TP# 20: 02/85
SOLAR WATER PUMPS
C. J. Swet
Paul E. Dorvel
John D. Furber
Wilson Boulevard, Suite 500
Arlington, Virginia 22209 USA
703/276-1800 * Fax:
This paper is one of a series published by Volunteers in
Assistance to provide an introduction to specific
technologies of interest to people in developing countries.
The papers are intended to be used as guidelines to help
people choose technologies that are suitable to their
They are not intended to provide construction or
details. People are
urged to contact VITA or a similar organization
for further information and technical assistance if they
find that a particular technology seems to meet their needs.
The papers in the series were written, reviewed, and
almost entirely by VITA Volunteer technical experts on a
Some 500 volunteers were involved in the production
of the first 100 titles issued, contributing approximately
5,000 hours of their time.
VITA staff included Maria Giannuzzi
and Leslie Gottschalk as editors, Julie Berman handling
and layout, and Margaret Crouch as project manager.
C.J. Swet, the author of this paper, has a background in
and is a consultant in solar and other "gentle"
with special emphasis on energy storage.
He has 20 years
experience in the field of solar energy, and has consulted
solar energy and other appropriate technology projects in
countries. He has
published several papers on solar energy
and other energy related topics.
Reviewers Paul E. Dorvel, John
D. Furber, and Daniel Ingold are also experts in the field
solar energy. Paul
E. Dorvel is currently Associate Principal
Engineer in the Power Systems Division of the International
He has over seven years experience in Africa
doing market research and field engineering for solar
John D. Furber is President of Pleasant
Valley Software Corporation and Starlight Energy
frequently lectures and consults overseas on solar energy
Daniel Ingold is a biophysicist by training and a
research engineer at Appropriate Technology Corporation.
VITA is a private, nonprofit organization that supports
working on technical problems in developing countries.
information and assistance aimed at helping individuals and
groups to select and implement technologies appropriate to
maintains an international Inquiry Service, a
specialized documentation center, and a computerized roster
volunteer technical consultants; manages long-term field
and publishes a variety of technical manuals and papers.
For more information about VITA services in general, or the
technology presented in this paper, contact VITA at 1815
Lynn Street, Suite 200, Arlington, Virginia 22209 USA.
UNDERSTANDING SOLAR WATER PUMPS
VITA Volunteer C.J. Swet
This paper examines water pumping systems that use solar
as a direct source of energy.
We look primarily at small-scale
rural applications in the Third World, where the potential
benefits are greatest and the near-term economics seem most
favorable. Two generic
technical approaches for solar water
pumping systems will be examined:
(1) thermodynamic (in which
the radiant energy is first converted to heat); and (2)
(in which it is first converted to electricity).
photovoltaic technology is more mature, it is used for
comparisons with other methods of pumping water.
of this complex subject is necessarily cursory; the aim is
provide prospective users with sufficient insight to
whether solar water pumping is a plausible option for their
specific situation, and to furnish a guide for further
The history of thermodynamic solar water pumping technology
back nearly four hundred years, when Solomon deCaux in
raised water for a fountain by the expansion of solar-heated
By the beginning of this century, many of the currently
design concepts had already been explored, and several
attempts at commercialization were under way when activity
due to the advent of the internal combustion engine and
cheap commercial fuels.
Much of this development had been aimed
at relatively large-scale applications.
It was not until the
resurgence of interest in solar energy caused by the
oil embargo that greater attention began to be directed at
rural applications in developing countries.
(*) Of particular interest to the serious reader in this
the the definitive and comprehensive work done in
with the United Nations Development Programme/World Bank
Solar-powered Irrigation Pumping Systems Project by Sir
William Halcrow & Partners and the Intermediate
Group, Ltd. The most
important reference for the prospective
solar pump user is Handbook on Solar Water Pumping (see
In contrast, the development of photovoltaic water pumps has
heavily emphasized small-scale rural applications since the
1960s. While both of
these technical approaches continue to
mature, neither can yet be judged inherently superior.
the recent efforts, however, have concentrated on
systems and a majority of solar pump field operating
has been with photovoltaic installations.
NEEDS SERVED BY THE TECHNOLOGY
Solar water pumps may be used for irrigation, community
supply, livestock watering, and in various industrial
In principle they may be used virtually anywhere, but the
compelling needs and opportunities are found in the
sun-rich rural areas of the Third World.
