ABSOLYTE

TECHNOLOGY

PROFESSIONAL PAPERS
 Table of Contents

Development of Totally Maintenance-Free

Lead-Acid Battery forTelecommunications Standby Power 1.1
Abstract 1.1
Introduction 1.1
Battery Design and Construction 1.1
Float Charge Characteristics 1.3
Deep Discharge Cyclic Performance 1.3
Partial State-of-Charge Operation 1.4
Conclusion 1.5
Acknowledgments 1.6
References 1.6
Operational Characteristics of a Sealed Gas-Recombinant
Lead-Acid Battery -An Update- 2.1
Abstract 2.1
Introduction 2.1
Deep Discharge Battery Operations 2.2
Flat Discharge 2.3
Cell Reversal 2.4
Float Life Projections 2.4
Gas Evolution and Water Loss 2.5
Summary and Conclusion 2.6
Acknowledgement 2.6
The Maturing of a Valve Regulated (VRLA) Battery Technology:
Ten Years of Experience 3.1
Abstract 3.1
Introduction 3.1
Commercial Product History 3.2
Absolyte II 3.3
Absolyte IIP 3.3
VRLA Battery Life 3.4
Manufacturing Improvements 3.5
VRLA Battery Testing 3.7
Summary and Conclusions 3.7
Acknowledgement 3.7
References 3.7
Product Design and Manufacturing Process Considerations
for the Application of a 10 Year Design Valve Regulated
Lead Acid Battery in the Outside Plant Environment 4.1
Abstract 4.1
Background 4.1
Volumetric Energy Efficiency 4.2
Thermal Management 4.3
Performance and Reliability 4.5
Conclusion 4.9
Real World Effects on VRLA Batteries in Float Applications 5.1
Abstract 5.1
Introduction 5.1
Purpose of Float Voltage 5.1
Response to Charging Current 5.2
Optimum Float Voltage 5.3
Optimum Temperature 5.4
Combined Effects 5.5
Thermal Stability 5.6
Summary 5.7
A Discussion About Water Loss, Compression
and the VRLA Cell 6.1
Abstract 6.1
Introduction 6.1
Water Loss Model 6.2
Positive Grid Corrosion 6.2
Vapor Transmission 6.3
VRLA Gassing Study 6.4
Results - Test 1 6.4
Results - Test 2 6.5
Saturation Effects Study 6.6
Results 6.6
Model vs. Test Conclusions 6.7
Separator Compression Effects 6.7
Conclusions 6.9
Acknowledgements 6.9
Operational Characteristics of VRLA Batteries
Configured in Parallel Strings 7.1
Introduction 7.1
Theory 7.2
Normal Parallel Configuration 7.2
Parallel Operation of Different Capacity VRLA Batteries 7.3
Parallel Operation of Batteries At Differing States of Charge 7.4
Parallel Operation at Higher Discharge Currents 7.5
Multiple Parallel Strings 7.6
Constant Power Discharge Loads 7.7
Testing Cable Configurations in Parallel Battery Systems 7.8
Conclusion 7.8
A Guideline for the Interpretation of Battery Diagnostic
Readings in the Real World 8.1
Abstract 8.1
Introduction 8.1
Baseline Values 8.2
Impedance as a function of State of Charge 8.3
Impedance as a function of accelerated float life 8.4
Impedance as a function of Cell Dryout 8.5
Impedance as a function of Loss of Compression 8.6
Conclusions & Guidlines for using impedance as a diagnostic tool 8.8
Intelligent Monitoring System satisfies customer needs for
Continuous Monitoring and Assurance on VRLA Batteries 9.1
Introduction 9.1
Diagnostic Tools 9.1
Continuous Monitoring: Physical Parameters 9.2
Continuous Monitoring: Usefulness of Output 9.3
Continuous Monitoring: Customer Requirements Satisfied
by Intelligent Monitoring System (IMS) 9.6
Conclusions 9.7
Acknowledgements 9.7
References 9.7
An Examination of High Rate Recharge on Absolyte IIP Batteries 10.1
Abstract 10.1
Introduction 10.1
Test One: Single Module Testing, Multiple Currents 10.1
Test Two: 48 Volt System Testing 10.4
Summary 10.6
Considerations for the Configuration and Arrangement of
Valve Regulated Monobloc Batteries in Enclosures 11.1
Abstract 11.1
Introduction 11.1
Test Results 11.2
Discusion of Results 11.4
Conclusion 11.5
Recommendations for Future Work 11.5
Examination of VRLA Cells Sampled from a Battery Energy
Storage System (BESS) after 30-Monts of Operation 12.1
Abstract 12.1
Background 12.1
Changes in Telecommunications Power 12.1
Metlakatla’s BESS Opportunity 12.2
Sampling Cells from the Metlakatla BESS 12.4
Electrical Characterization of Sampled Cells 12.5
Internal Examination of the BESS Cells 12.6
Summary and Conclusions 12.8
Acknowledgement 12.9
References 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
OCV’s Vent Opening Pressures and IR’s 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 - mΩ

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 GNB Technologies Robert J. Schmitt
Director of Marketing Lombard, Illinois U.S.A. 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
 Recharge Behavior at 2.35 VPC 150 A/100-Ah

120 120

100 % Recharge 100

Current - A/100-Ah

Percent Recharge
80 80

60 60

40 40
Current

20 20

0 0
0 1 2 3 4 5 6 7 8 9 10 11 12
Time - hours

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 Time with 12, 24 and

72 A/100 A-h Available Charge Current
120%
100% 72 A/100-Ah
Percent Recharge

80%
24 A/100-Ah
60%
40%
12 A/100-Ah
20%
0%
0 1 2 3 4 5 6 7 8
Time - hours

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 Cell 1 Cell 2 Cell 3 Cell 4 Cell 5 Cell 6
50 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

0.46
Before Testing
0.45
After Testing
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 5 Cell 6
Cell Number

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.

