Economic analysis of solar energy development in North Africa

1 Introduction

The global high level of solar irradiation intensity region mainly concentrated in the 10°north latitude to 35°north latitude, and the annual solar irradiation intensity is between 1800kWh/m2 to 2600kWh/m2. Hence, the resource of solar energy is rich in North Africa, and the potential is quite large to build solar power generation base in the most of North Africa region countries, such as Morocco Tunisia, Algeria, Egypt[1]. In recent years, North African economy is continued to grow steadily and energy demand is accelerated. But the existing energy and electric power situation in this region cannot adapt to the development state. The research of solar energy technical status,development trend and technical route will help solar power generation to be developed orderly in the world, especially in North Africa. The scheme of developing large scale solar energy resource in North Africa, and transmitting clean energy power through transmission channel between North Africa and Europe, which has the typical characteristics of global energy interconnection(GEI). It is an important opportunity to promote the interconnection of Asia, Africa and Europe and realize the GEI concept [2].

The main types of large scale solar energy power technology are photovoltaic(PV) power generation and concentrating solar power(CSP) generation. PV power generation technology is relatively mature, and its development objects are improving the efficiency and reducing the cost through material and basis technical innovation. CSP technology converts solar energy into thermal energy, and then drives generator through turbine to generate electric power. Compared to PV technology,CSP with thermal energy storage unit can smooth output power obviously, and generate power after sunset to meet night load demand or to meet the peak power regulation demand of power system. It can increase utilization hours and improve regulation performance. On the other hand,in terms of economy, CSP is still in the demonstration stage, less large scale application and engineering operation experience result in high cost of CSP project consulting,EPC and equipment.

Recently, the research of technical and economic comparison of PV with energy storage battery and CSP with thermal energy storage is less, particularly few studies in optimal capacity configuration and economic analysis of solar energy combined power generation. PV and CSP power stations can complement each other through the peak load shifting of energy storage system. From the point of view of power balance, the ratio of installed capacity of solar stations and ES system should be 1：1, but the economy was not considered[3]. Energy analysis method of PV and CSP combined generation system was proposed in [4], and the scheduling model and method of CSP, PV and wind power combined generation were given based on schedulability of CSP station with thermal storage in [5].Although the benefits of combined station connecting into grid were analyzed in these researches, the comprehensive economy was not considered. Some studies [6-7] compared technical economy and environmental evaluation between PV and CSP, but the comprehensive economy of combined generation was not assessed. For other forms of renewable energy generation, the complement of wind power and CSP was analyzed, as well as invest cost was evaluated,it can provide reference for CSP and PV combined power generation[8]. There is less research on how to configure CSP and PV stations capacity optimally from economy viewpoint.

It is necessary to study the influence of future solar energy large-scale development and utilization on the economy of CSP power generation, the advantage complementary of PV and CSP, as well as the technology and economy of combined power generation mode. It is of great significance to improve the utilization of solar energy and promote the development of solar energy industry

2 Solar power generation cost development trend

2.1 PV cost development trend

2.1.1 PV power generation cost composition

The initial investment of PV station can be divided into the cost of PV module, inverter, power distribution equipment,cable, station construction, etc. The cost in PV modules is about 50% of the initial investment; inverter including isolation cost is about 10% of the initial investment; the cost of power distribution equipment and cable is about 17%;the cost of station design, construction, installation and debugging, grid connection, labor charges is about 23%.

(1) Photovoltaic modules cost

Solar cell, which is also called Photovoltaic module,is the core part of photovoltaic power generation system.Recently, with the scale of solar battery is expanding rapidly, the efficiency of PV module is also greatly improved. The annual absolute efficiency of solar cells increases by about 0.3%. At the end of 2015, the average efficiency of polycrystal silicon and monocrystal silicon solar cells in China was 18.3% and 19.5% respectively.

The improvement of PV module efficiency and manufacturing process and the decrease of raw material price will lead to the future decrease of PV power generation cost. Some research and calculation results show that the efficiency of PV modules is increased by 1%, which is equivalent to 17% reduction in the price of PV power generation system. Therefore, with solar cell efficiency rising, PV module costs are also falling sharply.

