Report IEA PVPS T3-11: 2002

IEA PVPS International Energy Agency Implementing Agreement on Photovoltaic Power Systems Task 3 Use of Photovoltaic Power Systems in Stand-Alone and Island Applications Report IEA PVPS T3-11: 2002 <Testing of batteries used in Stand Alone PV Power Supply Systems> Test procedures and examples of test results October 2002 IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 1

Contents Contents...2 Foreword...4 Executive summary...5 Introduction...7 1 Scope and objectives...8 2 Description...8 3 Definitions, symbols and abbreviations...9 4 Cycling test procedures...9 4.1 Methodology...9 4.2 IEC 61427 Standard Cycling test...10 4.2.1 Objective...10 4.2.2 Description...10 4.2.3 Results...11 4.2.4 Comments...12 4.3 NF C 58-510 Standard Cycling test...12 4.3.1 Objective...12 4.3.2 Description...12 4.3.3 Results...13 4.3.4 Comments...14 4.4 PPER Cycling test...14 4.4.1 Objective...14 4.4.2 Description...14 4.4.3 Results...15 4.4.4 Comments...16 4.5 QUALIBAT Cycling test...16 4.5.1 Objective...16 4.5.2 Description...16 4.5.3 Results...17 4.5.4 Comments...18 4.6 Cycling test around 10 % SOC...18 4.6.1 Objective...18 4.6.2 Description...18 4.6.3 Results...19 4.6.4 Comments...20 4.7 Cycling test around 40 % SOC...20 4.7.1 Objective...20 4.7.2 Description...20 4.7.3 Results...21 4.7.4 Comments...22 4.8 DRE Cycling test...22 4.8.1 Objective...22 4.8.2 Description...22 4.8.3 Results...23 4.8.4 Comments...24 5 Comparison of cycling procedures...24 5.1 Overview of all the cycling procedures...24 5.1.1 Brief description...24 5.1.2 Comparison of degradation effects...25 5.2 Comparison of IEC 61427 and NF C 58-510...26 5.2.1 Results...26 5.2.2 Conclusion...26 5.3 Comparison of PPER and QUALIBAT...27 IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 2

5.3.1 Results...27 5.3.2 Conclusion...27 5.4 Comparison of cycling test around 10 % and 40 % SOC...28 5.4.1 Results...28 5.4.2 Conclusion...28 5.5 Evolution of C 10 capacity for tubular and flat plate solar batteries...29 5.6 Summary...30 5.6.1 Conclusion...30 5.6.2 Recommendations...31 6 Efficiency test procedures...32 6.1 NF C 58510 Efficiency test procedure...32 6.1.1 Description...32 6.1.2 Comments...33 6.2 IEC 62093 Efficiency test procedure...33 6.2.1 Description...33 6.2.2 Results...34 6.2.3 Conclusion...35 6.3 Comparison...35 7 Other procedures...35 7.1 Cycling test in the Sandia National Laboratory...35 7.1.1 Objectives...35 7.1.2 Description...35 7.2 Capacity test...36 7.2.1 Objectives...36 7.2.2 Description...36 7.2.3 Results and consequences...37 Conclusion...38 References...40 Annex A: degradation mechanisms...41 IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 3

Foreword The International Energy Agency (IEA), founded in November 1974, is an autonomous body within the framework of the Organisation for Economic Co-operation and Development (OECD), which carries out a comprehensive programme of energy co-operation among its 24 member countries. The European Commission also participates in the work of the Agency. The IEA Photovoltaic Power Systems (PVPS) Programme is one of the collaborative R&D agreements established within the IEA and, since 1993, its Participants have been conducting a variety of joint projects in the applications of photovoltaic conversion of solar energy into electricity. The overall programme is headed by an Executive Committee composed of one representative from each participating country, while the management of individual research projects (Tasks) is the responsibility of Operating Agents. Currently nine tasks have been established. The twenty-one members of the PVPS Programme are: Australia (AUS), Austria (AUT), Canada (CAN), Denmark (DNK), European Commission, Finland (FIN), France (FRA), Germany (DEU), Israel (ISR), Italy (ITA), Japan (JPN), Korea (KOR), Mexico (MEX), Netherlands (NLD), Norway (NOR), Portugal (PRT), Spain (ESP), Sweden (SWE), Switzerland (CHE), United Kingdom (GBR), United States (USA). This International Technical Report has been prepared under the supervision of PVPS Task 3 by: Olivier Bach, Hervé Colin, Daniel Desmettre and Florence Mattera. GENEC-CEA, France in co-operation with experts of the following countries: Australia, Canada, France, Germany, Italy, Japan, Norway, Portugal, Spain, Sweden, Switzerland and the UK. The report expresses, as nearly as possible, a consensus of opinion of the Task 3 experts on the subjects dealt with. IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 4

Executive summary In stand alone PV systems special attention must be paid to the battery bank, which is often said to be the weakest component of the system and the highest contributor to its life cycle cost. For this reason, designers, project managers are very concerned by the evaluation of the battery lifetime. This document makes an analysis of cycling test procedures for batteries used in stand alone photovoltaic power systems. The main objective of a cycling test procedure is to evaluate the battery lifetime, expressed in terms of reduced capacity, by reproducing a typical use of the battery in the field. This document presents an overview of frequently used battery test procedures. A description of seven cycling test procedures is made and the performance of more than 40 batteries of various types is presented. The test procedures induce ageing of the battery by accelerating the degradation (sulphation, corrosion, stratification and softening) of the battery grid and the active mass. The analysis of the results shows that most of the cycling procedures lead to significant battery sulphation, but almost none of them accelerate battery corrosion. In fact, there is no test procedure where the four degradation mechanisms lead to a significantly high level of degradation. It is therefore advisable for project managers to select several methods in order to ascertain the most likely battery degradation pattern for a given application and then be able to choose the right battery. In terms of battery technology, the results show that there is a wide range of efficiency and lifetime values for various batteries of the same technology. This is due to differences in the type of grid alloy, the active mass composition, the manufacturing process or the cell geometry. In addition, the longer and better service of tubular batteries was confirmed by the analysis. The behaviour of tubular batteries is, in general, not very dependent on the selected test procedure, as this kind of battery is more sensitive to sulphation than to softening and corrosion. However, flat plate solar batteries are sensitive to all types of degradation and their cycling life is much more dependent on the test procedure. Nevertheless the present results show that there is still a lot of work to do on the topic of accelerated battery cycling life evaluation. There is a strong need for a deeper international collaboration (Round Robin tests) to focus on a limited number of cycling procedures, gather more results on battery tests, search for other means to accelerate the tests (key parameters, modelling studies) and lead to the standardisation of these cycling tests. Résumé Dans les systèmes photovoltaïques isolés, une attention particulière doit être portée au banc de batteries car celles-ci sont souvent supposées constituer le composant le plus faible du système et être le plus coûteux durant la vie de ce système. De ce fait, les concepteurs et chefs de projets sont très concernés par l évaluation de la durée de vie d une batterie. Ce document analyse différentes procédures de test en cyclage de batteries, utilisées dans des systèmes photovoltaïques autonomes. Le principal objectif d une telle procédure est d évaluer la durée de vie de la batterie, exprimée en termes de perte de capacité, au cours du test, en reproduisant une utilisation typique ou spécifique sur le terrain. Ce rapport fait la revue de procédures de test de batteries fréquemment utilisées. La description de sept procédures de cyclage est faite et les résultats donnant la perte de capacité de plus de quarante batteries de différents types selon des profiles de charge/décharge divers sont présentés. Les cycles ont pour but de provoquer un vieillissement de la batterie en accélérant le processus de dégradation (sulfatation, stratification, corrosion, ramollissement) au niveau de IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 5

