A White Paper from the Experts
in Business-Critical Continuity™
Battery Optimization Services
A guide to optimize your battery maintenance.
by Peter Shore 1, Emerson Network Power and Geoffrey May 2, FOCUS Consulting
Executive Summary
All telecommunications nodes, from highly critical core sites to
remote access sites, rely on standby batteries to keep telecom
electronic systems up and running when the utility supply
either fails or suffers from other disturbances. As batteries are
subject to normal and abnormal aging and even occasional
manufacturing defects, a battery testing and maintenance
program is vital to ensure reliability. Several on-site and remote
battery testing techniques are available, each with different costs
as well as technical strengths and weaknesses. By combining
these different techniques and deploying them throughout the
battery life cycle in a dynamic fashion – taking site criticality,
age, battery size and previous test results into consideration – it
is possible to design an optimized maintenance regime that will
deliver improved battery reliability for each testing dollar spent.
Table of Contents
0.0Introduction...........................4
1.0Background............................4
2.0 Preventative, Predictive
and Corrective Battery
Maintenance...........................8
3.0 Battery Maintenance
Strategies...............................8
4.0 Battery Testing.......................9
5.0 Examples of Actual
Measurements......................15
6.0 Battery Maintenance............22
7.0 Optimized Corrective
Maintenance.........................31
8.0Conclusions..........................33
9.0References............................33
10.0Annex...................................34
For more information visit:
EmersonNetworkPower.eu/EnergySystems
1
P eter Shore is Service Product Manager for Emerson Network Power Telecom Services. He has over 30
years of experience in customer service and fault diagnostics in the electronics industry. Since 2001,
Peter has worked with Emerson on establishing battery service techniques for UPS and DC power
systems, as well as developing product enhancements that enable more effective diagnostics.
2
G
eoffrey May is a chartered engineer who obtained his primary and secondary degrees from the University
of Cambridge. He was Group Director of Technology for Hawker Batteries/BTR Power Systems (now EnerSys)
from 1991 to 2000 when he joined FIAMM as Chief Technology Officer. In 2003, he established FOCUS
Consulting to provide technical and business development services to battery users and manufacturers.
EN310TRA-BatOpt / 0113
1
Battery Optimization Services
Table of Contents
Introduction............................................................................................................................................ 4
1.0Background.............................................................................................................................. 4
1.1
Standby Battery Types.............................................................................................................. 5
1.2
Battery Failure Modes............................................................................................................... 5
1.2.1
Positive Grid Corrosion............................................................................................................. 5
1.2.2
Positive Grid Growth................................................................................................................ 6
1.2.3Sulfation................................................................................................................................... 6
1.2.4
Active Material Softening......................................................................................................... 6
1.2.5
Dry Out.................................................................................................................................... 6
1.2.6
Pillar Seal Leakage.................................................................................................................... 7
1.2.7
Lid Seal Leakage....................................................................................................................... 7
1.2.8Vents........................................................................................................................................ 7
1.2.9
Mechanical Damage................................................................................................................. 7
1.2.10 Group Bar Corrosion................................................................................................................. 7
1.2.11 Internal Shorts.......................................................................................................................... 7
2.0
Preventative, Predictive and Corrective Battery Maintenance.................................................. 8
2.1Preventative............................................................................................................................. 8
2.2Predictive................................................................................................................................. 8
2.3Corrective................................................................................................................................ 8
3.0
Battery Maintenance Strategies............................................................................................... 8
3.1
Replace on Age......................................................................................................................... 8
3.2
Replace When Faulty................................................................................................................ 8
3.3
Replace Based on Known Condition......................................................................................... 9
4.0
Battery Testing......................................................................................................................... 9
4.1
Battery Performance or Capacity Measurements..................................................................... 9
4.1.1
Full Discharge on External Load................................................................................................ 9
4.1.2
Partial Discharge On-Site or System Load.............................................................................. 10
4.2
Battery Condition or State-of-Health Testing......................................................................... 10
4.2.1
Float Voltage.......................................................................................................................... 10
4.2.2
Float Current and Temperature.............................................................................................. 11
4.2.3
Internal Ohmic Measurements............................................................................................... 11
4.2.3.1 Equivalent Circuit of a Battery................................................................................................ 11
4.2.3.2 Impedance Spectroscopy....................................................................................................... 12
4.2.3.3 AC Ohmic Techniques............................................................................................................ 12
4.2.3.4 DC Ohmic Techniques............................................................................................................ 12
EN310TRA-BatOpt / 0113
2
Battery Optimization Services
Table of Contents (continued)
4.2.8
IEC 60896-21: Manufacturers’ Test Method for Internal Resistance....................................... 13
4.2.9
Correlation Between Internal Resistance and Capacity.......................................................... 13
4.2.10 Baseline Reference Values...................................................................................................... 13
4.2.11 Measurement of Battery State-of-Health with Internal Resistance......................................... 13
4.2.12 Battery Life Cycle.................................................................................................................... 14
5.0
Examples of Actual Measurements......................................................................................... 15
5.1
Internal Resistance and Discharge Testing............................................................................. 15
5.2
Trend Analysis for Internal Resistance Measurements............................................................ 17
6.0
Battery Maintenance.............................................................................................................. 22
6.1
Key Requirements for a Battery Maintenance Program.......................................................... 22
6.1.1
Ideal Battery Maintenance...................................................................................................... 22
6.1.2
On-Site Discharge: Full Capacity Testing with an External Load.............................................. 23
6.1.3
On-Site Discharge: Partial Discharge Testing with the On-Site Load...................................... 24
6.1.4
On-Site Internal Resistance Testing........................................................................................ 25
6.1.5
Remote Testing: Partial Capacity Testing with the On-Site Load............................................ 26
6.1.6
Remote Testing with Dedicated Hardware: Partial Discharge and Internal Resistance........... 27
6.2
Battery Optimization: Selecting the Best Program to Suit Particular Battery Maintenance
Requirements......................................................................................................................... 28
6.2.1
Proposed Solution for Access Level of a Network (less critical nodes).................................... 28
6.2.2
Proposed Solution for Medium Critical Nodes........................................................................ 29
6.2.3
Proposed Solution for Medium to High Critical Nodes........................................................... 29
6.2.4
Proposed Solution for High Critical Nodes............................................................................. 30
7.0
Optimized Corrective Maintenance........................................................................................ 31
7.1
Replacement Policy for Single Cells or Monoblocs................................................................. 31
7.2
Connectors and Torque Settings............................................................................................ 31
7.3
Cell Leakage........................................................................................................................... 32
7.4Cleanliness............................................................................................................................. 32
8.0Conclusions............................................................................................................................ 33
9.0References.............................................................................................................................. 33
10.0Annex..................................................................................................................................... 34
10.1
Design Life.............................................................................................................................. 34
10.2
Cycle Life................................................................................................................................ 34
10.3
Service Life/End-of-Life........................................................................................................... 34
3
EN310TRA-BatOpt / 0113
Battery Optimization Services
Introduction
It is vital that batteries are maintained to ensure reliability.
However, cost pressures may dictate that the procedures used
to measure battery condition are either inadequate or infrequent, compromising the ability of the battery to support the
load. Therefore, understanding the key failure modes of standby
batteries, how to measure and estimate battery condition, the
different strategies for combining testing methods, and the
advantages and disadvantages of these methodologies will help
to determine the optimal battery strategy for the network.
1.0Background
All telecommunications networks, from
core central office sites to access sites,
rely on standby batteries to keep the
network running when the utility supply either fails or suffers from other
disturbances. More critical sites require
large batteries (>1000 Ah) configured to
provide autonomy times of an hour or
more. In some locations where legislation requires the system to be supported
through extended power outages in
emergency situations, the discharge
duration may be several hours. The
investment in batteries for these sites
can be very large, yet often the maintenance budget is not in proportion to the
battery cost or to the value of potential
lost business. In remote locations many
telecom sites are unmanned and subject
to higher variations in the local utility
supply; consequently, problems can go
undetected until the battery fails completely. The nature of cell sites is such
that overlapping coverage provides a
degree of redundancy and typical site
EN310TRA-BatOpt / 0113
traffic is relatively low volume and low
cost. However, the nature of network
traffic is changing rapidly, with mobile
data communications requiring more
and more bandwidth and users requiring better reliability for their connected
services. Consumers are less tolerant to
lost communication as we are becoming
more dependent on it in our daily life.
The reliability of the network depends
upon the integrity of the power provision
and the availability of reliable standby
batteries for secure power. Capital budgets (CAPEX) for batteries are a major
concern, and it is important to get the
maximum service life from the battery.
However, batteries must still be replaced
before they fail so that the network is
not left unprotected. Similarly, routine
operational expenditure budgets (OPEX)
need to be controlled and the costs of
battery inspection and maintenance
have to be balanced against the larger
costs in lost revenue and reputation if the
network is unreliable.
4
Battery Optimization Services
BATTERY TYPE
SEPARATOR / ELECTROLYTE
PLATE CONSTRUCTION
AGM (Absorptive Glass Mat)
Flat Pasted Plates
tion of 80% is the industry accepted limit
for end-of-life. Below 80% the retained
capacity falls off rapidly, and it is not
safe to assume there is any significant
capacity for any period beyond a few
months after this limit is reached.
VRLA
Valve Regulated Lead Acid
Tubular (OPzV)
GEL
Flat Pasted Plates
VLA (Flooded)
Tubular (OPzS)
Free Electrolyte
Vented Lead Acid
Flat Pasted Plates
Figure 1. Standby Battery Types
1.1 Standby Battery Types
Lead-acid standby batteries are either
valve-regulated lead-acid (VRLA) or
vented lead-acid (VLA) [Figure 1], also
referred to as flooded or wet types [1].
VRLA batteries come in two distinct
types; gel or absorptive glass mat (AGM).
