InTechOpen uses cookies to offer you the best online experience. By continuing to use our site, you agree to our Privacy Policy.

Engineering » Electrical and Electronic Engineering » "Development and Integration of Microgrids", book edited by Wen-Ping Cao and Jin Yang, ISBN 978-953-51-3400-8, Print ISBN 978-953-51-3399-5, Published: August 16, 2017 under CC BY 3.0 license. © The Author(s).

# FutureGrid: Use of Microgrids in Underserved Communities

By Andrew Hubble and Taha Selim Ustun
DOI: 10.5772/intechopen.68622

Article top

## Overview

Figure 1. Brazil’s, Sudan’s, and South Sudan’s existing electrical networks.

Figure 2. Selected countries.

Figure 3. Nyakabanda (left) and Rwamiko (right) load profiles.

Figure 4. HOMER microgrid diagram.

Figure 5. Breakeven distances for different load profiles.

Figure 6. A 40‐km wide corridor around existing transmission lines in Brazil (left) and Ghana (right).

Figure 7. Average and three standard deviations breakeven distance.

Figure 9. Torre de Lua community load profile.

Figure 10. Torre de Lua HOMER microgrid schematic.

Figure 11. Torre de Lua solar radiance.

Figure 12. Torre de Lua river flow rates.

Figure 13. Relationship between Nyakabanda community and closest grid access point.

Figure 14. Nyakabanda community HOMER microgrid setup.

Figure 15. Unused energy in storage‐less systems.

Figure 16. Rearrangement of loads to match SEMHC scheduling.

Figure 17. Probability of sufficient power for multiple clipping tiers.

Figure 18. Power consumed based on an amount of home clipped.

# FutureGrid: Use of Microgrids in Underserved Communities

Andrew Hubble and Taha Selim Ustun
Show details

## Abstract

Rural electrification in developing countries—especially Sub‐Saharan Africa—has trailed urban development drastically. The extreme costs associated with expanding traditional grid networks, and the relatively few people they serve, have proved to be a serious economic barrier. Being able to generate and distribute electricity at an affordable rate is crucial in order to effectively power homes, schools, health clinics, and private business. Through this continued cycle and lack of access to electricity, poverty only continues. If given access, quality of life increases through more educated, longer, and healthier lives as well as through developed entrepreneurship and business growth. Unfortunately, because of the remoteness of many communities they are often dismissed as unreachable. Furthermore, microgrids help address another global need: increased renewable energy penetration. Small‐scale energy production lends itself to solar installations, but depending on the location and available resources, wind and hydropower can also play an important role.

Keywords: microgrids, distributed generation, energy storage, grid extension, rural communities

## 1. Introduction

Often when we look at how technology has changed since its inception, it is difficult to imagine how the creators would react to the relentless progress and improvements on their original idea. Over a hundred years after Alexander Graham Bell invented the telephone, today’s smartphones are infinitely more complex, contain thousands of features, and possess processing power, Bell could never have imagined. Once luxury, cell phones are now found in every corner of the globe including the most remote villages in the developing world. If Thomas Edison and Nikola Tesla could see the state of electrification today, it is safe to assume they would be sorely disappointed. While electrification has certainly improved, it has severely lagged behind the growth of other technology. The generation and transmission of electricity looks much the same as it did over 130 years ago when the Vulcan Street Hydroelectric Plant in Appleton, Wisconsin began producing 12.5 kW of DC power. Over the next few years, more plants were constructed in both AC and DC, mostly powered by water or coal. While access to this electricity increased—as did the quality and economic viability—electricity never experienced the gigantic expansion in both availability and technology that other sectors did. Since its creation and original spurt of distribution, electricity has been slow to advance to a significant portion of the global population.

Unlike cell phones, electricity cannot be manufactured and shipped in discrete units. Because it is not a physical device, the infrastructure required to produce and distribute it is entirely unique. Due to the immense capital costs associated with electrification, individual business (and thus competition and natural advancement) have not developed in the same manner. As a result, electric utilities are slow to develop or expand, leaving no need for innovation.