la grille de la batterie et de la matière active. L analyse des résultats montre que la plupart des procédures conduit à une sulfatation significative de la batterie, mais que presque aucune n accélère le phénomène de corrosion. En fait, il n existe pas de procédure où les quatre types de dégradation atteignent simultanément un niveau important. De ce fait, les professionnels impliqués dans la sélection d une batterie peuvent être amenés à associer plusieurs méthodes de test afin de générer les dégradations attendues pour une application donnée et choisir ainsi les batteries «en bonne connaissance de cause». Certains résultats montrent qu il y a une grande dispersion des valeurs d efficacité ou de durée de vie pour des batteries de même technologie suite à des différences dans la nature de l alliage de la grille, la composition de la matière active, le processus de fabrication ou dans la géométrie des cellules. Ils confirment également la capacité des batteries tubulaires à fournir un service plus long et de meilleure qualité. Le choix d une méthode de cyclage pour une technologie de batterie donnée se fait en tenant compte du fait que les résultats des batteries tubulaires dépendent en général peu de la procédure car celles-ci sont peu sensibles à la corrosion de la grille et au ramollissement de la matière active mais plus sensibles à la sulfatation, tandis que les batteries solaires à plaques planes sont sensibles à tous les types de dégradation, ce qui rend leur comportement plus dépendant des caractéristiques de la procédure de test. Néanmoins, les résultats actuels montrent que le travail sur ce sujet n est pas achevé. Il y a un fort besoin d une collaboration internationale plus poussée pour se focaliser sur un nombre limité de procédures de cyclage, rassembler plus de données sur les batteries testées, chercher d autres moyens pour accélérer les tests (paramètres clés, étude de modélisation) et tendre vers une standardisation de ces tests de cyclage. Keywords: battery, cycling, degradation, lifetime IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 6

Introduction In stand alone PV systems special attention must be paid to the battery bank, which is often said to be the weakest component, in order to guaranty a satisfactory service to the end user. One of the aims of Task 3 is to increase the know-how on the storage function in SAPV systems in order to decrease its cost and increase the battery performance. Task 3 has therefore developed a specific subtask dealing with the storage function. This document describes the work being done within Task 3 to assess various procedures used to test batteries that are integrated in photovoltaic systems. There are a number of reasons to consider battery testing procedures: o The impact of battery on the PV system life cycle cost and reliability, o The choice of the appropriate battery for a specific application, o The management of the battery (this is studied in another Task 3 document [1] to be published), o Characteristics (efficiency, lifetime, etc.) may vary from one battery to another, even for the same design (especially for automotive and flat plate batteries, which are not well adapted to cycling, due to the industrial manufacturing process), o Due to specific constraints and operation conditions in the photovoltaic field, conventional battery tests are not appropriate, o Testing a battery using the existing procedures costs money and is very time consuming. Results of battery tests are often only available after a year or two. Today many testing procedures are available, but this makes it harder to choose the right procedure that suits the desired application. For specific applications, testing laboratories even develop proprietary procedures. This document aims to present an overview of frequently used battery test procedures [2], while emphasising the lessons that can be drawn in terms of typical battery degradation effects. Acknowledgements: This report integrates results coming from a contract funded by ADEME and other contracts such as the Qualibat project, funded by the European Commission. IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 7

1 Scope and objectives The general objective of this document is to summarise the experience of lead-acid battery testing. The report is intended to describe the main features of batteries, which can be measured in laboratories or in the field such as: o Initial capacity compared to rated capacity; o Efficiency; o Rate of capacity reduction with respect to certain charge/discharge profiles. The main results presented in the report are: o Details of seven cycling test procedures used internationally; o Performance of more than 40 batteries of various types in terms of capacity reduction versus charge/discharge profiles; o A comparison of the impact of various cycling procedures on similar batteries; o An evaluation of battery efficiency. The objectives of this report are to: o Provide project managers of photovoltaic applications with data to assist in battery selection and to help professionals involved in the design of PV systems; o Focus the activities of laboratories involved in battery testing on fewer selected test procedures; o Make more battery test data available. The main goal was to cover most of the technologies available on the market and to give a general overview of battery behaviour. It was not intended to select one model or one manufacturer, as this would have required testing of more samples of the same battery model. The batteries tested have mainly been purchased in small quantities from various retailers. 2 Description The first section of this report deals with cycling procedures. Each of them is described technically and summary test results are then presented. The base of the work is the cycling test study. These cycles induce ageing of the battery by accelerating the degradation of the battery grid and the active mass. The second part makes a comparison of the cycling procedures. Procedures that have a similar effect on battery ageing are compared with each other. Finally, an overview of all the cycling procedures that synthesise the data available to date is presented. The third part is dedicated to battery efficiency evaluation. Two procedures coming from the French Standard and an IEC Standard are described and compared. A fourth part gives some information about other battery test procedures. IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 8

3 Definitions, symbols and abbreviations Ah = Ampere hour C 10 = Battery capacity for a discharge time of 10 hours C 100 = Battery capacity for a discharge time of 100 hours DRE = Decentralised rural electrification IEC = International Electrotechnical Commission I 10 = Battery current for a discharge time of 10 hours I 100 = Battery current for a discharge time of 100 hours NF = French standard Ni-Cd = Nickel cadmium PPER = Rural Electrification Plan in Morocco QUALIBAT = QUicker Assessment of LIfetime and other characteristics of PV BATteries SOC = State of charge Vpc = Volt per cell VRLA = Valve-regulated lead-acid AGM = Absorptive (or absorbed) Glass Mat SLI = Starting Lighting Ignition SHS = Solar Home System 4 Cycling test procedures The main objective of a cycling test procedure is to evaluate the battery lifetime, expressed in terms of capacity reduction (rate at which the useable capacity declines), at each point in the course of the test procedure. The typical design of a cycling test procedure is defined by: o A sufficient number of cycles that allow a significant effect on battery capacity to be seen (in this report, a loss of 30 % is considered to be significant), o A periodic cycle profile aimed at representing the operating conditions of the battery in the field, with controlled parameters as close as possible in order to guarantee the reproducibility of the test, o The operating temperature. 4.1 Methodology The method adopted to present the results is detailed as follows: The y-axis of the graph, named evolution of C 10 capacity, represents the relative capacity at a given cycle, i.e. the ratio between the measured capacity and the rated capacity (the discrepancy between the rated capacity (given by the manufacturer) and the initial capacity (measured) explains why the ratio may be bigger than one, especially for tubular batteries, in the next graphs). The x-axis of this graph is usually expressed as a number of cycles. Although this approach is useful to compare different batteries within the same test procedure, it is meaningless for a comparison over different test procedures. So, in this report, the unit of the x-axis is the total number of Ah cycles through the battery divided by the rated capacity. Such a ratio represents the number of times the rated capacity is delivered by the battery. This ratio enables a comparison of the service delivered by a battery, regardless of the test procedures. On the x-axis the duration of the corresponding experiment is mentioned. Instead of plotting the curve for each battery tested, only a few of them are presented. Six types of technologies were selected: o Flat plate car batteries (flooded), o Flat plate stationary batteries (flooded), which are improved car batteries, IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 9

