Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Graduate School of the Environment
Centre for Alternative Technology
Machynlleth
Powys
Wales
SY20 9AZ tel: 01654 705981
School of Computing and Technology
University of East London
Docklands Campus
4‐6 University Way
London
E16 2RD tel: 020 8223 3000
VIABILITY STUDY OF A NEW LOW‐POWER
SOLAR THERMAL IRRIGATION PUMP FOR
SMALLHOLDER FARMERS IN LOW‐INCOME
COUNTRIES
by
N. T. Jeffries
A thesis submitted as part
of the requirements of a
Masters of Science
in Renewable Energy
at the
Centre for Alternative Technology
Wales
Jan 2010
ABSTRACT
Smallholder farmers in low income countries can benefit from affordable irrigation
pump systems as they enable cultivation of high value crops during dry season. This
generates extra income and contributes to an improvement in quality of life.
Currently the majority of small irrigation pumps are manually operated which is time
consuming and requires a high level of physical exertion. There is a potential market
for a low cost solar thermal pump that produces a high volume of water as well as
reducing the labour burden. This will allow more crops to be grown and free up time
for other productive tasks.
To assess the viability of the new pump, a monitoring evaluation study was designed
and implemented over a three month period in Ethiopia. The study assessed the
performance of nine solar pumps against a number of different indicators used to
measure viability. These were - performance, complexity, comparability and
affordability.
Despite the limitations of the study, the results proved that under the right conditions,
the pump could satisfy all of the criteria to different degrees. Performance measured
by daily volume and hydraulic energy output exceeded the target specification;
however it fell far short of the theoretical output. Most farmers were able to operate
the pump consistently and independently. Compared to the existing manual systems,
many hours could be saved each day through reduced labour input. By considering
daily water production and typical crop water requirements, it was shown that annual
revenue projections allowed a pay back time well below the conditions of local
microfinance institutions.
The study concluded by suggesting that even though all the viability criteria had
been satisfied, this alone may not be enough to engender widespread adoption of
the technology, especially in a risk adverse market with low labour costs. To improve
marketability, the perceived value and benefits of the system need to be increased
through different enhancement measures. These might include mechanical
modifications to raise the engineering performance closer to the theoretical potential;
as well as the design and promotion of compatible support infrastructure and training
material that will lead to more effective and efficient operation of the pumps.
N.T.JEFFRIES Masters Thesis REBE 1 | P a g e
ACKNOWLEDGEMENTS
In Ethiopia:
Gert-Jan Bom - for investing so much time and effort in the creation of a pump to
help people improve their lives, with no expectation of reward. For inviting me to
assist in its further development and for his many patient explanations of pump
theory.
Practica Foundation
Delta-T Systems, Cambridge - for the loan of an SPN 1 pyranometer and GP1 data
logger.
Bryce Gilroy-Scott - my supervisor at the Centre for Alternative Technology for
reviewing the draft thesis and providing ongoing supervision and support.
Csaba Zagoni – fellow REBE student for providing zenith angles
Ben Townsend and Nick Taylor of Squint Opera - for the visualisations (Fig 4, 5)
N.T.JEFFRIES Masters Thesis REBE 2 | P a g e
CONTENTS
ABSTRACT ............................................................................................................................. 1
ACKNOWLEDGEMENTS ....................................................................................................... 2
CONTENTS ............................................................................................................................ 3
GLOSSARY ............................................................................................................................ 5
LIST OF TABLES, FIGURES AND GRAPHS ......................................................................... 6
1. INTRODUCTION ................................................................................................................ 8
1.1 Background ................................................................................................................. 10
1.2 Study Objectives ......................................................................................................... 12
1.3 Additional aims (C5) .................................................................................................... 12
2. LITERATURE REVIEW ................................................................................................... 13
2.1 Technology .................................................................................................................. 13
2.2 Study approach ........................................................................................................... 15
2.3 Summary ..................................................................................................................... 17
3. CONTEXT ......................................................................................................................... 18
3.1 Factors that affect study .............................................................................................. 19
3.2 Physical context .......................................................................................................... 20
3.3 Socio-economic context .............................................................................................. 22
3.4 Farm sites.................................................................................................................... 23
4. STUDY APPROACH ......................................................................................................... 28
4.1 Methodology ................................................................................................................ 28
4.2 Study Limitations ......................................................................................................... 29
4.3 Method ........................................................................................................................ 29
4.4 Analysis ....................................................................................................................... 30
5. PUMP THEORY ................................................................................................................ 31
5.1 Steam pump assembly ................................................................................................ 31
5.2 Collector ...................................................................................................................... 32
5.3 Engine ......................................................................................................................... 32
5.4 Pump ........................................................................................................................... 33
5.5 Condenser and recirculation system ........................................................................... 34
5.6 Pump operation ........................................................................................................... 36
5.7 Limitations of steam engines ....................................................................................... 37
5.8 Efficiency for system components ............................................................................... 38
N.T.JEFFRIES Masters Thesis REBE 3 | P a g e
5.9 Theory of Prime Mover ................................................................................................ 39
6. DATASETS ....................................................................................................................... 44
6.1 Solar irradiance (D1) ................................................................................................... 44
6.2 Pump performance data .............................................................................................. 46
6.2.1 Continuous monitoring (D2) .................................................................................. 47
6.2.2 Independent Usage (D3) ...................................................................................... 49
6.3 Ground water level (D4) .............................................................................................. 50
6.4 Steam tests (D5) ......................................................................................................... 51
6.5 Measurements from other irrigation methods (D6) ...................................................... 53
6.6 Socio-economic data (D7) ........................................................................................... 54
6.7 Local cultivation practices and agronomic context (D8) .............................................. 55
7. RESULTS AND ANALYSIS .............................................................................................. 56
7.1 Performance (C1) ........................................................................................................ 56
7.1.1 Continuous monitoring results (D2) ...................................................................... 56
7.1.2 Performance analysis against required specifications .......................................... 57
7.1.3 Performance analysis against theoretical potential .............................................. 60
7.1.4 Potential errors ..................................................................................................... 64
7.2 Complexity/acceptance (C2) ....................................................................................... 65
7.3 Comparability (C3) ...................................................................................................... 68
7.4 Economic assessment (C4) ........................................................................................ 71
8. DISCUSSION .................................................................................................................... 75
8.1 Improvements to mechanical system .......................................................................... 75
8.2 Support infrastructure considerations .......................................................................... 79
8.3 Improved training and technical support ..................................................................... 80
9. CONCLUSIONS ................................................................................................................ 82
10. BIBLIOGRAPHY ............................................................................................................. 84
APPENDICES ....................................................................................................................... 86
Word count: 14, 950 (excluding appendices, tables, titles, photo descriptions etc.)
N.T.JEFFRIES Masters Thesis REBE 4 | P a g e
GLOSSARY
N.T.JEFFRIES Masters Thesis REBE 5 | P a g e
LIST OF TABLES, FIGURES, GRAPHS AND APPENDICES
Table 1 Pump irrigation options around Ziway Town
Table 2a Farm profiles – Northern Cluster (Edogojola)
Table 2b Farm profiles – Southern Cluster (Borchessa)
Table 3 Collector components
Table 4 Engine components
Table 5 Pump components
Table 6 Recirculation system components
Table 7 Stages during one cycle of engine
Table 8 Theoretical conversion efficiencies
Table 8a Theoretical engine performance
Table 8b Theoretical pump performance
Table 8c Theoretical steam consumption
Table 9 Summary of study approach
Table 10 Datasets for each study component
Table 11 Cycle functions and corresponding pump quantities
Table 12 Extract from typical calculation sheet:
Table 13a Extract from Independent Usage calculation sheet
Table 13b Independent usage calculation
Table 14 Pump 4 - Continuous monitoring (D2) summary
Table 15 Pump 1 - Continuous monitoring (D2) summary
Table 16 Summary of Independent Usage
Table 17 Estimate of time saved by solar pumping
Table 18 Typical duration and volumes for manual pumping
Table 19a Estimated revenue for hypothetical farm
Table 19b Estimated input costs for hypothetical farm
Table 20 Present and theoretical system efficiencies
N.T.JEFFRIES Masters Thesis REBE 7 | P a g e
1. INTRODUCTION
This thesis investigates the viability of a newly developed low-power solar steam
engine to provide pumping power for small scale irrigation. The work is based on a
3-month data collection and research period undertaken in Ethiopia in Oct- Dec
2010. During this time, the performance of a number of solar thermal pumps (STPs)
was closely monitored and relevant information collected on the physical, climatic
and socio-economic context.
Box 1 provides a real life example of the transformative power of a simple irrigation
pump for a poor farmer and his family.
According to estimates by Practica Foundation (Bom G. J., 2009) there are over 7
million diesel pumps sets in India, consuming about 3.3 billion litres of fuel.
Assuming 2.5 kg CO2/litre, this is equivalent to 8.6 million tonnes of CO2 emissions
or approximately 0.5% of India’s overall 2007 emissions (WorldBank, 2010). The
benefits of a reliable, low cost alternative can be summarised as follows:
• Reduce the capital and operating cost of irrigation
• Reduce the pressure on dwindling fossil fuel resources
• Reduce the emissions related to irrigation
As this thesis will demonstrate through theoretical projections, even at the current
scale, the solar thermal pump could produce enough water for the irrigation needs of
fields that are currently served by fossil fuel engines1.
Possible future –
low cost, solar
powered, zero
emissions
Irrigation schemes do not always bring positive impacts and benefits. Over extraction
can lead to ground water depletion and water logging or salinity issues. In other
areas, irrigation schemes have been known to increase the prevalence of malaria
(Kibret, 2010). These issues are recognised as important factors that need to be
considered in irrigation schemes, but are outside the scope of this study.
1
Note: this assumes highly efficient distribution systems.
N.T.JEFFRIES Masters Thesis REBE 9 | P a g e
1.1 Background
The solar powered steam engine has been available as a power source for irrigation
pumping for more than a century. The steam engine as a prime mover has been
around for much longer. In the final decades of the 19th C., the technology had
become so important that one historian acclaimed:
“It would be superfluous to attempt to enumerate the benefits which it has conferred upon the
human race, for such an enumeration would include an addition to every comfort and the
creation of almost every luxury that we now enjoy. The wonderful progress of the present
century is, in a very great degree, due to the invention and improvement of the steam-engine,
and to the ingenious application of its power to kinds of work that formerly taxed the physical
energies of the human race.” (Thurston, 1883 p. 1)
The discovery of abundant crude oil in the early 20th C. meant that solar steam
machines were no longer competitive or attractive, so interest in their further
development dwindled. In the present day, as fossil fuels become more expensive
and scarce, and concerns are raised about carbon-related climate change, there has
been a resurgence of interest once again being shown in these types of system.
Low-power steam engines, at a size that could be useful and affordable by farmers
in low income countries, have always presented a challenge because of their
inherent low efficiency (see section 4.7) and related high cost. To address these
issues, Practica Foundation has developed a new engine utilising modern materials,
such as plastics to replace metal components and rubber diaphragms in place of
conventional pistons. This has led to efficiency improvements at low power as well
as reduced manufacturing cost, meaning that the use of solar energy as a heat
source for steam engines is now a possible cost-effective opportunity.
Practica have designed and built a number of prototype solar collectors, steam
boilers, steam engines and pump types over the last decades. The development
chronology which spans over 25 years is detailed in Appendix 1 and summarised in
Boxout 2. Since 2009, the technology has now moved beyond proof of concept to
working prototypes2.
A previous version of the system had been tested in Ethiopia and the findings were
used to create the current model. Now, for the first time, multiple versions of the
same prototype have been produced allowing field trials to be carried out in real farm
locations with pumps operated by farmers. The steam engine and other specialised
parts were built in the Netherlands over the summer 2010 and then shipped to
Ethiopia for assembly. Well drilling and installations were managed locally, allowing
commencement of operation at the beginning of the dry season. The evaluation of
these pumps is the subject of this thesis.
2
For a general review of the technology refer to Wong (1998) and Delgado –Torres (2007).
N.T.JEFFRIES Masters Thesis REBE 10 | P a g e
Box 2 – Brief History of Development of Ethiopia Solar Thermal Pump
1874 – Beginnings – Agustine Mouchot exhibits a machine in Paris that can pump hundreds of
gallons per day.
1913 – Heyday – American Frank Shuman demonstrates a 10,oo0 sq.ft. collector in Meadi, Egypt
that pumped 6000 gallons/min from the Nile leading to much interest from the colonial powers.
1919 – The end of an era - discovery of cheap and abundant crude oil causes sharp decline in
interest in solar steam engines.
1983 – Fresh shoots – Dutchman Gert-Jan Bom return from Burkina Faso and considers how to
solarise his newly invented Volanta pump (deep well, manual, large momentum).
1980s – New developments - Number of companies from France, Switzerland, India, Finland etc
developing STPs from 90W – 10kW using different working fluids such as butane and Freon.
1993 – Entrance condensation (EC) - Bom publishes Energy Losses through Entrance
Condensation in Small Vapour Engines partly based on experiments with Japanese company using
rice husks to power steam engine. Work part funded by Dutch Ministry of Foreign Affairs (DGIS)
1990’s – Peristaltic engine - Bom and inventor-friend T. Visser experiment with a number of
engines to determine which configuration is least affected by EC. Decided that peristaltic engine may
provide best solution. A number were designed and built but although several showed initial promise
all suffered durability issues.
Late 1990’s – 2000’s (est) – Rolling diaphragm piston – focus reverted to solarisation of
volanta pump with a long (20cm) stroke length. A rubber rolling diaphragm was used for the piston.
However the rubber reacted badly with the hydrocarbon working fluid causing the diaphragm to fail
rapidly.
