SolarSteamPump.Ethiopia

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

N.T.JEFFRIES Masters Thesis REBE 1 | P a g e

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.

Low-power vapour engines have always presented problems to designers as they are inherently inefficient and therefore expensive. Additionally there are certain mechanical problems that are exacerbated with decreasing engine size. A new pump has been designed that addresses these issues through the use of new materials and a different engine configuration. The device is powered by steam produced by a boiler that is heated by a parabolic dish solar collector. A steam engine acts as the prime mover converting thermal energy into mechanical power to drive a reciprocating positive displacement piston pump.

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.


ACKNOWLEDGEMENTS

In Ethiopia:

Bob Yoder - for his wisdom, enthusiasm and encouragement.

International Development Enterprises (IDE) – for providing financial and logistical support and many other resources

Alena, Binyam and Jim Kaufman – for keeping the pumps running despite the heat and dust.

Kedir Gemechu – a very diligent and reliable enumerator and his wife Zabiba for walking many kilometres every day to bring Kedir lunch. .

Ababa - for translating Oromo-Amharic-English

The farmers and their families - who endured the pump monitoring and interviews with great patience and friendliness.

In Europe:

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)

All photos are the author’s own unless specified


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


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.)


GLOSSARY

STP Solar thermal pump

R & W Rope and washer pump

NGO Non-governmental organisation

M & E Monitoring and evaluation

MIT Micro-irrigation technology

IDE International Development Enterprises

BG Busa Ganofa (local microfinance provider)

FTC Farmer training centre

EC Entrance condensation

Ehyd Hydraulic energy output

DPV Daily pumped volume (litres)

GWL Ground water level

TDC Top dead centre

BDC Bottom dead centre

K Kelvin

Lsteam Latent heat of vaporization for steam at 100oC

GHI Global (total) horizontal irradiance (W/m2)

DHI Diffuse horizontal irradiance (W/m2)

DNI Direct normal irradiance (W/m2)

Z Zenith angles

H Solar irradiation (J/m2)

ηsystem System efficiency

ETB Ethiopian Birr (£1 = ETB25)


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

Fig 1 Distribution of categories for solar pumping academic papers

Fig 2 Location of Ziway Town

Fig 3 Pump test location plan

Fig 4 Main components of pump system.

Fig 5 Cut through of steam engine

Fig 6 Cut through of outlet pipe/condenser tube

Fig 7 Energy flow through pump system

Fig 8 Schematic of steam engine cycle


Fig 9 Schematic showing drawdown at Pump 2 (3 – 9 Nov)

Fig 10 Effect of half-full boiler

Fig 11 Schematic of boiler and buffer

Fig.12 Hypothetical farm plan

Graph 1 Theoretical S-curve for rate of technology adoption

Graph 2 Cloud hours during study period based on output of SPN 1

Graph 3 Carnot efficiency v. temperature

Graph 4 Theoretical pressure variation in steam engine for 50% cut off

Graph 5 Average daily variation in DNI during monitoring period

Graph 6 Steam production test for 3m2 collector

Graph 7 Hydraulic energy output for continuous monitoring days

Graph 8 Hydraulic energy output under ideal operating conditions

Graph 9 Average daily engine speed v. average pump efficiency

Graph 10 Effect of varying buffer to engine volume on working pressure

Graph 11 Potential irrigated area for current and expected performance

Appendix A Development of the solar thermal pump

Appendix B Farm layouts

Appendix C Steam Data

Appendix D Effect of varying cut off

Appendix E Single pressure cycle in steam engine

Appendix F Typical output from SPN 1 pyranometer

Appendix G Graph of one-week pump drawdown at Pump 2

Appendix H Steam production tests

Appendix J Detailed calculation sheets for continuous monitoring

Appendix K Overall pump efficiency analysis

Appendix L Kinetic energy losses during pumping

Appendix M Graph of one-week R&W usage at Tadessa farm

Appendix N Microfinance – limits and conditions

Appendix P Interview with technicians

Appendix Q Typical O & M tasks

Appendix R Risk issues


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.

Research and development into this type of renewable energy technology is particularly important in the present day context of countries such as Ethiopia, as it offers a low-cost, low labour input, non-fossil fuel technology that allows poor smallholder farmers to:

• Improve food security during the dry season • Create extra income • Reduce labour and fossil fuel inputs • Adapt better to potential climate change effects

The promotion and dissemination of small scale micro-irrigation pumps (MIT) is a particularly effective development intervention as the technology allows millions of individual poor farmers to help themselves. By irrigating small plots of land during dry season, farmers can cultivate high value vegetables and fruit allowing extra income to be earned. This directly leads to an improvement in living standards through access to better health care, education and other modern services.

This catalysing approach to development is the core activity of International Development Enterprises (IDE), an organisation that provided key assistance for this thesis. In Ethiopia IDE’s work is supported by Practica Foundation, a Dutch technology NGO that researches and builds appropriate technology systems.

Box 1 provides a real life example of the transformative power of a simple irrigation pump for a poor farmer and his family.

Box 1 – Wealth creation through adoption of micro-irrigation technology

Tadele Tiko and his family used to be crop share labourers on large landholding giving 50% of what he produced to the landowner. He was persuaded in March 2009 to buy a R&W pump using microfinance. After just over a year he produced about 25, 000 birr (£1000) of cash crops. Now he has extended cultivation on his land to almost 0.25 ha through the purchase of a 6700birr engine pump (cash purchase) supplied by a large, hand excavated ‘Sudanese’ open well. He also cultivates crops in a rented 0.25 ha near to the local lake.


The importance of the study in a wider context is that the results could offer support for the future potential of a low cost alternative to fossil fuel pumping. As a literature review revealed, small scale vapour engines are not at all common. There are clear reasons for this: firstly fundamental thermodynamic limitations; secondly scale factors that exacerbate as engine size gets smaller. Therefore if it can proved that a small size engine can operate successfully, it will reduce the risk in R & D investments aimed at scaling up the technology to a size and power output, which can compete with fuel consuming, carbon emitting diesel pump sets.

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.

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.

Possible future – low cost, solar powered, zero emissions


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).


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.

1928 – Key theory – first mention of entrance condensation (eintrittskondensation) by German Guttermuth in 1928, but phenomenon generally ignores as effect is less pronounced in large 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.


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:

• C1 - Performance - can the STP pump deliver sufficient water to farmer’s fields reliably and consistently?

• C2 - Complexity/acceptability - can a typical farmer operate and manage the technology in a reasonably independent way?

• C3 - Comparability - does the pump offer benefit over the existing alternatives?

• C4 - Economics – can the pump provide enough income to be affordable within the local economic context?

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.

1.3 Additional aims (C5) One of the guiding axioms of the STP design is that simplicity and cost is more important than efficiency, echoing the principles set out in the VLOM approach to development3. This is an established maxim for the development of sustainable community water systems aimed at avoiding the creation of ‘white elephant’ projects that stop working as soon as specialists have left (FWR, 2010). It has led to a simple, low cost pump that avoids expensive materials, utilises components that are easy to repair and as far as possible can be produced by local, low tech manufacturing processes.

However it is also recognised that efficiency cannot be totally ignored, as a better-performing system can lead to economic benefits. This could be either through: (i) lower costs as fewer materials needed; or (ii) increased potential revenue as more water pumped.

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:

• Relevant system monitoring to identify components that are underperforming, to identify future improvements to the pump.

• Observations on pump usage and farmer response to understand the most effective method of incorporating into irrigation schemes as well as the support infrastructure required to achieve this.

3 Village Level Operation and Management


2. LITERATURE REVIEW An extensive literary review has been carried out to uncover any publications and books related to the technology and study objectives. The broad aim of the review was to provide a robust academic foundation for the thesis investigation as well as demonstrating that the study offers an original insight to the challenge of providing affordable solar irrigation pumping.

The literature review was divided into two separate but overlapping areas:

1. Technology – small sized solar thermal pumps 2. Analytical approach – how best to measure viability

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


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:

• Methods to improve the efficiency even marginally would go a long way to improving the economics


• 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.

Smith (2008) is taking a novel approach to solar water pumping, utilising an oscillating common and a liquid piston. However the technology is immature and only suits very low head applications.

A study carried out by Burney (2009) focused on a different solar technology but in a broadly similar context, the introductory words echo this:

“Promotion of irrigation, particularly small holder irrigation – is therefore frequently cited as a strategy for poverty reduction, climate adaptation and promotion of food security” (Burney, 2010 p.1848)

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).

2.2 Study approach This aspect of the literature review investigated the academic foundation for assessing technology viability. The most influential reference source for developing the assessment approach was Rogers (2003) in which the Theory of the Diffusion of Innovations is introduced. The theory attempts to explain the spread of new ideas, particularly why certain innovations are more successfully adopted than others. The foundation of the theory originates from his 1943 diffusion study on the adoption of high yielding, corn seed in Iowa. The researchers recorded data and observations on 259 farmers from 1928 – 1951 in which time all but 2 farmers adopted the seeds. 6 International Crops Research Institute for the Semi-Arid Tropics


From this and other studies ideas such as early adopters, opinion leaders, change agents and many of the concepts that underpin modern diffusion and marketing theory originated, including the now familiar S-shaped for technology adoption:

Graph 1 – Classic S-curve for rate of technology adoption

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.