10.3
 Temperature Rise at Various Charge Currents

10
100 & 150 A/100-Ah
9

8
Temperature Rise - °C

4
50A/100-Ah
3

2
24 A/100-Ah
1

0
0 2 4 6 8 10 12 14
Time - hours

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

120 120%
2.35 VPC
Charge Current - A/100-Ah

100 100%
Percent Recharge

80 Recharge Percent 80%

2.25 VPC
60 60%

40 40%

20 20%
Current
0 0%
0 1 2 3 4 5 6 7 8 9 10
Time - hours

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.

T em p eratu re R ise D ue to C harging at 2.25 and 2.35 V P C

120 12
Charge Current - A/100 A-h

100 10
∆ T (2.35 V P C )

Temperature Rise - °C
80 8

60 6

40 4
∆ T (2.25 V P C )
20 2

0 0
0 1 2 3 4 5 6 7 8 9 10
T im e - h ours

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.

Maximum Attained Temperatures (°C)

module 2.35 VPC 2.25 VPC
1 Top 33.5 28.0
2 34.0 28.5
3 34.5 29.0
4 Bottom 32.5 28.0

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 lected due its growing popularity with system design-

ers. This popularity is a result of its favorable foot-
Battery and power system designers for current and print, and easy access to the terminations. In addition,
future telecommunications networks are faced with the packing or spacing of these types of batteries into
multiple choices. These choices such as power standard 19” or 23” rack/cabinet widths is also of
density, system life and safety, and cost are often concern to designers. Currently the IEC 896-2 stan-
competing. Understanding the impact of their dard for valve regulated lead acid batteries recom-
selections is often difficult due to the variety of mends a minimum gap of 5-10mm between batteries
conditions to which each application is exposed, and a when housed in a rack or cabinet.
lack of sufficient data to describe the outcomes of
their decisions. This is particularly the case with the Due to laboratory equipment constraints the batteries
batteries that are deployed in these systems. A clean were configured as two separate strings with 8 blocks
synthesis relating VRLA behavior to these choices of per string, thus yielding a 96 volt 100 Ah string or a
manufacturer, design and operating conditions does total cabinet system of 96 volts and 200 Ah’s of ca-
not exist, making good decisions difficult. pacity. The heat generation from this arrangement
would be identical to that for a configuration consist-
It is the purpose of this paper to examine one test sce- ing of four parallel strings, 100 Ah’s each, at 48 volts.
nario and present the results in a format that leads to a All testing was conducted in a laboratory environment
useful conclusion. The study conducted, involved ex- where the ambient conditions were maintained at 25
amining the behavior of sixteen, Marathon 100 Ah +/- 2oC. Figure 1 below is a photograph of the basic
front terminal batteries, assembled into one cabinet test set up.
system. This type of front terminal battery was se-

11.1
 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

Fig. 1 Test Set-up cabinet
A total of nine tests were conducted on the system.
The purpose of each test was to study the thermal The data in Figure 2 displays the battery temperature
conditions inside the cabinet, as a function of battery variation within the open cabinet where the surround-
spacing and ventilation, during a recharge at 2.35 ing ambient temperature was maintained at 25±2°C.
VPC with a current limit of 35A/100 Ah. There were The shelf located in position 4 at the bottom of the
three battery spacing arrangements considered; 10mm, cabinet reached a maximum temperature of 36°C.
5mm, and 0mm (i.e. batteries in contact) and three Shelves 1 and 2 located at the top of the cabinet
ventilation types; open ventilation (essentially a four reached a maximum temperature of 41°C, and the
shelf rack), natural convection only (small openings in shelf located in position 3 above shelf 4 and below
the side of the closed cabinet), and forced ventilation shelves 1 and 2 reached a maximum temperature of
(a fan in the top of the cabinet). Each recharge deliv- 42°C. The battery temperatures at the locations in the
ered 110% of the Ah’s removed, following an 8 hour, center of the strings within the cabinet varied by 5-
12.5 amp discharge. Note that the battery shelves did 7°C depending on shelf location and were as much as
not have openings that would permit air flow between 17°C above the ambient temperature. The battery sys-
the layers. There also was not any rectification equip- tem was able to return to a steady state temperature
ment located inside the cabinet that would influence after 40 hours.
the overall thermal behavior. A summary of the test-
ing protocol is shown in Table 1 below.
Figure 2. Open Cabinet
Table 1 10mm Space/Convection Cooling
Test # Battery Spacing Ventilation
50
Temperature (C)

1 10mm Open cabinet 45

Shelf 4 Battery
2 10mm Closed cabinet – natural 40
Shelf 3 Battery
35 Shelf 2 Battery
convection Shelf 1 Battery
30

3 10mm Forced ventilation 25

4 5mm Forced ventilation 20

0 10 20 30 40 50 60 70

5 5mm Closed cabinet – natural Time (Hours)

convection
6 5mm Open cabinet
7 0mm Open cabinet
8 0mm Forced ventilation
2.2 Test #2 (10mm Spacing – Closed
9 0mm Closed cabinet – natural
cabinet natural convection cooling
convection

The data in figure 3 displays the battery temperature

The thermal mapping of the cabinet was done by variation within the closed cabinet, natural convection
mounting a total of 9 thermocouples inside, two per cooling, 10-mm battery spacing and an ambient tem-
string (one in the center and one on the outside) and perature of 25±2°C. Again the shelf located in posi-
one for the ambient inside the cabinet. Temperature tion 4 at the bottom was the coolest reaching a maxi-
data was sampled every 5 minutes during charging. mum temperature of 41°C. Shelves 1, 2 and 3 located
The nomenclature used to describe the temperature above shelf 4 reached 43°, 42.5°C and 43°C respec-
(T) at the measured locations within the cabinet tively. The battery temperatures at the locations in the
throughout this paper is related to the battery shelf (1- center of the strings within the cabinet varied by only