o Flat plate solar batteries (flooded), which have thicker plates and additives in the electrolyte compared to car batteries, o Tubular batteries (flooded), o VRLA batteries (AGM and gel), o Ni-Cd batteries. Note: all the battery types have not been tested in each cycling test procedure. The battery capacity values are miscellaneous and vary from one test to another. For each type of battery the behaviour range is delimited by the best and the worst batteries and has been represented taking into account the accuracy of the measurements. All results have the same unit and the same colour code for a same battery technology in order to be compared easily. Afterwards the average lifetime for each battery family was estimated. This lifetime is defined as the number of times the capacity is delivered before the battery has lost 30 % of its initial capacity. Sometimes the experiment was not carried out long enough to see the 30 % capacity reduction, so in this case an extrapolation is made to estimate the lifetime of the batteries; but due to the non linear nature of the ageing curves, this extrapolation only gives a rough evaluation of it. To complete the analysis the four main types of degradation (softening, corrosion, sulphation and stratification) associated with each test procedure have been quantified. The evaluation of the weighting of each type of degradation for each test procedure has been done according to an analysis of the active mass at the end of the battery lifetime. This analysis consists of the quantity determination of chemical species, the characterisation of the sulphate crystals and the observation of the deposition and the corrosion layer. The four mechanisms are described in Annex A. 4.2 IEC 61427 Standard Cycling test 1 4.2.1 Objective This procedure uses the principle of performing cycles at two different states of charge [3]. Cycles are carried out successively around the low state of charge (20 %), then around the high state of charge (80 %) so as to reproduce the typical operating conditions of a battery exposed to a seasonal cycle. It tends to represent fault conditions, when the PV system sizing is not very good or the weather very bad and/or very good. 4.2.2 Description The test conditions are as follows: o Fifty cycles of 30 % depth of discharge are carried out between 5 and 35 % SOC, then a hundred cycles are carried out between 75 and 100 % SOC. These 150 cycles are immediately followed by a capacity measurement and form a test sequence, o A period lasts about 50 days. There are between 3 and 10 periods in a single test, requiring 5 to 16 months to complete the procedure. Several test sequences are repeated until an "end of test" criterion is reached, o The temperature of the battery bath is 40 C (The high temperature at which the test takes place aims at accelerating the degradation and thus reducing the time necessary to reach the end of life criterion). A period of this test procedure is presented in Figure 1: 1 The test procedure described here is slightly different from the IEC 61427 standard. The only difference is a 9h00 discharge in the phase A for the IEC 61427 standard instead of a 9h30 discharge for the studied cycling test procedure. IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 10

Figure 1 - IEC 61427 Standard cycling test procedure 4.2.3 Results Figure 2 shows the evolution of the C 10 capacity for tubular plate and flat plate batteries. Figure 2 - C 10 capacity evolution during IEC test procedure Figure 3 shows the expected number of times the rated capacity is delivered (life time) for each technology when this test procedure is used; this number is calculated when the C 10 capacity is equal to 70 % of the initial C 10 capacity. A minimum value and a maximum value, IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 11

obtained by extrapolation when the end of life criterion was not reached (cf. the checked bar), are presented according to this test and other experiments. 4.2.4 Comments Figure 3 - Service delivered over the battery lifetime (IEC test procedure) This test procedure is not very severe and requires a long time period in order to reach significant degradation. It slightly emphasises degradation by corrosion because of the 100 cycles at high state of charge during the phase B of the procedure. As tubular batteries are not very sensitive to corrosion, the number of times the rated capacity is delivered is very high and the test duration is long. In comparison flat plate solar batteries are rapidly affected by the degradation mechanisms and reach the end of test criterion in a short time. 4.3 NF C 58-510 Standard Cycling test 4.3.1 Objective This procedure [4] aims at simulating real operating conditions of a battery in a photovoltaic installation, with constant daily cycle configuration and deep seasonal cycles, causing changes in the state of charge during the different seasons. It is designed to represent the cycling conditions of a PV system in temperate countries, for which seasonal cycling are significant. 4.3.2 Description The test conditions are as follows: o A complete test sequence consists of three stages (stage A, stage B and stage A again) and two capacity measurements (C 10 and C 100 ) 2. In stage A (see Figure 4), the battery undergoes shallow cycles with 3 hours discharge at the 6.6*I 100 rate and recharge for 4 hours at the 4.85*I 100 rate until the low voltage threshold is reached. The charging regime is not sufficient to fully charge the battery so that the SOC of the battery progressively decreases during stage A. The number of cycles needed to reach the low voltage threshold ranges from 30 to 80 cycles, depending on the battery type; In stage B, the battery undergoes shallow cycles with the same discharge regime as in stage A and 4 hours recharge at the 5.45*I 100 rate so that the SOC of the battery progressively increases with cycling. The number of cycles in stage B is by definition the same as in stage A, o A test sequence lasts between 50 and 90 days. The total duration ranges from 3 to 10 test sequences, therefore from 3 months to two years. Several test sequences are repeated until an "end of test" criterion is reached, 2 Prior to the capacity measurements, the battery is discharged at I 100 down to 1.85 Vpc (flooded) or 1.80 Vpc (VRLA), then fully charged and maintained during 48 hours at 2.35 Vpc (flooded) or 2.30 Vpc (VRLA). IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 12

o The temperature of the battery bath is 40 C (this value is not representative of climate conditions in temperate countries but is used to accelerate the test procedure). A period of this test procedure is presented in Figure 4: 4.3.3 Results Figure 4 - NFC Standard cycling test procedure Figure 5 shows the evolution of the C 10 capacity for three types of battery: tubular plate, flat plate solar and flat plate car batteries. Sixteen batteries have been tested: 5 tubular plate batteries, 7 flat plate solar batteries and 4 flat plate car batteries. Figure 5 - C 10 capacity evolution during NFC test procedure Figure 6 shows the expected number of times the rated capacity is delivered (lifetime) for each technology when the NF C procedure is used; this number is calculated when the C 10 capacity is equal to 70 % of the initial C 10 capacity. A minimum value and a maximum value, obtained by extrapolation when the end of life criterion was not reached (cf. the checked bar), are presented according to this test. IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 13

Figure 6 - Rendered service over the battery lifetime (NFC test procedure) 4.3.4 Comments This procedure does not emphasise the corrosion ageing mechanism, but is a severe cycling test in term of sulphation and stratification mechanisms as the cycling test is operated at a medium average state of charge. Tubular batteries exhibit a better rendered service than flat plate batteries but, as they are sensitive to sulphation and stratification, the number of times the rated capacity is delivered is quite a lot smaller than the one obtained with the IEC procedure (maximum of 350 compared to 750). Figure 6 also shows that the flat plate car batteries can have better performance than flat plate solar batteries when car batteries are manufactured with a hand-made process compared to an automatic process. 4.4 PPER Cycling test 4.4.1 Objective This procedure has been developed by GENEC for a fountain application in Morocco [5]. In this situation, people carry their batteries to the fountain (PV system) to recharge them; so that the batteries are often deeply discharged and are not recharged immediately after use. Therefore, the test procedure uses deep cycles over a one-day period in order to be close to the real operating conditions of a battery in a solar home system. 4.4.2 Description The test conditions are as follows: o The battery is discharged at the I 10 rate until a cut-off voltage of 1.75 V per cell, left at free potential for 3 hours and recharged for 8 hours at the 1.5*I 10 rate, o Each cycle induces an ageing effect on the battery and measures the C 10 capacity at the same time (this allows the close control of the changes in the evolution of the capacity). o Several cycles are repeated until an "end of test" criterion is reached. The total duration ranges from 2 weeks to 4 months. o The temperature of the battery bath is 25 C. Some cycles of this test procedure are schematically represented in Figure 7: IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 14

