MICROFICHE

AT
A project of Volunteers
REFERENCE
LIBRARY
in Asia

Solar Photovoltaics for Irrigation Water Pumping

by Urs Rentsch

Published by:

Swiss Center for Appropriate Technology (SKAT)

Varnbuelstrasse 14
CH-9000 St. Gall
SWITZERLAND

Available from:

same as above

Reproduced by permission.

Reproduction of this microfiche document in any

form is subject to the same restrictions as those
of the original document.
 Working Paper WP 7182
Document de Travail
Document0 de Trabajo

SOLAR PHOTOVOLTAICS FOR IRRIGATION WATER PUMPING

Urs Rentsch

St. Gallen 1982

SlMiT VarnbiielstraLk 14, CH-9OLlOSt.Gallen,

Switzerland.Tel.071233481
 SOLAR PHOTOVOLTAICS FOR IRRIGATION WATER PUMPING

Contents:

a) Background information

b) Characteristics and constraints

C) Design optimization and matching

d) Reduc~ing solar array costs

e) Economic considerations 11

f) Conclusions 14

.g) Suggested further reading 15

h) Suppliers of photovoltaic pumps 16

I-~-“. ~~~~
SOLAR PHOTOVOL-TAICS FOR IRRIGATION

a) Background information
WATER PUMPING

The economic development of rural communities in most developing countries is

heavily dependent upon increased agricultural productivity. An:alarming fact is
that~the~ feeding of the worlds population will getmore and more difficult in
the near future, as populations is rapidly growing and the fertility~of the soil
is decreasing~ in many areas.

Irrigation is widely recognized as a feasible yield-increasing technology wh~ich

could play a key role in improy,ing food production. In ~areas; where irrigation
.~ with gravity-fed water through controlling the flow of r~ivers and suitable ca-
nal distribution. systems is not possible, the need for a cheap; readily availa-
ble energy source is obvious.

Since conventional mechanized irrigation through engine-driven pumps is getting

more and more expensive with climbing fuel prices, the~use of solar radiation
for water pumping is becoming interesting.

There are several technical options to~convert solar radiationto electricity

or mechanical energy. The most advanced option and probably the most. feasible
one for the iannediate future isthe use of solar (ore photovoitaic) cells;

klthough~'scientifically.complex, pbotovoltaic cells are conceptually quite

simple. They.convert the suni.ight:~directly to electricity. By j~oining large
numbers.of these cells ~together.{modules-and arrays);significant: amounts of
power~can be generated whenever the sun shines.

~The main advantageof the photovoltaic~ generator .is its reliability..,Its. life
timeis ex~pected to reach 20 years. Other advantagesare; that the electrical
energy can' be, generated~ right at the.site of.use and ,that it hasvery, low'run-
fling costs, i.e.~ low.maintenance,requirements and no fuel costs.

d use of this' technology, however, is its high

in the range.of US $,lO tom 15 per peak Watt

s are today competitive with othersystems only

.
sa high priority and other energy sources
plications a.re for telecommunication, auto-
recordingstations, water pumping,, soSartelevi,sion, solar re~frige-
ts. Of these applications,~wa.ter pumping has.
ive use if the prices are coming down.

photovoltaic pumping system is the solar

PV-pump are mainly the ,same, as of PV-ge-
re severalpoin,ts to'mention especially~:
 0 There is another point related to the varying demand of a year: A solar
pump has to be sized to be'able to satisfy the peak irri~gation demand for
,a given area and these conditions may only last for a relatively short
time. At other times, considerable surplus capacity will be available.
This is due mainly to variations in crop water demand, changeover between
crops, times when no crops are in the field or when rainwater is supple-
menting~irrigition water. The result of this mismatch between power avai-
lability and demand is that only .p$rt of the available solar energy can
be used for irrigation. In other words, these variations lead to an over-
sizing of the pump. In a ty ical, but rather good situation, the utiliza-
tion factor is about~ 40 7:. 17 I see fig. 1) Therefore, with high costs of
solar pumps, they can today only be successful in areas where substantial
oversizing due to extreme variation in irrigation water dema.nd can be
avoided.

