ABSOLYTE TECHNOLOGY

PROFESSIONAL PAPERS

Table of Contents
Development of Totally Maintenance-Free Lead-Acid Battery forTelecommunications Standby Power Abstract Introduction Battery Design and Construction Float Charge Characteristics Deep Discharge Cyclic Performance Partial State-of-Charge Operation Conclusion Acknowledgments References Operational Characteristics of a Sealed Gas-Recombinant Lead-Acid Battery -An UpdateAbstract Introduction Deep Discharge Battery Operations Flat Discharge Cell Reversal Float Life Projections Gas Evolution and Water Loss Summary and Conclusion Acknowledgement The Maturing of a Valve Regulated (VRLA) Battery Technology: Ten Years of Experience Abstract Introduction Commercial Product History Absolyte II Absolyte IIP VRLA Battery Life Manufacturing Improvements VRLA Battery Testing Summary and Conclusions Acknowledgement References Product Design and Manufacturing Process Considerations for the Application of a 10 Year Design Valve Regulated Lead Acid Battery in the Outside Plant Environment Abstract Background Volumetric Energy Efficiency Thermal Management Performance and Reliability Conclusion 1.1 1.1 1.1 1.1 1.3 1.3 1.4 1.5 1.6 1.6 2.1 2.1 2.1 2.2 2.3 2.4 2.4 2.5 2.6 2.6 3.1 3.1 3.1 3.2 3.3 3.3 3.4 3.5 3.7 3.7 3.7 3.7

4.1 4.1 4.1 4.2 4.3 4.5 4.9

Real World Effects on VRLA Batteries in Float Applications Abstract Introduction Purpose of Float Voltage Response to Charging Current Optimum Float Voltage Optimum Temperature Combined Effects Thermal Stability Summary A Discussion About Water Loss, Compression and the VRLA Cell Abstract Introduction Water Loss Model Positive Grid Corrosion Vapor Transmission VRLA Gassing Study Results - Test 1 Results - Test 2 Saturation Effects Study Results Model vs. Test Conclusions Separator Compression Effects Conclusions Acknowledgements Operational Characteristics of VRLA Batteries Configured in Parallel Strings Introduction Theory Normal Parallel Configuration Parallel Operation of Different Capacity VRLA Batteries Parallel Operation of Batteries At Differing States of Charge Parallel Operation at Higher Discharge Currents Multiple Parallel Strings Constant Power Discharge Loads Testing Cable Configurations in Parallel Battery Systems Conclusion A Guideline for the Interpretation of Battery Diagnostic Readings in the Real World Abstract Introduction Baseline Values Impedance as a function of State of Charge Impedance as a function of accelerated float life Impedance as a function of Cell Dryout Impedance as a function of Loss of Compression Conclusions & Guidlines for using impedance as a diagnostic tool

5.1 5.1 5.1 5.1 5.2 5.3 5.4 5.5 5.6 5.7 6.1 6.1 6.1 6.2 6.2 6.3 6.4 6.4 6.5 6.6 6.6 6.7 6.7 6.9 6.9 7.1 7.1 7.2 7.2 7.3 7.4 7.5 7.6 7.7 7.8 7.8 8.1 8.1 8.1 8.2 8.3 8.4 8.5 8.6 8.8

Intelligent Monitoring System satisfies customer needs for Continuous Monitoring and Assurance on VRLA Batteries Introduction Diagnostic Tools Continuous Monitoring: Physical Parameters Continuous Monitoring: Usefulness of Output Continuous Monitoring: Customer Requirements Satisfied by Intelligent Monitoring System (IMS) Conclusions Acknowledgements References An Examination of High Rate Recharge on Absolyte IIP Batteries Abstract Introduction Test One: Single Module Testing, Multiple Currents Test Two: 48 Volt System Testing Summary Considerations for the Configuration and Arrangement of Valve Regulated Monobloc Batteries in Enclosures Abstract Introduction Test Results Discusion of Results Conclusion Recommendations for Future Work Examination of VRLA Cells Sampled from a Battery Energy Storage System (BESS) after 30-Monts of Operation Abstract Background Changes in Telecommunications Power Metlakatlas BESS Opportunity Sampling Cells from the Metlakatla BESS Electrical Characterization of Sampled Cells Internal Examination of the BESS Cells Summary and Conclusions Acknowledgement References

9.1 9.1 9.1 9.2 9.3 9.6 9.7 9.7 9.7 10.1 10.1 10.1 10.1 10.4 10.6 11.1 11.1 11.1 11.2 11.4 11.5 11.5 12.1 12.1 12.1 12.1 12.2 12.4 12.5 12.6 12.8 12.9 12.9

Considerations for the Configuration and Arrangement of Valve Regulated Monobloc Batteries in Enclosures - Part II 13.1 Abstract 13.1 Introduction 13.1 Testing and Results 13.2 Test Summary 13.3 Tests 1, 2 and 3 (5 mm Spacing) 13.3 Tests 4, 5 and 6 13.4 Tests 7, 8 and 9 13.4 Tsts 1, 2 and 3 (10 mm Spacing) 13.5 Discussion of Results 13.5 Spacing of Blocks 13.6 Ventilation 13.7 String Location within the Cabinet 13.7 Top Terminal Battery Configuration 13.8 Conclusions 13.9 A Naturally Aged VRLA Battery: 18 Years Later 14.1 Abstract 14.1 Introduction 14.1 Initial Test Data: 14.1 OCVs Vent Opening Pressures and IRs Before and After Charge 14.2 Internal Examination 14.2 Electrical Testing 14.3 Internal Resistance and Capacity 14.4 Discussion of Initial Results 14.5 Capacity Recovery: Overcoming Compression Loss 14.5 Comparison of 3 Cells Performance: Compressed and Uncompressed 14.6 Compression: How Much? 14.6 Long Term Float 14.7 Conclusion 14.7 References 14.7

1.1

1.2

1.3

1.4

1.5

1.6

2.1

2.2

2.3

2.4

2.5

2.6

3.1

3.2

3.3

3.4

3.5

3.6

3.7

4.1

4.2

4.3

4.4

4.5

4.6

4.7

4.8

4.9

5.1

5.2

5.3

5.4

5.5

5.6

5.7

6.1

6.2

6.3

6.4

6.5

6.6
Percent Saturation Capacity Trial Number Impedance Saturation Impedance - m Percent Rated Capacity

6.7

6.8

6.9

7.1

7.2

7.3

7.4

7.5

7.6

7.7

7.8

8.1

8.2

8.3

8.4

8.5

8.6

8.7

8.8

8.9

9.1

9.2

9.3

9.4

9.5

9.6

9.7

AN EXAMINATION OF HIGH RATE RECHARGE ON ABSOLYTE IIP BATTERIES

Bruce A. Cole Director of Marketing GNB Technologies Lombard, Illinois U.S.A. ABSTRACT: Power needs in the telecommunication industry are presently undergoing rapid change. The coexistence of digital and analog systems, future build-out strategies and ever increasing data and Internet traffic throughout a network makes power planning a difficult task at best. As a reaction to this uncertainty, many in the industry are oversizing power panels today so they will meet their anticipated needs tomorrow. In such cases, after an outage and when the power plant comes back on-line, large amounts of current are available to the battery. Combined with this reality is a desire by users to rapidly bring their batteries to a high state of charge after an outage or a test. Both of these situations lead the user to value a battery which can accept high recharge currents without sustaining damage. GNB Technologies has conducted a series of tests to examine Absolyte IIP battery charge acceptance when the available current is essentially unlimited. The testing also endeavors to evaluate harmful effects on the battery such charging could have by examining heating effects, water loss, and capacity. INTRODUCTION: In the telecommunication industrys early years, power needs were substantial. The relays and tubes in the state of the art switching stations of the time required large amounts of electricity. As electro-mechanical switches were replaced by semiconductors, these devices became much more power efficient and power needs decreased. As evidence of this decrease, not so long ago, the power rooms built decades before went largely unused. Of course, that was then. This trend toward smaller power requirements has been reversed. Today digital systems overlay existing analog ones. Wireline that always existed now sits in parallel with wireless, broad band data for Internet resulting in large fiber network build-outs, cable TV, etc. The applications multiply. They all need power. Providers of these services are savvy but planning power requirements even a couple of years into the future under these circumstances is challenging. One reaction to the uncertainty is to oversize power panels relative to the battery in the expectation that the site will grow into the load. This strategy results in systems where the battery can see much larger inrush currents after a power outage than previous designs. On the other hand, battery manufacturers generally recommend that the charge current be limited to 18 to 30 amps per hundred ampere-hours of the batterys 8-hour capacity. These competing interests can cause uncertainty for the power planner. GNB Technologies has conducted a series of tests to characterize the behavior of its Absolyte battery subjected to conditions that would simulate high recharge currents. The testing also attempted to determine if the battery was harmed in any way by these large currents. TEST ONE: SINGLE MODULE TESTING, MULTIPLE CURRENTS The first test was designed to characterize the voltage and current acceptance behavior of an Absolyte IIP module. The battery tested was a single, series-connected, six-cell 264-Ah 50A11 module. The charge voltage was always 2.35 VPC although a number of different current limits were imposed. The testing also attempted to determine if the cells were sustaining damage from the high recharge currents. This was done by monitoring weight loss that would signal a decrease in saturation (i.e. water loss), impedance, and most importantly, capacity throughout the testing. In addition, temperature stability (i.e. signs of thermal runaway) and external physical appearance were observed. After discharging the 12 volt module 100% at its 8-hour rate, the battery was recharged at 2.35 VPC and with a current limit of 50, 100 and 150 amperes per 100 ampere-hours of capacity at the 8-hour rate respectively. Also examined were the behaviors at 12, 24, 36 and 72 A/100-Ah. A total of 9 cycles were put on the battery with six of them at 50 amps per 100 A-h or higher. A typical recharge curve at 150 A/100-Ah of available current follows: Robert J. Schmitt Technical Marketing Engineer