They both use the same principle of
oxygen recombination to reduce water
loss in floating service, either by using an
electrolyte that is made into a gel with
silica or a separator in the form of an
AGM which immobilizes the electrolyte.
Gel batteries may have tubular positive
plates (OPzV) or flat pasted plates. AGM
batteries have flat pasted plates. VLA
(Flooded) batteries may also have flat
pasted plates or tubular plates (OPzS).
For telecommunications service in
Europe, the vast majority of batteries
installed are VRLA types, mostly with
AGM separators, but OPzV types are
used in some circumstances. In North
America, large flat pasted plate flooded
batteries are most commonly used in
central office locations, OPzV and OPzS
types are rarely used, but the majority of batteries used across networks
are VRLA types with AGM separators.
5
Both VLA and VRLA battery types require
maintenance for reliable operation and
optimized service life. The main difference between VRLA and VLA with
respect to maintenance requirements is
that whilst VLA have transparent cases
and ‘free’ electrolyte, allowing routine
visual inspections and specific gravity
(SG) tests, the VRLA types have sealed
opaque cases and ‘fixed’ electrolyte
such that internal ‘visual’ inspections
and SG tests on electrolyte are no
longer possible. Therefore alternative
maintenance techniques need to be
developed for testing VRLA batteries.
The following maintenance procedures
primarily refer to VRLA batteries for
telecommunications service, although
many of the techniques are also suitable for VLA batteries used in either
UPS or telecom applications.
1.2 Battery Failure Modes
For standby service with a reasonably
reliable public electricity supply, batteries are designed such that corrosion
of the positive grid becomes the life
limiting parameter. Cycle life should not
be life limiting. At end-of-life, battery
capacity declines and capacity reten-
The principal causes of battery failure are:
1.2.1 Positive Grid Corrosion
The positive grid is polarized at the float
voltage, immersed in sulfuric acid, and will
corrode throughout the life of the battery.
The grid alloy, generally lead-calcium-tin
or lead-tin, is selected to have a high corrosion resistance, and the grid thickness
and other grid design parameters are
selected to provide sufficient grid metal
for the expected life of the battery. Grid
corrosion is accelerated by higher applied
voltages [Figure 2]. Typically, the float
voltage is set to maintain a fully charged
battery without excessive water loss,
and corrosion is kept at a level to achieve
New Grid
Aged Grid
Figure 2. Corrosion of Aged Grid
design life. Correct setting of the float
voltage is essential. Grid corrosion is
also sensitive to temperature such that
temperature compensation of the float
voltage will have a favorable effect. Grid
resistance increases during the life of the
battery, accelerating towards end-of-life.
There will also be some loss of connectivity between the grid and active material.
EN310TRA-BatOpt / 0113
Battery Optimization Services
Figure 3. Cracked Casing Around Pillar Seal
1.2.2 Positive Grid Growth
Changes in active material volume
and the volume of the corrosion product place stresses on the grids, which
become distorted at end-of-life. Some
battery designs can accommodate this
harmlessly, but it can lead to pressure on
the pillar seal, which may be distorted.
The case may become cracked [Figure 3]
or start to leak. Internally, this will cause
loss of connectivity between the grid and
the active material, increasing internal
resistance (IR) and reducing capacity.
1.2.3Sulfation
The normal discharge product on both
the positive and negative plates is lead
sulfate [Figure 4]. This is normally very
finely divided and easy to recover by
recharge, but over time and on cycling
it tends to coarsen and becomes more
difficult to recharge, eventually leading to capacity loss. There will be an
increase in the internal resistance of the
EN310TRA-BatOpt / 0113
cell. Sulfation will be increased if the battery is left in a partially or fully discharged
state for extended periods. In normal
operation, it is important to maintain
the correct float voltage and to apply
temperature compensation as recommended by the battery supplier. Shallow
cycling will not reduce sulfation as compared to leaving the battery on float.
1.2.4 Active Material Softening
Over time the active materials tend to
become softer and less cohesive. This will
be more rapid if the battery is regularly
cycled. This is associated with capacity
loss and changes in internal resistance.
1.2.5 Dry Out
At end-of-life or in certain fault conditions,
such as excessive float voltage, there will
be a loss of water from the cells, which will
lead to shrinking of the separator material, loss of compression, and reduced
contact between plates and separator.
These conditions result in loss of capacity
and increase in internal resistance. Higher
service temperatures also increase the
risk of drying out. If dry out occurs more
Figure 5. Cell Dry Out
rapidly, for example as a result of a high
applied voltage, it can lead to thermal
runaway. Cells are designed such that
dry out is not a normal failure mode. Dry
out is normally the result of a mechanical defect, high ambient temperature,
incorrect charging or a combination of these contributing factors.
Figure 5 shows a bloc with two cells,
separated electrically and physically
within the same case. The left hand cell
has dried out causing the white AGM
separator to shrink (gaps between AGM
now showing the top of the plates).
The discoloration on the right hand
AGM is due to copper contamination
of the electrolyte from the brass pillar inserts when the battery was cut
open. The lack of discoloration on the
left hand cell also illustrates the lack of
electrolyte due to dry out. This fault
was detected using Ohmic techniques.
Figure 4. Lead Sulfate Deposits
6
Battery Optimization Services
Figure 7. Vent Plugs Come in Different
Shapes and Sizes
Figure 6. Leak in the Pillar Seal
1.2.6 Pillar Seal Leakage
Pillar seals may develop leaks in service
[Figure 6]. This is usually a manufacturing defect and should be detected by
visual inspection. However, it might not
be detected by electrical measurements
until performance has deteriorated due
to dry out or corrosion caused by the
leaking electrolyte. Leakage and other
contaminants on the surface of a battery can lead to ground fault currents,
which in turn can lead to overheating and
thermal runaway. For this reason visual
inspection is a necessary complement
to electrical measurement and testing.
accelerated dry out, loss of electrolyte
or both. Some minor deposits around
a vent are harmless and normal, especially in the recombination ‘settling
in’ period when excess water is being
vented off. Excessive crystallized deposits, or liquid electrolyte may indicate
defects, overcharging or other defects.
1.2.9 Mechanical Damage
Cracks in the casing can lead to loss of
electrolyte by evaporation or direct
leakage of electrolyte [Figure 8]. Loss of
electrolyte will lead to dry out and loss
of capacity. Whatever the cause of a
leak, it needs to be discovered and corrected before it causes extensive damage
or even a complete battery failure.
1.2.7 Lid Seal Leakage
Lid seals may develop leaks, and this
is usually a manufacturing defect. As
for pillar seal problems, it will readily be seen visually but may not be
detected by electrical measurements.
1.2.8Vents
If the vents are missing, are ejected
or fail to close [Figure 7], there will be
a rapid run down in capacity due to
7
1.2.10 Group Bar Corrosion
Incorrect selection of the alloy for the
group bar can lead to rapid corrosion
of the negative group bar and to plates
becoming detached. At end-of-life
there will be corrosion of the positive
group bar and the pillar, which will
increase internal resistance. Improved
manufacturing techniques and correct
materials selection have substantially
eliminated this problem. Poor connections between the plate lugs and the
group bar, which can arise in manufacturing, also lead to higher internal resistance
and possibly, premature failure.
1.2.11 Internal Shorts
These can arise from damage or penetration of the separator. They may be
hard shorts leading to rapid failure or
soft shorts where the defect is relatively
minor and may not be initially apparent
as they can develop over time. Internal
shorts will cause loss of capacity by
self-discharge or a higher float current*;
although the latter may not be readily apparent in a series string, but the
affected cell voltage in a string may differ
from other cells. The best test for internal
shorts on installation is open circuit voltage. Any tendency toward internal shorts
will result in lower voltage; internal
resistance may be slightly lower as well.
* Float current will tend to increase towards endof-life or because of internal faults in the
battery, and there will be some correlation
with internal resistance measurements.
Normal maintenance will record float current
and track this over time.
Figure 8. Cracked Casing with Leakage
EN310TRA-BatOpt / 0113
Battery Optimization Services
2.0Preventative,
Predictive and Corrective
Battery Maintenance
2.2Predictive
3.1 Replace on Age
In order to realize the maximum benefits from any maintenance regime and
fully optimize the service life of any
standby battery, an approach combining
preventative with predictive and corrective procedures should be adopted.
These procedures measure changes
in battery condition and allow
trend analysis to predict the health
and life of the battery as well as
detecting faults. The various state
of health measurements are:
Many operators replace batteries on a
fixed timescale, regardless of the actual
condition of the batteries. Based on
their previous experience of failures,
they could, for example select a period
much less than the design life, such as
50% or every 5 years for a 10 year battery. This approach may at first seem
to be a good way to ensure battery
reliability, but healthy batteries may be
scrapped when there are still some years
of useful life or early undetected failures
may be present in the network giving
a completely false sense of security
about the available battery autonomy.
2.1Preventative
These procedures increase battery
reliability by taking action to prevent accelerated deterioration:
•Maintaining the correct temperature
•Maintaining the correct float voltage
•Checking and maintaining the correct
torque** settings on connectors
•Cleaning batteries to avoid the
possibility of current leakage paths
** It should be noted that many battery
manufacturers recommend two torque
settings in their manuals; an installation
torque value and a maintenance torque
value. The installation value is generally higher than the maintenance value
and too frequent resetting to the higher
value can cause damage to the posts.
