Electrifying the slums: Renewable Energy in Payatas
The potential for renewable electrification has been largely overlooked in urban slums. This project therefore aims to evaluate the extent to which new infrastructure could supply a target of 75% of electricity demand from renewable sources by 2030 in Payatas, a slum in the Philippines. Primary meteorological and survey data is used to supplement secondary data to estimate demand in 2030 at 11MW, and to calculate the possible contribution to supply of four renewable electricity sources; wind, solar, hydroelectric and landfill biogas generation. A total 12.66±4.91MW could be generated with three renewable technologies: wind energy is an unviable option in Payatas due to low wind speeds, and therefore does not contribute to this figure. Landfill biogas could generate base-load of 0.99±0.06MW, for which the infrastructure is already in place. Distributed solar photovoltaic generation could meet peak demand, supplying 2.59±0.41MW of electricity at the point-of-use, alongside a run-of-river micro-hydroelectric scheme, the potential output of which is the most significant of all technologies analysed, at 9.08±4.89MW. Whether or not 75% of electricity demand (8.3MW) can be supplied renewably in 2030 depends largely on hydroelectric output, which is highly uncertain, and to which the model developed is most sensitive. The significant uncertainty of the study therefore limits its conclusions. Many factors beyond the scope of this study are also influential, thus further research into economic, political, and technical aspects of this study is recommended.
Development; climate change; energy security; renewable electricity generation; Philippines
1.3 billion people globally lack access to modern energy services (IEA, 2012), the majority in developing countries like the Philippines, where three million households are without electricity (Webb, 2013). Most of these are in remote rural areas or squatter-settlements. Payatas is one such squatter-settlement, located northeast of Manila in Quezon City, near the landfill site for the capital (Figures 3.1 and 3.2). Although 99.98% national electrification is seen at barangay level (DOE, 2012), household electrification was just 83% in 2011 (Asian Power, 2011). Household electrification is particularly low in Payatas, and poverty is also high relative to the national average (Smyser et al, 2004). Quantification of this is difficult because specific data is lacking, so this study may begin to address this problem.
Energy access and climate change are interlinked, and are significant global problems. It is widely documented that renewable electricity can contribute to climate change mitigation by reducing dependence on carbon-intensive fossil fuel electricity generation (e.g. Ueckerdt et al, 2011), as well as having wider social and economic benefits such as increased energy security and human health (IPCC, 2011). There is no energy-related Millennium Development Goal (other MDGs pertain to environmental sustainability, health and education), despite the relevance of energy to all of them (Brew-Hammond, 2012). Electrification is an effective method of addressing energy poverty, and facilitating social and economic development in developing economies and deprived areas (Winkler et al, 2011). Renewable electrification therefore combines the dual goals of climate change mitigation and social development.
This project will bring together previous work on renewable electrification and apply it to an urban area, where renewables have been largely overlooked. Although the scope of the project is limited by practicality, the conclusions will be notable, as the potential for renewable electrification in Payatas has never been studied. Filipino government policy, like the introduction of Feed-in Tariffs in 2012, is particularly relevant to the four technologies assessed in this paper: wind, solar, biogas and hydroelectricity. Further research into factors beyond the remit of this study, such as political and economic themes, would be invaluable to assess other important facets of this interdisciplinary topic.
It is in this context that the study aims to quantify the extent to which Payatas can become electrified by 2030, and whether 75% of electricity can come from renewable sources. The Filipino Department of Energy (DOE) plans to increase renewable capacity nationally by ~50% by 2030 (Webb, 2013), but 75% is a realistic target for Payatas in light of the wealth of local resources, and the potential for development presented by the low current level of (fossil fuel) electrification.
To achieve this aim, four objectives were formulated:
- Quantify changes in electricity demand by 2030
- Calculate the potential contribution of each renewable resource to supply. Four renewable resources are considered; wind, solar, biogas and hydroelectricity.
- Develop a simple model to analyse the primary and secondary data on each of these sources, and perform simple statistical uncertainty and sensitivity analyses to further dissect the results.
- Synthesise this information and assess the feasibility of the aim to supply 75% of electricity renewably by 2030.
Each section will look specifically at each of the four renewable technologies in turn, as well as at other relevant factors. An analysis of the existing literature on the subject will be presented alongside the results of the data collection for context. Uncertainty and sensitivity will be evaluated and all relevant information synthesised to assess the feasibility of realising the 75% target, as stated in the aims and objectives.
2. Literature Review
There is a fairly wide body of literature about slum electrification (e.g. Malyshev, 2009; Karekezi et al, 2008; Scott et al, 2005), despite limited data on energy use in slums (Malyshev, 2009), and there is broad consensus that development of electricity infrastructure has largely positive impacts. However, a detailed scientific study on Payatas is lacking, as is a study concentrating on renewable electrification as opposed to the extension of traditional fossil fuel-based provision in slums. The majority of literature about renewable electrification is restricted to remote rural areas and decentralised renewable energy systems (e.g. Yadoo & Cruickshank, 2012; Deshmukh, 2009; Ferrer-Marti et al, 2011), or focuses on development outcomes.
This study seeks to synthesise previous work on renewable electrification and apply it to an urban area, where renewables have been largely overlooked. Moreover, these will be considered in tandem rather than separately, as is often the case in the wider literature. A more detailed literature review can be found in Kaundinya et al (2009). This research is relevant to the Filipino Department of Energy’s targets of increasing renewable energy generation by 100% by the end of 2013 and to increase household electrification to 90% by 2017, as well as the introduction of Feed-in Tariffs in 2012 (DOE, 2011).
It has been shown, for instance by Parikh et al (2012), that improving access of slum residents to electricity can have social benefits, such as improving aspirations. Lighting can extend working and studying hours, and increase public health, and safety (Ahmed & Menzies, 2012; Scott et al, 2005). Many scholars assert that renewable distributed generation is an effective method of electricity provision (Bayod-Rújula, 2009), minimising transmission losses, enabling the development of ‘smart’ grids (Ochoa & Harrison, 2011), and enhancing community engagement (Wolsink, 2012).
Renewable technologies (particularly large-scale hydroelectricity) have negative environmental impacts like flooding or chemical contamination (Boyle, 2004). However, these are minimal relative to the destruction caused by fossil fuel extraction, refining and combustion, which contribute to the enhanced greenhouse effect (Akella et al, 2009; Panwar et al, 2011; Ramanathan & Feng, 2009). Although some scientists believe the development of geo-engineering technologies such as carbon capture and storage lessens the urgency of developing low-carbon electricity sources (Lackner et al, 2012), they are not a “silver bullet” solution as detailed in Kreigler et al (2013). Decarbonising electricity generation is therefore essential, particularly in the long-term (Myhrvold & Caldeira, 2012). Furthermore, the development of renewables reduces the need for imports and increases energy security, which is especially pertinent for island states like the Philippines (Zafirakis & Chalvatzis, 2014), which imported 41% of its energy in 2011 (OECD/IEA, 2013)
Problems of legality, security of tenure and barriers to access, seen in Brazil (Coelho & Goldemberg, 2013), India (Baruah, 2012) and Sub-Saharan African countries (Rosnes & Vennemo, 2012), are important in any investigation of slum electrification, but a detailed examination is beyond the scope of this study. The assessment of slum electrification in Payatas by Scott et al (2005) is a more in-depth analysis, investigating electrification in a developmental context. They conclude that slum electrification is realistically achievable, evidently realising similar objectives to those of this study, which has a scientific rather than developmental focus. Although not quantitative, this is one of the most relevant studies to this project, and despite the considerable development that has occurred in Payatas, rough similarities can be seen between 2005 and 2013.
Ostojic et al (2013) present options for increasing the share of renewable electricity in Cebu, Philippines. Although significantly larger than Payatas, Cebu shares characteristics such as rapid economic growth, considerable unharnessed renewable potential, and being Filipino. This makes their conclusions applicable to Payatas, and some options, such as electricity generation from landfill biogas, extremely relevant.
However, the methodology used in Ostojic et al (2013) relies on bottom-up greenhouse gas (GHG) emissions inventories, which have been shown to be inaccurate compared with top-down estimates based on atmospheric observations (e.g. Miller et al, 2013; Nisbet & Weiss, 2010; Weiss & Prinn, 2011; Levin et al, 2010), suggesting their conclusions may not be based on the most accurate data.
Evidently, robust methodologies are essential in a study such as this. Methodologies for assessing resource potential are thus adapted from Kubik et al (2011), and Petersen et al (1998), which evaluate different metrics and equations to estimate wind power output, and Andrews & Jelley (2007) and Boyle (2004) to estimate output from other sources. These studies and books are more comprehensive, and involve larger datasets than this one, but the same principles are used. Numerous other studies of renewable resource potential also utilise the comprehensive NASA SSE/POWER databases, such as Lalwani et al (2012) and Ghasemi et al (2013), attesting to its reliability.
These methodologies are combined with data from geographically or contextually relevant studies such as Scott et al (2005) and Elliott et al (2001), which is a comprehensive assessment of wind characteristics in the Philippines. Through using robust, widely used methodologies and data, as well as primary data of its own, this study will contribute to scientific research in an area that has been previously neglected. It will begin to address some of the gaps in the literature by studying Payatas in depth, and by considering renewable electrification in a slum, which has not previously been done from a scientific perspective.
3. Study Design
To test the hypothesis that 75% of Payatas’s 2030 electricity demand can be supplied renewably, I collected primary and secondary data. Primary household data were collected in Payatas, while primary meteorological data were collected at a site in nearby Montalban, both shown in Figure 3.1. Secondary meteorological data were obtained from the Science Garden station in Quezon City, which is the closest SYNOP station to the study site, also shown in Figure 3.1. Primary data is used to supplement and contextualise secondary data, and provide additional depth where secondary data has a low temporal or spatial resolution. Qualitative data were used to guide the research, while quantitative data were used to realise the objectives and formulate conclusions.