10.1

Recharge Behavior at 2.35 VPC 150 A/100-Ah 120 100 Current - A/100-Ah 80 60 40
Current % Recharge

120 100 80 60 40 20 0 0 1 2 3 4 5 6 7 8 9 10 11 12 Time - hours Percent Recharge

20 0

As expected, the period of time that the battery accepts the highest charge current is brief. Charge efficiency is very high during this time and the state of charge rises rapidly. As the battery become full, its charge acceptance decreases and its current asymptotically approaches float level for the given voltage. It is interesting to note, that although 150 A/100-Ah was available, this battery only accepted 109 A/100-Ah. The data indicates that the cells internal resistance limited the amount of current that it could accept. In terms of the amount of time that it takes to recharge a battery at higher vs. lower available charge currents, the difference is significant. For example from the data below, 80% recharge at 12 A/100-Ah available current takes about 7.3 hoursnearly twice the amount time it takes with 24 A/100-Ah available (3.8 hours), and over four times that with 72 A/100-Ah (1.8 hours). If rapidly returning the battery to a high state of charge is desired, higher available charge current is clearly beneficial.

Recharge Time with 12, 24 and 72 A/100 A-h Available Charge Current 120% Percent Recharge 100% 80% 60% 40% 20% 0% 0 1 2 3 4 5 6 7 8 Time - hours 12 A/100-Ah
72 A/100-Ah 24 A/100-Ah

10.2

During these tests indicators of battery health did not suggest that the battery was harmed by the high in-rush charge currents. Impedance at the finish of testing decreased from initially recorded values an average of 5%. As these were brand new cells from the plant, it was expected that their relative cell saturation would be high and some water loss would be normal. Indeed after some initial loss, the saturation decrease stopped. The higher and lower numbers from trial to trial suggest the changes were so small that our equipment could not accurately measure the minute amounts of water lost. Impedance and saturation change data is presented below. Lastly and most importantly, capacities at the 8-hour rate, before and after testing increased from 100% to 106%.

Cumulative Saturation Change Current Trial Saturation1 50 A/100-Ah Saturation2 100 A/100-Ah Saturation3 150 A/100-Ah Saturation4 50 A/100-Ah Saturation5 150 A/100-Ah

Cell 1 -0.10% -0.17% -0.17% -0.13% -0.17%

Cell 2 -0.20% -0.30% -0.34% -0.30% -0.37%

Cell 3 -0.24% -0.37% -0.37% -0.34%

Cell 4 -0.24% -0.30% -0.27% -0.27% -0.24%

Cell 5 -0.13% -0.27% -0.13% -0.10% -0.13%

Cell 6 -0.10% -0.17% -0.13% -0.17% -0.13%

Impedance Before and After Testing 0.46 0.45 Impedance - milli-ohms 0.44 0.43 0.42 0.41 0.40 0.39 0.38 Cell 1 Cell 2 Cell 3 Cell 4 Cell Number Cell 5 Cell 6 Before Testing After Testing

Examination of thermal effects is examined in more detail in the next test but it was desired to compare the temperature rise at the various available current levels during the inherently exothermic recharge process. The results presented below are hardly shocking but nonetheless informative: Lower current levels resulted in less heating. It can be surmised that the 100 and 150 A/100-Ah rises are indistinguishable because as noted earlier, the latter did not take all of the current that was available, taking not quite 10% more than the former. The data demonstrated that the heat rise was a transient effect and that the maximum 9C temperature rise was gone in about 12 hours.

10.3

Temperature Rise at Various Charge Currents 10

100 & 150 A/100-Ah

9 8 Temperature Rise - C 7 6 5 4
50A/100-Ah

3 2 1 0 0 2 4 6 8 10 12 14 Time - hours
TEST TWO: 48 VOLT SYTEM TESTING

24 A/100-Ah

Where the first test concentrated on charge characterization at GNBs recommended equalize voltage, the second test sought to compare this charge behavior to that at the batterys recommended float voltage. It also would examine the heating effects of high charge on an entire four-module, 48 volt system, rather than just a single module. For this test, a 48 volt 50A11 264Ah battery was again discharged to 100% depth of discharge at its 8-hour to 1.75 VPC rate. The battery arrangement consisted of six cells in a module, four modules stacked to create the system. One thermocouple was placed into each of the four modules and two thermocouples monitored ambient temperature. Recharge current was limited to 100 A/100-Ah and occurred at 2.25 VPC and 2.35 VPC. The following graph compares the rate of recharge at the two voltages.

Recharge at 2.25 and 2.35 VPC; 100 A/100 A-h Current Limit

120
2.35 VPC

120% 100%
Recharge Percent 2.25 VPC

Charge Current - A/100-Ah

100 80 60 40 20
Current

80% 60% 40% 20% 0%

0 0 1 2 3 4 5 6 7 8 9 10 Time - hours

10.4

Percent Recharge

Again nothing terribly surprising happened here. As expected, the battery at higher charge voltage recharged more rapidly. Quantitatively, how much faster the battery recharged is of interest. For example, 80% recharge occurred in less than half the time at 2.35 vs. 2.25 VPC (1.6 vs. 4.0 hours) and 90% recharge is more than 3 times faster (2.3 vs. 7.8 hours). At 2.35 VPC, the charge current merely kissed the current limit line before receding. In all, the battery only dwelled for about 3 minutes at the current limit or within 5% of it. Contrasting , at 2.25 VPC, the battery never even approached the current limit, reaching a maximum of 61 A/100-Ah. As previously concluded, the batterys internal resistance effectively self-limits the charge current even if more current is available. As noted above, the test battery consisted of a single stack of four modules with a thermocouple placed into each module and two more monitoring ambient temperature. The difference between ambient and maximum battery temperature is depicted below. All testing occurred at a room temperature of approximately 25C.
T em p eratu re R ise D ue to C harging at 2.25 and 2.35 V P C 120 Charge Current - A/100 A-h 100 T (2.35 V P C ) 80 60 40 T (2.25 V P C ) 20 0 0 1 2 3 4 5 6 7 8 9 10 T im e - h ours 2 0 8 6 4 12 10 Temperature Rise - C

This portion of the testing demonstrated the intuitive notion that both temperature rise and charge acceptance is higher at a higher charge voltage. Defined as in-rush current admitted by the battery, charge acceptance at 2.25 VPC float was shown to be only 60% of what it is at 2.35 VPC while the temperature rise for the former was half that of the latter. In any event, the transient 10C rise from ambient is not of significant concern. Within 10 hours and while still on charge at 2.35 VPC, the battery temperature had dropped to within 4oC of ambient and was decreasing at a rate of approximately 0.9oC/hour. Temperature increases on the order of 8-10oC during 2.35 VPC recharges are typical. Having said this, it should again be emphasized that this testing occurred at approximately 25C. More caution must be applied to a high rate charge regime when the ambient temperature is already elevated, in order to avoid the point where the heat generated exceeds the batterys ability to dissipate it. Additional studies should be conducted to evaluate the impact of initially elevated ambient conditions. Mapping the location of the maximum temperature attained shows that the highest values occurred in interior modules, not at the top or bottom.