Due to the high capital.cost of the solar pump, there are at least today and in
the .immediate future eren more serious cons~traints, limiting its application
to fewer,areas. As the capital cost (C) is fairly proportional to the installed
power capacity, and the power demand is proportional to the product of the to-
tal head (H), (height by which the water must be lifted, plus pipe friction
and other energy losses), and the maximum flow rate of water (Q), the capital
cost is:

C = 'k . H . Q (where k is a constant)

o There is a proportionality between the price of water and the total head,
as the price of.water is roughly proportional toC/Q. Increasing the head
makes solar pumps increasingly 'uneconomic and also incompetitive with
motor pumps. Presently available systems costing overZO$/Wpkcan only be
economically viable for heads of little more than 1 m.21 (With a rather~
high assumption of a 6 US. cents/&water cost.)

. The relative capital costs of diesel and gasoline engine pumps are de-
creasing with increasing power demand, i.e< increasing land area or water
demand. On the other hand, the capital cost of photovoltaic pumps is pro-
portional to the installed power capacity. Therefore, solar pumps tend to
be more competitive with fuel engine pumps for quite small daily water de-
mands. In the near future;this will be in the range of 15 to 30 m3/day
(to suit about 0,3 to 1 ha).?)

~. The above two constraints indicate, that the most potential users of solar
pumps in the near future are small farmers living in areas with high
groundwater level. These water regimes exist imarily in the large allu-
vial deltas of Asia, North Africa and Arabia. % However, since these
people normallypossessno capital resources of their own, the capital
cost of the solar pumps are a crucial factor in their adoption by farmers.

1) Halcrow, Small-Scale
W., Solar-P&?ered Irrigation Pumpinq Systems, Teciizical
and Economic Review, London 1981
2~) Halcrow,. W., Technical and Economic Review
3) Tabors, R.D., The Economics of Water Lifting for Small Scale Irrigation in
the Third~World: Traditional and Photovoltaic Technologies, MIT, 1979
 4

The provision of finance to assisst the poorer farmers to meet this first
cost is an important aspect of the transfer of the technology. If govern-
ments or related national institutions don't take financial actions i?
this direction, the market for solar photovoltaic.irrigation pumping is
'very restricted.

It is rather difficult to say something on the social and cultural accep-

l tability of solar pumps, as until now very little experience is available
on this' aspect. Most actual installations are notrepresentative because
they are established in special situations. Obviously, solar,pumps are
alien elements in rural areas, representing a very high imported techno-
logy. The best chance to succeed is.given in situations, where irrigation
IS a familiar technique and a certain industrial infrastructure is al-
ready existing. There are a number of barriers to the local manufacture
or part-manufacture in developing countries. Particularly, photovoltaic
cells need to be manufactured on a rather large scale to bring unit costs
down,~and developing countries often lack the capability to establish
and sustain such an,industry.

Several factors lead to the ~expectation, that capital costs will fall and pho-
tovoltaic pumping systems will get more competitive in the future. From these
factors, the following will be discussed in the following sections:

. Edith more experience, design will improve and matching will be optimized,
leading to an increased efficiency and reduced required at-;-ay size.

. Several technical improvements in solar cells and their manufacturing

process will make them cheaper.

. Increasing fuel prices will make solar pumps more competitive with conven-
tional motor pumps.

. Economies are resulting from an increasing scale in manufacture and grow-

ing markets.

c) Design optimization and matching

The simplest photovoltaic pump consists of a photovoltaic array, a DC-motor and

a.pump: These elements have to be optimally chosen and matched to each other to
give an optimal efficiency.~To further increase efficiency in certain applica-
tions; electrical control devices and batteries could be used. Today, the opti-
mal achievable efficiency of the total system consisting only of a PV-array, a
motor and a pump is about 4,5 %. :!

