaluate the sediment discharge at that
time, the planners could only get sediment yield data from measurements in the
rivers of adjoining catchment basins because no entity had any sedimentometric
networks in the Doce river before 1974, except for the station of Tumiritinga. The
average values of water discharge and total sediment load were plotted to get
the relation presented in Table 1 and Figure 1.
As can be seen the
presented values are very scattered, giving a poor rating curve, due mainly the
very few data.
The resulting
equation
Gst
= 25434 Q0,7722 (1)
which gives for
Mascarenhas a total sediment load of 5,0x106 t y-1, for a
water discharge of 927 m3 s-1. It is not so far from the
value at Tumiritinga, upstream of Mascarenhas, on the same river, being a good value for the dam location.
|
River |
Station |
Period |
Water discharge |
Total sediment
load |
|
|
|
|
(m3 s-1) |
(106 t y-1) |
|
Santo Antônio |
Ferros |
1970/71 |
50 |
0,18 |
|
São Francisco |
Porto das Andorinhas |
1960/61 |
210 |
2,50 |
|
Pará |
Porto Pará |
1960/61 |
145 |
0,50 |
|
Paraopepa |
Ponte da Taquara |
1960/61 |
120 |
1,29 |
|
Indaiá |
Faz. Bom Jardim |
1960/61 |
46 |
2,05 |
|
Velhas |
Várzea da Palma |
1967/70 |
290 |
3,50 |
|
Borrachudo |
Faz. Matinha |
1960/61 |
190 |
0,51 |
|
Doce |
Tumiritinga |
1973/74 |
600 |
4,20 |
Table 1 - Total sediment load
in catchment areas in the vicinity of the Doce river (measurements before
Mascarenhas dam construction)
Figure 1 - Sediment transport curve obtained from data of adjoining
stations of the Doce river
EVALUATION OF THE
RESERVOIR SEDIMENTATION
The following
equation can be used to compute the sediment load entering the reservoir each
year:
(2)
which,
S = annual volume of sediment
entering the reservoir, in m3 y-1
Gst = total
sediment load, in t y-1
t = trap efficiency
of the reservoir
gap = apparent unit
weight, in t m-3
The trap
efficiency of the reservoir, using the Brune curve [4], shows a value near
zero, meaning that Mascarenhas could dispose of all the sediment load by the
spillway and conducts. However this is true for suspended sediment loads but
not for the heavy coarse bottom sediment loads.
This coarse
sediment load can have a value equal 2 to 150% of the suspended load [4]. Adopting
20% of the total, there results a value of 1.0x106 t y-1
for the bottom sediment load which remains in the reservoir. This value may
represents a volume of 0,77x106 m3 y-1,
considering an apparent unit weight for the coarse material equal to 1.3 t m-3.
As mentioned
before, the reservoir capacity is equal to 42.1x106 m3
and the dead volume to 8.0x106 m3. Dividing the volume of
bottom sediment load by this value there results a time of 55 years for total
reservoir sedimentation and 10 years to sedimentate a volume equal to the dead
volume.
These evaluated
values don?t consider the high sediment load during floods, however they
demonstrate that the useful lifetime of the Mascarenhas reservoir could be
small for the planned lifetime of operation for power generation.
If this evaluation
had been done before the dam construction a sediment control could have been
done, by providing the dam with an unloading bottom.
CONCLUSION
The results show
the importance of making sedimentological studies before the dam construction. Brazil
doesn?t have a tradition in sedimentological studies due to not very high
sediment loads in the rivers. However this is changing due to the crescent use
of the soil as a consequence of the high population increase. For a near future
several small and medium powerplants with the formation of their reservoirs are
being planned. So, it is necessary to improve sedimentological studies in order
to have better reservoir operation.
It is necessary to
review the basic sedimentometric
network and install a secondary network. Nowadays the brazilian basic network
is composed of less than 300 stations for suspended sediment measurements
presenting a very poor frequency of operation. It is necessary to improve the
frequency of measurement and to increase the number of stations. Also to get a
better geographic distribution of the stations besides making bottom sediment
load measurements.
The secondary
sedimentometric network can be installed on the rivers on which it is planned
to construct the dams. They may be operated only for a few years just to get
enough data to make the local sedimentologic studies. Several others studies
have to be made, as well as potential erosion investigation, for instance. These
studies will indicate the procedures for sediment control in the basin, in the
catchment area of the reservoir or even in the dam.
REFERENCES
1)
ALMEIDA, Sérgio Barbosa, e CARVALHO, Newton de Oliveira (1993). Efeitos do
assoreamento de reservatórios na geração de energia elétrica: análise da UHE
Mascarenhas, ES. X Simpósio Brasileiro de Recursos Hídricos e I Simpósio de Recursos
Hídricos do Cone Sul. ABRH. Gramado, RS. Brasil.
2)
CARVALHO, Newton de Oliveira (1998). Assoreamento e proteção de reservatórios.
VI Simpósio Nacional de Controle de Erosão. ABGE. Presidente Prudente, SP.
Brasil.
3) DNAEE,
Departamento Nacional de Águas e Energia Elétrica (1994). Disponibilidade Hídrica
do Brasil. Brasília, DF.
4).OLD,
International Commision on Large Dam (1989). Sedimentation Control of
Reservoir. Guidelines. Paris, France.
5)
STRAND, Robert (1974). Design of Small Dams. Sedimentation. US Bureau of
Reclamation. Washington, DC. USA