Feasibility of Substitution of the Conventional Street Lighting Installation by the Photovoltaic Case Study on a Municipality in Agadir in Morocco

Received Mar 23, 2017 Revised May 27, 2017 Accepted June 27, 2017 In this work, we present a technical and economic study of the solar powered street lighting system of a municipality in the south of Morocco. The state of the conventional street lighting system is first analyzed in a substation of street lighting. Then a sizing method is applied to the photovoltaic installation in the testing area. A financial study, by comparison between conventional and PV-based lighting, is carried out showing the feasibility of the PV street lighting.).


INTRODUCTION
The techno-economic study for solar street lighting system have been developed in many scientific research works [1][2][3]. These studies confirmed that the energy consumption in street lighting system is considerably reduced with PV power source. However, these works focused only on the technical feasibility [4][5][6][7][8][9][10][11], and it's much difficult to find studies on the sustainable and exhaustive feasibility in literature which require an analysis of electrical quality of power supply installation in accordance with EN50160 standard [12,13], and the quality of the electrical network consistent with the standards CE161000-2-2 [14] and CEI61000-3-4 [15].
In India, a study has been carried out in order to evaluate the energy and economic efficiency of the street lighting by the "Lithonia Visual Software" and "Homer Software" [16]. The first software is used in order to simulate the graphic structure of light points and the maximum luminous flux by adjusting the tilt angle and height of mast; the second one is used in for the conception and analysis of hybrid power generation systems.
Another study [2], fulfilled in Taiwan, has concerned an economic comparison between three types of systems: Solar street lighting with LED lamp, conventional street lighting with on the one hand LED lamp and mercury-vapor-lamp on the other hand. This study has shown that, despite a higher installation cost, the solar powered LED system is more economical due to an energy savings of up to 75% compared to the mercury lamp lighting system. In the case studied in this work, the street lighting of Dcheira El-Jihadia municipality, located in the suburb of Agadir city in Morocco, has consumed 3 364 300 kWh during the year 2015. This meant to the municipality a bill of around 4 178 750 MAD that represent 20% of its total budget. This was the main reason that led the municipality to move towards solutions to optimize the electricity consumption in street lighting, including solutions using PV energy.
Indeed, in the field of street lighting, the solar street light is a kind of street light which is supplied by the solar PV energy, that is, it's equipped with solar panels that collect sunlight throughout the day and transform it to electricity. Then, it insures the energy storage in batteries via a circuitry. Finally, to provide lighting, electrical energy is restored at night. This kind of street light thus becomes self-sufficient in energy. it will greatly contribute to reducing consumption costs as well as greenhouse gas emissions.
In this context, together with the great sunshine in the region of Agadir estimated at 3 144 kWh/m 2 [17], we have proposed to this municipality to study the feasibility of substitution of the conventional street lighting installation by the photovoltaic in order to minimize electrical energy consumption.
The present paper is organized as follows: after this introduction, we present in section 2 the tools of investigation used. Then, in section 3, we analyze different measurements carried out in one of the conventional street lighting substations in an area of Dcheira El Jihadia municipality. Then, we describe in section 4 a sizing method applied to the PV installation of street lighting in the testing area. Section 5 concludes the sizing with a financial study of the installation by comparison between the use of conventional and PV-based lighting. Then, the technical and economic study results will be presented, compared and discussed in section 6. Finally, a conclusion and references close this paper.

DEVELOPMENT TOOLS
A portable network analyzer [18] was used to study the state of a post of the street lighting. This device allows to make measures and to record certain electrical quantities as well as the parameters which identify the quality of an electrical network.
Otherwise, for sizing the proposed solar street lighting system, the tools used are: -PVSyst is a software dedicated to sizing photovoltaic systems [19].
-DiaLux software is used to calculate optimum brightness in public lighting installation [20].
-MapInfo which is a computer tool giving a representation of geographic data of light points on geographical/satellite maps [21].