Nowadays, power production finds its way into the public view as we battle the negative effects of climate change. Instead of the natural pressure in the industry to improve and out‐compete other companies, utilities are now being pressured externally via the government and general public. While a shift to renewable energy is undoubtedly important, it does not represent the only problem in this area. Access to any form of electricity in developing and rural areas is severely limited. At face value, it may seem that this is not an immediate problem, but there are innumerable secondary effects all stemming from a chronic lack of access.

Imagine a rural town in a developing country. There are 300 homes, a few grocery stores, a pharmacy, a general store, a school, a carpentry workshop, and a coffee milling station. What they do not have is electricity. While the government and utility are aware that the town exists, plans to offer electrification have never gone further than a Master Plan written years ago and shelved. Extending transmission lines are expensive, and if the utility thinks that there is not sufficient demand, they will not invest the money. The utility may also be unaware of the current size of the town, and thus the potential customers. Even if they did extend the grid, there are production shortages. Blackouts lasting hours or even days. This is not an unreasonable scenario, and is an accurate descriptor of a large portion of the unelectrified developing world. This lack of electricity means that at night families burn candles or kerosene lanterns which have harmful effects on the respiratory system when used in enclosed spaces; refrigeration is impossible, and food cannot be saved for long; water is pumped manually from boreholes, or carried from the nearest stream. All these activities which have to be done manually take an immense amount of time. Frequently, children are required to help their families in these tasks, and their studies suffer. By expending all this energy on the day‐to‐day tasks, it is difficult to develop and remove oneself from this cycle. Lacking access to electricity keeps people impoverished and uneducated.

## 2. Rural electrification and national grid distribution

Often microgrids are seen as solutions creating a more stable and reliable interconnected grid in urban settings [13], however, they need not be limited to these uses. Electrification in developing countries has trailed industrialized nations drastically and even more so in rural settings [4]. The vast majority of those without an electrical connection live in rural developing settings, where their access to resources in general is scarce. The eradication of poverty is on nearly every government agenda around the globe, and while on the surface, the undertaking is targeted and defined, in reality it is not. Access to reliable electricity is one large step in the correct direction and can no longer be considered a luxury. Lack of electrification contributes to the continued cycle of poverty, child mortality, chronic but otherwise treatable health issues, as well as suppressing education. Electricity is necessary for providing lighting without risking unnecessary smoke inhalation, for pumping water on anything other than an individual scale, and for refrigeration—which allows families to reduce food waste. Providing reliable and affordable electricity should be the top priority in tackling poverty eradication.

Unfortunately, the extreme costs associated with electrifying rural developing areas, as well as the relatively few people served, have caused come countries to exclude entire regions from their electrification schemes. With transmission lines costing up to $20,000 per km [5] in rough and rural terrain, shortening the transmission distances through the use of distributed generation and microgrids brings down the cost of rural electrification significantly. To understand the scope of the problem, transmission line coverage can best be depicted through utility maps. In Figure 1, large portions of Brazil, Sudan, and South Sudan are not currently serviced by existing transmission lines. In order to electrify these regions, millions of dollars would be required in infrastructure development. Fortunately, distributed generation microgrids can be utilized instead. #### Figure 1. Brazil’s, Sudan’s, and South Sudan’s existing electrical networks. Immediately, it becomes apparent that a more economical solution skips the long transmission lines and produces power closer to the users. The lower capital investment and varying sizes of communities present a wide array of customizable solutions, and as a result, there exists no uniform microgrid design which is applicable to all or even most potential microgrid sites. Despite this microgrids still hold a place in the global electrification scheme. This chapter aims to demonstrate that not only are microgrids closer to wide scale deployment in rural developing areas than may be commonly believed, but that there exist methods and technology making microgrids uniquely suitable for rural electrification. ## 3. Economic feasibility of microgrid breakeven distances Traditionally, there has been a singular approach to electrification: extend the national grid. When the utility considers this option for remote areas, little or no math goes into the determination. They are simply too far from the grid, and their demand is too low to justify the immense cost. With extension costs as high as$15,000 per km [6, 7] in rough terrain, the cost per kilowatt‐hour to achieve any kind of payback would have to be prohibitively expensive. As an alternative, residents of these rural areas can sometimes afford a few solar panels and batteries—especially if the cost is shared. This simple setup can sustain a little lighting and offer a place to charge cell phones, but stops well short of an acceptable solution. Instead, there exists a medium between these two solutions. Something, larger and more robust than a few panels linked together, but less expensive than grid extension: the microgrid. The microgrid can be sized and built according to demand and expanded with usage. The costs can be kept in check because the electricity is produced and consumed in the same area—no need for the expensive transmission lines. As a solution to rural electrification, the microgrid is new on the scene, and being largely untested requires some analysis to determine its feasibility.