11.2
1.5-2°C depending on shelf location and were as much
as 18°C above the ambient temperature. The battery Peak Temperatures: Tests 1,2 & 3
system was able to return to a steady state temperature Table 2. 10 mmSpacing Test Peak Temperatures
after 60 hours. 1= Top Shelf 4= BottomShelf
T 1C T 1O T 2C T2O T3C T3O T4C T4O CABINET
Figure 3 Closed Cabinet Open Cabinet 41.0 33.0 41.0 34.5 42.0 34.0 36.0 34.0 26.5
10mm Space / Fan Blocked Off
Closed Cabinet 43.0 37.5 42.5 38.5 43.0 37.0 41.0 36.5 35.0
50
Cabinet Forced Air 32.0 28.5 37.0 31.5 35.5 30.5 34.0 28.0 25.5
45
Shelf 4 Battery
Temperature (C)

40
Shelf 3 Battery
Shelf 2 Battery
35 Shelf 1 Battery
30

25 26.5 35.0 25.5

20 41 33 43 37.5 32 28.5
0 10 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 hav-
ing forced air-ventilation and an outside ambient tem-
perature of 25±2°C. During this test, shelf 1 located
2.4 Tests #4, 5, and 6 (5mm spacing
at the top was the coolest reaching a maximum tem-
trials)
perature of 32°C. Shelves 2, 3 and 4 located below
shelf 1 reached 37°C, 35.5°C and 34°C respectively.
Table 3 (and the diagram) summarizes the data for the
The battery temperatures at the locations in the center
second iteration of experiments which investigated the
of the strings within the cabinet varied by only 2-5°C
effects of decreased battery spacing from 10mm to
depending on shelf location and were as much as 9°C
5mm with the three differing types of ventilation. In
above the ambient temperature. The battery system
the open cabinet experiment the battery temperatures
was able to return to a steady state temperature after
in the center of the strings varied by 0.5-13°C with a
24 hours. Tabular comparisons of these tests are
peak temperature of 42°C seen in the top shelf which
summarized in Table 2 (and the diagram).
was 17°C above the ambient temperature. The system
was able to return to a steady state temperature after
Figure 4. Closed Cabinet 40 hours.
10mm Space / Forced Cooling

50
In the closed cabinet experiment the battery tempera-
45
tures in the center of the strings varied by only 0.5-
Temperature (C)

40
Shelf 4 Battery 1.5°C with a peak temperature of 43.5°C in the top
35 Shelf 3 Battery
30
Shelf 2 Battery
Shelf 1 Battery
shelf which was 18.5°C above the ambient tempera-
25
ture. The system was able to return to a steady state
20
temperature after 48 hours.
0 10 20 30 40 50 60 70

Time (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.

11.3
Peak Temperatures: Test 4,5 & 6 Peak Temperatures: Test 7,8, & 9
Peak Temperatures 10 mmSpacing Table 4. 0 mmSpacing Test Peak Temperatures
1= TopShelf 4= BottomShelf 1= Top Shelf 4= BottomShelf
T1C T1O T2C T2O T3C T3O T4C T4O CABINET
Open Rack 41.0 33.0 41.0 34.5 42.0 34.0 36.0 34.0 26.5 T 1C T 1O T 2C T2O T3C T3O T4C T4O CABINET
Cabinet 43.0 37.5 42.5 38.5 43.0 37.0 41.0 36.5 35.0 Open Cabinet 42.0 31.5 41.5 32.5 41.5 32.5 40.5 27.5 26.0
Cabinet Forced Air 32.0 28.5 37.0 31.5 35.5 30.5 34.0 28.0 25.5 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

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) 3 Discussion of Results
Table 4 (and the diagram) summarizes the data for the The results of the testing described in the preceding
third iteration of experiments that investigated the ef- text can be further examined by breaking out, and
fects of no space between batteries with the three dif- looking at, the three main variables being evaluated;
ferent types of ventilation. In the open cabinet ex- spacing of blocks, type of ventilation, and string loca-
periment the battery temperatures in the center of the tion within the cabinet.
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 3.1 Spacing of blocks
able to return to a steady state temperature after 40
hours. Shown below are three tables (5, 6, and 7) illustrating
the comparison of temperature behavior as a function
In the closed cabinet experiment the battery tempera- of battery spacing with each of the three types of ven-
tures in the center of the strings interestingly only var- tilation. Again, the temperatures shown are maximums
ied again by 0.5-1.5°C with a peak temperature of and the same nomenclature is used as earlier. Shelf 1
44.5°C again in the center of the top string which was is at the top of the cabinet, shelf 4 is at the bottom, C
19.5°C above the ambient temperature. The system is in the center of the cabinet, O is on the outside of
returned to a steady state after 55 hours. the string, Cbnt is the cabinet ambient, and the stable
conditions are after the temperature has returned to
In the final experiment investigating the effect of no steady state.
spacing between batteries with forced air ventilation,
the battery temperatures varied by a very minor 0- Table 5 - Open cabinet (no sides, only top and bottom)
Stable Time to
0.5°C attaining a peak battery temperature of 42.0°C T1C T1O T2C T2O T3C T3O T4C T4O Cbnt Temp Stable
again in the center strings which were 17.0°C above
ambient. The system returned to a steady state after
10mm 41C 33C 41C 34.5C 42C 34C 36C 34C 26.5C 25C 40hrs
40 hours.
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

11.4
Table 6 - Closed cabinet – no ventilation 3.3 String location within the cabinet
Stable Time to
T1C T1O T2C T2O T3C T3O T4C T4O Cbnt Temp Stable
Temperature profiles within the cabinet follow a fairly
intuitive pattern. In the tests where no forced ventila-
10mm 43C 37.5C 42.5C 38.5C 43C 37C 41C 36.5C 35C 27C 60hrs tion was present one would anticipate that the string
5mm 43.5C 36.5C 42.5C 37C 43C 36.5C 42C 31.5C 32C 24C 48hrs
on the bottom of the cabinet would be the coolest and
0mm 44.5C 37C 44C 37.5C 44C 37C 43C 32.5C 33C 25C 55hrs
temperatures would steadily rise throughout the cabi-
net with the maximum temperatures seen in the top
Table 7 - Closed cabinet – forced ventilation
string. In virtually all tests this was the case. In the
Stable Time to
T1C T1O T2C T2O T3C T3O T4C T4O Cbnt Temp Stable
tests where forced ventilation was present the top
string of batteries, or those closest to the fan, typically
10mm32C 28.5C 37C 31.5C 35.5C 30.5C 34C 28C 25.5C 26C 24hrs ran the coolest.
5mm 37.5C 30.5C 41C 33C 41C 33.5C 40C 28.5C 26C 24C 30hrs
0mm 41.5C 33C 42C 33C 42C 34C 41.5C 30.5C 26.5C 26C 40hrs

