Cost Estimation of Different Electric Systems for Rural Areas in Developing Countries D. Thirault, Student Member, IEEE, Y. Bésanger, Member, IEEE, and N. Hadjsaid, Member, IEEE A. Almeida, G. Huard The rural electrification current context in developing countries is nevertheless fundamentally different than in the developed countries. By an analysis of consumers needs (residential, commercial and micro-industrial) in rural and suburban areas, we can notice that the daily consumption will be much lower than in developed countries. The remoteness of one preexisting MV grid and the population density are also two different factors in relation to the current situation of developed countries. Additionally, new generation and storage technologies are emerging and their prices are decreasing. Current traditional architectures (large production ; high voltage system ; radial distribution system) are also called into question. These different reasons lead to a reflection on new ways and concepts of electric distribution systems [2]. Additionally, the electricity access seems to be an essential condition for a sustainable economic activity in these rural areas [3]. So we have to find better solutions for a lower cost electrification of developing countries. It is therefore necessary to think about the electric distribution topology, the distribution method (single-phase, three-phases, DC), the location and coupling of energy resources and about the power network’s control. It is also necessary to analyze consumers’ needs in these areas and energy resources available in these countries.

Abstract--The aim of this paper is to present a new methodology to design the electric power distribution network for rural areas of developing countries. Thus, in the context of the current emergence of new generation and storage technologies, this paper describes different means to electrify developing countries ; by MV systems or by autonomous systems with planned location of small generation and storage. There is also a cost quantification of these different architectures by the way of a cost calculation software which has been developed to size and compare the considered solutions. For each strategy, we calculate investment costs and yearly costs (maintenance and energy cost). So the updated assessment over 20 years can be calculated and the strategies comparison gives the best solution for each electrification problem. Index Terms--Costs, Developing nations, Distributed generation, Electrification, Power distribution, Rural areas.

I. INTRODUCTION

C

URRENTLY, more than half of the world population lives in rural areas including almost 90 % (approximately 2.5 billion) in developing countries. The major part of these people uses traditional combustibles like wood by using primary and not very effective technologies. Thus the basic needs of nutrition, heating and lighting are hardly satisfied ; all this contributes to maintain the cycle of poverty. Moreover, this problem is composed of three ways [1] : - Firstly, one billion people is very close to an existing electrical supply network, it is then necessary to determine an inexpensive technique of grid connection and electricity distribution, - Secondly, another one billion people is very far from a power network, an autonomous network operating with distributed production seems to be a good solution, - Thirdly, five hundred million people are not very far from a power network, so these two previous solutions are possible.

II. COST CALCULATION SOFTWARE DEVELOPED A cost calculation software is in the process of development to evaluate and compare the cost of different solutions. Input data are cost informations (equipment cost, energy cost, maintenance cost, …), technical informations (electric efficiency of different generation groups, lines characteristics, …), and case study informations (rural area topology with villages characteristics and different villages remoteness, population needs estimation). As we can see in the software organization (fig. 1), various electrification strategies are considered. A village electrification is made by a MV system (threephases for strategy 1 or two-phases systems for strategy 2) or by an autonomous system (strategy 3) with different dispersed generations (diesel or microturbine and renewable energy) [4]. For each strategy, we calculate the investment cost and yearly costs (maintenance and energy cost). So the updated assessment over 20 years can be calculated and the strategies comparison gives the best solution for the considered problem. The description of this software is given below :

This work was supported in part by the GIE IDEA. An economic corporation between EDF (Electricité de France), INPG (National Polytechnic Institute of Grenoble) and Schneider Electric. The research theme is on the electric distribution systems. D. Thirault is with GIE IDEA, ENSIEG, BP 46, Saint Martin d’Hères F38402 France (e-mail: [email protected]). Y. Bésanger and N. Hadjsaid are with LEG (Power Engineering Laboratory of Grenoble), ENSIEG, BP 46, Saint Martin d’Hères F-38402 France(e-mail:[email protected] and [email protected]). A. Almeida is with Schneider Electric, A2, 9 quai Paul Louis Merlin, F38050 Grenoble cedex 9 (e-mail: [email protected]). G. Huard is with EDF DRD, 1 av du Gal de Gaulle, F-92141 Clamart Cedex (e-mail: [email protected]).