Early 2000’s (est) – Flatplate collector and top hat diaphragm - the next development used
a smaller stroke diaphragm and an increase in force using a flat plate collector and a pentane working
fluid. After promising endurance testing in Holland, the pump was tested in Burkina Faso where it
did not have the same degree of success. There were also many problems in obtaining pentane from
local markets.
2003 – Evacuated tube (ET) collectors plus steam – for the first time steam was considered as
working fluid with new cheap, efficient ETs from China providing the necessary power. A system was
designed with 72 tubes (7.2m2) with an expected steam output of 72g/min. Once again testing in
Burkina Faso did not produce adequate or reliable results. The principal problem being steam
bubbles accumulating in the engine reducing the transfer of heat.
2008 – Concentrated collectors – by now Practica and IDE collaborating to develop a pump that
would suit the conditions of the treadle suction pump. First prototype used a Fresnel collector
comprising a 3 x 1.2m array of mirror slats focusing on a linear boiler. Engine-pump arrangement was
a diaphragm piston engine with top-mounted valves coupled to a diaphragm suction pump. Field
testing now moved to Ethiopia, identified focusing problems and reflector losses that meant pumping
only possible between 11am – 1pm. Engine also suffered water ingress thus exacerbating EC.
2009 – 2010 – Parabolic dishes – collector changed to dish using reflective material bought as
part of a German solar cooking kit. Outlet valve moved to bottom of engine and pump replaced with
reciprocating piston allowing deeper pumping. This version of the pump is currently being tested in
Ethiopia.
N.T.JEFFRIES Masters Thesis REBE 11 | P a g e
1.2 Study Objectives
The key purpose of the monitoring study has been to collect data and observations
that will help determine if the newly developed STP is a viable irrigation technology
option in low income countries. ‘Viability’ in the context of this study has been
defined using the following criteria:
The above criteria have been informed by the central aspects of theory of the
Diffusion of Innovations described in detail in the Literature Review chapter.
Another factor to consider is that just as a building alone does not make a school, a
pump alone does not make an irrigation scheme. Each type of pump requires a
certain set of conditions and support structures to allow it to operate most effectively.
In consideration of these two factors, the additional aims of the study will include:
The literature review was divided into two separate but overlapping areas:
2.1 Technology
The first thing to state is that solar thermal pumps are not all common. The ‘pump’
entry for Wikipedia states that steam pumps are: “mainly of historical interest”
(Wikipedia, 2010). Fraenkel and Thakes comprehensive manual on water pumps
“Water Lifting Devices” describes the systems as immature and non-commercial,
although does state that: “there is always the chance of some future breakthrough”
(Fraenkel, 2006 p.259)
Fig 1 provides a snapshot of the current academic focus in the field of solar powered
water pumping. The data for the chart is based on a key word search using “solar +
pumping” entered into the academic publication aggregator Science Direct4.
It is clear that the majority of academic effort (and by extension R & D activities) in
the field of solar pumping is focused on PV systems. However a closer examination
of reveals that these systems are generally only used for drinking water projects and
rarely if ever used for irrigation pumping. This is due to high system costs associated
with PV that can only justify their use in clean water applications where a higher
value is ascribed to the water. This supposition is supported by Ian Tansley, a
renewable energy expert at True Energy, who regularly carries out feasibility studies
for PV pumping installations on behalf of the UN and NGOs. In his experience, none
of these are for irrigation schemes5.
Some exceptions to this have been uncovered, for example Burney (2010) relates a
two-year study carried out on the adoption and impact of PV powered drip irrigation
systems for two farming communities in Benin. This study confirmed the high cost of
PV systems the installation estimate was $475 per 120m2, placing the technology in
a significantly higher cost bracket than STPs (Burney, 2010 p.1853).
4
http://www.sciencedirect.com/
5
Conversation at True Energy, Tywyn 30th August 2010
N.T.JEFFRIES Masters Thesis REBE 13 | P a g e
Fig 1 – Categories of solar pumping academic papers
In two reviews on solar thermodynamic pumps (Wong 1998, Torres Delgado 2007),
the researchers outlined the historical development and current state of the
technology. Wong’s (1998) study offered the most comprehensive version and
started by confirming the high cost of PV panels ($4-5/Wp) as the main challenge to
viability. The review divided the technology into conventional Rankine-cycle pumps
and more unconventional systems specially designed for developing countries that
avoid any high technology. The conventional systems investigated can be dismissed
on the basis that they are not comparable in terms of the scale of irrigation scheme
served. Furthermore the use of low boiling point fluids such as refrigerants and
organic fluids are not considered as sustainable as they are either difficult to procure
in the field or banned under the International Protocols (e.g. Freon)
For unconventional systems, the review described several very simple systems such
as the Savery pump, heat driven pumps and a fluidyne pump. However these are all
suction pumps and not applicable to the deeper GW conditions found in the field.
The only non-hypothetical data on solar thermal pumps that might be comparable
was Sumathy’s deep well pump that used a flat plate collector and pentane. This
recorded an efficiency equivalent to 0.12%
The two conclusions that can be drawn out of Wong’s very comprehensive
technology review might be construed as a validation for this thesis:
N.T.JEFFRIES Masters Thesis REBE 14 | P a g e
• Water is an economic choice of working fluid, but an optimisation study needs
to be carried out to achieve best results with minimum cost
Zeller (2003) carried out a study that purported to investigate a similar technology
under similar conditions. The system was also a small thermodynamic pump,
although it was designed as suction rather than a deep well pump. The pump used a
low boiling point refrigerant for the working fluid. The study objectives were broadly
similar in that Zeller was investigating whether his systems could improve the income
generation opportunity for small-scale farmers.
A closer reading revealed that the study was purely theoretical as the researchers
were not able to get the pump operational. Much of the performance and economic
analysis and therefore results were based on an assumed theoretical system
efficiency (2.5%), much larger than anything achieved in practice. The study was
dismissed as limited in usefulness, particularly as a key finding of this thesis was that
theoretical and actual performance can vary significantly.
A study carried out by Burney (2009) focused on a different solar technology but in a
broadly similar context, the introductory words echo this:
The study, funded by the World Bank and ICISRAT6, comprised an assessment of
the impact on food security of PV drip-irrigation systems in 2 treatment villages in the
Kalale district. The scope of the work was much larger but the analysis, included the
application of statistical models for a similar socio-economic setting could offer a
relevant reference source for element of this thesis, particularly in assessing relative
advantage of different technologies (C3- Comparability).
Source: Wikipedia
Diffusion theory has been subsequently built upon and reinforced by the results from
many other studies. The theory’s origins are in the sphere of rural sociology and
many study subjects’ are the kind of technology that this thesis is investigating. In
this lies one of the main criticisms of Rogers’ theory that is it may not be applicable
only to particular technology areas.
The five characteristics7 that Rogers identified that helps explain the relative success
of innovations are:
1. Relative advantage – the degree to which an innovation is better than the idea
it supersedes.
2. Compatibility – the appropriateness of the innovation to the particular
conditions
3. Complexity – the degree to which the innovation can be understood and used
4. Trialability – the degree to which the innovation can be experimented with on
a limited basis.
5. Observability – the exposure the innovation has to the potential adopters.
7
After several years he added to this initial list the attribute of reinvention - i.e. the degree to which an
invention is modified by the user post-adoption.
N.T.JEFFRIES Masters Thesis REBE 16 | P a g e
For obvious reasons, these attributes correlate with the criteria that have been
selected to assess the STP technology.
2.3 Summary
The literature review revealed the following key points:
• There are very few publications relating to small solar thermal pumps
• The publications that do exist represent theoretical analyses, systems that are
overly expensive or not appropriate to the study context
• Developing a system with water as a working fluid and raising the efficiency
compared to current working systems is seen as a key step to making STPs
viable
• For technology to be successful it needs to be better than existing
alternatives, appropriate to the setting and easily understood.
N.T.JEFFRIES Masters Thesis REBE 17 | P a g e
3. CONTEXT
IDE has been collaborating with Practica for a number of years in the development
and dissemination of appropriate irrigation technology to poor smallholder farmers.
The main technology areas have been low cost irrigation pumps and manual well
drilling techniques.
Since 2007, IDE has been in Ethiopia promoting foot operated treadle pumps for
shallow well pumping and hand operated rope and washer pumps for deeper wells.
From observations on ground water depths and cultivation practices, it became
evident that that there may be a niche for a low cost, small size deep well pump
powered by the sun. As Practica had been attempting to develop such a system for a
number of years, IDE asked them to design a prototype to suit the conditions in
Ethiopia.
The target specification suggested by IDE was a daily production volume of at least
2000litres from a depth of 15m. This can be equated to a daily hydraulic energy (E
3
hyd ) requirement of 300, 000 J/day, or in another form: E hyd = 30 m .m. This is a
useful quantity as it allows easy comparison between different hydraulic heads
(ground water depths). E hyd will be referred to throughout the text when comparing
pump performances.
N.T.JEFFRIES Masters Thesis REBE 18 | P a g e
In late summer 2010, ten improved solar steam pump systems were shipped to
Ethiopia for assembly and installation. The testing locations were identified from
existing IDE Ethiopia projects near Ziway. Model farmers were selected and offered
a free well in return for participation in the testing. A requirement for selection was at
least one or more years experience using a R & W pump for irrigating vegetables
thus allowing a comparison of the solar pump to manual alternatives.
The consequence of this was that for a good proportion of the first few weeks of
operation, there was a significant requirement for regular on-site attendance by
technicians and engineers. This included engine tuning, component replacements
and various other mechanical adjustments. For this reason the early ‘reliability’ of
the pumps were not factored into the assessment of viability as there was always an
expectation that certain components in the deployed system would need to be
replaced.
Another potential affecting factor is that on the test sites there are three activities
taking place, prioritised in this order:
1. Ongoing farming and domestic activities to support the needs of the families
2. Field testing of new technology
3. The monitoring and evaluation study
Each of these activities has its own objectives which can differ and overlap. If they
differ the above order of priority will determine where available energy and resources
are directed i.e. farmers will not use the solar pump if other pumps (or agricultural
activities) better serve their needs. These factors have been considered in the
design and analysis of the M & E Study.
N.T.JEFFRIES Masters Thesis REBE 19 | P a g e
3.2 Physical context
The project area is located in the Oromia Region around the market town of Ziway
(alt 1600m) approximately 150km south of Addis Ababa. The town lies to the west of
Lake Ziway, the most northerly of a chain of several large Rift Valley lakes that slices
through the centre of Ethiopia.
Ziway
The annual rainfall is approximately 1000mm, most of which falls in the rainy season
lasting from Apr – Sept (FAO, 2010). The rest of the year is predominately dry,
October and November being particularly hot and clear. During the field testing
period no rain was recorded. The temperature range recorded during this same
period was 11 – 35oC.
Cloudiness is an important variable for analysing the operation of the pump. As part
of the monitoring, solar irradiance was recorded for 48 consecutive days. Within this
period there were 57% cloudless days (27 full days) and over 80% of potential
operating hours were clear.
The solar data is analysed in a later section but has been summarised in Graph 2 to
provide a snapshot of the solar activity during the study period.
N.T.JEFFRIES Masters Thesis REBE 20 | P a g e
Graph 2 – Cloud hours during study period based on output of SPN 1 Pyranometer
8
cloudiness hours 7
6
5
4
3
2
1
0
21‐Oct
23‐Oct
25‐Oct
27‐Oct
29‐Oct
31‐Oct
2‐Nov
4‐Nov
6‐Nov
8‐Nov
10‐Nov
12‐Nov
14‐Nov
16‐Nov
18‐Nov
20‐Nov
22‐Nov
24‐Nov
26‐Nov
28‐Nov
30‐Nov
2‐Dec
4‐Dec
6‐Dec
The soil profile encountered is a sandy loam, made up of mainly hard consolidated
sand with bands of silty clay. To the south of the lake, the sand at depth was less
compacted. This type of soil strata led to problems with excessive fines entering the
well casing pump contributing to pump malfunctions and accelerated wear.
Interestingly, in these same areas the silt content of the surface soil meant that
percolation of water is slowed down allowing higher distributions efficiencies for
furrow irrigation.
N.T.JEFFRIES Masters Thesis REBE 21 | P a g e
The groundwater depth is determined by lake water level and therefore becomes
deeper as the land rises away from the lake. There is a hysteresis effect caused by
the slow percolation of the lake water through the soil strata, meaning that the
seasonal variations in ground water levels are slightly out of sync with rising and
falling of lake levels. The groundwater depth range for the test sites was between 2
– 13m.
During rainy season families cultivate maize, tef8 and other grains on plots of land
near to their homes. If they have a large enough plot of land, any surplus production
is sold to dealers in the local market. Mostly the production is subsistence level; the
harvest is stored near to the house to provide staple food during the dry season. In
the dry season, irrigation by ground or lake water is used to cultivate higher value
cash crops for sale to local dealers or at the local market. Crops types are
determined by demand and growing conditions. They are essentially limited to onion
seedlings, onions, peppers, chillies and kale.
Onion seedlings are grown either for transplanting to onion beds, or to sell on to
commercial onion growers. As seeds are expensive, farmers tend to wait to agree a
contract before starting cultivation. Seedlings are grown in 5 x 1m strips called
‘merips’ and the work is hard as they require daily watering with a sprinkler can.
However a contact to grow is perceived as a ’bonanza’ as harvesting time is only 45
days and the selling price allows high profits to be earned quickly. Good contacts
and proximity to a main road improves likelihood of a commission.
8
Eragrostis tef, a species of lovegrass used to make the food staple ‐ a flatbread called food called injera
N.T.JEFFRIES Masters Thesis REBE 22 | P a g e
Cultivation of onions
seedlings allows a high
profit to be made in a short
time. However not all
farmers are luck enough to
secure growing contracts.
