Energetic Complementarity Solar PV and Wind Power Based on Satellite Data

With the current global energy crisis renewable energy integration is expected to grow drastically in the near future. However, integrating high amount of renewables to the grid is subjected to various operational challenges. One solution for this challenge is looking for complementing intermittent renewable resources. In this paper solar PV and wind power complementarity analysis was carried out over the three topographic regions of Eritrea based on monthly satellite-based power generation data. Three different approaches (Pearson correlation coefficient, graphical and dimensionless index) were employed to investigate the complementarity of PV and wind power in the chosen sites. The analysis results showed that the southern coastal and central highlands sites have low complementarity behaviour. The solar PV and wind power synergy was however, found to be the maximum in the northern highlands and western lowlands sites of the country. It is interesting to find out that the high wind potential sites in southern coastal region are positively correlated to solar availability with peaks in winter and low in summer with no complementarity pattern.

Introduction

In addition to its environmental and socioeconomic benefits the demand for renewable energy is more critical than ever as a result of recent developments in the energy crisis. However, realization of high penetration of renewables is subjected to various challenges. The most pressing challenges are related to operational uncertainty due to renewable variability and non-dispatch-ability, inconsistency between load and generation, and others [1]. Several potential solutions were proposed by researchers to increase the variable renewable energy integration into the conventional grid, some of which are PV/wind complementarity, demand response, energy storage, and power forecasting [2].

Among the potential solutions resource complementarity plays a great role in smoothing the generation variability of renewables and improving grid matching capability [3]. Complementarity between Intermittent Renewable Energy Resources (IRER) can be defined as the ability of one resource to complement or compensate the other’s under-performance so as to stabilize the mean of the combined resource’s availability. Complementarity can occur in time, space or both between one or more IRER [4]. When two or more IRER complement each other in different regions at the same time it is called spatial complementarity while if the resources complement each other as a function of time in the same region it is called temporal complementarity. Spatio-temporal complementarity is the condition when one or more IRER complement each other in both time and space domains simultaneously [2]. The most commonly used approaches for evaluating the complementarity of IRER are Pearson correlation coefficients, graphical analysis, complementarity indices, and correlation maps. The Pearson correlation coefficient is an index that gives the indication if the resources can complement each other or not. If the resources have a negative correlation, then the resources can complement each other and are more reliable in combination than independent or in stand-alone mode [5]. Here it is important to mention that the selection of time resolution and data type are determinants on the complementarity analysis. Most studies use ground-based measurement while many others also use satellite and reanalysis data sets for analyzing complementarity of IRER [5].

In this paper, the complementarity between solar PV and wind power was investigated for the case of Eritrea based on the principle of dimensionless index [6] and Pearson correlation coefficient (PCC) using data from open-source data bases PVGIS and Global Wind Atlas. The power generation data were also presented graphically to better understand their complementarity behavior visually.

Materials and Methods

Solar and Wind Data Collection Methodology

The data used in this research was obtained from PVGIS a free web-based tool that provides PV generation and other weather parameters assimilated from satellite [7]. PVGIS is a data base which enable clients to get data in most parts of the world without any restriction or registry. The early version of for PVGIS is based on ground measurements of solar radiation collected throughout Europe but later versions shifted to satellite-based estimations. Using PVGIS users can download. csv files that contains hourly or monthly electricity generation for tracking or fixed grid connected PV systems for different PV technologies. Currently PVGIS can evaluate the performance of different grid connected PV systems by considering the effect of temperature and shadowing [8].

The downloadable Comma Separated Values (CSV) files contains PV power, irradiation, temperature, wind speed, and others [9]. However, the wind speed obtained from PVGIS is found to be unreliable in magnitude. For that reason, wind speed from PVGIS is scaled by the average wind speed from Global Wind Atlas [10]. Then using the log law equation, the scaled wind speed is extrapolated to the desired hub height. Standard Vestas wind turbine (V150) model power curve is used to calculate the power generation at all selected sites. The effect of air density on power production is incorporated in the analysis.

Site Selection

Eritrea is a country located in northeast Africa with a 1000-kilometer Red Sea coast. Eritrea’s landscape is divided into three topographic regions: highlands, western lowlands, and eastern lowlands or coastal areas [11]. The studied sites were carefully selected to include all topographical regions in order to shed light on the possibility of PV and wind power complementarity across the country. Located in the tropical region Eritrea has a year-round good solar intensity. The rainy and cloudy summer season somehow affects the solar intensity in the highlands. However, wind potential significantly varies throughout the country [11]. The most favourable sites are located in southern coastal areas and central highlands [12]. Table I contains geographical information on the sites. It should be noted that the geographic information given in the table does not represent the cities or towns themselves, but rather sites with favourable solar and wind potential in the vicinity of cities and towns.