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.


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 hyd ) requirement of 300, 000 J/day, or in another form: E hyd = 30 m3.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.

Based on this specification, a series of prototype solar collectors, steam boilers, steam engines and different types of pumps were built and tested. By 2009, a complete system was operational and test results confirmed that the concept worked as planned. The prototype was fabricated, tested for performance in the Netherlands and shipped to Ethiopia for durability testing. The results verified that the solar steam pump system could perform under field conditions but identified a number of necessary improvements that needed to be incorporated into the next version.

Photo: Early version of STP using flat plate collectors and pentane working fluid (Source: Practica)


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.

3.1 Factors that affect study The pump’s deployment and installation in Ethiopia was timed to coincide with the long dry season, during which solar energy and demand for irrigation is greatest. As well as allowing field testing, there was also a desire to demonstrate to funders some tangible process in the development of the technology to support future proposals. To meet the deadline required an accelerated deployment of the completed system into the field, this resulted in:

• Some components being designed and manufactured in haste • Certain parts did not undergo endurance testing before shipping • Pumps were installed quickly with limited training to farmers

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.


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.

Fig 2 – Location of Ziway Town

Map source: http://www.mazethiopia.com.et/eth_map.html

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.

Ziway


Graph 2 – Cloud hours during study period based on output of SPN 1 Pyranometer

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.

Typical soil profile taken at excavation at Pump 3

0

1

2

3

4

5

6

7

8

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

clou

dine

ss hou

rs


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.

3.3 Socio-economic context There is a large international rose growing operation to the south of town that provides employment for a large number of women in the area. Apart from this and the shops and small businesses in town, the remaining population rely on farming activities and small scale market trading to earn income.

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


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:

Table1 – Principal irrigation options around Ziway Town Economic circumstance Irrigation method Daily water volume Max irrigated area Poor economic status Rope and bucket,

unlined hand dug well < 1000litres <100 m2

Access to finance (GWL < 6m)

Treadle suction pump ~ 8000 litres 2000m2

Access to finance (GWL 5 – 18m)

Rope and washer pump

~ 4000 litres 1000m2

High economic status Membership of co‐operative

Engine pump >10000litres many hectares

Note: daily water volumes for manual irrigation methods based on 4hr pumping

3.4 Farm sites The pumps are installed in two clusters to the north and south of Ziway as indicated on Fig 3 - Location Plan. Northern cluster farms (Pumps 1 – 5) are easy to access being close to the main tarmac road. Southern cluster sites are more difficult as access is via a dirt road with a section that is seasonally flooded by an outflow from Lake Ziway, requiring a small boat crossing.

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.

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.


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.

The relationship between usage patterns and socio-economic profiles is discussed in detail as part of the assessment of the C2 - Complexity criterion.


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).

Fig 3 – Pump test location plan

Source: Base plan from Google Earth

5 km


Table 2a ‐ FARM PROFILES ‐ NORTHERN CLUSTER (EDOGOJOLA)

Pump #

Farmer Name Installed

Plot size (m2)

Irrigated Area (m2) GWL (m)

H/H number Livestock

Mobile phone Farmer Bio

1 Soado Abate

21/08 250 125 12.75 3 6 Y

Early adopter (Jan 2009) of R& W pump on IDE 50% subsidy scheme. In first few growing seasons was able to recoup investment and increase savings through significant sales of onion seedlings (30, 700 birr). As well smallholding, also rents another plot of land for additional cultivation. Very hard working.

2 Famer Training Centre

25/08 ‐ ‐ 6 ‐ ‐ ‐

Pump at FTC for demonstration, monitoring experiments and compatibility testing with other technologies e.g. raised header bags, drip irrigation.

3 Tadele Tiko 27/08 2750 1500 6.75 10 8 (estd) Y (2) Early adopter of R & W technology. Now successful model farmer. See Boxout 1 in main text for details.

4 Meketu Gurea

03/09 2500 300 7.45 7 22 N

Farmer is a widow of army soldier. Early adopter of R&W (IDE 50% subsidy scheme). Paid cash for pump. Her gender, marital status and proximity to the road, has made her a model farmer and a popular venue for demonstration technology as well as funder and stakeholder visits. However her h/h profile (made up of school children and elderly mother) mean that the availability of labour is the limiting factor.

5 Dereso Idao

09/09 3500 550 7.4 4 3 Y

Farmer used to be a crop share labourer and was initially resistant to the idea of buying a pump due to concerns about repayment of loan. Last year bought R & W pump with 2000 birr loan from BG and started cultivation on own land. Has already paid back 30% of loan and projected to pay back loan within scheduled period.


Table 2b ‐ FARM PROFILES ‐ SOUTHERN CLUSTER ‐ BORCHESSA

Pump #

Farmer Name Installed

Plot size (m2)

Irrigated Area (m2) GWL (m)

H/H number Livestock

Mobile phone Bio

6 Soado Duvay

11/10 6300 600 ‐ 5 5 N

Early adopter of R& W pump (IDE 50% subsidy scheme). Paid back loan quickly through cultivation and sale of onion seedlings. Plot of land includes large rain fed maize field. Proximity to lake has led to several nocturnal visits by hungry hippotamus. Household make up entirely women and children and therefore labour is limiting factor.

7 Medina Eba

15/10 500 250 6.35 6 1 N

Farmer is first wife of husband. No children but looks after 5 children of relatives. No significant irrigation before buying R & W pump in Feb 2010 for 1700 birr with loan of 2500birr (2950 birr with interest). So far paid back about 40% of loan from sales of peppers and kale.

8 Kufa Jeldo 20/10 700 400 5.15 9 6 Y

Part of Kebele management (village council). Bought R & W pump for 1600 birr about 8 months ago. So far paid back interest only on loan but expects to pay back on time based on revenue from home cultivation and additional parcel of land by lake (engine irrigated).

9 Chala Huluga

22/10 900 600 6 9 3 N

Farmer represents the ideal IDE customer – hard working, ambitious and open to new ideas. Bought R & W pump last year for 1700 birr with loan of 2500birr. Paid back over 25% through sales of peppers and onions. In first year of irrigation wore down metal support strut with nylon rope due to frequency of usage.


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:

• Careful selection of testing locations and individuals to ensure control of factors that may affect study e.g. soil conditions, family size, socio-economic factors

• Performance evaluation for multiple growing seasons in different locations that experience a broad range of solar conditions, water levels and soil conditions.

• Interval and daily volumes produced by all pumps over the course of a complete growing period

• Control data sets collected on main alternative methods allowing matched pair comparison to solar technology

• Economic assessment based on crop yields and market prices to determine relative benefits of different technologies.

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).


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.

System durability – deployment of pumps to Ethiopia was accelerated which led to the some system parts being designed hastily and not undergoing endurance testing. A good example is the feeder pump which recirculates water from the condenser back to the boiler. The pump suffered from continual problems related to the poor functioning of valves and rapid wear of parts (mainly the Teflon piston ring). This caused leakage from the top of the pump which meant that the boiler emptied quickly. The effect on the study was discontinuity in results as pumps waited to be repaired or parts replaced.

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


• 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.

Table 9 – Summary of study approach Study

component Data type Analysis Success indicator

1 Performance Measurements Calculate average daily volume produced based on collected data

Exceeds requirement for average landholding size

2 Complexity Measurements, observations and interviews

Collation of data. Formulation of opinion.

Independent, un‐assisted operation in more than 50% cases

3 Comparability Measurements, observations and interviews

Paired, cross‐sectional or matched pair comparison (see below)

Technology offers significant advantages over alternatives

4 Economics Plot sizes, cropping density, price analysis, interviews

Economic analysis of irrigated plot size versus prevailing market conditions

Potential revenue from pump satisfies typical pay back time and lending conditions

N.T.JEFFR

5. PUMThe purpsupport

• M• S• L

a• B

a

5.1 SteaThe comrecirculadetailed througho

Fig 4 – M

Each cowill presthe futurmanufac

RIES Masters T

MP THEOpose of thisthe experim

Mechanical System ope

imitations ond mechan

Basic analysnd experim

am pump amponents ofation system

and labelleout the doc

Main compon

omponent osent the relere this last fcturing stra

Thesis REBE

RY s section is mental appr

componentration - by dof the systenical propersis of the pr

mental resul

assembly f the steam

m. In this seed providingcument.

nents of pum

of the systemevant detailfactor will btegy and he

to provide roach and a

ts dividing one

em imposedrties of the erime moverts.

m pump are ection eachg a standar

mp system.

m is describls on mater

be an imporence produ

sufficient thanalysis, th

e engine cyd by thermoequipment.r to allow a

collector, eh of these crd nomencl

bed using vrial, dimensrtant considuction costs

heoretical ahis will inclu

ycle into disodynamic la. compariso

engine, pumomponentsature for co

visual aids asions and wderation in ds.

and technicde descript

stinct stageaws and the

n between

mp and a cos will be deconsistent re

and tables.where it wasdetermining

31 | P

cal details totions of:

es. e structural

theoretical

ondenser/ constructed

eferencing

The tabless made. In g a

a g e

o

d,

s


5.2 Collector Figure 4 shows the main parts of the collector; table 3 provides details on each of the parts: Table 3 – Collector components Component Material and other details Dimensions Local/

imported Dish 60 ‘leaves’ (thin triangular strips) of

treated aluminium (Al) arranged in a parabola, mounted on concentric 6mm steel hoops.