The first step in determining if a microgrid is suited to a particular rural site is to compare the cost of the stand‐alone microgrid to the cost of extending the existing grid structure. The cost of a stand‐alone system is dependent on the load generated from the community, but is also dependent on the available resources. Wind and solar are obvious options because of their availability in remote areas, but diesel for generators should also be considered due to its widespread availability and capacity for consistent energy generation. These three options, plus battery storage, are at the heart of the microgrid solutions examined here. While options like geothermal and hydroelectric are completely viable, they have been intentionally omitted due to their geographical restrictions.

A study of 200 fictitious communities was performed in 50 unique locations with varying load profiles. For simplicity, and to ensure a variety of locations, one location was chosen over each of the 50 countries. The countries are shown in Figure 2. The countries were furthermore subdivided into five categories based on their economic standing. The countries presented in Table 1 rank the selected countries based on estimates of annual income generated by the poorest 10% of the population. This is calculated by taking the GDP produced by the poorest 10% [8] and dividing it by a tenth of the population.Class 1 countries generate less than $100 of income per year, whereas class 5 earns more than$10,000 per year. Figure 2 highlights the countries chosen.

#### Figure 2.

Selected countries.

CountryPoorest 10%ClassCountryPoorest 10%Class
Central African Republic43.02Class 1Brazil*1138.44Class 3
Haiti49.45Indonesia*1187.26
Malawi56.11Argentina*2001.52
The Gambia84.37Chile2469.82Class 4
Guinea‐Bissau90.85Gabon2477.57
Dem. Rep. Congo*92.89Russia2975.52
Lesotho*93.08Mauritius3004.99
Liberia109.89Class 2Uruguay*3193.29
Mozambique111.27Belarus3296.42
Burundi114.40Croatia3368.81
Togo120.66Greece3654.73
Comoros121.51Romania3698.77
Rwanda*146.09Seychelles3730.37
Fed. Sts. Micronesia152.85Singapore*12,945.45Class 5
Guinea161.88Germany16,259.45
Niger162.40Ireland16,856.05
Uganda171.50Denmark16,998.03
Afghanistan*240.76Class 3The Netherlands17,738.54
Nepal*245.59Sweden18,860.41
Ghana*273.91Iceland19,241.66
Tanzania*287.47Finland19,431.24
India*553.53Switzerland28,246.13
Paraguay*706.92Luxembourg31,499.45

#### Table 1.

Rural GDP distribution.

*Countries were part of a previous study and not selected based on the same criteria discussed in the text.

#### Figure 14.

Nyakabanda community HOMER microgrid setup.

Both the Rwandan and Brazilian case studies have shown that isolated microgrids are economically feasible. These are rough snapshots of what is required for electrification, but they highlight the major aspects. Thus far we have seen the viability of microgrids compared to grid extension, and the actual “per kWh” costs of implementing the microgrid. But these solutions represent an out‐of‐the‐box electrification scheme not tailored to rural life. Because the way electricity is consumed in the West and in rural developing locations differs drastically, new technology, process improvements, and innovative uses can drastically reduce wasted energy and thus further bring down the costs. The next section in this chapter delves into these changes and analyzes how much of an impact they can have on a microgrid.