0-7803-7519-X/02/$17.00 © 2002 IEEE

23

Village characteristics and localisation

Daily load curve

Strategy 1

Strategy 2

Strategy 3

Str 2, (nbtr-2) : Str2, (nbtr-0) :

Str 2, (nbtr-1) :

-

-

nbtr transformers without decentralized production

-

nbtr-1 transformers, decentralized production

nbtr-2 transformers, decentralized production

Installed power calculation : transformers, decentralized production

LV power systems characteristics from a two phase transformer

For the 1st transformer, LV system calculation : (lenght, lossess, energy from the transformer, energy from the decentralized production)

For Str2, (nbtr-1) : cost calculation : - investment, - operation, - updated assessment.

Materials and energy cost

Strategies comparison and choice Fig. 1. Cost calculation software organization described here for Strategies 2.

2) Generation data In this study, we have considered generation by diesel group, generation by microturbine and wind power systems. These characteristics are described in table 3:

A. Input data In a standardization way of equipment, we have limited input materials which are power network data, power generation data, consumption data and variable data. 1) Power network data a) LV lines Considered LV lines are overhead twisted wires (three phases for section 35 mm², 70 mm² or 150 mm² and one-phase for section 16 mm²) which are the less expensive LV wires. MV system is overhead with a 70 mm² section. We have considered that the system is an European type. There are nevertheless two subsystems, a two-phases or a three-phases. LV and MV lines characteristics are described in table 1:

Section Cost ( NP

Diesel group Microturbine Wind system

TABLE IV HOUSEHOLD APPLIANCES CHARACTERISTICS Low consumption Normal consumption appliances appliances Power (W) Cost ( Power (W) Cost ( 1 light 10 5 40 1 1 radio set 20 15 20 15 1 TV set 70 150 70 150 1 fridge 30 l 40 150 100 100 1 fridge 150 l 80 250 250 200 household appliances 100 50 100 50 cosPhi 0,95 0,97

b) MV/LV transformers We have considered two MV/LV transformers types; a 100 kVA transformer for the three-phases MV system and a 25 kVA transformer for the two-phases MV system. Transformers characteristics are described in table 2: Transformers 100 kVA 25 kVA

Renewal cost 0.025 N:K 0.027 N:K 0.019 N:K

3) Consumption data In order to determine theoretical load diagrams for various types of villages, we define various services adapted to the users needs. Each service is composed of different household appliances. a) Different household appliances The table 4 describes domestic appliances input:

TABLE I LV AND MV LINES CHARACTERISTICS LV systems MV systems 3*150² 3*70² 3*35² 2*16² 3*70² 2*70² 13000 11000 5000 3000 18000 14000

TABLE II MV/LV TRANSFORMERS CHARACTERISTICS Cost ( io (%) Piron (W) Rsec (Ω) 8000 2,5 210 0,035 5000 3,1 75 0,11

TABLE III GENERATION CHARACTERISTICS Investment cost Maintenance cost 300 N: 0.018 N:K 800 N: 0.003 N:K 1000 N: 0.006 N:K

Xsec (Ω) 0,054 0,12

24 0-7803-7519-X/02/$17.00 (C) 2002 IEEE

b) Different services (1) Domestic services We have defined various types of domestic services :

-

TABLE V SERVICES PROPOSED TO DOMESTIC USERS Types of services S1 S2 S3 S4 S5 (number of …) Light 1 2 3 3 3 Radio set 1 TV set 1 1 Fridge 30 l 1 Fridge 150 l Household appliances

-

S6

S7

5

5

1

1

1

1 1

-

the fossil energy (diesel or gas) cost, it is the local energy cost and we have considered that this cost can vary between 0.0038 N:KDQG N:K the village geometric characteristics, households number, village type choice (1 to 6) which is an image of the area wealth, the use of low consumption appliances or not, updated assessment duration (20 years).

B. Various studied strategies We have studied 3 types of strategies to bring electricity to a rural area: - Str1 strategies : electrification by a MV three-phases system with the possibility to use dispersed generation, - Str2 strategies : electrification by a MV two-phases system with the possibility to use dispersed generation, - Str3 strategies : autonomous electrification by a LV local system with dispersed generation. 1) Str1 strategies These strategies are decomposed in sub-strategies str1,(nbtr-x) in which we install (nbtr-x) transformers and some distributed generation installed in LV systems for the peak power filling. Nbtr is the number of transformers to install when we don’t use dispersed generation (str1,(nbtr-0) strategy). MV system configuration is described in fig. 2. In this case, we have considered that the area has a peak power comprised between 300 kW and 400 kW.

