Topriska EV, Kolokotroni M, Dehouche Z, Notievo DT and Wilson E (2016). The potential to generate solar hydrogen for cooking applications: Case studies of Ghana, Jamaica and Indonesia, Renewable Energy. Vol 95, September 2016, Pages 495–509 (doi:10.1016/j.renene.2016.04.060)

The potential to generate solar hydrogen for cooking

applications: Case studies of Ghana, Jamaica and Indonesia

Evangelia Topriska1*, Maria Kolokotroni1, Zahir Dehouche1, Divine T. Novieto2, Earle A.

Wilson3

1 Brunel University, Uxbridge, United Kingdom2Ho Polytechnic, Volta Region, Ghana

3 University of Technology, Kingston, Jamaica

* Corresponding author: Evangelia.Topriska@brunel.ac.uk

Abstract

This paper evaluates one option to replace traditional cooking fuels in developing

communities with a flexible, modular and clean solution of solar hydrogen. The study focuses

on Ghana, Jamaica and Indonesia as characteristic examples of developing economies using

fossil fuels for domestic cooking. Statistical analyses are performed and the domestic cooking

demand profile is created for the three countries on available data and a specific quantitative

study in Ghana. The cooking demand profile is used to size solar hydrogen plant case studies

for rural communities based on a TRNSYS numerical model. The simulation results indicate

that the hydrogen plant sizing and management strategy satisfy the annual demands of the

communities which are 621.6kgH2 for Jamaica, 631kgH2 for Indonesia and 785kgH2 for

Ghana. The effect of the weather data on the simulation is estimated by a comparison

between TMY and recent weather data sets for Jamaica. Finkelstein-Schafer statistics indicate

differences between the composite and recent weather data sets, but these are proved to have

a minor effect on simulation results, with 0.9% difference in hydrogen generation. The

potential to establish solar hydrogen plants in the countries is further evaluated through the

creation of novel solar hydrogen potential maps, with Ghana presenting the greatest potential

of 140.5thousand kg/km2/year on average.

Keywords: Solar hydrogen, cooking energy in developing countries, renewable energy

simulation, weather data, hydrogen potential maps

1 Introduction

Energy security in developing economies is an important factor towards the

improvement of the living and social standards. Household energy use in the developing

world represents 10% of the world’s primary energy demand, according to the International

Energy Agency [1] and the main use of domestic energy (90%) is for cooking.

Almost 2.7 billion people depend on the burning of biomass as their primary cooking fuel [1,

2] and most of them live in Asia and sub Saharan Africa. This number is predicted to rise

further in the future; a fact that will have consequences in the increase of deforestation rates,

greenhouse gas emission and health problems related to indoor air pollution [3]. Improving

the currently used cooking systems and replacing them with renewable based ones can have

multiple advantages: reduced rates of respiratory problems, time saving due to less time spent

on fuel collection and cooking and environmental benefits [4].

At the beginning of the 21st century the United Nations (UN) Organisation launched the

Millennium Project that sets the target of reducing by 50% by the year 2015 the number of

households that use biomass as a primary source of fuel [5]. Moreover, the 17 sustainable

development goals set in 2015 further promote access to affordable, reliable and sustainable

energy for all, with a focus on the cooking fuels and energy [6] . Solar powered electric

cooking [7], improved wood stoves [8] and solar cooking projects [9] have been implemented

by the UN in many developing economies as sustainable, clean and emissions free cooking

methods. Beneficial outcomes have been reported by the use of solar cooking in refugee

camps in Africa and multiple locations in Asia and South America where energy poverty was

forcing people to use animal and agricultural waste [10, 11]. Solar cooking and modified

charcoal stoves are also highly promoted through the Global Alliance for Clean Cookstoves

[12] and in 2013 11.7 million modified stoves were distributed in seven countries. The solar

cooking projects however, which were the most ambitious and also consist a more sustainable

option, have not reported particular success. Their dependency on sunlight availability, as the

energy storage option is rarely combined, their low efficiency which sets cooking time-

consuming and the large space demand, especially in the case or parabolic or solar panel

cooker [13, 14], pose hindrances in their use. Furthermore, solar cookers have not been

widely accepted due to social and cultural reasons. As originally solar cooker projects where

targeting extreme energy poverty tackling, there is a strong association with social

discrimination and are faced with reluctance and criticism by local societies [15]. Moreover,

they do not correspond to traditional cooking habits since they limit the cooking place to

specific areas outside the house and also make it hard to cook in high temperatures and

prepare traditional fried meals [15, 16].

A successful alternative cooking system should be easy to adopt and should not pose disrupt

on the daily habits and cooking schedule of the locals that are traditionally used to cook in

stoves [1, 2]. Within this scope, this study evaluates the potential to replace the conventional

cooking systems in selected developing societies through the application of a solar hydrogen

system where the generated hydrogen is used as a cooking gas.

The solar hydrogen system consists of a photovoltaic panel array which powers a proton

exchange membrane (PEM) electrolyser plant. Hydrogen gas is generated and distributed to

households for use as an alternative clean and sustainable cooking fuel in modified gas stoves

with a simple adaptation to burn hydrogen. The current cooking systems in countries selected

for this study consist mainly on LPG, charcoal and firewood stoves and therefore the

introduction of a modified gas stove is a solution that can be easily accepted. The system

operation is evaluated through a numerical model in TRNSYS developed by the authors,

[17].

Safety of hydrogen use at home is an important issue considered in this project. Two

alternative methods of storage were developed; a relatively low cost cascading system where

hydrogen is stored in high pressure cylinders outside the house and is cascaded to lightweight

low pressure containers suitable for transportation in private vehicles and used inside homes

and (b) metal hydride storage which is more expensive and under development at present but

is considered as a safer option. The system described in this paper focuses on the metal

hydride storage. In addition, cookers need to be modified for burning hydrogen to which

odour and flame colour is added for safety. A prototype of the modified cooker has been

constructed and being tested and will be reported in a separate paper. Some visual

information of the developed system can be seen in [18] .

This paper is divided to four sections. Section 2 presents the country case studies

examined for the application of the solar hydrogen system. The process to calculate the daily

cooking profile of the typical household for each case is described. Section 3 presents the

simulation results for each case study based on the TRNSYS solar hydrogen numerical

model. Section 4 analyses the effect of weather data on the simulation results and, section 5

visualises the potential of solar hydrogen generation on a country level in the form of maps.

2 Case Studies Presentation

Three countries were selected as case studies for the evaluation of the solar hydrogen

system, Ghana, Jamaica and Indonesia. All three countries have developing economies [19]

and geographically belong to the near equatorial region. In addition, their prime proportion of

domestic energy demand comes from cooking and is dominated by fossil fuels, mainly

firewood, charcoal and petroleum by-products. Poor ventilation and out-dated cooking

methods result in numerous deaths every year directly related to the emissions of these

cooking fuels [1]. Moreover, these countries can be pushed to finance instabilities being

dependant on the importation of petroleum products for their energy needs.

The purpose of this study is to determine the size of a solar hydrogen system to provide

hydrogen as a cooking gas, that can replace the fossil fuel based cooking systems of selected

communities in the three case study countries. The communities belong to rural areas and

represent actual size small villages of 20 households. The evaluation of the cooking demand

profile for the households is essential for the correct and optimised sizing of the solar

hydrogen system, and this is done through quantitative studies, research and extensive

literature review.

