Embed Size (px)
Citation preview
ABSOLYTETECHNOLOGY
PROFESSIONAL PAPERS
Table of Contents
Development of Totally Maintenance-Free Lead-Acid Battery forTelecommunications Standby Power 1.1
Abstract 1.1Introduction 1.1Battery Design and Construction 1.1Float Charge Characteristics 1.3Deep Discharge Cyclic Performance 1.3Partial State-of-Charge Operation 1.4Conclusion 1.5Acknowledgments 1.6References 1.6
Operational Characteristics of a Sealed Gas-Recombinant Lead-Acid Battery -An Update- 2.1
Abstract 2.1Introduction 2.1Deep Discharge Battery Operations 2.2Flat Discharge 2.3Cell Reversal 2.4Float Life Projections 2.4Gas Evolution and Water Loss 2.5Summary and Conclusion 2.6Acknowledgement 2.6
The Maturing of a Valve Regulated (VRLA) Battery Technology:Ten Years of Experience 3.1
Abstract 3.1Introduction 3.1Commercial Product History 3.2Absolyte II 3.3Absolyte IIP 3.3VRLA Battery Life 3.4Manufacturing Improvements 3.5VRLA Battery Testing 3.7Summary and Conclusions 3.7Acknowledgement 3.7References 3.7
Product Design and Manufacturing Process Considerations for the Application of a 10 Year Design Valve RegulatedLead Acid Battery in the Outside Plant Environment 4.1
Abstract 4.1Background 4.1Volumetric Energy Efficiency 4.2Thermal Management 4.3Performance and Reliability 4.5Conclusion 4.9
Real World Effects on VRLA Batteries in Float Applications 5.1Abstract 5.1Introduction 5.1Purpose of Float Voltage 5.1Response to Charging Current 5.2Optimum Float Voltage 5.3Optimum Temperature 5.4Combined Effects 5.5Thermal Stability 5.6Summary 5.7
A Discussion About Water Loss, Compression and the VRLA Cell 6.1
Abstract 6.1Introduction 6.1Water Loss Model 6.2Positive Grid Corrosion 6.2Vapor Transmission 6.3VRLA Gassing Study 6.4Results - Test 1 6.4Results - Test 2 6.5Saturation Effects Study 6.6Results 6.6Model vs. Test Conclusions 6.7Separator Compression Effects 6.7Conclusions 6.9Acknowledgements 6.9
Operational Characteristics of VRLA BatteriesConfigured in Parallel Strings 7.1
Introduction 7.1Theory 7.2Normal Parallel Configuration 7.2Parallel Operation of Different Capacity VRLA Batteries 7.3Parallel Operation of Batteries At Differing States of Charge 7.4Parallel Operation at Higher Discharge Currents 7.5Multiple Parallel Strings 7.6Constant Power Discharge Loads 7.7Testing Cable Configurations in Parallel Battery Systems 7.8Conclusion 7.8
A Guideline for the Interpretation of Battery DiagnosticReadings in the Real World 8.1
Abstract 8.1Introduction 8.1Baseline Values 8.2Impedance as a function of State of Charge 8.3Impedance as a function of accelerated float life 8.4Impedance as a function of Cell Dryout 8.5Impedance as a function of Loss of Compression 8.6Conclusions & Guidlines for using impedance as a diagnostic tool 8.8
Intelligent Monitoring System satisfies customer needs forContinuous Monitoring and Assurance on VRLA Batteries 9.1
Introduction 9.1Diagnostic Tools 9.1Continuous Monitoring: Physical Parameters 9.2Continuous Monitoring: Usefulness of Output 9.3Continuous Monitoring: Customer Requirements Satisfied
by Intelligent Monitoring System (IMS) 9.6Conclusions 9.7Acknowledgements 9.7References 9.7
An Examination of High Rate Recharge on Absolyte IIP Batteries 10.1Abstract 10.1Introduction 10.1Test One: Single Module Testing, Multiple Currents 10.1Test Two: 48 Volt System Testing 10.4Summary 10.6
Considerations for the Configuration and Arrangement ofValve Regulated Monobloc Batteries in Enclosures 11.1
Abstract 11.1Introduction 11.1Test Results 11.2Discusion of Results 11.4Conclusion 11.5Recommendations for Future Work 11.5
Examination of VRLA Cells Sampled from a Battery EnergyStorage System (BESS) after 30-Monts of Operation 12.1
Abstract 12.1Background 12.1Changes in Telecommunications Power 12.1Metlakatla’s BESS Opportunity 12.2Sampling Cells from the Metlakatla BESS 12.4Electrical Characterization of Sampled Cells 12.5Internal Examination of the BESS Cells 12.6Summary and Conclusions 12.8Acknowledgement 12.9References 12.9