MEASUREMENT ANALYSIS OF CONVENTIONAL STREET LIGHTING SYSTEM
Prior to study and size of the proposed solution, we present the state of one of the street lighting substation, named "Substation 9", in the Dcheira El-Jihadia municipality.
The measurements were carried out as follow: Firstly, some electrical quantities characterizing the substation, such as voltage, current, power factor and fundamental frequency. Secondly, the parameters identifying the electrical grid quality of the conventional substation, such as total harmonic distortion, harmonic spectrum of the voltage and the current, and their compatibility levels. Finally, the power consumption involved in this conventional substation.

Tracing and tracking of the installation
Using MapInfo software, we have tracked coordinates of the different light points supplied by the "Substation 9" that was chosen for this study. This Substation is located in the region whose GPS coordinates are 30°21'46.8'' North and 9°31'42.2" West (Latitude 30.363 and Longitude -9.5284). Figure 1 presents the geographical distribution of 70 light points of various powers that "Substation 9" serves. These light points are listed in Table 1.   table 1 As an illustration, to evaluate the values of electrical quantities, as well as the electrical performance quality of "Substation 9", we make use of quantities adopted by EN50160 standard [12] summarized in Table 2.  Figure 2 shows the evolution of the voltage U of the three phases during one day; between 13h of 08 July 2015 and 15h of 09 July 2015. Following analysis of the above parameter, we find that the voltage U of the three phases is lower than ±10% of nominal voltage 230 V as stipulated by the standard EN50160 [12]. Indeed, according to this figure, the maximum value is approximately 250 V (< 253 V) and the minimum value is about 233 V (> 207 V) leading to a non-overloaded electricity network. We also note, in figure 2, a voltage imbalance between phases. But this imbalance does not exceed the EN50160 standard [12] limit of 2%. For example, on 09 July 2015 about 3h30, U max = U 1 -U 3 = 245.5 -244 = 1.5 V < 4.6 V. Figure 3 shows the temporal variation of the current I in the three phases of the "Substation 9" power supply. We observe that the three curves corresponding to phase currents present current peaks during starting of the street lighting lamps, at 19h45 of 08 July 2015 as shown in Figure 3. At around 11h00 of the following day, the lighting was forced to determine the light points supplied by the substation 9. These current peaks are approximately 1.4 times the nominal current I n (I n1 = 28 A, I n2 = 48 A, I n3 = 72 A). the various pulses occurring on the curves of phases 1 and 2 in the range of 10 A at 14h40 of 08 July 2015 shown in figure 3, are due to another additional load supplied by the phases 1 and 2. It corresponds to temporary supplying of maintenance engines by the "Substation 9". We also note that load current is stable throughout the lighting period, especially for phase 3 for which I 3  72 A. For the other two phases, the mean value of the currents I 1 and I 2 are respectively about 28 A and 48 A. the imbalance between I 1 , I 2 and I 3 is due to the light points inequality supplied by these lines. Indeed, the phases 1 and 2 supply 12 and 26 streetlights respectively, while the phase 3 alone supplies 32 streetlights. In order to correct this imbalance, the municipality must balance the three phases by charging them almost identically.

Power factor 
The statements issued by network analyzer include the evolution of power factor , as shown in Figure 4 for the three phases of the "Substation 9". We find that, for the two phases 1 and 3,  1 and  3 evolve almost in the same direction with an average value of about 0,7. On the contrary,  2 is unstable due to defect in the phase line 2, and possibly due to flickering street light, and also to temporary supplying of maintenance engines. Otherwise, the average total power factor measured in the installation of street lighting is approximately 0,7, which is not really optimal since most lamps installed don't have a power correction capacitor. The municipality must install capacitors in order to reach a power factor close to 0,95.  