3.1 Study Site
Payatas is a barangay in Quezon City, Luzon, Philippines, located at 14.65N 121.07E (Figure 3.1), and is a squatter-settlement built around a landfill site. It is divided into Payatas A and B, and further into four ‘phases’, as shown in Figure 3.2. The area was involved in the Depressed Areas Electrification Program 1990-92, but electricity access is still low on household level. Poverty is high relative to the national average (Smyser et al, 2004), though significant economic variation exists within the barangay. The 2010 census records 119,053 residents, and an average household size of 4.4 (NSO, 2012a).
Figure 3.1 Satellite imagery of the Metro Manila area, Philippines. Source: Google Earth/NOAA (2013)
Figure 3.2 Boundaries of Payatas, Quezon City, Philippines: 14.65N 121.07E, including geographical divisions and key locations. Underlying imagery obtained from Google Earth (2010)
3.2 Study Period
Fieldwork was conducted from July 21st to August 13th 2013, with additional meteorological measurements taken by a colleague from August 15th to September 12th.
3.3 Sampling Methods
Primary data collected were household surveys and meteorological observations, as outlined in Table 3.1a. These data were primarily used in the first two objectives: to quantify electricity demand in Payatas, and to calculate the potential contribution of local renewable resources to future electricity supply. Qualitative and quantitative primary data were collected in situ, and secondary data on biogas, solar, hydroelectricity and wind potential were compiled June-December 2013. Secondary data were mainly used to investigate the feasibility of the stated aims (objective four), to contextualise the results, and to strengthen the conclusions of the study. The data sources and methods are summarised in Table 3.1a/b.
The majority of the 15 household surveys conducted were in the Empire area, and all were within Payatas B. This is due to the use of a chain (‘snowballing’) method of finding respondents, whereby participants referred other respondents. Questions focussed on electricity consumption, legality and security of supply, and perceptions of renewables (the full questionnaire is reprinted in Appendix 5). Interviews with officials at the Department of Energy (DOE) and Quezon City Biogas plant were conducted by arrangement, and were largely technical.
Table 3.1a Primary data methodology
|Wind Speed (m/s)||– Windtronic 2 anemometer- 3.7m from surface- Measured at Montalban (Figure 3.1)- Measurements averaged over 3 minutes twice daily throughout the study period.- Log law derivation used to extrapolate to hub height|
|Cloud Cover (Oktas)||– Cloud cover estimated twice daily|
|River flow speed||– Primary data unavailable due to flooding|
|Roof aspect (see below)||– Selected random sample of dwellings along transect (Empire Rd.). Roof aspect, tilt and over-shading measured with a compass, and compared with secondary data from Google Earth (Appendix 1)- Accuracy of secondary assessment estimated and incorporated into calculations of solar output|
|Electricity consumption and demand data
||– Household surveys conducted with willing respondents,- ‘chain’ method used to find respondents|
Table 3.1b Secondary data sources
|Wind speed (monthly averages/ instantaneous m/s)||Elliott et al (2001)NASA Prediction of Worldwide Energy Resource (POWER)SYNOP Science Garden|
|Roof Tilt||Google Earth (Appendix 1)|
|Roof Aspect (% of each orientation)|
|Roof Over-shading (%)|
|Electricity Prices (PHP /month or /kWh)||MERALCO|
|‘Lifeline’ Tariffs (PHP)||MERALCODOE 21st EPIRA reportScott et al (2005)|
|Solar Irradiance (kWh/m2/day)||University of LowellNASA Surface meteorology and Solar Energy (SSE)|
|Biogas plant output (kWh) and capacity (kW)||Quezon City Biogas ProjectPangea Green Energy Philippines|
|Marikina River hydrological cross-sections (GIS shapefiles)||Personal Communication, Santillan (2013)Badilla (2008)|
|Hydrological flow data (m/s)||Badilla (2008)Tachikawa et al (2004)|
|Fuel Mix and expansion plans (qualitative)||DOE21st/13th EPIRA reports (2012) (2006)|
|Demographic data||Filipino NSO Census (2000) (2004) (2010)|
|DOE interviews (Qualitative)||Personal Communication, F.R. Domingo Jr.; R. Yambao, 2013)|
3.4 Data Analysis
Data were manipulated using the equations detailed in Section 4, and analysed using simple sensitivity analysis in Microsoft Excel. Monte Carlo simulations were run to model wind speed frequency distributions, and uncertainties were tested using the same program. Exploration of data was done graphically and mathematically using spreadsheet models, and compiled to facilitate analyses such as feasibility and economic assessments. Limitations of the methods are discussed in Section 5.
To investigate whether 75% of electricity can be supplied renewably by 2030, current and future electricity demand was investigated, and the potential contributions of four renewable technologies were assessed. The following sections relate to objectives one and two.
Current electricity demand was calculated in order to estimate future increases and thus required electricity supply. Average monthly consumption in Payatas was calculated from household survey data and compared with estimates of peak demand from DOE interviews. Expenditure data were used where consumption data were unavailable (Table 4.1). Average household electricity expenditure in Payatas is PHP1291/month, 1.5 times the value in Scott et al (2005) (PHP832/month), perhaps due to the smaller sample size (n=15), increased affluence by 2013, or the inclusion of small businesses, which use more electricity. Mean consumption is 163kWh/month (~2000kWh/yr), and most residents use 101-250kWh/month (Figure 4.1).
Consumption follows an approximately normal distribution around the mean, as expected, despite the small sample size of 15. The majority of demand is for lighting, cooling and entertainment (Table 4.2).
Table 4.1 Calculation of consumption from monthly expenditure
|Household||Expenditure(PHP/month)||Price/kWh(PHP)||Consumption(Calculated kWh)||Consumption (Actual kWh)||Error||Error (% Actual)|
Table 4.1 calculates consumption from expenditure data for each household surveyed, and compares it with actual consumption data, where available, to find the error associated with the method. Error is <11% in those cases where it was possible to determine. The price per kWh differs depending on usage because there are stepped subsidies in place. Further details, including subsidy structure and billing adjustments, are in Appendix 3.
Table 4.2 Common appliances used in Payatas, in descending order of importance.
Mobile phone charger
Current system peak demand for Payatas is ~150-200W per household (F. R. Domingo Jr., personal communication), meaning that system peak demand for 27,057 households is ~5.5 MW. Energy demand is expected to double nationally, and regionally (DOE, 2010), by 2030 on a 2007/09 baseline with the most pronounced growth in the residential sector (6.5%/yr) due to rising affluence and population growth (Mukherjee & Sovacool, 2012). As Payatas is ~88% residential, this is especially pertinent, meaning peak demand could reach 11MW by 2030. Any electricity provision should therefore be designed to provide 8.3MW of electricity (75% of peak demand), in keeping with the aims of this study. This figure is henceforth used to compare against potential output of each resource.
1.2.1. In Situ and Ex Situ measurements
Wind measurements obtained from in situ measurements at Montalban (hereafter PM) are consistently lower than those derived from Science Garden (hereafter SG), even after adjustment to 10m. Both datasets present lower values than NASA monthly averages (hereafter POWER), as shown in Figure 4.2. This could be due to differences in data collection methods – SG data follows internationally recognised SYNOP collection standards, whereas PM methodology was less rigorous due to limited facilities. Conversely, POWER data is derived from an atmospheric model, constrained by satellite observations, which may differ from on-Earth measurements, and atmospheric soundings (OpenEI, 2013).
There is significant variation in the PM and SG observations. This is not likely due to diurnal variations, but because the dataset is limited this cannot be disproven. Wind speeds were not recorded 3rd to 11th August, so data cannot be compared during this time period.
The precision of SG data (1.0m/s increments) is an order of magnitude larger than that of PM data (0.1m/s increments), as seen in Figure 4.2. This means extrapolation to hub heights, and thus wind power calculations, are less accurate and precise than those based on PM data. However, the small size of the PM dataset limits the conclusions that can be drawn because there are fewer data points to compare.
Figure 4.2 Wind Speed measurements at 10m. Blue line: POWER 22-year monthly average, green line: SG SYNOP (instantaneous) measurements, red line: PM data (3-minute averages)
Wind speeds are not predicted to change dramatically by 2030 (Stocker et al, 2013), so similar wind values are assumed to 2030.
4.2.2. Statistical wind speed frequency distribution
Wind speeds are expected to follow an approximately normal distribution around a mean. Wind speed frequency distributions (Figure 4.3) were created from the mean values of the SG and PM data, as well as that of Elliot et al (2001) using Monte Carlo simulations. The model was run to 1000 iterations to remove any bias from the original data, and predicts the frequency distribution of wind speeds. From the PM and SG datasets, wind speeds are statistically unlikely to frequently exceed ~7m/s, and are most often <4m/s, which is also reflected in the data. This figure correlates well with the mean wind speed estimates (3.0m/s) of Elliot et al (2001), and shows wind speeds may often be lower than cut-in speeds (see below).
Figure 4.3 Frequency distributions of wind speeds with mean value a) 0.33 m/s (PM mean), b) 0.99m/s (SG mean), and c) 3.0m/s (Elliot et al, 2001), all generated randomly in Excel using Monte Carlo simulations to 4 standard deviations.
4.2.3. Wind power output
Wind speeds from PM and SG data were extrapolated to hub heights of 65m, 80m and 100m for the Suzlon S64/950, Vestas V90 and Repower MM92 turbines respectively (see Table 4.3) using Eq. 4.1, derived from the log wind profile equation, assuming surface roughness of 1.0 (average for a “city”; Kubik et al, 2011).