Considerations for the Configuration and Arrangement of Valve RegulatedMonobloc Batteries in Enclosures - Part II 13.1
Abstract 13.1Introduction 13.1Testing and Results 13.2Test Summary 13.3Tests 1, 2 and 3 (5 mm Spacing) 13.3Tests 4, 5 and 6 13.4Tests 7, 8 and 9 13.4Tsts 1, 2 and 3 (10 mm Spacing) 13.5Discussion of Results 13.5Spacing of Blocks 13.6Ventilation 13.7String Location within the Cabinet 13.7Top Terminal Battery Configuration 13.8Conclusions 13.9
A Naturally Aged VRLA Battery: 18 Years Later 14.1Abstract 14.1Introduction 14.1Initial Test Data: 14.1OCV’s Vent Opening Pressures and IR’s Before and After Charge 14.2Internal Examination 14.2Electrical Testing 14.3Internal Resistance and Capacity 14.4Discussion of Initial Results 14.5Capacity Recovery: Overcoming Compression Loss 14.5Comparison of 3 Cells’ Performance:
Compressed and Uncompressed 14.6Compression: How Much? 14.6Long Term Float 14.7Conclusion 14.7References 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
Capacity
Impedance
Saturation
Per
cent
Sat
urat
ion
Per
cent
Rat
ed C
apac
ity
Impe
danc
e -
mΩ
Trial Number
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.
Robert J. Schmitt Technical Marketing Engineer
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 industry’s 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 semi-conductors, 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 in-rush 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 battery’s 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:
10.1
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 cell’s 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 hours—nearly 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 Behavior at 2.35 VPC 150 A/100-Ah
0
20
40
60
80
100
120
0 1 2 3 4 5 6 7 8 9 10 11 12
Time - hours
Cur
rent
- A
/100
-Ah
0
20
40
60
80
100
120
Per
cent
Rec
harg
e
% Recharge
Current
Recharge Time with 12, 24 and 72 A/100 A-h Available Charge Current
0%
20%
40%
60%
80%
100%
120%
0 1 2 3 4 5 6 7 8
Time - hours
Per
cent
Rec
harg
e
12 A/100-Ah
24 A/100-Ah
72 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%.
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 9°C temperature rise was gone in about 12 hours.
Cumulative Saturation ChangeCurrent Trial Cell 1 Cell 2 Cell 3 Cell 4 Cell 5 Cell 650 A/100-Ah ∆∆∆∆Saturation1 -0.10% -0.20% -0.24% -0.24% -0.13% -0.10%100 A/100-Ah ∆∆∆∆Saturation2 -0.17% -0.30% -0.30% -0.27% -0.17%150 A/100-Ah ∆∆∆∆Saturation3 -0.17% -0.34% -0.37% -0.27% -0.13% -0.13%50 A/100-Ah ∆∆∆∆Saturation4 -0.13% -0.30% -0.37% -0.27% -0.10% -0.17%150 A/100-Ah ∆∆∆∆Saturation5 -0.17% -0.37% -0.34% -0.24% -0.13% -0.13%
Impedance Before and After Testing
0.38
0.39
0.40
0.41
0.42
0.43
0.44
0.45
0.46
Cell 1 Cell 2 Cell 3 Cell 4 Cell 5 Cell 6Cell Number
Impe
danc
e -
mill
i-ohm
s
Before TestingAfter Testing
10.3
TEST TWO: 48 VOLT SYTEM TESTING
Where the first test concentrated on charge characterization at GNB’s recommended equalize voltage, the second test sought to compare this charge behavior to that at the battery’s 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 264-Ah 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
0
20
40
60
80
100
120
0 1 2 3 4 5 6 7 8 9 10
Time - hours
Cha
rge
Cur
rent
- A
/100
-Ah
0%
20%
40%
60%
80%
100%
120%P
erce
nt R
echa
rge2.35 VPC
2.25 VPC
Recharge Percent
Current
Temperature Rise at Various Charge Currents
0
1
2
3
4
5
6
7
8
9
10
0 2 4 6 8 10 12 14
Time - hours
Tem
pera
ture
Ris
e -
°C
24 A/100-Ah
50A/100-Ah
100 & 150 A/100-Ah
10.4
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 battery’s 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 25°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 10°C 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 25°C. 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 battery’s 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.