3.0 Battery Maintenance
Strategies
•Internal resistance
•Float current
•Capacity (full or partial discharge)
•Intercell connector resistance
•Individual cell or monobloc
temperature
2.3Corrective
These procedures provide remedies
to faults or problems that have been
detected:
•Replacing single cells or
monoblocs in a string
•Replacing whole battery strings
•Stripping and cleaning batteries
affected by leakage, corrosion
or dirty conditions
•Stripping and cleaning connectors
with corrosion or high resistance
contacts to the battery terminals
•Re-installing connectors and links to
the correct installation torque setting
EN310TRA-BatOpt / 0113
3.2 Replace When Faulty
Maintenance strategies based on replacement only once batteries have failed
will lead to loss of service with resulting
loss of traffic. Limitations in operational
budgets may restrict the level of maintenance and inspection, but it may actually
take several return visits to investigate
the cause and remedy a problem. In
practice, and in order to limit the cost of
further diagnostic work, a whole string
of batteries may therefore be replaced
even if only one bloc is the actual cause
of the problem. Clearly there has to be
a balance between the cost of battery
maintenance, the cost of battery replacements, and the alternative cost of lost
traffic. Simply replacing the battery
when it has failed may not be the optimal
strategy or the best economical solution.
8
Battery Optimization Services
4.0 Battery Testing
3.3 Replace Based on Known Condition
Optimum battery reliability can only
be achieved with a good level of maintenance and inspection. Condition
assessment of flooded cells has been
common practice for many years. Simple
internal and external visual inspections
of flooded cells can detect a wide range
of issues, for example, inspection of
the plates, electrolyte and separators
for contamination, shedding material,
sulfation, sedimentary deposits and low
electrolyte level which require regular
inspection and topping up with deionized or distilled water. Flooded cells
require a high level of routine maintenance to ensure reliable operation. As
a result, the majority of users switched
to VRLA batteries as they do not require
regular topping up with water.
Unfortunately, VRLA batteries acquired
a poor reputation soon after they were
introduced in the 1980s. They were
offered as fully maintenance free, but
the lack of maintenance and inspection resulted in batteries failing to meet
design life expectations. The maintenance free characteristics only refer to
the replenishment of water lost by electrolysis. VRLA batteries may not need
water addition, but they do need regular
maintenance and inspection in order to
assess their condition. Unfortunately,
many of the preventative and predictive techniques developed and long
established for flooded batteries cannot
be applied to VRLA batteries. The con-
9
struction of VRLA batteries with opaque
containers and AGM separators or gelled
electrolyte makes internal visual inspection impossible. This has provided the
stimulus for the development of new
techniques for predictive and preventative maintenance and, in particular,
internal resistance measurements using a
variety of commercially available instruments. External visual inspections are still
necessary to detect many failure modes
and to identify where preventative or
corrective maintenance is required.
For all battery types, discharge testing
remains the best way of verifying that
the retained capacity meets the specified
requirement. If the battery fails to meet
the design criteria it should be replaced.
For longer life flooded batteries (~20
years), a discharge test every 5 years
would usually be sufficient. However, due
to shorter life and lack of simple condition assessments for VRLA batteries,
annual discharge testing initially became
the preferred method for measuring battery capacity and assessing condition.
In recent years, improvements in VRLA
battery design and the high cost of discharge testing has resulted in many users
looking for ways to reduce their maintenance regime and reduce costs. VRLA
batteries will never be 100% maintenance free, and the level of maintenance
required for reliable performance is
an issue that has to be addressed.
Battery testing falls into two categories:
performance testing or state-of-health
testing. Performance testing is used
to determine the actual performance
capability of a battery; usually in
terms of capacity measured in Ah,
Wh, or in run-time on a given load.
State-of-health testing uses techniques
that measure one parameter which has
a known correlation to the health or
capacity of the battery. State-of-health
testing gives an indirect indication of
battery capacity; however, a formal
performance or capacity test is still
required to verify actual capacity.
4.1 Battery Performance or
Capacity Measurements
4.1.1 Full Discharge on External Load
The key performance measure of a battery is its capacity. This is determined by
a controlled discharge to verify capacity
at the specified discharge rate. The battery supplier will provide performance
data at constant power and constant
current at specified temperatures to
defined cut-off voltages. In practice this
will require external load banks, plus
time on-site to set up, carry out and clear
away the test afterwards. The battery will
need to be disconnected from the load,
which will result in a reduction of back-up
in the event of a mains outage, both during the test or immediately afterwards
while the battery is being recharged. This
risk of exposure can be minimized by
disconnecting no more than 50% of the
battery on one day in order to maintain
the ability to support the load (even if at
a reduced run-time). Discharge testing
to >90% depth-of-discharge (DoD) only
provides a reliable measurement of battery performance at the time of the test.
EN310TRA-BatOpt / 0113
Battery Optimization Services
52,9°C
50°C
45°C
40°C
35°C
31,6°C
Figure 9. A pair of connectors are
hotter than others carrying the
same discharge current.
4.1.2 Partial Discharge
On-Site or System Load
An alternative to full discharge testing
is to perform a partial discharge using
the system load. With appropriate monitoring and remote system control this
type of measurement can be initiated
remotely. Whether performed remotely
or with engineers on-site, the technique
involves reducing the load placed on the
rectifiers, allowing the entire battery to
support the load instead. The safest way
to do this involves reducing rectifier output voltage but leaving them on-line so
they still support the load in the event of
a battery failure. Alternatively the rectifiers may be disconnected or switched off,
which places the system at higher risk. In
either case the risk is mitigated by terminating the test either at a predetermined
time or at a predetermined cut off voltage higher than the system low voltage
disconnect setting. Performance of the
EN310TRA-BatOpt / 0113
Thermal imaging.
The ideal time to perform a thermal
image scan on the battery installation
is during a discharge test. Infrared
images using thermal imaging cameras can reveal issues that are not
visible. This example shows a pair of
connectors linking two cells together
that are hotter (red) than others carrying the same discharge current
[Figure 9]. Increased resistance at the
terminal due to inadequate torque or
oxidized contact between the connectors and posts is a probable cause
of this localized temperature difference and can be easily corrected.
battery may be estimated by comparing
the measured discharge with a look-up
table of discharge curves at different
rates. If the actual discharge is showing
a steeper decline of voltage than the
reference data, then the battery will
have a lower capacity. The accuracy will
be improved as the depth of discharge
increases, and typically a 60% DoD or
greater is a satisfactory procedure. If the
discharge is too shallow, for example, if
the cut-off voltage is set greater than 1.9
V per cell, it may not be possible to do
a proper comparison to published data
as most manufacturers’ published data
only covers discharges down to 1.9 V per
cell or below. While the risk exposure
of this technique is comparable to full
discharge, and the time required to run
the test (and consequently labor cost) is
lower – partial discharge does not give
the accuracy provided by full discharge.
4.2 Battery Condition or
State-of-Health Testing
4.2.1Float Voltage
Cell or monobloc (bloc) float voltage is
readily measured. If all the cells in a series
string are in a similar condition, the float
voltage will be uniformly distributed.
However, any cells that are showing significant float voltage deviation from the
string average could indicate defects. If
cell or bloc float voltage is consistently
outside of specified parameters, cells
could suffer from accelerated aging
by undercharging or overcharging. In
the early part of the life of a VRLA battery the cell float voltages in a string
vary significantly from cell to cell. This is
caused by variation in the recombination
reaction with the level of saturation in
each cell, which in turn is influenced by
the manufacturing process. With time
on float, the cells lose small amounts of
excess water, the recombination reaction increases in efficiency to a stable
level and the variation in float voltage
decreases. As a result, early data on
voltages needs to be interpreted with
caution; the period for stabilization is
at least 90 days. Measurements made
between 90 and 180 days will be representative of stable conditions, although
there may be very small changes after
6 months in service. A full record of cell
or monobloc voltages will highlight any
serious problems but will not provide any
indication of the battery capacity. Major
variations are likely to result from defective cells at any stage in battery life.
10
Battery Optimization Services
conductance, or DC techniques which
give a measure of internal resistance.
C
R1
R2
Figure 10. Simplified Equivalent Circuit for a Lead/Acid Battery
4.2.2 Float Current and Temperature
Float current is the steady current drawn
by a battery on float charge to maintain
its ‘state-of- charge’. It is much lower
than charge current (drawn by a battery
that has been discharged and is now
recharging to restore state-of-charge)
and also much lower and opposite polarity to the discharge or load current
drawn from the battery by the load.
Float current needs to be within the
specified range recommended by the
manufacturer. It will vary to a small
degree over life, but it is not a sensitive
measure of battery condition. At endof-life, float current will tend to rise as
the battery becomes dryer and more
current is applied to drive the recombination reaction. If the current is excessively
high, cells will become warmer. If they
are significantly above ambient temperature, and not in ‘discharge’ mode
or ‘recharge mode’ the battery is faulty.
Monitoring float current on each string
in a multi string battery could give useful indications of string faults; assuming
all conditions are equal, a significant
difference in string current may indicate the need for further investigation.
Monitoring string float current, ambient
and battery temperature can help to
warn against the onset of thermal runaway. Continuous monitoring of ambient
11
and battery temperature can also help
assess the impact of prolonged high
temperatures on the aging of batteries.
4.2.3 Internal Ohmic Measurements
When first introduced in the 1980s, VRLA
batteries were originally described as
maintenance free in regard to avoiding
the need for regular visual inspections
and topping up of electrolyte. However
it soon became apparent that VRLA
cells were not always performing to
their design life expectations. Internal
visual inspections of VRLA cells are not
possible due to the opaque walls of the
casing, so established maintenance
techniques and inspections developed
for flooded cells could not be transferred
to VRLA cells. Furthermore, even if the
containers were transparent, the battery construction is such that shedding
of active material cannot occur, corroded grid materials cannot be observed
and measurement of acid gravity is
not possible. New techniques had to
be developed to allow internal assessment of the condition of VRLA cells.