The average price of crystal module in China is about 0.568$/W (3.86yuan/W) in 2015, the direct manufacturing cost of monocrystal silicon module is about 0.5$/W(3.4yuan/W), while the manufacturing cost of polycrystal silicon module has fallen to below 0.48$/W (3.26yuan/W).Under the same conditions, the average manufacturing cost of module is 0.68 to 0.70$/W in America. The global PV industry is increasingly concentrated in a few countries and regions due to the impact of manufacturing costs. Chinese mainland, Taiwan, Malaysia and the United States are the world’s top four major PV manufacturing industries regions. Expected in the next three to five years, the cost of crystalline silicon solar cells will drop to about0.4$/W(2.72yuan/W) in China.

(2) Photovoltaic grid-connected inverter cost

The manufacture of inverter is very mature and the cost is not affected by raw material price, therefore, it is unlikely that the inverter will break the technology in the short term, and the cost of the manufacture will not change.Inverter price is mainly influenced by the market supply and demand and the outlook of the PV industry. China plans to achieve grid parity for solar energy in 2020, it will push downstream suppliers on lowering their price every year. The unit price of 500kW bulk centralized inverter is 0.029$/W, and the unit price of 50kW series inverter is about 0.044$/W.

(3) PV station construction cost

The initial investment of the ground PV power station accounted for most of the total cost of PV station. The cost of land relative to the whole cost of construction and maintenance is generally small, and its impact is not considered. According to the cost of PV module, inverter and other equipment, the unit kW investment of PV station has been also declining in recent years. The cost level of the PV power station is basically within the range of 7,500 to 9,000 yuan/kW in 2015[9].

The progress of PV power generation technology and the development and popularization of efficient batteries or other new batteries will improve module conversion efficiency and extend service life. It will lead to further drops in the cost of PV power generation. Based on the international price reduction of PV cell industrial chain and efficiency improvement of PV module, International Energy Agency(IEA) forecast that the average initial investment cost of PV station will fall to 4,500-6,000 yuan/kW in 2020, and 3,000-4,200 yuan/kW in 2030[10].

2.1.2 PV levelized cost of electricity(LCOE)development trend

According to the trend of PV power generation LCOE in the past decade, the international analysis agencies forecast the cost of PV power generation in the major countries by 2020 and 2030. The prediction data are shown in Table1 and Table 2 [11].

The large scale development effect is the driving force of the LCOE reduction of PV power generation.The impact of this cumulative production on the cost can be described using the learning curve which reflect the function relationship between the cost of an energy technology unit and its total output. The basic form is：

Table 1 Predicted value of the LCOE of PV generation in 2020

IEA Bloomberg New Energy Finance Europe EPIA Japan NEDO NERL China Renewable Energy Committee China9 cent/kWh7.36 cent /kWh— — —0.6-0.8 yuan/kWh Japan — 11.47 cent /kWh—14 yen/kWh— —India—6.83 cent /kWh— — — —USA10 cent /kWh7.06 cent /kWh— —10cent/kWh —Germany— 9.36 cent /kWh12 cent euro/kWh— — —

Table 2 Predicted value of the LCOE of PV generation in 2030

IEA Bloomberg New Energy Finance Europe EPIA Japan NEDO NERL China Renewable Energy Committee China 6 cent/kWh 4.72 cent/kWh — — — 〈0.6 yuan/kWh Japan— 7.24 cent/kWh—7 yen/kWh— —India—4.97 cent/kWh— — —USA7 cent/kWh4.07 cent/kWh— —8 cent/kWh—Germany— 6.49 cent/kWh6 cent euro/kWh— — —

Where C ( xt )is the unit cost for a certain technology in the tth year; C ( x0)is the initial unit cost for a certain technology in the benchmark year; xt is accumulative scale of a certain technology in the tth year; x0 is accumulative scale of a certain technology in the benchmark year; a is cumulative scale elastic coefficient.

Define technical learning rate, which represents the reduction proportion of technical cost when the cumulative scale doubles.

Define technical progress rate P R=2-α, which represents the proportion of technical cost to initial cost when the cumulative scale doubles.