Table 4.3 Characteristics of three turbine models used
|Turbine||Hub Height||Blade Length||Cut-in Speed||Rated Speed|
|Suzlon 950||65m||32m||3.5 m/s||11 m/s|
|Vestas V90||80m||45m||3.5 m/s||11 m/s|
|Repower MM92||100m||46.25m||3.0 m/s||11.2 m/s|
Eq. 4.1 Derivation of Log wind profile equation, allowing calculation of wind speed u2 at height z2 from known values of u1 at height z1, using value of surface roughness length z0, here assumed to be 1.0 (Kubik et al, 2011). All speeds measured in m/s and heights/lengths in m.
Wind power output was calculated using Eq. 4.2 assuming fixed pitch turbine blades, and Cp = 0.35; the theoretical maximum, Cp,max, is 0.59 (Andrews & Jelley, 2007) so this may be a conservative estimate.
Eq. 4.2 Kinetic energy in the wind (W), where A = total swept area (πr2) (m2), ρ = air density (1.2kg/m3), u = wind speed (m/s), cp = wind power coefficient. Cp,max is determined by the Lancaster-Betz limit, 16/27 or ~0.59
Cut-in speeds (at which turbines start operating), obtained from turbine rating curves (Figures 4.4 and 4.5) were compared against the results from both datasets. Rated speeds (maximum operating speeds) and half-rated speeds (half of the rated speeds, which are theoretical maxima and thus infrequently observed) were obtained from the manufacturers. The results are displayed in Table 4.4.
Figure 4.4 Power rating curve for Vestas V90 – 1.8MW turbine. Source: Vestas (undated)
Figure 4.5 Power rating curve for Repower MM92 – 2MW turbine. Source: Repower (undated)
Wind speeds above cut-in speed were observed in 26.44% of cases based on adjusted SG measurements for all turbines. Cut-in speed was observed in 3.85% of cases for the Suzlon S64/950, and 7.69% of data points for the Repower MM92 based on adjusted PM data. Cut-in speed was not achieved for the Vestas machine, probably due to its slightly higher value (3.5 m/s).
Table 4.4 How often wind speeds exceed cut-in speed, half-rated speed, and rated speed at hub height for each given turbine size and model, based on PM and SG data
|Turbine||>Cut-in Speed(% data points)||>Half-power rating (% data points)||>Rated speed (% data points)|
|Primary||Science Garden||Primary||Science Garden||Primary||Science Garden|
A power-rating curve for the Suzlon S64/950 was unavailable, so the required wind-speed for half-power generation was assumed to be 8.0m/s, similar to those of the Vestas V90 and Repower MM92 turbines. Based on adjusted PM and SG data, half-rated wind speeds were not observed for any turbine. Rated speed or above was therefore also not seen for any turbine based on either dataset. Available generation capacity, when above cut-in speed, would be low, for all turbine sizes. Only utility-scale turbines were investigated, as they are more economically and energetically efficient. Mean output calculated from SG/PM data with Eq. 4.2 was 72kW, making wind an unviable option for electricity generation.
The Pangea Green Energy Philippines (PGEP) Quezon City Biogas project currently operates at 840kW installed capacity at a load factor of 80-90% (PGEP, personal communication), including ~20kW provided to the municipality for street lighting and local development projects. This means it produces 6.25GWh/yr, with an average output of 712kW. Provided the municipality extends the contract, by 2030 an additional 320kW generator can be added, increasing capacity to 1.16MW (0.99MW output at 85% load factor). The low variability of supply means biogas could provide a reliable base load of 11.9% of total electricity required in Payatas by 2030, particularly as generators are regularly maintained, reducing the number of unplanned outages.
4.4.1. Marikina River Scheme
Hydroelectric potential was calculated for the nearby Marikina River using Eq. 4.3.
Eq. 4.3 Hydroelectric power output equation, where η = turbine efficiency (0.85 as stipulated by Andrews & Jelley, 2007), ρ = water density (1000kg/m3), g = acceleration due to gravity (9.81m/s2), h = head height (m), Q = flow rate (m/s)
A run-of-river scheme with minimal storage capacity was assumed rather than an impoundment system, which requires a dam and reservoir, because it would be more suitable in a mountainous area like the Marikina River catchment, though the variable flow rate could present problems. A schematic is shown in Figure 4.6.
Flow rates of 102.95m/s, 1487.90m/s & 6.39m/s (mean, maximum and minimum respectively) were taken from Tachikawa et al (2004). It was assumed that these would not change greatly by 2030. These were combined with estimated head heights, calculated from elevation data from Santillo et al (2012), cross-referenced with elevation profiles from Tachikawa et al (2004) and Google Earth maps, to derive a maximum possible head height along the Marikina River of 60m. Such heights would require large-scale flooding and would not be feasible, but were included for comparison.
Potential modelled hydroelectric output varies from 0.56-788MW, though a more realistic figure is 9.09MW±4.89MW. Figure 4.7 shows that output increases with head height for all three flow rates used in the analysis, as expected from Eq. 4.3. As absolute head heights are smaller than flow rates, increasing the flow rate has a greater effect than increasing head height, though the equation is mathematically equally sensitive to both parameters.
Figure 4.7 Hydroelectric output as a function of head height for minimum, maximum and mean flow values obtained from Tachikawa et al (2004) (see main text). The inset shows a close-up of the mean and minimum lines on the same axes.
4.4.2. La Mesa hydroelectric scheme
The potential for hydroelectricity in the La Mesa/Novaliches reservoir was assessed using Eq. 4.3, assuming a manufactured head of 10m and flow rate calculated from the recharge rate and watershed size taken from Tachikawa et al (2004). A lower value of η (0.8) corrects for transmission losses, frictional drag and turbulence (Sørensen, 2004).
The low flow rate (2.35m/s) and low head produced a maximum output of 185kW, making the proposal unviable. As the potential output of a run-of-river scheme is comparably high, this has little effect on the overall hydroelectric potential. For details see Appendix 2.
4.5.1. Roof area
Google Earth (GE) roof surveys estimated there to be 44,434m2 of potentially suitable roof area in 2013. In situ corroboration studies indicated that estimates were correct in ~93% of cases.
Figure 4.8 shows the proportions of roof orientations in Payatas, where the majority (53%) are flat. Over-shading, which reduces the total suitable area by 16.6%, is incorporated into adjustments in Eq. 4.8.
Figure 4.8 Roof orientation by percentage in Payatas. Dark turquoise: East/West-facing, dark blue: flat (i.e. no orientation), light blue: South-facing, grey-blue: North-facing. Percentages are rounded and as such do not total 100%.
Population growth estimates were used to estimate the useful roof area in 2030 for solar panels. The same ratios of aspects, tilts and over-shading from the GE roof survey were assumed.
Table 4.5 shows the area of roofs at different orientations and tilt, and the associated predicted output. A projected (medium) population growth rate of 34.84% (UN, 2011; cited in Racelis et al, 2012) was applied to the 2010 population of 119,053 (NSO, 2012), and falling household size, from 4.4 in 2010 to an estimated 4.0 in 2030, was considered. The number of households could rise from 27,687 in 2010 to 40,133 in 2030, thus increasing the potential roof area on which to place photovoltaic panels to ~64,385±4500m2.
Table 4.5 Roof characteristics and potential solar photovoltaic output for roofs in Payatas in 2010 and 2030. Optimum (highest incident sunlight) combinations are shaded. Values are shown to one decimal place.
|Orientation||Tilt (degrees)||Area (m2)||kWp||Incident Sunlight (kWh/m2/day)||Output (GWh/yr)|
4.5.2. Solar radiation
Solar potential was calculated using data from the University of Lowell and NASA. Solar irradiance at different orientations was calculated using Eq. 4.4 and Eq. 4.5, with the average annual radiation value adjusted for 10° and 20° tilt using Eq. 4.6.
Eq. 4.4 Radiation values for PV module facing the equator, where Sincident = average radiation value for Payatas (14.65 N 121.07 E) (1664.4 kWh/m2/yr), α = elevation angle (°), given by Eq. 4.6, β = module tilt angle (°), measured from the horizontal.
Eq. 4.4 assumes that the module is facing the equator (South in Payatas) to maximise the amount of energy received from the sun. To calculate the potential output of modules on East-/West-facing roofs, Eq. 4.5 was used:
Eq. 4.5 Radiation values for PV module at arbitrary tilt, where Sincident, α and β are as above, Θ = sun azimuth angle (°), ψ = azimuth angle that the module faces (°).
Eq. 4.6 Elevation angle, where ϕ = latitude (14.65 N) and δ = declination (°), given by Eq. 4.7
Eq. 4.7 Declination angle, where d = day of the year, with January 1st taken to be 1. The median day in July; the 16th, was used; d = 197
Maximum radiation was received on a flat roof (1664.4kWh/m2/yr), and minimum on an East-facing roof at 20° elevation (1422.8kWh/m2/yr), as shown in Table 4.5.
4.5.3. Potential photovoltaic output
The potential output of the total area estimated to be suitable for photovoltaic modules was calculated using Eq. 4.8.
Eq. 4.8 Energy output of solar panels adapted from Photovoltaic Software (2012), where A = panel area (m2), r = solar panel yield (%): ratio of one panel’s power (kWp), divided by its area, H = Average annual radiation (kWh/m2/yr), η = performance ratio (0.6) including factors such as over-shading (16.6% total area) and cloud cover (5.5%) derived from Google Earth survey, and losses due to e.g. temperature and transmission.
The performance ratio, η = 0.6, includes inverter losses (0.08), temperature losses (0.08), DC and AC cable losses (0.02 each), dust losses (0.02) and losses from over-shading and cloud cover (0.22) estimated from Google Earth roof survey.