This is derived from following (optimal ) efficiencies cf the individual compo-

nents:4)

-
4) Ha!alcrov, w., Small-Scale Solar-Powered Irrigation Pump.fng Systems. Phase I
Project Report, London 1981
 5

array cells 11 % at 25" C (10% at operating temperature)

connections 95%
motot- 85 %
pump 55 z
pipework 95 %
total system 4.6 %

However, existing installations mostly have a lower overall efficiency. 'This is

not due, in general, to a wrong calculation of the &-array size but rather to
a poor matching of the characteristics of motor, pump and the~generator to the
hydraulic requirements.

Another problem are bad pumps, being the weakest part in the whole system. At
present array costs, a one peer cent marginal increase in pump efficiency.(with
good pumps in the 40 - 50 % ef~ficiency range) would be'worth S
(electrical) system e::d 803fora.400 W system in arrav savings. gg :::r:f::: :he
choice of a good pump is very important to achieve a decrease of costs.

In the discussion of the criteria for a choice of the elements, only the most
promising applications for solar.irrigation pumping shall be regarded, i.e.
small-scale low-head solar pumping.

For the selection of the pump, the following ~criteria are most important:

s The pump should- have a very good efficiency being at least about 40 -
50 % (under field conditions).

. The pump should be able to establish and maintain prime on the suction
side of the pump. In existing applications the loss of prime (e.g.
through a cloudy period) is one of the major troubles. The pump should~
also be capable of running dry without damage.

. In many situations, the pump will be subject to wear'from particles su-

spended in the water. It must be able to pass suspended solids without
damage or clogging.

e Another factor are variations in head that might occur in the level of
groundwater. With low lift pumping in the 3 - 5 m range, even 1 m varia-
tion represents a large percentage of load change. The pump should be
able to cope with,this without requiring adjustments.

Low starting torque requirements enable the use of a larger part of the
l daily solar energy and exclude the need for a,battery.

. The characteristics of the pump under varying conditions (head and flow,
which can be drawn in a diagram as a current voltage characteristic of the
motor-pump), should match well with the optimal power points of the PV-
array under varying insolation.

From the large number of available types of pumps,, the submerged centrifugal

5) ffalcrow, W.; Techfiical and Economic Review

 6

pumpmeets
__ the above requirements best. Other pumps are excluded, because they
are not self-priming, have too low efficiencies for low-head applications (po-
sitive-displacement pumps) or cannot be matched well to the PV-array characte-
ristic (positive-displacement pumps). The, submerged version of the centrifugal
pump is preferred, because then no foot-valve is needed,

For proper selection of a pump motor, the following criteria should be observed:

. Asfor the pump, efficiency is of the utmost importance in minimizing the

required solar array size.

. The motor has to serve the needs of the centrifugal pump. Maximum opera-
ting speed is between 3000 - 4000 rpm.

. Maintenance -and reiiability are in remote areas and under tropical en-
vironment of major sign.ificance.

o Th,e matching of the motor input to the PV-array output should be as

simple as possible.

The last point is being best fulfilled by DC-motors, since no DC/AC-converter

would be needed. The lower cost of AC-motors~ is not profitable, because they
normally have lower efficiencies than DC-motors in the fractional horsepower
range. DC permanent magnet motors are most attractive, as they.are more e.ffi-
cient than series or shunt wound DC motors, particularly under part-load con-
ditions. The,main maintenance requirement is occasional brush, commutator and
bearing replacement. Brushless motors (with the magnets in the rotor and an
electronically commutated stator) are available, but in the tropics the' elec-
tronic circuitry ins still too vulnerable and needs 'good cooling. Compact sub-
mersible motor-pump units could prove valuable for solar pump applications.

Obviously optimal matching of the PV-array,, DC-motor and centrifugal pump is of

utmost importance for PV-array size reduction and predictability of operation.