## 6. Innovative enhancements for microgrid optimization

So far the microgrid has been treated as a response to a static load profile. Generally, this is acceptable because the electrical loads receive no feedback from the supply as to the amount of over/under production. If a new load is added to an already saturated system, voltage drops or blackouts may occur. Since these microgrids are designed and built from scratch, extra measures can be taken to reduce peak loads and distribute power consumption—easing the burden on the microgrid.

Batteries are often used in microgrids as a means of load shifting. When renewable energy generation is producing power in excess, the batteries can be charged and when the load outweighs the supply, the power stored in the batteries is used. Unfortunately, traditional batteries can be expensive and have limited cycle use before they need to be replaced. Alternative power storage methods are available—such as pumped hydro or flywheels—but a more grassroots solution reshapes the problem. Rather than store energy to match supply to meet demand change the nonessential demand to match supply.

### 6.1. Smart energy management for health Centers

Examples of how loads can be rearranged are readily available. One system, known as “smart energy management for health centers” (SEMHCs) clearly analyzes how this process works [32]. Rural health centers in particular are more prone to suffer from chronic power outages because of the nature of their loading. Clinics often have high‐power equipment that is only run for a short period of time. If many of these devices are powered simultaneously, it can overload the individual PV systems the run off. Since sufficient battery banks can be cost prohibitive at this scale, the clinics are at the mercy of whatever their arrays can produce. Alternatively, when the system is not being overloaded, energy that is not being used is lost. Instead, low‐tech scheduling systems can organize and shift loads to use the electricity more wisely.

Figure 15 illustrates this point further. The dotted line represents the power curve generated by a PV array throughout the course of a day, and the gray shaded blocks represent the loads. From t0 to t1 and between t4 and t5 the demand exceeds the generation, while the white space between them represents energy generated by the array that is lost.

### Figure 15.

Unused energy in storage‐less systems.

Scheduling patients’ services without the knowledge of power available ultimately hinder the services provided. Traditionally, clinics operate on a first‐come, first‐served basis, and if overloaded, the power will cut out with waiting patients. With the already low density of health centers in rural developing areas, patients sometimes have to travel large distances in order to arrive at the clinic. When the clinic then does not operate due to power issues, access to health services is essentially nonexistent.

SEMHC addresses this issue by scheduling services based on available energy and the priority of the service. The program starts by assessing the solar power production available, defined in Eq. (4) where s is the surface of the PV array, k is a constant for in‐line energy loss due to increased temperature, incident angle of radiation on the array, shading, and panel degradation, and R(t) is the average solar radiation on the PV panels over a specific period of time t. The values for R(t) can be pulled from satellite data from organizations such as NASA.

 E(Δt)=s·k·R(T)·Δt (4)

Next, the algorithm assigns a value Ci to the service required, where C represents the power rating and i is the device ranking or priority. The device is expected to operate for a closed time interval di. The process is optimized mathematically with Eq. (5).

 maxnϵℕ(∫tdtd+min(di)s·k·R(t)·Δt−∑i−1nCidi−A) (5)

The constant A represents the base load of the clinic—devices that are always on and consuming power. td represents the lowest time for which the generation power exceeds the nominal power of the devices. The final assurance is to guarantee that the demand does not exceed the power produced Pn by the panels, eliminating the possibility of overloading the system. Condition Eq. (6) then must be true at all times.

 ∑i=1nCidi≤Pn (6)

Since this is a priority‐driven system, any excess power not used by critical loads can be used low low‐priority low‐power demands. The system is shown in Figure 16 graphically. At the onset of the program, tmin, there are five loads already scheduled and running. Since the power available exceeds the demand, a new load NL3 is moved from standby to operating. NL3 is used because the higher priority loads would render condition (6) false.

### Figure 16.

Rearrangement of loads to match SEMHC scheduling.