The most widely used and effective
techniques for electronically performing
internal inspections can be generically
classed as Ohmic testing. These can be
further broken down into AC techniques
which give a value for impedance or
4.2.3.1 Equivalent Circuit of a Battery
To understand how Ohmic techniques
can be applied to a battery, it is first
necessary to recognize that a battery
can be represented as an equivalent
circuit showing the internal components of capacitance and resistance
[Figure 10]. The internal resistance of a
VRLA battery is comprised of the total
of the resistance of all of the internal
components of the battery: pillars,
group bars, intercell connectors, grids,
active materials and electrolyte. It may
be modeled as a network of resistors,
inductors and capacitors. The inductance is negligibly small and may be
ignored, but the capacitive component
is large at 1-2 F per 100 Ah of cell capacity. A simple model, referred to as the
Randles circuit, is shown in Figure 10.
In this circuit, R1 is the resistance of the
metallic components (pillars, group
bars, intercell connectors and grids)
and R2 is the resistance of the active
materials and electrolyte/separator,
which has a capacitor, C, in parallel. The
capacitance arises from the electrical
double layer that exists at the surface
of active material in contact with the
electrolyte. The electronic charge on the
surface of the electrode attracts ions of
the opposite charge in the electrolyte.
There is a layer of charge in the electrode
and a layer of opposite charge in the
electrolyte, referred to as the electrical
double layer. This capacitance value is
high because the specific surface area
of the electrodes in a battery is high
and the effect of aqueous electrolyte is
strong. In operation, the first small part
of discharge is provided by the electrical
double layer before the normal discharge
processes commence, which are rate
limited by diffusion in the electrolyte.
EN310TRA-BatOpt / 0113
Battery Optimization Services
4.2.3.2 Impedance Spectroscopy
While there are more complex models
for equivalent circuits that attempt to
describe the diffusion controlled processes, electrochemical impedance
spectroscopy has been used to separate
and understand the behavior of the different parts of the equivalent circuit. R2
may be separated into a component that
describes the charge transfer resistance
and another component known as the
Warburg impedance that attempts to
describe the diffusion of ions through
the pore structure of the active materials. However, it is sufficient to consider
a single resistor as the behavior of both
components changes with the age of the
battery. With impedance spectroscopy,
AC signals are injected with a known
current into the cells, and the voltage
response is measured over a range of
frequencies to characterize the values
for the different circuit elements [2].
It is desirable to identify any increase in
impedance or internal resistance, as this
indicates some internal changes which
could correspond to aging factors or fault
conditions. In practice, however, it is not
necessary to identify how a battery is failing; the fact that it is failing means that
action is required. Other factors such
as age, float voltage and temperature
will give background information on the
probable causes of battery deterioration.
4.2.3.3 AC Ohmic Techniques
Impedance or conductance readings are
derived from AC signal injection techniques. These techniques are used for
condition monitoring of VRLA batteries
with a single low frequency (generally
in the range 10-100 Hz, although some
test equipment operates at 1 kHz), and
the value obtained for the impedance
will be largely determined by R1 and
R2 with a small component from C.
Some equipment provides an impedance figure, others a conductance
figure; one is simply the reciprocal of
EN310TRA-BatOpt / 0113
Voltage
V
Open Circuit Voltage
R int = V2 – V1
I
Current
TIME
Figure 11. Calculating Internal Resistance from Ohms Law
the other. A battery, however, also has
an electrochemical element, or ionic
conductance, which does not follow
Ohms law. Consequently, a conductance
meter reading cannot simply be inverted
to get an impedance value to compare
with published impedance values or
measurements derived from other
impedance test equipment. Various
factors affect the reading, including
frequency, amplitude of the test signal
and the resolution of the test meter.
As the equivalent circuit illustrates,
there is a significant value of capacitance in a battery cell. Capacitors
have a high admittance for AC signals,
which is proportional to the frequency
of the test signal. The capacitor is in
parallel with a significant part of the
internal resistance elements of a cell.
Consequently, the results derived with
different equipment will vary considerably due to different test frequencies.
In general terms, lower test frequencies
provide better results for assessment
of internal resistance changes. No two
types of impedance or conductance
meters can be guaranteed to give
identical results, and so it is important
that the same type of test meter is
used throughout the life of a battery
in order to obtain consistent results.
4.2.3.4 DC Ohmic Techniques
Alternatively, the internal resistance
may be measured by DC techniques.
In this case the battery is subjected to
a momentary resistive load provided
by the test instrument, which lowers the surface charge on the plates
down to just above the open circuit
voltage level of the cell. The current
(I), the lowest voltage (V1), and the
voltage recovery (V2) are all measured when the load is removed. The
internal resistance (Rint = R1 + R2) is
then calculated from Ohms law.
Rint = (V2-V1)/I [Figure 11]
12
Battery Optimization Services
In practice, this technique will give reliably consistent results, not affected
by the capacitive value of the cell and,
therefore, not affected by frequency
or any spurious AC noise that may
be present on the battery or DC supply. Notwithstanding, this technique,
if applied with different loads (and
consequently test currents), will give
different results due to differences in
the electrochemical behavior of the
battery. Consequently, as with various AC techniques, it is important that
the same type of test equipment is
used throughout the life of the battery to obtain consistent results.
4.2.8 IEC 60896-21: Manufacturers’
Test Method for Internal Resistance
Manufacturers often report a DC internal
resistance measured to IEC 60896-21.
This is measured in a different way
to the voltage dip/voltage recovery
method. A first data pair for voltage
and current, Va and Ia, is recorded after
applying a current of 4C10 for 20 s to a
fully charged cell. A second data pair,
Vb and Ib, is then recorded after the
battery has stood on open circuit for 5
minutes with a current of 20C10 applied
for 5 s. The internal resistance is then
calculated as Ri = (Va – Vb)/ (Ib – Ia).
This figure will not necessarily be the
same as any figure measured by a voltage dip instrument, or an impedance or
conductance meter. It can be used to
estimate short circuit currents and for
the sizing of circuit protection devices.
4.2.9 Correlation Between Internal
Resistance and Capacity
There is considerable experience and a
large amount of published literature [3-7]
that shows that impedance or conductance measured at low frequencies or
DC resistance correlates reasonably well
with battery capacity. No significant
capacity loss should occur within the first
increase (<25%) in internal resistance.
With a 25-35% increase there is typi-
13
cally a small capacity loss, and with an
increase of 35% to 50% there is usually
a larger capacity loss approaching the
end-of-life figure of 20% lost capacity. At
an internal resistance increase of >50%,
the battery is highly likely to have lost
20% or more of the original capacity, and
would therefore be beyond end-of-life
as defined by battery manufacturers.
These measurements need to be carried
out with the battery in a fully charged
condition as they vary with state-ofcharge and at the same temperature or
compensated for temperature variation.
They are also dependent on the type of
instrument used. The most reliable data
will be obtained by using the same type
of instrument from installation to track
the change in internal resistance over
life. There is no satisfactory way of converting readings from one instrument to
another. For AC techniques, the variation
in frequency will give different readings
and will be different to DC techniques.
4.2.10 Baseline Reference Values
The baseline or initial values are
important for reference because stateof-health monitoring relies on tracking
changes to measured parameters over
time. Ideally a baseline value for internal
resistance (or impedance/conductance)
should be established shortly after batteries are commissioned and placed into
float service (typically 3 to 6 months
after commissioning). If surveillance
starts at a relatively early stage in life,
(typically 2 to 4 years) the initial condition can be estimated by extrapolation.
Baselines should be established for each
battery installation. However, if this is
not possible due to the age when testing
commenced, then it is possible to use
a peer group baseline which has been
established on the same type of battery
but in another installation. Battery manufacturers’ published Ohmic values are
useful as a guide but cannot be used as
accurate baseline reference values. This is
primarily because factory test techniques
are not the same as in service field test
techniques; therefore, values can vary
due to technique and equipment. In addition, a newly manufactured battery will
often need a period on floating charge
to stabilize the internal resistance, as the
recombination process will take some
time to reach a steady state. Baseline
management and technique for establishing reference data on all inspected
battery types is an essential part of
correct analysis of battery condition.
4.2.11Measurement of Battery Stateof-Health with Internal Resistance
Mismanagement or lack of understanding of battery baseline or reference
values has been one of the principal
reasons for the relatively slow adoption of Ohmic techniques. The industry
has seen many test equipment solutions developed, but there is no
recognized standard for how the measurement of internal resistance is made.
Consequently, many systems exist
that are equally valid but not compatible or interchangeable. When Ohmic
techniques are correctly managed and
applied, they are an invaluable tool in
assessing battery state-of-health.
EN310TRA-BatOpt / 0113
Battery Optimization Services
160
End-of-Life Zone
Battery replacement required
>50% increase
Internal Resistance (%)
150
Alert Zone
Between 35-50% increase in internal resistance
Recommend battery replacement
140
130
Between 25-35% increase in internal resistance
Battery capacity approaching end-of-life
120
Up to 25% increase in internal resistance
35-50% increase
Warning Zone
25-35% increase
Safe Zone
Baseline 100%
110
Internal Resistance (%)
100
Battery Capacity (%)
90
80
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Percent of Service Life
Figure 12. Battery Life Cycle Chart
4.2.12 Battery Life Cycle
The battery life cycle chart [Figure 12]
shows the change in internal resistance
over the service life of the battery, and
the actual capacity over the service life.
In the early part of life there is a modest
increase in capacity, and at end-of-life it
declines to reach the 80% cut-off. The IR
measurement should be tracked against
a baseline measurement for a new and
healthy battery. This measurement
increases progressively over the battery
life because of normal aging processes,
and then increases more rapidly towards
end-of-life. In the early part of life, the IR
will increase by up to 25% but capacity
will remain close to 100% of the nominal
value. As the IR increases to more than
25%, but below 50%, battery capacity will start to fall. When it reaches or
exceeds 50%, the retained capacity will
EN310TRA-BatOpt / 0113
have fallen to 80% or less. The decline
in capacity when the 80% threshold has
been reached is much more rapid, and
life beyond this point cannot be predicted with any certainty. The battery
must be replaced at this point for the
load to be supported. Any cells or monoblocs that have an IR growth rate above
the average level need to be investigated further, as they may have early life
defects such as manufacturing defects,
dry out, sulfation or other problems.