There are a number of pumping methods in the project area; choice is determined by
size of field, socio-economic status and ground water depth:
The selection criteria for the test farms were (i) an established relationship with IDE;
(ii) previous experience of irrigation; (iii) GW depth between 5 – 15m. Profiles for
each farm are presented in Table 2. They include physical characteristics of each
site, socio-economic information and brief bios for each farmer. Farm layout plans at
different stages in time, showing the changing configuration of the land are included
in Appendix 2.
N.T.JEFFRIES Masters Thesis REBE 23 | P a g e
Average smallholding area is 2175m2 and cultivated area about 550 m2 (25%). The
proportion of cultivated to smallholder area ranges between 10 – 60%. At the high
end is a large family with a petrol engine pump near to a main road, at the low end is
an almost exclusively female household with a small family and infant children.
The key factor that differentiates the two clusters is the proximity to the main road.
Farms that are close to the road are closer to markets and have greater exposure to
potential economic activity. All farms in the northern cluster have had ‘onion seedling
bonanzas’ which have greatly improved their financial situation.
Southern cluster farms do not have these opportunities and are generally poorer. In
some ways this makes them better monitoring sites (and future customers) as their
activities and crop choices are more suited to solar pumps.
N.T.JEFFRIES Masters Thesis REBE 24 | P a g e
Location plan
The plan below shows the locations of the pump installations relative to the main
road and the lake. The topography to the west of the lake rises more rapidly than to
the south. This explains the much lower ground water level in Pump 1 compared to
Pump 7, even though they are about the same distance from the lake edge. The
plan also shows the market where vegetables are traded.
The pyranometer is located at the IDE workshop. Note the cloud distribution on the
day that the image was taken (Aug 13th 2001). Under these conditions, if the sun
was in the east of the sky, then because of the cloud distribution the pumps may not
have been exposed to exactly the same solar conditions (see 6.1.4 Potential errors).
5 km
N.T.JEFFRIES Masters Thesis REBE 25 | P a g e
Table 2a ‐ FARM PROFILES ‐ NORTHERN CLUSTER (EDOGOJOLA)
Table 2b ‐ FARM PROFILES ‐ SOUTHERN CLUSTER ‐ BORCHESSA
Farmer represents the ideal IDE customer – hard
working, ambitious and open to new ideas. Bought
Chala R & W pump last year for 1700 birr with loan of
9 22/10 900 600 6 9 3 N 2500birr. Paid back over 25% through sales of
Huluga
peppers and onions. In first year of irrigation wore
down metal support strut with nylon rope due to
frequency of usage.
N.T.JEFFRIES Masters Thesis REBE 27 | P a g e
4. STUDY APPROACH
This section describes the approach taken to investigating the performance and user
response to the solar thermal pump. To begin a methodology will be outlined
describing an idealised approach to investigating the study objectives. Next the
constraints and limitations will be introduced, including those aspects of the physical
and socio-economic context that have a direct impact on the study. Finally based on
these limitations, the actual method and analytical approach that was employed will
be described. The next chapter details the data sets collected to support this stated
method.
4.1 Methodology
In ideal circumstances, it would be desirable to implement a study allowing cross-
sectional sampling of farmers, some with STPs (treatment groups) and other using
different of existing technologies (control groups). The physical and socio-economic
context of each of the farmers and pump installations would be carefully selected so
that key variables that may have an effect on the outcome of the experiment can be
controlled.
Farmers (and farms) would be monitored over a compete (or ideally several) growing
season(s) and data collected on solar inputs, pump outputs, changes in cultivation
practices, irrigated areas, crop yields, crop prices, farmers’ revenue and other
relevant areas. Combining all of this would allow a quantitative and statistical
analysis to be carried out to determine performance, comparability to other
technologies and economic benefit. The ideal scope might include:
A good model for such a study (albeit a different technology) that includes an in
depth evaluation of a technology and user response in a similar physical and socio-
economic context can be found in Burney (2009).
N.T.JEFFRIES Masters Thesis REBE 28 | P a g e
4.2 Study Limitations
Based on the available resources and the context of our study, the full
implementation of the idealised approach was not possible. Some of the limiting
factors have already been introduced in section 3.1. The issues that had the most
impact were:
Time and resources – time, equipment and human resources available for the study
were limited. The study area was over 100km2 and most of the work was carried out
by only one student, a translator and an enumerator.
Timing – the period available for data collection was Oct – mid Dec. The last pump
was not installed until 22nd Oct further reducing the available time. The installations
all require a ‘bedding in’ period, during which it was not always possible to collect
meaningful data. There was a delay in the farmer’s response to the new availability
of water, so that reconfiguring of land to accommodate the new water happened a
while after installation during which time there was no use for pumped water.
Other priorities – the operation and usage of the STPs require input from the
farmers. However farmers will always gravitate towards what is most lucrative and
important. Therefore if farmers have other income generating activity away from their
farm or in another part of their farm they will not use the STPs. It is felt that this effect
would be reduced if the monitoring period were longer as farmers gain more
confidence and knowledge about the operation of the STP.
4.3 Method
The evaluation of the solar thermal pumps was informed by collection and analysis
of relevant data and observations as well as reference to other similar studies. The
data collected is divided into two types:
• Measurable/quantitative data
o Pumped volume/hours of operation
o Solar irradiance
o Ground water level
o Small holding/garden size;
o Crop yield/prices/ profit
N.T.JEFFRIES Masters Thesis REBE 29 | P a g e
• Observed /interview/qualitative data
o Acceptance issues;
o Impact on irrigation methods/behaviour;
o Socio-economic factors;
o Perceived value by farmer.
The different data sets are described in detail in the next chapter.
4.4 Analysis
To analyse the results, no single individual model has been adopted, instead
different approaches have been used to suit the study component, data type, desired
result form and the success indicator. These results are then considered collectively
to produce an overall assessment of the technology.
Table 9 shows the general approach that will has been followed for analysing the
different sub-components of the study.
N.T.JEFFRIES Masters Thesis REBE 30 | P a g e
5. PUM
MP THEORY
The purp
pose of this
s section is to provide sufficient th
heoretical and
a techniccal details to
o
support the experim
mental apprroach and a analysis, thhis will include descripttions of:
• Mechanical componentts
M
• S
System operation - by dividing
d onee engine cy
ycle into disstinct stage
es.
• Limitations of
o the systeem imposed d by thermoodynamic laaws and the e structural
and mechan nical properrties of the equipment.
e .
• B
Basic analyssis of the prrime moverr to allow a comparison between theoretical
and experimmental results.
Fig 4 – Main
M compon
nents of pum
mp system.
N.T.JEFFR
RIES Masters TThesis REBE 31 | P a g e
5.2 Collector
Figure 4 shows the main parts of the collector; table 3 provides details on each of the
parts:
5.3 Engine
The parts of the engine are illustrated in the Fig 4 which is a cut through section at
approximately mid-cycle. The feed, outlet and connector hoses are not shown. Note
the engine position can be moved up and down the concrete support allowing the
stroke length to be varied for different GW depths.
N.T.JEFFRIES Masters Thesis REBE 32 | P a g e
Table 4 – Engine components
Component Material and other details Dimensions Local/
imported
9
Buffer Polypropylene (PP) – maximum V = 230cm3 Imp
chamber allowable temperature 130oC
5.4 Pump
The pump has a reciprocating positive displacement action that lifts water through
the relative movements of a piston and a foot valve. The down stroke closes the foot
valve and the piston moves through the water displacing a column of water equal to
the volume of the stroke; the upstroke displaces the water above the piston while
opens the foot valve drawing water into the pump cylinder.
The original piston design was a hollow steel cylinder with a diameter fractionally
below that of the pump casing cylinder. The idea was that it would be frictionless and
resistant to wearing by abrasive soils. However it soon had to be replaced, because
the sandy soils would not allow the piston to move freely. There are now two types of
pistons installed according to the amount of sand encountered in the initial well
development.
Type 1 - PVC Piston - Low sand conditions Type 2 - Rubber cup piston - High sand conditions
9
Higher temperatures will mean the use of engineering plastics leading to a large increase in cost.
N.T.JEFFRIES Masters Thesis REBE 33 | P a g e
Table 5 – Pump components
Component Material and other details Dimensions Local/
imported
Piston type 1 Perforated PVC piston cylinder, brass top Φ = 43mm Imp
valve
Piston type 2 Silicon diaphragm cup Φ = 43mm Imp
Piston rod Threaded steel rods. Φ = 8mm Local
3 m sections
Foot valve Standard steel ‐
N.T.JEFFRIES Masters Thesis REBE 34 | P a g e
Table 6 – Recirculation system components
Component Material and other details Dimensions Local/
imported
Outlet pipe PVC pipe Φ = 50mm (OD) Local
The performance of the feeder pump is critical to the effective operation of system.
During the monitoring there were frequent problems with the seals and valves, which
meant that the boiler sometimes had to be refilled several times in a day. This led to
gaps and discontinuities in the data collection (see 4.2 Limitations). The feeder pump
is currently being redesigned to create a more durable and reliable system.
Feeder pump being
used to manually fill
the boiler
N.T.JEFFRIES Masters Thesis REBE 35 | P a g e
5.6 Pump operation
The pump can be seen operating in the field by accessing the Practica Foundation
video channel on the internet10. The operational stages for one cycle are
summarised as follows:
2 Boiler located at focal point of collector absorbs solar energy which heats up water
and turns it into steam.
3 Steam conveyed to buffer vessel of steam engine via insulated hose. Pressure in
buffer allowed to build up until 1 - 1.5 bar (120 – 130oC)
4 When the desired pressure is reached, operator manually turns the flywheel
counter clockwise, opening the inlet valve allowing steam to enter the working
chamber.
5 The steam entering the cylinder pushes the diaphragm piston forward which moves
the rocker arm via a direct coupling with the connector rod.
A ‘scotch yoke’ mechanism converts the sliding motion of the connector rod into
rotational motion of the flywheel.
The rotational motion of two eccentric cams provides the timing of the opening and
closing of inlet and exhaust valves.
6 At bottom dead centre the outlet valve opens allowing the exhaust steam to be
discharged via the exhaust valve.
7 Lower temp steam leaves the engine and enters the condenser which is
submerged in the cool ground water thereby condensing the steam to water.
8 The rocker arm is connected to the top of the piston rod and provides the required
reciprocating motion and power to allow pumping of water from the well.
9 The condensate is sucked out of the condenser by the feeder pump which is also
powered by the rocker arm. The condensate is drawn into the feeder pump and
then pumped back to the boiler.
10 Flywheel momentum takes the diaphragm piston back to top dead centre where the
inlet valve opens again and the cycle continues.
10
http://www.youtube.com/user/Practicafoundation
N.T.JEFFRIES Masters Thesis REBE 36 | P a g e
5.7 Limiitations of steam eng gines
The fund
damental constraint on n the perforrmance of all
a vapour engines
e is the
t
thermoddynamic Ca arnot limit. This
T is calcu
ulated by co
onsidering the differennce in
temperaature (ΔT measured
m in K) between the incom ming (T2) and exhaustt steam (T1).
The equuation that relates
r effic mperature difference is
ciency to tem i as follow ws:
N.T.JEFFR
RIES Masters TThesis REBE 37 | P a g e
The phenomenon was the subject of a 1993 paper – Energy Losses through
Entrance Condensation in Small Vapour Engines (Bom 1993). As smaller engines
have a higher surface area to volume ratio, this means that entrance condensation is
proportionally more of a problem as engine size decreases. According to Bom, it is
for this reason that in small vapour engines, even accounting for collector and other
losses, where one might expect an overall efficiency of 3 – 4%, in fact:
“In practice, overall solar to water efficiencies as reported rarely exceed 1%” (Bom
G. ,1993, p.223)
The expected efficiencies of the individual stages are presented in Table 8. This
represents the performance of the current prototype as designed by Practica. It is
based on theoretical calculations as well as experimentation under controlled (but
not unrealistic) conditions in the Netherlands. The efficiencies listed represent
benchmark values for the performance of the system.
N.T.JEFFRIES Masters Thesis REBE 38 | P a g e
Table 8 - Theoretical conversion efficiencies
Stage Efficiency Details
Total 1.6%
If one applies these figures to typical design conditions, it is possible to calculate the
theoretical maximum daily pumped volume:
This exceeds the target specification put forward by IDE (Ehyd = 30m3.m) by a factor
of almost three.
As part of the analysis, volume measurements from the monitoring will be compared
to these theoretical values. Further investigation will also be carried out on individual
components of the system to determine the parts that are performing as expected
and those that are underperforming. This will create a focus for improvements and
the possibility of a more efficient system for the next design iteration.
The steam engine is a single cylinder engine. Each cycle of operation can be divided
into a pressure and expansion stroke which is delineated by the opening and closing
of inlet and exhaust valves. The engine cycle is very simply illustrated in Fig 8:
11
Based on 6hrs pumping
N.T.JEFFRIES Masters Thesis REBE 39 | P a g e
Fig 8 – Schematic of steam engine cycle
Inlet valve
TDC
opens
EXPANSION PRESSURE
o ff
STROKE
g cu t STROKE
n
asi
cre
De
Cut off = 50% Inlet valve
closes
I nc
rea
si n
g cu
t off
Exhaust
BDC valve opens
According to the design parameters, during ideal operation, when the inlet valve
opens at top dead centre (TDC), it admits steam at a pressure of between 1.3 – 1.5
bar (abs P = 2.3 – 2.5 bar) corresponding to a temperature of approximately 125 -
128oC (see Appendix C). When the exhaust valve opens (BDC), the working space
is exposed to the outside so pressure drops to atmospheric.