Solar pumps may also be
almost any size, but most small farms, villages, and animal
in developing countries require hydraulic output power of
than a kilowatt.
Many of these potential users are too far from
an electrical grid to economically tap that source of power,
engine-driven pumping tends to be prohibitevely expensive as
as unreliable due to the high cost of purchased fuel and
maintenance and repair capabilities.
Developing countries increasingly require less costly and
reliable methods of pumping water that do not rely on
Solar water pumping is one of several potentially
satisfactory alternatives, which also include wind, water,
biomass, animal, and human power.
Greater use of irrigation is clearly needed to increase the
from existing cultivated land and to permit cultivation of
marginal or unusable land.
Nearly all this additional
irrigation will have to use pumped water, since most of the
available sources of gravity-fed water are already fully
In Third World countries, most irrigated land is in family
plots of less than four hectares, a large proportion of
being under one hectare.
This practice can be expected to extend
to lands not yet under cultivation, since small plots have
found to be more productive than large farming units in
yield per hectare although more demanding in terms of labor
Daily demand for pumped irrigation water varies widely depending
on the season, crop, stage of growth, region, method of
distribution, and water management effectiveness, with
values ranging from about 20 to 120 cubic meters per hectare
day. Water sources
include ground water from open (dug) wells or
boreholes that are surface water from rivers, ponds, or
and typically at depths of two to 10 meters below ground
Irrigation water is seldom extracted from depths greater
meters because the value of its benefits is seldom high
justify the extra cost of deeper wells and additional
energy. To be
economically feasible for agricultural applications,
the cost of water delivered must be less than the value of
the benefits obtained through use of the irrigation water,
through improved yields or by enabling more crops to be
year. In 1982 a
global norm for the cost ceiling for water delivered
to the field (not to the crop) was approximately US$0.06
per cubic meter, although clearly the actual figure in a
situation will depend on the crops grown, the field
efficiency, and market prices.
If 60 percent of the pumped
water is used by the crop itself (a fairly typical condition
earth furrows), the cost ceiling for that water would be
per cubic meter.
Most villages in developing countries have fewer than 1,500
and in many of those villages the per capita water
consumption is far less than the 40 or more liters per day
desirable from a health standpoint.
A common reason for this low
consumption is that all the water is drawn from a single
resulting in lines, and in larger villages, the need to
water considerable distances.
Although multiple dispersed wells
will alleviate these problems, polluted sources become more
to avoid. Ground
water is usually extracted from depths
of 30 meters or more because its value for human consumption
much greater than that for irrigation; water sellers in
countries often command a price equivalent to more than
US$3.00 per cubic meter for 10 to 30 liters per day.
application, the economic feasibility of solar water pumping
much less a factor than its competitive position relative to
other methods of mechanized pumping.
For livestock-watering in remote areas, daily demand per
varies widely depending on breed and type of forage; about
liters is fairly representative for dairy cattle.
years many engine-driven borehole pumps have been installed
this purpose, pumping from depths as low as 30 meters.
herd of cattle and a village of the same population may have
comparable pumping power requirements.
However, for this application
it is often desirable to have multiple dispersed pumps in
order to minimize overgrazing near each watering place.
considerations accent the need for pumps that can operate
when unattended for long periods of time, and that do not
require secure fuel stores at each watering place.
Although thermodynamic and photovoltaic solar water pumping
systems are conceptually similar in that both are powered
directly by solar radiation, their operating principles are
following discussion highlights the distinctive
features of these systems.(*)
All thermodynamic systems use a solar collector to convert
radiation to heat and a heat engine to convert the heat to
power for pumping.
In heat engines a fluid or gas absorbs
heat at a higher temperature, which causes it to expand; it
then contracts upon removal of the heat at a lower
This expansion and contraction is harnessed to move a
piston in a cylinder, or it can expand against a turbine
Figure 1 illustrates the basic energy flows, showing
the necessary temperature differential across the heat
and the unavoidable losses associated with each stage of the
process. Of the two
indicated temperatures, the lower one cannot
be lower than that of the pumped water to which the unusable
degraded heat is typically rejected, while the upper one is
largely controlled by the type of collector.
upper temperature (within practical limits) raises the
system efficiency and reduces the required collector size,
usually at the cost of greater complexity or more expensive
broadly characterized conceptual approach can
have many different embodiments, with various types and
of collectors, working fluids, heat engine cycles, engines,
and pumps, as discussed in Section III.