The pump and motor characteristics can be drawn in different diagrams. For a
given head, a centrifugal pump will provide water, if the speed of rotation is
greater than a threshold value. With increasing speed, the torque will grow
steadily as the water output grow.s. Th~is situation is represented in fig. 2a.

When the motor.and pump are coupled, speed and torque of both motor and pump
must be coincident at any given moment. The points of coincidence can be drawn
as a curve also in 'the current-voltage (I/V) diagram of the motor-pump system.
Because in the operating range of the DC-motor, current is fairly proportional
to torque, and voltage to speed, fig. 2b is very similar to fig. 2a (figures
adopted from 6))

6) Matlin. R.W., Design Optimization and Performar~T Characteristics of a PV-

Micro-Irrigation system for Use in Developing Cnantries, MIT, 1979
 TORQUEb-in.)

I
10~ 20 30 40 50 60 70 80
v lvo\fs)
Fig:2a: Torque-speed for centrifugal Fig. 2b: Current-Voltage for ,motor-
pump at different heads pump system

I (amperes)
m
1000 W/m2

750 WlmZ

500 W/m'

?\
m

IO 20 30 40 50 60 70 Si,
v (volts!
Fig. 2c: Working curves of PV-array Fig.~~2d: Matching of PV-array and motor;
at different irradjance ~pump. system
levels

On the other hand, the possible working.points of the PV-array at different ir-
radiance levels can also be drawn in an :/V diagram (fig. 2 c).For ev~ery irra-
diance level, one point on the I/V curve signifies maximum power delivery. A
curve (m-m) can be drawn through the maximum power points of every irradiance
revel.

I'n~order to maximize the output of the system under all operating levels of
irradiance, the load~~(I/V curve of motor-pump system at a given head) should
coincide as closely as possible to the maximum power curve .of'the PV-array.
 (Thiscan be achieved by connecting the PV-modules in a certa.in manner;i.e. in
the right relation of parallel connections to in-series connections.) In prac-
tice it is common to match load and array for just one level of,irradiance,
such as say 800 ~W/m2. fit lower levels of irradiance, the subsystem power demand
causes the array~to operate off the optimum point (fig. 2d).~

The correct matching of subsystem and array is a complex analytical process

which is-greatly facilitated through the use of a computer model, which permits
an iterative approach to be used.

Two points are worth mentioning:

. For every application a new optimization must be done. This means also,
that the system is *very difficult to optimize, if the water headvaries
or not all relevant factors are known at the time of optimization.

. The system operates only atone irradiance level optimally, while having
a lower efficiency at all other irradiance levels.

These disadvantages ha&led to the development of an electronic optimizing

circuitry called maximum power point tracker (MPPT). The MPPT is an automati-
cally controlled DC-DC-converter (or impedance matching device) which changes
the voltage/current ratio (impedance) until the optima~l power point of the
array is found.

AS the MPPT consumes a certain amount of power (4 - 7 %17) the efficiency at'
the.design point will' be lower than without MPPT, but at other irradiance le-
vels it will be better.

Therefore, despite its high.price, MPPT's may be~profitable in ~situations where

conditions vary much (irradiance level, array temperatures, pumping head). The
MPPT allows manufacturers to sell standard pumping systems which still function
optimally under significantly different conditions.

Batteries could also be used for,power'conditioning. But because the require

relatively much maintenance (distilled water), are heavy (movable so f ar pumps:)
and expensive, their use is not recommended.

To maximize the use of available solar irradiation; tracking of the PV-array

could be considered. Usually the collector is fixed, facing south in the nor-
thern hemisphere or facing north in the southern hemisphere. At noon, the array
surface is normal to the sun's rays, collecting the most possible energy. In
the morning and Ian the evening, however, ,less of the available energy is collec-
ted, because the sun's rays don't fall normal on the collector.