2.1Ghana Case Study

Ghana is a West African country with tropical climate of a total land area of

238,500km2 and a population of 27 million [20]. It is positioned in the Gulf of Guinea and is

a country well-endowed with water resources, which are essential for the generation of

electrolytic hydrogen. The solar energy potential of the country is also very high, ranging

from 4-6 kWh/m2 providing excellent conditions for photovoltaic panel applications [21].

The greatest percentage of final energy use in the country corresponds to the residential

sector with energy for cooking accounting for almost 95% of the household energy use [22].

The majority of the population -57%- lives in rural areas and statistics show that most

Ghanaians rely completely on the use of biomass for cooking (firewood and charcoal) as well

as a small percentage on LPG. [23].

With an aim to create a cooking demand profile for a Ghanaian household and due to

lack of available data in the literature, a questionnaire survey was conducted with a

quantitative approach. Use and cost of fuels was evaluated as well as socio – economical

relation of household and fuel use and the potential to introduce a solar hydrogen system.

Local people from three communities were selected for the field study, Akatsi, Keta and

Sogakope, which are located in the south-east part of the country, in the Volta region [24].

These communities were selected as they represent the typical rural households of Ghana.

Raw data were collected through interviews and standardised questionnaires. The

questionnaires included factual questions with pre-determined answers and opinion based

questions with an open-ended form where the participants had the liberty to give their own

answers. The main subjects that were addressed are: sources of energy for cooking, socio-

economic background of the households and cooking behaviour.

The sample population consists of 155 adults who belong to distinct professional and

social groups. Cooking is a woman dominated activity in Africa [23], and most of the

participants are females who were contacted through visits in the local fuel markets of the

villages. The majority of the population forming the sample are engaged in trading and

owning their own business (in total 64%). The second largest category is teachers, followed

by clerks/ secretaries, farmers and fishermen/ fishmongers, bank workers and pensioners. The

results of this survey are presented by Topriska et al. [25] and a summary is included in this

paper for completeness and to facilitate comparison with Jamaica and Indonesian case-

studies. The results of the survey are analysed in sections 2.1.1 to 2.1.5.

2.1.1 Fuel Use

The first step of the analysis consists in clustering the participants according to the

fuel use. 67 participants use LPG as their main cooking fuel, 60 use charcoal and 28 use

firewood. In total, 56.8% of the participants use solid fuels as their main source for cooking

and 43.2% use LPG. The use of secondary cooking fuels is well established in the local

population. The additional fuel use has multiple purposes. They can act as a backup fuel, as

rural Ghana is an area where fuel supply is not guaranteed, and there are periods where there

can be shortages [26]. Moreover, they are used for the preparation of particular dishes.

Where charcoal is used as a back up to firewood or LPG, it is used specifically for the grilling

of fresh fish and for the cooking of black-eyed beans. The use of two fuels and consequently

two cooking devices enhances the energy security of the households.

Charcoal is the fuel mostly used by the majority of the households - 65.15%.

Following that, 52.90% of the households use LPG as main or additional cooking fuel and

last firewood is used by 36.77% of the households.

The choice of the cooking fuel presents trends that are associated to the area, the

educational attainment and the profession. Figure 1 presents the relation between the main

cooking fuel and the educational degrees of the participants, and the tendency that is

observed is that the use of LPG is more frequent among the participants of secondary school

and university degree.

The fuel use trend varies according to the area, the profession and income range in a

similar way to the educational attainment. Figure 2 shows that in the Sogakope population,

which consists mainly of fishermen and has the lowest income of the sample, the use of

firewood dominates, in contrast to Akatsi and Keta where LPG and charcoal are

predominantly used. This tendency is highlighted in Figure 3 which presents the trend

between fuel use and household income. It is shown clearly that the use of LPG increases as

the income increases whereas the use of charcoal and firewood decrease.

Therefore the poor population of the sample depends mainly on solid fuels, which has

an effect on their health and well-being. As Boadi and Kuitunen report [27] , the

socioeconomic status of local population in Ghana has a significant impact on the respiratory

health and the use of firewood and charcoal is highly associated. More particularly, they

highlight that poor households are more susceptible to respiratory health problems because of

their dependence on solid fuels [27]. This is also evident in this research with 70% of the

charcoal and firewood users stating that they prefer to cook in the outside areas of the house,

due to the smoke and the dirt emitted by the combustion of these fuels.

2.1.2 Daily cooking demand calculation

In most instances the participants provided information of the fuels in kg according to

the period of supply, which varied from a few days to 3 months. The values were averaged

per month, year, and day, and the daily values were used for the calculation of the daily

cooking demand. For each fuel, the adequate calorific value was used in order to convert the

amount of each fuel to the common base of kWhs. Further to this, for each fuel and therefore

cooking device used, the cooking efficiency is different, and what is important is to evaluate

the total kWhs related to the end use. Equation (1) is used to calculate the cooking demand

and is based on the daily amount of the fuel used by each household [25]. For each

participant, the daily amount of each of the three fuels is multiplied with the fuel calorific

value and the related fuel stove efficiency. The results of these calculations are summed for

all the participants and averaged for their size, which is 155, to result to a number that

represents the average cooking demand.

(1)

From the application of Equation (1), the average daily cooking demand of a typical

Ghanaian household is found to be 2.5kWh.

2.1.3 Fuel cost

According to the questionnaires data, the cost per kg of each fuel and the size of the

purchase unit can be seen in Table 1. Fuels are purchased in standard forms and sizes. The

Ghana Energy Outlook 2013 and Ahiataku-Togobo [28, 29] proclaim that a significant

increase in commercial fuel prices is observed in Ghana. This is chiefly due to the volatility

and rise in oil prices and imported LPG. Increased deforestation rates in the last decade that

led to scarcity of the produced firewood and charcoal have also contributed to the increase of

the fuel prices [28].

Feff)kg/kWh(CVF)kg(FDayCeff)kg/kWh(CVC)kg(CDay(

LPGeff)kg/kWh(CVLPG)kg(LPGDay(CDi

G

155

11551

According to the field data LPG is the most costly fuel on a cost per kWh basis, as

demonstrated in Table 2. However, to evaluate the actual expenses of the fuel use, the annual

satisfaction of the cooking demand by each one fuel is calculated. In this case charcoal is the

most expensive followed by LPG and firewood. Regardless of this fact, lower income

households insist on using firewood and charcoal. 42.37% of the charcoal users base their

selection on habit of using this particular fuel but mainly because they perceive it as low-cost.

This percentage is even higher for the people that use firewood, 92.6%.

2.1.4 Emissions Related to the use of the Cooking Fuels

The domestic cooking fuels, and in particular firewood, are the main contributors to

the emissions of the country. 68% of carbon monoxide, more than 50% of methane, over 70%

of sulphur oxides and almost 70% of nitrous oxide emissions result from the use of solid fuels

for cooking. Carbon dioxide accounted for 15% of the country’s emissions and the economic

growth of Ghana is expected to contribute to the rise of the GHG emissions by 200% by the

year 2020, in comparison to the 2000 levels [30].

The main contributor to the CO2 emissions of the cooking fuels of the field study is

firewood. It has a high CO2 emissions factor and it is the fuel with the greatest consumption.

CO is also an important aspect of the emissions that ought to be evaluated, as it can prove

lethal in high concentrations [31]. Charcoal and firewood combustion especially, lead to high

CO emissions and as it is odourless it is very hazardous. Taking into consideration the

emission factors of the LCA case study of Alfrane and Ntiamoah in Ghana [32], the

emissions of the fuels used in the case study can be seen in Table 3.

Thus, replacing the use of the currently used fuels with hydrogen could lead to the

reduction of almost 2 million kg of CO2 being emitted to the atmosphere each year.