module 2.35 VPC 2.25 VPC1 Top 33.5 28.02 34.0 28.53 34.5 29.04 Bottom 32.5 28.0
Maximum Attained Temperatures (°C)
Tem perature R ise Due to C harging at 2.25 and 2.35 VPC
0
20
40
60
80
100
120
0 1 2 3 4 5 6 7 8 9 10
Tim e - hours
Cha
rge
Cur
rent
- A
/100
A-h
0
2
4
6
8
10
12
Tem
pera
ture
Ris
e -
°C∆∆∆∆T(2.35 VPC)
∆∆∆∆T(2.25 VPC)
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 today’s 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.
1 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 sce-nario and present the results in a format that leads to a useful conclusion. The study conducted, involved ex-amining 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 design-ers. This popularity is a result of its favorable foot-print, 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 stan-dard for valve regulated lead acid batteries recom-mends 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 Ah’s of ca-pacity. The heat generation from this arrangement would be identical to that for a configuration consist-ing of four parallel strings, 100 Ah’s 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
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 deliv-ered 110% of the Ah’s 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 equip-ment located inside the cabinet that would influence the overall thermal behavior. A summary of the test-ing protocol is shown in Table 1 below.
Table 1 Test # Battery Spacing Ventilation
1 10mm Open cabinet
2 10mm Closed cabinet – natural
convection
3 10mm Forced ventilation
4 5mm Forced ventilation
5 5mm Closed cabinet – natural
convection
6 5mm Open cabinet
7 0mm Open cabinet
8 0mm Forced ventilation
9 0mm Closed cabinet – natural
convection
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-
4) and thermocouple location (C = center of string, O = outside of string). The term cabinet, or the abbre-viation cabnt, refer to the ambient temperature inside the cabinet.
2 Test Results
As expected, battery temperature variation was evi-dent in and dependent on the system configurations tested. A test by test summary of the results is pre-sented below. 2.1 Test #1 (10mm spacing/open
cabinet
The data in Figure 2 displays the battery temperature variation within the open cabinet where the surround-ing ambient temperature was maintained at 25±2°C. The shelf located in position 4 at the bottom of the cabinet reached a maximum temperature of 36°C. Shelves 1 and 2 located at the top of the cabinet reached a maximum temperature of 41°C, and the shelf located in position 3 above shelf 4 and below shelves 1 and 2 reached a maximum temperature of 42°C. The battery temperatures at the locations in the center of the strings within the cabinet varied by 5-7°C depending on shelf location and were as much as 17°C above the ambient temperature. The battery sys-tem was able to return to a steady state temperature after 40 hours.
2.2 Test #2 (10mm Spacing – Closed cabinet natural convection cooling
The data in figure 3 displays the battery temperature variation within the closed cabinet, natural convection cooling, 10-mm battery spacing and an ambient tem-perature of 25±2°C. Again the shelf located in posi-tion 4 at the bottom was the coolest reaching a maxi-mum temperature of 41°C. Shelves 1, 2 and 3 located above shelf 4 reached 43°, 42.5°C and 43°C respec-tively. The battery temperatures at the locations in the center of the strings within the cabinet varied by only
Figure 2. Open Cabinet10mm Space/Convection Cooling
20
25
30
35
40
45
50
0 10 20 30 40 50 60 70
Time (Hours)
Tem
per
atu
re (
C)
Shelf 3 BatteryShelf 2 Battery
Shelf 1 Battery
Shelf 4 Battery
11.2
1.5-2°C depending on shelf location and were as much as 18°C above the ambient temperature. The battery system was able to return to a steady state temperature after 60 hours.