Maximum Attained Temperatures (C) 2.35 VPC 2.25 VPC module Top 33.5 28.0 1 34.0 28.5 2 34.5 29.0 3 Bottom 32.5 28.0 4

10.5

It is important to note that as in the initial test using only a single module, these cells suffered no capacity or impedance degradation as a result of exposure to charge conditions of 2.35 VPC and currents up to 100 A/100-Ah. The final capacity of this battery system was in excess of 104% at the 8 hour rate. SUMMARY The two tests together reveal some interesting things about how an Absolyte IIP battery recharges from a worst case 100% depth of discharge from a controlled temperature baseline. A maximum in-rush charge current acceptance at a given voltage was limited by the battery. A discharged battery will not accept infinite amounts of current. Additional testing on larger capacity cells would be interesting but for the physical limitations (namely charge capacity). Both the current limit and charge voltage strongly affected recharge time. High rate recharge, up to 150 A/100-Ah, resulted in an acceptable transient temperature rise. The battery health indicators monitored in this test demonstrated that the GNB Absolyte IIP was not harmed by a high rate recharge practice. These indicators included impedance, water loss and capacity.

10.6

Considerations for the Configuration and Arrangement of Valve Regulated Monobloc Batteries in Enclosures
Director of Marketing, Bruce, Cole, GNB Technologies, USA Marketing Manager, Mark A. Jesko, GNB Technologies, USA Director of Value Engineering, Joe Szymborski, GNB Technologies, USA

Abstract
In todays rapidly changing and increasingly competitive world of telecommunications power, there is a continuous requirement for a reduction in costs while achieving increases in overall system power density. Much of this is being driven by the demand that has been created from the internet/data market explosion. The combination of these conditions has led to the design of increasingly smaller battery enclosures that are utilized in both wireless and outside plant fixed wired applications. Almost exclusively, these enclosures utilize valve regulated lead acid cells and batteries (VRLA), and frequently there are no temperature controls for the battery compartments. The magnitude of the installed base of cabinets and batteries, and the impact of uncontrolled environments, has created concern about the actual life of these batteries and the safety of the systems. As a result there have been a number of studies done reporting on the effect of high temperatures on lead acid batteries, and thermal models developed to characterize how ambient conditions impact the actual temperature performance of the batteries themselves. It is the goal of this paper to present the data and results of a series of tests run to characterize the thermal behavior of typical 10 year front terminal monobloc batteries in a standard cabinet enclosure under recharge conditions. The variables examined include; battery spacing, rack vs. cabinet systems (i.e. enclosed vs. non enclosed), and the effects of forced air flow. In addition the paper will consider current industry standards on these issues and explore recommendations for future studies.

Introduction

Battery and power system designers for current and future telecommunications networks are faced with multiple choices. These choices such as power density, system life and safety, and cost are often competing. Understanding the impact of their selections is often difficult due to the variety of conditions to which each application is exposed, and a lack of sufficient data to describe the outcomes of their decisions. This is particularly the case with the batteries that are deployed in these systems. A clean synthesis relating VRLA behavior to these choices of manufacturer, design and operating conditions does not exist, making good decisions difficult. It is the purpose of this paper to examine one test scenario and present the results in a format that leads to a useful conclusion. The study conducted, involved examining the behavior of sixteen, Marathon 100 Ah front terminal batteries, assembled into one cabinet system. This type of front terminal battery was se-

lected due its growing popularity with system designers. This popularity is a result of its favorable footprint, and easy access to the terminations. In addition, the packing or spacing of these types of batteries into standard 19 or 23 rack/cabinet widths is also of concern to designers. Currently the IEC 896-2 standard for valve regulated lead acid batteries recommends a minimum gap of 5-10mm between batteries when housed in a rack or cabinet. Due to laboratory equipment constraints the batteries were configured as two separate strings with 8 blocks per string, thus yielding a 96 volt 100 Ah string or a total cabinet system of 96 volts and 200 Ahs of capacity. The heat generation from this arrangement would be identical to that for a configuration consisting of four parallel strings, 100 Ahs each, at 48 volts. All testing was conducted in a laboratory environment where the ambient conditions were maintained at 25 +/- 2oC. Figure 1 below is a photograph of the basic test set up.

11.1

4) and thermocouple location (C = center of string, O = outside of string). The term cabinet, or the abbreviation cabnt, refer to the ambient temperature inside the cabinet.

Test Results

As expected, battery temperature variation was evident in and dependent on the system configurations tested. A test by test summary of the results is presented below. Fig. 1 Test Set-up A total of nine tests were conducted on the system. The purpose of each test was to study the thermal conditions inside the cabinet, as a function of battery spacing and ventilation, during a recharge at 2.35 VPC with a current limit of 35A/100 Ah. There were three battery spacing arrangements considered; 10mm, 5mm, and 0mm (i.e. batteries in contact) and three ventilation types; open ventilation (essentially a four shelf rack), natural convection only (small openings in the side of the closed cabinet), and forced ventilation (a fan in the top of the cabinet). Each recharge delivered 110% of the Ahs removed, following an 8 hour, 12.5 amp discharge. Note that the battery shelves did not have openings that would permit air flow between the layers. There also was not any rectification equipment located inside the cabinet that would influence the overall thermal behavior. A summary of the testing protocol is shown in Table 1 below. Table 1
Test # 1 2 3 4 5 6 7 8 9 Battery Spacing 10mm 10mm 10mm 5mm 5mm 5mm 0mm 0mm 0mm Ventilation Open cabinet Closed cabinet natural convection Forced ventilation Forced ventilation Closed cabinet natural convection Open cabinet Open cabinet Forced ventilation Closed cabinet natural convection
Temperature (C)

2.1

Test #1 (10mm spacing/open cabinet

The data in Figure 2 displays the battery temperature variation within the open cabinet where the surrounding ambient temperature was maintained at 252C. The shelf located in position 4 at the bottom of the cabinet reached a maximum temperature of 36C. Shelves 1 and 2 located at the top of the cabinet reached a maximum temperature of 41C, and the shelf located in position 3 above shelf 4 and below shelves 1 and 2 reached a maximum temperature of 42C. The battery temperatures at the locations in the center of the strings within the cabinet varied by 57C depending on shelf location and were as much as 17C above the ambient temperature. The battery system was able to return to a steady state temperature after 40 hours.
Figure 2. Open Cabinet 10mm Space/Convection Cooling
50 45 40 35 30 25 20 0 10 20 30 40 50 60 70

Shelf 4 Battery

Shelf 3 Battery Shelf 2 Battery Shelf 1 Battery

Time (Hours)

2.2

Test #2 (10mm Spacing Closed cabinet natural convection cooling

The thermal mapping of the cabinet was done by mounting a total of 9 thermocouples inside, two per string (one in the center and one on the outside) and one for the ambient inside the cabinet. Temperature data was sampled every 5 minutes during charging. The nomenclature used to describe the temperature (T) at the measured locations within the cabinet throughout this paper is related to the battery shelf (1-

The data in figure 3 displays the battery temperature variation within the closed cabinet, natural convection cooling, 10-mm battery spacing and an ambient temperature of 252C. Again the shelf located in position 4 at the bottom was the coolest reaching a maximum temperature of 41C. Shelves 1, 2 and 3 located above shelf 4 reached 43, 42.5C and 43C respectively. The battery temperatures at the locations in the center of the strings within the cabinet varied by only

11.2

1.5-2C depending on shelf location and were as much as 18C above the ambient temperature. The battery system was able to return to a steady state temperature after 60 hours.
Figure 3 Closed Cabinet 10mm Space / Fan Blocked Off
50 45

Peak Temperatures: Tests 1,2 & 3

Table 2. 10 m Spacing Test Peak Tem m peratures 1= Top Shelf 4= BottomShelf T 1C T 1O T 2C 41.0 33.0 41.0 Open Cabinet 43.0 37.5 42.5 Closed Cabinet Cabinet Forced Air 32.0 28.5 37.0 T2O 34.5 38.5 31.5 T3C 42.0 43.0 35.5 T3O 34.0 37.0 30.5 T4C 36.0 41.0 34.0 T4O CABINET 34.0 26.5 36.5 35.0 28.0 25.5