2.1.5 Reasons to switch to hydrogen as a cooking fuel

99.4% of the participants of the field study showed willingness to change the

currently used cooking fuels, for hydrogen cooking gas. The main reasons for the switch to

hydrogen gas would be if it was cheaper and safer in use than the current cooking fuels, and

especially for the users of firewood and charcoal, as shown in Figure 4. It should be noted

here that all participants show an interest in the hydrogen technology as something new and

innovative

2.2Jamaica Case Study

Jamaica is an island country of the Great Antilles in the Caribbean Sea of a population

of almost 3 million people. It has annual average solar irradiance values of 4.1-5.6

kWh/m2/day [33] and is a country well-endowed with natural resources. Nevertheless, it

remains dependant on imports of petroleum products, which account for 90% of its energy

mix [34]. This situation results in high energy import bills and cost of energy generation and

distribution. The oil refinery stations use aged technology and they are at the stage of

maintenance or replacement. The energy security of the country is subject to the global oil

prices and availability and in the past Jamaica has suffered greatly at periods of oil crisis.

Cooking is one of the activities that are highly affected by the instability of the energy

supply, as it mainly depends on LPG. As seen in Table 4, 86% of the overall Jamaican

households use LPG for cooking, with firewood and charcoal being the second most popular.

The estimation of the cooking profile of a typical Jamaican household is done through

the use of Equation (2) with data provided by the Planning and Statistical Institute of Jamaica

[35], as seen in Table 5. The average size of the LPG cylinders is multiplied to the LPG

calorific value and the LPG stove efficiency, and averaged over the period of supply.

(2)

The average daily cooking demand of a typical Jamaican household is found to be

1.98kWh.

2.3Indonesia Case Study

The third case study for the application of the solar hydrogen system is Indonesia, a

country in Southeast Asia, with a population of approximately 250 million. In terms of

weather conditions, Indonesia is characterized by a monsoonal climate, with a wet (December

– March) and a dry period (June – September). Solar irradiance range is averagely 4.8

kWh/m2/day [36, 37].

According to the Indonesia Energy Outlook for 2014, 72% of the total population of

Indonesia depends on biomass for cooking, i.e. almost 170 million [38]. The fuel mostly used

to satisfy the cooking needs has traditionally been kerosene which was subsidised by the

Indonesian government. This subsidy was a major cost burden for the state as it reached the

equivalent of 18% of the state expenditures and 57% of the total petroleum products subsidy

in 2008. Consequently, there has been an effort to swift away from the use of kerosene

through the promotion of LPG [39]. The use of firewood and charcoal is also widespread in

Indonesia especially in areas where kerosene and LPG supply is not guaranteed due to bad

weather and difficult access. Deforestation and the destruction of peat lands make Indonesia

the world's third largest emitter of greenhouse gases [40].

The estimation of the average daily cooking profile of a typical Indonesian household is

done through the use of Equation (3) in a similar way as described in Equation (2) and based

)days(LPGeff)kg/kWh(CVLPG)kg(LPGcyl

CD JJ

on the data provided by the kerosene to LPG gas conversion programme in Indonesia, [39],

and is found to be 2kWh.

(3)

3 TRNSYS model development of the Case Studies

The replacement of the conventional cooking systems in the case study countries by a

solar hydrogen system is examined. This is evaluated through simulations of a numerical

model of the system, developed by the authors in TRNSYS [17]. Three case studies are

developed for small communities of 20 households in each country. In each case the design

consists in hydrogen being produced in a central plant, where the electrolysis process is

powered by PV panels. Hydrogen is generated at low pressure (13.8bar), directly stored in

metal hydride cylinders and distributed to the households on a monthly basis. The systems

are designed and optimised as autonomous in order to meet the hydrogen demand at the

minimum cost of energy. Defining the optimal system configuration and components sizes is

pivotal and many researchers have reported work on the optimisation of similar systems.

Degiorgis et al. [41] have presented the techno economic analysis and optimisation strategy

of a renewable hydrogen based power system. Guinot et al., [42], Behzadi and Niasati, [43]

and Zhou et al. [44] describe optimisation techniques for power management strategy in

solar hydrogen systems. The importance in control and design strategies in PV powered

hydrogen systems is also highlighted in the work of Ulleberg [45, 46].

)days(Keff)kg/kWh(CVK)kg(Kcyl

)days(LPGeff)kg/kWh(CVLPG)kg(LPGcyl

CD

I

I 21

The methodology followed to size the systems is in line with the one described by

Zoulias and Lymperopoulos [47]. In each case study system, factors that change are

meteorological data and cooking load profiles, as calculated in section 2 of this paper. The

photovoltaic panels used for all the systems are the TrinaSolar TSM-180DA01, with a rated

power of 180W [48].

3.1Ghana case study

From the field data of Ghana, the cooking demand was estimated as 2.5kWh/day, for

the average household. With hydrogen stove efficiency of 60%, this is equivalent to

4.20kWh/day or 0.107kg of hydrogen per day. Therefore, for the 20 households 2.15kg of

hydrogen per day are necessary. The metal hydride storage tank selected to supply each

household once per month is designed for 3.35kg of hydrogen. The TRNSYS model for the

Ghana case study is shown in Figure 5.

The photovoltaic panels array calculated has maximum rated power 77.4kW. The

array is oversized to be able to provide the necessary power input so that the electrolyser can

achieve the required production even after 20 years. The performance degradation of the PV

panels over time is a factor that has to be taken into consideration [42, 44]. The electrolyser

works more efficiently when it operates continuously and is not subject to stops and start-ups

[17], therefore this factor is considered for the sizing as well.

Two electrolysers are designed for the Ghana case study and the parameters of each

are shown in Table 6. The results of the simulation show that the selected system size

satisfies the demand by almost 30kg extra of hydrogen per year at current conditions and can

be seen in Table 7.

3.2Jamaica case study

The cooking demand for a typical Jamaican household was estimated as

1.98kWh/day. With hydrogen stove efficiency of 60%, the equivalent in hydrogen is

3.3kWh/day or 0.0842kg per day, and the 20 households require 1.68kg of hydrogen per day.

The metal hydride storage tank selected to supply each household once per month is designed

for 2.65kg of hydrogen. A 63kW array is calculated for the Jamaica case study, which after

20 years will reduce to the level of 58.9kW and will still be sufficient to satisfy the annual

cooking needs, as shown in Table 8. Two electrolysers are utilised for the Jamaica case study

and are the same as in the Ghana case study and their characteristics as presented in Table 6.

The demand is satisfied by 31.3kg extra hydrogen at the current conditions.

3.3Indonesia case study

According to the analysis of section 2.3 the cooking demand for a typical Indonesian

household was estimated as 2kWh/day. With a hydrogen stove efficiency of 60%, this is

equal to 3.34kWh/day of hydrogen or 0.085kg of hydrogen per day. Therefore, to supply the

20 households, 1.7kg of hydrogen per day is needed with a metal hydride storage tank of

hydrogen storage potential of 2.7kg. The PV array is sized as 54kW, which after 20 years is

reduced to 50.4kW and the annual cooking needs are satisfied in all cases as shown in Table

9. The electrolysers utilised for the Indonesia case study are two and again their

characteristics are presented in Table 6. The current demand is over satisfied by almost 39kg.

3.4Simulation results discussion

Figures 6 to 8 present the monthly hydrogen generation in relation to the monthly

hydrogen demand, the metal hydride tank size for the monthly supply of the households and

also the associated monthly energy generated by the PV panels. The results are presented for

two scenarios, the current conditions and the conditions after the 20 year period, which takes

into consideration the panel degradation.