4.4.3 Results Figure 7 - PPER cycling test procedure Figure 8 shows the evolution of the C 10 capacity for two types of battery: flat plate solar and flat plate car batteries. Four batteries have been tested: 2 flat plate solar batteries and 2 flat plate car batteries. Figure 8 - C 10 capacity evolution during PPER test procedure Figure 9 shows the expected number of times the rated capacity is delivered (lifetime) for each technology when this test procedure is used; this number is calculated when the C 10 capacity is equal to 70 % of the initial C 10 capacity. A minimum value and a maximum value, obtained by extrapolation when the end of life criterion was not reached (cf. the checked part of the bar), are presented according to this test. IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 15

Figure 9 - Rendered service over the battery lifetime (PPER test procedure) 4.4.4 Comments This test procedure is favourable for the softening mechanism due to the large daily variations of the state of charge and for the sulphation mechanism as the battery is often at a very low state of charge. This test was dedicated to car batteries and solar batteries of poor quality that are often used in Morocco. As the test procedure allows 120% charge return to this kind of batteries, the weight of the stratification phenomenon is not important. Solar batteries have better performances than car batteries with the exception of the high quality car battery as seen in section 4.3.4. 4.5 QUALIBAT Cycling test 4.5.1 Objective This test, developed within the framework of the European project QUALIBAT [6], is used to rapidly assess the cycling ability of different battery designs for PV applications. The speed of this test is achieved by setting a high discharge rate and performing the test at 40 C. This new cycling procedure has been designed and adjusted by research and testing laboratories and a battery manufacturer. Two kinds of test were defined according to the two main battery types: flooded and VRLA. 4.5.2 Description The test conditions are as follows: o The cycling test for flooded batteries requires 3 cycles per day at a relatively deep depth of discharge (66 %) and high current. For the VRLA products 1.5 cycles per day are performed at a similar depth of discharge and lower current. A total of 50 or 100 cycles of charge/discharge are followed by a capacity measurement and this forms a test sequence, o Several test sequences are repeated until an "end of test" criterion is reached. A test sequence lasts between one and two months. Total duration ranges from one to seven months, o The temperature of the battery bath is 40 C. A period of this test procedure is presented in Figure 10: IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 16

Figure 10 - QUALIBAT cycling test procedure The charge and discharge rates are presented in Table 1. Table 1 - Charge and discharge rates (Qualibat test procedure) Battery type Discharge Charge Flooded 2 h at 3.3 I 10 2 h at 2.5 I 10 + 4 h at 0.71 I 10 VRLA 3 h at 2.2 I 10 10 h at 1.9 I 10 (2.28 Vpc limit) + 3 h at 0.1 I 10 The capacity measurement is made every 50 or 100 cycles with a discharge at I 10 to 1.8 Vpc. 4.5.3 Results Figure 11 shows the best and the worst evolution of the C 10 capacity for four types of battery: tubular plate, flat plate solar, VLRA and flat plate car batteries. Thirty batteries have been tested: 3 tubular plate batteries, 10 flat plate solar batteries, 2 flat plate car batteries and 15 VRLA batteries (AGM and gel types). Figure11 - C 10 capacity evolution during QUALIBAT test procedure IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 17

Figure 12 shows the expected number of times the rated capacity is delivered (life time) for each technology when this test procedure is used; this number is calculated when the C 10 capacity is equal to 70 % of the initial C 10 capacity. A minimum value and a maximum value, obtained by extrapolation when the end of life criterion was not reached (checked bar), are presented according to this test. Figure 12 - Rendered service over the battery lifetime (QUALIBAT test procedure) 4.5.4 Comments Chemical examination of several batteries tested with this procedure showed that some of them failed by shedding of their positive active material with enlarged pore size and some by sulphation. In contrast, there is little acid stratification, little grid corrosion and only the start of positive active material softening. In conclusion this procedure is specifically designed to emphasise softening. Flat plate car batteries are much affected by the degradation mechanisms caused by this procedure (sulphation and softening). Concerning the other types of batteries there is a wide dispersion of the results. 4.6 Cycling test around 10 % SOC 4.6.1 Objective At the time when no standard was available, an accelerated cycling test procedure was developed and proposed [7] to the professionals of the PV sector. This procedure is designed to test batteries in a low state of charge (between 0 and 20 % of SOC) as the battery often operates in those conditions in case of prolonged bad weather conditions and when the user consumes energy as soon as possible. The principle of this procedure was partially adopted by the IEC standard in the phase A (see section 4.2.2). 4.6.2 Description The test conditions are as follows: o Prior to the beginning of cycling, a discharge is made to 1.75 Vpc (0 % SOC), o Three hundred cycles of charge/discharge in a low state of charge (10 %) are followed by two capacity measurements (C 10 and C 100 ) and form a test sequence, o A correction is made after each cycle in order to avoid a drift in the state of charge (automatic calculation of the recharge coefficient), o Several test sequences are repeated until an "end of test" criterion is reached. A test sequence lasts about three months. The total duration of the test procedure ranges from three months to more than one year, o The battery bath temperature is 40 C. The first cycles of this test procedure are presented in Figure 13: IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 18

Figure 13 - Cycling test around 10 % SOC procedure The capacity measurements are performed after a full recharge (120 % of charge) of the battery. 4.6.3 Results Figure 14 shows the evolution of the C 10 capacity for four types of battery: tubular plate, VRLA (AGM), flat plate stationary and Nickel-Cadmium batteries. Figure 14 - C 10 capacity evolution during the cycling test around 10 % SOC procedure Figure 15 shows the expected number of times the rated capacity is delivered (life time) for each technology when this test procedure is used; this number is calculated when the C 10 capacity is equal to 70 % of the initial C 10 capacity. A minimum value and a maximum value, obtained by extrapolation when the end of life criterion was not reached (checked bar), are presented according to this test and other experiments. IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 19

Figure 15 - Rendered service over the battery lifetime ( cycling test around 10 % SOC procedure) 4.6.4 Comments This procedure is specifically designed to emphasise the stratification and sulphation mechanisms as the battery is maintained at low states of charge (0 up to 20 % SOC). In this test, stationary batteries exhibit good performances because they derive from flat plate car batteries that were improved (for instance, thicker plates) for solar purposes. Tubular batteries, which are sensitive to sulphation and stratification, reach the end of life criterion in a rather short time (300 cycles). 4.7 Cycling test around 40 % SOC 4.7.1 Objective As for the cycling test around 10 % SOC, this accelerated cycling test procedure was developed at the time when no standard was available, and it was proposed [7] to the professionals of the PV sector. This procedure is designed to test batteries in a medium state of charge (between 30 and 50 % of SOC) as the battery often operates in those conditions in case of prolonged bad weather conditions. 4.7.2 Description The test conditions are as follows: o Prior to the beginning of cycling, a discharge at I 10 is made to 30 % SOC (amperehour counting), o Three hundred cycles of charge/discharge in a medium state of charge (40 %) are followed by two capacity measurements (C 10 and C 100 ) and form a test sequence, o Several test sequences are repeated until an "end of test" criterion is reached. A test sequence lasts about three months. The total duration ranges from three months to more than one year, o The battery bath temperature is 40 C. The first cycles of this test procedure are presented in Figure 16: IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 20