These systems exploit the photovoltaic effect to convert
radiation to direct current electricity, which powers a
pump. A basic
photovoltaic system layout is shown in
Photovoltaic conversion occurs when light falls upon a
thin, flat material called a solar cell.
One side of the cell
becomes electrically positive, and the other electrically
This is a solid-state, electronic effect.
Like a transistor,
the solar cell has no moving parts except electrons.
(*) Detailed explanations of these operating principles can
in the publications listed in the bibliography.
As long as light falls on the cell, the electrons flow as an
electrical current through an external circuit containing
solar cells are connected in series strings to
obtain the desired output voltage.
Series strings can be connected
in parallel to obtain the desired output current of a
modules are then interconnected and mounted.
Photovoltaic array output current and power--to the extent
is constant--vary linearly with solar irradiance.
and power output decrease with increasing cell temperature on
order of 0.5 percent per [degrees] C above 28 [degrees] C.
Figure 3 shows the performance characteristics of the
of a typical photovoltaic pumping system, illustrating the
of proper matching of the electical source and the hydraulic
load over a range of operating conditions.
components and configurations of these systems are discussed
Most of the small-scale systems that have been developed beyond
the prototype stage use Rankine cycles similar to the one
schematically in Figure 4, with organic working fluids such
Freon 11 and slow-speed reciprocating engines that directly
piston pumps. Many
developing regions are familiar with Rankine
systems because of experience with steam engines.
fluids can produce higher heat-to-work conversion
than steam at temperatures up to their stability limit of
about 150 [degrees] C, but extreme care must be taken to
ensure zero leakage
since very small amounts incapacitate the system and
is difficult in the field.
A reciprocating engine is virtually
the only choice,
since turbines and rotary expanders are
excessively expensive in the small sizes of interest.
reciprocating (piston) pumps tend to be more efficient than
conventional high-speed centrifugal pumps at heads greater
about 10 meters, although single stage centrifugal pumps
are easy to make) are well suited for very low-head irrigation.
The system depicted in Figure 4, which was designed by a
company, has a trough-type concentrating solar collector
follows the sun by rotating about a horizontal north-south
Sun following is automatic, powered by the shifting weight
solar-heated Freon and controlled by a sun shade mounted on
collector, but the orientation must be reset manually each
The collector has a 12 square meter aperture, occupies 16.5
square meters of ground space, and weighs 170
kilograms. All of
the other above-ground components occupy a 0.4 cubic meter
and weigh about 50 kilograms.
Aluminum is used extensively for
both weight reduction and corrosion resistance.
At a Freon
temperature of 107 [degrees] C the reported delivery rate is
liters per second (roughly equivalent to 40 cubic meters per
if the pump operates eight hours) against a total pumping
14 meters. The
reported rate is nearly five liters per second
(100 cubic meters per day) against a head of three meters.
above-ground pump location limits the use of this system to
shallow wells or surface water sources involving suction
greater than about eight meters.
A somewhat similar system from West Germany has about 40
meters of stationary flat plate collectors that can heat the
Freon 11 working fluid to about 90 [degrees] C.
Its pump can be located
below grade and is adaptable to wells up to 60 meters deep.
Preliminary testing in India indicates a delivery rate of 40
cubic meters per day against total pumping heads of 15 to 20
meters. The pump is
sized to permit greater output when larger
collectors are used.
Water (steam) has some important advantages as a Rankine
working fluid. It
can be used at higher temperatures than are
possible with organic fluids, to achieve higher
Also, the consequences of leakage are far less severe.
firm has developed a 2-kilowatt uniflow reciprocating steam
engine powered by a glass strip reflector trough collector.
However, higher temperatures require greater optical and
precision, which increases the cost per unit collector area
tends to offset the size reduction made possible by improved
economic competitiveness of high-temperature
Rankine cycle solar pumps is still a under contention.