There are occasions that the monthly hydrogen generation is more than the equivalent

demand, but there are also months that the production does not meet the demand. The system

is designed so that the extra hydrogen production acts supplementary for the months that

there is shortage, and a regulated operation and fuel supply is guaranteed for both scenarios.

This energy management strategy was chosen as the most optimised in terms of total system

costs. As Zoulias and Lymperopoulos explain [47], over dimensioning a power system should

be avoided in such cases where the shortage effect is minimal but the cost implications are

significant.

The replacement of the conventional cooking systems in the case studies with the solar

hydrogen system has a significant projected capital cost but the operational costs are minimal.

The communities can benefit from the provision of a clean fuel which can be locally

produced. Furthermore, it can have a significant impact in CO2 emissions reduction; it is

calculated that based on [49] 175tn of CO2/year can be saved for the Jamaica case study,

256tn of CO2/year for the Ghana case study and 180tn of CO2/year for the Indonesia case

study.

4 Weather data analysis and effect on the simulation

Weather data is a significant factor for the simulation of energy systems and suitable

weather data allow the simulation to give solid results and understanding of the energy

system’s behaviour [50]. An analysis and a comparison of the weather data sets used for the

Jamaica case study are performed in this section.

Recent weather data for Jamaica and typical weather data available from the Meteonorm

7.0 database are analysed through the use of Finkelstein-Schafer Statistics. Moreover, in

order to evaluate the effect of the weather data on the system operation, simulations with the

different weather data sets are performed in the last part of this section.

4.1Composite weather data

According to Song and Haberl, the TMY2 format is the most suitable for energy systems

simulations where solar variation is critical for the results [51]. To this Crawly also adds that

the TMY2 method shows good results and provides users with energy simulation results that

most closely represent weather patterns [52]. Therefore the TMY2 weather data format is

selected as the most suitable for the simulation of the solar hydrogen system in TRNSYS

software.

Weather data for Jamaica that is compatible with TRNSYS was derived from

Meteonorm 7.0 database. The typical weather year for Jamaica consists of a mixture of data

that is measured on the weather stations of the island, and satellite data [53]. The weather

stations of Jamaica record temperature, wind speed and relative humidity but the irradiance

data is GOES satellite data of the years 2010 - 2013 with a resolution of 8km on the

horizontal, adapted to regional stations (Cuba, Puerto Rico and Central America) for the

period 1971-1990 [54].

4.2Recent weather data

The composite weather data of Meteonorm is compared to recent weather data for

Jamaica from a weather station installed in the University of Technology in Kingston [55].

The recent weather data sets include data of:

Global horizontal radiation

Dry bulb temperature

Wind speed

Relative humidity

The comparison between the recent weather data of Kingston and the composite

Meteonorm weather data will assist towards the identification of variations and possible

differences in certain periods. Especially in this case that the composite irradiance data

originate from satellite measurements, extra inaccuracies might be introduced. For a solar

hydrogen system big variations in the irradiance of the composite weather data set in relation

to the actual recent irradiance can result in oversizing or under sizing the PV array that will

supply the electrolyser.

4.3Weather data sets comparison

The analysis is based on the Finkelstein-Schafer (FS) statistical method which

constitutes in the generation and comparison of Cumulative Distribution Functions (CDF).

The FS statistic is applied as a mean of comparing two sets of distributions and indicating

their similarities or dissimilarities. The smaller the FS the more similar the two distribution

sets, and in the case it is zero the two sets are identical [56]. The FS statistical method has

been greatly used in the TMY generation process. Wilcox and Marion [57], Levenmore and

Chow [58], Lee, Woo and Levenmore [59] are some of the researchers that present the

method and discuss the technical advantages of the FS statistics in the TMY generation

process. Furthermore, the FS statistical method has been the main assessment method in the

process of comparing different sets of weather data and generating TMY for various locations

around the world. Kalogirou [60] and Argiriou et al. [61], examine the methodologies and

describe the process for the generation of TMY in south east Europe. Rahman and Dewsbury

[50], Zang et al. [62] and Chan et al. [63] present work regarding the generation of TMY files

in locations in Asia. Similarly, Fagbenle [64] and Ohunakin et al. [65] report the process of

creating TMY in Africa.

Quarterly values for a three year period (2011, 2012, and 2013) are acquired for the

recent weather data. From these quarterly values the daily average values are generated.

Regarding the composite weather data, the hourly average values are available in the

Meteonorm weather data set for Jamaica. From these hourly values the daily average values

are generated.

In order to compare these sets of weather data the FS statistical method is used and

adapted so as to make three comparison processes:

2011 weather data set and the Meteonorm typical year data

2012 weather data set and the Meteonorm typical year data

2013 weather data set and the Meteonorm typical year data

The CDF curves for the Meteonorm weather data set and for the recent weather data sets

are generated for the daily average values of global horizontal radiation, dry bulb

temperature, relative humidity, and wind speed for every month. They are compared and the

similarities of the distribution sets are measured as per the Equations (4) to (6):

(4)

(5)

(6)

The CDF curves of global horizontal radiation, dry bulb temperature, relative

humidity, and wind speed for each of the three years of recent data in comparison to the CDF

curves for Meteonorm typical year are given in Figure 9. The calculated FS statistics of each

weather parameter for each of the total years of the three years are given in Table 10. For the

total weighted FS, equal weighting factors are used.

N

iitcentDataSeReiataSetMeteonormDx )x(CDF)x(CDF

NFS

1

1

M

xxxweighted FSWF

MFS

1

1

M

xxWF

1

1

4.4Discussion of the weather data analysis

For global horizontal radiation, the CDF curve of 2011 is more similar to the

Meteonorm CDF curve and 2013 is the most different. This is also confirmed by the FS

values. The maximum global horizontal radiation value for Meteonorm is greater than the

maximum values for all the three recent years by 4.6% for 2011, 12.15% for 2012 and

12.34% for 2013. Thus, it is concluded also through the CDF curves that Meteonorm

overestimates global horizontal radiation.

For dry bulb temperature the CDF curve of 2013 is closer to the CDF for Meteonorm.

2011 and 2012 have similar average yearly temperatures, 25.82 ˚C and 25.76 ˚C, whereas

2013 has an average temperature of 26.25˚C which is closer to the Meteonorm average of

27.72 ˚C. The FS also indicates that 2012 is the year most different than Meteonorm.

Overestimation of dry bulb temperature has an effect on the operation of PV panels as high

temperatures affect the panels’ voltage and cause it to drop, thus causing a consequent

reduction in the power output .

For relative humidity the CDF curves for all three years are almost similar and the one

closest to the Meteonorm CDF curve is for 2013, with the smallest FS. This is highlighted

even more from the average yearly values that are 74.30% for 2011, 74.91% for 2012 and

74.05% for 2013. The average value for the Meteonorm year is 69.61%. It can be thus

concluded that Meteonorm underestimates the relative humidity values.

For wind speed the CDF curve of Meteonorm differs significantly from the CDF curves

of the recent weather years. The FS for 2011 is the biggest showing that this year is the most

dissimilar. The average yearly wind speed of Meteonorm is by average 25% greater than the

average yearly values for all three years.

The total weighted FS statistics between the recent weather data and the Meteonorm

data indicate that the year that is most different to the Meteonorm year is 2013 with FS=12.5.

The effect of the dissimilarities between the Meteonorm file and the recent weather data will

therefore be examined through simulations with the Meteonorm and the 2013 weather data

set.