Since the technology development process of the PV industry is different, the falling proportion of technology cost during the doubling of the scale of the PV industry is changing in different period. Hence, the LCOE of PV power generation in 2020, 2025 and 2030 are predicted by using the variable learning rate method, learning rate are L1=0.25 from 2008 to 2012, L2=0.29 from 2013 to 2017 and L3=0.27 from 2020 to 2030. The input data for formula(1)are shown in Table 3, and the forecast results are shown in Fig. 1.

Table 3 Input data for LCOE prediction of PV

Year 2009 2010 2011 2012 2013 2014 2015 2016 2017 2020 2025 2030 Accumulative scale（GW） 23.19 40.34 70.47 100.48 138.83 179.03 229.13 299.13 384.43 540 756 1721 LCOE（$/kWh） 0.3 0.24 0.19 0.15 0.145 0.135 0.115 0.095 0.08 0.072 0.062 0.042

Fig. 1 Forecast curve of the LCOE of PV

2009 is chosen as benchmark year for prediction.From 2009 to 2017, the LCOE calculated value by formula(1) is close to Bloomberg New Energy Finance statistic value. Meanwhile, the LCOE predicted value is 0.072 $/kWh(about 0.49yuan/kWh) in 2020, and 0.042$/kWh(about 0.286yuan/kWh) in 2030. These predicted values are also close to Bloomberg New Energy Finance predicted values in Table1 and Table2. It is shown that this method can predict the development trend of PV power LCOE accurately, and the prediction curve can be used as a useful reference to describe the future trend of PV power generation LCOE.

2.2 CSP cost development trend

2.2.1 CSP power generation cost composition

The construction cost of CSP power station is divided into three parts： cost of equipment, installation and civil construction. Take the 50MW tower CSP power station as an example, the whole investment can be divided into the cost of the solar island, thermal power generation island, thermal storage system, the preparation fee of the site, the supporting facilities of the power station and the cost of infrastructure facilities and indirect expenses.The proportion is 61%, 15%, 17%, 3%, 3% and 1%,respectively. As the installed capacity of CSP power station increases, the proportion of solar island cost is higher and higher (When the station installed capacity is 300MW and 600MW, the proportion of solar island cost can reach 68%and 70% respectively)[13].

(1) Solar island construction cost

The largest part of the construction cost of the tower solar thermal power station is the heliostat. Its cost has a great impact on the cost of CSP power station as well as LCOE. The cost of heliostat is divided into three parts：material cost is about 50%, processing cost is about 40%and transportation cost is about 10%. In terms of material cost, the heliostat is developing in a less material and lighter direction, in order to increase reflectivity of mirror and reduce raw material cost of reflector. For processing cost,with the mature manufacturing process and the scale effect of batch manufacturing, the processing cost of the parts will be reduced greatly. Meanwhile, the scale effect also reduced the cost of heliostat installation and transportation.It is estimated that when the installed scale reaches 2GW/year, the cost of the heliostat is expected to decline by more than 55%.

In the manufacturing cost of heat absorber, the proportion of materials and installation is 80%, and the processing cost is about 20%. Therefore, the scale effect can only bring about 10% to 20% of the cost reduction.

Considering the cost expectation and proportion of each subsystems of solar island, , the unit thermal power cost of solar island will be reduced from 3,600-4000 yuan/kW to 1,800yuan/kW, with the emergence of CSP power generation scale effect.

(2) Thermal power generation island cost

The construction cost of thermal power generation island and site and infrastructure can be calculated according to the traditional coal-fired power station. For small units of coal-fired power stations, the current cost is generally about 6,000 yuan/kW. For the medium and large units of 300MW and above, the cost can be reduced to 3,000 to 4,000 yuan/kW. But there is no unit capacity of 300MW CSP stations recently, because of some technical bottlenecks.It is expected that the cost of thermal power generation island with using capacity of 100MW unit is about 3,000 yuan/kW.

(3) Thermal energy storage(TES) system cost

The cost of the thermal energy storage system is related to the installed capacity of the power station and thermal storage full load hours(FLH). There are only a few cases of running molten salt TES for tower CSP power stations,and the process of molten salt system still has great improvement space. Recently, the unit thermal energy capacity cost of TES system is about 160-200 yuan/kWh.

