Protection of photovoltaic systems against lightning

Transcription

1 Protection of photovoltaic systems against lightning Experimental verifications and techno-economic analysis of protection H. E. Sueta, A. Mocelin, R. Zilles and P. F. Obase Institute of Energy and Environment / University of São Paulo-Brazil Abstract This paper presents some experimental results of a study on photovoltaic modules submitted to several impulses and direct current applications simulating components of lightning. It also presents two types of lightning protection for photovoltaic generation system: a 1 MWp and a minigrid of 15.6 kwp installed in a remote site (Amazonia), away from the major centers. Finally, the paper presents a technical and economic analysis for implementation of lightning protection systems in photovoltaic installations. Keywords photovoltaic systems; lightning protection; earthing I. INTRODUCTION This paper presents the results of simulations of direct lightning strokes applied in photovoltaic modules. It can see in the various tests that even being hit several times by the test currents (impulsive and continuing current), modules can withstand to the damage and often still generate energy with little loss of power. In the tests, the modules have been subjected to very severe conditions that do not happen in reality. We evaluate that for each type of project, a cost-benefit study should be done to provide or not an external Lightning Protection System (LPS) for the photovoltaic power plant. For large power stations that occupy large areas and with many modules, the cost of deploying an external lightning protection generally is not competitive with the cost of replacing some modules affected by lightning. For minigrids, the deployment of an external lightning protection can be a good option once its cost can be compared to the cost of the labor to a possible exchange of the module and the shipping costs for this module, since many of these systems are in locations distant from urban centers. For large power stations, a good option of LPS is the analysis of the use of the metallic structures as a natural component of the LPS. These metallic structures are used for fixation and positioning of photovoltaic modules and are normally buried in the soil, creating an electrical path between the frames of the modules and the ground. This paper presents a study in a photovoltaic power station of 1MWp, where the metallic structure was considered as a natural component of the LPS. This study was composed by measurements of the soil resistivity, stratification of the soil and calculation of the grounding subsystem; this last was calculated using the earthed structures connected between each other in a defined set of connections. Depending on the amount of metallic structure used, the depth that they will be buried and of the soil conditions, the set of structures can be considered as the LPS of the power station. E. Boemeisel Empresa Brasileira de Energia Solar São Paulo, Brazil For minigrids, depending on the distance between the place of installation and the urban center (where additional modules and skilled labor are found), the installation of an external LPS may be the best option. We present in this paper an example of this type of installation. This work also presents a technical and economic analysis of some options of LPS for two types of photovoltaic installations: power station of the order of 1 MWp and minigrids installed in remote locations far from the major centers. II. EXPERIMENTAL STUDY IN PV MODULES This paper presents an initial study to estimate the behavior of photovoltaic modules in the face of direct lightning strokes. Initially the modules were inspected and measured using the test procedure from IEC [1] (Crystalline silicon terrestrial photovoltaic (PV) modules - Design qualification and type approval), which are: Visual Inspection, Maximum Power determination and Insulation Test (procedures 10.1, 10.2 and 10.3 respectively). In total were tested three modules: modules 1 and 2 with dimensions 45 cm x 96 cm, 36 polycrystalline cells, 50 Wp, with metallic material on the bottom side (the module 2 was only used for the calibration of test parameters, see Figs. 1a and 1b). The module 3 with dimensions 45 x 88.2 cm, 45 Wp, 36 polycrystalline cells with insulating material on the bottom side. Fig. 1a and 1b: Front and back of the modules 1 and 2 For the Maximum Power determination (SOLAR tests), the modules were tested using the Flash Solar Simulator of the IEE-USP. This Flash Solar Simulator can simulate the Standards Tests Conditions (STC: 1000 W/ m², 25 C and A.M 1.5G spectrum) while the electronic system measures the current and voltage to trace the characteristic curve of the module (I x V curve). For the Insulation tests (IT tests), the modules were tested using a MEGOHMMETER. After the IT tests, the modules were subjected to many current applications: impulse currents with waveform 7.4/15µs and amplitudes of 13 ka (AT tests), see Fig. 2a. Between each current application,