Table 4.5 shows the potential output of each roof type.
Run-of-river hydroelectricity could generate 9.08±4.89MW, while potential for a reservoir scheme and wind is so low (185kW and 72kW mean, respectively) that they are disregarded. Solar and landfill biogas electricity generation represent 33% and 12%, respectively, of the target 8.3MW, a figure derived from household surveys and interviews with DOE representatives. A total of 12.66±4.91MW could be generated using these three technologies, making the study’s aims potentially achievable, though uncertain. Figure 4.9 summarises these findings.
Figure 4.9a Potential output of each renewable electricity source (MW), including uncertainty ranges.
Figure 4.9b Solar, Wind and Biogas on an enlarged scale. Colours are as above
The sample size of the surveys used to estimate demand was very small (n=15), which somewhat limits the conclusions drawn, but the qualitative data obtained guided the rest of the research. Preliminary results, showing widespread public support of community renewable energy, are supported in the literature on the subject (e.g. Rogers et al, 2012; Warren & McFadyen, 2010; Wolsink, 2012). In addition, the mean household electricity consumption (Figure 4.1) is similar to figures presented in similar studies, such as Scott et al (2005), suggesting that the conclusions of the survey are generally accurate.
The inclusion of subsidy rates in the calculation of consumption from expenditure confused the data because prices per unit vary depending on subsidies received, and how many. Expenditure was frequently the only data respondents were able to provide, however, and respondents were often unsure about subsidies they received. The model developed to calculate consumption from expenditure is shown in Appendix 3, and the results are presented in Table 4.1. The error of this method was calculated from households that provided comparable consumption and expenditure data, and was estimated to be <11%, deemed acceptable for the purposes of this study
The error bounds of the uncertainty analysis (discussed further in Section 5.6.1) were used to generate scenarios of high, medium and low output, illustrated in Figure 5.1.
Figure 5.1 Projected output of renewable technologies in 2030 (GWh), using different scenarios based on the mean, minima and maxima of uncertainty boundaries specified in Section 5.7.1, against projected demand in Payatas for 2030.
Total electricity demand for Payatas in 2030 is estimated at 96.97GWh. Using the mean predicted output for each renewable electricity source (wind is not included as it is unviable), output is predicted to reach 110.87GWh, which is greater than 100% of predicted demand, and allows a reserve margin of 14.33%. This is lower than the reserve margin for Luzon observed in 2009 (16.41%) and projected for 2030 (18.34%) (DOE, 2009). However, the target (75% of total demand) set by this study would be met even if the additional 3.88GWh required reserve were generated with grid capacity.
Using the maximum projected capacity, the reserve margin rises to 57.67% and this problem is eliminated. However, under the minimum projected output scenario, the amount of electricity generated is less than the projected demand (69.54GWh), leading to a deficit in production of 27.44GWh. This figure is 71.7% of projected demand and therefore fails to meet the target of 75% renewable electricity generation in 2030. This energy gap would need to be filled using fossil fuel-based electricity generation via the national grid.
Whether or not the hypothesis that 75% electricity (8.3MW) can be met with renewables by 2030 is true depends on the level of uncertainty, and accurately quantifying output. This is discussed further in Section 5.6.1.
5.2.1. Data manipulation
A derivation of the log profile equation was used to extrapolate measured wind speeds up to hub height. Equations and coefficients were adapted from Kubik et al (2011). The log wind profile (Eq. 5.1) and power law (Eq. 5.2) equations perform roughly as well as each other on average, though in practice the Power Law yielded estimates that were inconsistent with studies like Elliott et al (2001). Furthermore, the wind shear coefficient, α, is a more sensitive parameter than surface roughness, z0,which means that errors in α would be more significant in the final calculations.
Eq. 5.1 Log wind profile equation, allowing extrapolation up to hub height based on two wind speeds, where k = von Karman’s constant (~0.4), u* = friction velocity (m/s), ūz = mean wind speed (m/s) at height z (m), and z0 = surface roughness (m)
Eq. 5.2 Empirical Power Law equation, where u and z values are as in Eq. 4.1, and α = 1.43 as per the 1/7th power law (Kubik et al, 2011)
The full log wind profile equation was not used because there was not enough data to estimate the profile properly, or to correct for atmospheric stability. Eq. 4.1 was used instead; complications regarding the calculation of friction velocity u* were also avoided, and the profile was calculated from measurements at just one height.
5.2.2. Potential wind power output
Many studies, such as Elliott et al (2001) use the Weibull probability distribution function (Weibull, 1951) as an alternative to analysis using Monte Carlo simulations, which this study has employed (Section 4.2.2).
Elliott et al also use Wind Power Density (WPD), which can be a better indication than wind speeds of the true potential at a particular site. Although WPD was calculated from PM and SG data, inherent inaccuracies and imprecision were amplified. Furthermore, air density was assumed to be constant, while in truth it can vary seasonally by 10-15% (ibid) because air density is related to temperature and pressure as per the ideal gas law.
Figure 5.2 Output of wind turbine (kW) as a function of wind speed at hub height (m/s) derived from adapted SG SYNOP data. Three turbine models are shown; Blue line: Suzlon S64/950 (65m hub, 32m blades), Red line: Vestas V90 (80m hub, 45m blades), Green line: Repower MM92 (100m hub, 46.25m blades).
Higher wind speeds and a larger blade swept area produce higher potential wind output. Despite the 20m difference in hub height between the Vestas V90 and Repower MM92, output per m/s wind speed is only fractionally higher for the larger turbine, as shown in Figure 5.2. This is because output (Eq. 4.2) depends more on swept area than hub height. The MM92’s swept area is only 5.6% larger than the V90’s because its blade length is only 1.25m longer.
Even maximum potential output of 610kW (at 100m hub height and 7.56m/s adjusted wind speed) would not be great enough to justify the costs of installation, which are significant. Timilsina et al (2013) use data from NEA/IEA (2010) and Lazard (2011) to estimate the Levelised Cost of Electricity (LCOE) of utility scale onshore wind turbines to be between £313,045 – £1,441,088/MW installed nameplate capacity, meaning the mean projected output (2.3 – 76.8kW) would cost between £42.29 – £3130.78/MWh over an 22 year lifetime, nearly 250 times greater than the cost of the theoretical maximum output, £7.48/MWh (see Appendix 4).
5.2.4. Data analysis
Although underestimations of the wind resource could have been made because of the limitations in sampling, it is apparent that wind speeds are too low at any height considered to make wind an economically or practicably viable option for electricity generation in Payatas. The focus will hereafter be on biogas, hydroelectricity and solar.
Biogas could be used as a reliable base-load generator because the variation in output is very low (PGEP, personal communication). The 2030 output could be 8.64GWh, providing 8.93% of the estimated 96.97GWh projected 2030 usage, or 8.91% of the 11MW projected peak capacity. The sensitivity of the output to changing load factor is moderate (see Section 5.6.1), suggesting it is unlikely there will be significant deviation from this figure.
Furthermore, the infrastructure to generate electricity from landfill biogas is already in place, and would require little further development to increase the output. The scheme might struggle to remain viable if UNFCCC Clean Development Mechanism credits, around which the project is structured, were withdrawn, but this depends on economic factors like the export price of electricity and rising costs of fossil fuel generation.
Potential output of 9.09MW±4.89MW from a run-of-river scheme in the Marikina River would supply a large proportion of the projected electricity demand in Payatas, even assuming the lower end of the uncertainty range. Hydroelectric output is seasonally variable as it depends upon the hydrological cycle and meteorology (Poff et al, 1997), but this makes it reasonably predictable with accurate observation and modelling. Flow rates increase when precipitation is highest during August and September (Figure 5.3): in 2013 monthly precipitation was 614.0mm and 637.3mm, respectively (Mundomanz, 2013). Furthermore, the times when flow rates are lowest, during the hottest, driest months, is when solar output is highest, thus shoring the gap between supply and demand.
Although the data used by Tachikawa et al (2004) is from 1995, and absolute flow rates may have changed, the relative patterns of the flow regime in the Marikina River are unlikely to have altered. Unfortunately it was impossible to corroborate this data with either primary observations or data obtained from the Filipino climatic research department, PAGASA, so this uncertainty has simply been incorporated into estimates of potential output, as shown in Figure 4.9. Sensitivity is highest for hydroelectricity, the most uncertain resource, as detailed in Section 5.7. Changes in flow rate throughout the year may be greater than the uncertainty range specified, as discharge varies by up to 3000% from maximum to minimum, but output calculated from annual mean flow is likely to fall within the uncertainty bounds specified. Uncertainties in flow estimates could be reduced with accurate hydrological modelling such as is undertaken by Badilla (2008).
Figure 5.3 Daily variation in discharge (m3/s) in the Marikina River for the year 1995, showing 12-month mean, flow duration and variations in daily discharge throughout the year. Source: Tachikawa et al (2004).
5.5.1. Roof characteristics
As shown in Figure 4.8, flat roofs, which make up the majority in Payatas, yield the highest electricity output, potentially producing 6772.7kWp by 2030. However, this can be a misleading measure of output because it describes the maximum output possible under ideal conditions (Boyle, 2004), and is thus rarely reached. Table 5.1 shows the output per unit area of module for each roof type, which is a more realistic estimate of output throughout the year. It is evident from Table 5.1 and Figure 5.4 that lower tilt yields higher annual output at this latitude, because the sun is close to being directly overhead for a large portion of the year.