Network frequency
Among the electrical quantities in the "Substation 9", we illustrate in figure 5 the fundamental frequency F measured and that is less than ±1% of the nominal value 50 Hz imposed by the standard EN50160 [12]. Indeed, the measured values, F max and F min are equal to 50.06 Hz < 50.5 Hz and 49.95 Hz > 49.5 Hz respectively, which represents the acceptable variations. Figure 5. Fundamental frequency F of the supply voltage on the "Substation 9"

Parameters identifying the electricity network quality 3.3.1. Total harmonic distortion factor
In order to assess the electricity network quality of "Substation 9", we use one of the dedicated parameters, named total harmonic distortion THD f . This is defined as the ratio between the total effective value of the harmonics of signal U or I, and the effective value of its fundamental component. This is a parameter defining the overall distortion of the periodic quantity.
The THD f reflects substantially increase of Joule effect in the lines and devices. We distinguish two total harmonic distortion:  THD V corresponding to the total harmonic distorsion in voltage.  THD I representing the total harmonic distorsion in current. According to measurement results shown in figure 6-a, we find that the total harmonic distortion in voltage VTHD is less than 5%, which is acceptable under the EN50160 [12]. Thus, this electrical grid is environmentally non-polluting in voltage, which is not the case for current. Indeed, the spectrum on figure  6-b shows that the total harmonic distortion in current ITHD is about 13,8% from 20h to 06h15 of 09 July 2015, exceeding 10% required by the standard. To minimize the ITHD, we must set up a passive harmonic filter [22][23][24], which is envisaged in the near future by the municipality. There is also a high ITHD between 19h30 and 20h which is not due to the street lighting: Since the lighting begins only at 20h, then the ITHD is not caused by the ignition of the lamps, but by an impact of other loads supplied by the phase 2; namely the maintenance works near the "Substation 9".

Harmonic spectrum of voltage and current
The lamps used in street lighting are a non-linear load corresponding to a deformed current draw, and then to an harmonic rich current.
From the measurements taken on three phases L1, L2 and L3 of "Substation 9", as shown in figure  7, the current harmonics H In are more predominant than the voltage harmonics H Vn . This is due to the impact of loads whose current exceeds 16 A. The spectral lines H In as will as H Vn correspond to the odd harmonics ranks of order 3, 5, 7, 9, 11, etc. In fact, the harmonics are characterized by their odd/even rank. The even harmonics are often negligible in industry, because they are cancelled by the signal symmetry. In addition, these even harmonics only exist in the presence of a DC component, contrary to the odd harmonics (3,5,7) which are ampler on the public electricity grid.  Table 3 and Table 4.  Figure 7, the current harmonics H I3 as well as the voltage harmonics H V5 [13] are ampler than the others. Indeed, H I3 and multiples of 3 are generated by single-phase loads since, in this case, the HPS lamps (High Pressure Sodium) are powered by one phase line. These H In are particularly damaging to threephase power at the station, because of the high current produced in the neutral conductor leading to production of an excessive heating of neutral wire.

Energy consumption
The power consumption E c (en kWh) is the basis data for improvement work of a street lighting system. It's related to the lighting time and equipments state.

PV SYSTEM DESIGN
Hereinafter, we propose a sizing method of a stand-alone PV system, which will provide electricity to various lines supplied by the "Substation 9".
The method consists of dimensioning, in PVsyst 5.0 software [19], the PV panels to be installed in this solar street lighting system. This software requires some characteristics of the study area, such as; geographical data, photometric data and the energy requirement. Subsequently, the software generates a type of PV module and of suitable battery.
The approach to be taken for the sizing is as follows [25]:  Tracking and tracing of the installation;  Sizing of the lamps;  Definition of the energy requirement;  Determination of the peak power of the photovoltaic panels;  Calculation of the battery capacity;  Choice of charge controller. Sizing a PV system requires knowledge of daily average irradiance I rd at the place of installation. Table 5 presents the meteorological data in the testing area. These data are obtained using simulation tool PVgis-CMSAF [26], incorporating irradiance data I rd , during the period from 1998 to 2011.