Figure 16 - Cycling test around 40 % SOC procedure The capacity measurements take place after a full recharge (120 % of charge) of the battery. 4.7.3 Results Figure 17 shows the evolution of the C 10 capacity for four types of battery: tubular plate, VRLA (AGM), flat plate stationary and Nickel-Cadmium batteries. Figure 17 - C 10 capacity evolution during the cycling test around 40 % SOC procedure Figure 18 shows the expected number of times the rated capacity is delivered (life time) for each technology when this test procedure is used; this number is calculated when the C 10 capacity is equal to 70 % of the initial C 10 capacity. A minimum value and a maximum value, obtained by extrapolation when the end of life criterion was not reached (checked bar), are presented according to this test and other experiments. IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 21

Figure 18 - Rendered service over the battery lifetime ( cycling test around 40 % SOC procedure) 4.7.4 Comments This procedure is specifically designed to emphasise the stratification and sulphation mechanisms as the battery is maintained at low states of charge (30 up to 50 % SOC). In this test, stationary batteries exhibit good performances because they derive from flat plate car batteries that were improved (for instance, thicker plates) for solar purposes. Tubular batteries, which are sensitive to sulphation and stratification, reach the end of life criterion in a rather short time (320 cycles). In comparison with the test around 10 % SOC, the results are very similar because, even if the sulphation and stratification mechanisms are supposed to be less favoured when the SOC is near 40 %, it is sufficient to degrade the performances of the batteries. 4.8 DRE Cycling test 4.8.1 Objective This test procedure was developed by GENEC for the Decentralised Rural Electrification (DRE) Directives [8]. The objective was to accelerate the sulphation and stratification degradation mechanisms in very few cycles and rehabilitate the battery also in few cycles. 4.8.2 Description The test conditions are as follows: o Stratification is established by applying 5 charge/discharge cycles with a recharge limited by a high voltage threshold (2.35 Vpc). Such a threshold voltage is insufficient for fully recharging the batteries. These cycles are followed by 5 cycles with a recharge coefficient of 1.2 C 10. The applied overcharge thus enables the battery to recover from the sulphation induced during the 5 first cycles with low recharge. The development of the restored capacities allows an evaluation of the ability of the battery to recover its initial state. The series of 10 cycles is followed by a capacity measurement and form a test sequence, o Several test sequences are repeated until an "end of test" criterion is reached. A test sequence lasts about 12 days. The total duration ranges from 3 to 16 test sequences, therefore from 1 to 7 months, o The battery bath temperature is 25 C. Both types of charge/discharge cycles are presented in Figure 19: IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 22

Figure 19 - DRE cycling test procedure The characteristics of each type of cycle are given hereafter: 4.8.3 Results Figure 20 shows the evolution of the C 10 capacity for three types of battery: flat plate solar, flat plate car and VRLA (AGM) batteries. Figure 20 - C 10 capacity evolution during the DRE test procedure IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 23

Figure 21 shows the expected number of times the rated capacity is delivered (life time) for each technology when this test procedure is used; this number is calculated when the C 10 capacity is equal to 70 % of the initial C 10 capacity. 4.8.4 Comments Figure 21 - Rendered service over the battery lifetime (DRE procedure) This test does not emphasise a particular degradation, i.e. the sulphation, stratification and softening mechanisms are accelerated at a same level. The three types of batteries tested with this procedure are sensitive to these degradations and reach rapidly the end of life criterion. 5 Comparison of cycling procedures In this chapter an overview of the battery degradation modes emphasised by each respective cycling test procedure is presented in order to synthesise the available data (see annex A for a short description of these degradations). Subsequently, the cycling test procedures, which induce similar degradations, are compared with each other more in detail. The objective of the comparison is to identify the best procedure that will emphasise the various types of degradation in the battery, but which can be achieved and give a reliable result in a relatively short period of time. Another objective is to determine how the different test procedures may complement each other. Finally, this section presents a summary table, which details the parameters of each test procedure and summarises the outstanding results obtained in terms of battery technologies and degradation effects. 5.1 Overview of all the cycling procedures This section gathers the main characteristics of the seven test procedures that have been described in this document. 5.1.1 Brief description A brief description of the seven test procedures is given in Table 2: Table 2 - Overview of the seven test procedures Logo Name Description Total duration IEC 61427 Standard On page 10 From 5 months to more than 1 year NF C 58-510 Standard On page 12 From 3 months to 2 years PPER on page 14 From 2 weeks to 4 months IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 24

Table 2 (concluded) Logo Name Description Total duration QUALIBAT on page 16 From 1 month to 7 months Test around 10 % SOC On page 18 From 3 months to more than one year Test around 40 % SOC On page 20 From 3 months to more than one year DRE on page 22 From 1 month to 7 months The test duration depends on the battery technology and may vary from one sample to another. 5.1.2 Comparison of degradation effects Figure 22 shows the results of the relative comparison of the degradation effects for the seven test methods: Figure 22 - Comparison of degradation effects Most of the cycling procedures lead to significant battery sulphation. In contrast, almost none of them accelerate battery corrosion (none of the procedures really overcharge the batteries more than 20 % per cycle). There is no test procedure where the four mechanisms lead to high overall degradation. Therefore it is necessary to use at least two test methods in order to observe the development of almost all the types of degradation. IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 25

5.2 Comparison of IEC 61427 and NF C 58-510 5.2.1 Results Figure 23 shows the best and worst changes in the C 10 capacity for two technologies of battery (tubular plate and flat plate solar batteries). Three samples of identical batteries are tested with both procedures. Four types of batteries have been tested: 1 flat plate solar battery and 3 tubular plate batteries. Figure 23 - C 10 capacity changes during IEC and NF test procedures This graph clearly shows that the ageing effect of an identical battery can be quite different depending on the test used to evaluate it (see tubular battery 2 in Figure 23). The impact of the four degradation effects can be visualised in the Figure 24. A post mortem examination of the active mass, the analysis of the lead sulphate crystals size and the quantity determination of chemical species lead to the following quantitative results: 10 Weight of degradations 5 0 Softening Corrosion Sulphation Stratification IEC NFC Figure 24 - Impact of degradation (IEC and NF test procedures) The scale of the weight of degradation varies from 0 (no impact) up to 10 (high impact). 5.2.2 Conclusion Both procedures allow the evaluation of battery behaviour when operating at different states of charge, which are close to real operating conditions. Each test emphasises a different type IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 26

of degradation. However, both tests show that tubular batteries are better since they can cycle longer than flat plate batteries. Both procedures require a few months to two years, and are similar in duration. 5.3 Comparison of PPER and QUALIBAT 5.3.1 Results Figure 25 shows the best and worst evolutions of the C 10 capacity for two technologies of battery (flat plate car and flat plate solar batteries). Three samples of identical batteries are tested with both procedures. Four types of batteries have been tested: 2 flat plate solar batteries and 2 flat plate car batteries. Figure 25 - C 10 capacity changes during PPER and QUALIBAT test procedures The impact of the four degradation effects can be visualised in Figure 26: 10 Weight of degradations 5 0 Softening Corrosion Sulphation Stratification PPER QUALIBAT Figure 26 - Impact of degradation (PPER and QUALIBAT test procedures) 5.3.2 Conclusion Both procedures are designed to cycle the batteries deeply so that they emphasise the ageing by softening. The higher current in the QUALIBAT procedure compared to the PPER procedure accelerates the ageing rate of batteries, especially for flat plate car batteries (see Figure 24). IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 27