4 Conclusions
The data clearly shows that spacing the batteries de-
creases the maximum temperature at the center of the The work that was done in this test yields several main
cabinet only when combined with forced air ventila- conclusions.
tion. This benefit was the greatest across all four lev- 1. Forced convection within a closed cabinet will
els of batteries when the spacing was increased from yield significant benefits by reducing maximum
5mm to 10mm. Here the spacing resulted in a 4-5oC temperature increases above ambient conditions.
reduction in maximum temperatures in the center of In this testing up to a 10oC decrease was realized.
the cabinet. 2. At least a 10mm spacing between blocks is re-
quired to realize the maximum benefits of forced
We can see that under conditions without forced venti- air ventilation.
lation (Tables 5 and 6), spacing distance does not sig- 3. In conditions where forced air ventilation is not
nificantly impact the maximum temperature (<2oC) or present battery spacing plays a limited role in re-
the time to stable conditions. There is however an ducing maximum battery temperatures.
anomaly in the outside temperature of the bottom 4. Under normal recharge profiles battery tempera-
string in the 10mm configuration where the tempera- tures internal to a cabinet will see anywhere from
ture ran approximately 5oC higher and where this en- a 7oC to 20oC rise above ambient conditions
tire cabinet configuration took longer to reach a stable (25+/-2oC). These temperatures will however re-
temperature. The authors do not have a solid theory turn to a stable ambient within 60 hours.
to propose as to why this condition occurred other 5. Battery strings located at the bottom of a cabinet
than possibly more efficient heat transfer was realized or nearest a fan will remain the coolest during
with the greater battery spacing. charging.

3.2 Ventilation 5 Recommendations for Future

Work
Tables 2, 3, and 4 more clearly show the impact of
ventilation. Under all spacing conditions a fully closed This testing draws some interesting conclusions but
cabinet with only natural convection and airflow will also raises several points for further investigation.
always result in higher (as much as 7%) peak tempera- Some of these are:
tures than if the cabinet was fully open or if forced air 1. Repeat of the work at elevated ambient tempera-
was applied. tures and under normal float conditions for ex-
tended periods of time.
With regards to forced ventilation, we concluded that 2. Variations in the initial charge current rate.
unless the battery spacing was at the 10mm level, sig- 3. Location of electronic equipment inside the cabi-
nificant reductions in battery temperatures were not net to monitor the effect of an additional heat
observed. Again, this is verified by examining the source.
data in Tables 5, 6, and 7 above. 4. The impact of using standard top terminal mon-
oblocs.
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: Recently, after nearly three years of continuous operation,

several cells were sampled at random from the battery and
Valve-Regulated Lead-Acid (VRLA) batteries continue to be examined. When these samples were taken, the battery’s
employed in a wide variety of applications for monitoring system indicated that the battery had been
telecommunications and Uninterruptible Power Supply (UPS).
maintained at about 75-85% state-of-charge over the entire
With the rapidly growing penetration of Internet services, the
requirements for standby power systems appear to be changing. time period and had received only three equalization charges.
For example, at last year’s INTELEC, high voltage standby This paper will review electrical testing conducted on the
power systems up to 300-vdc were discussed as alternatives to battery samples as well as the results of extensive physical
the traditional 48-volt power plant. At the same time, battery analyses performed on the battery materials and components.
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- 1.1 Changes in Telecommunications Power
line” charging or intermittent charge to the battery have been
proposed. Some of these techniques go against the widely Power strategies within the telecommunications industry
accepted rules of operation for batteries to achieve optimum are changing. As suppliers broaden their product offerings to
lifetime. Experience in the telecom industry with high voltage include Internet service, cable TV and other communications
systems and these charging scenarios is limited. However, GNB media, their power needs are moving away from the
has several years of experience in the installation and operation traditional 48-volt dc power plant. More of the equipment
of large VRLA battery systems that embody many of the power
used to provide these new services is powered at utility
management philosophies being proposed. Early results show
that positive grid corrosion is not accelerated and battery provided voltages. As a result, service providers are
performance is mantained even when the battery is operated at installing high voltage, ac power plants that, more and more,
a partial state-of-charge for long periods of time. resemble Uninterruptible Power Supplies (UPSs).

For example, at last year’s INTELEC, papers were

1. Background presented discussing high voltage power systems, up to 300-
volts, as alternatives to the traditional 48-volt power plant [2-
In 1996-97, GNB installed and commissioned a large 4]. Rather than operate two power plants, both a low voltage
VRLA battery system for a Battery Energy Storage System and a high voltage system, operators are considering
(BESS) at the island village of Metlakatla, Alaska [1]. The converting all of their equipment to operate at the higher
battery’s function is to stabilize the island community’s voltage levels. Experience in the telecom industry with high
power grid providing instantaneous power into the grid when voltage power plants is limited however, and there is concern
demand from a local sawmill is high, and absorbing excess about understanding the operational considerations associated
power from the grid to allow its hydroelectric generating with these high voltage systems.
units to operate under steady-state conditions. This nominal
756-volt system is capable of providing 1.4-MWh at about Valve-Regulated Lead-Acid (VRLA) batteries continue to
the battery’s 90-minute discharge rate. Because the battery is be the dominant battery technology that many of the
required to randomly accept power as well as to deliver telecommunications suppliers are installing into their new
power on demand to the utility grid, it was decided to and expanding power sites. In an attempt to improve VRLA
continuously operate the battery at between 70 and 90% battery reliability, to extend VRLA battery lifetime, and to
state-of-charge. Like some of the recently proposed overcome the sensitivity of VRLAs to charging conditions,
alternative charge regimes for telecommunications alternative charging regimes to the traditional “float”
installations, this battery’s operations raised several concerns operation of batteries are being proposed. Off-line charging
regarding long-term performance and life.