If the array is tracked from east to west, almost the fully intensity of sun
!ight (around lOOOW/m2) can be harnessed from the moment when the sun rises to
its setting. Automatic tracking is not suitable in developjng countries, as hit
complicates then system,~ increases cost and reduces reliability. But it has been
shown, that through manual tracking at least 40~% extra energy could become

7) Natlin, R.N., 1979

 available!) Moving,the collector twice per day is only 5 % less efficient
than automatictracking. As this manual tracking is probably necessary only
during a~short period (month of peak water demand), the array could be sized
smaller and the utilization factor increased without much effort and cost.

d) Reducing solar array costs

Solar cells have been known since 25 years, used mainly for spacecraft, and in
the past few years are being~increasingly used for electricity generation on
earth. The. basic material for solar cell manufacture is silicon, a material
derived from silicon dioxide which is found abundantly in nature as sand.

Silicon is a semiconductor which usually does not conduct electricity, but can
do so, if a sufficient amount of light or heat is present. Through the inciden-
ce of a photon, an electron can get enough energy to.escape from its place in
a silicon-atom/An electrical field is necessary, to transport this electron
to the surface of the semiconductor and thus to permit a current of,electrons
to flow through the silicon crystal. This can be achieved, by doping the.sili-
con crystal on both sides with different specific impurities (the result is
called pn-junction). For collecting the generated charge-carriers at the surface
of the silicon, metallic electrodes are soldered onto the cell. The front side
electrode has the form of a grid, to cover as little of the light-absorbing sur-
face as possible.

Most solar cells used in present applications are fabricated from high-purity
monocrystalline silicon by the following process: The purified silicon is mel-
ted and a perfect crystal is;grown drawing a seed crystal slowly from the melt.
The silicon crystal is~then sawed into thin wafers of'about 0.25 mm thickness.
A pn-junction is created in each wafer by ,exposing its surface to a gas of
,doping material. Finally the electric contacts are attached and the cells are
encapsulated with a transparent substance to.protect them from air and water.

The described producti,on'method yields cells with efficiencies between IO % and

15 %. The th oretical maximum efficiency of the monocrystalline silicon cell is
about 23 %. 97
I

I- The methods to produce solar

in the~electronic
immature. However,
industry
celis from monocrystalline
since many years.
the trend for manufacturing
silicon have beenused
Thus, they cannot be said to be
electronic components was to
pack more and more elements on one chip to reduce the required silicon area.
For this reason, the price of silicon doesn't play the same role as in the so-
lar cell application. Certain stepsin the manufacturing process could be opti-
mized to reduce solar cell costs. The most expensive steps are the following:

. Very pure monoc,rystalline starting material is needed to achieve a good

efficiency. The reasons is, that the charge carriers in the silicon are
trapped by impurities and grain boundaries (grains are small crystals with
various orientations in a polycrystalline silicon).

87 Halcrow, W., Technical and Economic Review

9) Howell, Y., Adler, D., How Sil,icon Cel'ls brk, Sunworld, Vol 4/.No:l. 1980
 10

. About 50 % of the expensive starting material is lost during fabrication,

mainly through cutting the silicon crystal to wafers.

. The crystal growth and the diffusion process for doping the wafers are
.very slow.

. The production and fabrication methods are very energy-intensive. The

energy payback period of monocrystalline cells is said to be in the range
of its lifetime. The solar cell costs will be~much dependent on fuel
price increases without marked reduction of the energy input.

. Many steps in the production process are not yet automated.

Several methods being developed at the moment, could lead to a cost reduction
of the solar cells. They aim at decreasing processing energy requirement, re-
ducing the amount of s;arting material, automating manufacture etc., while
maintaining or even increasing the solar cell efficiency. Among the most pro-
mising and most discussed are the following:

. In single-crystal silicon cells manufacture, much of the expensive star-

ting material could be saved, if the~cuttinq of the crystal to wafers
could be changed, ,by drawing a ribbon of crystal silicon from the melt
instead of a~cylindrical crystal. This technique has been demonstrated
in the laboratory and would not only reduce costs but enable automatic
(and therefore large-scale) production of solar cells. Ion-implantation
of dopants instead of the sloti.high-temperature diffusion process would
reduce energy requirement. both methods increase damages in the crystal
structure (which can partially be repaired) and therefore cause a re-
duced cell efficiency.