2 the maximum power determination and the insulation tests were repeated to check for any changes in the module. In the first AT test sequence, currents are directed to strike the metallic framework of the photovoltaic module, see Fig. 2b, as an electric arc. Fig. 2a: Waveform of one impulse current application and Fig. 2b: AT test scheme in the H.V. Lab. In the second AT test sequence, the cables of the photovoltaic module (length of 2 meters) were connected to a Surge Protection Devices (SPD) suitable for photovoltaic modules and in a high voltage divider. In these applications, the impulses are still applied in the frame of the modules and the impulse voltages were measured at the end of the cables with the SPD connected in two ways: close to the junction box of the module (Fig. 3a) and close to the voltage divider (Fig. 3b). insulation (test with voltage applied equal to 1000 V plus twice the maximum system voltage, in this case, 2259V), however within the limits of the standard. After the application of the impulse current in the center of the module, the rupture of the insulation occurred with 1366V. In the test with the voltage applied equal to the maximum system voltage of the module (625V) during 2 minutes, the module supported all tests, even after the test with impulses in the center. b) In the module 3, after 3 applications of current impulses on the frame of the module, we verify the rupture of the resistance of isolation during the test with voltage applied equal to 1000 V plus twice the maximum system voltage (2668 V). The rupture of insulation occurred with about one minute of application of test voltage. After of the application of the impulse in the center of the module, this test was repeated and the module supported the essay. In the tests with the voltage applied equal to the maximum system voltage of the module (829 V) for 2 minutes, the module supported all essays. The modules were subjected to pulses of direct current (DC) as an electric arc. The objective of the DC applications was to verify any physical damage on the photovoltaic modules, since these pulses cause more thermal effects. In each module were applied 6 DC pulses at different positions in the frame. The gap between the aluminum electrode connected to the source and the frame of the module was approximately 3 cm. To initiate the arc was used a copper wire that vaporizes when a current flow keeping the arc at the gap. For each module, were applied 2 pulses of approximately 70 C (550 A/0,130 s), 2 pulses of approximately 150 C (550 A/0,275 s), see Fig. 4a, and 2 pulses of approximately 300 C (550 A/0,55s), see Fig. 4b. After all applications of continuing current, the measurements of Maximum Power and Insulation were repeated. Fig. 3a: SPD near module Fig. 3b: SPD near divider In another AT test sequence, currents are directed to strike the center of the photovoltaic module, as an electric arc. At the end of all current applications, the maximum power determination and the insulation tests were repeated and the results compared with others previous tests. We note that the current impulses applied in the frame of the two modules did not left significant marks and no visual deterioration was observed. When the impulse current was applied in the center of the module 1 (with metallic bottom side) its glass broke. Applying the current impulse in the center of the module 3 (with insulating bottom side) it was verified a mark in the module, but without breaking the glass. In relation to the tests with impulsive current, we can see that, in both modules, the parameters relating to Maximum Power determination remained within the ranges of uncertainty before and after all impulse applications. In relation to the Insulation Test, we verify that: a) In module 1, after the current impulse test in the frame of the module, there was a decrease of the resistance of Fig. 4a: Mod As Fig. 4b: Mod As Finally, the modules were subject to DC pulses applications, but this time in the center of them. For this, it was necessary to break the isolation of the module due the limitation of the value of the continuous source voltage; this was done through a metal pointer in the center of the module. In this case, we also used a copper wire that vaporizes when a current flow to initiate the arc at the gap. Figs. 5a and 5b shows the damages in the modules.