(4) Operation and maintenance cost

Operation and maintenance expenses mainly include repair expenses, personnel salary and welfare, materials,water and other expenses. At present, the O&M cost rate of CSP power station is estimated based on the experience of traditional thermal power station, and it accounts for 2% of fixed asset investment each year.

2.2.2 CSP levelized cost of electricity(LCOE)development trend

According to the international renewable energy agency(IRENA)research, during 2010-2015, the cost of CSP power station decreased mainly due to falling investment costs, which fell by 18% to 22%, whereas the improvement of efficiency lead to cost decreased during 2015-2020, falling by10%-15%. Based on prediction during 2020-2030, the scale economies effect will be the main reason for cost reduction, and the decline range will about from 21% to 33%[14].Table 4 gives the predicted value of the cost of CSP power station with different FLH of TES from NREL[15].

Table 4 Predicted value of the cost of CSP power station with different FLH of TES from NREL ($/kWh)

Y e a r 2 0 1 5 2 0 2 0 2 0 2 5 2 0 3 0 9 F L H 8 2 0 0 7 1 5 0 6 5 5 0 6 1 0 0 6 F L H 6 9 0 0 6 0 0 0 5 5 0 0 5 1 5 0 3 F L H 6 2 0 0 5 6 0 0 5 0 0 0 4 6 0 0

Table 5 Predicted value of the LCOE of CSP generation in 2020

T y p e A g e n c y U S A C h i n a P a r a b o l i c t r o u g h s t a t i o n I E A 0.1 0-0.1 4$/k W h —D O E 0.1 0-0.1 1$/k W h —

Continue

Type Agency USA China Parabolic trough station IRENA 0.13$/kWh 0.10$/kWh Tower station DOE 0.08-0.09$/kWh —IRENA 0.16$/kWh 0.08$/kWh

Agency such as IEA, DOE and IRENA choose different scenarios of cost reduction to predict CSP LCOE in 2020 with low level and high level. The predicted value are shown in Table 5[11]. We can see that the LCOE of parabolic trough station will drop to 0.1-0.14$/kWh, while tower station will reach about 0.08-0.16$/kWh.

Due to the effect of scale is greater, learning curve also can be used to predict construction cost and LCOE of CSP power station. The input data for prediction are shown in Table 6. The learning rate for cost prediction is L1=0.25 and for LOCE prediction is L1=0.12

Table 6 Input data for cost and LCOE prediction of CSP

Year 2015 2016 2017 2020 2025 2030 Accumulative scale（MW） 4940 5017 6200 10200 41000 69000 Cost（$/kW）3h 6200 6173 5810 5553 4494 3801 6h 6900 6870 6466 6180 5002 4230 9h 8200 7684 7621 7344 5945 5027 LCOE($/kWh)3h 0.140 0.139 0.134 0.125 0.102 0.091 6h 0.128 0.127 0.123 0.115 0.093 0.084 9h 0.120 0.119 0.115 0.108 0.087 0.078

Fig. 2 Prediction curves of the unit cost of CSP with different FLH of TES

Fig. 3 Prediction curves of the LCOE of CSP with different FLH of TES

We can see from the two figures, the unit cost and LCOE predicted values before 2020 are fairly close to NREL’s predicted values, whereas the LCOE predicted value in 2020 is close to IRENA’s predicted value for parabolic trough station. These results show that the predicted costs of the CSP power station from different agencies are similar in the short term.

From 2020 to 2030, the predicted results based on learning curve are lower than NREL’s predicted values.Hence, taking all factors and predicted values into account,the unit cost will fall to 5,000 to 6,000 $/ kW around 2030,and the LCOE will fall to below 0.08$/kWh(the FLH of TES is more than 9 hours ).

3 Solar resource development potential in North Africa

3.1 Overview of solar resources in North Africa

According to the regional classification of solar thermal utilization, North Africa is one of the highest level solar energy resources in the world[16]. Morocco, Algeria,Tunisia and Libya are all have huge potential to exploit solar energy. Fig. 4 shows the average DNI distribution map in North Africa. The total solar annual irradiation in Algeria is 2,700kWh/m2, and the amount of development technologically is about 169,440TWh per year. In Morocco,Egypt and Tunisia, the total solar annual irradiations are 2,600kWh/m2, 2,800kWh/m2, and 2,300kWh/m2 respectively.