2.3 Test #3 (10mm spacing – forced air ventilation)
The data in Figure 4 displays the battery temperature variations measured during the final test iteration with 10mm battery spacing within the closed cabinet hav-ing forced air-ventilation and an outside ambient tem-perature of 25±2°C. During this test, shelf 1 located at the top was the coolest reaching a maximum tem-perature of 32°C. Shelves 2, 3 and 4 located below shelf 1 reached 37°C, 35.5°C and 34°C respectively. The battery temperatures at the locations in the center of the strings within the cabinet varied by only 2-5°C depending on shelf location and were as much as 9°C 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).
Peak Temperatures: Tests 1,2 & 3
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 0.5-13°C with a peak temperature of 42°C seen in the top shelf which was 17°C 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 tempera-tures in the center of the strings varied by only 0.5-1.5°C with a peak temperature of 43.5°C in the top shelf which was 18.5°C above the ambient tempera-ture. The system was able to return to a steady state temperature after 48 hours. The final experiment investigated the effects of a 5mm battery spacing in a closed cabinet with forced air ven-tilation. The battery temperatures in the center of the strings varied by 2.5-3.5°C while attaining a peak bat-tery temperature of 41.0°C in the two middle shelves, which equated to a rise of 16.0°C above the ambient temperature. The system was able to return to a steady state temperature after 30 hours.
T 1C T 1O T 2C T2O T3C T3O T4C T4O CABINET
Open Cabinet 41.0 33.0 41.0 34.5 42.0 34.0 36.0 34.0 26.5
Closed Cabinet 43.0 37.5 42.5 38.5 43.0 37.0 41.0 36.5 35.0
Cabinet Forced Air 32.0 28.5 37.0 31.5 35.5 30.5 34.0 28.0 25.5
Table 2. 10 mm Spacing Test Peak Temperatures
1= Top Shelf 4= Bottom Shelf
41
36
41
34
33
34.5
42
44
26.5
43
41
42.5
37
37.5
38.5
43
36.5
35.0
32
34
37
30.5
28.5
31.5
35.5
28
25.5
Figure 3 Closed Cabinet 10mm Space / Fan Blocked Off
20
25
30
35
40
45
50
0 10 20 30 40 50 60 70
Time (Hours)
Tem
per
atu
re (
C) Shelf 4 Battery
Shelf 1 BatteryShelf 2 Battery
Shelf 3 Battery
Figure 4. Closed Cabinet10mm Space / Forced Cooling
20
25
30
35
40
45
50
0 10 20 30 40 50 60 70
Time (Hours)
Tem
per
atu
re (
C)
Shelf 4 Battery
Shelf 1 BatteryShelf 2 Battery
Shelf 3 Battery
11.3
Peak Temperatures: Test 4,5 & 6
2.5 Tests 7, 8, and 9 (0mm spacing trials)
Table 4 (and the diagram) summarizes the data for the third iteration of experiments that investigated the ef-fects of no space between batteries with the three dif-ferent types of ventilation. In the open cabinet ex-periment the battery temperatures in the center of the strings varied by only 0.5-1.5°C with a peak tempera-ture of 42°C again realized in the top string which was 17°C 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 tempera-tures in the center of the strings interestingly only var-ied again by 0.5-1.5°C with a peak temperature of 44.5°C again in the center of the top string which was 19.5°C 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 0-0.5°C attaining a peak battery temperature of 42.0°C again in the center strings which were 17.0°C above ambient. The system returned to a steady state after 40 hours.