12.1
[5, 8] and intermittent charging [6] of the battery, for example, alone island electric utility, operates the BESS to supplement
have been suggested. its generating facilities. The primary generation source for
the utility is 4.9-MW of rain-fed hydroelectric capability. In
In addition, some telecommunications providers are addition to its residential customers, MP&L supplies power
looking for ways to better utilize their facilities and to a commercial cannery and cold storage facility, and to the
equipment, and to reduce their operating costs. At least one Annette Hemlock Mill, a commercial lumber mill operation.
major telecommunications company is considering using its The BESS is used to provide instantaneous power to the
standby battery banks for load-leveling and load-sharing utility system to satisfy the random instantaneous load
during periods when utility rates are high. These new battery demands of a log chipper at the mill without causing
operating options all appear inviting; however, experience brownouts or overvoltage conditions to the remainder of the
with operating batteries under these conditions in the utility’s customers.
telecommunications industry is, like operating high voltage
power plants, also limited. 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.
1.2 Battery Energy Storage Systems (BESS)
With the addition of the diesel generator, MP&L’s total
generating capacity came to just over 8-MW, twice the
For the past several years, GNB has been intimately
average base load on the utility. But to achieve reasonable
involved, in conjunction with Sandia National Laboratories,
efficiency for the diesel, a greater portion of the utility’s base
in developing the concept of battery energy storage as a way
load had to be shifted from the less expensive hydro
to supplement and improve the quality of power received
generation to the more expensive diesel. Even with the
from the utilities. In the past, battery energy storage systems
addition of the diesel, electrical frequency often drooped to
were used simply to supplement generation capacity for the
less than 57-Hz, and system voltage remained very erratic.
utility to meet demands during periods of high electrical
usage. Operation in this limited mode alone could not justify
Operation and maintenance costs for the diesel added to
the cost of a BESS.
the problem. Fuel cost was $360,000 to $400,000 per year.
Transporting 475,000 gallons per year of diesel fuel by ferry
The more recent uses of battery energy storage in utility
from the mainland, and then through pipe across the island
application have been to correct on-going power quality
increased both the environmental risk and the financial
issues, in addition to providing a reserve of energy for
burden to the community. Each fuel shipment required an
uninterrupted power supply, peak shaving and load leveling.
average cash outlay of $100,000 – a significant amount for a
By introducing these additional functions, a BESS can be a
small local utility. In addition, minor overhauls to the diesel
viable alternative to other power management solutions. To
cost $150,000 every three years; and major overhauls every
achieve these objectives however, it is necessary to construct
six years cost $250,000.
and operate the battery in an unconventional manner.
A techno/economic feasibility study was conducted by
First, the BESS has to be integrated with the utility power
GNB Technologies and General Electric Company with
feed so that its operation is seamless to the power user. This
assistance from Sandia National Laboratories that compared
requires that the battery system be at a relatively high
battery energy storage to other options using only the existing
voltage. Second, the battery would have to be operated in a
hydro and diesel units. The study indicated that a 1-MW,
partially discharged state, often for months without any
1.4MWh battery energy storage system (BESS) could provide
equalization recharge, so as to efficiently accept power from
the spinning reserve, frequency control, and power quality
the grid as well as to rapidly deliver power to the utility grid
improvements that Metlakatla needed. The study concluded
to instantaneously correct power quality issues. Unique as
that the cost of the BESS could be recovered within three
these operating conditions are, they bear a very strong
years based on operational cost savings alone.
resemblance to the conditions being considered by
telecommunications operators. Data from these BESS
batteries can help telcos evaluate and better understand these
2.1 The Metlakatla BESS Battery
high voltage battery systems.
The battery at the Metlakatla BESS facility (Figure 1)
consists of 378 GNB ABSOLYTE IIP 100A75 modules
2. Metlakatla’s BESS Opportunity
arranged in a single series-connected string providing the
The Metlakatla BESS is located in the community of system with its nominal 756-volt rating. The 100A75 cell has
Metlakatla on the Annette Island Reserve at the southern tip a nominal C/8 capacity rating of 3,600 Ampere-hours; its
of Alaska. Metlakatla Power and Light (MP&L), a stand- rating at the intended 90-minute discharge rate for this

12.2
application is approximately 2,000-Ah / 3.87-kWh. The A heater is provided to warm the facility during the colder
entire battery system is rated 1.4-MWH at a 1.0-MW winter months; however, only outside air is circulated for
discharge rate. Each 100A75 cell is comprised of three forced convection cooling.
individual 100A25 cells connected in parallel within the
cell’s modular container, thus providing a statistical Operation of the Metlakatla BESS battery started in
population base of 1,134 samples. February, 1997, and except for a few short periods when the
system has been purposely shut down for maintenance to
The battery connects to a General Electric power either the battery or the system’s electronic inverters, it has
conversion system (PCS), based on gate-turn-off (GTO) operated essentially continuously since its commissioning.
thyristors, that can support a continuous load of 800-kVA and Warranty on the battery is based on an 8-year service life.
pulse loads of up to 1200-kVA. The PCS allows bi-
directional power flow between the ac system and the battery 2.2 Typical Operation of the BESS Battery
in less than a quarter-cycle. A 900-kVA filter bank removes
harmonics and compensates the voltage of the electrical The BESS was designed to be connected continuously to
signal. The BESS connects to the MP&L grid at a 12.47-kV the MP&L grid. From a fully charged condition, the battery
substation. The battery is housed in a 40 x 70-ft steel Butler is first allowed to be discharged to about an 80% state-of-
building that sits on a concrete pad at the substation. An charge (SOC). After reaching that point, the battery is then
automatic generation control (AGC) system provides allowed to accept recharge from the grid when load demand
computerized control and dispatch of MP&L’s hydro and is less than the output of the hydroelectric units. The BESS
diesel units as well as the BESS for optimum efficiency. The PCS inverters draw power from the battery to instantaneously
AGC can be remotely accessed to monitor the status of the satisfy surge events on the grid. The BESS AGC computer
battery bank. monitors the current flowing out of or into the battery and
automatically adjusts the output of MP&L’s hydroelectric
The battery’s 378 cells are arranged in two back-to-back units to essentially maintain the battery at about its 80% SOC
rows, each row comprised of twelve stacks of ABSOLYTE point. The control algorithm assumes a 100% charge
modular trays eight high, separated by an aisle. The battery acceptance efficiency of the battery as it accepts charge at an
is positioned to minimize cable runs between rows of battery SOC less than 90%. The charge algorithm that is used to
stacks and to the power conversion equipment. Pilot cell and control the recharge of the battery during operation is
temperature measurements are made at locations strategically summarized in Table 1. The power limitations of the PCS
positioned throughout the battery bank. Air is circulated by a equipment itself is the only factor that limits battery
fan to maintain consistent temperatures within the building.