5.4 Comparison of cycling test around 10 % and 40 % SOC 5.4.1 Results Figure 27 shows the best and worst changes in the C 10 capacity for four battery technologies (tubular plate, flat plate car and Ni-Cd batteries). Three samples of identical batteries are tested with both procedures. Four types of batteries have been tested: 2 tubular plate batteries, 1 flat plate stationary battery and 1 Ni-Cd battery. Figure 27 - C 10 capacity changes during cycling test around 10 and 40 % SOC procedures The impact of the four degradation effects can be visualised in Figure 28: 10 Weight of degradations 5 0 Softening Corrosion Sulphation Stratification 10% SOC 40% SOC Figure 28 - Impact of degradation (cycling test around 10 and 40 % SOC procedures) 5.4.2 Conclusion Both procedures are equivalent and effectively emphasise the degradation of the batteries by causing a high rate of stratification associated with a sulphation of the electrodes. Operating at 10 % SOC is a little bit more aggressive but the difference is quite small compared to an operation at 40 % SOC. IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 28

5.5 Evolution of C 10 capacity for tubular and flat plate solar batteries A comparison of the worst tubular batteries tested in each procedure can be seen in Figure 29: Figure 29 - C 10 capacity changes for the worst tubular batteries The evolution of the worst flat plate solar batteries tested in each procedure is displayed in Figure 30. Figure 30 - C 10 capacity evolution for worst flat plate solar batteries In all of the test procedures it can be seen that the behaviour of the worst type of tubular batteries is quite similar. In fact tubular batteries are not very sensitive to softening and corrosion but more to sulphation and stratification. As the importance of sulphation in the different test procedures is quite constant, the tubular batteries behaviour in general (not only the worst types) is not very dependent on the test procedure that is used. In contrast, flat plate solar batteries can react differently depending on the type of degradation emphasised by the test procedure. In fact, flat plate solar batteries are sensitive to all kinds of degradation effects: corrosion, softening, sulphation and stratification. IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 29

5.6 Summary 5.6.1 Conclusion Table 3 presents the parameters and results of the seven test procedures. Table 3 - Battery cycling test procedure characteristics IEC 61427 NF 58510 PPER Qualibat 10 % SOC 40 % SOC DRE Number of periods 3 to 10 3 to 10 Not defined 50 to 100 300 300 3 to 16 Total duration (months) 5 to 12 3 to 24 0.5 to 4 1 to 7 3 to 13 3 to 13 1 to 7 Depth of discharge (%) 30 20 90 60 20 20 90 Current High Low High High Moderate Moderate High Bath temperature ( C) 40 40 25 40 40 40 25 Number of C 10 capacities released Degradation emphasised Tubular 450 to 750 220 to 350 250-400 250 to 375 250 to 380 Flat solar 70 to 200 80 to 120 90 to 95 35 to 335 80-120 Flat car 60 to 135 10 to 95 70 25-70 Stationary 350 to 400 350 to 450 VRLA 60 to 350 250 to 300 250-270 120 to 140 Ni-Cd 50 to 120 55 to 120 Corrosion Sulphation Softening Softening Stratification Stratification Stratification Sulphation Sulphation Sulphation Sulphation Stratification Sulphation Softening The first observation is that there is a very high dispersion between batteries in the same technological family. This is easily explained by the differences within a same family; the grid alloy, active mass composition, manufacturing process and cell geometry are all factors, which give the batteries quite different performance attributes. The IEC and NF cycling test procedures emphasise the differences between the technologies most strongly. These two procedures have the common characteristic of being standard procedures for the accelerated ageing of PV batteries. The battery lifetime is similar when using procedures PPER, QUALIBAT and DRE. These three procedures include large-span cycles, representing a PV system with a very low autonomy, as can be the case for SHS where the user consumption is so high that the full capacity of the battery is discharged every day. In contrast, procedures around 10 % and 40 % SOC do not give a clear classification of the technologies, and the lifetime of the different batteries is quite similar. In conclusion, all the procedures studied are reasonably representative of the field conditions, at least from a degradation perspective. Table 4 presents a summary of the advantages/disadvantages of the cycling procedures: Table 4 - Advantages/disadvantages of the seven cycling procedures Procedure Main interests Disadvantages IEC 61427 Standard o Reproduces seasonal cycles (low & high o Not very severe SOC) o Long duration for tubular o Accelerates the corrosion mechanism batteries o Control of capacity loss at end of sequence NF C 58-510 o Reproduces daily and seasonal cycles o Long duration for tubular Standard o Cycling conditions of temperate countries batteries o Control of capacity loss at end of sequence IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 30

Table 4 (concluded) Procedure Main interests Disadvantages PPER o Representative of SHS operating conditions o Designed for flat plate batteries o Short duration o Control of capacity loss during sequence QUALIBAT o Short duration o Control of capacity loss at end of sequence Test around 10 % o Strongly emphasises stratification and o Specific for operating at very SOC sulphation low SOC o Reproduces situations of prolonged bad weather conditions o Control of capacity loss at end of sequence Test around 40 % o Emphasises stratification and sulphation o Control of capacity loss at SOC o Representative of SHS operating end of sequence conditions DRE o Short duration o Control of capacity loss at o Accelerates stratification and sulphation in very few cycles end of sequence (but the sequence is short) o Shows ability of a battery to recover from stratification o Doesn t emphasise a particular degradation 5.6.2 Recommendations According to the test procedure that is chosen, batteries that are sensitive to a given type of degradation will respond differently in terms of both degradation and lifetime. Therefore the most appropriate battery cycling test procedure must be chosen to suit the intended conditions of use (daily use, seasonal use, use at low or medium state of charge, etc). This selection enables an evaluation of the battery ageing with a test that is optimised in terms of duration. As has been seen in Section 5.1.2 it may be necessary to associate at least two test methods to establish different degradation effects. This must be done in order to deliver realistic results concerning the expected lifetime of a given type of battery for a given application. For instance, the QUALIBAT procedure and the test around 10 % of SOC could be used together because they are both aggressive (and rapid) and complementary from the perspective of obtaining a balanced degradation. Table 5 identifies which operating conditions favour which degradation mechanisms for a given type of battery. It indicates what type of battery is sensitive to a degradation mode caused by given conditions. Used together with Figure 22, which illustrates the mechanisms generated by each cycling test, it offers the possibility to choose the right testing procedure. Table 5 - Sensitivity of batteries to the degradation modes Conditions Sensitivity of each battery type Preferential degradation mode Explanation Tubular Flat plates VRLA SLI Stationary Deep cycles X X X Softening / Shedding Active mass submitted to high mechanical stresses No electrolyte mixing Negative plate sulphation Insufficient recharge X X X X X Stratification / sulphation Low SOC X X X X Sulphation Dissolution / recrystallisation of the lead sulphates High SOC X X Corrosion Oxygen release High temperature X X X Corrosion Enhanced kinetic / Higher oxygen generation IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 31