Φ = 2m, A = 3 m2 Al – Imp Frame ‐local

Boiler Steel cylindrical can, painted black, capped with 15mm concrete insulation.

Φ = 150mm, H = 150mm

Local

Steam hose Silicon hose. Internal hose Φ = 0.6cm, Wall thickness = 4mm, L = ~ 130cm

Imp

Frame Steel ‐ Local

Tracker PV Cell and shade. DC Motor. Nylon tracker cord

‐ Imp

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.

Fig 5 – Cut through of steam engine


Table 4 – Engine components Component Material and other details Dimensions Local/

importedBuffer chamber

9Polypropylene (PP) – maximum allowable temperature 130oC

V = 230cm3

Imp

Working space

PP V = 160cm3 Imp

Diaphragm Nitrobutyl (NBR) rubber ‘top hat’ diaphragm.

Eff. Φ = 10cm, eff. surface A = 80cm2, Stroke = 4cm. Displacement = 320cm3

Imp

Valves Spring action, PP/rubber seals ‐ Imp

Scottish yoke and cam mechanism

Laser cut steel ‐ Imp

Engine support

Reinforced concrete ‐ Local

Engine frame Mild steel Local

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.


Table 5 – Pump components Component Material and other details Dimensions Local/

importedPiston type 1 Perforated PVC piston cylinder, brass top

valve Φ = 43mm Imp

Piston type 2 Silicon diaphragm cup Φ = 43mm Imp

Piston rod Threaded steel rods. Φ = 8mm 3 m sections

Local

Foot valve Standard steel ‐

Pump cylinder Reamed PVC section Φ = 43mm (ID) Local

Pump casing Glued and sleeved PVC pipe sections 6m sections, Φ = 50mm (OD)

Local

Well casing Glued and sleeved PVC pipe sections 6m sections, Φ = 75mm (OD)

Local

5.5 Condenser and recirculation system This part of the system takes low temperature steam from the exhaust outlet, condenses the steam and returns it to the boiler. The condenser is integrated into the pump outlet to utilise the cooling capacity of the ground water. A feeder pump connected to the rocker arm provides the power for the recirculation. Fig 6 – Cut through of outlet pipe/condenser tube


Table 6 – Recirculation system components Component Material and other details Dimensions Local/

importedOutlet pipe PVC pipe Φ = 50mm (OD) Local

Condenser Copper pipe with openings for incoming steam and outgoing condensate

Φ = 20mm Imp

Recirculation pump Copper cylinder, brass piston, Teflon piston seals, o‐rings.

Φ = 13mm, A = 1.1cm2. Imp

Recirculation hose Silicon hose Small diameter (4x8mm)

Imp

The feeder pump is an important system component with two main purposes (i) return the condensate to the boiler; (ii) allow manual topping up of the boiler. The design of the pump is challenging because the required volume per stroke is so small - about 0.4cm3 (based on 100rpm, 40g/min steam). For practical construction and operational reasons it has an oversized capacity and a minimum of ‘dead space’ so each stroke pumps mainly air. This air/water pump needs to be designed more carefully to avoid air leaks.

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


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:

Table 7 – Stages during one cycle of engine Stage Stage description 1 Solar collector oriented towards sun using PV tracker and small electric motor.

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.JEFFR

5.7 LimiThe fundthermodtemperaThe equ

C

Graph 3the over

Graph 3

Note: ass

In small condenswalls of condensspace. Treachesengine talso neecausing accumuprovidedwalls to each cyc

RIES Masters T

itations of damental cdynamic Caature (ΔT muation that r

arnot efficie

3 shows thisrall efficienc

– Carnot eff

umes ideal en

steam engsation (EC)the working

ses when it The steam c saturation to move theeds to be hethe pressulated conded by the wadrop. The hcle.

Thesis REBE

steam engonstraint on

arnot limit. Tmeasured inrelates effic

ency %

s graphicallcy limit is eq

ficiency v. te

ngine, no fric

gines there . Entrance g space in tcomes into

continues ttemperatu

e piston. Aseated up. A

ure to drop tensate on t

alls of the cyheating and

gines n the perforThis is calcu K) betwee

ciency to tem

T2 – T1 /

y indicatingquivalent to

emperature

tion losses, p

is an additicondensatithe engine.o contact wo condensere. Before t

s the piston At the end oto atmosphhe cylinderylinder, whid cooling co

rmance of aulated by con the incommperature

/T2

g that for tho 7.4%.

perfect insulat

onal limitinion is the cy. When ste

with the relae and heat this point nomoves, ne

of the workheric. The pr walls to evich thereforontinues wi

all vapour eonsidering

ming (T2) adifference i

e design te

tion; exhaust

g factor calyclical heat

eam enters tively cool wup the wallo pressure

ew cylinder k stroke, theressure drovaporate. Tre causes thth associat

engines is tthe differennd exhaustis as follow

emperature

temperature

lled entrancting and coothe enginewalls of thels until the scan build uwall is expo

e exhaust vop causes t

The energy he temperated energy

37 | P

the nce in t steam (T1

ws:

of 130oC,

100oC

ce oling of the, it

e working steam up in the osed whichalve opensthe for this is

atures of thelosses at

a g e

).

e

h s

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)

5.8 Efficiency for system components A simple way of visualising the theory of the pump is as a series of conversion stages as energy flows through each stage of the system finally emerging in the form of hydraulic energy (pumped water). Figure 7 is a simplified schematic, as there are other minor losses; but it does offer a useful starting point for understanding the system.

Fig 7: Energy flow through pump system

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.


Table 8 - Theoretical conversion efficiencies Stage Efficiency Details

Collector 45% Estimated through steam tests in Netherlands.

Carnot 7.4% Assumes ideal engine, working pressure = 1.5bar, temp = 130oC

Engine 70% Attributable to entrance condensation and friction losses

Pump 70% Product of volumetric, friction and kinetic efficiencies

Total 1.6%

Source: (Yoder, 2010)

If one applies these figures to typical design conditions, it is possible to calculate the theoretical maximum daily pumped volume:

• Incoming solar irradiance = 850W/m2 • Collector size = 3m2 • Water depth = 7.5m

Therefore the expected maximum pump output should be:

• Q = (3000 x 0.016)/(7.5*10) = 0.54l/s or 11,750 l/day11 • Ehyd= 84m3.m

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.

5.9 Theory of Prime Mover The steam engine is the prime mover of the steam pump system. It is the mechanism through which the thermal energy of the sun is converted into mechanical energy available for water pumping.

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


Fig 8 – Schematic of steam engine cycle

BDC

Inlet valve closes

TDC Inlet valve opens

Exhaustvalve opens

PRESSURESTROKE

EXPANSIONSTROKE

Decreasing cut off

Increasing cut off

Cut off = 50%

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.

Note: work done = P x V, and is proportional12 to the area bounded by pressure change, total stroke (black dotted) and atmospheric pressure (red dotted).

This is a highly idealised representation but serves to illustrate the following features of the engine operation.

• High working pressure leads to more work done by engine. • Optimal cut-off determined by design working pressure. For maximum work

output, pressure line should intersect the atmospheric pressure line at the end of the stroke (e.g. 50% cut off appears optimal for working pressure = 2bar)

12 To work out actual work done multiply by diaphragm area.

N.T.JEFFR

Graph 4

Note:

1. Stroke 2. Atmosp

The ‘barvapour tin Appen

During tattemptsby insertthe rock

RIES Masters T

– Theoretica

length = 4cmpheric pressu

rk’ is the enthrough nonndix D.

he monitoris were madting sensor

ker arm to d

Thesis REBE

al pressure v

m, therefore anre varies with

nergy lost thn-optimal c

ing period, de to measurs in the bufdelineate ea

variation in s

nything to theh altitude. In Z

hrough not ut off. The

but outsideure preciseffer and wo

ach working

steam engin

e right of this vZiway (al = 16

extracting tissue of va

e of the scoly the press

orking spaceg cycle.

ne for 50% cu

value is not in600m), atmP =

the full exparying cut of

ope of this ssure change. A strain g

Expeanaperf

ut off

ncluded in the= 0.836bar

ansive eneff is discuss

study, a numes during egauge was

erimental rig ualyse the engineformance.