(2) Collective services We consider that the different collective services for a village are: - tworship and meeting places illumination, - water’s pumping station and water’s purification (basic service of 10 liters of water per homes per day), - the medicine preservation. (3) Micro-industrial and tertiary services We consider that micro-industries needs are : - the stores’ illumination, - the food preservation and the cooking, - the ice blocks making, - the use of small engines (crushing, husking equipment). Thus, all these services lead to a profitable economic activity. So we have estimated the consumption of this activity as a 20 % increase of the global village consumption between 8:00 am and 8:00 pm. c) Different types of villages We have defined various theoretical types of villages (very poor village to fairly rich village) by choosing various repartitions of services by village.

S1 S2 S3 S4 S5 S6 S7

Village 1 0,6 0,2 0,1 0,099 0,001 0 0

P = 100 kW

TABLE VI PERCENTAGE OF SERVICES BY VILLAGE Village 2 Village 3 Village 4 Village 5 Village 6 0,1 0,1 0,05 0,05 0 0,1 0,1 0,05 0,05 0,01 0,2 0,3 0,1 0,05 0,02 0,5 0,3 0,3 0,15 0,02 0,09 0,1 0,45 0,4 0,8 0,01 0,05 0,025 0,1 0,1 0 0,05 0,025 0,2 0,05

Length of MV connection

Zone 1

Zone 2

Zone 3

Zone 4

Fig. 2. MV system configuration, str1,(nbtr-0) for a 300 to 400 kW area

In Str1 strategies, LV systems are composed by a three phases framework and by single phase 16² branches.

LV

d) Internal equipment We have considered that the internal household equipment is installed and paid by the utility which manage the local electricity market. For each home, we have installed : - a differential switch or a differential circuit breaker for persons’ protection, - fuses for short circuits protection, - a metering system to guarantee that the customer is in agreement with the signed contract. The considered investment cost is 30  4) Variable input data Variable input data of the software are: - the MV line length to connect the area to be electrified, - the electric power cost ; it is the cost at the level of a substation which can be varied between 0.038 N:K and 0.152 N:K

35² or 70² or 150²

16² MV

Transformer 100 kVA

House Fig. 3. LV system configuration for str1 strategies

2) Str2 strategies Str2 strategies are an electrification by a MV two-phases systems with the use of dispersed generation for the peak power filling. Consequently, MV system configuration is the same than in str1 strategies (there are only more transformers: 25 kVA compare to 100 kVA). LV systems are composed by a single phase framework and by single phase 16² branches as it can be seen in fig. 4.

25 0-7803-7519-X/02/$17.00 (C) 2002 IEEE

- lengths of LV lines (three-phases and single-phase), - lengths of MV lines, - transformers number, - production group. The yearly operation cost can be calculated from the previous calculation of: - electric energy from the MV systems, - fossil energy from microturbines and diesel groups, - wind energy. To compare the different strategies, we have to calculate the updated assessment over 20 years: 20 Yearly _ cos t (1) year Updated _ assessment = Investments +

Distributed generation LV 16² 16²

Transformer 25 kVA

House

Fig. 4. LV system configuration for str2 strategies

3) Str3 strategy Str3 strategy is an electrification by an autonomous system with dispersed generation [5] [6] (diesel group and wind system). There is no MV system and LV system have the same configuration than in str1 strategies (fig. 2. without transformers).

∑ (1 + actu )

year =1

year

In (1), actu is the actualization rate (actu = 0.1). This updated assessment gives for each strategy the updated kWh cost for the utility.

C. Calculation principles

III. ANALYSIS AND RESULTS

The principle software operation is described in fig. 1. where we put the example of Str2 strategies. 1) Load curve calculation Firstly we calculate the daily load curve of the rural area. This load curve is defined by a database on various appliances consumptions, the choice of the village type (image of the average incomes) and the number of households. From this load curve, we can define the maximum power of the area (Pmax). Pmax will be the dimensioning element of LV systems. 2) LV systems design The design of LV system is based on a predefined configuration of the power network (see fig. 5.) and depends on Pmax, and on the homes’ number. The design is automatically implemented in the way of respecting voltage drop constraints and accepted current constraints.