4.5The effect of the Composite and the Recent Weather Data on the

Simulation Results

The monthly and annual hydrogen generation for the simulation with Meteonorm and

the 2013 weather data sets can be seen in Table 11.

As can be seen the annual hydrogen generation is higher for the simulation with

Meteonorm weather data than the 2013 weather data, by 5.81kg H2. This corresponds to a

0.9% difference in the hydrogen generation between the two weather data sets. The

overestimation of the global horizontal radiation in Meteonorm, leads to a smaller size of PV

array than if sizing would be based on the recent weather data. Nevertheless, this difference is

insignificant and even though the recent weather data simulation gives a smaller result in

terms of hydrogen generation, the demand is still covered.

Therefore, it can be concluded that even though the FS statistics between the Meteonorm

weather data set and the 2013 recent weather data set indicates a difference, the simulation

results between the two data sets are very similar. It is thus considered correct to use the

Meteonorm file for simulation, as the inaccuracy of the irradiance data which are GOES

satellite data is proved to have a negligible effect.

5 Solar Hydrogen Potential Maps

A step further to the analysis of the solar hydrogen system in the case study countries is

the evaluation of the country potential of solar hydrogen generation through the creation of

solar hydrogen potential maps. The generation procedure is partly based on the methodology

of the National Renewable Energy Laboratory (NREL) for the estimation of the potential for

hydrogen production from key renewable resources in the United States [66] :

I. The country is separated in specific areas. In this research, the areas are represented

by the administrative areas.

II. The solar potential for each country in kWh/m2/day is evaluated for each

administrative area

III. In each area 10% of the land is assumed to be used for solar panel installation, and

only 30% of this area to be actually covered with solar panels [66].

IV. The solar panels utilised are assumed to have an efficiency of 15%.

V. The electricity demand of the electrolysis process is 61kWh/kg of hydrogen, for the

PEM electrolysers used [17]

These steps are combined in Equation (7) which calculates the normalised solar hydrogen

potential in (thousand kg/km2/year).

(7)

The above calculations were followed for Ghana, Jamaica and Indonesia. In order to

create the solar hydrogen potential maps the results of these calculations were used as input

to the ArcMap software which is a professional Geographic Information Software (GIS) for

the generation of maps [67]. The software receives the information (numbers) from the

)kg

kWh(

)yearday()

daymkWh(Solar...

lSHpotentia61

365101503010 23

inputted database and lays it on the country map as features. A GIS system is the most

appropriate way to generate maps as it is designed to analyse, manage and present any type of

spatial or geographical data. Rumbayan et al. describe how using this software can prove to

be a very effective in the generation of solar energy potential maps and present an application

study in Indonesia [68].

The generated solar hydrogen potential maps for Ghana, Jamaica and Indonesia are

presented in Figures 10 to 12. The map sizes appear in auto-zoom scale to show the

variations within the countries clearly, but in reality Indonesia is the biggest country,

followed by Ghana and last Jamaica. Each administrative area is characterised by a specific

range of hydrogen potential in the normalised unit of thousands of kg of hydrogen per km2

per year. The solar hydrogen potential values are grouped into five classifications: 110-120,

120-130, 130-140, 140-150 and 150-160. The darkest colours on the maps indicate the higher

solar hydrogen potential. It is observed that Ghana includes areas with higher solar hydrogen

potential than Jamaica and Indonesia, and this is due to the fact that the global horizontal

radiation in Ghana (4-6 kWh/m2/day) presents higher values than in the other two countries.

In total country area indicators, Table 12 shows that Ghana has the greatest solar

hydrogen potential per area unit of the three countries. However in total country area,

Indonesia presents by far the greatest potential as it is the biggest country of the three,

followed by Ghana and last Jamaica.

These results and the generated maps can be used as a first step for the evaluation of

solar powered electrolysis investments in the selected countries. The maps act as an indicator

of the most appropriate areas for the solar hydrogen system application and can be used as a

mean to assess the replacement of the currently used fossil fuels by hydrogen.

6 Conclusions

This paper presented the application of a solar hydrogen system as one option of

potential cooking system in developing societies. Three countries were examined for the

application, Ghana, Jamaica and Indonesia, all three currently presenting high dependence on

fossil fuels for cooking.

The calculation of the domestic daily cooking demand profile resulted in 2.5kWh/day for

Ghana, 1.98kWh/day for Jamaica and 2kWh/day for Indonesia. These values were inputted in

a numerical model in TRNSYS which is used for the design and sizing of a solar hydrogen

plant. Hydrogen generation rates were calculated according to the available PV output for the

weather conditions of each country, and the annual cooking demands were satisfied by 815kg

H2 of hydrogen for Ghana, 653kg H2 for Jamaica and 670kg H2 for Indonesia.

Furthermore, the effect of the weather data on the simulation was evaluated through the

comparison of the composite Meteonorm weather file used for the simulation, to a recent

weather data set of Jamaica. Finkelstein-Schafer statistic indicates a difference between the

datasets. Nevertheless, the results of the simulation with the two weather data sets were found

to differ only by 0.9% in the annual hydrogen production, thus confirming that the use of the

composite weather file is reliable.

Finally, the potential of the development of solar hydrogen plants in the case study

countries, was further estimated through the creation of novel solar hydrogen potential maps,

that can be used as an initial decision making tool for investments.

7 Acknowledgments

This work is partly funded by the ACP Caribbean & Pacific Research Programme for

Sustainable Development of the European Union (EuropeAid/130381/D/ACT/ACP).

8 Works Cited

[1] IEA, “Energy For Cooking in Developing Countries,” 2006. [Online]. Available:

http://www.iea.org/publications/freepublications/publication/cooking.pdf. [Accessed

November 2013].

[2] W. Foell, S. Pachauri, D. Spreng and H. Zerriffi, “Household cooking fuels and

technologies in developing economies,” Energy Policy, vol. 39, no. 2011, pp. 7487-

7496, 2011.

[3] World Health Organization, “Indoor Air Pollution, Household Energy and the

Millenium Deevelopment Goals,” 2004. [Online]. Available:

http://www.who.int/indoorair/info/iabriefing1rev.pdf. [Accessed November 2013].

[4] World Health Organisation, “Fuel For Life - Household Energy and Health,” WHO ,

2006. [Online]. Available: http://www.who.int/indoorair/publications/fuelforlife.pdf.

[Accessed October 2014].

[5] “UN Millenium Project Goals, Targets and Indicators,” 2006. [Online]. Available:

http://www.unmillenniumproject.org/goals/gti.htm#goal1. [Accessed November 2013].

[6] United Nations, “Goal 7: Ensure access to affordable, reliable, sustainable and modern

energy for all,” 2015. [Online]. Available:

http://www.un.org/sustainabledevelopment/energy/. [Accessed November 2015].

[7] Inited Nations in India, “Sustainable Energy for all,” 2015. [Online]. Available:

http://in.one.un.org/task-teams/sustainable-energy-for-all. [Accessed November 2015].

[8] United Nations, “A cooking stove improves lives in Kenya,” 2015. [Online]. Available:

http://www.un.org/en/sustainablefuture/stories_kenya_cooking_stove.shtml. [Accessed

November 2015].

[9] U. N. D. Programme, “The Mekhe Solar Cooker Project,” 2015. [Online]. Available:

http://www.undp.org/content/undp/en/home/librarypage/environment-energy/

climate_change/the_mekhe_solar_cookerproject.html. [Accessed November 2015].

[10] The UN Refugee Agency, “Solar cooker offers ray of hope for refugee environment,”

2004. [Online]. Available: http://www.unhcr.org/40c08d4b4.html. [Accessed November

2015].