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 1.0-
MW discharge rate
at a nominal 756-
volts.

12.3
discharge current. Equalization charges are scheduled twice output from the battery of 745,735-Ah; total charge input to
each year. the battery was reported by the computer to be 751,468-Ah.
Four individual 100A25 cells were selected from various
TABLE 1 locations within the battery to represent positions where
BESS Recharge Control Algorithm
variations in temperature had been observed and recorded by
Step Mode Control Parameter Limit / Transition the battery monitoring system. The monitor system indicated
that the battery was at about 78-81% state-of-charge when the
1 Current 35A/100Ah to 2.25 vpc cells were sampled. The samples were purposely taken prior
2 Current 25A/100Ah to 2.32 vpc to an equalization charge in order to assess the accuracy of
3 Voltage 2.32 vpc 18A to 2A/100Ah
4 Voltage 2.25 vpc Continuous the monitoring system’s state-of-charge algorithm.
Eq Voltage 2.35 vpc 12 Hours
The open circuit voltage measured on the cells (2.089 –
2.099 volts) correlated well with the monitor’s approximation
The PCS provides both active and reactive power to of battery state-of-charge. Previous testing at GNB had
counter load swings created by the log chipper. The BESS shown that new ABSOLYTE cells allowed to self-discharge
sources watts and VArs when the system load jumps higher to an open circuit voltage of 2.09-volts were able to deliver
than the average, and sinks watts and VArs when the load 78.5% of their nominal 1-hour capacity rating. The internal
falls below the average. Because the BESS’ resultant net impedance of the cells was measured at an average value of
output is nearly zero, the batteries require little additional 262-µohms; the nominal impedance for this size cell is 229-
charging. When required, the AGC dispatches the hydro µohms.
units to provide the minimal overcharge the battery requires.
Operation of the battery is shown in Figure 2, which shows After the sample cells were removed from the battery
the printout of the battery system screen for a typical day. 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
3. Sampling Cells from the Metlakatla BESS BESS facility were used as replacements for the sampled
cells. After arriving at the laboratory, the open circuit
In October 1999 (approximately 32 months after system voltage, impedance and weight of each of the sample cells
start-up), GNB and Sandia conducted a planned surveillance was recorded. No significant change in open circuit voltage
sampling of cells from the Metlakatla BESS battery. When or internal impedance occurred during transit, and cell
the cells were sampled, the AGC computer indicated a total weights were within the accepted tolerance range for this size

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” a telecommunications application, and result in a permanent
condition for teardown, visual inspection and chemical reduction in the battery’s capacity.
analyses. The remaining three cells were reassembled into
appropriately sized module trays for further electrical testing On the next several charge / discharge cycles, the amount
and characterization. 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
4. Electrical Characterization of Sampled Cells this way it would be possible to determine the efficiency of
charge acceptance of the battery even when the amount of
The three cells sampled from the Metlakatla BESS were available recharge was limited. The actual C/8 discharge
connected together in series to form a 6-volt battery. The capacity as well as the percentage increase over the previous
battery was then discharged, without any refreshening or cycle is summarized in the following table where the
boost charge, at 150-Amps, its nominal C/8 rate. On this recharge ampere-hours have been intentionally limited.
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 TABLE 2
Charge Acceptance Test Results
less than the residual capacity estimated to be available from
the cells by the BESS monitoring system. However, since the Discharge # Discharge Ah % Rated Recharge Ah Capacity Increase
monitoring system estimates battery state-of-charge simply 1 766.5 63.9 858.6 -
by summing ampere-hours discharged and charged, some 2 856.9 71.4 959.8 111.8%
error is to be anticipated especially considering the 6-month 3 945.0 78.8 1,058.5 110.3%
4 1,023.0 85.3 N/A 108.3%
interval over which the estimate was made. The monitor
resets to 100% state-of-charge following an equalization
charge. 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
2.15
of the cells essentially being equivalent to the amount of
2.10 overcharge provided. As the cells’ capacity increases and
Cell # 1 approaches a fully charged state, overcharge acceptance
2.05 Cell # 2 efficiency starts to decrease as might be expected.
2.00
Cell # 3 Interestingly however, these experiments demonstrate an
almost 100% ampere-hour charge acceptance efficiency for
VPC

1.95 these samples even after having been operated in a partially

discharged state for over 6-months. It is important to
1.90
appreciate that the recharge voltage during these experiments
1.85 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
1.80 from a typical recharge where recharge is limited to 112% of
1.75
the capacity discharge is shown in Figure 4.
0 1 2 3 4 5 6
TIME (HOURS) 4.2 Discharge Performance after Equalization Charge

Figure 3: Cells sampled from the Metlakatla BESS Following the last of the four “limited recharge” cycles, the
delivered 64% of their rated capacity without any sample cells were given a standard equalization charge in
boost charge after being operated in a partially
discharged state for over 6 months.
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
4.1 Charge Acceptance Test capacity to a cutoff of 1.75 volts per cell. The samples were
subjected to four additional discharge cycles, and the battery
The next objective was to determine if operating the continued to deliver, on average, 101.9% of its rating.
battery in a partially discharged state for extended periods
had caused any permanent deterioration of the battery. It is In addition to these capacity discharges, the sampled cells
widely thought that failing to adequately recharge a lead-acid were subjected to a series of discharges to verify capacity
battery can cause “hard” sulfate to form within the cell’s conformance at various discharge rates. Average compliance
active materials. This would hinder recharge acceptance, to published specifications for this size cell was 105.6% for
especially at low charge voltages as might be encountered in discharge rates ranging from C/1 to C/24.