Stirling cycle heat engines offer perhaps the most promising
means of exploiting the very high temperatures (over 500
[degrees] C) that
can be obtained with point focusing collectors, such as
Sunpower Inc. in the United States has developed
a free piston Stirling engine with an integral diaphram
using helium as the working fluid.
In tests by the manufacturer
with a simulated solar thermal input of 1 kilowatt
to the output of a dish approximately 1.4 meters in
the Stirling engine delivered 2 liters per second at 560
against a head of four meters.
At its present stage of development,
however, it is easily damaged, and test results have been
Another promising Stirling engine pump is the
"Fluidyne" liquid piston system being developed by
company, but no solar version has yet been demonstrated.
Many other technically intriguing and potentially useful
have been or are being developed, including:
smaller organic Rankine systems;
very small (about 25 watts) steam Rankine
an organic vapor liquid piston pump;
a heated air liquid piston pump;
a fluid overbalancing rocking beam engine
various solid state systems based on
and the differential expansion of bimetal
Some of these systems have become commercially available,
must be emphasized that none of them (or of the other
described above) is known to have successfully undergone the
extensive testing under field conditions that characterize a
Ability to manufacture and repair such technology often depends
on the region. The
several systems presently being developed in
India presumably would be manufactured there, and an Indian
affiliate of the West German company is evaluating the
of locally producing all or part of the German system.
does not mean, though, that these systems of Indian design
or would be manufactured elsewhere in the Third World.
frames, conventional heat exchangers, and some types of
collectors could be made and repaired in many developing
but reciprocating engines and piston pumps of high
call for close tolerances that may not be readily achievable
with available skills and equipment.
A number of types and sizes of photovoltaic systems are
commercially, in various stages of product development, that
the range of pumping needs outlined in Section I.
design variations of these systems are fewer and more easily
presented than those of the relatively immature
variations center mainly on:
the choice of solar cell material;
the choice between stationary and
the choice between planar and concentrating
the type of electric motor;
the type of pump; and
the method of source/load matching.
All commercially available systems use crystalline silicon
cells, of either the single crystal or polycrystal
types of solar cells, which may be less expensive, are under
use thin films of semiconductor materials,
such as amorphous silicon or cadmium sulfide.
solar arrays produce roughly 100 watts per square meter
most favorable conditions.
Specific pumping needs do not influence
the choice among these competing designs.
In most systems the solar arrays have a fixed orientation;
are tilted permanently toward the equator at an angle that
energy collection during the season of maximum demand (or
for the year if demand is fairly constant).
This is the simplest
and most affordable configuration, but not necessarily the
in terms of cost per unit of delivered water.
slightly greater first cost and complexity, the orientation
be manually adjusted several times during the day, thereby
the daily output by up to 30 percent.
This tends to be
cost effective provided that manual labor is available and
inexpensive for highly seasonal irrigation
applications. If the
system is used over most of the year, a fully automatic
device may be justified.
Although such systems have not yet
demonstrated sufficient reliability under field conditions,
recent field operating experience with gravity-driven Freon
trackers on pumping systems has been encouraging.(*)
As long as solar cells remain the dominant cost item there
incentive to reduce the required area not only through sun
but also through concentrating the intercepted solar
radiation. The cost
decrease due to further cell area reduction
tends to be offset by the added cost of concentrating optics
the need for better cooling of the cells and more precise
If solar cell prices diminish as predicted, the incentive
will become much less compelling.
Permanent magnet direct current motors are the most commonly
pump drivers for small-scale systems.
Alternating current motors
cost less but are much less efficient in the sizes of
Linear actuators have been used to drive piston pumps, but
concept requires considerably more development.
Many of the
direct current motors in current use are of the conventional
brush type, which is efficient but poorly suited for
operation and needs brush replacement after every few
(*) See for example Dankoff, W., "Pumping
Water," Solar Age, February
1984, pp. 29-35.
hours of use.
Electronically-commutated brushless direct current
motors are finding favor because they require less maintenance
and are more readily adapted to submerged operation,
they are slightly less efficient.
Single-stage centrifugal pumps are frequently used when the
pumping head is less than 10 meters, and are either
or (if the suction lift is too great) submerged.
With open wells
or surface water sources, these pumps and the motors can
thereby minimizing the suction lift.
For higher heads, either
multistage centrifugal or positive displacement (piston or
cavity) types are most efficient.