3 Fig. 5a: Mod.1 center Fig. 5b: Mod.3 center Evaluating the results after the direct current pulses that simulate the continuing current, we note that in applications in the frames, in spite of the damage in the metal, there was a very small loss of the power of the modules. The applications in the center of the modules caused extensive damage and loss of power, being that the module 3 did not respond to maximum power determination. In relation to the Insulation test, the module 3, that already had presented rupture of the isolation in previous test with voltage applied equal to 1000 V plus twice the maximum system voltage, again presented isolation rupture in this voltage. After all tests in the module 3, it endured the essays despite having lower values of insulation resistance. The module 1, despite having presented a rupture of the isolation earlier, after these tests, endured the Insulation test showing smaller values of insulation resistance. In general, the modules were submitted to various applications of current pulses in order to simulate lightning in different points of them. We note that, even being submitted to more severe conditions than those found in the nature, the modules can still generate power, despite losing electrical insulation and efficiency. III. STUDY OF LPS IN A LARGE POWER STATION The large photovoltaic plants are often built on plan land and occupy large open areas. The installation of an external lightning protection in that areas requires an even greater area because it is necessary the installation of interception rods making sure there is no shading on the PV modules at any time of the year, in order to avoid loss of efficiency in the generation of energy. Generally in the construction of these power stations, metallic structures are mounted directly buried in the soil for fixation and positioning of the photovoltaic modules (see Fig. 6). In this study, we present the use of this structure together with the interconnection between the parts no connected of these structures as natural component of the LPS. We began this study with the measurement of soil resistivity in order to calculate the soil stratification in the place where the power station will be installed. After that, we are able to calculate the resistance of the grounding system composed by steel bars buried in the soil and interconnected by the structure. Fig. 6: Structures of the power station Now, we presented the results of the measures of the soil resistivity (Wenner arrangement) realized in the power station and the study of the soil stratification. Based in the area of the location (15200 m²), it was verified the need of a minimum of 6 lines of measurement (see Fig. 7: A, B, C, D, E, F). PLA NTA SUN EDISON PLANTA EBES D E A F B C NORTE SALA ELÉTRICA Fig. 7: Lines of measurements

4 For the soil stratification, the collected data of soil resistivity were used as input in the CDEGS software, module RESAP. This software stratified the soil in two layers: ρ1 = 1147 Ω.m with a thickness of first layer equal to 3.2 meters and ρ2 = 379 Ω.m with the second layer until the infinity. The calculation of earth resistance of the mesh was realized by software CDEGS, module HIFREQ. Initially was considered the mesh of 80 m by 80 m, where the rods are connected to 30 cm above the ground by a copper cable of 50 mm 2. The steel rods with 12 mm of radius are buried at a depth of 1.5 m. The distance between the rods is 10 m. Fig. 8 show the mesh. Due the limitation of the computer program it was not possible simulate the actual setting of the mesh. Fig. 8: simulation It was considered that a current of 1000 A, 60 Hz was applied in this mesh, then, with these data, the module HIFREQ calculated the voltage drop on the surface of the earth. Finally, it was calculated the resistance of the mesh, dividing the voltage by current of 1000 A, obtaining: R 1 = / Ω. Additionally, another computer simulation was carried out with the same characteristics of previous simulation, but considering a mesh of 20.8 m by 20.8 m with rods buried every 2.6 m, then it was calculated the earth resistance of the mesh, obtaining: R 2 = / Ω. Comparing the values of R 1 and R 2, it is possible to note that the larger the dimension of the mesh lower is the value of the earth resistance. In the case of the simulation with value of R 1, we have the actual mesh size (80 m by 80 m), however the distance between the rods (10 m) is greater than the real situation. For the case of simulation with value of R 2 the distance between rods is the real (2.6 m), however the size of the mesh (20.8 