Table I. Selected Sites for Analysis

Site	Geographical location	Altitude (m)	Region
Assab	13N, 42.7E	10	Southern coastal
Dbarwa	15.1N, 38.7E	2340	Central highland
Nakfa	17.1N, 38.27E	1724	Northern highland
Teseney	15.36N, 36.68E	1412	Western lowland

Complementarity Index

In this paper we apply an easily and simplified concept of complementarity analysis based on Beluco’s method presented in [13]. The main challenge in complementarity analyses is the availability of time series data on the chosen locations. Here in this paper, we used a simplified method and data collection methodology which can be replicated at any location without any data associated problems since PV power and wind data were obtained from free web-based data bases. There are different approaches for evaluating the energetic complementarity of IRER, but complementarity analysis based on monthly data can give quick and reasonable implication for a more precise analysis. This approach provides a fundamental understanding of the basic principle of complementarity.

The first step in evaluating energetic complementarity is selecting and screening the time series data.

Based on [13], time complementarity kt is determined as: (1)kt=|m1−m2|6 (2)kt=|m2−m1+12|6where m1 and m2 refers to the months in which minimum availability of solar PV and wind power occurs. Here the smallest value of kt is taken from (1) and (2). If kt is 0 implies that the minimum value for both resources occur at the same month and there is no complementarity but if the result is 1 it implies that there is maximum time complementarity between the solar PV and wind power. If the result in between there is some kind of linear relation between the two resources.

The mean value of solar PV and wind power is calculated as a1 and a2, then the energy complementarity ke is computed as [13]: (3)ke=1−|a1−a2a1+a2|

ke will be 1 if the mean of solar PV and wind power is equal and 0 when one is much greater than the other.

The next step is to quantify the amplitude of variation of both PV and wind power around the mean. Here we denote v1 for PV and v2 for wind power variability. The value of v1 and v2 is the difference between the maximum and minimum values of PV and wind power, respectively.

The Energy complementarity is then computed from the following relations [13]: (4)α1=1−v1a1 (5)α2=1−v2a2 (6)ka={1−(α1−α21−α2)2ifα1≤α2(1−α2)2(1−α2)2+(α1−α2)2ifα1≥α2

The final complementarity index is therefore computed by multiplying kt,ke,ka: (7)k=ktkeka

Pearson correlation coefficient rsw is computed as [14]:

(8)rsw=∑i=1n(si−s¯)(wi−w¯)∑i=1n(si−s¯)2.∑i=1n(wi−w¯)2where si and wi are time series measurements, s¯ and w¯ are the average value of PV and wind power respectively and n is the sample size.

Result and Discussions

Here, we present the analysis results for each of the four sites listed in Table I. Because it is not convenient to compare two energy resources generated using different conversion technologies, we normalize the time series using their respective maximum capacities, with the result falling between 0 and 1. The normalized time series is therefore the basis for all results in the discussion that follows.

Graphical Interpretation of Complementarity

Fig. 1 shows how the resources can complement each other. It is intriguing to learn that the best places, such Assab and Dbarwa in the southern coastal areas and central highlands, do not complement one another.

Fig. 1. Solar PV and wind power monthly generation: (a) Assab, (b) Dbarwa.

Sites in northern highlands and western lowlands, however, show a promising complementarity behavior in the monthly averages. Fig. 2 shows the monthly resource availability in Nakfa and Teseney. As seen from the figure the two sites can complement each other to some extent and further precise complementarity analysis is recommended.

Fig. 2. Northern highland and western lowland sites: (a) Nakfa, (b) Teseney.

Correlation Coefficient

Correlation between IRER gives a good indication if resource availability can complement each other or not. Sites with negative correlation can complement each other, the deficit of one resource can be compensated by the availability of the other. Table II gives the correlation coefficient of solar PV and wind for all four sites.

Table II. PCC between Solar PV and Wind Power

Site	Correlation
Assab	0.2389
Dbarwa	0.4890
Nakfa	−0.4521
Teseney	−0.1668

Complementarity Index

Using a dimensionless index, this section evaluates the energetic complementarity of solar PV and wind power in time for the three topographic regions of Eritrea. The partial, time, energetic, and amplitude complementarity of all sites is given in Table III.

Table III. Dimensionless Index of Solar PV and Wind Power

Site	kt	ke	ka	k
Assab	0.33	0.96	0.85	0.27
Dbarwa	0.17	0.97	0.95	0.16
Nakfa	0.33	0.87	0.73	0.21
Teseney	0.50	0.90	0.73	0.33

As shown in Fig. 3 the monthly power generated by both resources behaves the same and the possibility of compensating each other is less. This is also validated by the low value of dimensionless index tabulated in Table III.

Fig. 3. Monthly average of power generation in southern coastal and central highland sites: (a) Assab, (b) Dbarwa.

The northern highlands and western lowlands sites, however, have good complementarity characteristics. As shown in Fig. 4, Nakfa and Teseney showed good complementarity behavior both analytically and graphically, but dimensionless index-based approach returns unexpected small value of k to say less possibility for complementarity.

Fig. 4. Monthly average power generation in Northern highland and western lowland sites: (a) Nakfa, (b) Teseney.

Conclusion

Energetic complementarity between solar PV and wind power for four sites representing the three topographic regions of Eritrea was presented. The analysis was done using three different approaches. The results showed that the northern highland and western lowland sites have good complementarity characteristics, while the southern coastal and central highlands show no possible complementarity behaviour. Although wind and solar were not in synergy in the most favourable wind sites, Eritrea can still benefit from solar PV and wind hybrid plants to some extent.