Temperature (C)

40 35 30 25 20 0 10

Shelf 4 Battery Shelf 3 Battery Shelf 2 Battery Shelf 1 Battery

26.5

35.0

25.5

41

33

43

37.5

32

28.5

20

30

40

50

60

70

Time (Hours)
41 34.5 42.5 38.5 37 31.5

42

34

43

37

35.5

30.5

2.3

Test #3 (10mm spacing forced air ventilation)

36

44

41

36.5

34

28

The data in Figure 4 displays the battery temperature variations measured during the final test iteration with 10mm battery spacing within the closed cabinet having forced air-ventilation and an outside ambient temperature of 252C. During this test, shelf 1 located at the top was the coolest reaching a maximum temperature of 32C. Shelves 2, 3 and 4 located below shelf 1 reached 37C, 35.5C and 34C respectively. The battery temperatures at the locations in the center of the strings within the cabinet varied by only 2-5C depending on shelf location and were as much as 9C above the ambient temperature. The battery system was able to return to a steady state temperature after 24 hours. Tabular comparisons of these tests are summarized in Table 2 (and the diagram).
Figure 4. Closed Cabinet 10mm Space / Forced Cooling
50 45

2.4

Tests #4, 5, and 6 (5mm spacing trials)

Table 3 (and the diagram) summarizes the data for the second iteration of experiments which investigated the effects of decreased battery spacing from 10mm to 5mm with the three differing types of ventilation. In the open cabinet experiment the battery temperatures in the center of the strings varied by 0.5-13C with a peak temperature of 42C seen in the top shelf which was 17C above the ambient temperature. The system was able to return to a steady state temperature after 40 hours. In the closed cabinet experiment the battery temperatures in the center of the strings varied by only 0.51.5C with a peak temperature of 43.5C in the top shelf which was 18.5C above the ambient temperature. The system was able to return to a steady state temperature after 48 hours.

Temperature (C)

40 35 30 25 20 0 10 20 30 40 50 60 70

Shelf 4 Battery Shelf 3 Battery Shelf 2 Battery Shelf 1 Battery

Time (Hours)

The final experiment investigated the effects of a 5mm battery spacing in a closed cabinet with forced air ventilation. The battery temperatures in the center of the strings varied by 2.5-3.5C while attaining a peak battery temperature of 41.0C in the two middle shelves, which equated to a rise of 16.0C above the ambient temperature. The system was able to return to a steady state temperature after 30 hours.

11.3

Peak Temperatures: Test 4,5 & 6

Peak Tem peratures 10 m Spacing m 1= TopShelf 4= BottomShelf T1C T1O T2C T2O T3C T3O T4C 41.0 33.0 41.0 34.5 42.0 34.0 36.0 O Rack pen 43.0 37.5 42.5 38.5 43.0 37.0 41.0 Cabinet Cabinet Forced Air 32.0 28.5 37.0 31.5 35.5 30.5 34.0 T4O CABINET 34.0 26.5 36.5 35.0 28.0 25.5

Peak Temperatures: Test 7,8, & 9

Table 4. 0 mmSpacing Test Peak Temperatures 1= Top Shelf 4= BottomShelf T 1C T 1O T 2C T2O T3C T3O T4C 42.0 31.5 41.5 32.5 41.5 32.5 40.5 Open Cabinet 44.5 37.0 44.0 37.5 44.0 37.0 43.0 Closed Cabinet Cabinet Forced Air 41.5 33.0 42.0 33.0 42.0 34.0 41.5 T4O CABINET 27.5 26.0 32.5 33.0 30.5 26.5

26.5

32.0

26.0

42

31.5

43.5

36.5

37.5

30.5

26

33.0

26.5

42

31.5

44.5

37

41.5

33

41.5

33.5

42.5

37

41

33.5
41.5 32.5 44 37.5 42 33

33

40.5

43

36.5

41

33.5
41.5 32.5 44 37 42 34

29

27

42

31.5

40

29.5

40.5

27.5

43

32.5

41.5

30.5

2.5

Tests 7, 8, and 9 (0mm spacing trials)

Discussion of Results

Table 4 (and the diagram) summarizes the data for the third iteration of experiments that investigated the effects of no space between batteries with the three different types of ventilation. In the open cabinet experiment the battery temperatures in the center of the strings varied by only 0.5-1.5C with a peak temperature of 42C again realized in the top string which was 17C above the ambient temperature. The system was able to return to a steady state temperature after 40 hours. In the closed cabinet experiment the battery temperatures in the center of the strings interestingly only varied again by 0.5-1.5C with a peak temperature of 44.5C again in the center of the top string which was 19.5C above the ambient temperature. The system returned to a steady state after 55 hours. In the final experiment investigating the effect of no spacing between batteries with forced air ventilation, the battery temperatures varied by a very minor 00.5C attaining a peak battery temperature of 42.0C again in the center strings which were 17.0C above ambient. The system returned to a steady state after 40 hours.

The results of the testing described in the preceding text can be further examined by breaking out, and looking at, the three main variables being evaluated; spacing of blocks, type of ventilation, and string location within the cabinet.

3.1

Spacing of blocks

Shown below are three tables (5, 6, and 7) illustrating the comparison of temperature behavior as a function of battery spacing with each of the three types of ventilation. Again, the temperatures shown are maximums and the same nomenclature is used as earlier. Shelf 1 is at the top of the cabinet, shelf 4 is at the bottom, C is in the center of the cabinet, O is on the outside of the string, Cbnt is the cabinet ambient, and the stable conditions are after the temperature has returned to steady state. Table 5 - Open cabinet (no sides, only top and bottom)
Stable Tim to e p T1C T1O T2C T2O T3C T3O T4C T4O Cbnt Tem Stable

10m 41C 33C 41C 34.5C 42C 34C 36C 34C 26.5C 25C m 5m 42C 31.5C 41.5C 33.5C 42C 33C 40.5C 29C 27C 25C m 0m 42C 31.5C 41.5C 32.5C 41.5C 32.5C 40.5C 27.5C 26C 25C m

40hrs 40hrs 40hrs

11.4

Table 6 - Closed cabinet no ventilation

T1C T1O T2C T2O T3C T3O T4C T4O Stable Time to Cbnt Temp Stable

3.3

String location within the cabinet

10mm 43C 5mm 0mm

37.5C 42.5C 38.5C 43C 37C

41C 36.5C 35C 27C

60hrs 48hrs 55hrs

43.5C 36.5C 42.5C 37C 44.5C 37C 44C

43C 36.5C 42C 31.5C 32C 24C 43C 32.5C 33C 25C

37.5C 44C 37C

Table 7 - Closed cabinet forced ventilation

Stable Time to T1C T1O T2C T2O T3C T3O T4C T4O Cbnt Temp Stable

10mm32C 28.5C 37C 31.5C 35.5C 30.5C 34C 28C 25.5C 26C 5mm 37.5C 30.5C 41C 33C 41C 33.5C 40C 28.5C 26C 24C 0mm 41.5C 33C 42C 33C 42C 34C 41.5C 30.5C 26.5C 26C

24hrs 30hrs 40hrs

Temperature profiles within the cabinet follow a fairly intuitive pattern. In the tests where no forced ventilation was present one would anticipate that the string on the bottom of the cabinet would be the coolest and temperatures would steadily rise throughout the cabinet with the maximum temperatures seen in the top string. In virtually all tests this was the case. In the tests where forced ventilation was present the top string of batteries, or those closest to the fan, typically ran the coolest.

The data clearly shows that spacing the batteries decreases the maximum temperature at the center of the cabinet only when combined with forced air ventilation. This benefit was the greatest across all four levels of batteries when the spacing was increased from 5mm to 10mm. Here the spacing resulted in a 4-5oC reduction in maximum temperatures in the center of the cabinet. We can see that under conditions without forced ventilation (Tables 5 and 6), spacing distance does not significantly impact the maximum temperature (<2oC) or the time to stable conditions. There is however an anomaly in the outside temperature of the bottom string in the 10mm configuration where the temperature ran approximately 5oC higher and where this entire cabinet configuration took longer to reach a stable temperature. The authors do not have a solid theory to propose as to why this condition occurred other than possibly more efficient heat transfer was realized with the greater battery spacing.