3.2 Solar resources in Tunisia

The solar energy resources in Tunisia are abundant,and the total amount of solar irradiation is greater than 2300kWh/m2. The annual average DNI distribution in Tunisia is shown in Fig. 5.

Fig. 4 Distribution diagram of annual average DNI in North Africa region

Fig. 5 Distribution diagram of annual average DNI in Tunisia

The most abundant solar resources in Tunisia are in the south, where average DNI is over 2,400kWh/m2. It’s generally higher than in northern Tunisia. Some researchers suggested the threshold value of 2000 kWh/m2/year for the annual DNI for solar thermal power generation[17]. Therefore, the most area of the center and the south of Tunisia have annual DNI suitable for the CSP plant running.

CSP plants require large areas of land for deployment of solar field, power block and storage component.Requirement of land may vary according to the CSP technology used and the extent of storage with the plant.The required land is expected to be available in abundance(and perhaps at reasonable cost) in arid and semi-arid areas of the country with high DNI [18]. As the center and the south regions of Tunisia are semi-arid and arid with about 40% of the country territory are desert, there are very large waste land. Most of these lands are with slope less than 2%and therefore suitable for CSP plants [1].

4 The economy of solar energy development in North Africa under the background of Africa-Europe interconnection

4.1 Conception of Africa-Europe interconnection

We can take Tunisia as a sample country of solar energy development, because solar resource is abundant and developable scale is large. But Tunisian economic development and load growth are slow, and result in local consumption of electric power is difficult. Besides, Tunisia is closer to Europe, and the conception of transmitting solar energy to European countries as clean power through Africa-Europe Interconnection is feasible.

4.1.1 Power generator planning scheme

Considering same factors, such as development potential of solar energy, technology maturity of energy storage and CSP, PV with energy storage batteries and PVCSP combined power generation can be chosen as planning schemes. On the one hand, it can relieve the intermittent power of solar power generation and improve the power quality. On the other hand, compared with only CSP power generation system, it can also reduce the cost of power stations.

4.1.2 Boundary conditions of planning

(1) Designing operation curve of DC transmission engineering based on solar energy resource, European load characteristic and load curve;

(2) The transmission capacity is about 2 GW based on development scale of solar energy in Tunisia;

Fig. 6 Operation curve of DC transmission with 5500h

(3) The utilization hours of DC transmission engineering are about 5500h, and the curtailment rate of PV power is less than 5%;

(4) Ensuring the technical economy of the project with configuration of power generators reasonably and economically.

4.2 PV and energy storage(ES)

PV and ES can store the spare solar energy of the PV generation during the day, and use the ES to meet the demand of the electricity from the DC project at night when PV is out of power. According to some relevant data, the average power curve of PV generation in Tunisia is shown in Fig. 7.

Fig. 7 Typical power curve of PV generation in Tunisia

Considering utilization hours are 5500h, a typical daily load curve with sample points per hour is chosen as research scenario. The minimum energy and power capacity of ES should meet night load demand. Then the output power of PV will meet daytime load. Meanwhile,ES will charge if the solar irradiation is enough. According to this production simulation method, the power generators configuration scheme based on analysis and calculation is PV with 6.1GW and ES with 2.27GW. The discharge electric energy from ES is about 14.39GWh every day. The simulation curve of PV and ES generation system is shown in Fig. 8.

Fig. 8 Simulation curve of PV& ES generation system

Referring to the cost prediction of PV and ES combined generation system after 2020[15], the predicted cost of PV plant is about 700$/kW, and the LCOE is about 0.072$/kWh, while the cost of ES(Li-battery ) is 273.5 $/kW,and the LCOE is 0.125$/kWh based on electric energy production. Hence, the LCOE of PV and ES combined generation is about 0.09$/kWh, equal to about 0.63yuan/kWh.

4.3 PV and CSP combined generation scheme

This scheme replaces a part capacity of CSP to PV generation. With reducing costs, it still ensures a relatively steady 24 hours of electricity.