41 | P

e analysis.

rgy of the sed further

mber of each stroke attached to

used to e

a g e

e o


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:

Table 8a - Theoretical engine performance ENGINE unit totalPrinciple dimensions Effective diameter diaphragm cm 10.0Effective surface area cm2 80.0Stroke cm 4.0Inlet cut off % 50%Pressure stroke Full pressure stroke cm 2.0Pressure kg/cm2 1.5Work done on full pressure Nm 24.0Expansion stroke Average expansion pressure kg/cm2 0.5Stroke expansion cm 2.0Work done during expansion Nm 8.0Total work done per cycle Nm 32.0Speed rpm 100.0Cycle time sec 0.6Power output W 53.3

Note: 1. Cut off 50% is first guess at best value. Later experimentation may reveal more optimal value. 2. Engine speeds between 100 - 140 rpm represent a well-operating engine. Actual speed

determined by solar irradiance, fullness of boiler and position of engine. 3. Max temperature in engine limits allowable working pressure. Polypropylene may suffer structural

damage above 130oC. By converting engine power to hydraulic power we can estimate the corresponding potential daily water output:


Table 8b - Theoretical pump performance PUMP Engine input W 53.3Pump efficiency % 70%Hydraulic power input W 37.3Depth of water m 7.5Flow per second l/s 0.50Output per hour litres 1804.0Hours per day hrs 6.0Output 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:

Table 8c - Theoretical steam consumption STEAM CONSUMPTION unit totalDisplacement cm3 320.0Cut off 50%Displacement before cut off cm3 160.0Steam density at 1.5 bar cm3/gr 714.0Steam requirement per cycle gr 0.22Steam requirement per minute gr/min 22.4Entrance condensation, loss factor 1.5Real steam requirement gr/min 33.61Real steam requirement gr/cycle 0.34Latent heat of vaporization J/g 2260Steam power input W 1266

Note: Entrance condensation estimate assumes 50% more steam required 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:

Efficiency η = 53.3/1266 = 4.2%

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.


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

6.1 Solar irradiance (D1) Global and diffuse horizontal irradiance (GHI and DHI) was recorded at 1-min intervals using a Delta-T SPN-1 solar pyranometer and GP1 data logger. A typical graph output from the equipment is shown in Appendix F.

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

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


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:

DNI GHI – DHI /cos Z

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.

Graph 5 – Average daily variation in DNI during monitoring period

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

400

500

600

700

800

900

1000

W/sq.m. Average DNI Ziway 28 Oct ‐ 5 Dec


• DNI = 853 W/sq.m • Solar Irradiation (energy), H = 26 000kJ/m2 or 7.2 kWh/ m2.

6.2 Pump performance data The key engineering quantity recorded in the course of the field testing was daily pumped volume (DPV). This allows the calculation of important indicators such as:

• System efficiency – how well the pump is performing • Potential cultivated area – which enables the potential revenue that can be

generated by the pump to be calculated

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:

1. Continuous monitoring (D1) - hour by hour monitoring of key installations 2. Independent Usage (D2) – from rotating farm visits

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.


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 pumped

Indication of overall volume over a long period.

DST Distance elapsed Volume pumped since last resent

Independent usage since last visit.

TM Time elapsed Pump duration since last reset

How long the pump has been running since last visit.

AV Average speed Average rpm since last reset

Average engine speed since last visit

MAX Maximum speed Max rpm since last reset

Limited use ‐ but high rpm indicates 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.

6.2.1 Continuous monitoring (D2) For the continuous monitoring, an enumerator was employed from the local technical college to keep records for complete pumping days over the duration of several weeks’ operation. The scope of the enumerator’s work was as follows:

• Fill boiler tank at beginning of day; top-up if necessary during day • Start pump when pressure allows. • Run pump as long as solar power allows • Re-start pump if stoppages occur • Maintain a pump logbook that includes:

o Weather description including wind direction, cloud cover and any changes in the day.

o Start and end time o DST, TM, RPM and pressure meter14 every 15 min o 20 litre bucket test every half hour15 o Time, duration and reason for any stoppages o Total daily TM, DST and average RPM o Log any stoppages or issues during day’s operation o Reset computer at end of day

The enumerator’s activities can be understood in the context of the daily calculation sheets, for which an extract has been included in Table 12. The highlighted boxes are measurements recorded by the enumerator:

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.


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

DISTANCE CYCLES Vcalc Efficiency Stroke 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 top section of the table shows:

• Pump characteristics – for calculation of maximum vol/stroke = 0.136l/s • Time start/end – duration of pumping day, required to relate volume data to

solar irradiance. Not necessarily equal to total pumping hours, as there are stops and starts during the day.

• Cycle circumference (2m) – set during computer calibration. It is used to convert DST into revolutions/strokes.

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)

The bottom section of the table is a 15-min interval pumping log:

• TM is pump operating hours – gives an indication of daily pumping hours giving (or pump downtime)

• DST - distance elapsed since last record allowing calculation of 15-min interval volumes

• Speed (km/h) – can be converted easily into RPM giving an instantaneous indication of how well the pump is running at that time.


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:

Table 13a - Extract from Independent Usage calculation sheet

Pump 9 17‐Nov

Dia 43 mm Stroke 11.6 cm Setting 3 BUCKET TEST Volume 20 ltr

DISTANCE CYCLES Vcalc Stroke 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:

Table13b – Independent usage calculation

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


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.

GWLs were measured in two ways:

• 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.

16 The two measurements agreed within +/- 5mm

HOBO water level logger for measuring well drawdown during pumping. The device records pressure changes which can be converted into depths.


Fig 9 – Schematic showing drawdown at Pump 2 (3 – 9 Nov 2010)

Atm P = 83.6kPa

GWL = - 6.25m

GL = - 0.0m

Pump casing = -17.95m

Pressure sensor = -18.5m

Well casing = -22 m (est)

Static P = 203.5kPa

Pump rate = 0.2l/s

Drawdown P = 198.1kPaDrawdown = 0.54m

HOBO

The tests indicate a potential 540mm drawdown in water levels during pumping. This will need to be considered in the later analysis.

6.4 Steam tests (D5) Steam tests are a simple way of estimating the efficiency of the collector and investigating what factors contribute to optimal collector performance. The set up for the test was very simple: the boiler was allowed to start producing steam and then the steam hose was inserted into the spout of a watering can sitting on an electronic scale. After a short while, once steam production has stabilised, the balance was zeroed and 1-min mass readings were taken. The experiment was repeated for consistency and with different variables.

Steam test experimental set‐up


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:

Graph 6 – Steam production test for 3m2 collector

Key findings:

• For an average DNI = 870W/sqm: o Steam production = 37g/min o Collector efficiency = 53%

• Cleaning the collector improves the steam production • The fullness of the boiler has a significant effect on steam production

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.

y = 0.0235x + 0.0109

y = 0.0456x + 0.0893

y = 0.0335x + 0.0075

y = 0.0369x + 0.0427

0

0.1

0.2

0.3

0.4

0.5

0.6

0 2 4 6 8 10 12 14

steam (grams)

time (mins)

Steam production tests ‐ 1st Dec 2010, 14:00 ‐ 15:00

boiler not full

boiler full

dirty collector

clean collector

in focus


The results of the steam tests are referred to later when considering the overall efficiency of the system.

Fig 10. – Effect of boiler being partially full

6.5 Measurements from other irrigation methods (D6) Data was collected on other pumping methods to allow a comparison between technologies. For the physical context of the test farms (growing area = 100 – 1000m2, GWL = 5 – 15m), the main alternatives are rope and washer (R & W) pumps and hand drawn open wells. Treadle pumps are suction and therefore cannot be used below 6m. Diesel engine pumps are used on much larger fields.

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:

• 20 litre bucket filling time by a mix of people (gender, ages) at a number of different locations

• Motion data from a HOBO Accelerometer attached to the wheel spoke of the R& W pump. The tests were carried out at different locations and under different circumstances. The devices collected 10-sec data over a one-week period allowing the time (of day) and duration of pumping to be determined.

Hand drawn open wells

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


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:

• Irrigation amount and frequency – 15 ltrs/row every 3 days • Total volume required – 1050 ltrs • Time to fill 15 litre watering can – 2min 40 secs • Total continuous lifting and filling time - > 3hrs (not incl. distribution) • Harvest – 1 kuntal/2weeks • Potential revenue 50 – 70birr

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.

6.6 Socio-economic data (D7) The main socio-economic data collected were information on the physical and socio-economic context of each farm site. The aim was to create profiles that can be referred to when comparing usage patterns and other success indicators for the technology. For example, if certain pump sites are used more frequently and effectively than others, profile data can be compared to investigate patterns or similarities between the users:

The main characteristics used to create these profiles were:

Rope and washer pump at test site 7. The low cost manual technology is the closest comparable technology to the STPs.