. Attempts have been made'to weaken the single-crystal requirementby pro-

ducing solar cells from thin layers of polycrystalline silicon. Poly-
crystalline silicon is comprised of many small silicon crystals with
various orientations. However, as grain boundaries trap the'charge car-
riers, both, current through the .solar cell and efficiency,dqcrease. This
problem is reduced -if the grain boundaries are.perpendicular to the pn-
junction. Today, polycrystalline silicon cells are available (in pilot-
scale) with efficiencies.around 10 %.

. Crystalline silicon is a rather poor absorber of visible light. Thus, the

solar cells should have a thickness of at least 100 micrometers. Onthe
other hand, amorphous silicon (Si-atoms are not ordered in a crystal
structure) is a good absorber of visible light so that a thin film of
1,micrometer would be sufficient to absorb the light. Amorphous silicon
cells need much less starting material and do not pose the same problems
concerning purity and crystallinity as the crystalline solar cells. The
manufa~cturing process is~ much.cheaper,.because amorphous silicon is pre-
pared by the condensation of silicon vapour onto a cold substrate. How-
ever, the transport of electric charges through the amorphous solid is
more difficult, r sulting in a poor efficiency of about 5 % in the la-
boratory today.lOe

10) for further details, see Adler, D.,, Amorphous Silicon,Solar Cells, Sunworld,
vol. ‘g/No. 1, 1980
 11

. A fourth possibility for array cost reduction is concentration of the

sunTight onto a small spot, which needs less area of solar cell surface.
for s,mall-scale applications in developing countries, however, only flat
mirrors without tracking mechanism are feasible. This solution will lose
its attractivity if cell costs decrease.,

In all methods discussed above, reduction of the cell costs below ,a certain
limit is only possible at the expense of a good efficiency. Many cost factors
such asland, encapsulation , 'structure, wiring, installation and transport are
strictly proportional to the installed panel area, and therefore inversely pro-
portional to the efficiency of the array. Thus;it is doubtful if solar cells
with an efficiency.below say 10 %~become economic, even if they are very
cheap.11)
.

e), Economic considerations'

Most presently available solar pumping systems are costing more than 30 $/Wpk
iprices for delivered systems; one american supplier sells his systems much
cheaper). For a typical system of 250 Wpk, CIF-prices would be about IO'OOOUSS
of which two thirds are solar array costs.

Of course, these prices will decrease, ~if solar pumping is becoming more attrac-
tive. However, 'the prices cannot fall short of a certain limit. With solar cell
costs of zero, the price of the solar array will be ate least in the range of
0,5 - 1 $/Wpk~. (At 15 % efficiency and a price of 75 - 150 S per m2 of array
area for encapsulation,, structure, wiring etc.) In addition to this, about
2 $/Wpk will have to be:paid for motor and pump. (about 300 S each for 300 W
rated ~power).~ Therefore, including transportation and installation, the price
of solar ~pumps will hardly drop below 3 - 5 $/Wpk in the future.

Irrigations Ts economically viablelonly if the cost of water can be ~covered by

an ~increased,incoaie from crop production. The economically acceptable price of
water depends very much on local variables such as, crops and climate, and is
economically viable ,today'in the range of 1 - 5 US cents per m3 pumped water.
In fig. 3 water prices for different costs of~solar pumps are drawn, varying.
with head.