Table 5 (concluded) Conditions Sensitivity of each battery type Preferential degradation mode Low temperature X X X Enhanced sulphation Reduced available High charge current Low charge current X = sensitive to capacity X X X Drying out Shedding Explanation Lower lead sulphate solubility Reduced diffusion rate High gassing Mechanical disconnection of active mass particles by gas bubbles X X X X Sulphation Crystal growth is favoured in comparison with germination 6 Efficiency test procedures This part is dedicated to battery efficiency evaluation. This procedure evaluates the faradic (Ah) and energy efficiencies for a battery. The ampere-hour efficiency is the ratio of the ampere-hours restored to the ampere-hours received; it indicates the energy or number of ampere-hours required to bring the discharged battery to a given state of charge at a given temperature. Two procedures are presented in this section: the efficiency test procedure available in the NF C 58510 Standard and the procedure provided by the IEC 62093 Draft Standard. They mainly differ in the capacity range of the cycles, as explained below. This is due to the fact that the IEC 62093 Draft Standard is aimed at avoiding the gassing phenomenon. 6.1 NF C 58510 Efficiency test procedure 6.1.1 Description This procedure evaluates the faradic and energetic efficiency for a state of charge between 0 and 75 % of the initial capacity at C 100. After an initial discharge, the retained value is that obtained after one charge/discharge cycle (the end of recharge is determined by amperehour counting). The efficiency is calculated as the ratio of the restored energy during the cycle to that supplied during the same cycle. The test is conducted at 25 and 40 C and is completed within 200 hours. The test procedure is described in Figure 31. IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 32

6.1.2 Comments Figure 31 - NF C 58510 efficiency test procedure With this method, the battery is cycling between 0 and 75 % of state of charge. Around 75 % of SOC, part of the current supplied is not used to recharge the battery but leads to the hydrolysis of the water and hydrogen and oxygen release. The value of the battery efficiency can then be distorted. Moreover, calculating the efficiency on the basis of only one cycle can give inaccurate results. 6.2 IEC 62093 Draft Efficiency test procedure 6.2.1 Description This procedure [9] evaluates the faradic and energy efficiency for a state of charge between 0 and 50 % of the initial capacity at C 10. The retained values are those obtained after a few charge/discharge cycles (the end of recharge is determined by ampere-hour counting) considering that a new battery is then close to its optimal formation while avoiding the effects due to stratification. This is because with these types of cycling conditions, the effects of stratification appear around the 5 th cycle and strongly distorts the results beyond this 5 th cycle. The efficiency is calculated as the ratio of restored energy during a cycle to that supplied during the same cycle. If the efficiency values are stable after four cycles, then an average value is calculated from the efficiency values at the 3 rd and 4 th cycle. If the efficiency values are not stable after four cycles, additional cycles will be needed until two consecutive values are constant to within ±3 %. The test is conducted at 25 C and is completed within 56 hours. The test procedure is described in Figure 32. IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 33

6.2.2 Results Figure 32 - IEC 62093 Draft Efficiency Test Procedure Figure 33 shows the best and the worst efficiencies for three types of battery: tubular plate, flat plate solar and VRLA batteries. Sixteen batteries have been tested: 6 tubular plate batteries, 8 flat plate solar battery and 2 VRLA batteries. Energy efficiency Faradic efficiency 1.05 1.05 1 1 0.95 0.95 0.9 0.9 0.85 0.85 0.8 0.8 Flat plate solar batteries VRLA batteries Tubular batteries Flat plate solar batteries VRLA batteries Tubular batteries Figure 33 - Energy and faradic efficiencies (IEC 62093 Standard) IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 34

6.2.3 Conclusion Noticeable differences (around 10 %) appear between batteries of the same technology for either faradic or energy efficiencies. Also tubular batteries have lower (2 and 4 % respectively) faradic or energy efficiency than flat plate solar batteries. 6.3 Comparison The IEC draft standard seems to offer four advantages compared to the NF standard: o The test duration is shorter (56 hours compared to more than 200 hours); o The result of the efficiency determination is more precise as it is calculated on the basis of at least three cycles; o It avoids the gassing zone where the current is not entirely used for the battery charge, which may distort the result; o It is more important to know the efficiency of a battery at low states of charge when there is a real need to charge the battery. 7 Other procedures This section presents some information about battery test procedures developed by other organizations. 7.1 Cycling test in the Sandia National Laboratory 7.1.1 Objectives According to previous indoor testing, the Sandia National Laboratory identified three main points that need to be considered in a battery test procedure [10]: o The battery recharge strongly depends on the regulation voltage and the charge and discharge rates, o The battery charge acceptance degrades in cycling with low depth of discharge, o The regulation voltage depends on the battery temperature. In order to identify the effects of these parameters on the battery cycle-life, a test procedure has been developed. 7.1.2 Description An initial battery charge and capacity test are performed at the beginning of the procedure. Then cycles that are similar to daily charge/discharge cycles of a PV system are defined as follows: o 25 shallow cycles (DOD of 20 %), o 4 deficit charge cycles (DOD of 60 %), o 2 deficit charge cycles to 11.4 V (low voltage disconnect threshold), o Recovery charge cycles, o Sustaining shallow cycles (DOD of 20 %) The first and second phases abort if the capacity loss reaches 20 %. The other test conditions are: o The battery bath temperature is 25 C, o The total number of cycles per test sequence is 91, o The test is terminated when the capacity loss reaches 20 % of the initial capacity (this capacity loss is measured by an end of load voltage EODV ). The evolution of the battery capacity during the test sequence is presented in Figure 34: IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 35

7.2 Capacity test 7.2.1 Objectives Figure 34 - PV battery cycle-life test sequence in Sandia Laboratory A particular procedure has been designed [11] to show the difference between the available capacity for the user and the real state of health of a VRLA battery. 7.2.2 Description The procedure is made of a first discharge after a typical solar charge, and a second discharge after a rather intensive charge regime. This intensive charge is in fact composed of: o A I 10 discharge, o A battery charge, called IUI charge, that ends when 112 % of the nominal capacity has been recharged and is composed of three phases: Charge at I 10 ; Voltage maintained at 2.35 Vpc; Charge at 0.08 * I 10, o Another discharge at I 10, o An IU charge (charge at I 10, voltage maintained at 2.35 Vpc) to bring back the battery into operation. Figure 35 shows the voltage profile of this test sequence: IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 36

7.2.3 Results and consequences Figure 35 - Capacity test cycle for VRLA batteries Some results show there is a difference of up to 20 % between the two discharges. It outlines the fact that the user has no access to the complete capacity of the battery because a complete recharge is rarely made in PV systems. IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 37