• Proximity to a main road • Household size and make up • Number of livestock • Mobile phone ownership • Plot size

6.7 Local cultivation practices and agronomic context (D8) This data informs the C3 Comparability and C4 Economics components of the study. Broadly it allows an understanding of the value of the irrigation water, the basis of choosing particular crops and where energy and resources for irrigation activities are likely to be directed.

The data that has been collected from observations and discussions with farmers, IDE field staff and market traders and includes:

• Crop water requirements • Frequency and volume of watering activities • Crop types, density, time to harvest etc. • Market value of different crops • Types and details on other inputs e.g. distribution systems, seeds fertilisers,

pesticides, insecticides etc. • Microfinance - limits and conditions on local credit facilities

This data is not presented explicitly in one place; rather it has been integrated into several aspects of the overall analysis.


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)

7.1.1 Continuous monitoring results (D2) The locations for continuous monitoring were selected based on accessibility and ground water depth. The main sites represented the typical and maximum water depths encountered in the study area. The monitoring durations were as follows:

• Pump 4: GW depth = 7.5m; duration = 21 days (28 Oct – 15 Nov) • Pump 1 : GW depth = 13m; duration = 12 days (17 Nov – 28 Nov)

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 4 DST

Pump 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 22.529‐Oct 69.2 80% 0.109 0.157 3763 28.230‐Oct 78.2 88% 0.120 0.194 4695 35.231‐Oct 83.4 86% 0.118 0.189 4918 36.91‐Nov 69.7 84% 0.115 0.190 4002 30.02‐Nov 70.4 88% 0.121 0.198 4241 31.83‐Nov 62.0 84% 0.115 0.172 3557 26.74‐Nov 67.4 83% 0.113 0.210 3799 28.55‐Nov 77.3 93% 0.127 0.230 4911 36.86‐Nov 78.0 88% 0.120 0.199 4685 35.17‐Nov 78.7 86% 0.117 0.194 4618 34.68‐Nov 50.9 80% 0.110 0.152 2792 20.99‐Nov 50.8 88% 0.120 0.200 3059 22.910‐Nov 57.0 69% 0.094 0.143 2677 20.111‐Nov 68.7 80% 0.110 0.170 3765 28.212‐Nov 66.3 79% 0.108 0.160 3597 27.013‐Nov 46.7 70% 0.096 0.158 2233 16.715‐Nov 64.9 76% 0.103 0.147 3343 25.1Average 83% 0.113 0.179 3758 28.2Maximum 93% 0.127 0.230 4918 36.9

Note: Full cloud cover on 14th Nov


Table 15 – Pump 1 - Continuous monitoring (D2) summary

PUMP 1 DST Efficiency Vol factor Flow Daily volume Hyd. energy

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

7.1.2 Performance analysis against required specifications If we combine both sets of results (graph 7), there is no immediate reason for optimism as a large proportion of testing days fall below the target specification for hydraulic energy output of 30m3.m.

Graph 7 – Hydraulic energy output for continuous monitoring days

Note: Significant cloud cover on 25th and 26th Nov

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

10

15

20

25

30

35

40

45

50

55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Ehyd (m3.m)

day

Daily hydraulic energy output ‐ all days

Pump 4 Pump 1 Target specification


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:

Graph 8 – Hydraulic energy output under ideal operating conditions

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:

20

25

30

35

40

45

50

55

1 2 3 4 5 6 7 8 9 10 11

Ehyd (m3.m)

Test

Daily hydraulic energy output ‐ non‐ideal days removed

Pump output Target specification Average


• 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.

Graph 9 – Average daily engine speed v. average pump efficiency

This can be explained as a slower pumping rate allows more time for leakage between the piston seals and valves (slippage).

KEY FINDINGS

• Average daily pumped volume (DPV): o 7.5m depth = 3750 ltrs o 13m depth = 2500 ltrs

• Maximum DPV: o 7.5m depth = 5000 ltrs o 13m depth = 4000 ltrs

• On clear days no mechanical issues, DPV exceeds target specification by average of 25%.

• Best day’s pump performance exceeded the specification by 75%. • High rpm increases the pump volumetric efficiency.

R² = 0.7311

60

65

70

75

80

85

90

95

100

105

110

60% 65% 70% 75% 80% 85% 90% 95%

Average RPM

Volumetric efficiency

RPM v. Pump efficiency at Pump 4


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:

• Average ηsystem = 0.43% • Maximum ηsystem = 0.69%

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%.

Pump efficiency – includes 3 elements:

1. Volumetric – measured in previous section. Average = ~ 83% 2. Kinetic – in a displacement pump the column of water (mass = m) is

accelerated up the well casing, to a peak velocity = v. For this acceleration there is an associated kinetic energy (KE) requirement = 0.5 x mv2. The KE calculation is included in Appendix K. Result = ~ 7% reduction.

3. Friction – between piston rod and packing gland, rocker arm hinge etc. This can only be estimated, say - 10% reduction.

This gives a total pump efficiency = ~ 70% which agrees with expectations.


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 pressure meter informs on two levels:

• The pressure reading itself, Pa which rarely exceeds 1 bar. • The flicking of the pressure needle during each cycle which drops to a lower

pressure (Pw) each time the inlet valve is opened

The interpretation and implication of the pressure data can be explained as follows:

Pressure is proportional to temperature (see Appendix C). For Pa = 1 bar, the corresponding steam temperature (T2) = 120oC, reducing Carnot to:

Carnot %= (393-373)/393 = 5.1%

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.

Fig 11. Schematic of boiler and buffer

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.

230cc160cc 2700cc

Workingspace

X

Bufferspace

PaPb

Boiler

Steam hose

Inletvalve

Pressure drops fromPa to Pb as inlet valve opens


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.

The quantities are related by the equation:

Pw Vb / Vb Vw x Pw

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%)

Graph 10: Effect of varying buffer to engine volume on working pressure

Note assumes: Pa = 2.5bar (absolute pressure)

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

0.0

0.5

1.0

1.5

2.0

2.5

0 1 2 3 4 5 6 7 8 9 10 11Volume ratio

Effect of buffer:working space volume ratio on working pressure


Key findings:

• Average system efficiency = 0.73%, maximum = 0.69% • This is much lower than expected value which is attributable to the

underperformance of the engine • Maintenance of a high working pressure in the working space is critical. • Current small buffer volume causes unacceptably large drop in working

pressure. • Several other factors need to be addressed to improve the operational

performance of the engine.


7.1.4 Potential errors The following have been identified as possible sources of error in the analysis of the pump performance:

• Well drawdown – during tests at Pump 2 drawdown was measured as 540mm at a flow rate of 0.2l/s. This means that pumps are operating at a higher discharge head, meaning that pumps are performing slightly higher than that indicated in the analysis.

• Feeder pump losses: a proportion of the system power is required to pump the condensate.

This can be estimated as follows:

o Piston Area = 1.1cm2, Stroke = 2cm

o System P = 1 bar (100,000Pa) o F = P x A = 11N o Work done = F x D = 11 x 0.02 = 0.22J o At 60rpm = 0.2W; at 100 rpm = 0.33W

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%.

• Engine diaphragm shape – assumes cylindrical but actually a truncated cone (see fig 4). As material is rubber, there may be heat and elasticity effects that distort shape. This may impact analysis on engine performance as volume may be over or underestimated.

• Pump piston diameter – there were variations between pistons due to manufacturing differences and material wear. For ηvol calculations a design value of 43mm was used; this sometimes yielded results greater than 100%. Although this is not impossible19 it is unlikely to be happening here. The error is about 5% for every 1mmm that the true diameter deviates from 43mm.

• 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


• 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)

7.2 Complexity/acceptance (C2) Complexity is a measure of how easily farmers can understand and operate the system. The assessment method uses measurements collected during rotating visits to the farms. The key data set is D3 Independent Usage. The evolving farm layout plans presented in Appendix 2 also give a good visual representation of the different responses by individual farmers to the solar pumps.

The assessment of complexity highlights certain conflicts embodied in the study:

• 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.

Table 16 – Summary of Independent Usage Pump Dates DST Total Vol Daily Vol

km litres l/day

1 2 ‐ 10 Nov 169 7593 950 1 11 ‐ 16 Nov 181 8157 1359 4 21 ‐ 29 Nov 79 4355 543 5 14 ‐ 21 Oct 115 6574 940 5 27 ‐ 31 Oct 64 3645 911 5 1 ‐ 17 Nov 304 17388 1022 5 18 ‐ 23 Nov 92 5251 525 8 23 ‐ 29 Nov 48 2720 495 9 18 ‐ 23 Nov 119 11108 2020

9 23 ‐ 29 Nov 145 7890 1435

20 8 pumps x 45 days = 360 days less 10 cloudy days, 33 x D2 days = ~ 250


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:

1. Land configured to receive irrigation water – at pump 6 it took a long time to prepare land for irrigation and procure/grow seedlings (see photo)

2. Crop types need to suit solar irrigation – if commercial seedling are being cultivated, farmers prefer to irrigate early morning and night with R&W pump.