Table 5.1 Yearly output (kWh/m2/yr) of solar panels on roofs of different aspect and tilts per m2 area
Figure 5.4 shows the variations in radiation received on modules with differing tilt at 15 N (14.65 N to nearest degree). At lower tilt angles from early April to late September, incident power is closer to power on the horizontal because the sun is (nearly) directly overhead. Maximum module power is generated when the sun’s rays are perpendicular to the module, so when tilted, a module cannot extract the maximum power from the sun’s radiation if it is directly above. However, during the months of the year (October – March) when the sun is not at its zenith (directly overhead), the tilted module power is greater than that of a flat one, and is closer to the theoretical maximum of a module that perfectly tracks the sun. These two effects are more pronounced at 20° tilt, but are still minimal because the Philippines is close to the equator.
Because the potential output is higher during summer when the sun is closest to its zenith, output averaged over the year is higher for modules at lower angles because the losses during summer, when potential output is closest to its maximum, are minimal compared with gains in the winter.
Optimum tilt and orientation should be more properly assessed to gain a more precise, site-specific estimate of output, perhaps using a model like that of Kacira et al (2004).
Figure 5.4 Solar radiation received at 15’N on surfaces with varying tilt. Module power: solar radiation received by a (tilted) module; incident power: what would be received by a sun-tracking module, perpendicular to the sun’s rays; power on horizontal: what would be received for a flat module, i.e. the radiation reaching the ground. Source: Honsberg & Bowden (2013)
There were several assumptions made when calculating the radiation and output of solar modules. Solar irradiance data were taken from NASA’s SSE database, which is of a significant size and reputation. Data is compiled from the World Climate Research Program/Surface Radiation Budget (WCRP/SRB) shortwave data set (Version 1.1), ISCCP-C1 (International Satellite Cloud Climatology Project) 3-hourly satellite data, and ISCCP-C2 monthly mean cloud fraction data on a 2.5° grid cell (~280km2).
Day 197 (July 16th) was chosen because the average monthly value for July (4.55kWh/m2/day) was closest to the yearly average (4.56kWh/m2/day), and the 16th is the median day in the month.
The yearly average radiation value (1664.4kWh/m2/day) was used to calculate the incident, horizontal and module irradiances, and elevation, azimuth and zenith angles. This is often not sufficient to evaluate temporal variations in output but simplification is necessary given the limited scope of this study. A more detailed method is found in Eerme (2012).
Module tilt (β) is measured from the horizontal, and was only calculated for 10° and 20° because steeper roofs were rarely observed in Payatas (<1%).
Solar radiation received, calculated from Eq. 4.4, is without including the effects of cloud-cover (and is hence the maximum potential power). The effects of over-shading and cloud cover from primary data were included in Eq. 4.8 as part of the η coefficient. The geographical distance between the sampling site and study site, although small, could affect the results, especially with respect to cloud cover. Convective cells can be very small and thus spatial disparities could exist between the Science Garden and the study site. Furthermore, the measurements were taken in July and August during the wettest, and thus cloudiest, part of the year, which could lead to an overestimation of the shading of modules.
5.5.3. Electricity output
Solar could supply 23.40% of projected 2030 demand, providing a fairly reliable, consistent load. A significant proportion of electric demand is during the hours of daylight, when PV operates, so could it be used to fit load curve peaks, as discussed by El-Khattam & Salama (2004). Diurnal variability would need to be smoothed, for instance using biogas generation to meet nocturnal demand, or storage could be installed at the point-of-use as detailed in Meyer & Ernest van Dyk (2004).
5.5.4. Distributed generation
The majority of load in Luzon (70%) occurs in the Metro Manila area, and most of the electricity currently supplied to Payatas is imported from generators in the North of the island. PV would be a suitable technology for distributed generation (DG) as it could provide a reasonable proportion of the electricity required at the point-of-use, thereby minimising brownouts, reducing transmission losses, improving energy security, and thus removing the need for further, complicated, grid extension.
Several studies have shown DG can increase community involvement in energy provision and demand reduction (e.g. Wolsink, 2012). Survey respondents were enthusiastic about the idea of community-controlled solar electricity, and some were already involved in energy co-operatives. This suggests community DG would be a popular, realistic and effective method of renewable electricity provision.
Storage of energy, perhaps with batteries, would make the system more robust because demand could be met during the hours of darkness and problems of intermittency could be minimised, enabling integration with the grid and maximising efficacy (c.f. Toledo et al, 2010; Hill et al, 2012). A wealth of literature exists on the integration of PV into ‘smart’ grids, assessing storage and active demand management (e.g. Kanchev et al, 2011; Zahedi, 2011), which would make Payatas an example of responsive and innovative electricity provision.
5.6. Sensitivity and Uncertainty Analyses
The sensitivity and uncertainty of the model, which are highly linked, were tested as part of objective three. The results are summarised briefly in Table 5.2.
Uncertainty analyses were conducted to ascertain the error of observations, and therefore confidence in the conclusions drawn.
Table 5.2 Sensitivity and uncertainty of various parameters
|Variable||Sensitivity||Uncertainty||Change in variable||Change in output of relevant technology|
|Hub height||Low||Negligible||±1m (1-1.5%)||±<0.10m/s|
|Blade length||Moderate||Negligible||±1m (2-3%)||±<2.20W/m/s|
|Head height||High||±5m||±1m (7%)||± 10%|
|Flow rate||Moderate||Unknown||±20.59m/s (20%)||+7% -5%|
|Roof area||Considerable||±4500m2 (7%)||±1%||±1%|
|Radiation||Low||±<166kWh/m2/yr (10%)||±166kWh/m2/yr (10%)||±<166kWh/m2/yr (<10%)|
|Performance Ratio||Moderate||±0.1 (12%)||±0.1 (10%)||±24.97kWh/m2/yr|
|Biogas Load Factor||Moderate||±0.05 (5%)||±0.05 (5%)||±0.06MW|
Uncertainty is highest for variables affecting hydroelectric output; head height and flow rate (Figure 4.9 and Table 5.2), so confidence in the conclusions is affected most by this. Head height is highly dependent on the location of a run-of-river project, and in the absence of a thorough study into ideal sites, error is high. Reducing uncertainty in this area would have a significant effect on the overall uncertainty of hydroelectric output because the model is highly sensitive to variations in head height; cutting error by just ±1m would reduce uncertainty by ±10%. The combined uncertainty for hydroelectricity is ±53.85%, including that of head height and flow rate.
The uncertainty of NASA radiation data is less than ±10% when compared with surface data for all areas relevant to the Philippines (NASA, 2005). Although the data is relatively old (1985-1988) radiation figures do not change greatly enough for this to be a significant factor in the final output; the IPCC notes that solar irradiance and radiative forcing changed by only -0.04W/m2 from 1986-2008 (Stocker et al, 2013), 41,610 times less than the average yearly radiation received at the surface in Payatas.
Error in roof area (~±7%) is derived from in situ corroboration studies performed to validate secondary estimates made from Google Earth as described in Section 3. The combined uncertainties of roof area, performance ratio and radiation produce a total of ±15.78% for solar electricity. Figures for turbine characteristics were obtained from manufacturers, so uncertainty in hub height and blade length is assumed to be negligible.
There is unknown uncertainty associated with flow rates from Tachikawa et al (2004) as a detailed methodology and uncertainty analysis is missing from the study. As the sensitivity of the model to this variable is considerable, this could present a limitation on the conclusions of this study.
Uncertainty in other variables is fairly minimal, with similarly minor effects on the final conclusions, dictated by the sensitivity of the model.
Simple sensitivity analysis was conducted to dissect the data and explore the robustness of the model to changes in various factors.
Although power output increases linearly with increases in solar flux (Borbely & Kreider, 2001) the model is fairly robust to changes in radiation; a change of ±166kWh/m2/yr (±10%) produces a change in solar output of just ±1.6kWh/m2/yr, significantly lower than the error of radiation. However, because the uncertainty of radiation is relatively high, the sensitivity must be quoted as <166kWh/m2/yr. The model is also insensitive to variations in hub height: as shown in Section 5.2.2, power output is more sensitive to changes in blade length, and thus the swept area of the turbine. Changes in hub height have comparably little effect on the wind speed, particularly when they are already low. A 1m increase in blade length (or πr in swept area) produces a change of 2.2W/m/s wind speed at hub height, while wind speed changes by <0.1m/s for each 1m increase in hub height. This is to be expected from Eq. 4.2, which shows power to be proportional to swept area (πr2), and the cube of wind speed, u3. Initial wind velocity therefore has a greater effect on output than hub height because output increases non-linearly with respect to wind speed.
The model is most sensitive to changes in head height of the run-of-river hydroelectric scheme. A ±5m change in head height produces a ±50% change in output, because it significantly alters the total energy per unit mass of water, as per Bernoulli’s equation (Twidell & Weir, 2006). The flow rate is also an important variable; a ±20% change from the values in Tachikawa et al (2004) yields a 7% increase or 5% decrease in hydroelectric output.
The model is considerably sensitive to changes in roof area; output increases linearly with changes, which is predictable from Eq. 4.8. The performance ratio, η, is also fairly important in determining final output. A change of ±0.1 (roughly equivalent to the error) produces a change of nearly ±25kWh/m2/yr, around a sixth of the solar output (see Table 5.2). The load factor of landfill biogas generators also has a moderate impact on final output.
Load can be expected to be relatively constant in Payatas because appliances are required consistently diurnally and annually. Peaks would likely occur during the evening and summer when most lighting, cooling and entertainment are required simultaneously.
Figure 5.5 shows the weekly demand forecast for 22nd-28th November 2013, a typical week on the large Luzon grid. Predictable dips in system load occur over the weekend, which are fairly straightforward to forecast because the majority of demand is residential.
Flexible demand management, via mechanisms like smart grids or direct price incentives, might smooth or even change the timing of peak demand as discussed by Everett (2007). More efficient appliances could also reduce demand and save residents money, though this would likely lead to a rebound effect, whereby efficiency savings encourage increased use of electricity as the price per unit falls (Sorrell et al, 2009).