Conclusions

The work that was done in this test yields several main conclusions. 1. Forced convection within a closed cabinet will yield significant benefits by reducing maximum temperature increases above ambient conditions. In this testing up to a 10oC decrease was realized. 2. At least a 10mm spacing between blocks is required to realize the maximum benefits of forced air ventilation. 3. In conditions where forced air ventilation is not present battery spacing plays a limited role in reducing maximum battery temperatures. 4. Under normal recharge profiles battery temperatures internal to a cabinet will see anywhere from a 7oC to 20oC rise above ambient conditions (25+/-2oC). These temperatures will however return to a stable ambient within 60 hours. 5. Battery strings located at the bottom of a cabinet or nearest a fan will remain the coolest during charging.

3.2

Ventilation

Tables 2, 3, and 4 more clearly show the impact of ventilation. Under all spacing conditions a fully closed cabinet with only natural convection and airflow will always result in higher (as much as 7%) peak temperatures than if the cabinet was fully open or if forced air was applied. With regards to forced ventilation, we concluded that unless the battery spacing was at the 10mm level, significant reductions in battery temperatures were not observed. Again, this is verified by examining the data in Tables 5, 6, and 7 above.

Recommendations for Future Work

This testing draws some interesting conclusions but also raises several points for further investigation. Some of these are: 1. Repeat of the work at elevated ambient temperatures and under normal float conditions for extended periods of time. 2. Variations in the initial charge current rate. 3. Location of electronic equipment inside the cabinet to monitor the effect of an additional heat source. 4. The impact of using standard top terminal monoblocs. 5. Impact of using perforated trays that would allow air movement between layers.

11.5

Examination of VRLA Cells Sampled from a Battery Energy Storage System (BESS) after 30-Months of Operation
Joseph Szymborski, George Hunt and Angelo Tsagalis
GNB Technologies Lombard, Illinois USA

Rudolph Jungst
Sandia National Laboratories Albuquerque, New Mexico USA

Abstract: Valve-Regulated Lead-Acid (VRLA) batteries continue to be employed in a wide variety of applications for telecommunications and Uninterruptible Power Supply (UPS). With the rapidly growing penetration of Internet services, the requirements for standby power systems appear to be changing. For example, at last years INTELEC, high voltage standby power systems up to 300-vdc were discussed as alternatives to the traditional 48-volt power plant. At the same time, battery reliability and the sensitivity of VRLAs to charging conditions (e.g., in-rush current, float voltage and temperature), continue to be argued extensively. Charge regimes which provide offline charging or intermittent charge to the battery have been proposed. Some of these techniques go against the widely accepted rules of operation for batteries to achieve optimum lifetime. Experience in the telecom industry with high voltage systems and these charging scenarios is limited. However, GNB has several years of experience in the installation and operation of large VRLA battery systems that embody many of the power management philosophies being proposed. Early results show that positive grid corrosion is not accelerated and battery performance is mantained even when the battery is operated at a partial state-of-charge for long periods of time.

Recently, after nearly three years of continuous operation, several cells were sampled at random from the battery and examined. When these samples were taken, the batterys monitoring system indicated that the battery had been maintained at about 75-85% state-of-charge over the entire time period and had received only three equalization charges. This paper will review electrical testing conducted on the battery samples as well as the results of extensive physical analyses performed on the battery materials and components. 1.1 Changes in Telecommunications Power

Power strategies within the telecommunications industry are changing. As suppliers broaden their product offerings to include Internet service, cable TV and other communications media, their power needs are moving away from the traditional 48-volt dc power plant. More of the equipment used to provide these new services is powered at utility provided voltages. As a result, service providers are installing high voltage, ac power plants that, more and more, resemble Uninterruptible Power Supplies (UPSs). For example, at last years INTELEC, papers were presented discussing high voltage power systems, up to 300volts, as alternatives to the traditional 48-volt power plant [24]. Rather than operate two power plants, both a low voltage and a high voltage system, operators are considering converting all of their equipment to operate at the higher voltage levels. Experience in the telecom industry with high voltage power plants is limited however, and there is concern about understanding the operational considerations associated with these high voltage systems. Valve-Regulated Lead-Acid (VRLA) batteries continue to be the dominant battery technology that many of the telecommunications suppliers are installing into their new and expanding power sites. In an attempt to improve VRLA battery reliability, to extend VRLA battery lifetime, and to overcome the sensitivity of VRLAs to charging conditions, alternative charging regimes to the traditional float operation of batteries are being proposed. Off-line charging

1.

Background

In 1996-97, GNB installed and commissioned a large VRLA battery system for a Battery Energy Storage System (BESS) at the island village of Metlakatla, Alaska [1]. The batterys function is to stabilize the island communitys power grid providing instantaneous power into the grid when demand from a local sawmill is high, and absorbing excess power from the grid to allow its hydroelectric generating units to operate under steady-state conditions. This nominal 756-volt system is capable of providing 1.4-MWh at about the batterys 90-minute discharge rate. Because the battery is required to randomly accept power as well as to deliver power on demand to the utility grid, it was decided to continuously operate the battery at between 70 and 90% state-of-charge. Like some of the recently proposed alternative charge regimes for telecommunications installations, this batterys operations raised several concerns regarding long-term performance and life.

12.1

[5, 8] and intermittent charging [6] of the battery, for example, have been suggested.

In addition, some telecommunications providers are looking for ways to better utilize their facilities and equipment, and to reduce their operating costs. At least one major telecommunications company is considering using its standby battery banks for load-leveling and load-sharing during periods when utility rates are high. These new battery operating options all appear inviting; however, experience with operating batteries under these conditions in the telecommunications industry is, like operating high voltage power plants, also limited. 1.2 Battery Energy Storage Systems (BESS)

alone island electric utility, operates the BESS to supplement its generating facilities. The primary generation source for the utility is 4.9-MW of rain-fed hydroelectric capability. In addition to its residential customers, MP&L supplies power to a commercial cannery and cold storage facility, and to the Annette Hemlock Mill, a commercial lumber mill operation. The BESS is used to provide instantaneous power to the utility system to satisfy the random instantaneous load demands of a log chipper at the mill without causing brownouts or overvoltage conditions to the remainder of the utilitys customers. In an attempt to solve their power quality problems, MP&L 15 years ago installed a $2 million, 3.3-MW diesel generating system to work in conjunction with their hydroelectric units. With the addition of the diesel generator, MP&Ls total generating capacity came to just over 8-MW, twice the average base load on the utility. But to achieve reasonable efficiency for the diesel, a greater portion of the utilitys base load had to be shifted from the less expensive hydro generation to the more expensive diesel. Even with the addition of the diesel, electrical frequency often drooped to less than 57-Hz, and system voltage remained very erratic. Operation and maintenance costs for the diesel added to the problem. Fuel cost was $360,000 to $400,000 per year. Transporting 475,000 gallons per year of diesel fuel by ferry from the mainland, and then through pipe across the island increased both the environmental risk and the financial burden to the community. Each fuel shipment required an average cash outlay of $100,000 a significant amount for a small local utility. In addition, minor overhauls to the diesel cost $150,000 every three years; and major overhauls every six years cost $250,000. A techno/economic feasibility study was conducted by GNB Technologies and General Electric Company with assistance from Sandia National Laboratories that compared battery energy storage to other options using only the existing hydro and diesel units. The study indicated that a 1-MW, 1.4MWh battery energy storage system (BESS) could provide the spinning reserve, frequency control, and power quality improvements that Metlakatla needed. The study concluded that the cost of the BESS could be recovered within three years based on operational cost savings alone.

For the past several years, GNB has been intimately involved, in conjunction with Sandia National Laboratories, in developing the concept of battery energy storage as a way to supplement and improve the quality of power received from the utilities. In the past, battery energy storage systems were used simply to supplement generation capacity for the utility to meet demands during periods of high electrical usage. Operation in this limited mode alone could not justify the cost of a BESS. The more recent uses of battery energy storage in utility application have been to correct on-going power quality issues, in addition to providing a reserve of energy for uninterrupted power supply, peak shaving and load leveling. By introducing these additional functions, a BESS can be a viable alternative to other power management solutions. To achieve these objectives however, it is necessary to construct and operate the battery in an unconventional manner. First, the BESS has to be integrated with the utility power feed so that its operation is seamless to the power user. This requires that the battery system be at a relatively high voltage. Second, the battery would have to be operated in a partially discharged state, often for months without any equalization recharge, so as to efficiently accept power from the grid as well as to rapidly deliver power to the utility grid to instantaneously correct power quality issues. Unique as these operating conditions are, they bear a very strong resemblance to the conditions being considered by telecommunications operators. Data from these BESS batteries can help telcos evaluate and better understand these high voltage battery systems.