The change in pressure in the course of one engine cycle is represented on Graph 3
which plots absolute pressure against stroke distance.
This is a highly idealised representation but serves to illustrate the following features
of the engine operation.
12
To work out actual work done multiply by diaphragm area.
N.T.JEFFRIES Masters Thesis REBE 40 | P a g e
Graph 4 – Theoretica
al pressure variation
v in ssteam engin
ne for 50% cu
ut off
Note:
1. Stroke length = 4cm
m, therefore an
nything to thee right of this value
v is not in
ncluded in the
e analysis.
2. Atmosppheric pressure varies with
h altitude. In Z
Ziway (al = 16600m), atmP = 0.836bar
The ‘barrk’ is the en
nergy lost th
hrough not extracting the
t full expa ansive energy of the
vapour through
t n-optimal cut off. The issue of varying cut offf is discusssed further
non
in Appenndix D.
Expeerimental rig u
used to
ana
alyse the enginee
perfformance.
N.T.JEFFR
RIES Masters TThesis REBE 41 | P a g e
An extract from one of the tests is in included in Appendix E – Single Pressure cycle
in Steam Engine. The graph shows the correlation and differences between the ideal
model and real operation. For example in the ideal model, working pressure is
assumed from the very beginning of each stroke, but in reality it takes time to reach
this pressure, The complete objectives of this experiment were not ultimately
reached due to the equipment limitations; however the preliminary results have been
used to support some of the study findings.
We can make an estimate of the theoretical power output of the engine by assuming
a working pressure and applying a standard mechanical equation for work done (WD
= F x V) to the principal dimensions of the engine:
N.T.JEFFRIES Masters Thesis REBE 42 | P a g e
Table 8b - Theoretical pump performance
PUMP
Engine input W 53.3
Pump efficiency % 70%
Hydraulic power input W 37.3
Depth of water m 7.5
Flow per second l/s 0.50
Output per hour litres 1804.0
Hours per day hrs 6.0
Output per day m3 10.8
Note:
1. Assumed water depth = 7.5m
2. Pump efficiency estimated and includes volumetric efficiency of piston, friction losses through
mechanical movements and kinetic losses due to acceleration and de-acceleration of water
column.
Finally, by considering the engine dimensions and the thermodynamic properties of
steam it is possible make an estimate of expected steam consumption:
Dividing the latent power in the steam with the power output from the steam engine
allows us to calculate a value for the combined Carnot and Engine efficiency:
There is a small discrepancy between this value and those presented in Table 7 due
to the different sources and assumptions, but they are close enough to provide a
good order of magnitude estimate for potential engine performance.
N.T.JEFFRIES Masters Thesis REBE 43 | P a g e
6. DATASETS
This section details the data that was collected during the M & E period. For ease of
referencing in the analysis and later sections, codes have been assigned to each
dataset.
Table 10 – Datasets for each study component
STUDY COMPONENT DATASET
1 Performance D1. Solar irradiance and weather data
D2. 5 weeks continual monitoring at different ground water depths
D3. Independent Usage of all pumps
D4. Ground water level drawdown
D5. Steam production tests
Pressure sensor tests in engine cylinder (limited)
2 Complexity/acceptance D3. Independent Usage of all pumps
D7. Socio‐economic data
Observations/ interviews ‐ rotating visits to all installations
D8. Local cultivation practices and agronomic context (see below)
3 Comparability D6. Measurements from other irrigation methods
e.g. accelerometer data from R& W pumps
D8. Local cultivation practices and agronomic context (see below)
4 Economics D7. Socio‐economic data
D8. Local cultivation practices and agronomic context e.g: Plot size
measurements, typical crop choice and proportions, Crop water
requirements, density and time to harvest, market visits and trader
interviews to determine crop prices
SPN 1 Pyranometer (loaned by
Delta T systems, Cambridge). In
photo device is shown near to
Pump 5, however for security
reasons during monitoring was
mounted on IDE workshop roof
Due to the high cost of the equipment, the solar pyranometer had to be located on
the roof of the IDE workshop compound in the centre of Ziway town (see Fig 2). The
N.T.JEFFRIES Masters Thesis REBE 44 | P a g e
roof allowed unobstructed exposure to the passage of the sun from dawn to dusk
and the maximum distance from the most distant installation was 6km. The adjacent
tin roofs were covered up to reduce any potential albedo effects.
As the solar collector is a tracked system, to calculate the incoming solar energy it
has been necessary to convert from horizontal to direct normal irradiance (DNI). This
is the component of irradiance normal to the receiving plane of the collector. The
equation that relates horizontal to normal irradiance is:
The solar zenith angle (Z) is the angle that the sun makes with the vertical and is a
function of many parameters principally time, day number and latitude. To calculate
Z at such small time resolution, a SMARTS13 model was created with the assistance
of a student at the Centre for Alternative Technology, Wales (Zagoni, 2010).
DNI datasets for each day were generated by applying the above cosine relationship
to the horizontal data and then integrating over the time period of the pump tests to
determine the total solar energy available for pumping.
900
800
700
600
500
400
Key findings:
Based on solar data collected from Oct 28 – Dec 5, considering the 6 hr period, 9 –
3pm, the following average solar values were estimated:
13
Simple Model of the Atmospheric Radiative Transfer of Sunshine, NREL Laboratories
N.T.JEFFRIES Masters Thesis REBE 45 | P a g e
• DNI = 853 W/sq.m
• Solar Irradiation (energy), H = 26 000kJ/m2 or 7.2 kWh/ m2.
For practical reasons, it was not possible to monitor DPV production at all the
installations continuously. This was because of limitations on resources, meaning
that a suitable individual could not be present at each pump all the time to record
measurements. The two types of volume data collected were:
The method for measuring volume was using a cycle computer (Giant 8), comprising
a magnet, sensor and digital counter. The magnet is attached to the inside of the
flywheel, so that on each rotation it passes within 5mm of the sensor fixed to the
concrete engine support. The computer is calibrated to a wheel circumference of 2m,
so for every 5 rotations the counter records a distance of 10m. As each revolution
corresponds to one stroke of the pump piston rod, it follows that that distance
recorded on the computer is direct proportional to volume pumped.
Cycle computers recorded
speed, duration and
distance data for the
engine flywheel, which
was converted into
volumetric data for the
pumps.
N.T.JEFFRIES Masters Thesis REBE 46 | P a g e
Table 11 - Cycle functions and corresponding pump quantities
Code Cycle function Pump variable Utility
Speed Speed Engine speed Instantaneous measurement of RPM, which
indicates how well the engine is running.
ODO Odometer Cumulative volume Indication of overall volume over a long
pumped period.
DST Distance elapsed Volume pumped since Independent usage since last visit.
last resent
TM Time elapsed Pump duration since How long the pump has been running since
last reset last visit.
AV Average speed Average rpm since last Average engine speed since last visit
reset
MAX Maximum speed Max rpm since last Limited use ‐ but high rpm indicates
reset possible runaway incidents
TIME Time of day Time Synchronised with other data loggers for
time stamped analysis
The parameters directly corresponding to volumetric data (DST, ODO) were the
most useful and frequently collected quantities. However other readings were useful,
for example ‘speed’ readings gave an instantaneous indication of engine rpm.
14
If available, as there was only a few pressure meters it was not always attached to buffer.
15
The bucket test, measures the duration and distance (DST) corresponding to the filling of a 20ltr bucket.
N.T.JEFFRIES Masters Thesis REBE 47 | P a g e
Table 12 - Extract from typical calculation sheet:
Pump 4 Nov‐4
Dia 43 mm ODO time
Setting 4 Start 538.3 9:04
Stroke 9.4 cm End 607.8 15:11
Stroke vol 0.136 ltr Cycle 2 m
GWL 7.45 m
A. BUCKET Volume 20 ltr
Stroke
DISTANCE CYCLES Vcalc Efficiency vol. TIME flow
TEST TIME start end total l/str (s) l/s
11:30 27.02 27.33 0.31 155 21.1 95% 0.129 84 0.24
12:00 33.51 33.84 0.33 165 22.5 89% 0.121 74 0.27
12:30 39.84 40.2 0.36 180 24.6 81% 0.111 93 0.22
B. LOG
TM DST vol/str Vol Vol cum SPEED P
mins km litres litres km/h rpm bar
11:30 142 26.9 0.129 210.3 1667.3 13 108
11:45 157 30.12 0.129 207.7 1875.1 13.1 108
12:00 173 33.34 0.121 195.2 2070.2 12.9 108
The mid-section relates to the Bucket Test, which is carried out every 30 min. The
total time and DST to fill a 20ltr bucket are recorded, allowing calculation of:
• Volumetric efficiency
• Stroke volume - to convert DSTs into volumetric data.
• Instantaneous flows (l/s)
N.T.JEFFRIES Masters Thesis REBE 48 | P a g e
6.2.2 Independent Usage (D3)
For the pumps that were not being continuously monitored, pump usage and
volumetric data were determined by spot readings of DST, ODO and TM collected
during rotating visits to test sites (at least one visit/week). After taking
measurements, the computer was reset and the time/date noted allowing data to be
linked to specific time intervals.
In the course of the study, every pump underwent a longer duration monitoring
session (2 hrs +). Pump characteristics (stroke length, ground water depth) were
measured and several bucket tests carried out. This allows calibration and validation
of future results. The process is illustrated through the following calculation sheet:
Pump 9 17‐Nov
Dia 43 mm Stroke 11.6 cm
Setting 3
BUCKET
TEST Volume 20 ltr
Stroke
DISTANCE CYCLES Vcalc vol. TIME flow
TEST TIME start end total l/str (s) l/s
11:40 3.01 3.22 0.21 105 17.7 0.190 74 0.27
12:50 14 14.22 0.22 110 18.5 0.182 75 0.27
0.215 0.186 0.268
During each visit to a test site, an ODO reading is taken and then the counter was
reset. On the next visit, a new reading is taken, giving an indication of volume
pumped in the intervening period and therefore a metric for independent usage. The
calculation is as follows:
PUMP 9
Interval 21‐Nov 23‐Nov
ODO (km) 21.65 141
Dist 119.35 km
Days 5.5 days
Vol. factor 0.186 l/str
Volume 11108 litres
Daily volume 2020 litres
N.T.JEFFRIES Masters Thesis REBE 49 | P a g e
6.3 Ground water level (D4)
Accurate GWLs were required at each pump location to calculate the hydraulic
energy (Ehyd) associated with the pumped volume. Pump drawdown was also
measured, as a drop in level corresponds to an increase in discharge head,
therefore increasing the hydraulic power requirement.
• Static GWL – an electronic dip level was lowered into a nearby open well. The
sensor at the end of a measuring tape beeps when it contacts water.
• Pump drawdown– a HOBO water level logger (photo below) was suspended
between the pump cylinder and above the bottom of the borehole (see fig 9).
The sensor measures changes in water pressure during pumping which can
be converted into variations in water depth.
Static GWLs were measured at the beginning and the end of the study to check for
changes during the field testing period16. The results are shown in Table 2 and have
been factored into the pump performance calculations.
Drawdown tests require the pump (piston and pump casing) to be physically
removed from the well and therefore were not carried out many times. The main test
was carried out between 3 – 9th Nov at Pump 2. In this period the pump was
operated on several occasions and maximum drawdown was calculated. The output
graphs from the logger have been included in Appendix G. The well schematic (fig.
9) provides a visual representation of the results.
HOBO water level logger for
measuring well drawdown
during pumping. The device
records pressure changes
which can be converted into
depths.
16
The two measurements agreed within +/- 5mm
N.T.JEFFRIES Masters Thesis REBE 50 | P a g e
Fig 9 – Schematic showing drawdown at Pump 2 (3 – 9 Nov 2010)
GWL = - 6.25m
Static P = 203.5kPa
Pressure sensor = -18.5m HOBO
The tests indicate a potential 540mm drawdown in water levels during pumping. This
will need to be considered in the later analysis.
Steam test
experimental set‐up
N.T.JEFFRIES Masters Thesis REBE 51 | P a g e
By plotting steam mass against time one can determine the rate of steam production.
The energy required for this can be calculated using the latent heat of vaporization
(Lsteam) = 2260kJ/kg at 100oC. This energy amount can then be compared to solar
irradiation H to estimate collector conversion efficiency (ηcollector)
Steam production tests were carried out over two days at Pump 4. Tables of results
have been included in Appendix H. Graph 6 below illustrates some of these results:
steam
(grams) Steam production tests ‐ 1st Dec 2010, 14:00 ‐ 15:00
0.6
y = 0.0456x + 0.0893
0.5
in focus y = 0.0369x + 0.0427
0.4 boiler not full
boiler full
y = 0.0335x + 0.0075
0.3
dirty collector
0.2 clean collector
y = 0.0235x + 0.0109
0.1
0
0 2 4 6 8 10 12 14 time (mins)
Key findings:
The last finding has a big implication on overall pump performance, (as well as the
continuity of the pump monitoring) and is a key issue that will need to be addressed
for the next design iteration. The likely explanation is that if the boiler walls are not in
contact with water (see fig 10), the radiative surface is decreased and therefore less
energy will be transferred to heating the water.
From previous tests carried out in Netherlands, the reduction in steam production is
most pronounced when the boiler 30% full17.
17
Telephone conversation with Gert Jan Bom, 14th January 2010.
N.T.JEFFRIES Masters Thesis REBE 52 | P a g e
The results of the steam tests are referred to later when considering the overall
efficiency of the system.
Hand drawn open wells are discounted for the reasons described at the end of this
section. This leaves R & W pumps as the only realistic alternative for comparison.