Sizing light fixtures
In the case of the conventional street lighting, the loads to be supplied are mainly lamps in powers of 150, 250 and 400 W. For our solution, both for environmental and economic reasons, we opt for LEDs lamps (Light Emitting Diode) replacing conventional lamps. These LEDs lamps were chosen by means of the DiaLux 4.12 software [20]. Figures 9 and 10 show a comparative example, given by the manufacturers [27], between the photometric results of a conventional HPS lamp in power of 150 W and that of equivalent LEDs lamp [28].
With respect to photometric properties given in Figures 9 and 10, the manufacturers define on the one hand, two "C-plane"; C0-C180 named the transverse plane perpendicular to the light fixture axis and the longitudinal plane C90-C270 passing through the axis of light fixture. On the other hand, they define the parameter  which represents the viewing angle wherein we can observe the light fixture. Figures 9 and 10 illustrate the luminous intensity distribution in Cd/kLm, emitted in a given direction within the two planes C0-C180 and C90-C270. We note that the light fixture efficiency  of the HPS lamp in power of 150 W up to 82%, is closer (84%) to that of the light fixture of equivalent LEDs lamp. This efficiency is defined as the ratio of the luminous flux emitted by the light fixture and that of the lamp.
Following our comparative study of different types of HPS lamps used in conventional street lighting, table 6 summarizes the simulation results obtained in the DiaLux software.

Calculation of the electrical energy requirement
This energy is required for the design of photovoltaic generator (PVG) and the storage capacity of the battery with an adequate series/parallel mode selected for their interconnections. Considering the light point that consumes 33 W, for an average of 13 h/day, the electrical energy requirement is approximately E c = 430 Wh/d.

Sizing of the PV Modules
With the data in the previous paragraph, it is possible to know the number of PV modules needed to cover electricity demand. This involves firstly calculating energy which the PV modules have to produce every day [25]- [33]:

Energy produced by a PV panel
According to the influence of several parameters, for an expected energy requirement E c to supply a single light point, the energy E p to produce by a single PV module is given by the following relation.
Where k represents a coefficient dependent on the meteorological uncertainty, the uncorrected tilt of the PV modules according to the season, the operating point of the modules that is rarely optimal, the performance of charge-discharge of the batteries (80%), the cable losses, etc. For the stand-alone systems, that is with storage unit such as batteries, this coefficient k is generally between 0.55 and 0.75 [34]. Considering in our case, an average value k = 0.65, this allows us to estimate the value of E p at 660 Wh/d.

Calculation of the peak power of the PVG to install
The peak power P c of a single PV module is calculated by the following relation: Where I r is the average annual daily irradiation and G = 1 kW/m 2 is the radiation in standard measurement conditions (STC: Standard Test Conditions). In our study, according to data of the table 5, I r is about 5.86 kWh/m 2 /d, which gives us a peak power P c of approximately 113 Wp (Peak Watt).

Determination of number of solar panel
Our choice went to polycristalline PV modules type SPP81-12, whose technical specifications are given in Table 7 [28].
As the radiation G does not reach 1 kW/m 2 and that these PV panels have a peak power P MPP of 80 Wp, a single module will be sufficient to supply a light point of power 33 W according to equation (3).

Calculation of the capacity of battery
For PV based systems, storage batteries are needed. In order to design them, we proceed as follows:  Determination of the number of days battery life required N j  Evaluation of the depth of discharge D (en %) acceptable for the type of chosen battery  Calculation of battery capacity C (Ah) to implement whose voltage is U (V), = For daily energy consumption E c of about 430 Wh with autonomy N j estimated 2 days, for a voltage U = 12 V, there should be a battery capacity C of about 90 Ah admitting a depth of discharge D in the order of 80%. In our case, we opted for the battery of the type GEL/12V of a capacity of 90 Ah [25,35].