[11] Solar Cookers International, “How Many Solar Cookers?,” 2015. [Online]. Available:

http://www.solarcookers.org/our-work/solar-cooker-distribution/. [Accessed November

2015].

[12] Global Alliance for Clean Cooking, “Cookstoves2015,” 2015. [Online]. Available:

https://www.cleancooking2015.org. [Accessed November 2015].

[13] “Solar Cookers: Pros and Cons of the Different Types,” 2015. [Online]. Available:

http://www.solarcooker-at-cantinawest.com/solarcookers-prosandcons.html. [Accessed

November].

[14] E. Cuce and P. M. Cuce, “A comprehensive review on solar cookers,” Applied Energy,

vol. 103, pp. 1399-1421, 2012.

[15] P. Piroschka Otte, “Solar cookers in developin gcountries—What is their key to

success,” Energy Policy, vol. 63, no. 2013, pp. 375-381, 2013.

[16] G. L. Simon, A. G. Bumpus and P. Mann, “Win-win scenarios at the climate–

development interface: Challenges and opportunities for stove replacement programs

through carbon finance,” Global Environmental Change, vol. 22, no. 2012, pp. 275-287,

2011.

[17] E. Topriska, M. Kolokotroni, Z. Dehouche and E. Wilson, “Solar hydrogen system for

cooking applications: Experimental and numerical study,” Renewable Energy, vol. 83,

no. November 2015, pp. 717-728, 2015.

[18] University of Technology, Jamaica, “Sustainable Hydrogen Cooking Gas Project,”

2015. [Online]. Available: http://www.solarhydrogen.utechsapna.com/. [Accessed

November 2015].

[19] The International Statistic Institute, “Developing Countries,” 2015. [Online]. Available:

http://www.isi-web.org/component/content/article/5-root/root/81-developing. [Accessed

May 2015].

[20] Ghana Statistical Service, “2010 Population and Housing Cencus,” 2014. [Online].

Available: http://www.statsghana.gov.gh/docfiles/2010phc/National%20Population

%20Projection_2010%20to%202014.pdf. [Accessed November 2014].

[21] M. Seth and O. Essandoh, “Ghana Country Report,” Ministry of Energy Ghana, 2011.

[22] C. G. Abavana, “Ghana: Energy and Poverty Reduction Strategy,” Government of

Ghana, Ouagadougou, Burkina Faso, 2004.

[23] Ministry of Energy, “Gender Assessment of the Ghana Energy Sector,” 2006. [Online].

Available: http://www.energia-africa.org/fileadmin/files/media/reports/Ghana/Ghana

%20Audit%20final%20report.pdf. [Accessed May 2015].

[24] N. S. S. o. Ghana, “Population by region, district, locality of residence, age groups and

sex, 2010,” 2012. [Online]. Available:

http://www.statsghana.gov.gh/docfiles/population_by_region_district_locality_of_resid

ence_age_groups_and_sex,_2010.pdf. [Accessed November 2014].

[25] E. Topriska, M. Kolokotroni, Z. Dehouche, D. T. Novieto and E. A. Wilson, “Analysis

of domestic cooking fuels demand in Ghana: field work and case study application of

solar hydrogen cooking system,” in WREC XIV, University POLITECHNICA of

Bucharest, Romania, June 8-12, 2015, Bucharest, 2015.

[26] W. Ahiataku-Togobo, “National Cooking Energy Strategies-Lessons from Ghana,”

Ministry of Energy Ghana, Ouagadoukou, 2013.

[27] K. O. Boadi and M. Kuitunen, “Factors affecting the choice of cooking fuel, cooking

place and respiratory health in the Accra metropolitan area, Ghana,” Journal of

Biosocial Sciences, vol. 38, no. 2006, pp. 403-412, 2005.

[28] Energy Commission Ghana, “Energy Outlook for Ghana 2013,” Energy Commission,

Accra, 2013.

[29] W. Ahiataku-Togobo, “Ghana Household Energy Project,” Abuja, 2007.

[30] Strategic National Energy Plan 2006-2020 Annex 1, “Strategic National Energy Plan

2006-2020 Annex 1,” Energy Commission Ghana, Accra, 2006.

[31] United Nations, “Global Alliance for Clean Cookstoves,” 2014. [Online]. Available:

http://www.cleancookstoves.org/our-work/the-solutions/cookstove-fuels.html.

[Accessed November 2014].

[32] G. Alfrane and A. Ntiamoah, “Analysis of the Life-Cycle Costs and Environmental

Impacts of Cooking Fuels Used in Ghana,” Applied Energy, p. 6, 2012.

[33] D. Loy and M. Coviello, “Renewable Energies Potential in Jamaica,” Ministry of

Commerce Science & Technology Government of Jamaica, Santiago, 2005.

[34] “National Renewable Energy Policy 2009-2030,” Ministry of Energy and Mining,

Kingston, 2010.

[35] “Residential Consumer End Use Survey Volume 1,” Planning Institute and Statistican

Institute of Jamaica, Kingston, Jamaica, 2007.

[36] MEMR, “Directorate General of Mineral and Coal, Indonesian Coal Resources

Development and Future Direction of Coal Export,” Japan, 2011.

[37] Weather Online, “Weather Online- Indonesia,” 2015. [Online]. Available:

http://www.weatheronline.co.uk/reports/climate/Indonesia.htm. [Accessed February

2015].

[38] BBPT, “Indonesia Energy Outlook 2014,” Jakarta, 2014.

[39] Pertamina, “Kerosene to LP Gas Conversion Programme in Indonesia,” World LP Gas

Association, France, 2008.

[40] Rainforest Action Network, “Indonesia Climate Change and Rainforests,” Rainforest

Action Network, San Fransisco, 2014.

[41] L. Degiorgis, M. Santarelli and M. Cali, “Hydrogen from Renewable Energy: A Pilot

Plan for Thermal Production and Mobility,” Journal of Power Sources, vol. 171, pp.

237-246, 2007.

[42] B. Guinot, B. Champel, F. Montignac, E. Lemaire, D. Vannucci, S. Sailler and Y.

Bultel, “Techno-Economic Study of PV-Hydrogen-Battery Hydrogen System for Off-

Grid Power Supply: Impact of Performances' Ageing on Optimal System Sizing and

Competitiveness,” International Journal of Hydrogen Energy, vol. 2015, no. 40, pp.

623-632, 2014.

[43] M. S. Behzadi and M. Niasati, “Comparative performance analysis of a hybrid

PV/FC/battery stand-alone system using different power management strategies and

sizing approaches,” International Journal of Hydrogen Energy, vol. 40, no. 2015, pp.

538-548, 2014.

[44] K. Zhou, J. A. Ferreira and S. de Hann, “Optimal Energy Management Strategy and

System Sizing Method for Stand-Alone Photovoltaic-Hydrogen Systems,” International

Journal of Hydrogen Energy, vol. 2008, no. 33, pp. 477-489, 2007.

[45] O. Ulleberg, Stand-Alone Power Systems for the Future: Optimal Design, Operation &

Control of Solar Hydrogen Energy Systems, Trondheim: Department of Thermal

Energy and Hydropower, Norwegian University of Science and Technology, 1998.

[46] O. Ulleberg, “The Importance of Control Strategies in PV-Hydrogen Systems,” Solar

Energy, vol. 76, no. 2004, pp. 323-329, 2004.

[47] E. I. Zoulias and N. Lymperopoulos, Hydrogen-Based Autonomous Power Systems, 1

ed., London: Springer, 2008.