The data collected to allow a comparative study was as follows:
The only data on hand drawn open wells was collected from a family who cultivate
kale on a plot of land adjacent to Pump 1 using this method. In this farm irrigation
required lowering half a tyre inner tube attached to a nylon rope into the well, flicking
N.T.JEFFRIES Masters Thesis REBE 53 | P a g e
the rope so the tube submerges and then hauling it up. Each half tube was enough
to half fill the watering can. The following observations and measurements were
collected on this irrigation method:
This is very small revenue compared to the effort invested, essentially representing a
subsistence level activity. Additionally the physical posture required to scoop and lift
the water would mean that it would be impossible to sustain the rate of filling
continuously.
Rope and washer pump
at test site 7. The low
cost manual technology
is the closest comparable
technology to the STPs.
N.T.JEFFRIES Masters Thesis REBE 54 | P a g e
• Proximity to a main road
• Household size and make up
• Number of livestock
• Mobile phone ownership
• Plot size
The data that has been collected from observations and discussions with farmers,
IDE field staff and market traders and includes:
This data is not presented explicitly in one place; rather it has been integrated into
several aspects of the overall analysis.
N.T.JEFFRIES Masters Thesis REBE 55 | P a g e
7. RESULTS AND ANALYSIS
7.1 Performance (C1)
There are two aspects of pump performance:
1. Performance against target specifications
2. Performance against theoretical potential
The first will inform the assessment of viability for the conditions in Ethiopia; the
second indicates the degree to which the system performance deviates or correlates
with the theory and design. This will help inform the additional aims of the study
(section 1.3)
Tables 14 & 15 summarise the results; a calculation sheet has been included in
Appendix J containing all the raw data and illustrates the process through which the
results were generated.
Table 14 – Pump 4 - Continuous monitoring (D2) summary
Pump
PUMP 4
DST efficiency Vol. factor Flow Daily volume Hyd. energy
km % l/str l/s litres cubm.m
28‐Oct 51.4 86% 0.117 0.158 2998 22.5
29‐Oct 69.2 80% 0.109 0.157 3763 28.2
30‐Oct 78.2 88% 0.120 0.194 4695 35.2
31‐Oct 83.4 86% 0.118 0.189 4918 36.9
1‐Nov 69.7 84% 0.115 0.190 4002 30.0
2‐Nov 70.4 88% 0.121 0.198 4241 31.8
3‐Nov 62.0 84% 0.115 0.172 3557 26.7
4‐Nov 67.4 83% 0.113 0.210 3799 28.5
5‐Nov 77.3 93% 0.127 0.230 4911 36.8
6‐Nov 78.0 88% 0.120 0.199 4685 35.1
7‐Nov 78.7 86% 0.117 0.194 4618 34.6
8‐Nov 50.9 80% 0.110 0.152 2792 20.9
9‐Nov 50.8 88% 0.120 0.200 3059 22.9
10‐Nov 57.0 69% 0.094 0.143 2677 20.1
11‐Nov 68.7 80% 0.110 0.170 3765 28.2
12‐Nov 66.3 79% 0.108 0.160 3597 27.0
13‐Nov 46.7 70% 0.096 0.158 2233 16.7
15‐Nov 64.9 76% 0.103 0.147 3343 25.1
Average 83% 0.113 0.179 3758 28.2
Maximum 93% 0.127 0.230 4918 36.9
th
Note: Full cloud cover on 14 Nov
N.T.JEFFRIES Masters Thesis REBE 56 | P a g e
Table 15 – Pump 1 - Continuous monitoring (D2) summary
DST Efficiency Vol factor Flow Daily volume Hyd. energy
PUMP 1
km l/str l/s litres cubm.m
17‐Nov 68.78 72.1% 0.073 0.129 2508 32.6
18‐Nov 59.77 64.2% 0.065 0.102 1942 25.2
19‐Nov 56.18 69.0% 0.070 0.114 1950 25.3
20‐Nov 95.42 83.5% 0.085 0.162 4025 52.3
21‐Nov 81.5 71.6% 0.073 0.129 3091 40.2
22‐Nov 69.4 71.8% 0.073 0.124 2620 34.1
23‐Nov 66.97 66.3% 0.067 0.124 2230 29.0
24‐Nov 70.9 63.8% 0.065 0.111 2286 29.7
27‐Nov 68.1 63.8% 0.065 0.115 2386 31.0
28‐Nov 67.4 64.4% 0.065 0.120 2152 28.0
Average 0.690 0.070 0.123 2519 32.7
Maximum 0.835 0.085 0.162 4025 52.3
Note: Significant cloud cover on 25th, 26th Nov
55
50
45
40
35
30
25
20
15
10
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
day
Pump 4 Pump 1 Target specification
However if we are interested in the true future potential of the pumps, it is more
appropriate to look at the maximum output values. These represent the days where
there were fewest impediments to the operating of the pumps such as cloudy
N.T.JEFFRIES Masters Thesis REBE 57 | P a g e
weather and pump malfunctions. In this context the maximum hydraulic energy
output exceeds the target specification by 25 and 75% (Pump 4 and 1 respectively).
Another way to express this is to ‘clean’ the data by removing any days where there
were significant (>1 hr) stoppages due to cloud or technical issues such as:
• Pump malfunctions
• Technician servicing or replacement of parts
• Multiple boiler fills because of leaky feeder pump
This is possible as part of the enumerator’s duties was to log all stoppages and other
events detailing the reason and duration. By eliminating days where (i) multiple
stoppages occurred; (ii) cloud covers permitted < 5 hrs pumping, it is possible to
isolate ‘ideal’ operating days. The graph below shows this new data set:
55
50
45
40
35
30
25
20
1 2 3 4 5 6 7 8 9 10 11
Pump output Target specification Average Test
It can now be seen that 100% of the results exceed the target specification by an
average of 20%. Note: the peak on test day 9 where output was almost twice the
target specification. For this day the collector was had been cleaned in the morning,
the pump had undertaken a full technical service the previous day including the
replacement of several seals and valves and the weather was clear throughout i.e.
system and weather was optimal.
The pump log also recorded engine speeds at 15-min intervals throughout the day.
Maintenance of high engine speed is identified as an important factor in engine
efficiency, because:
N.T.JEFFRIES Masters Thesis REBE 58 | P a g e
• High engine frequency reduces entrance condensation (Bom G. , 1993, p.227)
• Lower rpm leads to a higher steam consumption per stroke
There is another reason: if we plot average daily RPM v. volumetric efficiency, it can
be seen that engine speed also has an effect on the performance of the water pump.
Average RPM RPM v. Pump efficiency at Pump 4
110
105
100
95
R² = 0.7311
90
85
80
75
70
65
60
60% 65% 70% 75% 80% 85% 90% 95%
Volumetric efficiency
This can be explained as a slower pumping rate allows more time for leakage
between the piston seals and valves (slippage).
KEY FINDINGS
N.T.JEFFRIES Masters Thesis REBE 59 | P a g e
7.1.3 Performance analysis against theoretical potential
Combining the volumetric, ground water depth and solar irradiance data allows an
estimate of overall system efficiency (ηsystem) to be calculated using the equation:
ηpump Ehyd/E sol
Solar irradiation (E sol) is the total incoming solar energy available for conversion to
hydraulic energy (Ehyd). To calculate E sol values for each pump test, average t-min
interval irradiances (DNI t) were calculated and multiplied by the time interval and
collector size, for each time interval between the start and end for the pump test and
the area of the collector (A = 3m2). This can be expressed as the following equation:
E sol DNIt x t x 60 x A
By applying these equations, system efficiencies have been calculated for all
pumping days. A full table of results is presented in Appendix K, the key findings are:
Referring to section 5.8, the expectation for the current design is ηsystem = 1.6%.
Hence it is immediately clear that the system is performing well below expected.
Before exploring the reasons, it is again necessary to ‘treat’ the data to eliminate the
effects of clouds and technical stoppages; this has been done by taking the average
of the top ten results giving ηsystem = 0.55%. Referring back to the section on pump
theory, system efficiency is the product of collector, Carnot, engine and pump
efficiencies. Assessing each of these using monitoring data, allows us to investigate
which part of the system is underperforming.
Collector efficiency - steam tests indicate that the collector is performing a little better
than expected (Appendix H). Average = ~ 53%.
This gives a total pump efficiency = ~ 70% which agrees with expectations.
N.T.JEFFRIES Masters Thesis REBE 60 | P a g e
The conclusion is that the lower than expected system performance is attributable to
Carnot and engine factors. This supposition is supported by pressure readings taken
by the enumerator (30, 31st Oct, 13th Nov) from a meter attached to the buffer
cylinder (see Appendix J).
The interpretation and implication of the pressure data can be explained as follows:
The situation is worse because Pa is boiler pressure, and Pw is the pressure that
drives engine output. The schematic below illustrates the reason that Pa drops to Pw
giving a much lower working pressure than desired.
Working Buffer
Boiler
space space
Steam hose
X
160cc 230cc 2700cc
Inlet
valve
Pa
Pb
The buffer volume (Vb) stores at least one cycle’s volume of steam (Vw) plus an
allowance for entrance condensation. The reason for having a buffer space is to
minimise the diameter of the steam hose. Without a buffer the hose would have to be
a larger diameter to allow the required amount of steam to be delivered for each
cycle. A large steam hose is undesirable because it increases: (i) heat losses (ii)
entrance condensation as more condensate enters the engine.
N.T.JEFFRIES Masters Thesis REBE 61 | P a g e
However a buffer means that each opening of the inlet valve leads to an increase in
the volume that the steam can occupy, reducing the working pressure proportional to
the % increase in volume.
In other words a large buffer is required to maintain a high working pressure in the
engine. For our system Vb/Vw = (230/(160+230)) = 1.43 which is too low. The effect
is shown below with the current volume ratio dotted on. The results suggest an
absolute pressure in the engine of 1.5 bar (T = 111oC, Carnot = 2.8%)
Effect of buffer:working space volume ratio on working pressure
2.5
2.0
1.5
1.0
0.5
0.0
0 1 2 3 4 5 6 7 8 9 10 11
Volume ratio
Actually the reality is not this bad, as evinced by actual Pw readings observed on the
pressure meter which rarely dropped this low18. However it does show clearly that a
buffer size at least 6x larger than the working space is needed to maintain a high
working pressure.
This is not the only reason that the engine does not perform as well as expected.
Other factors include:
• Non-insulated steam hose - loses heat thereby reducing efficiency and
increases amount of condensation entering the engine i.e. increases entrance
condensation.
• Optimal cut off % - still to be determined (see Appendix D)
18
This can be explained by the fact that the boiler continues to provide steam pressure when the inlet
valve is open
N.T.JEFFRIES Masters Thesis REBE 62 | P a g e
Key findings:
N.T.JEFFRIES Masters Thesis REBE 63 | P a g e
7.1.4 Potential errors
The following have been identified as possible sources of error in the analysis of the
pump performance:
• Feeder pump losses: a proportion of the system power is required to pump the
condensate.
For a flow rate of 0.2l/s from a depth of 7.5m assuming ηpump = 70%, hydraulic
power = 21W. Therefore parasitic losses are between 1 – 1.5%.
• Bucket volume – the 20ltr bucket was a simple but important piece of equipment
in many of the monitoring activities, helping calculate flow rates and validate
measurements. The buckets, which were the cheap plastic type bought from the
local market, had a scale on the inside walls with the top notch (20ltr) just below
the rim. The bucket volume was validated once on an electronic scale and
weighed in at 19.7kg. Filling the bucket was not always precise, as the water
pulsed out of the outlet making it unclear exactly when the bucket was full.
Averaging the bucket fills over a large data set should have had the effect of
reducing reduced the errors related to this potential inconsistency.
19
Could be explained by negative slippage
N.T.JEFFRIES Masters Thesis REBE 64 | P a g e
• Pyranometer location – the maximum distance between furthest installations and
pyranometer was 6km. It has been assumed that within this radius all farm
locations experienced broadly similar solar irradiance, although this may not
always be true (see notes on Fig 3)
• Farmers will pump water only when they need it, whereas the study would like
the farmers to operate as often as possible.
• Due to durability issues and other equipment issues not all pumps were
operational all the time
For these reasons, in this part of the study it was both (i) difficult to get a full and
continuous data set; (ii) necessary to exercise care in attributing the correct reason
for pumps not being used.
Having made these caveats, table 16 summarises all the independent usage data
collected during the monitoring period. The data represents 75 days of
measurements. Discounting cloudy days, pump 2 (rarely operated as not a farm site)
and continual monitoring days, we can say that there were 25020 possible monitoring
days. Therefore the dataset represents a not insignificant 30% of the total.
20
8 pumps x 45 days = 360 days less 10 cloudy days, 33 x D2 days = ~ 250
N.T.JEFFRIES Masters Thesis REBE 65 | P a g e
The results indicate that at least five out of the eight farmers are able to operate the
system independently.
Even if the pump is operational, from observations, the following conditions were
required to improve the likelihood of independent usage:
Photo was taken on 29th Nov at
Pump 6. The land had only recently
been furrowed and seedlings had
not been planted yet. Therefore
there was no use for solar water.
The farm has a small family
composed of women and young
children which explains the slow
preparation of the land.
Of all the test sites, Pump 9 demonstrated the most extended and consistent level of
independent use. The specific circumstances that created this are surmised as:
21
On one occasion farmer ran from his farm, across the whole village to ask technicians working at
Pump 7 to come back to his farm as his pump had a problem
22
This area is far from the main road so does not have commercial onion growers commissioning
lucrative seedling production.