Peak Temperatures: Test 7,8, & 9
3 Discussion of Results
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 loca-tion 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 ven-tilation. 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)
T1C T1O T2C T2O T3C T3O T4C T4O Cbnt
Stable Temp
Time to Stable
10mm 41C 33C 41C 34.5C 42C 34C 36C 34C 26.5C 25C 40hrs
5mm 42C 31.5C 41.5C 33.5C 42C 33C 40.5C 29C 27C 25C 40hrs
0mm 42C 31.5C 41.5C 32.5C 41.5C 32.5C 40.5C 27.5C 26C 25C 40hrs
T 1C T 1O T 2C T2O T3C T3O T4C T4O CABINETOpen Rack 41.0 33.0 41.0 34.5 42.0 34.0 36.0 34.0 26.5Cabinet 43.0 37.5 42.5 38.5 43.0 37.0 41.0 36.5 35.0Cabinet Forced Air 32.0 28.5 37.0 31.5 35.5 30.5 34.0 28.0 25.5
Peak Temperatures 10 mm Spacing1= Top Shelf 4= Bottom Shelf
42
29
41.5
40.5
31.5
33.5
33
27
26.5
43.5
42
42.5
36.5
36.5
37
43
31.5
32.0
37.5
40
41
33.5
30.5
33.5
41
29.5
26.0
T 1C T 1O T 2C T2O T3C T3O T4C T4O CABINET
Open Cabinet 42.0 31.5 41.5 32.5 41.5 32.5 40.5 27.5 26.0
Closed Cabinet 44.5 37.0 44.0 37.5 44.0 37.0 43.0 32.5 33.0
Cabinet Forced Air 41.5 33.0 42.0 33.0 42.0 34.0 41.5 30.5 26.5
Table 4. 0 mm Spacing Test Peak Temperatures
1= Top Shelf 4= Bottom Shelf
42
40.5
41.5
32.5
31.5
32.5
41.5
27.5
26
44.5
43
44
37
37
37.5
44
32.5
33.0
41.5
41.5
42
34
33
33
42
30.5
26.5
11.4
Table 6 - Closed cabinet – no ventilation
Table 7 - Closed cabinet – forced ventilation
The data clearly shows that spacing the batteries de-creases the maximum temperature at the center of the cabinet only when combined with forced air ventila-tion. This benefit was the greatest across all four lev-els 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 venti-lation (Tables 5 and 6), spacing distance does not sig-nificantly 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 tempera-ture ran approximately 5oC higher and where this en-tire 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.
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 tempera-tures 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, sig-nificant reductions in battery temperatures were not observed. Again, this is verified by examining the data in Tables 5, 6, and 7 above.
3.3 String location within the cabinet
Temperature profiles within the cabinet follow a fairly intuitive pattern. In the tests where no forced ventila-tion 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 cabi-net 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.
4 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 re-quired 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 re-ducing maximum battery temperatures.
4. Under normal recharge profiles battery tempera-tures internal to a cabinet will see anywhere from a 7oC to 20oC rise above ambient conditions (25+/-2oC). These temperatures will however re-turn 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.
5 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 tempera-
tures and under normal float conditions for ex-tended periods of time.
2. Variations in the initial charge current rate. 3. Location of electronic equipment inside the cabi-
net to monitor the effect of an additional heat source.
4. The impact of using standard top terminal mon-oblocs.
5. Impact of using perforated trays that would allow air movement between layers.
T1C T1O T2C T2O T3C T3O T4C T4O CbntStableTemp
Time toStable
10mm32C 28.5C 37C 31.5C 35.5C30.5C 34C 28C 25.5C 26C 24hrs
5mm 37.5C 30.5C 41C 33C 41C 33.5C 40C 28.5C26C 24C 30hrs
0mm 41.5C 33C 42C 33C 42C 34C 41.5C 30.5C26.5C 26C 40hrs
T1C T1O T2C T2O T3C T3O T4C T4O CbntStable Temp
Time to Stable
10mm 43C 37.5C 42.5C 38.5C 43C 37C 41C 36.5C 35C 27C 60hrs
5mm 43.5C 36.5C 42.5C 37C 43C 36.5C 42C 31.5C 32C 24C 48hrs
0mm 44.5C 37C 44C 37.5C 44C 37C 43C 32.5C 33C 25C 55hrs
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 TsagalisGNB Technologies Lombard, Illinois USA
Rudolph JungstSandia 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 year’s 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 “off-
line” 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.
1. Background
In 1996-97, GNB installed and commissioned a largeVRLA battery system for a Battery Energy Storage System
(BESS) at the island village of Metlakatla, Alaska [1]. The
battery’s function is to stabilize the island community’s
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 battery’s 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 tocontinuously operate the battery at between 70 and 90%
state-of-charge. Like some of the recently proposed
alternative charge regimes for telecommunications
installations, this battery’s operations raised several concerns
regarding long-term performance and life.
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 battery’s
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 year’s INTELEC, papers werepresented discussing high voltage power systems, up to 300-
volts, as alternatives to the traditional 48-volt power plant [2-
4]. 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 thetelecommunications 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
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 andequipment, 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)
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 qualityissues, 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 apartially 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. Metlakatla’s 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-
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 theAnnette 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
utility’s 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&L’s total
generating capacity came to just over 8-MW, twice theaverage base load on the utility. But to achieve reasonable
efficiency for the diesel, a greater portion of the utility’s 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 islandincreased 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.