[48] Civic Solar, “Trina Solar TSM-180DA01 180 Watts Solar Panel,” 2015. [Online].

Available: http://www.civicsolar.com/product/trina-solar-tsm-180da01-solar-panel.

[Accessed May 2015].

[49] EIA, “Total Carbon Dioxide Emissions from the Consumption of Energy 2012,” 2012.

[Online]. Available: http://www.eia.gov/beta/international/data/browser/#?

cy=2012&start=1980&end=2012&ord=SA&vs=INTL.44-8-GHA-

MMTCD.A~INTL.44-8-IDN-MMTCD.A~INTL.44-8-JAM-

MMTCD.A&v=H&vo=0&so=0&io=0&c=00000000000000001000102&pa=00000000

000000000000000003&f=A&ug=g&ct=0&tl_type=p. [Accessed May 2015].

[50] I. A. Rahman and J. Dewsbury, “Selection of Typical Weather Data for Subang,

Malaysia,” Building and Environment, no. 42, pp. 3636-3641, 2007.

[51] S. Song and J. S. Haberl, “Analysis of teh Impact of Using Synthetic Data Correlated

with Measured Data on the Calibrated as-built Simulation of Commerical Building,”

Energy and Buildings, p. 11, 2013.

[52] D. B. Crawley, “Which Weather Data Should You Use dor Energy Simulations of

Commercial Buildings?,” ASHRAE, Washington D.C., 1998.

[53] Meteotest, “Weather Stations in Meteonorm,” 2014. [Online]. Available:

http://meteonorm.com/en/. [Accessed July 2014].

[54] Meteonorm Version 7, “Handbook Part1: Review and Software,” Meteonorm, Bern,

2014.

[55] University of Technology, Jamaica, “Faculty of Engineering & Computing Receives

Portable Weather Station for Solar Energy Research,” 2010. [Online]. Available:

http://www.utechjamaica.edu.jm/news/vol14iss11/. [Accessed July 2014].

[56] J. M. Finkelstein and R. E. Schafer, “Improved Goodness of Fit-Tests,” Biometrika, vol.

58, no. 3, pp. 641-645, 1971.

[57] S. Wilcox and W. Marion, “User's Manual for TMY3 Data Sets,” May 2008. [Online].

Available: http://www.nrel.gov/docs/fy08osti/43156.pdf. [Accessed July 2014].

[58] G. J. Levenmore and D. H. Chow, “Solar Irradiation and Coincident Dry-bulb

Temperature Selection for Load Calculations,” ASHRAE, USA, 2006.

[59] K. Lee, H. Yoo and G. J. Levenmore, “Generation of Typical Weather Data Using the

ISO Test Reference Year (TRY) Method for Major Cities of South Korea,” Building

and Environment, vol. 45, pp. 956-963, 2009.

[60] S. A. Kalogirou, “Generation of Typical Meteorological Year (TMY-2) for Nicosia,

Cyprous,” Renewable Energy, vol. 28, pp. 2317-2334, 2003.

[61] A. Argiriou, S. Lykoudis, S. Kontoyiannidis, C. A. Balaras, D. Asimakopoulos, M.

Petrakis and P. Kassomenos, “Comparison of Methodologies for Generation Using 20

Years Data for Athens, Greece,” Solar Energy, vol. 66, no. 1, pp. 33-45, 1999.

[62] H. Zang, Q. Xu, D. Pengwei and K. Ichiyanagi, “A Modified Method to Generate

Typical Meteorological Years from Long-Term Weather Database,” International

Journal of Photoenergy, vol. 2012, p. 9, 2012.

[63] A. L. Chan, T. T. Chow, S. K. Fong and J. Z. Lin, “Generation of Typical

Meteorological Year for Hong Kong,” Energy Conversion and Management, vol. 47,

pp. 87-96, 2006.

[64] R. L. Fagbenle, “Generation of a Test Reference Year for Ibadan, Nigeria,” Energy

Conversion and Management, vol. 36, pp. 61-63, 1995.

[65] O. S. Ohunakin, M. S. Adaramola, O. M. Oyewola and R. O. Fagbenle, “Generation of

Typical Meteorological Year for North-East Nigeria,” Applied Energy, vol. 112, pp.

152-159, 2013.

[66] A. Milbrandt and M. Mann, “Potential for Hydrogen Production from Key Renewable

Resources in the United States,” NREL, Sprinfield , USA, 2007.

[67] ESRI, “ArcGIS for Desctrop,” 2015. [Online]. Available:

http://www.esri.com/software/arcgis/arcgis-for-desktop. [Accessed April 2015].

[68] M. Rumbayan, A. Abudureyimu and K. Nagasaka, “Mapping of Solar Energy Potential

in Indonesia using artificial neural network and geographical information system,”

Renewable and Sustainable Energy Reviews, vol. 16, no. 2016, pp. 1437-1449, 2012.

[69] xe, “The World's Trusted Currency Authority,” 2015. [Online]. Available:

http://www.xe.com/. [Accessed November 2015].

None PS MS SH TertiaryLPG 2 3 4 23 35Firewood 14 0 13 1 0Charcoal 12 6 24 15 3

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Number of Participants

Main Fuel Use per Educational Attainment

Figure 1: Main fuel use per number of participants and percentage according to the educational attainment

AKATSI KETA SOGAKOPELPG 37 45 0Firewood 9 18 32Charcoal 40 41 20

0%10%20%30%40%50%60%70%80%90%

100%

Number of Participants

Total Fuel Use per Area

Figure 2: Total fuel use per number of participants and percentage for each area

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0-200 200-400 400-600 600-800 800-1000GHS per month income

Total fuel use per incomeCharcoal Firewood LPG

Figure 3: Trends in fuel use according to the household income

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Charcoal Firewood LPG

Perc

enta

ge fo

r eac

h fu

el c

ateg

ory

Fuel

Reasons to switch to hydrogen cooking gas

Fuels cost increasing Safer Interested in the new technology

Figure 4: Reasons to switch to hydrogen gas of participants per category

Figure 5: Ghana case study TRNSYS print screen

0

50

100

150

200

250

300

350

400

450

500

550

600

650

700

750

800

0

5

10

15

20

25

30

35

40

45

50

55

60

65

70

75

80

1 2 3 4 5 6 7 8 9 10 11 12

(MW

h)

(kg

of h

ydro

gen)

Months

Monthly Demand Production Corresponding to DemandProduction Corresponding to Demand after 20 years MH TankAverage PV Energy Average PV Energy after 20years

Figure 6: Monthly demand and production of hydrogen and PV yield for the current conditions and after the panel

degradation period for the Ghana case study

050100150200250300350400450500550600650700750800

05

101520253035404550556065707580

1 2 3 4 5 6 7 8 9 10 11 12

(MW

h)

(kg

of h

ydro

gen)

Months

Monthly Demand Production Corresponding to DemandProduction Corresponding to Demand after 20 years MH TankAverage PV Energy Average PV Energy after 20years

Figure 7: Monthly demand and production of hydrogen and PV yield for the current conditions and after the panel

degradation period for the Jamaica case study

050100150200250300350400450500550600650700750800

05

101520253035404550556065707580

1 2 3 4 5 6 7 8 9 10 11 12

(MW

h)

(kg

of h

ydro

gen)

Months

Monthly Demand Production Corresponding to DemandProduction Corresponding to Demand after 20 years MH TankAverage PV Energy Average PV Energy after 20years

Figure 8: Monthly demand and production of hydrogen and PV yield for the current conditions and after the panel

degradation period for the Indonesia case study

Figure 9: Comparison between the CDF curves for GHR, temperature, relative humidity and wind speed for all years