If the pump is above
ground or floating, it usually is closely coupled to the
if submerged, the pump may either be closely coupled to a
motor or driven by a vertical shaft.
pumps ordinarily are submerged except in cases where the
small but the total pumping head is high.
Single-stage centrifugal pumps can be made with
characteristics that fairly closely the solar array
characteristics, so that the array can operate at near-peak
efficiency over a wide range of operating conditions.
matching cannot take place with multistage centrifugal or
For systems that are not inherently
compatible in this respect, it is possible to install an
impedance matching device between the array and the motor
that will automatically optimize the load on the array.
devices, called maximum power point trackers or maximum
controllers (MPCs), will increase daily pumped output and
allow pumping to start under low moring irradiance.
controllers add to the complexity and cost of a system, in
to creating an approximate five percent power drain on the
are that MPCs are most cost-effective in
systems over about one kilowatt peak capacity.
Below this level,
it may be more cost-effective to substitute extra array
for an MPC.
Local production of nearly all components except the solar
appears possible in many developing countries.(*) India and
have begun cell manufacture and some other countries are
the assembly of modules from imported cells.
cell technology is advancing so rapidly, and crucial choices
among the candidate semiconducting materials have yet to be
(*) For an in-depth discussion of the potential for local
see Small-Scale Solar-Powered Pumping Systems:
Its Economics and Advancement, by William Halcrow and
Partners, and Intermediate Technology Power, Ltd., and its
documents concerning manufacture of solar water pumps in
the less developed countries (June 1983).
it may be prudent for most of the Third World to await the
emergence of clearly superior design approaches before
in production facilities.
Other than solar cells, the most
doubtful items for potential local manufacture appear to be
piston pumps, because of their close machining tolerances.
IV. COMPARING THE
To gain widespread acceptance, small-scale water pumps must
only deliver water at a cost below the value of that water;
must also do so at a lower cost than alternative methods of
pumping water. Most
prominent among the currently available
alternatives are diesel or kerosene engines, wind power,
and humans. A good
basis for comparing them with solar is the
cost per unit volume of water delivered under like
over a like number of years.
This takes into account costs of
purchasing, financing, delivery, installation and start-up,
operating and maintenance labor, repairs, and
United Nations Development Programme studies, comparative
of delivered water have been estimated for irrigation,
water supply, and livestock watering in Kenya, Bangladesh,
Thailand (see bibliography).
Based on 1982 prices, some typical results are shown in
Figures 5, 6, 7, and 8.
There are many qualifications to these results,
too numerous and complex for exposition here.
It must be pointed
out, though, that the solar costs are based on photovoltaic
systems in which the dominant cost item is the solar module
its 1982 price of about US$8/ peak watt.
That price is likely to
become much lower within a few years, making the solar
more competitive. It
should also be noted that the attractively
low wind power costs are based on average mean wind
for each country; within those countries there are regions
totally inadequate wind.
The costs for animal power are somewhat
optimistic because they do not include the cost of diverting
animals from other activities, and the "low case"
are based on unrealistically optimistic values for fuel cost
these points in mind it seems fairly clear
that within a few years solar will be quite competitive in
sunny regions that have little wind.
GLOSSARY OF TERMS
Aperature. The solar
Dish collector. One
in which the reflecting surface is a paraboloid
that concentrates direct solar radiation
absorber at its focal point. Usually
above 250 [degrees] C, with two-axis tracking.
distance the water level in a well is temporarily
Flat plate solar collector.
One in which the aperture is essentially
the area of the absorber surface, the
surface is essentially planar, and no concentration
employed. Usually for temperatures
below 100 [degrees] C.
Hydraulic output power.
The power imparted by the pump to the
proportional to the product of the flow rate and the
pumping head. In watts, roughly equal
to liters per
meters times ten.
Irradiance (radiation intensity).
The energy flux density in the
radiation, usually expressed in watts per square meter.
Static head. The
vertical distance between the water source
level at no
flow and the point of discharge.
Suction lift. The
height that water must be lifted from the
to the pump.
Total pumping head.
The static head plus drawdown and flow pressure
One with a cylindrical parabolic reflecting
concentrates direct solar radiation onto an
(usually a tube) at its focal line.
from 100 [degrees] to 250 [degrees] C, tracking about one axis.