3. Solar irrigation needs to offer the most attractive returns in terms of time invested. At pump 3 the farmer’s time was mainly consumed with his new engine pump that was projected to yield large returns.

4. Farmer needs to be motivated and have enough family labour to engage in the irrigation work (e.g. damming and opening furrows)

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:

• Large motivation and work ethic of farmer and family21. • Main crop of onions suits furrow irrigation during the day22. • No livestock or small children to tend.

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.

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.


• 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:

• Over 50% of farmers demonstrated independent usage of pump • Some farmers pumped over 2000l/daywithout assistance • Several factors increase likelihood of independent usage

23 The evidence is a hole in the metal pump frame worn down by the friction of the nylon rope


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.

The obvious advantage of a non-manual pumping method is that it frees up time to allow other productive activities to take place. Economists quantify the benefits using the term opportunity costs. Manual pumping is generally carried out using a watering can which is used for direct application to crops. Hence there an STP offers potential for time savings both in filling the can as well as distributing the water.

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:

Table 17 – Estimate of time saved by solar pumping Pump location Water depth Bucket fill

(20ltr) Max volume

(l/day) Hours saved

4 7.5m 60 5000 4.2 hrs 1 13m 100 4000 5.5 hrs

This is a significant saving and does not include time taken for water distribution. To determine how long farmers currently devote to pumping activities, measurements were collected at different sites over 1-week periods. The results in table 18 were recorded at a farm near to Pump 1 that had approximately 800m2 under cultivation. The farm relies solely on manual pumping to irrigate crops. Data was collected between 10 – 17 Nov using an accelerometer, a device which checks and logs motion at 10-sec intervals.

Table 18 – Typical duration and volumes for manual pumping Date in motion time equivalent volume equivalent

10‐sec intervals hrs ltrs

11‐Nov 951 2.64 2377.5 12‐Nov 1262 3.51 3155 13‐Nov 1382 3.84 3455 14‐Nov 556 1.54 1390 15‐Nov 883 2.45 2207.5 16‐Nov 894 2.48 2235

17‐Nov 1213 3.37 3032.5

Note: 14 Nov was Sunday hence reduction in pumping activity


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.

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.

This is supported by observations at Pump 4 and Pump 5 at the end of the monitoring period. At this time the farmers were contracted to grow 40+ ‘merips’ of

24 Estimate based on 63 ‘merips’ of onions seedlings each with a daily water requirement of 50 litres.

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.


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

• Large potential time and energy saving • Labour requirement not totally eliminated as farmer presence required on site • Investment in efficient distribution system required to fully benefit from STP • Convenience of usage could be more important than saving • Storage, preferably elevated, key to increasing STP potential

25 This assumes that for STPs farmers use the water differently than with manual pumping (as observed)


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


Fig.12. Hypothetical farm to estimate revenue potential

ONION BEDS(100 sq.m.)

ONION BEDS(100 sq.m.)

PEPPERS(60 sq.m.)

CABBAGES(60 sq.m.)

SEEDLINGS(100 sq.m.)

600 sq.m plot

onions seedlingssold or transplantedto onion beds at 1:30ratio

hose distributionsystem to minimiseinfiltration losses


Table 19a – Estimated revenue for hypothetical farm

REVENUE

Crop Irrigated area Daily waterTime to harvest Crop density Productivity Revenue per unit Revenue

(m2) (litres) plants/m2

Onions 200 1000 4.5 months 45 1 kg/10 plants 6 ‐ 15 birr/kg ETB 5,400.00

Onions seedlings 100 500 45 days 1 merip/ 10sqm.

2 crops/dry season 150 ‐ 260 birr ETB 3,600.00

Cabbages 60 300 Continual 2 2‐4 kuntal/2

weeks 30 birr/kuntal ETB 1,040.00

Peppers 60 300 3 ‐ 5 months 4 5 kilos/plant 6 birr/kg ETB 7,200.00

TOTAL 420 2100 Gross income ETB 17,240.00

Note: 1. Information based on interviews and observations at project farms. 2. Type and proportion of crops based on typical farms using irrigation 3. Onion seedling can be either sold for a high return, or transplanted at an area ratio of 1:30 to onion beds. 4. Onion density and yield based on drip feed layout i.e. 30cm between 10 plant cluster 5. Cabbage density and yield from interview with farmers 6. Pepper plants yield 3 harvests per plant. Density and costs from Meketu and Dereso farms. Dry peppers 25 birr/kg or more 7. Daily water based on 5l/sq.m/day 8. Kuntal is a big plastic sack (holds 100kg onions)


Table 19b – Estimated input costs for hypothetical farm

COSTS Description Unit Unit price No Total

Seeds Imported indian seeds tin ETB 150 12.00 ETB 1,800

Fungicide Rediomil Gold kg ETB 250 1.00 ETB 250

Insecticide Selecron 0.5ltr ETB 290 1.00 ETB 290

Distribution system Drum and hose complete ETB 750 1.00 ETB 750

Labour n/a ETB 0

Transport to market Donkey cart journey ETB 5 30.00 ETB 150

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


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:

• Enhance the current system to respond to the cautious psychology of potential buyers

• Market and sell the device where there is less aversion to risk.

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.

8.1 Improvements to mechanical system The pump system as currently installed in the field, achieves a max solar to water efficiency of about 0.7%. For a 15m well, this equates to about 3000litres/day which, assuming some distribution losses (~25%), allows the cultivation of about 450m2. However the results indicate that farmers irrigate twice this area using only manual methods. Therefore the relative advantage may not be strong enough to provoke a switch to the more expensive alternative.

28 Only 327 R&W pumps have been sold in the Ziway area in 3 years, despite the overwhelming evidence that such devices can transform fortunes. Source: Internal data from IDE Ethiopia, provided via email by Sophia Musa on Fri 14 2011


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

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:

Table 20 – Present and theoretical system efficiencies Efficiencies Present Expected

Collector 53.0% 60.0% Carnot 2.5% 7.4% Engine 60.0% 70.0%

Pump 70.0% 80.0%

Overall 0.6% 2.5%

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

0

2000

4000

6000

5 10 15 20

Area(m2)

Water depth (m)

Present

Expected

Field size currently beingirrigated by engine pump GWL = 6.75m


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.

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:

• Improve collector efficiency o Improve the accuracy of the collector focus o Modified boiler placement and shape and improve insulation

• Increasing Carnot cycle efficiency o Shorter and better insulated steam line to reduce heat loss o Minimizing pressure drop between boiler and engine

• Increasing engine efficiency o Minimize condensed water in steam pipe and engine entrance o Optimize cut off

• Increasing water pump efficiency, o Improve piston seal o More effective pump rod seal for pressure applications

• Improved condenser system and feeder pump

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.


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.

In practical terms this could entail:

• 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

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:

• Self-start mechanism – to eliminate need for manual starting • Redesigned pressure release valve – to minimise steam hose blowouts • Fold-away handle – allows pumps to be used in manual mode when the sun

is not shining • More reliable tracking – improve current system which has problems with the

tracker cord slipping on the motor shaft

Flatter collector mounted on the engine will improve the overall efficiency of system.

Photo source: Practica


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.

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).

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


8.3 Improved training and technical support The experience of personally operating the pumps as well as extensive observations of farmers revealed a whole spectrum of different response to the technology. As the analysis on system complexity demonstrated, there were a lot of positive examples of consistent, independent usage. However there were also many occasions when farmers could not start the pumps, for reasons that transpired to be very simple and fixable.

Some reasons that can be offered for the operational difficulties can be attributed to:

• Inadequate and limited training • Original recipient of training did not pass information to other operators • No instruction/operations manual • Mechanical issues with the current prototype.

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:

• Standard start-up and shut down procedures • Safe method of operation (based on risk assessment) • Commonly encountered problems with corresponding solutions29

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.

Experimental low cost elevated storage structure being tested at FTC. Photo source: Bob Yoder, IDE


• Schedule of basic maintenance checks to ensure good operation • Phone number/details of local service provider

Of course, with self-maintenance, comes the risk of accidental damage. However this would be a factor in any purchase of mechanical equipment. The instruction manual should point out the risks related to each maintenance task and clearly state any mechanical activities that should not be carried out by the user.

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.


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:

• Performance and Economics – both quantifiable and objective concepts • Comparability and Complexity –more qualitative and subjective entities.

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.

As the study progressed it became apparent that other considerations, contained within the additional aims of the study, were more important and useful:

• What modifications need to be made to the current system to make it more reliable and attractive to potential buyers?

• What set of conditions will lead to successful diffusion of the technology?

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.


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.


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.


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.