After the completion of any load, the system reevaluates which waiting loads can be fit into the unused capacity. With this low‐cost, low‐tech solution to power management, unnecessary outages can be avoided and wasted energy can be minimized. This is only one specific case of a computer‐based microgrid improvement. With the advent of smart meters, control at a household level can be more than binary (on/off) as seen with the health clinic example. With built‐in wireless communication hardware, individual loads inside a household can be monitored and controlled.

### 6.2. The load attenuating stochastic simulator

The load attenuating stochastic simulator (LASS) utilizes the abilities of networked smart meters to control loads at a fine granularity, delivering or cutting power at the appliance level depending on the customer’s desires or requirements. This creates a tiered electrification scheme in which customers can pay a reduced rate under the condition that their power may be throttled first in the event, the demand exceeds generation [33]. While the results in this field suggest that the most economical solution is to provide sufficient capacity to meet all demand, once there is a limitation in supply, customers are best served by having their loads clipped versus having the entire system overload and experience blackouts. This works in much the same way as SEMHC, except it is applied over an entire microgrid, not an individual consumer.

To prove that load clipping increases overall electricity distribution, a simulation is used whereby LASS uses a Probability Mass Function (PMF) as an input over each time step for the possible loads and generation over a microgrid system for a traditional weekday and Saturday [33]. The simulator generates demand and loads for each time step based on a fictional microgrid setup. Inside this simulation, there exist two tiers of customers: customers whose power can be clipped and customers whose power cannot be clipped. The goal is to reduce the probability of power outages by determining what percentage of customers must be in the clipped category.

The base case (where no customers are clipped) and four other scenarios where 70, 75, 80, and 100% of homes can be clipped are examined. In the control case with no clipping, there are instances where power will have to be cut for multiple hours per day for both weekday and weekend profiles. The severity of this can be seen in Figure 17. The reverse is obviously true as well. If you cut power to 100% of homes, then the probability that there will be an outage is essentially nonexistent, since instead of overloading the microgrid, it has just been shut down completely.

### Figure 17.

Probability of sufficient power for multiple clipping tiers.

Whether or not power has been clipped, a power outage that has been avoided is only part of the problem. For any form of electricity distribution system to be viable, the maximum amount of electricity needs to be sold without shutting down the entire system. Protecting the grid by turning power off to everyone means that the utility (or distribution owner) loses income. Instead, there exists a balance between providing electricity to the maximum amount of people without increasing the risk of an outage. Figure 18 highlights the amount of power consumed (or sold) respective to each clipping scheme. No clipping is still the worst performing, since without clipping an outage is nearly guaranteed, and therefore no power is consumed by anyone. Again, the inverse is not ideal. If you cut power to everyone, then again, no power is sold or consumed. When 75% of customers’ power is clipped, the probability of an outage is less than 1%—nearly the same as any of the higher schemes. However, clipping only 75% of homes compared to 80% generates more revenue through more distributed power.

### Figure 18.

Power consumed based on an amount of home clipped.

The figures provide excellent visuals for hour‐by‐hour outage probability, but the root of the issue focuses on the total time, or overall percent chance of an outage. Table 9 shows the solutions for the weekday and weekend times.

Percentage of homes clippedExpected total power cut duration (hours/day)Expected energy sold (kWh/day)
Weekday0 (No Clipping)11.2303197.17
70%1.9474360.59
75%0.0137392.31
80%0.0005385.33
100%0.0005356.29
Saturday0 (No Clipping)9.4835240.20
70%0.0005396.87
75%0.0005392.89
80%0.0005388.92
100%0.0005373.13

### Table 9.

Optimum clipping rates.

Clipping to 75% on weekdays and 70% on Saturdays results in the highest amount of energy sold. While the power outages required for weekdays is higher than if more homes were clipped and the overzealous clipping actually leads to less power sold. This does not increase the robustness of the system, but instead takes an overly cautious approach and shuts down power unnecessarily.