N.T.JEFFRIES Masters Thesis REBE 66 | P a g e
• The alternative was a R & W pump at far end of plot. Before solar pump
farmer spent up to 4hrs/day using this method.23
The farmer at pump 9 is not educated, technical or sophisticated (using the simple
metric of mobile phone ownership). However with only a minimal amount of
guidance, he has managed to operate the pump day after day. It was even observed
on occasions that the farmer has made small modifications and repairs to keep the
pump going. This is a strong indication that the pump is not too complicated to
operate by the target demographic.
It is likely that future iterations of the pump will be even simpler to operate and the
purchases will be accompanied by proper training and illustrated local language
training manuals. This will further improve the ability for farmers to maintain and
operate the systems.
Key findings:
23
The evidence is a hole in the metal pump frame worn down by the friction of the nylon rope
N.T.JEFFRIES Masters Thesis REBE 67 | P a g e
7.3 Comparability (C3)
The relative advantage of the STP over existing irrigation technologies will be a
strong part of the decision making process when considering a pump purchase. For
the plot size and water depths encountered on the test farms, the only real
alternative is the R&W pump. The datasets and methods used to assess relative
advantage have already been described; the following section presents the results
and analysis.
From timing different individuals at different locations (author included), the average
time to fill a 20 litre bucket was estimated. Combining this with maximum recorded
pumping volumes at the two main test sites, an estimate of time saved can be made:
This is a significant saving and does not include time taken for water distribution.
N.T.JEFFRIES Masters Thesis REBE 68 | P a g e
To validate the above results the irrigation water requirement for the crops growing
on the farm at the time of monitoring was estimated as 3150 litres24.
Accelerometer being attached
to rope and washer pump at
Pump 7. The devices checks for
movement at a pre‐set interval
allowing pumping duration and
volumes to be estimated.
On the surface STPs seem to offer key advantage over manual pumping of savings
in labour input, both in the pumping and distribution of water. This saving is both in
terms of time and expended energy.
However STPs do not totally eliminate the need for manual input as they still require
the farmer to be present on site during operation. The key difference is that it does
not require a complete immersion in the activity. The exertion is much less and if the
STP is operating well, it allows other tasks to be carried out in parallel.
There are some other important factors to consider when assessing the relative
advantage. If we look at the raw data for the above farm (Appendix M) in more detail,
it can be seen that most pumping is carried out between 6 – 8am and 4.30 – 6.30pm.
This is because farmers choose to irrigate certain crops when the sun is low and
thus evaporation losses minimised. The cooler temperatures at this time are almost
certainly an additional factor in this choice.
24
Estimate based on 63 ‘merips’ of onions seedlings each with a daily water requirement of 50 litres.
N.T.JEFFRIES Masters Thesis REBE 69 | P a g e
onion seedlings for commercial farmers. Even though the STPs were operational at
this time, the technology was ignored in favour of a more labour intensive but
convenient method that allowed the farmers to choose their irrigation time.
A final point worth considering is that although manual irrigation methods are more
labour intensive, they do offer a more efficient use of water25 as it is containerised at
the source and then directly applied to the crops with little or no wastage. If STPs are
to achieve high water efficiency, the method of distribution will have to be
considered. Unlined furrow irrigation can reduce the outlet-to-crop volume by 40% or
more (Fraenkel, 2006, p.18) depending on soil type. Therefore future dissemination
of the technology will need to go hand in hand with compatible and efficient
distribution methods.
Key findings
25
This assumes that for STPs farmers use the water differently than with manual pumping (as
observed)
N.T.JEFFRIES Masters Thesis REBE 70 | P a g e
7.4 Economic assessment (C4)
In Ethiopia, the costs and payback time for equipment purchases needs to be
minimised. Potential buyers are risk adverse and micro-finance institutions limit
payback period for loans to 1 year (see Appendix N).
As the solar pump is not yet a consumer product, the cost is not yet known, therefore
it is not possible to undertake a traditional payback analysis. However a cost
estimate has been made by the designer, based on a breakdown of materials and
labour which values the pump at about $250-300 (Bom G. J., 2009).
For the economic analysis, it is assumed that farmers will not buy technology that
requires more than 1 year to payback. Comparing the potential annual revenue with
the pump cost estimate will give an indication of the economic viability as well as any
potential flexibility on manufacturing cost target.
Annual revenue has been estimated by considering a hypothetical farm with a 600m2
growing area, this is the average size encountered across the 8 test farms (refer to
table 2). The crop mix was chosen based on observations of existing growing
practices and current market demands. Fig. 12 is a schematic of the hypothetical
farm; tables 19 a/b present estimates of potential revenue and input costs, based on
the stated assumptions. The analysis produced the following economic variables:
• Potential gross revenue = 17,240 birr
• Input costs = 3240birr
• Potential net revenue = 14,000birr = $87526
• Production cost estimate = 4000 – 5000 birr ($250 -300)
The analysis indicates that the net annual revenue greatly exceeds the estimated
production cost and under these circumstances suggests a financially attractive
proposition.
Although the analysis is based on a hypothetical plot of land, the estimated revenue
is not unrealistic if compared to the income of farmers that have using solely manual
irrigation methods27:
• Soado Abate - 2008/9 – 30, 700 birr (2 years)
• Tadessa Mekuria (R & W monitoring farm) – 2009/10 – 18, 000 birr
• Tadele Tiko - 2009/10 - 25, 000 birr (3 growing seasons)
Finally, several farmers were asked what they would be prepared to pay for a STP
system based on the perceived benefits and the cost of other equipment. Only two
farmers responded and offered a figure between 5000 -7000 birr as a reasonable
price. This combined with the annual revenue projection, suggests that there is some
upward flexibility in the production cost of the system.
26
Dec 2010 exchange rate $1 = 16 birr
27
Information from interviews with farmers and agricultural extension officers from IDE
N.T.JEFFRIES Masters Thesis REBE 71 | P a g e
Fig.12. Hypothetical farm to estimate revenue potential
PEPPERS
(60 sq.m.)
onions seedlings
sold or transplanted
to onion beds at 1:30
ratio
CABBAGES
(60 sq.m.)
SEEDLINGS
(100 sq.m.)
600 sq.m plot
N.T.JEFFRIES Masters Thesis REBE 72 | P a g e
Table 19a – Estimated revenue for hypothetical farm
REVENUE
Time to
Crop Irrigated area Daily water harvest Crop density Productivity Revenue per unit Revenue
(m2) (litres) plants/m2
N.T.JEFFRIES Masters Thesis REBE 73 | P a g e
Table 19b – Estimated input costs for hypothetical farm
ETB 3,240
Note:
1. Information based on interviews with IDE staff and observations at project farms.
2. Onion seed requirement 2 tins/merip
3. Fungicide/insecticide not mandatory, only required if problem occurs
4. Fungicide for purple blotch; insecticide for thrips ‐ on onions only
5. Distribution system important for efficient water usage,
6. Distribution cost based on half 200 litre drum, 20m of 1" hose + labour
7. Labour costs zero as undertaken by family
N.T.JEFFRIES Masters Thesis REBE 74 | P a g e
8. DISCUSSION
The data collected from the monitoring study indicates that the solar pump system
can produce a volume of water that is useful to the conditions found on the test
farms. Furthermore, the level of independent usage recorded shows that the
equipment is simple enough to be operated by farmers without specialised skills or
knowledge. Compared to current manual alternative, the new pump offers an
irrigation method that is much less physically exerting and time consuming. Finally
using a simple economic analysis, it has been demonstrated that the potential
revenue from utilisation of the pump will payback the likely investment of the pump
within a short period.
All of this together suggests that the pump is a viable micro-irrigation technology for
this scale of cultivation. But will farmers buy the equipment? To attempt an answer
this question, it is worth considering two facts about the local population:
• There is a high aversion to risk. The evidence for this is in other attempts
to promote and sell irrigation technology28.
• The value attributed to time-saving is not so high, particularly if your family
is large and you have an abundant labour pool.
Therefore even though the four tests for viability have been satisfied, there is a
sense that more is needed to increase its attraction to potential buyers. The solutions
to this could be to:
Assuming that the best course of action is to create the best possible product, the
following section discusses the ways that the perceived value of the system might be
improved.
N.T.JEFFRIES Masters Thesis REBE 75 | P a g e
However it has already been proved that the current system is underperforming
significantly (section 7.1.3). What happens if we address this? Graph 11 shows the
potential of the pump if system improvements were made:
Graph 11 - Potential irrigated area at different GWLs for current and expected performance
6000
Area
(m2) Present
Expected
4000
Field size currently being
irrigated by engine
pump GWL = 6.75m
2000
0
5 10 15 20 Water depth (m)
Note:
1. Based on average solar irradiation 27Oct ‐ 4Nov in Ziway, Ethiopia
2. Daily water requirement 5litres/sq.m.
The above graph is based on measured and theoretical efficiencies shown below:
The results are clear - if the mechanical issues to the pump system are addressed,
there is a potential for a three-fold increase in output. This may have a further
implication. The dotted line on the graph shows the total area of land that a farmer
currently irrigates with a diesel engine pump. So if the theoretical performance could
be achieved, it would make the STP a potential alternative to a diesel pump.
This has obvious and important implications to the diffusion of the technology. If a
reliable alternative existed that can perform as well as a diesel pump, but is cheaper
N.T.JEFFRIES Masters Thesis REBE 76 | P a g e
and has no operating costs, then suddenly the relative advantage of the system
becomes much more pronounced.
Note: the above comparison is not strictly accurate as we are comparing gross solar
pump output against net crop water requirement. Distribution efficiency is the factor
that differentiates the two.
Improvements to the pump
that bring the system efficiency
in line with the theory will
allow areas of land that are
currently irrigated using diesel
pumps to be considered,
making the technology more
attractive to potential
customers.
From observations and measurements during field trials and discussions with the
designers and technicians (see Appendix P), the following are identified as potential
system improvements:
N.T.JEFFRIES Masters Thesis REBE 77 | P a g e
All of these are quite simple changes that individually make a small contribution, but
collectively amount to something substantial. Practica is already implementing some
of these changes, which will be incorporated into the next prototype.
• A flatter collector with a boiler placement further from the dish will focus solar
power just onto the base of the boiler leading to more efficient heat transfer
and allowing the other sides of the boiler to be insulated.
• Placing the collector closer to the engine will reduce the length of the steam
hose, reducing heat losses and eliminating the need for a buffer
Flatter collector mounted on
the engine will improve the
overall efficiency of system.
Photo source: Practica
The above modifications will improve the hydraulic energy output of the pump. There
are other changes that are being considered to improve the overall user experience
and therefore improving the overall quality of the product:
N.T.JEFFRIES Masters Thesis REBE 78 | P a g e
8.2 Support infrastructure considerations
There are other considerations beyond mechanical changes that may enhance the
system. As already mentioned an efficient distribution is critical in making most
effective and efficient use of water.
Efficient distribution systems allow
more water to reach the crops. The
photo shows a section of 50mm
PVC pipe with emitter holes spaced
to suit the width of furrows. The
pipe is connected to the solar pump
with some home made lay flat
hose. The farmer controls which
furrows watered using band valves
made from tyre inner tube
Another intervention that will greatly enhance the system will be a means of water
storage. If farmers can pump and use water at a time that is convenient for them, the
perceived value of the technology will be greatly increased. The simplest way of
storing water is the excavation and lining of underground storage tanks around the
plot. These tanks can be filled during the day so in the evening and early morning,
the farmers can fill their watering cans and water their crops. Ignoring the safety
issues and increase in mosquito breeding habitats, this does make things easier.
A more effective method could be by creating elevated storage, which eliminates the
exertion associated with scooping water and carrying cans. Elevated storage gives
the farmers total control and convenience in the method and time of day that they
utilise their water. It also opens up the potential for advanced distribution methods
such as high efficiency drip irrigation (see photo).
N.T.JEFFRIES Masters Thesis REBE 79 | P a g e
Experimental low cost
elevated storage structure
being tested at FTC. Photo
source: Bob Yoder, IDE
Some reasons that can be offered for the operational difficulties can be attributed to:
In the future, many operational issues will be eliminated as the engineering design of
the pump is improved. However to be sustainable there will need to be appropriate
training as well as an easily understandable operations manual that includes:
29
An obvious analogy is the bicycle for which there is a whole raft of self-fix it problems that the owner
can address and the more complex problems that are left to a bike mechanic.
N.T.JEFFRIES Masters Thesis REBE 80 | P a g e
• Schedule of basic maintenance checks to ensure good operation
• Phone number/details of local service provider
Appendix Q - Schedule of O & M tasks is a first attempt to set out some of these self-
maintenance tasks. Appendix R shows the format and content of a typical
Operational Risk Assessment.
N.T.JEFFRIES Masters Thesis REBE 81 | P a g e
9. CONCLUSIONS
The central question posed by this thesis was whether solar thermal pumps can be a
viable option for small scale irrigation in Ethiopia. To examine this question, the
concept of viability was deconstructed into four smaller constituent parts. The
selected criteria were:
Early on in the study it was clear that the pump, under the right circumstances,
produces fair quantities of water, generates sufficient revenue and is simple enough
to be operated by a typical farmer. The relative advantage over existing systems was
also proven, although less convincingly. As the report explains, this was due to
cultivation practices and socio-economic factors within the local community.
In this context, there are certain issues that need to be addressed, so that the
perceived value and advantages of the pumps can be increased. A better, more
valuable product is required to overcome the inherent cautiousness of the local
community and persuade ‘early adopter’ individuals to begin the process of
technology adoption (see Graph 1). The questions highlighted and investigated in
this part of the analysis included:
1. Although the system fulfils the basic design specification, why does it fall so
far short of the performance predicted by theory?
2. Even when a pump is fully operational, why do farmers sometimes still use
manual methods?