m by 20.8 m) is smaller than the real situation. Thus, the simulations indicate that for a larger mesh and with smaller rods distance the earth resistance value will be lower. In the case of power station under study, there are approximately 1400 rods buried, that corresponds more than 17 times the amount of rods used in both simulations. This ratio gives us the security that the grounding resistance of the system, with 1400 rods buried on the soil analyzed, all interconnected (through the metallic structures of the support of the photovoltaic modules or through the equipotentialization cables), will have a value lower than 10 Ω, as recommended by Brazilian standard [2]. In this way, the rods, metallic structures and equipotentialization connections compose the grounding subsystem for electrical purposes and lightning protection. To verify if the conclusions above can be used for other types of soil with resistivity larger, we repeated the simulations for other values. In these new simulations, we used two different conditions for three resistivity larger values: the condition 1 considering a homogeneous soil and the condition 2 with a soil stratified in two layers, where we changed the resistivity of the first layer of 3.2 meters depth with the new values and the second layer with a fixed value of 379 Ω.m until the infinity. The new resistivity values used in the simulation were: 3000, 5000 e Ω.m. The Table I presents the results obtained for the two types of mesh for which simulations were made (mesh size 80 m by 80 m with distances between the rods 10 m and mesh size 20.8 by 20.8 m with distances between the rods 2.6 m). Table I: Resistance of the meshes obtained in simulations. Cond. 1: Homogeneous soil Cond. 2: Two layers soil Resistivity (Ω.m) Resistance of mesh (Ω) Resistivity of the Resistance of mesh (Ω) 80 x x20.6 first layer (Ω.m) 80 x x Analyzing the results of the simulations, we can see that with the increase of soil resistivity the values of the resistances of the meshes increase considerably, therefore, the use only of buried rods may not be sufficient to obtain a good grounding system. In this way, we can see that each case must be studied separately, mainly for soil with high resistivity. IV. EXAMPLE OF PROTECTION IN A MINIGRID In this part is presented an example of protection for a minigrid consisting of six photovoltaic arrays with 20 photovoltaic modules each, totaling 120 modules, 15.6 kwp. These arrays are connected to the racks of the inverters, controllers and batteries that are in a building protected by an LPS. The photovoltaic modules are protected by an airtermination system using Franklin rods according the Rolling Sphere Method, LPL I [3, 4, 5, 6], making sure there is no shading on the PV array at any time of the year (see Figs. 9 and 10). In this case, there are 5 Franklin rods with 3 meters of height above of the ground. Besides that there are 6 mini rods with 60 cm each installed on the roof of the room (where are the inverters, controllers and batteries), interconnected with each other forming a ring. This air-termination ring was connected to the earth-termination system with bare copper cable of 35 mm² of gauge.

5 and charge controllers must be essential, regardless of the type of photovoltaic generation system. Fig. 9: LPS project of a minigrid Fig. 10: LPS project details This minigrid was installed on a flat terrain, in an area of 625 m², surrounded by a metal fence that was also connected to earth-termination system. This earth-termination system is formed by rings with bare copper cable of 50 mm² of gauge buried 0.50 cm depth. The Surge Protection Devices (SPD) were specified for the particular arrangement of this minigrid installed in the Amazon region. For the protection of the photovoltaic modules and the electrical equipment s (controllers and inverters) were specified 3 different types of SPD specific for photovoltaic systems. In this way, the external LPS prevent the direct discharges and the internal system (SPD) guards against the induced effects of discharges, within of the probabilities and normalized efficiency. In this case, where the majority of the minigrids is installed far from big cities and electrical grid, the protection described above can be a good solution, because the replacement of a damaged PV module involves a complicated logistics operation. However, the cost of the PV module is much smaller in comparison to the cost of the inverters and charge controllers. In this manner, surge protection of the inverters V. TECHNICAL-ECONOMIC ANALYSIS OF LPS IN PV SYSTEMS Finally, a