APPENDICES


APPENDIX A

Development of the solar thermal pump

In the Paris Exposition of 1874, Augustine Mouchot demonstrated a machine that could pump hundreds of gallons of water each hour. By the turn of the century, inventor and solar visionary, Frank Shuman had designed a machine that pumped steadily through the summer of 1907/8 and even in cold Pennsylvania winters: “with snow piled up around the collectors”. (Butti and Perlin 1980).

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


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.

Low-powered thermodynamic systems are already inherently inefficient because of the Carnot limit. Entrance condensation provides the explanation that even in well-designed small steam pumps, where taking into account Carnot and other system losses there should still be efficiencies of 3-4 %, steam-water conversion efficiencies greater than 1% are rarely exceeded (Bom 1993).

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.

Universal peristaltic engines

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.JEFFR

rotary vaentranceengine mseparatifrequenccondenspotentiaand testsufferedlong eno

Rolling

After thisattentiondrive theone endpump wwas neenitro-butunforeseslightly ethe diap

Flat pla

Having ethought the force

RIES Masters T

ane, turbinee condensamay provideon betweency – seemesation avoidl to becometed over a 5d durability iough to be w

diaphragm

s lengthy dn was againe Volanta pd to a crank ith the othe

eded to trantyl rubber (Neen problemexpanding thragm to fa

te plus pen

encountereoccurred th

e. This led

Thesis REBE

es) to deteration. Aftere the best sn hot and ced to best odance. At te a new typ5 year perioissues in thworthwhile

m pistons

eviation, amn focused opump was th

on the pumer end. For nsfer the poNBR) seemm arose – tthe rubber ail rapidly.

ntane

ed this probhat it shouldthe design

rmine whichr some delibsolution to tcold cycles, obey the nethe time it wpe of univerod – all of thhat it was no.

mbitions abon solarisinghought to bmp shaft anthis it was t

ower efficienmed to satis

he NBR reacausing sm

lem, the ned be possiber to start l

h configuratberation it wthis problemno valves wly discove

was also thorsal enginehem workedot possible

bout universg the Volan

be a verticalnd connectethought thantly. A rolli

sfy this requacted to the

mall creases

eed for a lonble to shorteooking at s

diaphragmaterial was moddiaphragcollector which waconversioconsideracarried oover a onwere reasmall expdesignerHoweverout in Bu

tion would bwas decidedm. The rubband a reasoered rules oought that i. A numberd for a whilto get the w

sal enginesnta pump. Tl oscillatinged to the coat a relativeng diaphra

uirement. Oe hydrocarbs in the diap

ng stroke wen the stroksmaller-stroms that woexpansion

dified to accm and attawith a pent

as the curreon technoloation. End

out in Hollanne month p

asonably proplosion thatr’s garage inr subsequeurkina Faso

be least affd that a perber construonably highof entrancet may even

r were desige but in theworking tub

s were put oThe simples cylinder co

oncrete basly long strogm piston m

Once again abon workingphragm wh

was questioke length an

oke ‘top hat’ould not be issues. Th

commodateched to a fltane workin

ent solar poogy under durance tesnd 24hour period and tomising, det destroyedn Papendrent field test did not hav

89 | P

fected by ristaltic ction,

h operating

n have the gned, built e end they be to last

on hold andst way to oupled on se of the oke(20cm) made of an g fluid by ich caused

ned – the nd increase’ so prone to

he engine e the new lat plate ng fluid ower

sting was pumping he results

espite a d the echt. ting carried ve the sam

a g e

d

d

e

o

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.

Evacuated tube plus water- 2003

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.

Concentrating collectors- 2008

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.

Parabolic collectors - 2009

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


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


APPENDIX B – Farm Layouts


APPENDIX C – Steam Data

Pressure Temp Specific volume Steam enthalpy kg/cm2 oC cm3/gr kJ/kg

0.1 102.7 1533 2680 0.2 105.1 1414 26840.3 107.4 1312 2687

0.4 109.6 1225 2691 0.5 111.6 1149 2694 0.6 113.6 1088 2697 0.7 115.4 1024 2699 0.8 117.1 971 2702 0.9 118.8 923 2705 1 120.4 881 2707

1.1 122 841 2709 1.2 123.5 806 2712 1.3 124.9 773 2713 1.4 126.3 743 2715 1.5 127.6 714 2717 1.6 128.9 689 2719 1.7 130.1 665 2721 1.8 131.3 643 2722 1.9 132.5 622 2724 2 133.7 603 2726

Data source: http://www.simetric.co.uk/si_steam.htm

100

105

110

115

120

125

130

135

140

400

600

800

1000

1200

1400

1600

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Pressure (bar x 10)

Temp (oC)Specific volume

(m3/kg)Steam characteristics

Specific volume Temperature

N.T.JEFFR

APPEND

This sugwill mak

RIES Masters T

DIX D – Eff

ggests that ke better us

Thesis REBE

fect of vary

for a workie of the exp

ying cut of

ng pressurepansive en

ff

e of 2.5bar ergy of the

(absolute),steam.

, a lower cu

108 | P

ut off (40%)

a g e

)

N

A

N.T.JEFFRIES Mas

APPENDIX E

0.6

0.8

1.0

1.2

1.4

1.6

1.8

6.00

AbsolutePressure (bar)

sters Thesis REBE

– Single Pres

7.00

Inlet opens

ssure cycle in

8.00 9.00

Rocker arm di

Steam Engin

10.00

splacement

Inlet closes

ne

11.00 1

Buffer

P at

P a

BDC

109 | P a

12.00 13.00

Work

tm (measured)

atm (chart)

a g e

0 14.00

king

15.00 1

Stroke

Inlet opens

16.00

0

20

40

60

e (cm)

Displ'mnt(mm)

degC

W.m

2

TEMP longTEMP short

TOTALDIFFUSE

Appendix F SPN1 Ouptut Data .dt6

DeltaLINK-PC Copyright (C) 2005 - 2007 Delta-T Devices

10/28/201012:00:00 AM

10/28/201012:00:00 PM

10/29/201012:00:00 AM

10/29/201012:00:00 PM

10/30/201012:00:00 AM

10/30/201012:00:00 PM

10/31/201012:00:00 AM

degC

W.m

2

10.0

15.0

20.0

25.0

30.0

35.0

0

100

200

300

400

500

600

700

800

900

1000

1100

1200

1300

1400


APPENDIX G – Graph of one-week pump drawdown at Pump 2 (data: HOBO Pressuremeter)


APPENDIX H – Steam production tests

Dec 3rd 2010 Test 1 Test 2

Test 3 ‐Dirty

Test 4 ‐ Clean

Solar Irradiance

Pump 4 Time Mass Time Mass Time Mass time DNI (W/sq.m)

0 14:02 0 14:23 0 15:25 0.26 14:00 955 1 14:03 0.175 14:24 0.055 14:51 0 15:26 0.345 14:05 958 2 14:04 0.23 14:25 0.105 14:52 15:27 0.375 14:10 953 3 14:05 0.275 14:26 14:53 0.07 15:28 0.41 14:15 945 4 14:06 0.31 14:27 0.185 14:54 0.105 15:29 0.475 14:20 959 5 14:07 0.345 14:28 0.22 14:55 0.13 15:30 0.51 14:25 948 6 14:08 0.385 14:29 0.255 14:56 0.17 15:31 0.535 14:30 938 7 14:09 0.42 14:30 0.29 14:57 0.195 15:32 0.57 14:35 938 8 14:10 0.445 14:31 0.33 14:58 0.22 14:40 927 9 14:11 0.495 14:32 0.37 14:59 0.24 14:45 925 10 14:12 0.53 14:33 0.41 15:00 0.265 14:50 916 11 14:13 0.565 14:34 0.45 15:01 0.28 14:55 911 12 14:35 0.485 15:02 0.305 15:00 916 13 14:36 0.52 15:05 911

15:10 907 Lsteam 2.258 MJ/kg 2.258 MJ/kg 2.258 MJ/kg 2.258 MJ/kg 15:15 908 Steam 0.18 kg/5mins 0.19 kg/5mins 0.11 kg/5mins 0.195 kg/5mins 15:20 906

Steam rate 36 g/min 38 g/min 22 g/min 39 g/min 15:25 891 Ereq 406 kJ 429 kJ 248 kJ 440 kJ 15:30 883 DNI 945 W/sqm 938 W/sqm 916 W/sqm 883 W/sqm

Solar 851 kJ 844 kJ 824 kJ 794 kJ

Efficiency 48% 51% 30% 55%


APPENDIX J – Detailed calculation sheets for Continuous monitoring

LOCATION PUMP 4 31st October 2010

Dia 43 mm Time ODO SUMMARY Setting 4 START 8:44 247.9 Pump TM 439 mins Stroke 9.4 cm END 16:15 332 Total Vol 4958 litres Stroke vol 0.136 ltr Computer Flow 677.6 l/hr