Choosing the correct percentage of homes to clip is critical in the cost effectiveness and reliability of microgrids. While this study has provided an excellent starting point or generalization, each microgrid will have to be fine‐tuned to its particular power generation and loading curves. The goal of this section was to demonstrate that demand‐side changes can have a meaningful impact on both the reliability and affordability of microgrid setups.

The final section of this chapter explores a rural‐specific energy source not often considered in microgrids: agriculture. About 78% of the world’s poor lives in rural areas and relies on agriculture for both food security as well as household income [34]. Biomass is a large and often untapped rural resource that can provide a significant portion of on‐demand power.

Burning agricultural waste in small steam furnaces allows for localized generation utilizing an abundant and proximal resource. Generation units 10 and 50 kW are already in production for exactly this kind of use [35]. Village industrial power (VIP) operates in East Africa and offers an off‐the‐shelf 10 kW unit, which can be transported in the back of a small pickup truck. Agricultural waste is burned directly and the self‐contained unit can generate electricity. Used in tandem with solar or wind generation, a biomass unit can act in the same way that a diesel generator traditionally assists microgrids. Biomass can be stored and the unit brought online when demand is high.

The viability of biomass as a generation source primarily hinges on its availability, which is what makes this option suitable for rural and not urban use. Case studies have been carried out which are generally geographically limited. Specifically, the Punjab region of India was analyzed and it was determined to house vast untapped biomass resources. If the Punjab example is followed, the first step is determining what types of biomass are available. The six major biomass options from crops grown in the region are outlined in Table 10, and are categorized into four sections [36].

CategoryType of biomassName of crop
A1StrawWheat
Barley
Seasum
Pulses
A2StalkCotton
Maize
Arhar
Rapeseed and Mustard
A3Bagasse*Sugar Cane
Tops and LeavesSugar Cane
A4Cobs*Maize
Shells*Groundnuts

#### Table 10.

Identification of available biomass [36].

*Indicates processing residue.

The A1–A4 category ratings simply separate the styles of biomass, where A1 represents generic unused dry biomass and A2 represents woody biomass. It is important to note that the energy reserves of biomass do not come directly from the sum of biomass itself, rather, the actual unused biomass represented here accounts for the subtraction of biomass used for domestic purposes, animal fodder, heating, etc. These values here represent biomass burned by farmers in the field. This biomass is truly unused and serves no other purposes.

The amount of energy available comes from the product of the present supply and the lower heating value (LHV). This is taken through the amount of cultivated land with each of the above‐described crops, as well as the reduction due to moisture. The final available amounts are shown in Table 11.

CategoryType of biomassName of cropCultivated area (km2)Moisture content (%)Total biomass (dry basis) (kt)Used (%)Unused biomass (dry basis) (kt)
A1StrawWheat34,7659.214,317.30802863.45
Barley7045.282036.22
Seasum20619.352015.48
Pulses27431.92806.88
Total23,187.9910,339.73
A2StalkCotton604312707.4731.3486.03
Maize164911.5800.4424.2598.70
Arhar9139.007011.70
Rapeseed and Mustard498142.497042.75
Total1689.431139.18
A3Bagasse*Sugar Cane1441151154.1440‐50577.07
Tops and LeavesSugar Cane59.2940.9960376.40
Total2095.13953.47
A4Cobs*Maize16498.6207.7824.2154.70
Shells*Groundnuts379.870.92360.49
Total2625.881307.95

#### Table 11.

Available energy stores derived from unused biomass [36].

*Indicates processing residue.

With tens of millions of tonnes of unused biomass going to waste, the potential for on‐demand electricity generation is immense. In total, there exists over 200 TJ of unused energy from all the biomass over the entire Punjab region. Obviously, this is only one specific instance, and many other areas may come in well below this, but even at a fraction of the potential, the remaining energy is immense. From this point, the major hurdles become collection and storage. Fortunately, at the microgrid scale, the volumes required are not overwhelming.