Under the same conditions as in section 4.2, the minimum power of CSP must meet the maximum load demand in night, and the thermal storage capacity(FLH)should support all-night load. In daytime, PV as the main power source will meet daily load and the renewable power curtailment is limited to less than 5%.

The configuration scheme is the capacity of PV is 1.48GW and CSP is 1.9GW with 10FLH of thermal energy storage. The output power simulated curve of combined power generation station is shown in Fig. 9.

Fig. 9 Simulation curve of PV&CSP generation system

According to the daily power simulated curve, the annual power energy production of PV and CSP can be calculated are 2,682GWh and 8,454GWh respectively.

The LCOE can be calculated according to equation(2)[19]：

Where It is construction cost of plant, Mt is the operation and maintenance cost in the tth year, Et is power energy production in the tth year, r is discount rate, n is plant operation period.

For CSP power generation system, the investment cost of molten salt thermal energy storage with 10FLH for tower CSP station is about 40thousand yuan per kilowatt [20].The utilization hours of CSP are about 4400h according to power energy production. When the plant auxiliaries load rate is 8%, the net power energy is about 7,691GWh. Based on these conditions, the LCOE of CSP generation is about 0.13$/kWh (equal to about 0.89yuan/kWh). According to the cost reduction trend in the future, the LCOE can fall by about 20%, and it will reach to about 0.104$/kWh (equal to about 0.71yuan/kWh).

Considering PV and CSP annual power energy production and predicted LCOE around 2020 in this scenario, the LCOE of PV and CSP combined power generation will reach to about 0.097$/kWh (equal to about 0.66yuan/kWh).

If the utilization hours of DC transmission project are decreased to 4500h, the predicted LCOE of PV with ES and PV-CSP combined generation are about 0.082$/kWh and 0.107$/kWh respectively according to this analysis method.

4.4 Comprehensive economic comparison

Compared economy of the two solar energy development schemes, the LCOE of PV and CSP combined generation is higher under the condition of 4500h utilization hours of DC transmission project, whereas it almost equal to the LCOE of PV and ES power generation when the utilization hours of DC transmission project are 5500h. As the rapid development of CSP technology, the LCOE of PV and CSP scheme will be more advantage.

Recently, the LCOE of PV generation is about 0.074-0.088$/kWh in Europe, it also has competitive strength to PV-CSP combined generation. The development route of transmitting solar energy power from North Africa to Europe through transcontinental transmission channel will be constrained by different electricity prices between Africa and Europe. But the regulating function of PV-CSP combined generation is better than PV generation station,and it can participate in the power system auxiliary service.The combined generation scheme can benefit from this to make up for the high cost disadvantage in the future electricity market.

5 Conclusion

The North African region is rich in solar energy and is close to European continent. It is in line with the concept of the global energy interconnection to transport clean energy and electricity through transcontinental power connection.The development mode and scale of each country in North Africa are determined by economy of solar development.The LCOE development trend of PV and CSP are analyzed and predicted in this paper. Then solar development in Tunisia is as an example under background of transcontinental power connection, and two schemes of PV with ES and PV-CSP combined generation are analyzed respectively. We can conclude that：

(1) Scale effect is the motivation of solar thermal power generation equipment cost, construction cost and LCOE reduction.

(2) According to the prediction from IEA and other consultative agencies, learning curve method is adopted to forecast LCOE and construction cost of PV and CSP power generation. In 2030, the LCOE of PV and CSP will fall to 0.042$/kWh (equal to 0.286yuan/kWh) and 0.078-0.091$/kWh(equal to 0.53-0.619yuan/kWh) respectively based on learning curve method.

(3) Two solar energy development schemes in Tunisia are given under Africa-Europe interconnection background.The LCOE of PV and CSP combined generation is almost equal to LCOE of PV with ES when the utilization hours of DC transmission project are 5500h. The PV and CSP combined generation scheme can benefit from its flexible regulation performance in the future electricity market.

Acknowledgements

This work was supported by National Key Research and Development Plan (2016YFB0900100); State Grid Corporation Science and Technology Program(SGQHJY00GHJS1700078);Youth Fund of China Electrical Power Research Institute (NY84-17-003).