2.1 The Metlakatla BESS Battery The battery at the Metlakatla BESS facility (Figure 1) consists of 378 GNB ABSOLYTE IIP 100A75 modules arranged in a single series-connected string providing the system with its nominal 756-volt rating. The 100A75 cell has a nominal C/8 capacity rating of 3,600 Ampere-hours; its rating at the intended 90-minute discharge rate for this

2.

Metlakatlas BESS Opportunity

The Metlakatla BESS is located in the community of Metlakatla on the Annette Island Reserve at the southern tip of Alaska. Metlakatla Power and Light (MP&L), a stand-

12.2

application is approximately 2,000-Ah / 3.87-kWh. The entire battery system is rated 1.4-MWH at a 1.0-MW discharge rate. Each 100A75 cell is comprised of three individual 100A25 cells connected in parallel within the cells modular container, thus providing a statistical population base of 1,134 samples. The battery connects to a General Electric power conversion system (PCS), based on gate-turn-off (GTO) thyristors, that can support a continuous load of 800-kVA and pulse loads of up to 1200-kVA. The PCS allows bidirectional power flow between the ac system and the battery in less than a quarter-cycle. A 900-kVA filter bank removes harmonics and compensates the voltage of the electrical signal. The BESS connects to the MP&L grid at a 12.47-kV substation. The battery is housed in a 40 x 70-ft steel Butler building that sits on a concrete pad at the substation. An automatic generation control (AGC) system provides computerized control and dispatch of MP&Ls hydro and diesel units as well as the BESS for optimum efficiency. The AGC can be remotely accessed to monitor the status of the battery bank. The batterys 378 cells are arranged in two back-to-back rows, each row comprised of twelve stacks of ABSOLYTE modular trays eight high, separated by an aisle. The battery is positioned to minimize cable runs between rows of battery stacks and to the power conversion equipment. Pilot cell and temperature measurements are made at locations strategically positioned throughout the battery bank. Air is circulated by a fan to maintain consistent temperatures within the building.

A heater is provided to warm the facility during the colder winter months; however, only outside air is circulated for forced convection cooling. Operation of the Metlakatla BESS battery started in February, 1997, and except for a few short periods when the system has been purposely shut down for maintenance to either the battery or the systems electronic inverters, it has operated essentially continuously since its commissioning. Warranty on the battery is based on an 8-year service life. 2.2 Typical Operation of the BESS Battery

The BESS was designed to be connected continuously to the MP&L grid. From a fully charged condition, the battery is first allowed to be discharged to about an 80% state-ofcharge (SOC). After reaching that point, the battery is then allowed to accept recharge from the grid when load demand is less than the output of the hydroelectric units. The BESS PCS inverters draw power from the battery to instantaneously satisfy surge events on the grid. The BESS AGC computer monitors the current flowing out of or into the battery and automatically adjusts the output of MP&Ls hydroelectric units to essentially maintain the battery at about its 80% SOC point. The control algorithm assumes a 100% charge acceptance efficiency of the battery as it accepts charge at an SOC less than 90%. The charge algorithm that is used to control the recharge of the battery during operation is summarized in Table 1. The power limitations of the PCS equipment itself is the only factor that limits battery

Figure 1 The battery at the Metlakatla BESS consists of 378 GNB ABSOLYTE IIP 100A75 modules connected in series to deliver 1.4-MWH at a 1.0MW discharge rate at a nominal 756volts.

12.3

discharge current. Equalization charges are scheduled twice each year.

TABLE 1 BESS Recharge Control Algorithm Step 1 2 3 4 Eq Mode Current Current Voltage Voltage Voltage Control Parameter 35A/100Ah 25A/100Ah 2.32 vpc 2.25 vpc 2.35 vpc Limit / Transition to 2.25 vpc to 2.32 vpc 18A to 2A/100Ah Continuous 12 Hours

output from the battery of 745,735-Ah; total charge input to the battery was reported by the computer to be 751,468-Ah. Four individual 100A25 cells were selected from various locations within the battery to represent positions where variations in temperature had been observed and recorded by the battery monitoring system. The monitor system indicated that the battery was at about 78-81% state-of-charge when the cells were sampled. The samples were purposely taken prior to an equalization charge in order to assess the accuracy of the monitoring systems state-of-charge algorithm. The open circuit voltage measured on the cells (2.089 2.099 volts) correlated well with the monitors approximation of battery state-of-charge. Previous testing at GNB had shown that new ABSOLYTE cells allowed to self-discharge to an open circuit voltage of 2.09-volts were able to deliver 78.5% of their nominal 1-hour capacity rating. The internal impedance of the cells was measured at an average value of 262-ohms; the nominal impedance for this size cell is 229ohms. After the sample cells were removed from the battery string, they were shipped from Metlakatla to GNBs laboratories in Lombard, Illinois a suburb of Chicago. Spare cells that were being maintained at the Metlakatla BESS facility were used as replacements for the sampled cells. After arriving at the laboratory, the open circuit voltage, impedance and weight of each of the sample cells was recorded. No significant change in open circuit voltage or internal impedance occurred during transit, and cell weights were within the accepted tolerance range for this size

The PCS provides both active and reactive power to counter load swings created by the log chipper. The BESS sources watts and VArs when the system load jumps higher than the average, and sinks watts and VArs when the load falls below the average. Because the BESS resultant net output is nearly zero, the batteries require little additional charging. When required, the AGC dispatches the hydro units to provide the minimal overcharge the battery requires. Operation of the battery is shown in Figure 2, which shows the printout of the battery system screen for a typical day.

3.

Sampling Cells from the Metlakatla BESS

In October 1999 (approximately 32 months after system start-up), GNB and Sandia conducted a planned surveillance sampling of cells from the Metlakatla BESS battery. When the cells were sampled, the AGC computer indicated a total

Figure 2 Printout of the Metlakatla BESS battery monitor screen showing system operation on a typical day. The dark line in the center of the screen represents the current flowing into and out of the battery.

12.4

cell. One of the cells was retained in the as received condition for teardown, visual inspection and chemical analyses. The remaining three cells were reassembled into appropriately sized module trays for further electrical testing and characterization.

a telecommunications application, and result in a permanent reduction in the batterys capacity. On the next several charge / discharge cycles, the amount of recharge was intentionally limited to 112% of the previous capacity discharged. The recharge profile used was the same as programmed for the BESS at Metlakatla (see Table 1). In this way it would be possible to determine the efficiency of charge acceptance of the battery even when the amount of available recharge was limited. The actual C/8 discharge capacity as well as the percentage increase over the previous cycle is summarized in the following table where the recharge ampere-hours have been intentionally limited.

4.

Electrical Characterization of Sampled Cells

The three cells sampled from the Metlakatla BESS were connected together in series to form a 6-volt battery. The battery was then discharged, without any refreshening or boost charge, at 150-Amps, its nominal C/8 rate. On this first, as sampled discharge, the test battery delivered 766Ah or 63.9% of its rated capacity. All three of the cells performed similarly (Figure 3). The delivered capacity was less than the residual capacity estimated to be available from the cells by the BESS monitoring system. However, since the monitoring system estimates battery state-of-charge simply by summing ampere-hours discharged and charged, some error is to be anticipated especially considering the 6-month interval over which the estimate was made. The monitor resets to 100% state-of-charge following an equalization charge.

TABLE 2 Charge Acceptance Test Results Discharge # 1 2 3 4 Discharge Ah 766.5 856.9 945.0 1,023.0 % Rated 63.9 71.4 78.8 85.3 Recharge Ah Capacity Increase 858.6 959.8 111.8% 1,058.5 110.3% N/A 108.3%

2.15 2.10 Cell # 1 2.05 2.00

VPC

Cell # 2 Cell # 3

1.95 1.90 1.85 1.80 1.75 0 1 2 3 TIME (HOURS) 4 5 6

The data shows the partially discharged cells accepting almost the full amount of overcharge provided them during these limited recharge experiments, with the capacity increase of the cells essentially being equivalent to the amount of overcharge provided. As the cells capacity increases and approaches a fully charged state, overcharge acceptance efficiency starts to decrease as might be expected. Interestingly however, these experiments demonstrate an almost 100% ampere-hour charge acceptance efficiency for these samples even after having been operated in a partially discharged state for over 6-months. It is important to appreciate that the recharge voltage during these experiments was limited at 2.32 volts per cell, and that the actual recharge time for each of the recharges was less than 4 hours. Data from a typical recharge where recharge is limited to 112% of the capacity discharge is shown in Figure 4. 4.2 Discharge Performance after Equalization Charge

Figure 3: Cells sampled from the Metlakatla BESS delivered 64% of their rated capacity without any boost charge after being operated in a partially discharged state for over 6 months.