The capital costs and operation can be handled in a similar way to renewables or diesel generation, but with a few caveats. Whereas diesel would be purchased, stored, and used by the microgrid operator, the biomass is locally sourced. Two prominent incentives exist for engaging the community and collecting the biomass required to operate the generation unit. Either cash can be directly paid out per kilogram of biomass collected by farmers, or discounts/vouchers for electricity can be distributed.

## 8. Conclusions

It is often said that no single renewable energy source will be able to entirely replace our dependence on fossil fuels, but instead it will take a combination of solar, wind, hydro, biomass, and others to wean our dependence on pollution‐heavy and unsustainable fuels. The idea of a multipart solution is not a new one, and it is certainly not limited to power production. Much in the same way that our energy needs must come from multiple sources, there are multiple levers to pull in order to reduce the cost of rural electrification and bring it within the reach of developing countries.

Just like solar, wind, hydro, and biomass come together to offer a solution to our power production needs, microgrids equally rely not only on multiple generation sources, but also on multiple consumption strategies to increase viability. The microgrid by itself is expensive and clunky and does not utilize new technological developments or advancements to improve on itself. At its core, it is a scaled‐down technology that has not changed in over 100 years.

However, when combined with the techniques and technologies discussed in this chapter, the microgrid takes on a new form. It is no longer wasteful—carrying extra capacity which cannot be used—or allowing itself to be stretched and overloaded leading to total failure. This smart microgrid recognizes its own capacity, stores as much excess energy as possible, and recommends the shifting or clipping of noncritical loads for the benefit of the entire community. Tightening down and preventing electrical waste directly translates to a less expensive, more efficient distribution of energy. Cheaper electricity opens the door to rural electrification where it was once too costly to distribute power, and direct access to inexpensive power provides a boost in quality of life unimaginable to those who have been fortunate enough to have access to a near unlimited supply of inexpensive power.