Two:conclusions may be drawn from this figure:

. Today solar irrigation pumping 1s not economically viable for all heads
practically occuring.

l With a price of about 1 US cent/m3 per m pumping head, which corresponds

with rainimal solar pump costs of~3 - 5 S/Wpk, solar irrigation ~?umps
could become economically viable only for pumping heads below 5'- 6 m.
Fig. 3: Effect of pumping head on water unit,costs for a daily output of.
40 m3/day .CGXC,~: H.&~ow, w., TeChnical and EconomiF Review

In the near future, the solar pump will compete with mainly two groups of pum-
ping systems: conventional motor pumps (diesel, ,gasoline) and traditional pumps
(human and animal powered). This competition depends very much on the power re-
quirement, thus on the area of land to be irrigated.

Today, solar pumps cannot compete with diesel pumps for areas above a quarter
ha. However, this situation will change, if array c~osts'decrease and fuel costs
increase. Fig. 4 shows the relation between water costs and irrigated area for
different fuel prices and solar pump costs. Solar pumps will first become com-
petitive for irrigation of land areas smaller than ,I ha. Later, they could be-
come competitive even for all areas.

For poor farmers on small land holdings, the important~ questionis not, whether
the so1a.r pump could compete with.diesel pumps, but whether it could compete
with traditional pumping methods. Recently, efficient low-cost manual irriga-
tion pumps have been developed. Width the Rower Pump for example, two people
could pump 40.m3 of water with a 5 m head in 8 hours. i*) This would be enough

I21 Rcwer Pump, Mirpur Agricultural Workshop and Training School, Dacca,
Bangladesh (l,eaflet)
Fig 4: Effects ~of water demand on waterunit costs for 5 m pumping head

Source: Halcrow, W.: Technical and Economic Review

to irrigate an average area of 1 ha. Animal powered irrigation pumping' is prac-

tised since centuries in many areas of~the world. An economical comparision
with.these techniques is difficult, since it 'is very dependent on local'condi-
tions (fodder, labour costs, bullock rent etc.). However, the,following consi-
derations indicate, that solar pumps will hardly, be able to compete with tra-,
ditional methods:

. The attitude of poor farmers. towards risk and uncertainty (high discount
rate) favours pumping systems with low capital costs. The Rower Pump for
example costs about 10 - 13 $, a corresponding,solar pump at least 600~ 5
(with a 3 S/Wpk minimal assumption).

. For the same reason, operating costs (labour, fodder, rent) count less in
the farmer's.view,than capital costs.

., In many rural areas labour and animal power is abundantly available.

The above considerations reflect the point of view of potential buyers and thus
represent a financial analysis. An economic analysis (from the point of iew
of governments), however, could showy more.benefits for the solar pump. Id If

13) French, D., The Econo!nics of Renewable Energy Systems for Developing'
Countries, Washington, i97V .
 prices of solar pumps decrease enough, governments could take measures to intro-
duce credit mechanisms or subsidies in favour of solar pumps. Unless this is
done,,solar irrigation pumps will hardly come into.widespread use in the near
future.

f) Conclusions

Solar pumps are today not economically viable for irrigation purposes. Neverthe-
less,'the market for solar photovoltaics will~grow as there are other economi-
cally viable applications in remote areas, where reliability is very important.
Pumping~ of drinking water for instance belongs to this group of applications.
The growing market and technological improvements will lead to decreasing solar
cell prices. With increasing fuel costs, solar pumps will become competitive
with motor pumps at least for small land holdings~(but only for low pumping
heads).

-Small farmers with little purchasing power will hardly be in a position to af-
ford solar pumps, even if the solar cell costs decrease dramatically. As is the
case with many other technologies, financing mechanisms will have to be devised
and subs~tantial subsidies will be required to make a',new technology accessib~!e
to the small and the very small farmer, who constitute the majority of the
rural population in the developing worl~d.
 15

g) Suggested further reading:

French, D., The Economics of Renewable Energy Systems for Developing Countries,
Washington D.C., 1979