4.1

Charge Acceptance Test

The next objective was to determine if operating the battery in a partially discharged state for extended periods had caused any permanent deterioration of the battery. It is widely thought that failing to adequately recharge a lead-acid battery can cause hard sulfate to form within the cells active materials. This would hinder recharge acceptance, especially at low charge voltages as might be encountered in

Following the last of the four limited recharge cycles, the sample cells were given a standard equalization charge in accordance with the recommendations in the battery operating manual. The test samples were then discharged at the nominal C/8 discharge rate delivering 99.3% of rated capacity to a cutoff of 1.75 volts per cell. The samples were subjected to four additional discharge cycles, and the battery continued to deliver, on average, 101.9% of its rating. In addition to these capacity discharges, the sampled cells were subjected to a series of discharges to verify capacity conformance at various discharge rates. Average compliance to published specifications for this size cell was 105.6% for discharge rates ranging from C/1 to C/24.

12.5

Typical Limited Amp-Hr Recharge of Metlakatla Cells Amps 450

Current (Amps) / % Recharge

cell had not been changed as the result of operating in a partially discharged state for such an extended period.
2.38 2.36 2.34 2.32 2.30 2.28 2.26 2.24 2.22 2.20
Cell Voltage (vpc)

% Rchg

Avg VPC

4.4

Total Capacity Discharged

400 350 300 250 200 150 100 50 0 0.0 0.5 1.0 1.5 2.0 Time (Hours) 2.5 3.0 3.5 4.0

Based on cycle life testing conducted at GNB, the data suggests that each unique battery design has a certain lifetime discharge throughput that is somewhat independent of discharge rate or depth of discharge [7]. For the ABSOLYTE IIP design this discharge throughput is equal to approximately 1,000 times the nominal C/8 capacity of the cell in ampere-hours; approximately 3.6 million Ah for the 100A75 configuration installed at Metlakatla. In order to maintain the power quality of MP&Ls grid, the BESS battery is required to be alternately discharged and charged to supplement the fixed output of the hydroelectric generating units to meet the variable customer demand on the utility. The BESS monitoring system continuously monitors and integrates current flow into and out of the battery string. The total capacity discharged from the battery since its commissioning was reported by the monitoring system as being 745,735-Ah. Thus it is possible to estimate the amount of cycling the cells have experienced and to estimate the discharge throughput for the cells at the time they were sampled as being approximately 21% of its lifetime capability. This information will be helpful when the condition of the cell plates is examined.

Figure 4: Data recorded during a limited ampere-hour recharge of the Metlakatla sample cells shows their ability to readily accept high levels of charge current at relatively low charge voltages.

These discharge data indicate that the cells sampled from the Metlakatla BESS after over 30-months of operation, most of which was continuously cycling between about 70 and 90% state-of-charge, met or exceeded all of the performance specifications for this size cell. Based on these electrical characterizations, it was concluded that there had been no deterioration or damage done to the cell as the result of the unusual manner in which these cells were operated. 4.3 Float Behavior of Sampled BESS Cells

5.

Internal Examination of the BESS Cells

One further electrical test was conducted on the cells sampled from the Metlakatla BESS; that being a float charge test to establish the cells float current over the range of float voltages recommended for the ABSOLYTE IIP design.

Two of the sampled cells were torn down for physical and chemical analyses the cell retained in its as received condition, and one of the cells that had completed the electrical testing described previously. Both cells were examined at the same time to allow visual comparisons to be made. Sandia personnel assisted during these examinations. Overall Observations. Both cells appeared normal after the covers were removed. The cells were tightly compressed within the cell jar. All plates were completely encapsulated by the glass mat separator. There was no evidence whatsoever of any strap, terminal post or plate lug corrosion. There was no free liquid electrolyte in either of the cell jars once the cell elements were removed. Negative Plates. Negative plates sampled from three locations within the cell element stack were examined from both cells. All of the negative plates exhibited normal wetness as demonstrated by pressing the plates to observe a halo of liquid electrolyte around the finger being pressed onto the plate. Negatives from the cycled cell exhibited a shiny metallic streak when tested by striking across the surface of the plate with a hard object. Plates from the as

TABLE 3 Float Current Behavior at Float Voltage Float Voltage ( vpc ) 2.22 2.25 2.28 Float Current ( mA/100Ah) 20 55 117

Specification at 2.25vpc is 45 55 mA/100Ah

The float current at the various voltages tested was exactly as would be predicted for this size cell after having stabilized in a pure float charge application, suggesting that the surface properties and morphology of the electrodes in the

12.6

received cell also exhibited the metallic sheen, although not quite as shiny as that observed on the cycled cell. Chemical analysis of the as received cell negative plates indicated a lead sulfate content of approximately 28% which correlates well with the discharge capacity delivered by the other cells in their initial capacity test. The lead sulfate content in the negative plates from the cycled cell was approximately 10% which is typical for a fully charged VRLA cell. Positive Plates. Positive plates from neither of the cells showed any visible signs of surface sulfation that might have developed over the time during which these cells were operated at Metlakatla in a partially discharged state. Plates from both of the cells were dark brown to black in color. Although the positive paste on the as received cell visually appeared to be slightly drier, the active material on both cells plates was firm and crispy. X-ray diffraction analyses of the active materials indicated greater than 89% PbO2 for the cycled cell and 66% for the as received cell. Wet chemical analyses matched with the x-ray data and indicated 12% lead sulfate in the cycled cell and 29% lead sulfate in the as received sample. Both the negative and the positive active materials were what should be expected for cells considering their operational history and treatment prior to analysis. Positive Grids. Samples of the positive grids from both cells were taken, cross sectioned and polished to determine the amount of corrosion that had occurred over the lifetime of these cells. Metallurgical photos of cross-sectioned samples of these grids are shown in Figures 5 and 6.

Figure 6 Cross section of a positive grid from the as received cell sampled from the Metlakatla BESS after 30 months of operation in a partially discharged state shows a minimal amount of corrosion.

Measurement of the corrosion layers indicate a corrosion thickness of approximately 0.13 0.18mm. Compared to the corrosion rate basis of 0.08mm per year used to determine the design life for the ABSOLYTE IIP product, the actual corrosion rate for these cells was 0.05 to 0.07 mm/yr. These measurements indicate that the rate of positive grid corrosion with the Metlakatla operating and charge regime is less than that experienced under pure float conditions. It is important to note that the accelerated corrosion that was of concern because of the batterys operation in a partially discharged condition (as suggested by the shape of the Landers curve), is not being observed. If anything, the amount of positive grid corrosion that the samples exhibited was less than what would have been expected even under ideal float charge conditions with temperature and float voltage strictly maintained. Dimensions of positive grids from both cells were measured to assess the extent of positive plate growth. Growth in the long dimension of the grid was less than 0.4%; growth in the short dimension was less than 0.12%. In either case, the amount of positive grid growth was much less than the 6% allowance provided in the design of the cell. Separator and Electrolyte. The glass mat separator on both cells was adequately wetted. There was no excess free liquid electrolyte observed in either cell.

Figure 5 Cross section of the positive grid sampled from a cycle tested cell from the Metlakatla BESS shows a minimal amount of corrosion.