technical-economic analysis on lightning protection systems in photovoltaic systems is presented. Analyzing the installation of an external Lightning Protection System for a large photovoltaic central of approximately 1MWp, the cost involved are approximately: Materials: R$90.000,00 ~ US$ 41, Services: R$ ,00 ~ US$ 62, Total: R$ ,00 ~US$ 104, Considering the installation of: 162 Franklin type masts without base of 4 meters each; 2200 meters of cooper cable of 50 mm² of gauge for the grounding system, buried in ditches of 0.5 m depth. The cost to replace 1 PV module for this case is R$ 650, equivalent to approximately U$$ 300, considering the cost of the module and the logistics for the replacement. The module considered was a 130 Wp, that in Brazil costs R$ 650 (~US$ 300), and the cost for the logistics, in this case, was not considered because the central is inside of the city. In the case of Minigrids built in remote areas, deployment costs of an LPS are approximately: Materials: R$7.000,00 ~ US$ 3, Services: R$13.000,00 ~US$ 6, Total: R$20.000,00 ~US$ 9, Considering the installation of: 5 masts of 4 meters; 6 air terminals of 60 cm; 150 meters of cooper cable of 50 mm² of gauge for the earth-termination and air-termination subsystems in the container of control, buried in ditches of 0.5 m depth or installed in the container as air-termination and down-conductors subsystems. The cost to replace 1 PV module for this case is R$ 1.650, equivalent to approximately US$ 770, considering the cost of the modules and the logistics for the replacement. The module considered was a 130 Wp, that in Brazil costs R$ 650 (~US$ 300), and the cost for the logistics was R$ (~US$ 470) [8]. VI. CONCLUSIONS This work presents a series of laboratory experiments in photovoltaic modules submitted to current pulses simulating some components of lightning discharges. In these experiments were verified that even being submitted to more severe conditions than those found in the nature, the PV modules can still generate power, despite losing efficiency.

6 The studies realized in large and small photovoltaic generation systems showed that spending on the implementation of a LPS can be compensated by lower costs in replacing damaged photovoltaic modules. The photovoltaic installations, both large power stations and minigrids, are usually located in places surrounded, with restricted access of people and with little intervention of operators. Therefore, the risk of possible lightning that reach these facilities to cause injuries in people is small. The grounding system for the PV modules and gridconnected inverters does not involve high short-circuit current, making this system simpler compared to those used in substations. However, bidirectional inverters used in minigrids require special attention, because they are connected in batteries that have high values of short-circuit current. It is relevant to mention that the surge protection measures are very important for these systems, because the inverters and charge controllers are expensive and the logistics requirements are complicated to replace them in the field. ACKNOWLEDGMENT We thank the staff of the IEE-USP for the great help in the tests: High Voltage Laboratory - Clovis Kodaira, Paulo Marcos Furtado; Laboratory of Photovoltaic Systems Tadeu Osano de Oliveira, Marcelo Pinho Almeida, Aimé Fleury de Carvalho Pinto Neto; Laboratory of High Currents Ricardo Santos D'Avila, Ivan Bueno Raposo and Eduardo Chinen. We also thank the company TERMOTÉCNICA that provided the costs for deployment of LPS. REFERENCES [1] IEC 61215, Crystalline Silicon Terrestrial Photovoltaic (PV) Modules [2] ABNT NBR Proteção de estruturas contra descargas atmosféricas, in portuguese. [3] IEC , First edition, Protection against lightning Part 1: General principles. [4] IEC , First edition, Protection against lightning Part 2: Risk Management. [5] IEC , First edition, Protection against lightning Part 3: Physical damage to structures and life hazard. [6] IEC , First edition, Protection against lightning Part 4: Electrical and Electronic Systems within Structures. [7] M. P. ALMEIDA, Qualificação de sistemas fotovoltaicos conectados à rede Dissertação de mestrado apresentada no Programa de Pósgraduação em Energia da Universidade de São Paulo, in portuguese. [8] A. R. MOCELIN, Instalação e Gestão de Sistemas Fotovoltaicos Domiciliares: Resultados Operacionais de um Projeto Piloto de Aplicação da Resolução ANEEL 83/2004 Dissertação de mestrado apresentada no Programa de Pós-graduação em Energia da Universidade de São Paulo, in portuguese.