The cause of the poor engineering performance can be attributed to the prime mover
(steam engine). This is mainly due to low working pressure in the working cylinder
caused by an undersized buffer. Low pressure leads to a reduced Carnot efficiency,
slower engine speed, higher steam consumption and lower pump efficiency.
Pump underutilization can be explained because the system does not give total
convenience and flexibility for farmers to choose their preferred irrigation schedule.
The solutions to these issues will be found both in modifications to the mechanical
system as well as by promoting the pump along side appropriate support
infrastructure.
N.T.JEFFRIES Masters Thesis REBE 82 | P a g e
In practical terms this could entail an improved collector dish located much closer to
the engine. This will lead to an increase in thermal conversion efficiency and allow a
shortening of the steam hose, thus eliminating the requirement for a buffer. The
development and promotion of effective distribution and storage systems specifically
suited to the solar pumps will create a more effective and convenient irrigation
product. The development and dissemination of suitable training material will
contribute to an increased understanding and confidence for pump users.
The duration and scope of the monitoring study was limited by available time and
resources. Timing was also a factor as the newness of the prototype, the late
installation dates and the slow response of some of the farmers to the technology
meant that there was a ‘settling in’ period as mechanical issues were tweaked and
farmers reconfigured their land to accommodate the new water resource. As more
data is collected it will be possible to reinforce or adjust the findings of this thesis.
N.T.JEFFRIES Masters Thesis REBE 83 | P a g e
10. BIBLIOGRAPHY
Bom, G. (1993). Energy losses through entrance condensation in small vapour engines. Solar Energy ,
Vol. 50 (No. 3), pp. 223‐228.
Bom, G. J. (2009). Solar Pump Development ‐ History of Past 2 years. Papendrecht: Practica Internal
Memo.
Boyle, G. (2003). Energy Systems and Sustainability. Oxford: Oxford University Press.
Boyle, G. (2004). Renewable Energy for a Sustainable Future. Oxford: Oxford University Press.
Burney, W. B. (2010). Solar‐powered irrigation enhances food security in the Sudano‐Sahel.
Proceedings of the National Academy of Sciences , Vol 107 (No. 5), 1848 ‐1853.
Butti, J. P. (1954). A Golden Thread ‐ 2500 Years of solar architecture and technology.
London/Boston: Marion Boyars.
Delgado‐Torres, A. M. (2009). Solar thermal heat engines for water pumping: An update. Renewable
and Sustainable Energy review , Vol. 13, 462 ‐472.
FAO. (2010). Evapotranspration Data. Retrieved Aug 10, 2010, from FAO Cropwater:
http://www.fao.org/nr/water/infores_databases_cropwat.html
Fraenkel, J. T. (2006). Water Lifting Devices (Third ed.). Rugby: Intermediate Technologies and FAO.
FWR. (2010). The Development of Effective Community Water Supply Systems Using Deep and
Shallow Well Handpumps. Retrieved January 15, 2010, from Foundation for Water Research:
http://www.fwr.org/wrcsa/tt13200.htm
Kibret. (2010). The impact of a small‐scale irrigation scheme on malaria in Ziway, Central Ethiopia.
Tropical Medicine and International Health , p 41‐50.
Rogers, E. M. (2003). Diffusion of Innovations (5th ed.). New York: Free Press.
SELF. (2008). A cost and reliability comparison between solar and diesel powered pumps.
Washington: Solar Electric Light Fund.
Thomas Smith, C. M. (2005). NIFTE Solar water Pump: A Technical Brief. Oxford: Thermofluidics Ltd.
Thurston, R. H. (1883). A History of the Growth of the Steam‐Engine. London: Kegan, Paul. Trench
and Company.
Weir, J. T. (2006). Renewable Energy Resources. Abingdon: Taylor and Francis.
Wikipedia. (2010). Pumps. Retrieved Jan 15, 2010, from Wikipedia: Wikipedia
WorldBank. (2010, Jan 10). Carbon Dioxide Emissions (Kt). Retrieved Jan 10, 2010, from World Bank
Development Indicators: http://data.worldbank.org/country/india
Yoder, B. (2010). Solar steam pump development and testing . Addis Ababa: IDE.
N.T.JEFFRIES Masters Thesis REBE 84 | P a g e
YW Wong, K. S. (1999). Solar thermal water pumping systems: a review. Renewable and Sustainable
Energy reviews (3), 185 ‐ 217.
Zagoni, C. (2010, November 10‐30). Email: Solar Zenith Angles. London.
Zeller, B. (2003). Assessment of the economic performance of a solar thermal water pump for
irrigation in semi arid India. University of Hohenheim: Institute for Agricultural Economics and Social
Sciences.
N.T.JEFFRIES Masters Thesis REBE 85 | P a g e
APPENDICES
N.T.JEFFRIES Masters Thesis REBE 86 | P a g e
APPENDIX A
The idea for a small solar powered water pump was first conceived by the designer,
Gert-Jan Bom in 1983, after the successful deployment of the Volanta pump - a
medium to deep well, easy to service, easy to mechanize, flywheel-powered hand
pump. Having installed a number of these pumps in Burkina Faso, Niger, Cameroon
and other places in Africa – the question was asked whether the pump could be
powered by solar energy. Three options were considered – photovolatics , Stirling
Engines and Rankine cycle. As PV costs were very high at that time, and it was felt
that the high temperature differences required for Stirling Engines would be difficult
to achieve, the focus from early on was on Rankine cycle vapour engines.
Entrance condensation
Shortly afterwards a relative in Japan involved in the rice industry approached the
designer and asked Bom to consider the use of rice husks to produce steam to
power an engine. A prototype was designed based on existing steam theory and
then constructed and tested. Immediately it became clear that the amount of steam
consumed per cycle far exceeded the theoretical amount calculated.
This was a first encounter with the phenomenon known as entrance condensation.
Since then it has become the key design consideration in the design of all versions
of small solar steam pumps. Once the phenomenon was identified by the designers,
an historic literature review was undertaken and it was found that entrance
condensation (eintrittskondensation) had been mentioned in books dating back to
1928. However as the issue is less pronounced in larger engines, due to lower
volume/surface area ratios, for the 5 – 100Hp engines built at that time designers
were happy to accept the associated 5 – 10% energy losses.
Entrance condensation is the cyclical heating and cooling of the walls of the working
space in the engine. When steam enters the engine the steam condenses as it
comes into contact with the relatively cool walls of the engine. The steam continues
to condense and heat up the walls until the steam reaches saturation temperature.
Before this point no pressure can build up in the engine to move the piston. As the
piston moves, new cylinder wall is exposed which also needs to be heated up. At
the end of the work stroke, the exhaust valve opens causing the pressure to drop to
atmospheric. The pressure drop causes the accumulated condensate on the cylinder
N.T.JEFFRIES Masters Thesis REBE 87 | P a g e
walls to evaporate, the energy for which is provided by the walls of the cylinder,
which therefore drop in temperature. The heating and cooling continues with
associated energy losses at each cycle.
At the time DGIS funded B & R Consultants to carry out further research on entrance
condensation with the ultimate aim of assisting in the solarisation of the Volanta
pump. This led to the publication of the paper: Energy Losses through Entrance
Condensation in Small Vapour Engines (Bom 1993) which identified the key factors
that affecting factors these include: engine frequency, internal area/volume ratio, use
thermal properties of cylinder material and others.
During this study a literature review was carried out to see what others were doing in
the same field. A number of companies from Germany, France and Finland had
developed vapour engines – most of them much larger in size
Sofretes, France, a three cylinder, up to 10kW all metal engine using Butane as
working fluid heated by a 75-3000m2 flat plate collector array, overall efficiency low
but not quantified.
Indo-Swiss solar pump an experiment in India, overall efficiency 0.25-0,5%, all metal
single cylinder linear engine directly coupled to a reciprocating pump, flat plate
collector array of 10m2, working fluid Freon.
Dornier Systems, a 1kW two cylinder boxer engine, all metal, Freon as working fluid,
heated by a 25m2 array of vacuum tube collectors, with an hydraulic output of about
400W. Testdata from a prototype in Indonesia claims 1-3% efficiency.
Wrede Ky, Finland, a 90W hydr. double acting single cylinder all metal engine,
organic working fluid (type inknown) claimed peak efficiency 0.8%
However the lower efficiencies, use of multi- cylinder metal engines and combined
inlet and outlet valves indicated that entrance condensation had not been
considered.
At this point Bom approached an inventor-friend (ing. T.Visser), an expert in seals for
dredging pumps and a keen model steam engine enthusiast. Visser has been
working closely with Bom ever since on the solar steam pump and has contributed
greatly to its design. Between them they considered all types of engine (piston,
N.T.JEFFRIES Masters Thesis REBE 88 | P a g e
rotary va
ane, turbinees) to deterrmine which h configurattion would bbe least afffected by
entrancee condensa ation. Afterr some delib beration it was
w decided d that a perristaltic
engine may
m provide e the best solution
s to tthis problem
m. The rubb ber construction,
separation between n hot and cold
c cycles, no valves and a reaso onably highh operating
frequenccy – seeme ed to best obey
o the newly discove ered rules oof entrance
condenssation avoid dance. At the
t time it was
w also tho ought that it may even n have the
potential to become e a new typ pe of univerrsal engine. A numberr were desig gned, built
and testted over a 5 year perio od – all of th
hem worked d for a while but in the
e end they
suffered
d durability issues
i in th
hat it was no ot possible to get the w
working tubbe to last
long enoough to be worthwhile
w .
Rolling diaphragm
m pistons
After thiss lengthy deviation, am mbitions abbout universsal enginess were put on o hold and d
attentionn was again n focused on o solarisingg the Volannta pump. T The simples st way to
drive thee Volanta pump
p was thhought to bbe a verticall oscillating cylinder co
oupled on
one end d to a crank on the pum mp shaft annd connecte ed to the co oncrete basse of the
pump with the othe er end. For this it was tthought tha at a relatively long stro
oke(20cm)
was nee eded to trannsfer the po ower efficien
ntly. A rolling diaphra gm piston made
m of
nitro-buttyl rubber (N
NBR) seem med to satissfy this requ
uirement. O Once again ana
unforese een problem m arose – the NBR rea acted to thee hydrocarb bon workingg fluid by
slightly expanding
e the
t rubber causing sm mall creasess in the diap phragm which caused d
the diaphragm to fa ail rapidly.
Having encountere
e ed this problem, the neeed for a lonng stroke wwas questioned – the
thought occurred thhat it should
d be possib
ble to shorteen the strokke length annd increase
e
the force
e. This led the designer to start looking at smaller-stro
s oke ‘top hat’’
diaphragms that wo ould not be so prone too
material expansion issues. Th he engine
was mod dified to acccommodate e the new
diaphragm and attached to a fllat plate
collector with a penttane workinng fluid
which waas the curreent solar po
ower
conversioon technolo ogy under
consideraation. End durance tes sting was
carried out
o in Hollan nd 24hour pumping
p
over a onne month period and the results
were reaasonably pro omising, deespite a
small expplosion thatt destroyed
d the
designerr’s garage inn Papendre echt.
Howeverr subsequent field testting carried
urkina Faso did not havve the same
out in Bu
N.T.JEFFR
RIES Masters TThesis REBE 89 | P a g e
degree of success. Engine outputs were much lower than expected, valves operated
badly and problems occurred with the pentane working fluid in that the ground water
temperature was too high for easy condensing and any topping up required due to
leakages were difficult because of lack of availability in the local markets. From this
experience a key decision was made to switch to water as the working fluid.
To create steam at a high enough temperature and pressure required for the engine
meant a new type of collector had to be considered. At this time evacuated tube
collectors were starting to become more affordable as China began to manufacturer
them in large quantities for domestic solar thermal installations. Working now with
the newly formed Practica Foundation, he designed a 72 evacuated tubes collector
system equivalent to a solar surface of 7.2sq.m. The calculated output for such a
system was 75 g/min that would give an engine output of 75 RPM.
Again field testing in Burkina Faso did not produce an adequate or reliable
performance – the principal problem caused by steam bubbles in the tubes
preventing heat transfer and pooling of water at the bottom of the engine which
created unacceptably high levels of entrance condensation.
By now Practica Foundation and IDE had started collaborating with the aim of
developing a solar pump that would operate in similar conditions to IDE’s low-cost
treadle suction pump. The collector that was first considered was a Fresnel
collector, comprising a 3 x 1.2m array of mirror slats each of which can be adjusted
to focus onto a linear boiler giving a peak efficiency of about 30%. The key
advantage for such a system was that if oriented E-W, there should be no need for
active tracking. The engine-pump arrangement comprised a diaphragm piston
engine at the top connected to a diaphragm suction pump by a ‘scottish yoke’30. The
mechanism both rotates a flywheel to produce a smooth pumping motion as well as
synchronises the top mounted inlet and outlet valves.
Field tests, now moved to Ethiopia identified problems with collector focusing and
reflector losses leading to reflector efficiencies that rarely exceeded 10%. The effect
was that pumping was only possible between 11 – 1pm. As well as low collector
performance, the engine continued to experience water ingress from the outlet side
once again leading to high entrance condensation.
The most recent collector type is a parabolic dish. The first dish tested was a $180,
1.4m diameter collector bought as part of a solar cooker kit from a technical school in
30
http://en.wikipedia.org/wiki/Scotch_yokemechanism
N.T.JEFFRIES Masters Thesis REBE 90 | P a g e
Germany31 (Sun & Ice, Kirchweidach, Germany). Having determined that it could
produce adequate steam, the next stage was to reduce the high cost. Research was
carried out on availability of appropriate reflective materials so that proprietary
collectors could be produced rather than buying off the shelf. Appropriate material
was procured from a company in Germany32 (Alanod) that produces thin sheets (0.3,
0.4, 0.5 mm) of treated coated aluminium
The other challenge was achieving accurate focusing through the production of a
precisely shaped dish. This issue was addressed using a concrete mould formed
using a scraper attached to a central pivot. To form the dish, thin triangular strips (or
‘leaves’) are cut out from the sheets of reflective material and these are fanned out
around the mould to form the parabolic shape. Concentric steel rings are then placed
over to fix the leaves in place. To avoid shape distortion, rivets or nails are avoided,
instead a polymer is used to attach the leaves to the steel frame.