Conclusion This report presents typical battery behaviour caused by seven frequently used cycling test procedures. From this brief survey, the main features of a candidate battery for a given stand-alone photovoltaic application are: o Its efficiency, which characterises the ability of the battery to be recharged with a minimum of energy (this is important for the overall efficiency of the PV system), o Its capacity reduction due to the daily cycles of operation, which could degrade the quality of the rendered service to the user, o Its lifetime, which will deeply impact on the system life cycle cost, o The temperature of the battery during operation in the field. In terms of battery technology, some results confirmed the capability of tubular batteries to provide a longer and better service, but a deeper analysis showed that: o There is a great dispersion of efficiency and lifetime values for various batteries of the same technology, o The tubular battery behaviour in general is not very dependent on the test procedure as this kind of battery is not very sensitive to softening and corrosion; it is more sensitive to sulphation, o All types of degradation (corrosion, softening, sulphation and stratification) have much more effect on flat plate solar batteries, and then the behaviour of the battery is much more dependent on the test procedure. Concerning the ability of the test procedures to meet their objectives by emphasising battery degradation effects, it appears that: o Most of the cycling procedures lead to significant battery sulphation, o Almost none of them accelerate battery corrosion (it may be necessary to develop a procedure to evaluate corrosion because this phenomenon is quite significant in lighting kits and SHS), o There is no test procedure where the four mechanisms lead to a high level of overall degradation (the test procedures were designed to simulate field conditions and not especially to accelerate specific types of degradation). Finally, the information delivered by this report meets the objectives listed at the beginning of the document: o It provides data and recommendations about the most common battery cycling test procedures that are used by project managers to evaluate the expected behaviour for a given application and can help them to select the required procedure, o It may focus the activities of battery testing laboratories on fewer selected procedures: a standard procedure such as the IEC or NFC standard is still representative of real operating conditions and so constitutes a reasonable integrated cycling test; for a specific use of a battery, the five other test methods can be complementary to each other, o It presents only a few laboratory test results about capacity reduction, efficiency measurement and cycling profiles. Therefore there is a strong need for a deeper international collaboration on this matter. It could be done by: Round Robin tests to focus on a limited number of cycling procedures and search for other means to accelerate the tests (for instance, the impact of test temperature must be taken into account); Modelling studies that can enable to reduce the test durations by a better understanding of the chemical phenomenon; Standardisation of these cycling tests. IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 38

The study has shown that it is possible to select a cycling test procedure, or a set of procedures (that are complementary from the degradation point of view), in order to choose the best battery. This can be done according to the various categories of PV systems and the future operating conditions, whilst minimising the test duration (for instance, with the NF C 58510 Standard, the test of a battery of 5 year lifetime can last only 6 months). Nevertheless these tests are still time-consuming, and other methods should be developed to complete the range of procedures presented. IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 39

References [1] F. Mattera and H. Colin, Management of batteries used in stand alone PV power supply systems, report IEA PVPS T3-10: 2002 to be published [2] E. Potteau, D. Desmettre, F. Mattera, O. Bach, J-L Martin, P. Malbranche, Results and comparison of seven accelerated cycling test procedures for the photovoltaic application, LABAT conference, Sofia 2002 [3] IEC 61427 Standard, Edition 1: Secondary cells and batteries for solar photovoltaic energy systems General requirements and methods of test. [4] NF C 58-510 standard: Lead-acid secondary batteries for storing photovoltaically generated electrical energy. [5] D. Desmettre, JL. Martin, Essais de batteries à usage photovoltaïque, note finale de la convention CEA/ADEME 94-95, NT DER/SCC/LVT-GENEC/95-022 [6] D. Desmettre, F Mattera, P. Malbranche, S. Métais, publishable report of the QUALIBAT project EU JOR3-CT97-0161 (1999). [7] D. Desmettre, JL. Martin, Essais de vieillissement accéléré des éléments d accumulateurs électrochimiques à usage photovoltaïque, NT DER/SCC/LVT- GENEC/93-014 and 94-016 [8] D. Desmettre, Validation d un protocole d essai pour batteries en usage dans les kits photovoltaïques, NT DER/SSAE/LVT-GENEC/2000.002 [9] IEC 62093 draft standard: Balance of system components for photovoltaic systems _ design qualification natural environments. [10] T. Hund, Battery testing for photovoltaic applications, Sandia National Laboratory, NM 87185-0753 [11] R. Wagner and D.U. Sauer, Charge strategies for VRLA batteries in solar power applications, Journal of Power Sources 95 (2001) 141-152 IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 40

Annex A: degradation mechanisms Four main degradation mechanisms can occur in solar battery in operating conditions: o Sulphation, o Stratification of the electrolyte, o Softening of the active mass, o Corrosion of the grid. They are briefly described: Softening When a battery is submitted to successive cycles of charge/discharge, repeated volume variations of the active mass create some changes in its morphology such as a modification of the cohesion, the distribution of lead oxides (PbO 2 ) and the size of crystals. These changes lead to a loose connection between the aggregates of lead dioxide and a softening of the active mass. The first consequence is a loss of capacity caused by a lower participation of this active mass in the electrochemical reactions. The extreme situation is obtained when there is no more connection between the aggregates and they fall to the bottom of the battery container (see Figure A1). This process is called shedding. Corrosion Figure A1 - Shedding of the active mass (positive flat plate) When a lead-acid battery is in a high voltage condition (end of charge or overcharge), the oxygen produced at the positive plate leads to the formation of an oxide layer at the interface between the current collector, or grid, and the active mass (see Figure A2). A similar corrosion layer is formed when the battery is left in open circuit conditions for a long time. The corrosion layer is resistive which affects the current collection by the grid. There are many consequences such as: o o o o o o Decrease of electronic exchanges (the layer forms a barrier to ionic diffusion), Increase of the internal resistance, Decrease of the charge acceptance, Decrease of the battery capacity, Mechanical consequences, Grids becoming thin and fragile. IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 41

Grid Corrosion layer Sulphation Figure A2 - Corrosion layer (positive collector of a flat plate battery) The so-called irreversible sulphation phenomenon is the presence of non-rechargeable lead sulphate crystals in positive and negative active masses. Lead sulphate is formed during the discharge phase of the battery. When the battery remains at low states of charge, a process of re-crystallisation of the lead sulphate crystals happens and affects their characteristics: PbSO 4 crystals at the positive and negative plates become bigger (see Figures A3 and A4) and are less connected with the active mass; thus they are no more rechargeable. This leads to a loss of capacity. Figure A3 - PbSO 4 small crystals Figure A4 - PbSO 4 big crystals (Scanning Electron Microscope (SEM) microphotography of charged positive plates) Stratification The stratification of the electrolyte is the presence of a vertical gradient of the concentration of sulphuric acid due to the operation of the battery. Pure sulphuric acid formed during the charge phase has a higher density than the bulk electrolyte and tends to fall down to the bottom part of the battery. This phenomenon is favoured by deep discharges and recharges, but decreases during overcharge as the electrolyte is mixed by oxygen and hydrogen gas bubbles. The stratification also depends on the characteristics of the battery such as: the grid alloy (batteries with lead-antimony grids are less sensitive to stratification); the plate IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 42

geometry (stratification is favoured with the height of the plates), and the compression level of the plates stack (separators slow the descent of acid to the bottom of the battery). The stratification of the electrolyte leads to a loss of capacity (bottom part of the battery less recharged) and to the forming of irreversible lead sulphate crystals (see Figure A5). The stratification phenomenon does not occur, or only at a low level 3, in VRLA batteries because the electrolyte is immobilised in Absorptive Glass Material (AGM) separators or in a gel. Figure A5 - Irreversible lead sulphate crystals due to stratification (positive plates) This photography is obtained through the use of a radioactive tracer that enables to display the distribution of irreversible lead sulphates at the interface active mass/electrolyte: the lead sulphate crystal density increases from white to red, showing here a clear difference of concentration between the lower and upper parts of the plates. This methodology has been developed by Genec. 3 In some AGM batteries the electrolyte stratification may be significant and manufacturers recommend placing the batteries horizontally to avoid this phenomenon. IEA PVPS Task 3 - Testing of batteries used in stand alone PV power supply systems 43