Internal resistance measurements have
been widely adopted as an industry
recognized state-of-health indicator for
VRLA batteries. These measurements
alone, however, cannot accurately
measure the actual battery capacity.
Consequently, at various threshold levels
(plus 20-25%, 30-35%, 40-45%) discharge
tests need to be carried out if absolute
determination of capacity is required. In
practice, a combination of IR measurements, full discharge or partial discharge
tests and physical inspection of batteries
is required to maintain batteries in good
condition and to ensure that they are
always ready to support the load within
their design limits. Financial limitations
may restrict testing budgets for some
installations, but an appropriate selection of lower cost test techniques can still
provide a user with greater confidence
in battery condition and reliability.
14
Battery Optimization Services
5.0 Examples of Actual Measurements
5.1 Internal Resistance and Discharge Testing
The following series of charts show the
results of two types of tests performed
on the same battery; internal resistance
test per cell [Figure 13], discharge test
of the full battery [Figure 14] and discharge test per cell [Figure 15]. In the
third chart the results of the internal
resistance and the discharge tests are
combined [Figure 16] and in the last
chart the combined results are sorted
in ascending IR order [Figure 17].
1. Results of an IR test on a battery of 2 x 24 x 2 V cells – Some anomalies are evident
(flagged as black, red and amber) [Figure 13].
1600
Internal Resistance Values Per Cell
Internal Resistance (µΩ)
1400
Internal Resistance
1200
Base Line
1000
Start of Amber Internal Resistance
Warning Zone
Start of Red Internal Resistance
800
Alert Zone
600
Start of Black Internal Resistance
Critical (or End-of-Life) Zone
400
200
0
Battery Blocs
Figure 13. IR Test Anomalies (Black, Red and Amber)
2. Result of a discharge test performed shortly after the IR test – The battery performance just met the run time criteria, so it passed. In many cases this test would
be the only source of information, and it would be assumed the battery is healthy. A
partial discharge test would not have tested the battery deeply enough to identify
the actual capacity [Figure 14].
55
Total Battery Voltage
53
End of Discharge Voltage
51
Voltage
49
47
45
43
41
39
37
Time
Figure 14. Discharge Test Curve Shortly After IR Testing
15
EN310TRA-BatOpt / 0113
Battery Optimization Services
3. Discharge data shown on a cell by cell chart – This chart shows that several cells
did not meet the required capacity. Any one of these cells could be on the point of
outright failure, which would result in a catastrophic loss of capacity and failure to
support the load [Figure 15].
2.40
Individual Cell Voltages
2.20
Voltage
2.00
End of Discharge Voltage
1.80
1.60
1.40
1.20
1.00
Time
Figure 15. Discharge Data of Individual Cells
4. IR and depth of discharge (DoD) data combined – The capacity of each bloc is
calculated as a DoD percentage. The discharge test was to 1.8 V per cell, which
delivered a 95% depth of discharge. Cells above 1.8 V at termination are >95%
capacity. 100% capacity can be inferred but only proven if the discharge actually
went down to 1.75 V per cell. [Figure 16].
100%
Internal Resistance (µΩ)
1400
1200
90%
Internal Resistance Values Per Cell
80%
Depth of Discharge Per Cell
70%
1000
60%
800
50%
40%
600
30%
400
20%
200
Depth of Discharge
1600
Internal Resistance
Base Line
Start of Amber Internal Resistance
Warning Zone
Start of Red Internal Resistance
Alert Zone
Start of Black Internal Resistance
Critical (or End-of-Life) Zone
10%
0%
0
Battery Blocs
Figure 16. Discharge Data and Internal Resistance Combined
EN310TRA-BatOpt / 0113
16
Battery Optimization Services
5. IR and DoD data combined and sorted in ascending order – Displaying DoD on the
same chart as IR in ascending order demonstrates a very clear relationship between
IR and capacity. The majority of cells in the green zone are >95% DoD. The majority of cells in the amber zone are <95% DoD. All cells in the red zone are between
70 and 80% DoD. The two cells in the black zone are <70% DoD. [Figure 17].
100%
Internal Resistance (µΩ)
1400
1200
90%
Internal Resistance Values Per Cell
80%
Depth of Discharge Per Cell
70%
1000
60%
800
50%
40%
600
30%
400
Depth of Discharge
1600
20%
200
Internal Resistance
Base Line
Start of Amber Internal Resistance
Warning Zone
Start of Red Internal Resistance
Alert Zone
Start of Black Internal Resistance
Critical (or End-of-Life) Zone
10%
0%
0
Battery Blocs
Figure 17. Discharge Data and Internal Resistance in Ascending Order
This correlation between IR and capacity enables Ohmic testing to be used
as an indicator of battery condition.
Individual cell faults or suspect faults
can be identified from the IR measurements. However, care should be taken
when evaluating only a single set of IR
data; it is possible to see a wide range of
IR values, especially on new cells, while
capacity is still within acceptable limits.
17
5.2 Trend Analysis for Internal
Resistance Measurements
For a full and proper evaluation of
battery condition using IR data, the
readings should be taken on a regular
basis (minimum one test per year), and
the collected data should be plotted
in a trend report to evaluate change
over time. Individual bloc data should
be evaluated to identify bloc faults,
premature aging, potential low capacity issues and other faults. Overall
battery performance can be evaluated
by plotting the trend on string maximum, minimum and average values.
The string average trend is a useful
indicator for projecting service life;
the maximum value trend will indicate
premature aging faults that can impact
overall service life if left uncorrected.
The minimum value gives a comparison
for the other two and also can indicate
when new blocs have been fitted.
When developing the battery optimization program, historical data from
many battery tests was reviewed.
Techniques for analyzing the data were
developed, and the trends for IR measurements were carefully examined.
In this section we will look at four distinct examples with IR tests performed
on different batteries. Each example
consists of six data sets.
EN310TRA-BatOpt / 0113
Battery Optimization Services
Example 1 – Typical battery behavior
expectancy of 8 years or more, while
baseline value must be established as
In Figure 18, the maximum, minimum
the trend line for the maximum value
accurately as possible to determine the
and average IR values for a string
shows a potential failure projected
start point for service life projections.
have been plotted against time with
between years 5 and 6. This potential
a baseline (green area), 25% increase
failure became evident after the first
warning threshold (amber area), and a
three readings, and corrective action
Leaving the defective bloc in the string
would risk premature failure of the
can be planned before the string is
whole string. The health and service
at risk of failing. Trend analysis for IR
life of the remaining blocs can also be
data is essential to assess the behav-
affected due to accelerated aging caused
The IR values show a near linear rate of
ior of the entire battery string or
by the defective bloc. Early detection
increase with time. The trend line for the
individual cells or monoblocs in com-
allows for corrections to be made while
average readings show a potential life
parison to the whole battery. The
batteries are still under warranty.
>35% alert threshold (red area), >50%
End-of-Life threshold (black area).
1800
Internal Resistance (µΩ)
1600
Data Sets:
1
2
3
4
5
End-of-Life Zone
6
>50% above baseline
Alert Zone
1400
>35% above baseline
1200
Warning Zone
>25% above baseline
1000
Safe Zone
800
Baseline 970 µΩ
600
String Max (
400
String Average (
200
String Min (
trend)
trend)
trend)
000
Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4
Year 0
Year 1
Year 2
Year 3
Year 4
Year 5
Year 6
Year 7
Date / Year
Figure 18. Maximum and Minimum IR Values Plotted Over Time
EN310TRA-BatOpt / 0113
18
Battery Optimization Services
Example 2 – IR testing
starting near end-of-life
Figure 19 shows battery IR data collected
when the battery was close to end-oflife. IR maximum, minimum and average
values are plotted with the ambient
temperatures that were also recorded.
measurement anomalies. This example
shows that the IR anomaly at data set
5 is not correlated with temperature,
and therefore, may have been due to
inaccuracies in measurement, change
of equipment type or operator error.
Six sets of data were plotted in total.
There were anomalies between consecutive readings as illustrated by the
fifth set of data, where the minimum,
maximum and average all fell below the
trend established by the other plots.
Plotting temperature at the same time
helps to identify if this is a factor in IR
The average values showed a near
linear trend, but the maximum values were increasing rapidly. The lack
of early life data makes fault prediction difficult, but this example does
show that some cells have failed at
year 7 in a battery that could otherwise reach a ten year service life.
2400
2000
12
3
4 5
1800
30.0
1600
1400
1200
25.0
1000
800
22.0 22.0
600
21.0 21.0
20.0
20.0
400
18.6
200
•While data collected in the latter years
of a battery life needs to be carefully
analyzed, especially with regards
to establishing the baseline value,
meaningful analysis can still be made.
>50% above baseline
6
Ambient Temperature Deg C
Internal Resistance (µΩ)
Data Sets:
•Consistency in measuring instruments
and technique is essential for the
data to be used for trend analysis.
End-of-Life Zone
35.0
2200
From this example the following can
be concluded:
Alert Zone
>35% above baseline
Warning Zone
>25% above baseline
Safe Zone
Baseline 1200 µΩ
String Max
String Average (
trend)
String Min
Temp °C
0
15.0
Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4
Year 0
Year 1
Year 2
Year 3
Year 4
Year 5
Year 6
Year 7
Year 8
Year 9
Year 10
Date / Year
Figure 19. Trend/Regression Data – IR Testing Starting Near End-of-Life
19
EN310TRA-BatOpt / 0113
Battery Optimization Services
to one cell rather than the entire battery,
and early replacement of a bad cell will
bring the battery back to a good condition with many years potential service
life still remaining.