12.5
 Typical Limited Amp-Hr Recharge of Metlakatla Cells
cell had not been changed as the result of operating in a
partially discharged state for such an extended period.
Amps % Rchg Avg VPC

450 2.38
4.4 Total Capacity Discharged
400 2.36
Current (Amps) / % Recharge

350 2.34 Based on cycle life testing conducted at GNB, the data

Cell Voltage (vpc)

300 2.32 suggests that each unique battery design has a certain lifetime
250 2.30 discharge “throughput” that is somewhat independent of
discharge rate or depth of discharge [7]. For the ABSOLYTE
200 2.28
IIP design this discharge throughput is equal to
150 2.26
approximately 1,000 times the nominal C/8 capacity of the
100 2.24 cell in ampere-hours; approximately 3.6 million Ah for the
50 2.22 100A75 configuration installed at Metlakatla.
0 2.20
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 In order to maintain the power quality of MP&L’s grid,
Time (Hours) 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
Figure 4: Data recorded during a limited ampere-hour
recharge of the Metlakatla sample cells shows their utility. The BESS monitoring system continuously monitors
ability to readily accept high levels of charge current and integrates current flow into and out of the battery string.
at relatively low charge voltages. 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
These discharge data indicate that the cells sampled from of “cycling” the cells have experienced and to estimate the
the Metlakatla BESS after over 30-months of operation, most discharge “throughput” for the cells at the time they were
of which was continuously cycling between about 70 and sampled as being approximately 21% of its lifetime
90% state-of-charge, met or exceeded all of the performance capability. This information will be helpful when the
specifications for this size cell. Based on these electrical condition of the cell plates is examined.
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. 5. Internal Examination of the BESS Cells

4.3 Float Behavior of Sampled BESS Cells Two of the sampled cells were torn down for physical and
chemical analyses – the cell retained in its “as received”
One further electrical test was conducted on the cells condition, and one of the cells that had completed the
sampled from the Metlakatla BESS; that being a float charge electrical testing described previously. Both cells were
test to establish the cells’ float current over the range of float examined at the same time to allow visual comparisons to be
voltages recommended for the ABSOLYTE IIP design. made. Sandia personnel assisted during these examinations.

Overall Observations. Both cells appeared “normal” after

TABLE 3 the covers were removed. The cells were tightly compressed
Float Current Behavior at Float Voltage within the cell jar. All plates were completely encapsulated
by the glass mat separator. There was no evidence
Float Voltage Float Current
( vpc ) ( mA/100Ah)
whatsoever of any strap, terminal post or plate lug corrosion.
There was no free liquid electrolyte in either of the cell jars
2.22 20 once the cell elements were removed.
2.25 55
2.28 117
Negative Plates. Negative plates sampled from three
Specification at 2.25vpc is 45 – 55 mA/100Ah 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
The float current at the various voltages tested was exactly “halo” of liquid electrolyte around the finger being pressed
as would be predicted for this size cell after having stabilized onto the plate. Negatives from the cycled cell exhibited a
in a “pure” float charge application, suggesting that the shiny metallic streak when tested by striking across the
surface properties and morphology of the electrodes in the surface of the plate with a hard object. Plates from the “as

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’ Figure 6
plates was firm and crispy.
Cross section of a positive grid from the “as received”
cell sampled from the Metlakatla BESS after 30
X-ray diffraction analyses of the active materials indicated months of operation in a partially discharged state
greater than 89% PbO2 for the cycled cell and 66% for the “as shows a minimal amount of corrosion.
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 Measurement of the corrosion layers indicate a corrosion
negative and the positive active materials were what should thickness of approximately 0.13 – 0.18mm. Compared to the
be expected for cells considering their operational history and corrosion rate basis of 0.08mm per year used to determine
treatment prior to analysis. the design life for the ABSOLYTE IIP product, the actual
corrosion rate for these cells was 0.05 to 0.07 mm/yr. These
Positive Grids. Samples of the positive grids from both measurements indicate that the rate of positive grid corrosion
cells were taken, cross sectioned and polished to determine with the Metlakatla operating and charge regime is less than
the amount of corrosion that had occurred over the lifetime of that experienced under pure “float” conditions.
these cells. Metallurgical photos of cross-sectioned samples
of these grids are shown in Figures 5 and 6. 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.
Figure 5
Concentration of the sulfuric acid electrolyte solution from
Cross section of the positive grid sampled from a cycle the cycled and charged cell averaged at 1.309 s.g. Variation
tested cell from the Metlakatla BESS shows a minimal in concentration between the “top” section of the separator
amount of corrosion.