Overall system efficiency.
The fraction of intercepted solar radiation
imparts pumping energy to the water, i.e., pump
output power per unit aperture/irradiance.
BIBLIOGRAPHY/SUGGESTED READING LIST
Halcrow, William and Partners, and Intermediate Technology
Small-Scale Solar-Powered Pumping
Systems: The Technology,
and Advancement (United Nations Development
Project GLO/80/003). Washington,
1983. Available through the World Bank,
following supporting documents:
Performance tests on improved photovoltaic
Economic evaluation of solar water pumps
Potential for improvement of photovoltaic
Review of solar thermodynamic pumping
Manufacture of solar water pumps in
Small-Scale Solar-Powered Irrigation Pumping Systems
Development Programme Project GLO/78/004,
report). Washington, D.C.:
World Bank, July 1981.
Small-Scale Solar-Powered Irrigation Pumping System
Economic Review (September 1981), amplifying
Handbook on Solar Water Pumping (United Nations Development
Project GLO/80/003). Washington, D.C.:
February 1984. This handbook directly
issues and methods of selecting, evaluating,
and specifying a
solar water pumping system.
Kreider, J., and Kreith, F., eds.
Solar Energy Handbook.
1981. The reader is referred to the
Chapter 1 for
history of solar thermodynamic water pumping
Chapter 7 for
non-concentrating solar-thermal collectors
Chapter 8 for
intermediate concentration collectors
Chapter 9 for
high concentration collectors
Chapter 22 for
solar powered heat engines
Chapter 24 for
United Kingdom Section of
International Solar Energy
Proceedings of Conference on Solar Energy
Volume on Refrigeration and Water Pumping.
ISES, January 1982.
United Nations Development Programme; World Bank; and
Energy. Proceedings of Workshops on
Countries. Washington, D.C.:
World Bank, June
LIST OF SUPPLIERS AND MANUFACTURERS
OF SOLAR WATER PUMPING SYSTEMS
THERMODYNAMIC SYSTEMS (not necessarily mature products):
organic Rankine flat plate,
approx. 500 watts output
7990 Friedrichshafen 1
FEDERAL REPUBLIC OF GERMANY
fluid overbalancing beam engine,
c/o A. de Beer
flat plate, approx. 200 watts
P.O. Box 349
REPUBLIC OF SOUTH AFRICA
fluid overbalancing beam engine,
c/o Pelegano Village Industries
flat plate, approx. 200 watts
P.O. Box 464
organic Rankine trough
P.O. Box 42
approx. 300 watts output
PHOTOVOLTAIC SYSTEMS (commercially available and fairly
AEG--Telefunken Raumfahrttechnik und Neue Technologien
2000 Wedel, Holstein
FEDERAL REPUBLIC OF GERMANY
Via Bergano, 21
ARCO Solar, Inc.
20554 Plummer Street
Chatsworth, California 91311 USA
A.Y. McDonald Corp.
P.O. Box 508
Dubuque, Iowa 52001 USA
133 Enterprise St.
Evansville, Wisconsin 53536 USA
Grundfos Pump Corp.
2555 Clovis Ave.
Clovis, California 93612 USA
Caixa Postal 8085
Sao Paulo 01000
11901 West Cedar Avenue
Lakewood, Colorado 80228 USA
11511 New Benton Hwy.
Little Rock, Arkansas 72201 USA
Mobil Solar Energy Corp.
16 Hickory Dr.
Waltham, Massachusetts 02254 USA
Pompes Gitnard Etablissements
179, Boulevard Saint Denis
Philips GmbH, Unternehmensbereich Licht
2000 Hamburg 1
FEDERAL REPUBLIC OF GERMANY
Solar Electric International
31 Queen Anne's Gate
London, SW1H 9BU
Solar Usage Now Inc.
420 East Tiffin St.
Bascom Ohio USA
1335 Piccard Dr.
Rockville, Maryland 20850 USA
3646 E. Atlanta Ave.
Phoenix, Arizona 85040 USA
12533 Chadron Avenue
Hawthorne, California 90250 USA
10 DeAngelo Dr.
Bedford, Massachusetts 10730 USA
2821 Mays Ave.
Amarillo, Texas 79109 USA
P.O. Box 6015
Santa Fe, New Mexico 87502 USA