Design of the charge controller
Any stand-alone photovoltaic system requires a charge/discharge regulation of the batteries. It's the role of the solar controller which ensures the protection of batteries against surcharge as well as deep discharge. It also allows to be protected against a reverse current while ensuring the temperature control; this is why the charge controller should always be installed next to the batteries. This charge controller will be sized according to the following parameters:  The nominal voltage U: is the battery voltage: 12 V.  The maximum current of the charge controller I e : is the maximum operating current of the generator. It must be backed easily by the charge controller. To estimate the current, we take a coefficient of 1,5 to ensure a safety margin. As a result, for the maximum current I max = 4.6 of the PV panel, I e is nearly 7 A. As standard value, we choose I e = 10 A [36].  The output current I s = 10 A [36].
Considering the above, we opted for the type of charge controller STECA PRS 1010 Solarix 10A-12V whose technical specifications are summarized in table 8 [36].

ECONOMIC EVALUATION OF THE PROJECT
In order to carry out and to justify the chosen solution in our technical study of the street lighting system of Dcheira-El Jihadia municipality, an economic study is required. This is covered in some detail in this section.
In order to get a sense for the amount of the investment for the PV system, we start by calculating the original investment. This calculation is based on the price of each component of the installation while integrating the installation cost of the lights. Then, we determine the annual cost of energy consumption using basic data. In fact, this data is the product of operating hours, the power consumption and the unit cost. Finally, we estimate the annual operating cost after calculating the annual cost of maintenance.

Original Investment
It is defined as the sum of the acquisition cost and the labour cost for the installation. Indeed, the PV system materials is about 905 830 MAD, while the labour is approximately 5 250 MAD. Whereas the cost of equipment of the conventional installation did not exceed 391 700 MAD, with a labour cost of the installation of approximately 15 750 MAD.

Annual cost of the energy consumption
The cost of annual energy consumption is the product of total power of the lamps by the lighting hours per year (4530h) and the price of the Kilowatt-hour (1,11071 MAD). This average annual cost for the conventional installation is about 69 200 MAD. In the case of PV system where consumption is free, the equivalent of annual cost is about 18 000 MAD.

Annual operating cost
It includes the annual cost of energy consumption C a and the annual cost of maintenance C m which corresponds to the replacement cost. The conventional system has generated a significant cost of about 73 280 MAD, which C m is about 4 080 MAD. On the contrary, the operating cost for the PV system is only 14 250 MAD corresponding only to the annual cost of maintenance.

Total operating cost for 25 years
The operating cost is calculated for a given period. This period is considered as the maximum lifetime of PV panels, which is equal to 25 years [33][34][35][36][37]. For that purpose, we proceed first to update the prices according to the following equation [38]: The present value (here = 1.11071 MAD). n: Number of years to the future value.

:
Discount rate (8%). The total operating discounted cost will be 570 400 MAD for the conventional installation for a period n of 25 years and a discount rate  of 8%. While for the PV system, it does not exceed 167 400 MAD. Table 9 summarizes the results of different cost calculation.

RESULTS AND DISCUSSION
The technical and financial studies of street lighting system presented above have allowed us to compare the quantities of energy operated in the PV and conventional installations, which will lead us to substantial gain by substituting PV cells supply for conventional electricity supply. The results thus obtained are summarized in table 10. The energy consumed during 2015 for the conventional installation is in fact 54 200 kWh, whereas for PV system it does not exceed 16 300 kWh.  According to table 10, we note that the solar street lighting system generates, during a 25 years period, a large gain of about 403 000 MAD corresponding to 70% of the budget and of energy savings compared to conventional system.

CONCLUSION AND PERSPECTIVES
The investigations presented in this paper allowed us to study, analyze, design and compare the annual energy consumption, then the cost of two types of street lighting installation implemented or envisaged in a municipality of Agadir in Morocco. The two types are the conventional street lighting and that based on solar energy, especially the photovoltaic. According to the results, we can confirm the feasibility of substituting conventional street lighting technology for renewable energy sources such as PV. In fact, it's far more cost-effective both on the energy and economic levels. Indeed, following our study, the solar street lighting solution provides for the municipality, a large gain of about 70% of the public lighting budget, and of energy savings, compared to the conventional street lighting. In perspective, a monitoring application of energy efficiency of the solar street lighting installation will be developed.