## References

1 - Kanchev H, Colas F, Lazarov V, Francois B. Emission reduction and economical optimization of an urban microgrid operation including dispatched PV‐based active generators. IEEE Transactions on Sustainable Energy. 2014;5:1397‐1405. DOI: 10.1109/TSTE.2014.2331712
2 - Schmitt L, Kumar J, Sun D, Kayal S, Venkata S. Ecocity upon a hill: Microgrids and the future of the European City. IEEE Power and Energy Magazine. 2013;11:59‐70. DOI: 10.1109/MPE.2013.2258286
3 - Manfren M, Caputo P, Costa G. Paradigm shift in urban energy systems through distributed generation: Methods and models. Applied Energy. 2011;88:1032‐1048. DOI: 10.1016/j.apenergy.2010.10.018
4 - World Energy Outlook Electricity Access Database. World Energy Outlook [Internet]. 2015. Available from: http://www.worldenergyoutlook.org/resources/energydevelopment/energyaccessdatabase [Accessed: October 16, 2016]
5 - Hubble AH, Ustun TS. Scaling renewable energy based microgrids in underserved communities: Latin America, South Asia, and Sub‐Saharan Africa. In: IEEE PES Power Africa Conference; 28 June‐03 July 2016. Livingstone: IEEE; 2016. pp. 134‐138
6 - CORE International Inc. Zambia Rural Electrification Master Plan. Phase 1: Rapid Resource Assessment. Final Report. Technical Report, Core International, 2005
7 - Holland CW, Derbyshire W. Electrifying 20 million households an Indonesian rural electrification strategy. In: EEA Conference and Exhibition; 19‐20 June 2009. Christchurch: EEA; 2009. pp. 19‐20
8 - Income Share Held by Lowest 10%. World Bank [Internet]. 2014. Available from: http://http//data.worldbank.org/indicator/SI.DST.FRST.10 [Accessed: October 01, 2016]
9 - Cruz M, Foster J, Quillin B, Schellekens P. Ending Extreme Poverty and Sharing Prosperity: Progress and Policies. World Bank Group: Policy Research Note [Internet]. 2015. Available from: http://documents.worldbank.org/curated/en/801561468198533428/pdf/101740‐NWP‐PRN03‐Oct2015‐TwinGoals‐Box393265B‐PUBLIC.pdf [Accessed: November 13, 2016]
10 - Murenzi J, Ustun TS. The case for microgrids in electrifying sub‐Saharan Africa. In: IEEE Renewable Energy Congress (IREC); 24‐26 March 2015. Tunisia: IEEE; 2015. pp. 1‐6
11 - Buchana P, Ustun TS. The role of microgrids & renewable energy in addressing sub‐Saharan Africa’s current and future energy needs. In: IEEE Renewable Energy Congress (IREC); 24‐26 March 2015. Tunisia: IEEE; 2015. pp. 1‐6
12 - Solar and Wind Energy Resource Assessment. UNEP [Internet]. 2015. Available from: http://en.openei.org/apps/SWERA/ [Accessed: November 03, 2016]
13 - Pump Price for Diesel Fuel (USD per Litre). World Bank [Internet]. 2015. Available from: http://data.worldbank.org/indicator/EP.PMP.DESL.CD [Accessed: September 20, 2016]
14 - Kagimu V, Ustun TS. Novel business models and policy directions based on SE4ALL global framework for minigrids. In: IEEE International Conference on Technological Innovations in ICT for Agriculture and Rural Development (TIAR); 10‐12 July 2016. Chennai: IEEE; 2016. pp. 1‐6
15 - Bhatia M, Angelou N. Beyond Connections Energy Access Redefined. Energy Sector Management Assistance Program. World Bank; 2015
16 - Modi V, McDade S, Lallement D, Saghir J, editors. Energy Services for the Millennium Development Goals. Washington, DC: World Bank; 2005. p. 116
17 - Barnes DF, Khandker SR, Samad HA. Energy poverty in rural Bangladesh. Energy Policy. 2011;39:894‐904. DOI: 10.1016/j.enpol.2010.11.014
18 - Nussbaumer P, Nerini FF, Onyeji I, Howells M. Global insights based on the multidimensinal energy poverty index (MEPI). Sustainability. 2013;5:2060‐2076. DOI: 10.3390/su5052060
19 - Energy for Community Services. Poor People’s Energy Outlook. Warwickshire: Practical Action Publishing; 2013. p. 71. ISBN: 978 1 85339 789 9
20 - Tenenbaum B, Greacen C, Siyambalapitiya T, Knuckles J. From the Bottom Up: How Small Producers and Minigrids can Deliver Electrification and Renewable Energy in Africa. World Bank [Internet]. 2104. Available from: http://www.esmap.org/sites/esmap.org/files/ESMAP_From%20the%20Bottom%20Up_Optimizedpdf.pdf [Accessed: October 16, 2016]
21 - Franz M, Peterschmidt N, Rohrer M, Kondev B. Mini‐grid Policy Toolkit. Africa‐EU Renewable Energy Cooperation Programme [Internet]. 2014. Available from: http://www.ren21.net/Portals/0/documents/Resources/MGT/MinigridPolicyToolkit_Sep2014_EN.pdf [Accessed: September 30, 2016]
22 - Groh S, Taylor H. The role of microfinance in energy access: Changing roles, changing paradigms, and future potential. Enterprise Development and Microfinance. 2015;26:307‐324. DOI: 10.3362/1755‐1986.2015.026
23 - Ustun TS. The Importance of Microgrids & Renewable Energy in Meeting Energy Needs of the Brazilian Amazon. Unpublished Manuscript. 2015
24 - G. Roberta, T. Esteves. Energy supply on Amazon Remote Communities in Brazil: A case study for the Communities of Terra Nova, Mourao and Sao Sebastiao Methods, Calculatio Formulae. In: International Conference on Microperspectives for Decentralized Energy Supply; 28 February‐01 March 2013. Berlin. 2013. pp. 100‐103
25 - SIN Maps. Operador Nacional do Sistema Eletrico [Internet]. 2011. Available from: http://www.ons.org.br/conheca{\_}sistema/mapas{\_}sin.aspx [Accessed: November 17, 2016]