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
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
cell’s modular container, thus providing a statisticalpopulation 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 bi-
directional 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. Anautomatic generation control (AGC) system provides
computerized control and dispatch of MP&L’s 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 battery’s 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 andtemperature 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 inFebruary, 1997, and except for a few short periods when the
system has been purposely shut down for maintenance to
either the battery or the system’s 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-of-
charge (SOC). After reaching that point, the battery is thenallowed 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&L’s 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 issummarized in Table 1. The power limitations of the PCS
equipment itself is the only factor that limits battery
Figure 1
The battery at theMetlakatla BESS
consists of 378GNB ABSOLYTEI I P 1 0 0 A 7 5
modules connectedin series to deliver1.4-MWH at a 1.0-
MW discharge rateat a nominal 756-volts.
12.3
discharge current. Equalization charges are scheduled twice
each year.
TABLE 1BESS Recharge Control Algorithm
Step Mode Control Parameter Limit / Transition
1 Current 35A/100Ah to 2.25 vpc 2 Current 25A/100Ah to 2.32 vpc 3 Voltage 2.32 vpc 18A to 2A/100Ah
4 Voltage 2.25 vpc Continuous Eq Voltage 2.35 vpc 12 Hours
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 loadfalls 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 systemstart-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
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 bythe 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 system’s state-of-charge algorithm.
The open circuit voltage measured on the cells (2.089 –
2.099 volts) correlated well with the monitor’s 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 internalimpedance of the cells was measured at an average value of
262-µohms; the nominal impedance for this size cell is 229-
µohms.
After the sample cells were removed from the battery
string, they were shipped from Metlakatla to GNB’s
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 circuitvoltage, 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
Figure 2
Printout of the Metlakatla
BESS battery monitorscreen showing systemoperation on a typical day.
The dark line in the centerof the screen representsthe current flowing intoand 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.
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 766-
Ah 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 fromthe 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.
Figure 3: Cells sampled from the Metlakatla BESS
delivered 64% of their rated capacity without anyboost charge after being operated in a partiallydischarged 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 cell’s
active materials. This would hinder recharge acceptance,
especially at low charge voltages as might be encountered in
a telecommunications application, and result in a permanent
reduction in the battery’s capacity.
On the next several charge / discharge cycles, the amount
of recharge was intentionally limited to 112% of the previouscapacity 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.
TABLE 2Charge Acceptance Test Results
Discharge # Discharge Ah % Rated Recharge Ah Capacity Increase 1 766.5 63.9 858.6 -
2 856.9 71.4 959.8 111.8% 3 945.0 78.8 1,058.5 110.3% 4 1,023.0 85.3 N/A 108.3%
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 experimentswas 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
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 atthe 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.
1.75
1.80
1.85
1.90
1.95
2.00
2.05
2.10
2.15
0 1 2 3 4 5 6T I M E ( H O U R S )
V P
C
C e l l # 1
C e l l # 2
C e l l # 3
12.5
Figure 4: Data recorded during a limited ampere-hour
recharge of the Metlakatla sample cells shows theirability to readily accept high levels of charge currentat 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 theunusual manner in which these cells were operated.
4.3 Float Behavior of Sampled 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.
TABLE 3Float Current Behavior at Float Voltage
Float Voltage Float Current ( vpc ) ( mA/100Ah)
2.22 20 2.25 55
2.28 117
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
cell had not been changed as the result of operating in a
partially discharged state for such an extended period.
4.4 Total Capacity Discharged
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&L’s grid,
the BESS battery is required to be alternately discharged andcharged 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.
5. Internal Examination of the BESS Cells
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
Typical Limited Amp-Hr Recharge of Metlakatla Cells
0
50
100
150
200
250
300
350
400
450
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Time (Hours)
Curr
ent (A
mps)
/ %
Rech
arg
e
2.20
2.22
2.24
2.26
2.28
2.30
2.32
2.34
2.36
2.38
Cell
Volta
ge (
vpc)
Amps % Rchg Avg VPC
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% whichcorrelates 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 visuallyappeared 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 5
Cross section of the positive grid sampled from a cycletested cell from the Metlakatla BESS shows a minimalamount of corrosion.