Example 3 – Early life failure
In this example [Figure 20], the maximum values show that at least one cell
failure has occurred before the end of
year two. The second data set shows
the maximum IR value was above the
warning level. The third reading was
substantially above the alert level and
off the scale in the remaining data sets.
The slower increase of the average values
shows that the fault is likely to be limited
If this faulty cell remained in service,
the whole string would be at risk of
failure and subject to accelerated aging
or reduced service life. This is a good
example of an instance where replacing
50,000
45,000
End-of-Life Zone
Data Sets:
1
2
3
4
5
>50% above baseline
6
Alert Zone
40,000
Internal Resistance (µΩ)
a single cell, probably while still covered
by manufacturers’ warranty, protects the
load and enhances the service life of the
whole string. This also illustrates how the
IR rate of change can rapidly accelerate
as a cell fails or approaches end-of-life. If
the interval between inspections is too
long, a rapidly developing fault in a single
cell or monobloc can affect string performance and a perfectly good battery may
be prematurely ruined.
>35% above baseline
35,000
Warning Zone
30,000
>25% above baseline
Safe Zone
25,000
Baseline 14,580 µΩ
20,000
String Max
15,000
String Average (
10,000
String Min (
5,000
trend)
trend)
0
Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4
Year 0
Year 1
Year 2
Year 3
Year 4
Year 5
Year 6
Year 7
Date / Year
Figure 20. Internal Resistance Over Time
EN310TRA-BatOpt / 0113
20
Battery Optimization Services
approaching end-of-life, as both
average and maximum internal resistance values are above the warning
level and alert level respectively.
Example 4 – Short battery service life
Figure 21 demonstrates that the entire
battery has a short service life. Data
sets 3, 4, 5 and 6 in years 4 and 5 show
an accelerated increase in internal resistance for both the maximum values and
the average values. This shows that many
cells are increasing at the same time.
This is a good example of where total
battery replacement is required,
and possibly a review of a suitable
battery type for the application so
a longer life can be achieved.
As early as half way through year 4, it
is evident that the battery is quickly
22,000
Data Sets:
20,000
1
2
3
4 5
End-of-Life Zone
6
>50% above baseline
Alert Zone
Internal Resistance (µΩ)
18,000
>35% above baseline
16,000
Warning Zone
14,000
>25% above baseline
12,000
Safe Zone
10,000
Baseline 9500 µΩ
8,000
String Max
6,000
String Average
4,000
String Min (
trend)
2,000
0
Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4
Year 0
Year 1
Year 2
Year 3
Year 4
Year 5
Year 6
Year 7
Date / Year
Figure 21. Trend/Regression Data – Short Battery Service Life
21
EN310TRA-BatOpt / 0113
Battery Optimization Services
6.0 Battery Maintenance
Ideally, battery maintenance should
be tailored to the criticality of the site
and to the user’s budget requirements.
A dynamic, intelligently scheduled
maintenance program that delivers
preventative, predictive and corrective
maintenance and uses both state-ofhealth and discharge measurements
will provide the optimum information on overall battery health. While
site criticality and cost will most likely
determine the solution, a flexible
approach using a combination of remote
or site based capacity and state-ofhealth testing will meet most needs.
6.1 Key Requirements for a
Battery Maintenance Program
In order to make an informed decision
about the best method for testing a
battery installation or network of battery installations, a clear vision of what
is expected from a battery maintenance
program is required. This vision will
then help in defining which features
to include in the ideal test program.
In this section, a variety of test techniques will be examined in close detail
and a method for objective selection
based on their relative merits, strengths
and weaknesses will be described:
1.Ideal battery maintenance
(setting the benchmark)
2.On-site discharge: full discharge
test using an external load
3.On-site discharge: partial discharge
test with the on-site load
4.On-site internal resistance test
5.Remote discharge: partial discharge
test with the on-site load
6.Remote testing with dedicated
hardware: partial discharge
and internal resistance
7.Battery optimization: selecting
the best program to suit particular
battery maintenance requirements
EN310TRA-BatOpt / 0113
Figure 22. Ideal Battery Maintenance
6.1.1 Ideal Battery Maintenance
To provide a technique for identifying
optimal battery maintenance methods,
a set of ideal parameters needs to be
defined; i.e. what you expect or need to
achieve with the service delivered. In our
analysis we have identified 6 key parameters. The individual characteristics of
several maintenance techniques are
then assessed against these parameters.
The ideal battery maintenance program
would score 10 on each parameter
and can be conveniently presented by
a simple spider diagram [Figure 22].
1. Capacity estimation: The most
fundamental battery requirement is to provide the capacity
needed to support the load for
the designed back up time.
2. Faulty cell detection: Even in a new
battery, a single cell or monobloc failure can lead to loss of critical power,
and detecting these faults enables
optimum protection for the load.
3. Remaining life estimation: The
service life of a VRLA battery is
often far less than the design life,
as they are affected by a number
of variables; temperature, charging, early life defects and other
problems. If the remaining service life can be estimated, then
replacements can be planned
before a battery failure causes a
loss of available backup time.
4. Mechanical issue correction:
Links, straps and connecting
hardware require regular checking and correction, as do physical
issues such as leaking acid, swollen cells or monoblocs, high
temperature and other factors.
5. Cost effectiveness: The results
from each technique also need to
provide a good return on the OPEX
invested. The CAPEX on batteries
is a significant investment, often
higher than the DC plant it supports.
An ideal service program should be
able to maximize battery service
life so that CAPEX and the total cost
of ownership can be reduced.
6. Test risk avoidance: An important
priority in any battery test technique
is that the load is not put at any
unnecessary risk. Techniques and
procedures used should minimize
this risk and maximize available back
up power during the test procedure.
Different maintenance and diagnostic
techniques will score between 0 and
10 for the various features required,
where 10 is a perfect score. These are
described on the following pages.
22
Battery Optimization Services
6.1.2 On-Site Discharge: Full Capacity
Testing with an External Load
Battery discharge testing is acknowledged in the industry as the most
effective method of determining a
battery’s actual capacity and ability
to provide a reliable source of power.
Many battery manufacturers will only
accept these tests when questioning
battery performance for warranty issues.
Regular test results will clearly identify
when the battery capacity is falling
below specification, allowing the user
to target battery replacement actions
on the most critical systems. Load testing also verifies the integrity of the
DC conduction path without placing
the plant in jeopardy of failure during
testing. Against all the advantages,
however, it is also the most time consuming and costly method of testing.
This off-line test method discharges
the battery into an external load and
provides an accurate measure of capacity at the time of the test. The on-site
discharge test is assessed against the
parameters as follows [Figure 23]:
1. Capacity estimation: (10)
Determines the actual battery
capacity measured to industry
standards or to the battery manufacturer’s specification.
2. Faulty cell detection: (10) Any
cell or monobloc that is defective will be detected; the voltage
of each cell or monobloc will
be measured and clearly identified during the discharge.
3. Remaining life estimation: (4) As
a battery ages, the capacity starts
to decline below 100%. This can be
detected before the capacity drops
below a critical level (80% of specified capacity is generally accepted
as end-of-life). However, early-life
discharge testing gives little indication of the likely service life.
5. Cost effectiveness: (1.5) Performing
capacity tests to the manufacturer’s
specification requires skilled engineers and time; each test will run
for 2, 3 or 4 hours according to the
selected parameters. Often this
test needs to be performed out
of normal hours to reduce the risk
to critical traffic. This is the most
expensive testing technique.
6. Test risk avoidance: (7) To protect
critical loads, no more than 50% of
the installed batteries are tested
on one day and the string(s) under
test are isolated from the system to
ensure that customer equipment
is not subjected to decreased voltage levels. Discharged strings are
re-introduced to the system in a
controlled manner to ensure a seamless integration of standby power.
4. Mechanical issue correction:
(10) Before each discharge test, a
thorough visual inspection of the
battery and a check of torque settings are performed. Issues can
be corrected before discharge.
Figure 23. Full Discharge
23
EN310TRA-BatOpt / 0113
Battery Optimization Services
6.1.3 On-Site Discharge:
Partial Discharge Testing
with the On-Site Load
This technique is widely used by
telecommunications operators, as it
demonstrates the system’s ability to
support the load in the event of a mains
failure. Since all connected strings are
discharged at the same time, partial
discharge can be performed in a much
shorter time than a full discharge on a
system with multiple batteries. Partial
discharge measures the run time on site
load to an agreed voltage level above
the systems low voltage disconnect
limit, but only estimates run time to the
systems specified lower voltage limit.
This estimate assumes that partially discharged cells will continue to discharge
along the same performance curve. This
assumption is risky; a nominal 60% depth
of discharge only proves a battery can
deliver 60% of capacity. To prove it can
deliver at least 80%, a discharge to >80%
depth of discharge is needed. Service
life is difficult to forecast from partial
discharges, and often a failed cell or
monobloc will only be found after it has
reached end-of-life. At that point, a battery system failure may be imminent.
It should be noted that this is a functional performance test rather than a
capacity test, and battery manufacturers may not accept partial discharge
test results when questioning battery capacity for warranty issues. This
test uses the actual service load
to partially discharge the battery
and is assessed against the defined
parameters as follows [Figure 24]:
Figure 24. Partial Discharge
1. Capacity estimation: (6) Partial discharge testing is typically limited to
1.90 V per cell, which gives approximately a 60% depth of discharge;
the capacity for a full discharge
can be estimated by extrapolation
from the battery manufacturer’s
published specification. It should
be noted that the load is live and,
therefore, cannot be adjusted.
2. Faulty cell detection: (8.5) This will
detect battery cells or monoblocs
with a significant loss of capacity performance; early detection of faults
is difficult and may not be possible.
3. Remaining life estimation: (3)
Partial discharge gives limited
data for estimating the remaining service life. Deeper discharges
are required to provide early
indications of capacity loss.
