Nyamasheke Micro Hydro
Initial Report Draft.
Dartmouth HELP Worldwide
Rural electrification in Rwanda has been determined as by varies development agencies
to be one of the main drivers of economic development in areas not served by the
national electricity grid (Brown)i. The area in and around Nyungwe National Park, with
large elevation change, and relatively high rain fall, is well suited to micro hydro
systems. There are many perennial rivers with potential to supply power to local villages
not connected to the grid or to supplement those with electricity.
For Microhydro Basics see Phillip Brown USAID report (P2).
Microhydro systems can range in size from several hundred watts to several hundred kW
and in cost from several hundred to a million dollars (USD). Currently in Rwanda there
are several engineering companies doing large scale microhydro instillations (>100 kW).
GTZ projects cost around $2000 to $4000 per kW. So a 100 kW system could cost about
$300,000. The drawbacks, aside from the high initial cost, of the larger systems is that
they require higher tolerances, needing skilled and expensive maintenance for non local
components, extensive civil works and equipment for construction affecting the
environment, and if they fail to perform due to technical (mechanical or electrical failure)
or natural problems (floods, less flow than expected, etc) a very large investment is lost.
These large systems also require a river with a large flow.
Smaller systems may be more appropriate for many areas but the large scale contractors
are generally not interested in them.
A hydro assessment was done of the rivers around Banda, Gisakura (in the Kamiranzovu
valley) and Rangiro. The hydro potential is:
Nyrabanda (above Kingange) 7.3
Nyrabanda (lower down, near
Buriri River (@ road bridge) 20.2
Sempundu (@ foot bridge) 0.5
Zihari (@ foot bridge) 1.5
Zihari (@ Sempundu Hydro) 11.6
Kamiranzoru (above Zihari
unnamed river (@ foot bridge) 1.0
unnamed river 2 (@ road
Rangiro Hydro 15.6
Rangiro River 15.6
*See appendix for detailed river descriptions and seasonal rainfall variation.
P=gHQ, assume =50%, and Q is half the available flow in the river.
It is obvious that there is plenty of potential available in the area. It seems that small
picohydro (<10kW) schemes are the most appropriate for the available rivers and for the
local population. There are many rivers with high head and low flow coming off the hills,
that are nearby to where people live. The village of Banda is the obvious choice for an
initial implementation as there is plenty of hydro potential, and it would have large
benefits to the community.
There are two locally implemented picohydro systems in the area, Sempundu Leonards
near Gisakura and at the Rangiro School.
Mr. Leonards system is very small (~1kW) system taking water from the Zihare River. It
has been working on and off for the last several years, at the moment it not working due
to erosion and collapse of the inlet channel on the river bank. The system powers, 10
lights in two houses and a radio charger. When he doesn’t need power the water can be
diverted out of the inlet channel to irrigate his fields. The whole system cost 800,000
The Rangiro school has designed and built their own hydro system to power lights for the
school dormitories, classroom, several houses and a church. The system has .4 cms of
flow (exactly half of the river flow) and 80 meters of head. A locally fabricated turbine
drives a 7.5 kW AC generator. The cost estimates varied hugely. The main components
(turbine, generator, penstock, transmission lines) cost 3.5 million fr (see appendix for cost
break down). The other quotes from Jeremie, the school headmaster in charge of the
project, varied from 5 million to 20 million to 50 million (A variation from $1,200/kW to
$12,000/kW). They school and town leaders are interested in a similar project on the
Nyirakesha River, several kilometers out of town.
At this point it seems like the best solution is to design a picohydro system that can be
made, repaired and installed with local labor and components, creating electricity
primarily for illumination. The design would seek to maximize efficiency, and robustness
while minimizing cost, and environmental impact. The goal would be to encourage
systems like Mr. Leonards through improved design, micro finance, incentives, and/ or
subsidies. Having many small systems is much more robust than relying on one large
system, resolves the issue of ownership, reduces environmental damage and could
stimulate the local economy instead of having an outside contractor do the construction.
A pilot project to demonstrate the manufacture, installation, use and repair of picohydro
systems in the Nynamashake region is recommended, Banda Village would be ideal for
this. If successful it would be easy to spread naturally to others areas, without large
investments.This would also act as an introduction to the benefits of hydro power to
locals. The environmental, technical and social issues could be determined and then used
to implement a larger system with more success. Multiple systems would also give a very
accurate representation of river flow and rain fall over time, important for agriculture and
A 10kW system is capable of powering:
120 25W bulbs of 500 6W bulbs
1 water pump
For the community such as Banda this would go a long way towards development.