The cost evaluation of various strategies can be determined according to variable inputs. We choose various parameters for the results analysis as the electric power cost, the fossil energy cost, and the MV connection length. All these strategies are compared with the French traditional solution ie a connection to three-phases MV systems and a three-phases LV system. In this part, we show principal results obtained as : - Light low consumption interest and first general results, - Electrification choice for various strategies. A. Light low consumption interest In this part, we quantify the light low consumption interest for each strategy. To show this interest, we recall that the lamps cost is supported by the energy operator. 1) Description of the considered village The considered village is a type 3 village of 1000 households. It’s an area of 700 m by 400 m. The length connection to a MV system is variable. 2) Daily load curve Fig. 6. show the daily load curve for this village with or without the use of low consumption lights (LCL).

y

LV 35² or 70² or 150²

H LV 16² LV 16²

Daily load curve

Power (kW) 160 140 120 100 80 60 40 20 0

x

U niform repartition of n households

1

3

5

7

9

11 13 15 17 19 21 23

Daily load curve

Power (kW) 160 140 120 100 80 60 40 20 0 1

3

Hours

L

5

7

9

11 13 15 17 19 21 23 Hours

Fig. 5. LV systems design for Str1 and Str3 strategies

Without LCL Fig. 6. Daily load curve for the considered village

3) Losses calculation We have designed the LV systems and the load curve calculation ; so we can estimate losses in the LV systems and in the MV/LV transformers (for str1 and str3 strategies). The energy produced by the dispersed generation and MV systems is then estimated for each strategy. 4) Cost calculation This design allows us to calculate the initial investment cost which takes into account: - internal equipment,

As we can see, the peak power is three times lower when LCL are used and the daily energy consumed is two times lower. The following results show impacts of these differences on the different strategies cost. 3) Impacts on the initial investments Fig. 7. and fig. 8 show LCL impacts on the initial investments for the considered village. These bar charts resume the investment from the bottom to the top in household appliances, LV three phases wires, LV single phase wires, MV lines, transformers and distributed generation.

26 0-7803-7519-X/02/$17.00 (C) 2002 IEEE

With LCL

Initial investments costs

Updated assessment for different strategies

Cost (k )

Cost (k )

140

800 120

600

100 Small generation

North European systems

Transformers

80

MV lines

400

LV single phase lines

60

Str 1 Str 2 Str 3

LV three phase lines Household appliances

40

200

20

0 0

0 European systems

Str 1

Str 2

5

Initial investment cost

Fig. 10. Updated assessment with LCL

The updated assessment comparison show an important decrease of the final updated cost (30 %) for each strategy. So, the use of low consumption appliances seems to be a very interesting solution for the electrification of developing countries. Fig. 10. show that we will electrify this village: - by a two-phase MV systems if the connection length is less than 4 km, - by an autonomous system if the connection length is higher than 4 km (constant cost of 240 k  As we can see, a two-phases system seems to be always advantageous than a three phases systems (Str1 or North European type).

140 120 100

Small generation Transformers MV lines Single-phase LV lines

80 60

Three phases LV lines Household appliances

40 20 0 European systems

Str 1

Str 2

Str 3

Fig. 8. Initial investment with LCL

These figures show the investments repartition for the different strategies. In these cases, we consider that there is a preexistent MV systems which crosses in the area. So the MV system cost is not very high. The comparison of the initial investment with and without LCL show a little decrease of the initial cost even if the internal equipment cost is higher. This cost reduction is due to a less constraining dimensioning of the production system and of the distribution system. 4) Impacts on the updated assessment Fig. 9. and fig. 10. show LCL impacts on the updated assessment for this village and nearby, the significant cost reduction given by LCL installation. These results are shown for medium energy cost (0.076 N:KIRUHOHFWULFLW\FRVWDQG 0.0076 N:KIRUIRVVLOHQHUJ\FRVW 

B. Optimal strategy choice for different villages As we know, there is a limit distance for a village to be connected at a MV system. In this part, we show the results obtained for the electrification of different villages. We have considered three villages : - village a: 50 households, - village b: 200 households, - village c: 1500 households, which are type 3 villages. We consider that we install low consumption appliances and we show the results for different energy cost: - energy 1 : Electricity = 0.038 N:KIRVVLOHQHUJ\ 0.038 N:K - energy 2 : Electricity = 0.076 N:KIRVVLOHQHUJ\ 0.0076 N:K - energy 3 : Electricity = 0.152 N:KIRVVLOHQHUJ\ 0.0038 N:K

Updated assessment for different strategies Cost (k )

800 600

North European systems Str 1 Str 2 Str 3

400

200

Energy 1 Energy 2

0 0

5

Energy 3

15

10

15 Length of MV connection (km)

Fig. 7. Initial investment without LCL Cost (k )

10

Str 3

TABLE VII OPTIMAL STRATEGY CHOICE FOR DIFFERENT VILLAGES Village a Village b Village c L13000m : Str3 L3500m : Str3 L700m : Str3 L>400m : Str3

This table show the limit connection distance to a MV system for different villages as a function of the energy cost. So when the electricity cost increases and the fossil energy decreases, the maximum length of connection decreases. For a very small village (like village a), this critical distance is very small (from 2 km to 0.7 km).