5. Cost effectiveness: (3) Performing
partial capacity tests requires skilled
engineers and time; each test will run
for typically two hours according to
selected parameters. Cost and time
is reduced by discharging all connected batteries at the same time.
6. Test risk avoidance: (7) To protect
the critical load, the depth of discharge is typically restricted to 1.90
V per cell rather than a full discharge
to 1.75 V per cell. The rectifier voltage is reduced, not switched off, so
that the rectifiers will still support the
load even if a battery fails. If a utility
outage occurs during or shortly after
the test, then the available battery
autonomy time will be significantly
lower than a fully charged battery.
4. Mechanical issue correction:
(10) Before each discharge test,
a thorough visual inspection
of the battery and a check of
torque settings are performed.
EN310TRA-BatOpt / 0113
24
Battery Optimization Services
6.1.4 On-Site Internal
Resistance Testing
This technique is gaining wider acceptance in the industry by both battery
manufacturers and users as a reliable
method to determine a battery’s stateof-health. Internal resistance testing is
quicker and more cost effective than
discharge testing, and the battery capacity remains 100% available to the critical
load throughout the test. It is also a
useful technique to detect faulty cells
and monoblocs; furthermore, in regular
maintenance it is used to predict battery failures before they lead to a loss of
power to the critical load. Several battery manufacturers have prepared white
papers or technical bulletins on Ohmic
test techniques [8,9], and practically all
battery manufacturers will publish an
internal resistance, impedance or conductance value for their products.
State-of-health internal resistance testing should be considered a valuable
supplement to discharge testing. Many
users consider it to be reliable enough
to replace discharge testing; however,
if absolute proof of battery capacity is
required, only a capacity discharge test
will provide this.
The use of on-site internal resistance
measurements is assessed against the
defined parameters as follows [Figure 25]:
1. Capacity estimation: (3) Internal
resistance can be correlated to
capacity by comparing measured
state-of-health values to known
healthy values for IR, but a direct
measure of capacity can only be
obtained by discharge testing.
5. Cost effectiveness: (7.5) The
internal resistance battery meter
automatically records each parameter as the test is performed. Each
cell or monobloc test takes a few
seconds, and an entire battery test
can be fast and very cost effective.
2. Faulty cell / monobloc detection:
(8.5) Individual cell or monobloc
values can be compared against
the known healthy baseline and
against other cells or monoblocs
in the same battery to identify
anomalies that may be attributed
to low capacity or defective cells.
The accuracy of this improves
with more data for trending.
6. Test risk avoidance: (10) The internal resistance battery meter can be
used while the battery under test
is still on-line, with full availability
to support the critical load. The
test is non-invasive in that it does
not affect system performance.
3. Remaining life estimation: (7)
Trending state-of-health data
enables forecasting of service life
and early detection of faults. The
accuracy of this improves when
combined with discharge test data.
4. Mechanical issue correction: (10)
The full conducting path, including
straps, inter-tier and inter-cell or
monobloc links can be tested. This
can identify high resistance links due
to incorrect torque settings, corrosion and other problems. These
can be corrected while on-site.
Figure 25. Internal Resistance
25
EN310TRA-BatOpt / 0113
Battery Optimization Services
6.1.5 Remote Testing: Partial Capacity
Testing with the On-Site Load
With minimal additional hardware,
useful battery information can be
extracted from the DC system controller with remote access, either through
an Ethernet connection or a wireless
modem. As with on-site partial discharge, this technique demonstrates the
system’s ability to support the load in the
event of a mains failure. This test can be
performed quickly and cost effectively,
although data at the bloc level is not
recorded with basic hardware. Remote
monitoring allows these tests to be
performed as often as required and also
provides information 24/7, thus improving the chances of early detection.
Capacity estimates are not as accurate
as estimates from a full discharge test.
While the site load is actually being used
for this test, it is initiated and monitored
remotely with basic hardware only. It
may be assessed against the defined
parameters as follows [Figure 26]:
1. Capacity estimation: (6) Partial discharge testing is typically limited to
1.90 V per cell, which gives approximately a 60% depth of discharge.
Capacity for a full discharge can be
estimated by extrapolation from battery manufacturer’s specification.
2. Faulty cell detection: (2) With
only battery voltage and string current readings available, individual
cell or monobloc faults may not
be detected. String faults will indicate a potential cell or monobloc
fault, but not identify which one.
3. Remaining life estimation: (5)
Partial discharge gives limited
data for estimating the remaining service life. Deeper discharges
are needed to identify the early
stages of capacity loss.
4. Mechanical issue correction: (0)
Remote monitoring alone cannot identify potential mechanical,
torque setting or corrosion issues.
Site visits will be required for
visual and physical inspection.
5. Cost effectiveness: (9.5) Remote
partial capacity tests are performed
by skilled engineers and operators at a 24/7 monitoring center.
No engineer attendance to site
is required. Each test will run for
typically two hours according to
selected parameters. Cost and time
is reduced by discharging all connected batteries at the same time.
6. Test risk avoidance: (7) To protect
the critical load, the depth-of-discharge is typically restricted to 1.90
V per cell rather than a full discharge
to 1.75 V per cell. The rectifier voltage is reduced but not switched off,
and if a battery fails, the rectifiers
still support the load. If a utility
outage occurs during or shortly
after the test, the available battery
autonomy time will be significantly
lower than a fully charged battery.
Figure 26. Remote Partial Discharge
EN310TRA-BatOpt / 0113
26
Battery Optimization Services
6.1.6 Remote Testing with
Dedicated Hardware: Partial
Discharge and Internal Resistance
State-of-the-art monitoring hardware
can deliver 24/7 information on battery
condition. This can be an ideal solution
for highly critical sites, or sites where
physical location makes site visits either
difficult or prohibitively expensive.
Apart from discharge data and IR data,
information on alarms, temperature,
utility reliability, system load and many
other parameters can be used to assess
complete system performance and
health. A remote monitoring solution
cannot, however, provide physical and
visual inspections, check and adjust
torque settings or clean leaking cells.
Prevention by cleaning and re-torquing
of connectors is an important part of
battery maintenance, and routine site
visits should still form part of a comprehensive battery maintenance program.
With appropriate hardware, remote discharge testing may be combined with
remote internal resistance measurements and assessed against the defined
parameters as follows [Figure 27]:
1. Capacity estimation: (7) A dedicated monitoring solution can be
used to control and record a partial
discharge, and with 24/7 monitoring
any unplanned outage event may
also be captured. Data from these
discharges will be at bloc level and
can be used to estimate capacity
or verify capacity if the unplanned
event provides a sufficiently deep
enough discharge. (3) Remote IR
does not add to capacity estimates
but helps identify which blocs are
affecting battery performance.
2. Faulty cell detection: (7 / 9.5)
Single bloc or monobloc faults can
be detected as they develop and
tracked by frequent monitoring to
confirm anomalous behavior before
a critical fault occurs. Multiple readings increase the accuracy of fault
detection, and the analysis of both
IR and discharge data increases
the accuracy to detect faults.
3. Remaining life estimation: (5 / 8)
Trending state-of-health data enables
forecasting of service life and early
Remote Internal Resistance Test
Remote Discharge Test
detection of faults. Multiple readings
and analysis of both IR and discharge
data increases the accuracy of service
life prediction and fault detection
4. Mechanical issue correction: (4)
With appropriate IR monitoring
hardware, the full conducting path
including straps, inter-tier and intercell links can be tested remotely.
This can identify high resistance
links due to poor torque or corrosion. Corrective maintenance
visits can then be scheduled based
on actual condition of the battery and connectors. (0) Remote
discharge tests cannot identify
or correct mechanical issues.
5. Cost effectiveness: (7 / 8)
Monitoring hardware cost can pay
for itself by reduced CAPEX, as
batteries are only replaced when necessary, OPEX is also reduced as site
visits are only required for essential
maintenance and remedial work. In
addition, there is reduced downtime.
6. Test risk avoidance: (10) The IR
testing can be performed while the
battery is on-line with full availability to support the critical load;
the test is non-invasive and does
not affect the system performance.
(7) The discharge test is performed
by reducing the DC bus voltage,
not switching the rectifiers off
or disconnecting the battery.
Figure 27. Remote IR with Remote Discharge
27
EN310TRA-BatOpt / 0113
Battery Optimization Services
6.2 Battery Optimization: Selecting
the Best Program to Suit Particular
Battery Maintenance Requirements
By defining parameters and values for
features of various battery maintenance
techniques, it becomes clear that no
single technique can deliver optimum
battery health information and score
high on all desirable features. To gain
maximum benefit from battery maintenance, a holistic, life cycle approach is
required. Rather than selecting a single
technique that may not deliver on all
required features, a combination of
techniques could be used. Combining
techniques in a dynamic way can achieve
results that are greater than any single
technique. For example, service life
predictions are far more accurate when
internal resistance data and capacity
data are analyzed together. If the current
regime is annual discharge testing, then
the frequency of discharge tests can be
reduced from a fixed annual schedule to
a condition- based schedule. This gives
opportunities for OPEX budgets to go
further and allow testing on parts of
the network previously lacking in maintenance due to budget restrictions.
The criticality of each node in a network
would typically be defined by the traffic flowing through the node. This is
normally based on the volume of traffic,
as the revenue streams are generally
proportional to volume. However, the
nature of the traffic also plays a significant part in assessing the criticality of
the node. For example remote nodes
used for emergency services traffic
would be more critical than a similar
node with overlapping coverage used
for non-essential traffic. The nature
of telecommunications traffic is a rapidly changing landscape. Video and
data transmission (including financial
transactions and other applications)
are creating more and more demand
for bandwidth and availability from all
networks. Network resilience requires
redundancy and reliability of the power
supplies. When the utility supplies fail,
the network relies on the emergency
back-up power, initially from the battery
and then, in many cases, from generators. Network reliability is therefore
inextricably linked with battery reliability. All layers of a network should have
a maintenance program that is right
for the various applications. A selection of proposed solutions is outlined
below, and the flexibility of the battery
optimization concept allows for many
alternative solutions to be defined.