12.7
and the “bottom” section was approximately 0.006 specific The negative plate lugs are all firmly embedded within the
gravity units. The measured concentration is at the design strap, forming continuously bonded connections. There were
concentration for this type cell and indicates that there has no signs of any oxidation corrosion on the negative strap
been no loss of water from the cell that would have caused material itself. Similarly, there no signs of any corrosion of
the electrolyte concentration to be increased. The the negative terminal post material or the cover seal bushing.
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 electrolyte’s 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 Figure 8
stability and corrosion resistance of the internal lead busbars,
Cross section of a negative terminal post from a
straps and terminal posts of VRLA designs. These concerns Metlakatla BESS cell shows no corrosive attack.
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 6. Summary and Conclusions
corrosion.
It has long been held that to achieve optimum life and
Samples of the negative plate busbar strap and the negative performance from a lead-acid battery, it is necessary to float
terminal post were taken, cross sectioned and polished to the battery under rigid voltage conditions to overcome self-
determine if these internal lead parts were experiencing discharge reactions while minimizing overcharge and
unusual corrosion under the operating conditions for the corrosion of the cell’s positive grid. This has resulted in
Metlakatla BESS. A section of the negative plate strap is batteries being used in telecommunications applications
shown in Figure 7; a cross section of one of the cell’s strictly in a standby mode. This may have been acceptable
negative terminal posts is shown in Figure 8. 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
Figure 7 of VRLA cells in this type of use. As telecommunications
power requirements change, it is conceivable that battery
Cross section of a negative plate strap from one of the power plants in the telecommunications industry will take on
Metlakatla BESS battery samples shows no indication
of corrosion or oxidation that could reduce life. 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 8. References
cost.
[1] Miller, N. W., et al., “A VRLA Battery Energy
Detailed examination of cells sampled from the battery Storage System for Metlakatla, Alaska”
system at the Metlakatla BESS after over 30 months of Proceedings of the 11th Annual Battery Conference
operation showed no unusual conditions that would signal an on Applications and Advances, Long Beach, CA
early degradation of the cell’s components. Positive and 1996.
negative active materials composition was consistent with the
state of charge of the cell when sampled. Active material [2] Marquet, D., et al. “New Power Supply for New
structure was essentially in “as new” condition. Positive grid Telecom Networks and Services” Proceedings of
corrosion as evidenced by metallurgical examination and Intelec 99, Copenhagen.
dimensional measurements to assess growth appeared to be
even less than what one would expect in a perfectly [3] Eklund, S. & Montin, S. “Custom Designed Power
controlled float charge environment. The degree of wetness Supply DC – Applications for the Telecom
of the cell’s separator materials was appropriate for this Industry” Proceedings of Intelec 99, Copenhagen.
design cell and there was absolutely no indications of
electrolyte stratification or concentration variations that [4] Akerlund, J. “48V DC Computer Equipment
would suggest excessive self discharge of the cell or loss of Topology – An Emerging Technology”
water from the cell. Furthermore all hard lead components Proceedings of Intelec 98, San Francisco.
within the cell (i.e., straps, plate lugs and terminal posts) were
in pristine condition showing no evidence whatsoever of any [5] Jones, R., et al. “Recharging VRLA Batteries for
unusual corrosive attack. Maximum Life” Proceedings of Intelec 98, San
Francisco.
Overall the condition of the cell could be described as
“unremarkable”. For those in the telecommunications [6] Sideris, T., et al. “Battery Aging and the Case for
industry who are considering broadening the scope of Stopping Float Charging” Proceedings of Intelec
operation of their battery systems to possibly supplement and 99, Copenhagen.
enhance the quality their high voltage power supply, these
observations provide encouragement. Admittedly, the data [7] Deshpande, S., et al. “Intelligent Monitoring
and observations discussed in this paper is but a single point. System Satisfies Customer Needs for Continuous
However, it is GNB’s and Sandia’s plan to continue the Monitoring and Assurance on VRLA Batteries”
surveillance of the Metlakatla battery both by continually Proceedings of Intelec 99, Copenhagen.
monitoring its electrical performance during operation and to
further sample cells from the battery throughout its lifetime. [8] Kakalec, R. J. & Kimsey, T. H. “A New Battery
Plant Configuration that Eliminates Thermal
The telecommunications industry requires this information Runaway in Valve Regulated Lead-Acid Batteries”
so that it can make enlightened decisions about how best to Proceedings of Intelec 2000, Phoenix.
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 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-AC04-
94AL85000.

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 Industry Leader in Network Power…

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
Büdingen, Germany
TEL: 49.6042.8170
FAX: 49.6042.81233

MIDDLE EAST/AFRICA
Exide Technologies
Abu Dhabi, U.A.E. The Network Power Division of Exide Technologies is the global leader in stored
TEL: 971.2.226235 electrical energy solutions for all major critical reserve power applications and needs.
FAX: 971.2.227644
Such network power applications include communication/data networks, UPS systems
JAPAN for computers and control systems, and electrical power generation and distribution
GNB Industrial Power Japan
Tokyo, Japan/Pacific Rim systems. With a strong manufacturing base in both North America and Europe and a
TEL: 81.3.5325.6281 truly global reach (operations in greater than 80 countries) in sales and service, the
FAX: 81.3.5325.2063 Network Power Division has all of the tools necessary to satisfy your power needs.
AUSTRALIA/NEW ZEALAND
Exide Technologies
Padstow, N.S.W. Australia
TEL: 61.2.9722.5700
Global Brands…
FAX: 61.2.9774.2966

SOUTH EAST ASIA

Exide Technologies S.E. Asia
Singapore GNB Flooded
TEL: 65.546.2866
FAX: 65.546.2966

CHINA Based on over 100 years of technological innovation the Network Power Division
Exide Technologies
Hong Kong, China continues to lead the industry with such recognized global brands as Absolyte,
TEL: 852.3106.2668 Sonnenschein, Marathon, Sprinter, and Flooded Classic. These products and brands
FAX: 852.3106.0260
are synonymous with quality, reliability, performance and excellence in all markets
Exide Technologies served.
Beijing, China
RE
AD

TU

TEL: 86.10.6510.2910
LE

RN

FAX: 86.10.6510.2912

LATIN AMERICA
Total Battery Management… RECYCLE

GNB Industrial Power

Lombard, Illinois U.S.A. In addition to being the leader in delivering premium products to the market, Exide
TEL: 1.630.629.5200 Technologies takes pride in its commitment to the environment. As part of a
FAX: 1.630.629.2635
complete approach to manufacturing, distributing, and recycling lead acid batteries,
INDIA the Total Battery Management program has been developed to ensure a safe and
GNB Industrial Power
Bangalore, India responsible life cycle for all of our products.
TEL: 91.80.550.0581
FAX: 91.80.550.0582

www.gnb.com

GB-3819 REV 4/02

Printed on recycled paper.