From the initial testing period in Ethiopia, it was decided that a reciprocating piston
pump would be more useful as it allowed deep well pumping to suit local ground
water conditions. The engine was rotated to a horizontal reciprocating action,
allowing the outlet valve to be placed at the base to reduce the water build up in the
engine.
The parabolic dish collector, horizontal action diaphragm steam engine, reciprocating
piston pump described above is the system that is currently being tested in Ethiopia.
---------------------------------------------------
Wong (1998) and Delgado –Torres (2007) offer general reviews of the development
of solar thermal pumps.
31
http://sun-and-ice.sdrom.ru/state/AA:navID.36/AB:navID.36/AC:-1.163261631/
32
http://alanod-solar.com/opencms/opencms/index.htmlname, reference
N.T.JEFFRIES Masters Thesis REBE 91 | P a g e
APPENDIX B – Farm Layouts
N.T.JEFFRIES Masters Thesis REBE 92 | P a g e
N.T.JEFFRIES Masters Thesis REBE 93 | P a g e
N.T.JEFFRIES Masters Thesis REBE 94 | P a g e
N.T.JEFFRIES Masters Thesis REBE 95 | P a g e
N.T.JEFFRIES Masters Thesis REBE 96 | P a g e
N.T.JEFFRIES Masters Thesis REBE 97 | P a g e
N.T.JEFFRIES Masters Thesis REBE 98 | P a g e
N.T.JEFFRIES Masters Thesis REBE 99 | P a g e
N.T.JEFFRIES Masters Thesis REBE 100 | P a g e
N.T.JEFFRIES Masters Thesis REBE 101 | P a g e
N.T.JEFFRIES Masters Thesis REBE 102 | P a g e
N.T.JEFFRIES Masters Thesis REBE 103 | P a g e
N.T.JEFFRIES Masters Thesis REBE 104 | P a g e
N.T.JEFFRIES Masters Thesis REBE 105 | P a g e
N.T.JEFFRIES Masters Thesis REBE 106 | P a g e
APPENDIX C – Steam Data
Specific volume Steam characteristics
(m3/kg) Temp (oC)
1600 140
135
1400
130
1200
125
1000 120
115
800
110
600
105
400 100
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Pressure (bar x 10)
Specific volume Temperature
N.T.JEFFRIES Masters Thesis REBE 107 | P a g e
APPEND
DIX D – Efffect of vary
ying cut offf
This sug
ggests that for a working pressure
e of 2.5bar (absolute),, a lower cu
ut off (40%))
ke better use of the exp
will mak pansive energy of the steam.
N.T.JEFFR
RIES Masters TThesis REBE 108 | P a g e
A
APPENDIX E – Single Pres
ssure cycle in Steam Engin
ne
Absolute Displ'mnt
Pressure (bar) Inlet Inlet Inlet (mm)
opens closes BDC opens
1.8 60
1.6
1.4 40
1.2
1.0 20
P attm (measured)
0.8 P aatm (chart)
0.6 0
6.00 7.00 8.00 9.00 10.00 11.00 1
12.00 13.00
0 14.00 15.00 1
16.00
N.T.JEFFRIES Massters Thesis REBE
N 109 | P a g e
Appendix F SPN1 Ouptut Data .dt6
1400
1300
1200
1100
1000
900
800
W.m2
TOTAL
700
DIFFUSE
600
500
400
300
200
100
0
35.0
30.0
25.0
degC
TEMP long
TEMP short
20.0
15.0
10.0
10/28/2010 10/28/2010 10/29/2010 10/29/2010 10/30/2010 10/30/2010 10/31/2010
12:00:00 AM 12:00:00 PM 12:00:00 AM 12:00:00 PM 12:00:00 AM 12:00:00 PM 12:00:00 AM
N.T.JEFFRIES Masters Thesis REBE 111 | P a g e
APPENDIX H – Steam production tests
N.T.JEFFRIES Masters Thesis REBE 112 | P a g e
APPENDIX J – Detailed calculation sheets for Continuous monitoring
Pumped volume cumulative and incremental (31 Oct)
250 6000
Total daily volume
15‐min volumes
200 5000
4000
150
3000
100
2000
50 1000
0 0
9:00 10:00 11:00 12:00 13:00 14:00 15:00 16:00
time
N.T.JEFFRIES Masters Thesis REBE 114 | P a g e
APPENDIX K – Overall system efficiency analysis
Pump Overall
Date Location duration Volume E hyd E sol efficiency
hrs litres J J/3 sq.m. %
28‐Oct Pump 4 5.4 2957 220279 51539161 0.43%
29‐Oct Pump 4 6.6 3763 280327 66434928 0.42%
30‐Oct Pump 4 6.7 4682 348842 67879338 0.51%
31‐Oct Pump 4 7.3 4955 369163 76560556 0.48%
1‐Nov Pump 4 6.2 4075 303560 70033992 0.43%
2‐Nov Pump 4 6.0 4235 315528 72102793 0.44%
3‐Nov Pump 4 5.9 3614 269260 66982533 0.40%
4‐Nov Pump 4 5.8 3986 296952 63963812 0.46%
5‐Nov Pump 4 6.2 4973 370466 64208527 0.58%
6‐Nov Pump 4 6.8 4712 351069 70828834 0.50%
7‐Nov Pump 4 7.1 4612 343572 70634166 0.49%
8‐Nov Pump 4 6.8 2793 208050 63577449 0.33%
9‐Nov Pump 4 4.5 3046 226931 53708262 0.42%
10‐Nov Pump 4 5.9 2915 217181 64891673 0.33%
11‐Nov Pump 4 6.6 3788 282171 66023865 0.43%
12‐Nov Pump 4 6.5 3603 268391 64726901 0.41%
13‐Nov Pump 4 5.7 2242 166999 52427054 0.32%
15‐Nov Pump 4 6.37 3348 249435 63475574 0.39%
17‐Nov Pump 1 6.1 2493 317888 64129562 0.50%
18‐Nov Pump 1 6.0 1949 248505 61946819 0.40%
19‐Nov Pump 1 6.0 1937 246930 68901597 0.36%
20‐Nov Pump 1 7.2 4047 516005 74256848 0.69%
21‐Nov Pump 1 6.65 2926 373042 61982396 0.60%
22‐Nov Pump 1 5.5 2519 321118 59482831 0.54%
23‐Nov Pump 1 5.5 2368 301870 56431077 0.53%
24‐Nov Pump 1 5.9 2306 294027 64869604 0.45%
25‐Nov Pump 1 5.3 1368 174415 32014724 0.54%
27‐Nov Pump 1 5.8 2173 277063 67893741 0.41%
28‐Nov Pump 1 5.6 2214 282232 67998446 0.42%
29‐Nov Pump 5 5.4 2722 204146 61600798 0.33%
30‐Nov Pump 3 4.7 2790 188301 65021054 0.29%
1‐Dec Pump 3 4.9 2902 195893 65650909 0.30%
2‐Dec Pump 3 4.9 3032 204631 70139264 0.29%
4‐Dec Pump 5 6.4 3448 258564 75499468 0.34%
5‐Dec Pump 5 4.4 1819 136428 68176213 0.20%
Average 0.43%
Maximum 0.69%
N.T.JEFFRIES Masters Thesis REBE 115 | P a g e
APPENDIX L – Kinetic energy losses during pumping
KINETIC ENERGY CALCULATION
Pump output
Flow l/s 0.227
6 hr day litres 4912
Pump pipe
pump pipe diam. cm 4.4
area cm2 15.2
length m 18.0
Water volume cm3 27356
Displacement per stroke
piston diameter cm 4.3
area cm2 14.5
displacement cm3 136.4
Crank speed
rpm rpm 100
time for up stroke sec 0.300
stroke length cm 9.4
mean upward speed m/s 0.313
Water velocity
mean m/s 0.150
peak m/s 0.299
Kinetic energy
Mass kg 27.4
Required energy Watt 1.225
Potential energy
lift head m 7.5
displacement per sec. cm3/s 227
Energy W 17.1
Total required energy W 18.3
Kinetic energy component % 6.7
N.T.JEFFRIES Masters Thesis REBE 116 | P a g e
APPENDIX N – Microfinance
N.T.JEFFRIES Masters Thesis REBE 118 | P a g e
APPENDIX P - Interview with Technicians - 13/11/2010
They were interviewed at Pump 6 on 13th November to learn from their experiences
including any modifications they have made to improve the system.
Collector
‐ Concrete mould, collector support/frame and boiler were all procured and
fabricated in Addis based on drawing sent from Holland.
‐ Large outside circular frame, originally L-section was changed to a 25mm
square box section
‐ Collector support material changed to 8mm concentric steel bar.
‐ Polymer glue from Holland augmented with locally bought silicon to reduce
overall amount of expensive imported polymer.
‐ Collector assembly approximately 7 hours over 3 days.
‐ Bottom hinge of collector frame (to control tilt) had 2 plates removed to allow
easier tightening. FTC frame has not been modified and has tendency to tip
over.
Tracker
Engine
N.T.JEFFRIES Masters Thesis REBE 119 | P a g e
‐ Experience so far has been that precision steel cutting and shaping, although
possible in Ethiopia is probably more expensive costs of making in Holland
and transporting.
‐ Local steel fabrication should be confined to milling, sawing and punching.
‐ Steam inlet position changed to bottom of buffer chamber to allow better
piping arrangement
‐ Steel frame which attaches engine to concrete support reduced in size and
cross bars removed to reduce material costs.
‐ Cross bars put back as new frame was determined to be not strong enough.
‐ Valve configuration slightly altered – pivot centres moved to reduce wear and
get better performance
‐ Roller bearings changed to wheel bearings (cost driven) and then back to
roller bearings as brass wearing down quickly.
‐ Springs for valve operations too tight on shaft. Tendency to stick when rusting
occurs. Lesson – either lubricate as part of regular maintenance; or change to
slightly larger spring.
‐ Greasing of cam mechanism needs to be balanced with advantages of
smooth-running but disadvantages in attracting dust.
‐ Light oiling inside piston casing of engine – reduces noise of engine. However
must be careful as oil may deteriorate rubber diaphragm.
‐ Original diaphragm a little small (4mm), which meant it slipped out of place
‐ New diaphragm produced that occasionally split on the edge, but this problem
has not recurred for last 4 weeks. Not sure of the cause of the splitting.
‐ Valve gaskets – 1st version 0.5mm tended to split, new 1.0mm gasket doesn’t
seem to have this problem. Lessons – (1) instruct people not to hand push the
valves as this is a greater movement than normal operation; (2) provide a few
extra gaskets in spare parts kits.
‐ Safety valve needs rethinking as never really effective.
Pump
‐ Original piston was hollow steel cylinder with brass valve – performed well in
workshop trials in Addis
‐ In sandy conditions it was presumed that there would be friction issues
between piston and pump cylinder – so replaced with perforated PVC piston
and piston ring
‐ High levels of wear has been experienced in some PVC rings (see main text)
‐ In very sandy conditions (e.g. Borchessa), a silicon piston cup used as seems
to tolerate sand better.
‐ Intention is to use this arrangement during initial well cleaning period and then
revert back to PVC piston.
‐ Further investigation needs to take place and other options explored as
nothing conclusive has been determined yet.
N.T.JEFFRIES Masters Thesis REBE 120 | P a g e
‐ Care needs to be employed in the positioning and fixing the relative position
of well head and rocker arm to reduce friction in packing gland.
‐ Stability of riser pipe-concrete pipe connection as well as packing gland seal
and pipe outlet arrangement all needs to be re-thought especially for
pressure discharge systems.
Recirculation system/condenser
‐ Pump stroke set so that bottom of piston stroke does not go beyond bottom of
cylinder at the highest engine block setting.
‐ Inlet valve comprising ball bearing sitting in rubber o-ring – works well.
‐ Outlet valve is small plastic cone attached to small spring. This is a little
fragile and sometimes becomes stuck open because of debris. Spring force
needs increasing from time to time by stretching out.
‐ Pump mechanism, a Teflon ring wrapped around a brass piston cylinder
prone to wear. Very little wear (0.1mm) is sufficient to lose the piston seal
resulting in loss of water from the cylinder top of pump.
‐ Condenser currently copper cylinder contained in pump outlet. The whole
mechanism need to be re-thought to make more robust and user friendly.
N.T.JEFFRIES Masters Thesis REBE 121 | P a g e
APPENDIX Q – Typical O & M tasks
Of course issues of self-maintenance also come with the risk of accidental damage
however this is inherent in piece of mechanical equipment. The instruction manual
should indicate the risks related to each maintenance task and clearly state which
mechanical activities should not be attempted out by the user.
General tasks:
• Manually pump to remove any potential grit between piston ring and pump
cylinder
N.T.JEFFRIES Masters Thesis REBE 122 | P a g e
APPENDIX R – Risk issues
The table below is an example of the format of a Risk Assessment. The content and
messages could be included as part of the training sessions as well as being
incorporated into manuals associated with purchases of the pump.
A number of the risks identified below will be designed out mechanical modifications
e.g. automated starting and better functioning pressure valve will reduce/eliminate
steam hose blow outs. The possible inclusion of a cage or barrier will reduce risks
associated with the flywheel
N.T.JEFFRIES Masters Thesis REBE 123 | P a g e