Length of MV connection (km)

Fig. 9. Updated assessment without LCL

27 0-7803-7519-X/02/$17.00 (C) 2002 IEEE

IV. CONCLUSION

VI. REFERENCES [1]

This paper deals with the problem of rural electrification in developing countries. We have several ways of electrification as by MV systems with different means of connection (threephases or two phases) or by an autonomous system with distributed generation (diesel groups, microturbines and renewable energies). A cost calculation software is developed. It can compare these different strategies of electrification. For a considered village, the tool builds automatically the structure adapted to a daily load curve in the way to respect constraints in voltage drops and admissible currents. After this calculation we can compare each strategy and choose the best one. The choice criteria is the updated assessment over 20 years. So we can determine a critical length of connection to a MV system for each strategy. From the different results of this paper, one can see the interest of using low consumption appliances which reduce the investment and operation costs of the power delivery system (for a MV system or for an autonomous system). We can also determine a critical length of connection to a preexistent MV system for different villages and different variable inputs. So this critical length gives the optimal solution for the considered problem. We continue the development of this software and we will analyze others solutions like the use of storage or others renewable energies. Generation systems should be also optimized in a better way and after we want to extend calculations to a higher area of electrification.

[2]

[3]

[4]

[5]

[6]

World Energy Council, “The challenge of rural energy poverty in developing countries,” October 1999, Tech. Rep., Available: http://www.worldenergy.org/ D. Thirault, Y. Bésanger, N. Hadjsaid, F. Dumas and G. Huard, “A meshed distribution system for the electrification of rural areas in developing countries,” presented at the PSCC’02 Conference, Sevilla, Spain, June 24-28, 2002. R. Ramakumar, T.J. Hammons, A. Obozov, V. Kirilov, M. Berdybaeva, J. Gutierrez-Vera, “Renewable energy technology alternatives for developing countries,” IEEE Power Engineering Review, April 1998. R. Ramakumar, I. Abouzahr and K. Ashenayi, “A knowledge-based approach to the design of integrated renewable energy systems,” IEEE Transactions on Energy Conversion, Vol. 7, No. 4, December 1992. A. G. Bakirtzis, P. S. Dokopoulos, “Short term generation scheduling in a small autonomous system with unconventional energy sources”, IEEE Transactions on Power Systems, Vol. 3, No. 3, August 1988. R. Chedid, S. Rahman, “Unit sizing and control of hybrid wind-solar power systems”, IEEE Transactions on Energy Conversion, Vol. 12, No. 1, March 1997.

VII. BIOGRAPHIES Damien Thirault (Student Member) received the Electrical Engineering degree from Ecole Nationale Supérieure d’Ingénieurs Electriciens de Grenoble (France). He is presently working for the Dr. Eng. Degree in Electrical Engineering from the Institut National Polytechnique de Grenoble (INPG) at the Laboratoire d’Electrotechnique de Grenoble (LEG). His research interests are Distribution Power Systems and Distributed Generation particularly for developing countries. Yvon Besanger (Member) received the Dr. Eng. in Electrical Engineering from the INPG in 1996. He is currrently an Associate Professor at ENSIEG (Ecole Nationale Supérieure des Ingénieurs Electriciens de Grenoble) and LEG. His research interests are Distribution networks operation and reliability, FACTS Devices, and Power System Dynamic Stability,. Nouredine Hadjsaid (Member) received the Dr. Eng. in Electrical Engineering and “Habilitation à Diriger des Recherches” Degrees from the INPG in 1992 and 1998, respectively. He is currrently a Professor at ENSIEG and LEG. His research interests are power system operation and security. A. Almeida is graduated in Electrical Engineering from Ecole Nationale Supérieure de l’Electronique et de ses Applications (ENSEA), CergyPontoise, France. He received the Diploma Degree (DEA) in Electrical Engineering from Université Pierre et Marie Curie, Paris, France, in 1992. He is working as a research engineer with Schneider Electric, Grenoble, France. His fields of interest are distributed power systems and power electronics.

V. ACKNOWLEDGMENT The authors gratefully acknowledge the contributions of F. Dumas and G. Marboeuf for their work on this research subject.

G. Huard is working as a research engineer with EDF. He works at the “Division de la Recherche et du Développment”. His research interests is the use of renewable energies for Decentralized Rural Electrification.

28 0-7803-7519-X/02/$17.00 (C) 2002 IEEE