Figure 6
Cross section of a positive grid from the “as received”cell sampled from the Metlakatla BESS after 30months 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 battery’s operation in a
partially discharged condition (as suggested by the shape of
the Lander’s 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.
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. Theconsistency 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 electrolyte’s concentrationdemonstrates 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 polarizationhave 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 cell’s
negative terminal posts is shown in Figure 8.
Figure 7
Cross section of a negative plate strap from one of the
Metlakatla BESS battery samples shows no indicationof corrosion or oxidation that could reduce life.
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 self-
discharge reactions while minimizing overcharge and
corrosion of the cell’s 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 isdemanding 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 viabilityof 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
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 batterysystem at the Metlakatla BESS after over 30 months of
operation showed no unusual conditions that would signal an
early degradation of the cell’s 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 cell’s separator materials was appropriate for this
design cell and there was absolutely no indications ofelectrolyte 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 andenhance 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 GNB’s and Sandia’s 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.
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 tocollect 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-AC04-
94AL85000.
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.
[2] Marquet, D., et al. “New Power Supply for New
Telecom Networks and Services” Proceedings of
Intelec 99, Copenhagen.
[3] Eklund, S. & Montin, S. “Custom Designed Power
Supply DC – Applications for the Telecom
Industry” Proceedings of Intelec 99, Copenhagen.
[4] Akerlund, J. “48V DC Computer Equipment
Topology – An Emerging Technology”
Proceedings of Intelec 98, San Francisco.
[5] Jones, R., et al. “Recharging VRLA Batteries for
Maximum Life” Proceedings of Intelec 98, San
Francisco.
[6] Sideris, T., et al. “Battery Aging and the Case for
Stopping Float Charging” Proceedings of Intelec
99, Copenhagen.
[7] Deshpande, S., et al. “Intelligent Monitoring
System Satisfies Customer Needs for Continuous
Monitoring and Assurance on VRLA Batteries”
Proceedings of Intelec 99, Copenhagen.
[8] 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.
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 AMERICAGNB Industrial PowerChicago, Illinois U.S.A.TEL: 1.630.629.5200FAX: 1.630.629.2635
GNB Industrial PowerMaple, Ontario CanadaTEL: 1.905.669.9326FAX: 1.905.669.7688
EUROPEExide TechnologiesBudingen, GermanyTEL: 49.6042.8170FAX: 49.6042.81233
MIDDLE EAST/AFRICAExide TechnologiesAbu Dhabi, U.A.E.TEL: 971.2.226235FAX: 971.2.227644
JAPANGNB Industrial Power JapanTokyo, Japan/Pacific RimTEL: 81.3.5325.6281FAX: 81.3.5325.2063
AUSTRALIA/NEW ZEALANDExide TechnologiesPadstow, N.S.W. AustraliaTEL: 61.2.9722.5700FAX: 61.2.9774.2966
SOUTH EAST ASIAExide Technologies S.E. AsiaSingapore TEL: 65.546.2866FAX: 65.546.2966
CHINAExide TechnologiesHong Kong, ChinaTEL: 852.3106.2668FAX: 852.3106.0260
Exide TechnologiesBeijing, ChinaTEL: 86.10.6510.2910FAX: 86.10.6510.2912
LATIN AMERICAGNB Industrial PowerLombard, Illinois U.S.A.TEL: 1.630.629.5200FAX: 1.630.629.2635
INDIAGNB Industrial PowerBangalore, IndiaTEL: 91.80.550.0581FAX: 91.80.550.0582
www.gnb.com
GB-3819 REV 4/02
Printed on recycled paper.
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 systemsfor computers and control systems, and electrical power generation and distributionsystems. With a strong manufacturing base in both North America and Europe and atruly global reach (operations in greater than 80 countries) in sales and service, theNetwork Power Division has all of the tools necessary to satisfy your power needs.
Global Brands…
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 brandsare synonymous with quality, reliability, performance and excellence in all marketsserved.
Total Battery Management…In addition to being the leader in delivering premium products to the market, ExideTechnologies 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 andresponsible life cycle for all of our products.
Industry Leader in Network Power…
GNB Flooded
RECYCLE
LEA
D
RETU
RN