Figure 10: Solar hydrogen potential map in Ghana

Figure 11: Solar hydrogen potential map in Jamaica

Figure 12: Solar hydrogen potential map in Indonesia

Table 1, Type and costs of standard fuel purchased units, conversion source: [69]

Fuel Purchas

e Unit

Weigh

t (kg)

Average

Cost of

Purchas

e Unit

(GHS)

Averag

e Cost

per kg

(GHS

per kg)

Average

Cost of

Purchas

e Unit

(US$)

Averag

e Cost

per kg

(US$

per kg)

Average

Cost of

Purchas

e Unit

(€)

Averag

e Cost

per kg

(€ per

kg)

Charcoal

Small

Bag

10 17 1.70 4.4 0.44 4.1 0.41

Medium

Bag

20 26 1.30 6.8 0.34 6.2 0.31

Large

Bag

50 65 1.30 16.9 0.34 15.6 0.31

LPG

Small

cylinder

6 20 3.33 5.2 0.87 4.8 0.80

Medium

cylinder

13 45 3.45 11.7 0.90 10.8 0.83

Large

cylinder

15 50 3.33 13.0 0.87 12.0 0.80

Firewoo

d

1 tie 18 5 0.28 1.3 0.07 1.2 0.07

Table 2, Fuel costs per kWh for the total fuel use of each main fuel and total cost for the average household,

conversion source: [69]

Fuel Cost of

fuel

(GHS/kW

h) Case

Study

Data

Cooking fuel

consumption to

satisfy

demand(kWh/ye

ar)

Cooking fuel

consumption to

satisfy

demand(kg/yea

r)

Annual

cost per

househol

d (GHS)

Annual

cost per

househol

d (US$)

Annual

cost per

househol

d (€)

Firewoo

d

0.08 6518 1671.3 521.4 135.6 125.1

Charcoa

l

0.14 5070 596.5 709.8 184.5 170.4

LPG 0.26 2028 156 527.3 137.1 126.6

Table 3, Annual emissions that correspond to the used fuels

Fuel Annual (tonnes of

fuel)

Annual Emissions (tonnes of

CO2)

Annual Emissions (tonnes

of CO)

LPG 6.07 34.1 0.3

Charcoal 29.72 472.9 54.6

Firewood 109.88 1479.5 123.4

Total 145.67 1986.5 178.3

Table 4, Distribution of households by source and use of energy for cooking, source: [22]

% of Households

Electricity 1.3

Kerosene 0.4

LPG 86.0

Charcoal 5.2

Firewood 7.1

Solar 0

Total 100

Table 5, Size, Price and Durability of LPG Cylinders in Jamaica, source: [22]

Cylinder

Size

Means

Urban Rural All

Unit Price

(JM$)

Weeks

Lasted

Unit Price

(JM$)

Weeks

Lasted

Unit Price

(JM$)

Weeks

Lasted

20lb 949.5 6.85 1101.82 6.90 1003.55 6.87

25lb 1144.66 7.71 1154.54 7.56 1149.27 7.64

30lb 1236.92 8.74 1297.89 7.31 1268.06 8.01

100lb 3744.17 22.97 3911.78 21.64 3802.64 22.53

Table 6, PEM electrolysers’ parameters in the case studies (per electrolyser)

Maximum hydrogen production rate (Nm3/h) 2

Number of stacks 4

Number of cells 20

Cell area (cm2) 92

Stack current (A) 110

Maximum stack power (kW) 10.25

Hydrogen generation pressure (bar) 13.8

Stack efficiency 63.6%

Table 7, Ghana case study hydrogen production and PV yield at current and future conditions

Annual hydrogen demand (kg) 785

Annual hydrogen production (kg) at current

conditions

815

Annual hydrogen production (kg) at the end of

20 years

783.4

Monthly average PV energy (MWh) at current

conditions

601.1

Monthly average PV energy (MWh) at the end

of 20 years

559.2

Table 8, Jamaica case study hydrogen production and PV yield at current and future conditions

Annual hydrogen demand (kg) 621.6

Annual hydrogen production (kg) at current

conditions

653.0

Annual hydrogen production (kg) at the end of

20 years

619.6

Monthly average PV energy (MWh) at current

conditions

467.2

Monthly average PV energy (MWh) at the end

of 20 years

436.5

Table 9, Indonesia case study hydrogen production and PV yield at current and future conditions

Annual hydrogen demand (kg) 631.0

Annual hydrogen production (kg) at current

conditions

670.0

Annual hydrogen production (kg) at the end of

20 years

635.4

Monthly average PV energy (MWh) at current

conditions

452.4

Monthly average PV energy (MWh) at the end

of 20 years

422.2

Table 10, Total FS Statistics of recent weather data of Kingston compared with Meteonorm weather data

GHR (W/m2) T (°C) RH (%) WS (m/s) Weighting

Factor for

Each

Parameter

TOTAL

Weighted FS

2011 29.89 2.31 5.32 0.89 0.25 9.6

2012 40.46 1.95 5.33 0.69 0.25 12.1

2013 43.18 1.46 4.75 0.65 0.25 12.5

Table 11, Hydrogen generation results of the simulation with Meteonorm and 2013 weather data

Meteonorm hydrogen generation (kg) Recent weather hydrogen generation (kg)

Jan 47.27 47.01

Feb 51.50 49.49

Mar 56.28 52.93

Apr 61.27 57.50

May 54.16 56.14

June 55.57 60.92

July 63.25 61.58

Aug 55.67 59.01

Sep 57.12 47.71

Oct 53.65 54.01

Nov 49.89 48.17

Dec 47.27 52.61

Year 652.89 647.08

Table 12, Total country solar hydrogen potential indicators

Country Ghana Jamaica Indonesia

Average solar hydrogen potential (thousand

kg/km2/year)

140.5 139.2 133.5

Total area (km2) 238,535 10,991 1,904,569

Country annual solar hydrogen potential (million

kg/year)

33,514 1,530 254,260

Nomenclature

Symbols Subscripts

CD Average daily cooking demand (kWh/day) G Ghana

LPGDay Daily amount of LPG used by each household (kg)

J Jamaica

CDay Daily amount of charcoal used by each household (kg)

I Indonesia

FDay Daily amount of firewood used by each household (kg)

i day of the parameter reading in Finkelstein-Schafer statistics

CVLPG Calorific value of LPG, equal to13kWh/kg Abbreviations

CVC Calorific value of charcoal, equal to 8.5kWh/kg

LPG Liquefied Petroleum Gas

CVF Calorific value of firewood, equal to 3.9kWh/kg

TMY Typical Meteorological Year

CVK Calorific value of kerosene, equal to 12.7kWh/kg

FS Finkelstein-Schafer

LPGeff Efficiency of the gas burner, equal to 45% CDF Cumulative Distribution Function

Ceff Efficiency of the charcoal pot, equal to 18% PS Primary school

Feff Efficiency of the firewood stove, equal to 14%

MS Middle school

Keff Efficiency of the kerosene stove, equal to 35%

SH Senior high school

LPGcyl LPG cylinder size (kg) GHS Ghanaian Cedi

Kcyl Kerosene cylinder size (kg)

x Parameter in Finkelstein-Schafer statistics

N number of daily readings in a month

M Number of parameters for the weighted Finkelstein-Schafer statistic

WF Weighting factor for a specific weather parameter

SHpotential Normalised solar hydrogen potential (thousand kg/km2/year)

Solar

Solar potential ( kWh

m2 day)