Halcrow, We., ITDG, Testing and.Demonstration of Small-Scale Solar-Powered Pum-

ping Systems, State of,the Art Report, UNDP/World Bank, London,
December 1979

Halcrow, W., ITDG, Small-Scale Solar-Powered Irrigation Pumping Systems, Pha-'

use I Project Report, C:b!DP/World Bank, London, July 1981

Halcrow, W., ITDG,Small Scale Solar-Powered Irrigation Pumping Systems, Tech-

nical and Economic Review, UNDP/World Bank, London, September 1981

ISES, Sunworld, Vol. ~/NO. 1, 1980, Special Issue on Solar Photovoltaics

Lysen, E.H.; Pumping Water with Solar Cells, "Wind and Sun Compendium", No. 6,
Amsterdam, April 1981

Matlin, R.W., Design Optimization and Performance Characteristics of a Photo-

~voltaic Micrr-Irrigation System for Use in Developing Countries,
MIT Lincoln Laboratory, Lexington, 1979

Saunier, G., Folea, D., Electrical Solar E:tergy AppTications in Remote Rural
Area, Proc. of the Regional Asia and Pacific Workshop on the Appli-
cations of Solar Energy in Agricultural and Post Harvest Activities,
Bandung, 12 - 15 Jan. 1981

Smith, D.V.,Photovoltaics in the Third World, MIT, Massachusetts, 1979

Tabors, R.D., The Economics of Water Lifting for Small-Scale Irrigation in the
Third World: Traditional and Photovoltaic Technologies, MIT, Massa-
chusetts, 1979
 16

h) Suppliers of photovoltaic pumps (not complete)

Motorola Solar System

AEG-Telefunken
Solar,Operations, MG
Neue Technologien
P.O. Box 20924
Industriestrasse 29
Phoenix; Arizona 85036
D-2000 Wedel
U.S.A.
Western Germany
Omera-Segid
Applied Solar Energy
ll,, Boulevard du General Delambre
Corporation
F-95101 Argenteuil
Don Julian Road
France
15251 P.O. Box 1212
-City of Industry
Philips-Nedet-land B.V.
CA 91749
Afdeling Elonco
U.S.A.
Boschdijk 525
NL 4510 Eindhoven
Arco Solar, Inc.
Nederland
20554 Plummer Street
Chatsworth, CA 91311
Photowatt
USA 125, Rue du President Wilson
F-92302 Levallois Perret
Briau S.A.
France
B.P. 0903,
F-37009 Tours Cedex
Photowatt International S.A.
France
98ter, Boulevard Heloise
95,102 Argenteuil
E.N.E.
France
Van der Meerschenlaan '88
1150 Brussel
Pompes Guinard
Belgique
179, Boulevard Saint-Denis
F-92406 Courbevoie
France
Ferranti Ltd.
Electronic Component Division
RTC,
Gem Hill, Chadderton
La Radiotechnique Compelec
Oldhain OL9 &VP
130 Avenue Ledru Rollin
England
F-75540 Paris - Cedex 11
France
Klein, Schanzlin & Becker AG
adept. 'VHEZ
Siemens AG
P.O. Box 360 Unternehmensbereich Bauelemente-
D-6650 Homburg (Saar)
Balanstrasse 73
Western Germany
D-8000 Munchen 80
~t.lor+Qcn
,.L.,*I_I II a;@r,T#ar,~y~
Lucas Energy Systems~ Ltd.
Windmi~ll Road
Sensor Technology Inc.
Haddenham~Aylesbury
21021 ~Lassen Street
Buckinghamshire HP 178 JP
Chatsworth, California 91311
United Kingdom
Area Code (213)882-4100
U.S.A.
Solar~Electric International
483i Del Ray Avenue
Washington, D.C. 20014
U.S.A.

.Solar'Power Corp.
20 Cabot Road
Woburn, MA. 01801
,U.S.A.

'Solarex
1335 Piccard Drive
Rockville
Maryland 20850
U.S.A.

Spectrolab Inc.
12500 Gladstone Avenue
Sylmar
California~91342
~U.S.A.