Concentration of the sulfuric acid electrolyte solution from the cycled and charged cell averaged at 1.309 s.g. Variation in concentration between the top section of the separator

12.7

and the bottom section was approximately 0.006 specific gravity units. The measured concentration is at the design concentration for this type cell and indicates that there has been no loss of water from the cell that would have caused the electrolyte concentration to be increased. The consistency of the electrolyte concentration across the separator demonstrates the ability of the cell to resist electrolyte stratification, even after being operated for an extended period of time in a partially discharged state. These cells were operated in a horizontal orientation. The concentration of the sulfuric acid electrolyte solution from the as received cell was lower, as would be expected for a partially discharged cell. The average concentration measured was 1.241 s.g. with a variation between the top and the bottom of the separator of 0.002 specific gravity units. The consistency of the electrolytes concentration demonstrates the excellent capabilities of the AGM material used in this cell design to support diffusion to prevent electrolyte stratification. Internal Top Lead. At several of the past INTELEC meetings, concerns have been expressed regarding the stability and corrosion resistance of the internal lead busbars, straps and terminal posts of VRLA designs. These concerns are especially associated with the negative plate hardware internal to the cell. Reaction of these lead parts with oxygen gas, and the minimum amount of negative plate polarization have been identified as contributing to this unusual type of corrosion. Samples of the negative plate busbar strap and the negative terminal post were taken, cross sectioned and polished to determine if these internal lead parts were experiencing unusual corrosion under the operating conditions for the Metlakatla BESS. A section of the negative plate strap is shown in Figure 7; a cross section of one of the cells negative terminal posts is shown in Figure 8.

The negative plate lugs are all firmly embedded within the strap, forming continuously bonded connections. There were no signs of any oxidation corrosion on the negative strap material itself. Similarly, there no signs of any corrosion of the negative terminal post material or the cover seal bushing.

Figure 8 Cross section of a negative terminal post from a Metlakatla BESS cell shows no corrosive attack.

6. Summary and Conclusions It has long been held that to achieve optimum life and performance from a lead-acid battery, it is necessary to float the battery under rigid voltage conditions to overcome selfdischarge reactions while minimizing overcharge and corrosion of the cells positive grid. This has resulted in batteries being used in telecommunications applications strictly in a standby mode. This may have been acceptable when the battery supported a 48-volt dc power plant. However, as telecommunications providers expand their horizons to supply video and Internet services in addition to conventional voice services, equipment architecture is demanding that, more and more, high voltage ac power be supplied for standby purposes. Thus battery power plants become extensions of the ac power grid. GNB in conjunction with Sandia National Laboratories has been active in the design, installation and monitoring of large battery strings used in conjunction with traditional utility sources. One of these programs has been the BESS at Metlakatla, Alaska. An important part of these efforts is follow-on analysis of battery lifetime in these applications. Data has been provided to demonstrate the long-term viability of VRLA cells in this type of use. As telecommunications power requirements change, it is conceivable that battery power plants in the telecommunications industry will take on a similar complexion; and that the battery in these power

Figure 7 Cross section of a negative plate strap from one of the Metlakatla BESS battery samples shows no indication of corrosion or oxidation that could reduce life.

12.8

plants will perform additional functions such as load leveling, peak shaving and power quality enhancement to justify its cost. Detailed examination of cells sampled from the battery system at the Metlakatla BESS after over 30 months of operation showed no unusual conditions that would signal an early degradation of the cells components. Positive and negative active materials composition was consistent with the state of charge of the cell when sampled. Active material structure was essentially in as new condition. Positive grid corrosion as evidenced by metallurgical examination and dimensional measurements to assess growth appeared to be even less than what one would expect in a perfectly controlled float charge environment. The degree of wetness of the cells separator materials was appropriate for this design cell and there was absolutely no indications of electrolyte stratification or concentration variations that would suggest excessive self discharge of the cell or loss of water from the cell. Furthermore all hard lead components within the cell (i.e., straps, plate lugs and terminal posts) were in pristine condition showing no evidence whatsoever of any unusual corrosive attack. Overall the condition of the cell could be described as unremarkable. For those in the telecommunications industry who are considering broadening the scope of operation of their battery systems to possibly supplement and enhance the quality their high voltage power supply, these observations provide encouragement. Admittedly, the data and observations discussed in this paper is but a single point. However, it is GNBs and Sandias plan to continue the surveillance of the Metlakatla battery both by continually monitoring its electrical performance during operation and to further sample cells from the battery throughout its lifetime. The telecommunications industry requires this information so that it can make enlightened decisions about how best to utilize one of its most underutilized facilities resources the battery.

8. References [1] Miller, N. W., et al., A VRLA Battery Energy Storage System for Metlakatla, Alaska Proceedings of the 11th Annual Battery Conference on Applications and Advances, Long Beach, CA 1996. Marquet, D., et al. New Power Supply for New Telecom Networks and Services Proceedings of Intelec 99, Copenhagen. Eklund, S. & Montin, S. Custom Designed Power Supply DC Applications for the Telecom Industry Proceedings of Intelec 99, Copenhagen. Akerlund, J. 48V DC Computer Equipment Topology An Emerging Technology Proceedings of Intelec 98, San Francisco. Jones, R., et al. Recharging VRLA Batteries for Maximum Life Proceedings of Intelec 98, San Francisco. Sideris, T., et al. Battery Aging and the Case for Stopping Float Charging Proceedings of Intelec 99, Copenhagen. Deshpande, S., et al. Intelligent Monitoring System Satisfies Customer Needs for Continuous Monitoring and Assurance on VRLA Batteries Proceedings of Intelec 99, Copenhagen. Kakalec, R. J. & Kimsey, T. H. A New Battery Plant Configuration that Eliminates Thermal Runaway in Valve Regulated Lead-Acid Batteries Proceedings of Intelec 2000, Phoenix.

[2]

[3]

[4]

[5]

[6]

[7]

[8]

7. Acknowledgement GNB Technologies acknowledges and appreciates the support and technical assistance provided by Sandia National Laboratories in advancing the use of batteries, and in particular VRLA designs, to support and improve the reliability and quality of utility provided electrical power. Furthermore, GNB acknowledges MP&L for allowing us access to the battery system, to monitor its operation and to collect samples for these aging and surveillance studies and examinations. Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the United States Dept. of Energy under Contract DE-AC0494AL85000.

12.9

13.1

13.2

13.3

13.4

13.5

13.6

13.7

13.8

13.9

14.1

14.2

14.3

14.4

14.5

14.6

14.7

GLOBAL OPERATIONS NORTH AMERICA GNB Industrial Power Chicago, Illinois U.S.A. TEL: 1.630.629.5200 FAX: 1.630.629.2635 GNB Industrial Power Maple, Ontario Canada TEL: 1.905.669.9326 FAX: 1.905.669.7688 EUROPE Exide Technologies B dingen, Germany u TEL: 49.6042.8170 FAX: 49.6042.81233 MIDDLE EAST/AFRICA Exide Technologies Abu Dhabi, U.A.E. TEL: 971.2.226235 FAX: 971.2.227644 JAPAN GNB Industrial Power Japan Tokyo, Japan/Pacific Rim TEL: 81.3.5325.6281 FAX: 81.3.5325.2063 AUSTRALIA/NEW ZEALAND Exide Technologies Padstow, N.S.W. Australia TEL: 61.2.9722.5700 FAX: 61.2.9774.2966 SOUTH EAST ASIA Exide Technologies S.E. Asia Singapore TEL: 65.546.2866 FAX: 65.546.2966 CHINA Exide Technologies Hong Kong, China TEL: 852.3106.2668 FAX: 852.3106.0260 Exide Technologies Beijing, China TEL: 86.10.6510.2910 FAX: 86.10.6510.2912 LATIN AMERICA GNB Industrial Power Lombard, Illinois U.S.A. TEL: 1.630.629.5200 FAX: 1.630.629.2635 INDIA GNB Industrial Power Bangalore, India TEL: 91.80.550.0581 FAX: 91.80.550.0582

Industry Leader in Network Power

The Network Power Division of Exide Technologies is the global leader in stored electrical energy solutions for all major critical reserve power applications and needs. Such network power applications include communication/data networks, UPS systems for computers and control systems, and electrical power generation and distribution systems. With a strong manufacturing base in both North America and Europe and a truly global reach (operations in greater than 80 countries) in sales and service, the Network Power Division has all of the tools necessary to satisfy your power needs.

Global Brands
GNB Flooded

Based on over 100 years of technological innovation the Network Power Division continues to lead the industry with such recognized global brands as Absolyte, Sonnenschein, Marathon, Sprinter, and Flooded Classic. These products and brands are synonymous with quality, reliability, performance and excellence in all markets served.
RN TU RE

Total Battery Management

LE AD

RECYCLE

In addition to being the leader in delivering premium products to the market, Exide Technologies takes pride in its commitment to the environment. As part of a complete approach to manufacturing, distributing, and recycling lead acid batteries, the Total Battery Management program has been developed to ensure a safe and responsible life cycle for all of our products.

www.gnb.com

GB-3819

REV 4/02

Printed on recycled paper.