GWL 7.45 m 1 cycle 2 m Flow 0.188 l/s BUCKET TEST V = 20 ltr DST CYCLES Vcalc Efficiency Stroke vol. TIME flow TEST TIME start end total l/str (s) l/s

9:00 1.97 2.38 0.41 205 28.0 72% 0.098 157 0.13 9:30 7.73 8.06 0.33 165 22.5 89% 0.121 93 0.22 10:00 13.15 13.49 0.34 170 23.2 86% 0.118 102 0.20 10:30 19.28 19.59 0.31 155 21.1 95% 0.129 101 0.20 11:00 25.55 25.86 0.31 155 21.1 95% 0.129 100 0.20 11:30 31.79 32.11 0.32 160 21.8 92% 0.125 117 0.17 12:00 37.84 38.16 0.32 160 21.8 92% 0.125 131 0.15 12:30 44.19 44.52 0.33 165 22.5 89% 0.121 89 0.22 13:00 48.22 48.57 0.35 175 23.9 84% 0.114 73 0.27 13:30 53.01 53.35 0.34 170 23.2 86% 0.118 83 0.24 14:00 59.13 59.47 0.34 170 23.2 86% 0.118 90 0.22 14:30 64.53 64.87 0.34 170 23.2 86% 0.118 106 0.19 15:00 70.07 70.42 0.35 175 23.9 84% 0.114 105 0.19 15:30 75.94 76.3 0.36 180 24.6 81% 0.111 176 0.11 16:00 81.16 81.51 0.35 175 23.9 84% 0.114 177 0.11

AVERAGE 0.340 170 23.194 0.866 0.118 113 0.189


0

1000

2000

3000

4000

5000

6000

0

50

100

150

200

250

9:00 10:00 11:00 12:00 13:00 14:00 15:00 16:00

Total daily volum

e

15‐m

in volum

es

time

Pumped volume cumulative and incremental (31 Oct)

PUMP LOG 31st October 2010 TIME TM DST vol/str Vol Vol cum SPEED P

mins km litres litres km/h rpm bar 9:00 16 1.87 0.098 91.2 91.2 8 67 1 9:15 30 4.18 0.098 112.7 203.9 11.3 94 1.19:30 45 7.11 0.121 177.6 381.5 11.6 97 1.1 9:45 60 10.2 0.121 187.3 568.8 11.6 97 1.1 10:00 75 13.06 0.118 168.2 737.0 11.7 98 1.1 10:15 91 16.16 0.118 182.4 919.3 12.4 103 1.110:30 105 19.27 0.129 200.6 1120.0 12.4 103 1.1 10:45 120 22.24 0.129 191.6 1311.6 11.7 98 1.1 11:00 135 25.34 0.129 200.0 1511.6 11.6 97 11:15 150 28.48 0.129 202.6 1714.2 12.7 106 1.311:30 165 31.61 0.125 195.6 1909.8 12.5 104 1.1 11:45 180 34.66 0.125 190.6 2100.4 12.3 103 1.1 12:00 195 37.72 0.125 191.3 2291.7 12.5 104 1.112:15 210 40.9 0.125 198.8 2490.4 12.1 101 1.212:30 225 44.01 0.121 188.5 2678.9 12.3 103 1.1 12:45 240 47.18 0.121 192.1 2871.0 11.7 98 1.1 13:00 245 47.99 0.114 46.3 2917.3 ‐ ‐ ‐13:15 257 50.19 0.114 125.7 3043.0 11.8 98 1.1 13:30 272 52.9 0.118 159.4 3202.4 11.4 95 1.1 13:45 287 55.93 0.118 178.2 3380.7 12.7 106 1.3 14:00 303 58.88 0.118 173.5 3554.2 11.5 96 1.114:15 318 61.67 0.118 164.1 3718.3 11.3 94 1 14:30 332 64.39 0.118 160.0 3878.3 11.2 93 1 14:45 347 67.19 0.118 164.7 4043.0 11.2 93 1 15:00 362 69.97 0.114 158.9 4201.9 11.4 95 115:15 377 72.82 0.114 162.9 4364.7 11.6 97 1 15:30 392 75.71 0.111 160.6 4525.3 10.5 88 1 15:45 408 78.66 0.111 163.9 4689.2 10.5 88 1 16:00 423 81.16 0.114 142.9 4832.0 10.1 84 116:15 439 83.36 0.114 125.7 4957.8 6.6 55 0.9 TOTAL 439 4957.8 95


APPENDIX K – Overall system efficiency analysis

Date Location Pump

duration Volume E hyd E sol Overall 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%


APPENDIX L – Kinetic energy losses during pumping

KINETIC ENERGY CALCULATION Pump output Flow l/s 0.2276 hr day litres 4912 Pump pipe pump pipe diam. cm 4.4area 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 0.313Water velocity mean m/s 0.150 peak m/s 0.299Kinetic 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


APPENDIX N – Microfinance

Local Microfinance organisation : Busa Ganofa

Terms and conditions

• Max first loan 2500 birr • If you have a good repayment history loan size will be increased • MIT (micro irrigation technology) – 1 year repayment • Other inputs (chemicals seeds etc.) – 8 months • Interest ~ 20% • Age 20 – 65; at least 2 years previous irrigation experience • IDE undertakes due diligence • Discussion with IDE staff, family • Determine size of loan • Prepare business plan, plan crop types • Prepare land based on agreed business plan incl. excavate well • Busa Ganofa representative visits farmer • Must demonstrate a serious determination and seriousness to his business

plan • Need to specify a guarantor:

o Must have lived in local area for over 5 years o No debts to other institutions o Land certificate proving ownership and tax payments for land

• Final visit by IDE


APPENDIX P - Interview with Technicians - 13/11/2010

Jim Kauffman is an experienced machinist from Pennsylvania (US) and local technicians Binyam and Alena have been based in Ethiopia been working closely with the solar pump for over 2 months. Their work has involved the production of the locally fabricated elements of the pump system, the overall assembly and field installations.

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

‐ PV panel/shading device seems to work – shade must be facing west. ‐ Original arrangement had an intermediate electronic box (charge controller?)

which seemed to operate well in terms of motor control. ‐ Direct PV connection to motor seems to work less well. ‐ Tracker cord was rubber to start which had degrading problems (UV

weakening?). Nylon string with springs made out of rubber cord seems to work much better.

‐ Inclusion of on/off switch would eliminate the need for disconnecting wires. ‐ DC motor seems to have reliability issues ‐ While connected to PV panel and with cord removed from drive mechanism,

will cause motor to run all day ‐ If there is only a small current going through the DC motor, not enough for

operation – is this an issue?

Engine

‐ Currently whole engine including steel parts made in Holland.


‐ 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.


‐ 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.


APPENDIX Q – Typical O & M tasks

For effective and sustainable operation, certain self-maintenance tasks need to be carried out regularly by the user. Once all of these factors have been checked, only then should it be necessary to call the local servicing provider. A schedule of tasks should be included in the instruction manual.

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:

• Weekly cleaning of collector with soap and clean soft cloth • Clear any debris from return hose • Tighten hinges on collector frame • Check hoses are not twisted or bent • Check connections between the boiler and the engine • Check pressure release valve is lubricated and free from rust • Check inlet and outlet valve free and smooth operation • Check connections in and out of the condenser • Check tilt of condenser/outlet pipe close to horizontal

Common encountered problems:

Boiler emptying issues:

• Clean/unblock filter from condenser to feeder pump • Consider changing piston ring in feeder pump

Steam escaping from engine:

• If from valves - check gaskets on valves • If from main engine cylinder - check diaphragm

Pump piton not smooth:

• Manually pump to remove any potential grit between piston ring and pump cylinder


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

# HAZARD RISK MITIGATION 1 Hot steam Scalding from

steam Collector should be faced away from the sun during pump shutdown. All children and untrained personnel to be discouraged from handling hoses or any part of the system

2 Blow out of steam hose during start‐up

Scalding from steam

All children and unnecessary people to stand away from the device between collector being aimed at sun and pump start up

3 Blow out of steam hose during collector close down

Scalding from steam

Options:A ‐ Collector rotated slowly away from the sun B ‐ Carefully disconnect steam hose from engine and point steam towards the ground All children and unnecessary people to stand away from the device at close down

4 Blow out of steam hose boiler refilling

Scalding from steam

If filling up boiler after a period of inoperation steam hose should be disconnected from engine and aimed away towards the ground

5 Rapidly rotating fly‐wheel

Risk of hand or limb injuries if trapped or placed in spokes

All children and untrained personnel to stand away from the device while in operation

6 Engine runaways If engine speed too fast danger exacerbates no.6 as well as increase possibility of engine damage

Control the start up speed using a glove or a cloth on the flywheel. If the flywheel continues to speed up , shut the pump down and check that the system is correctly set up e.g. is the rocker arm attached to the pump piston?

SolarSteamPump.Ethiopia

Documents

Transcript of SolarSteamPump.Ethiopia