The cost of a system large system implemented by GTZ is about $3000/kW. Systems
around 10 kW can be developed for between $1500 and $2000/kW. A goal of this project
would be to minimize the cost/kW for small systems through improved design and use of
local material and labor.
The cost of the Dartmouth HELP project, to implement a hydro system in Banda
producing around 10kW is estimated to be between $10,000 and $20,000, based on
previous similar projects and the cost of components.
At the moment the cost of electricity (Electrogaz) is $.24/kWh. 1kW generator would
produce about 8500 kWh/ year which if sold at current market prices would generate
$2000 annually. If this were possible then a $1000 system would pay off in 6 months.
Although net metering like this is not possible for small systems in Rwanda it shows the
value of the electricity produced in a microhydro system.
See Appendix E for project cost variation and estimation.
Appendix A: River Data
Ben Koons (Dec 07)
H P P to
Altitude Q low H low high Town
Name X Y (m) (m^3/s) (m) high (kW) (kW) (km)
Kinyange 746128 9729776 1923 0.12 10 50 2.92 14.6 1.4
Nyiragagigo 746450 9730296 1917 0.08 10 50 1.86 9.3 1.9
Nyrabanda (above Kingange) 746008 9729417 1762 0.20 5 15 2.45 7.3 1.2
Nyrabanda (lower down, near
church) 745768 9729433 1762 0.13 3 10 0.97 3.2 1.0
Buriri River (@ road bridge) 743656 9728947 1740 0.55 5 15 6.73 20.2 1.2
Sempundu (@ foot bridge) 732911 9730509 1802 0.03 3 6 0.26 0.5 0.5
Zihari (@ foot bridge) 733250 9730659 1739 0.10 1 6 0.25 1.5 0.75
Zihari (@ Sempundu Hydro) 733513 9731146 1635 0.30 12 16 8.72 11.6 1.3
Kamiranzoru (above Zihari
junction) 733769 9731134 1593 5.82 5 40 71.24 569.9 1.5
unnamed river (@ foot bridge) 734083 9730228 1772 0.03 2 12 0.17 1.0 1.3
unnamed river 2 (@ road
bridge) 731751 9729198 1848 0.09 2 12 0.42 2.5 1.4
Rangiro Hydro 739746 9735254 1770 0.04 40.0 80.0 7.81 15.6
Rangiro River 739746 9735254 1770 0.08 40.0 80.0 7.80 15.6
Nyirakesha River 741733 9734097 1590 1.45 4.0 40.0 14.25 142.5
Appendix B: Monthly Rainfall Variation (10 year average).
Monthly Rainfall (10 year average)
* Compiled from Nyungwe National Park Rain and Temperature Data.
Appendix C: River Variation
The following rivers were measured by Ben Koons in December 07 and by Phillip Brown
in August 07.
River ∆Q (fraction)
Buriri River (@ road bridge) 0.15
Buriri River (closer to town) 0.20
Zihari (@ Sempundu Hydro) 0.20
The rivers have about one quarter of the flow.
Appendix D: River Log
Travel to Banda. Long steep windy road in poor condition down to the village. Pop-5100.
Most subsistence farmers (beans, corn). Met cell (local area) leader, Nzabagurira
Emmanuel (speaks no English). Martin (forest tracker) showed me around the village and
to several rivers and waterfalls.
X 744795, Y 9729424
Elevation: 1741 m
Pop~5100 (cell leader), 4300 households (USAID)
Most subsistence farmers (beans, corn).
Cell leader, Nzabagurira Emmanuel (speaks no English).
The village center sits on top of a small hill and has a community center (Kageno) and
Primary school down by the Nyrabanda River. No electricity or running water. To get
there from WCS headquarters you have to take a long steep windy road in poor condition
down to the village.
The Kinyange river is very steep coming out of Nyungwe Forest until it joins the
Nyrabanda river. The steep waterfall section (coordinates marked) is inside the park and
the flow measurement is taken from just outside where it starts to level off. Possible
generator position 1.3km from Banda center (.7 km from edge). The terrain would make
it difficult for penstock placement.
Similar to Kinyange but assessed section entirely outside the park. Drops made up of
clean waterfalls and very steep sections. 1.9 km from Banda.
This river goes right by the town of Banda, the Primary School and Community center
are located near it. This river was assessed by Phillip Brown (USAID) in Nov 2007.
During the dry season it had a flow of .05 cms, while I measured .13 cms. This
discrepancy could be due to lower water levels or different locations (the river splits apart
and come back together many times, I summed parallel sections) or both. Although I
didn’t find much elevation drop the USAID paper reports a “head drop of approximately
15 meter is easily achievable”. If so then there is reasonable hydro potential. The
advantage of the Nyrabanda over the other 2 above is that it is closer to town, minimizing
Very similar situation as the Nyrabanda River. It enters the Nyrabanda downstream of
Very low flow, (.03 cms) and not very steep, very limited potential. Meets Zihari river
Initial flow measured at foot bridge was .10 cms, with a gradient of 1:10. Lower down (at
the intake of Sempundu Leonards system) the flow was .3 cms with a gradient of 1:5.