Figure 5.5 Weekly demand forecast 22/11/2013-28/11/2013 for Luzon grid, all in MW. Adapted from NGCP (2013). Black line: capacity, dark green: coal, aquamarine: diesel, orange: natural gas/combined cycle, yellow: geothermal, dark blue: hydroelectric, purple: system load, red: gross reserve/deficiency, grey: required regulating reserve, lime green: required contingency reserve, turquoise: required dispatchable reserve.
Because Payatas is so small, DG should suffice for base- and mid-range loads, whereas plants with different response times (to fluctuations in demand) would be required on a larger (grid) scale, as in Laughton (2007).
Increasing the reserve margin of the electricity system could minimise problems associated with renewable intermittency. On a small scale (Payatas) intermittency could be countered through use of the grid as a back-up or storage mechanism, as described by Bayod-Rújula (2009).
Landfill biogas electricity could provide a reliable base load of 0.99±0.06MW, with 9.08±4.89MW from hydroelectricity, and 2.59±0.41MW from solar.
Each household is assumed to have 1.87m2 photovoltaic modules on the roof, operating at 32W capacity, roughly 16-21% of peak household load. This would require significant, rapid investment because this is equivalent to roughly half of installed national capacity in 2013 (Clover, 2013). A further 12-16% of this load would come from landfill biogas generation, and the remainder met with hydroelectric or grid capacity. However, because peaks are likely to occur during darkness, unless storage is incorporated into the system these will have to be supplied without using solar electricity.
Fluctuations in the power output of renewables such as photovoltaics are difficult to forecast. Rapid-response back-up generation capacity would be necessary to supplement an increasing number of renewables, such as pumped storage. The mountainous terrain would easily accommodate a small hydroelectric pumped storage facility nearby, or the nearby Kalayaan facility could be used.
5.8. Further Research
Although technically feasible, many influential factors beyond the scope of this study, such as the political and economic environment, affect whether the 75% (8.3MW) target is realistically achievable.
From household surveys of ‘jumpers’, people who illegally abstract electricity, it is apparent that economic and political factors significantly affect their access to legal electricity. The terms of Meralco’s (the local electricity utility) supply require a discernible bathroom and kitchen if a dwelling is to be connected to a meter. Many poor residents do not fulfil the requirements and are therefore forced either to purchase electricity at an elevated price from an illegal intermediary, or to steal it directly from high-voltage lines, at great personal risk (as in Figure 5.6).
Furthermore, the government is not likely to extend infrastructure to an area it is actively trying to relocate people from. It is also unlikely that electrification will be a political priority in the aftermath of Typhoon Haiyan, which killed over 5000 people and destroyed large parts of the country (Laguardia, 2013).
Although preliminary survey results and evidence from wider literature (e.g. Rogers et al, 2012; Warren & McFadyen, 2010; Wolsink, 2012) suggest that renewable community electricity would be welcome, a more detailed assessment of social acceptability of any proposed scheme would need to be undertaken to verify this.
The economics of wind generation, discussed in Section 5.2.4, make the option particularly unviable when considering the low projected output. Calculations are shown in Appendix 4.
Although there is little consensus, some studies assert that solar PV will reach grid parity (costs equal to grid generation) by the middle of this decade: 83% of the Asia-Pacific residential electricity market segments are predicted by Breyer & Gerlach (2012) to be beyond grid-parity by 2020. The Levelised Cost of Electricity of some types of PV dropped by almost 50% from 2009 to 2012 (Bazilian et al, 2013), suggesting that residential distributed solar generation will be an economically attractive option before 2030, even without subsidy (Reichelstein & Yorston, 2012) and despite residential installations being expensive (Hernandéz-Moro & Martinéz-Duart, 2013).
Figure 5.6 Jumping the meter wall in Phase III. Tangled (orange) telephone wires are used to abstract electricity from legal meters (bottom right) – Meralco conduct around 100 raids per week to disconnect such connections.
Hydroelectricity is the most economically viable renewable, although there are other negative environmental impacts. A small scheme such as that proposed, however, is likely to have minimal negative impacts while providing sufficient electricity for Payatas.
As the Quezon City Biogas project is already functional, its economic viability has been proven. However, without Clean Development Mechanism (CDM) credits, the project may struggle to maintain viability. This depends on the cost of wholesale electricity – the generator currently exports to Meralco at PHP4.5/kWh, though this could change in future.
Demand is projected to at least double by 2030 based on a baseline of 2007/09 (Mukherjee & Sovacool, 2012; DOE, 2010). It is technically feasible that the target 8.3MW could be generated, as per objective four of this study, though political and economic factors further affect whether it is achievable. A total of 12.66±4.91MW could be generated with the technologies discussed. The largest proportion of electricity supply would be met with hydroelectricity, generating 9.08±4.89MW. Wind generation is an economically and practically unviable resource, and could not supply any of the demand in Payatas. A base-load of 0.99±0.06MW could be provided by landfill biogas generation, the infrastructure for which already exists. Distributed solar photovoltaic generation would generate nearly a quarter of total peak demand (2.59±0.41MW) or up to 21% of household peak demand. DG would increase the resilience of the electricity system, increase community involvement, and minimise transmission losses. Combined with demand-side management, DG could address some problems of intermittency and variability in renewable generation. Uncertainty would need to be reduced with further data collection and scrutiny, as uncertainty is highest (±53.85%) regarding hydroelectricity, where the model is most sensitive. Further investigation beyond the scope of this study is required, such as an economic assessment of the proposed system, and investigation of political and social factors such as acceptability. These would allow stronger conclusions to made, and improve confidence in projections.
This study has presented data on renewable slum electrification in the Philippines, an area that has not previously been researched or widely documented in scientific literature. It has shown that the aim of supplying 75% of Payatas’s growing electricity demand with renewable sources by 2030 is possible, though significant uncertainty limits the conclusions drawn. Better quality data collection and targeted research could improve this. Demand was quantified in Payatas, fulfilling objective one, and is projected to double regionally (DOE, 2010) and nationally (Mukherjee & Sovacool, 2012) by 2030. This trend is likely to be replicated in Payatas, bringing peak demand to 11MW, 75% of which is 8.3MW. The potential contribution of four renewable sources was analysed in accordance with the second objective, and it emerged that wind is a practically and economically unviable option for electricity generation, due to the low wind speeds measured in Payatas (frequently <0.5m/s), and hence low potential output (72kW on average). The total potential generation capacity of 12.66±4.91MW is therefore comprised solely of solar, landfill biogas and hydroelectricity. Primary data collected was useful to add depth to secondary data, which had a low temporal resolution or was imprecise.
Distributed solar generation could generate nearly a quarter of total peak demand (2.59±0.41MW) and could remove many problems associated with current grid provision, such as brownouts and transmission losses. It would be conducive to autonomous community ownership of electricity systems, which is locally relevant, as shown by household surveys, and has been shown to increase acceptability of projects (Wolsink, 2012). Peak output from hydroelectric and solar generation occurs at different times of year, so some issues with variability could be avoided by balancing supply of these sources. However, problems of diurnal intermittency and variability would have to be reduced using demand-side management and by using predictable and reliable capacity such as landfill biogas generation to provide a base-load, with DG meeting peaks in demand at the point-of-use. Biogas could generate 0.99±0.06MW, and the infrastructure already exists, making development uncomplicated. This suggests that achieving the 75% target is technically feasible, as per objective four.
Hydroelectricity could supply the most significant proportion (50-168%) of the 8.3MW target, even using the lower bounds of the specified uncertainties (4.19MW). However, the model is most sensitive to variables pertaining to hydroelectric output, where uncertainty is also highest. Reducing uncertainty here would therefore have the most significant impact on improving the confidence in conclusions drawn, an important finding of work on objective three. Steps taken to fulfil objective four showed that realisation of the 75% target depends greatly on the uncertainties discussed, and on economic and political conditions, analyses of which are outside the scope of this study. It is technically probable that 8.3MW could be generated renewably by 2030, but whether or not this is a political priority will determine if this is the case. The Filipino government’s targets to improve efficiency and dramatically increase renewable energy generation indicate, however, that some degree of renewable electrification may be politically possible in Payatas during the coming decades.
Further research is necessary to solidify the rough conclusions of this study. Areas of particular importance are the location of any potential run-of-river hydroelectric scheme along the Marikina River, which would allow necessary quantification of the uncertainty associated with head height and flow rate, and a detailed assessment of the political, economic and social factors associated with any renewable electricity delivery system in Payatas.
Special thanks go to my supervisor, Dr. Jane Powell, and Prof. Andrew Lovett for constructive criticism and guidance on the project. I thank Mr. and Mrs. Wilkinson for their sponsorship, without which my research would not have been possible. I also thank Engr. Jojene Santillo, and the team of Dr. Enrico C. Paringit for the kind use of cross-section data from their study of Flood Parameterisation on the Marikina River. The advice and collaboration of Engr. F. R. Domingo Jr. and Ric Yambao at the Department of Energy in Manila was also invaluable.
 Barangays are the lowest administrative level in the Philippines, comparable to local councils
 PHP = Filipino Peso. PHP10 ≈ £0.14
 Load Factor: the ratio of the average load supplied by a system to the maximum load over a specified time period; 0.9 would show a system is operating 90% of the time.
 The distance the water falls to the turbine
 Loss values do not add up linearly to η=0.6 because losses are weighted within the model.
 At 2008 USD exchange rates; US$1 = £0.616133788 (XE Currency, 2013)
 Distributed generation is widely defined (e.g. Ackermann et al, 2001) but generally refers to small scale, in situ generation on the demand side of a network.