Sempundu Leonards Hydro System.
Sempundu Leonard has constructed a small hydro system to power his and his sons
houses. It consists of an inlet channel (~100m) to take water from the Zihari River to a
settling tank (1m X 1m X 2 m concrete tank) protected by a trash race (roofing iron with
2 square cm holes punched in it. At the bottom of the settlings tank there is a penstock
(75 mm dia) leading down about 14 vertical meters to a shed containing the turbine and
generator. When the generator is not in use the water can be diverted to irrigate his grops.
The system was said to have worked well for 2 years, then there was electrical trouble,
then mechanical trouble with the alternator, now it doesn’t work because the bank at the
inlet has been eroded away such that water no longer enters the system. The whole
system cost 800,000 fr (~$1500).
It would be great to get Mr. Leonards system working so the technology is accepted in
The Kamiranzovu River is the main drainage in the valley. The flow was measured at a
point just above where the Zihari River enters to be 5.82 cms. From Mr. Leonards house
on the hill you can see a very obvious drop upstream of where we took flow
measurements, with approximately 20 meters of head over a very short distance,
unfortunately due to terrain and time constraints we could not make it to this exact point
to examine it more closely. We did make it further upstream to where the river exits the
park where there are more falls and rapids. The total head from where the flow
measurements were taken to the park boundary was 50m.
The Kamiranzovu river has great potential to be a fairly large generator of power. It
would not be hard to generate 200-400 kW of power. Unfortunately the area measured is
very hard to get to and for such an operation where there would be a lot of civil work
getting equipment and materials to the site would prove costly.
Unnamed River 1
This river flows down into the Kamiranzovu valley. The flow measurements were taken
at X 734083, Y9730228. It has very similar flow and vertical drop as the Sempundu river.
Unnamed River 2
This river is on the opposite side of the drainage of the Kamiranzovu Valley (X 731751,
Y 9729198). With reasonable flow and gradient it could produce between .2 and 2 kW,
depending on the head. It is quite close to Gisagura.
Sunday 12/16. Stay at WCS headquarters in Gisakura and work on river data/ report.
Monday 12/17. Travel to Rangero to see school hydro system. The system has been
working off and on for several years, it is working now. It was designed, built, and is now
used by the students of the school in town. It seems to be a quite good design.
Rangiro Hydro system
Alfa AC synchronous generator
Type: ST 7.5, 7.5kW, 230V, 32.8A, 50Hz
Cost: 600000 fr ($1100)
Approx 12” runner
Cost: 500,000 fr ($920)
L=19*6m = 114m (generating 80m of head)
Cost: 1,520,000 fr ($2800)
16mm *120m (@4000/m)
10mm* 120m (@3000/m)
Cost: 840000 fr ($1540)
Stone and concrete
Main components: 3,460,000 fr ($6350)
The flow into the penstock is .04cms, exactly half of the flow that was measured
upstream of the intake. In the dry season more water is diverted into the intake.
With 80m of head and .04 cms of flow the system should have a theoretical output of
15kW (assuming a 50% turbine), twice of what the generator is rated for.
This river is 1.9 km away from the center of Rangiro, it was selected as a potential hydro
site by members of Rangiro. It has similar head potential too Rangiro river and more
flow. Placement for the penstock and generator would be significantly harder though as
the area is steep and undeveloped. It is also further away from the town, and very
Was told on my return by Rick that this river has significant potential. It is between
Rangiro and Banda.
Appendix E. Cost Estimation.
Option 1: Standard ~10kW system
Based on experience in the United States and elsewhere, the cost of the 10 kW system:
Turbine and generator: $6,000
intake, pipeline, and housing $2,000.
The breakdown for the Rangiro School system (7.5kW) is;
The total cost was quoted at being between $10,000 and $40,000
The cost of the systems outlined above is heavily dependent on local labor cost,
community involvement, and availability of materials.
Option 2: Low Cost Picohydro.
For the smaller system, it is estimated that this system would be $500 to a maximum of
$1,000. These costs could be considerably less if used materials and equipment are
Team size: 3-6
Plane ticket: ~$1700 each
Length of stay: 4-8 weeks.
$7000 + food, accom. and transport.
$27000 + food, accom. and transport
Brown, Hydropower Electricity Generation Evaluation, Nyungwe National Park and
Vicinity. USAID-IRG. Aug. 2007.