Ackermann, T., Andersson, G. & Söder, L. (2001) Distributed generation: a definition. Electric Power Systems Research 57 (3) 195-204
Ahmed, W. & Menzies, I. (2012) Using Output-Based Aid in Urban Projects.
Ahrens, C. D. (2003) Meteorology Today: An introduction to weather, climate and the environment (7th Ed.) Brooks Cole, Thompson, USA.
Akella, A. K., Saini, R. P. & Sharma, M. P. (2009) Social, Economical and Environmental impacts of renewable energy systems. Renewable Energy 34 (2) 390-396
American Wind Energy Organisation (undated) Size specifications of common industrial wind turbines [online] available at http://www.aweo.org/windmodels.html accessed 2nd October 2013
Andrews, J. & Jelley, N. (2007) Energy Science: principles, technologies and impacts. Oxford University Press.
Asian Development Bank (2011) Long-term Projection of Asia Development Bank. Asian GDP and Trade. ADB, Mandaluyong City, Philippines
Asian Power. (2011) The Philippines is on the road towards 90% household electrification. IN FOCUS Asian Power.
Badilla, R. (2008) Flood Modelling in Pasig-Marikina River Basin. Thesis; International Institute For Geo-Information Science And Earth Observation Enschede, The Netherlands
Baruah, B. (2010) Energy Services for the Urban Poor: NGO Participation in Slum Electrification in India Environment and Planning C: Government and Policy 28 (6) 1011 – 1027
Bayod-Rújula, A. A. (2009) Future development of the electricity systems with distributed generation. Energy 34 (3) 377-383
Bazilian, M. Onyeji, I., Leibreich, M., MacGill, I., Chase, J., Shah, J., Gielen, D., Arent, D., Landfear, D. & Zhengrong, S. (2013) Re-considering the economics of Photovoltaic power. Renewable Energy 53 (1) 329-338
Borbely, A. M. & Kreider, J. F. (2001) Distributed Generation: The Power Paradigm for the New Millennium. CRC Press
Boyle, G. (2004) Solar Photovoltaics in Boyle, G. (ed.) (2004) Renewable Energy. Oxford University Press.
Brew-Hammond, A., (2012) Energy: the Missing Millennium Development Goal. Energy for Development – Environment and Policy 54 (2012) 35-43
Breyer, C. & Gerlach, A. (2012) Global Overview on Grid Parity. Progress in Photovoltaics: Research and Applications 21 (1) 121-136
Clover, I. (2013) Philippines to reach 5 MW solar capacity by year’s end. P Magazine: Photovoltaic Markets and Technology, 24th September 2013.
Coelho, S. T. & Goldemberg, J. (2013) Energy access: Lessons learned in Brazil and perspectives for replication in other developing countries. Energy Policy 61 (2013) 1088-1096
Department of Energy (2009) Philippine Power Development Plan 2009-2030. DOE, Philippines.
Department of Energy (2011) National Renewable Energy Program (NREP) DOE, Philippines.
Department of Energy (2012) 21st EPIRA Implementation Status Report (Period covering May 2012 to October 2012) DOE, Philippines.
Deshmukh, A. (2009) The Role of Decentralized Renewable Energy for Rural Electrification: Maharashtra case study, India. IIIEE, Lund University
Eerme, K. (2012) Interannual and intraseasonal variations of the available solar radiation in E. B. Babatunde (ed.) (2012) Solar radiation. InTech, Croatia.
El-Khattam, W. & Salama, M. M. A. (2004) Distributed generation technologies, definitions and benefits. Electric Power Systems Research 71 (2) 119-128
Elliott, D., Schwartz, M., George, R., Haymes, S., Heimiller, D. & Scott, D. (2001) Wind Energy Resource Atlas of the Philippines. National Renewable Energy Laboratory, Colorado.
Everett, B. (2007) Demand Flexibility, Micro-Combined Heat and Power and the ‘Informated’ Grid. In: Boyle, G. (ed.) Renewable Electricity and the Grid: the challenge of variability. Earthscan, London.
Fabritz, G. (1954) Wasserkraftmaschinen in Hiitte Maschinenbau II (A) Wilhelm Ernst and Sohn. Berlin. 865-961
Ferrer-Martí, L., Garwood, A., Chiroque, J., Ramirez, B., Marcelo, O., Garfí, M., & Velo, E. (2012) Evaluating and comparing three community small-scale wind electrification projects. Renewable and Sustainable Energy Reviews 16 (7) 5379-5390.
Ghasemi, A., Asrari, A., Zarif, M., & Abdelwahed, S. (2013) Techno-economic analysis of stand-alone hybrid photovoltaic–diesel–battery systems for rural electrification in eastern part of Iran—A step toward sustainable rural development. Renewable and Sustainable Energy Reviews 28 456-462
Google Earth Imagery (2010) 14º42’43.91”N 121º06’06.96” E Imagery date 26/2/2010. Aerometrex.
Hernandéz-Moro, J. & Martinéz-Duart, J. M. (2013) Analytical model for solar PV and CSP electricity costs: Present LCOE values and their future evolution. Renewable and Sustainable Energy Reviews 20 (1) 119-132
Hill, C. A., Such, M. C., Chen, D., Gonzalez, J. & Mack Grady, W. (2012) Battery Energy Storage for Enabling Integration of Distributed Solar Power Generation. IEEE Transactions on Smart Grid 3 (2) 850-857
Honsberg, C. & Bowden, S. (2013a) Terrestrial Solar Radiation: Solar Radiation on a Tilted Surface [online] available at http://pveducation.org/pvcdrom/properties-of-sunlight/solar-radiation-on-tilted-surface accessed 9th September 2013
Honsberg, C. & Bowden, S. (2013b) Terrestrial Solar Radiation: Arbitrary Radiation and Tilt [online] available at http://pveducation.org/pvcdrom/properties-of-sunlight/arbitrary-orientation-and-tilt accessed 9th September 2013
IPCC (2011) (Edenhofer, O., Pichs‐Madruga, R., Sokona, Y., Seyboth, K., Matschoss, P., Kadner, S., Zwickel, T., Eickemeier, P., Hansen, G., Schlömer, S., von Stechow, C. (eds.)) Summary for Policymakers. In: IPCC Special Report on Renewable Energy Sources and Climate Change Mitigation. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.
International Energy Agency (2012) World Energy Outlook. IEA
Kacira, M., Simsek, M., Babur, Y. & Demirkol, S. (2004) Determining optimum tilt angles and orientations of photovoltaic panels in Sanliurfa, Turkey. Renewable Energy 29 (8) 1265-1275
Kanchev, H., Di, L. Colas, F., Lazarov, V. & Francois, B. (2011) Energy Management and Operational Planning of a Microgrid With a PV-Based Active Generator for Smart Grid Applications. IEEE Transactions on Industrial Electronics 58 (10) 4583 – 4592
Karekezi, S., Kimani, J., & Onguru, O. (2008) Energy access among the urban poor in Kenya. Energy and Sustainable Development 12 (4) 38–48
Kaundinya, D. P., Balachandra, P. & Ravindranath, N. H. (2009) Grid-connected versus stand-alone energy systems for decentralized power—A review of literature. Renewable and Sustainable Energy Reviews 13 (2009) 2041–2050
Kreigler, E., Edenhofer, O., Reuster, L., Luderer, G. & Klein, D. (2013) Is atmospheric carbon dioxide removal a game changer for climate change mitigation? Climatic Change 118 (1) 45-57
Kubik, M. L., Coker, P. J. & Hunt, C. (2011) Using meteorological wind data to estimate turbine generation output: a sensitivity analysis. Wind Energy Applications. World Renewable Energy Congress, 8-13th May 2011, Linköping, Sweden.
Lackner, K. S., Brennan, S., Matter, J. M., Alissa Park, A. H., Wright, A. & van der Zwaan, B. (2012) The urgency of the development of CO2 capture from ambient air. Proceedings of the National Academy of Sciences 109 (33) 13156-13162
Laguardia, C. (2013) Philippines disaster diary: desolation and bleakness has engulfed Tacloban. The Guardian Poverty Matters Blog, 23rd November 2013. [online] available at: http://www.theguardian.com/global-development/poverty-matters/2013/nov/26/philippines-disaster-relief-diary-tacloban accessed 27th November 2013
Lalwani, M., Kothari, D. P., & Singh, M. (2012) Viability analysis by techno-economic aspects of grid interactive solar photovoltaic project in India. IEEE 2012 International Conference on Advances in Engineering, Science and Management 769-772
Laughton, M. (2007) Variable Renewables and the Grid: an overview. In: Boyle, G. (ed.) Renewable Electricity and the Grid: the challenge of variability. Earthscan, London.
Lazard (2011) Levelized Cost of Energy Analysis—Version 5.0. Lazard, US
Levin, I., Naegler, T., Heinz, R., Osusko, D., Cuevas, E., Engel, A., … & Zimov, S. A. (2010) The global SF6 source inferred from long-term high precision atmospheric measurements and its comparison with emission inventories. Atmospheric Chemistry and Physics 10 (6) 2655-2662
Malyshev, T. (2009) Looking ahead: energy, climate change and pro-poor responses. Foresight 11 (4) 33-50.
Meralco AppCal (2010a) Basis for computing the estimated electricity consumption of home appliances for the Meralco Appliance Calculator (Meralco AppCal) [online] available at: http://www.meralco.com.ph/mac/formula.html accessed 6th September 2013
Meralco (2010b) Summary Schedule of Rates June 2010 Effective June 2010 Billing Month [online] available at: http://www.meralco.com.ph/pdf/rates/summary_schedule_rates_June2010.pdf accessed 6th September 2013
Meralco (2013) Summary Schedule of Rates June 2013 Effective June 2013 Billing Month [online] available at: http://www.meralco.com.ph/pdf/rates/2013/June/06_2013_GC_Table.pdf accessed 6th September 2013
Meyer, E. L. & Ernest van Dyk, E. (2004) Assessing the reliability and degradation of photovoltaic module performance parameters. IEEE Transactions on Reliability 53 (1) 83-92
Miler, S. M., Wofsy, S. C., Michalak, A. M., Kort, E. A., Andrews, A., E., Biraud, S. C., Dlugocencky, E. J., Eluszkiewicz, J., Fischer, M. L., Janssens-Maenhout, G., Miller, B. R., Miller, J. B., Montzka, S. A., Nehrkorn, T. & Sweeney, C. (2013) Anthropogenic emissions of CH4 in the United States. Proceedings of the National Academy of Sciences 2013, published ahead of print November 25th 2013 [online] available at: http://www.pnas.org/content/early/2013/11/20/1314392110.abstract accessed 2nd December 2013
Mukherjee, I. & Sovacool, B. (2012) Sustainability principles of the Asian Development Bank’s (ADB’s) energy policy: An opportunity for greater future synergies. Renewable Energy 48 173-182
Mundomanz (2013) Yearly SYNOP Reports: Science Garden, Philippines [online] available at: http://www.mundomanz.com/meteo_p/yearrep?countr=PHILIPPINES&ind=98430&year=2013&l=1&action=display accessed 8th November 2013
Myhrvold, N. P. & Caldeira, K. (2012) Greenhouse gases, climate change and the transition from coal to low-carbon electricity. Environmental Research Letters 7 (1) 014019
National Aeronautics and Space Agency (NASA) (2005) Global Change Master Directory: Surface Solar Energy (SSE) Monthly Data in ASCII. [online] available at: http://gcmd.nasa.gov/records/GCMD_SSE_MONTHLY.html accessed 15th November 2013
National Grid Corporation of the Philippines (2010) 2010 Transmission Development Plan. Volume 01: Major Network Expansions Final Report. NGCP, Philippines
National Statistics Office (2012) National Statistics Office Special Release: 2010 Census of Population and Housing – Final Results. NSO, Republic of the Philippines, NCR District II – Quezon City No. 2012-04
NEA/IEA (2010) Projected Costs of Generating Electricity, 2010 Update. Nuclear Energy Agency/ International Energy Agency, OECD, Paris.
Nisbet, E., & Weiss, R. (2010) Top-down versus bottom-up. Science 328 (5983) 1241-1243
Ochoa, L. F. & Harrison, G. P. (2011) Minimizing Energy Losses: Optimal Accommodation and Smart Operation of Renewable Distributed Generation. IEEE Transactions on Power Systems 26 (1) 198 – 205
OECD/IEA (2013) Energy Statistics Yearbook. Organisation for Economic Co-operation and Development/International Energy Agency.
OpenEI (2013) NASA Surface meteorology and Solar Energy (SSE) [online] available at http://en.openei.org/datasets/node/728 accessed 1st November 2013
Ostojic, D. R., Bose, R. K. Krambeck, H., Lim, J. & Zhang, Y. (2013) Energizing Green Cities in Southeast Asia: Applying Sustainable Urban Energy and Emissions Planning. World Bank, Washington D.C.
Parikh, P., Chaturvedi, S. & George, G. (2012) Empowering change: The effects of energy provision on individual aspirations in slum communities. Energy Policy 50 (2012) 477-485\
Parthasarathy, S. (2010) Breaking the expertise barrier: understanding activist strategies in science and technology policy domains. Science and Public Policy 37 (5) 355-367
Petersen, E. L., Mortensen, N. G., Landberg, L., Højstrup, J. & Frank, H. P. (1998) Wind power meteorology, Part I: Climate and turbulence, Wind Energy 1 2- 22
Photovoltaic Software (2012) Photovoltaic Design Tools ‘How to calculate the annual solar energy output of a photovoltaic system’ [online] available at: http://www.photovoltaic-software.com/PV-solar-energy-calculation.php accessed 7th September 2013
Poff, N. L., Allan, J. D., Bain. M. B., Karr, J. R., Prestegaard, K. L., Richter, B. D., Sparks, R. E. & Stromberg, J. C. (1997) The Natural Flow Regime. BioScience 47 (11) 769-784
Racelis, R. H., M.R.M. Abrigo, and J.M.I. Salas (2012). Philippines 2007 National Transfer Accounts (NTA): Consumption, Labour Income and Lifecycle Deficit by Income Group. Philippine Institute for Development Studies, Discussion Paper Series No. 2012-32. PIDS, Makati City, Philippines
Ramanathan, V. & Feng, Y. (2009) Air pollution, greenhouse gases and climate change: global and regional perspectives. Atmospheric Environment 43 (1) 37-50
Reichelstein, S. & Yorston, M. (2012) The Prospects for Cost Competitive Solar Power. Energy Policy (2012)
Repower (undated) Repower Systems: MM92. Factsheet. Repower Systems SE, Hamburg, Germany
Rogers, J. C., Simmons, E. A., Convery, I. & Weatherall, A. (2012) Social impacts of community renewable energy projects: findings from a woodfuel case study. Energy Policy 42 (2012) 239-247
Rosnes, O. & Vennemo, H. (2012) The Cost of Providing Electricity to Africa. Energy Economics 34 (5) 1318-1328
Scott, N., McKemey, K., Batchelor, S., Cowan, B., Awasthi, P., Mohlakoana, N., … & Singru, N. (2005) Energy in Low-Income Urban Communities (Vol. 8146). Final Technical Report for DFID KaR Project.
Smyser, C., Anneke, W., Maia, S., Vitelli, M. L. & Sullivan, J. (2004) Innovative Approaches to Slum Electrification. Advanced Engineering Associates International. USAID
Sørensen, B. (2004) Renewable Energy: Its physics, engineering, use, environmental impacts, economy and planning aspects (3rd ed.) Elselvier Academic Press
Sorrell, S., Dimitropoulos, J. & Sommerville, M. (2009) Empirical Estimates of the Direct Rebound Effect: a Review. Energy Policy 37 (4) 1356-1371
Stocker, T., Dahe, Q. & Plattner, G. K. (co-ordinating lead authors) et al (2013) Working Group I Contribution to the IPCC Fifth Assessment Report (AR5), Climate Change 2013: The Physical Science Basis. IPCC, Geneva
Timilsina, G. R., van Kooten, G. C. & Narbel, P. A. (2013) Global wind power development: Economics and policies. Energy Policy 61 (2013) 642-652
The Wind Power: wind turbines and wind farms database (2013) Suzlon S64/950 Factsheet [online] available at http://www.thewindpower.net/turbine_en_219_suzlon_s64-950.php accessed 2nd October 2013
Toledo, O. M., Filho, D. O. & Diniz, A. S. A. C. (2010) Distributed photovoltaic generation and energy storage systems: A review. Sustainable and Renewable Energy Reviews 14 (1) 506-511
Twidell, J. & Weir, T. (2006) Renewable Energy Resources (2nd Ed.). Taylor & Francis, London, New York.
Ueckerdt, F., Brecha, R., Luderer, G., Sullivan, P., Schmid, E., Bauer, N., & Böttger, D. (2011) Variable renewable energy in modelling climate change mitigation scenarios. In: Proceedings of the 2011 International Energy Workshop, Stanford, US.
United Nations (2011) World Population Prospects: The 2010 Revision (Medium fertility variant, 2011-2100). Population Division, Department of Economics and Social Affairs, United Nations.
Vestas (undated) Vestas V90-1.8MW Factsheet: the North American Variant. Vestas – American Wind Technology, Inc., Portland, USA
Vestas (2013) V90 – 1.8/2.0MW Gridstreamer™ Operational Data [online] available at: http://www.vestas.com/en/wind-power-plants/procurement/turbine-overview/v90-1.8/2.0-mw-gridstreamer™.aspx#/vestas-univers accessed 2nd October 2013
Warren, C. R. & McFadyen, M. (2010) Does community ownership affect public attitudes to wind energy? A case study from south-west Scotland. Land Use Policy 27 (2) 204-213
Webb (2013) Renewable Energy in Asia Pacific: a Legal Overview (3rd Ed.) DLA Piper
Weibull, W. (1951) A Statistical Distribution Function of Wide Applicability. Journal of Applied Mechanics 18 (3) 293-297
Weiss, R. F., & Prinn, R. G. (2011) Quantifying greenhouse-gas emissions from atmospheric measurements: a critical reality check for climate legislation. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences 369 (1943) 1925-1942
Winkler, H., Simões, A. F., Rovere, E. L. L., Alam, M., Rahman, A., & Mwakasonda, S. (2011) Access and affordability of electricity in developing countries. World Development 39 (6) 1037-1050
Wolsink, M. (2012) The research agenda on social acceptance of distributed generation in smart grids: Renewable as common pool resources. Renewable and Sustainable Energy Reviews 16 (1) 822-835
World Bank (2013) World Bank Databank, World Development Indicators database [online] available at: http://databank.worldbank.org/ddp/home.do accessed 15th March 2013
XE Currency (2013) Current and Historical Rate Tables: USD – US Dollar [online] available at: http://www.xe.com/currencytables/?from=USD accessed 15th November 2013
Yadoo, A. & Cruickshank, H. (2012) The role of low carbon electrification in reducing poverty, and climate change strategies. Energy Policy 42 591-602
Zafirakis, D. & Chalvatzis, K. J. (2014) Wind energy and natural gas-based energy storage to promote energy security and lower emissions in island regions. Fuel 115 (2014) 203-219
Zahedi, A. (2011) Maximizing solar PV energy penetration using energy storage technology. Renewable and Sustainable Energy Reviews 15 (1) 866-870