6.2.1 Proposed Solution for Access
Level of a Network (less critical nodes)
For a typical access node in a telecom
network, the usual solution is zero
maintenance. Batteries are replaced
either on a predetermined schedule based on age, or on a break/
fix system which only reacts after
a node has suffered downtime.
As mentioned above, these strategies
are risky, and the potential losses could
be high enough to justify a low cost
maintenance program. For the access
level of a network, it is proposed that
a basic remote monitoring service
with annual remote battery discharge
tests could generate enough savings in
CAPEX and reduced down time to give
a return on investment and enhanced
reliability for the nodes [Figure 28].
It was discussed earlier that the selection of maintenance techniques
should be based on criticality of the
supported load. To understand this,
a typical network consisting of an
outer layer of access sites and other
nodes, an inner layer or backbone of
more critical sites and localized hubs,
and a core layer of centralized critical hubs, which ultimately carry all of
the network traffic, is considered.
Figure 28. Optimized Maintenance
EN310TRA-BatOpt / 0113
28
Battery Optimization Services
6.2.2 Proposed Solution for
Medium Critical Nodes
Even at this level of criticality some operators still chose the
risky zero maintenance option;
full discharge is too expensive.
Some will use the lower cost partial discharge technique, but this has limited
value in fault detection and prediction.
Annual On-Site Internal Resistance Test
Annual Remote Partial Discharge Test
The proposed solution with the optimization concept is to combine annual
remote partial discharge tests with
annual site IR tests [Figure 29].
This combination adds all site visit related
benefits at a cost which is much lower
than full discharge, and gives significant
value with enhanced fault detection
and service life prediction. All visual,
environmental, mechanical and physical
issues are dealt with during the IR test.
6.2.3 Proposed Solution for
Medium to High Critical Nodes
Remote monitoring and remote testing may not be possible on all sites,
or more accurate discharge testing is
required. This would be due to more
critical sites or higher value batteries where absolute certainty about
battery condition is required before
committing to large CAPEX. In this
situation the proposed solution would
be a combination of annual IR tests
with the addition of condition-based
full discharge tests and conditionbased additional IR tests as the battery
approaches end-of-life [Figure 30].
Condition based testing uses the stateof-health information from the annual IR
tests to set trigger points for additional
tests. Discharge tests are most beneficial
when they are performed during the
declining capacity part of the battery
life cycle. As illustrated in the battery
life cycle chart [Figure 12], this decline
in capacity typically corresponds to an
increase in IR of about 25% to 50% above
29
Figure 29. Annual On-Site IR with Annual Remote Partial Discharge
baseline, where typically the capacity
will be at 80% of the nominal specified
capacity. To accurately assess the decline
in capacity, the proposed optimized
solution would perform a discharge test
at 20 to 25% increase in IR, then at 30
to 35% increase in IR, and finally at 40
to 45% increase in IR. The battery life
cycle chart also shows that the rate of
change of internal resistance accelerates
closer to end-of-life; so the optimized
maintenance solution would perform
additional IR tests to capture the accelerated aging. The recommended trigger
point for increasing the frequency of
IR tests would be when the average IR
value for the battery is 25% above the
baseline. Going from one IR test a year to
two IR tests a year should be sufficient,
but the flexibility of the optimization
concept allows for more tests if criticality and budget make it a requirement.
Annual On-Site Internal Resistance Test
Conditional Discharge Test
Figure 30. Annual On-Site IR with Conditional Discharge
EN310TRA-BatOpt / 0113
Battery Optimization Services
Remote Internal Resistance Test
Remote Discharge Test
Annual On-Site
Internal Resistance Test
Conditional Discharge Test
Figure 31. Remote Discharge and IR with On-Site IR and Conditional Discharge
6.2.4 Proposed Solution
for High Critical Nodes
For the most critical of sites where
downtime is just not an option the solution which offers highest confidence
in battery condition and reliability is a
combination of remote discharge tests
and remote IR tests [Figure 31], where
the monitoring also gives 24/7 visibility of any battery or system related
issues, with on-site IR and site discharge
EN310TRA-BatOpt / 0113
tests for all essential visual, mechanical and environmental checks coupled
with absolute proof of capacity.
This solution would use annual remote
partial discharge tests and four remote
IR tests per year as the basis for regular
condition assessment. To ensure all site
visit related tasks are also covered, this
solution would also provide one annual
site IR visit. This combination provides
ongoing state-of-health assessment
along with good and regular estimates
of capacity. However, if absolute con-
firmation of capacity is required prior
to committing to a full battery replacement then an on-site full capacity test
can be added close to end-of-life.
These combinations will give the full
benefits of 24/7 monitoring coupled
with the confirmation from site visits
and the essential visual, physical and
mechanical checks that are required
for maximum battery reliability.
30
Battery Optimization Services
7.0 Optimized Corrective
Maintenance
The predictive and preventative
elements of an optimized battery
maintenance program are an essential
part of a complete package but for
maximum effectiveness the corrective maintenance elements should
form part of the overall program.
7.1 Replacement Policy for
Single Cells or Monoblocs
Single cells or monoblocs with identified faults should not be left in a battery
string as they will often cause issues for
the other cells in the string. If a cell fails
to short circuit, the float voltage applied
to the other cells will increase, the current will increase and the aging process
will be accelerated. Conversely, if high
internal resistance causes a cell to fail,
the string will not be able to sustain the
correct discharge within the normal
voltage limits. Healthy cells in the string
could be undercharged and would not
be able to reach 100% state-of-charge.
Sustained exposure to undercharging
can lead to sulfation and permanent
damage. Single cell faults may be the
result of mechanical or manufacturing defects or accidental damage, and
may not be detected until months
after installation when the battery is
needed to perform a full duty cycle.
As a battery ages, the failure modes are
more likely to be caused by factors that
are affecting all the cells in the string.
A decision to replace larger numbers of
cells needs to be based on the remaining
service life of the healthy cells against
the cost of replacing the whole string.
New cells inserted into old strings will
tend to age faster than the existing cells.
This is principally because as batteries
age, the float current increases. This float
current must flow through all the series
connected cells, even the new cells in the
string. This leads to overcharging and
accelerated aging of the new cells. The
decision to replace the whole string or
31
individual cells needs to be judged on a
case-by-case basis but generally if more
than a quarter of the cells are at end-oflife, the whole string should be replaced.
One technique that is often used for
critical UPS back-up systems is to have
‘hot standby’ batteries with a number
of identical blocs kept on float service
but not connected to any loads. These
hot standby batteries will age at the
same rate as the in-service batteries.
When a fault occurs with an in-service
battery it can be quickly replaced with
a hot standby battery, which will have
minimal impact on the string equilibrium. In practice, this approach is rarely
adopted in telecommunications applications. Should a single string fail, it is
more common to have multiple parallel strings connected to the system to
provide redundancy. If cell or bloc faults
develop across multiple strings, the load
could be exposed to drastically reduced
battery back-up, as it only takes a single
cell fault to prevent a whole string from
delivering power. In certain circumstances it is advisable to consolidate all
healthy blocs into contiguous strings
and then remove or isolate all faulty
blocs in string groups. Managing healthy
and faulty blocs in this way improves
battery reliability and helps prevent premature failure of battery installations.
7.2 Connectors and Torque Settings
The correct torque setting for all connectors is essential for reliable performance
from the battery during a discharge.
Some manufacturers recommend two
torque settings: an installation torque to
be used when the battery is new and has
just been installed and commissioned;
and a maintenance torque setting to be
used when performing regular maintenance on the battery. Maintenance
torque is often lower than the installation
torque setting. Bolts and connecting hardware are subjected to routine
mechanical stresses, vibration, temperature changes, and with some older post
designs, the lead will creep under load.
Over time the connections will loosen.
If this is not corrected, the contact resistance will increase at the connector to
post interface. Increased resistance will
lead to reduced battery performance
and, most significantly, heat. The current flowing through a poorly torqued
connection during a discharge can
cause severe overheating. In extreme
cases, battery fires can be directly
caused by high resistance connectors.
Figure 32. Effects of Overheating
In Figure 32, the effects of overheating
can be seen as a discolored connector
link, washer and bolt (dulled instead
of bright). The red deposits around
the terminal are where the red plastic
cover that identifies the positive post
has melted. The blue deposits are copper sulfate where electrolyte has leaked
and reacted with the copper in the connectors. The pillar may have developed
a leak, which in turn has caused the
corrosion observed, or increased temperature from a bad connection may
have led to the pillar developing a leak.
EN310TRA-BatOpt / 0113
Battery Optimization Services
With appropriate techniques and test
equipment, it is possible to measure the
resistance of each link in a string. Any significant change should be investigated,
as it can be caused by either a loosening of the connectors or possibly by
contamination from leaking electrolyte
and subsequent corrosion or oxidation.
Occasionally re-tightening to maintenance torque is not enough, especially if
the surface has become contaminated
and oxidized. Care is required, however,
as re-setting the torque can actually
lead to post damage. If maintenance
personnel routinely re-torque to the
higher installation setting, there is a
risk that the post becomes distorted
and damaged over time. If the battery
manufacturer specifies one torque
value for installation and another value
for maintenance, then the installation
value should only be used when either
the battery is first installed, or if after a
re-torque to maintenance values the link
resistance is still too high. In this case
the link and connecting hardware must
be stripped, cleaned until bright, coated
with a suitable conductive, oxide-preventing grease, and then re-assembled
using the installation torque value.
7.3 Cell Leakage
Cells or monoblocs may develop leaks
from defective pillars, container-to-cover
defects or from mechanical damage.
If the damage and leak is obvious
[Figure 33], as in this bloc with a cracked
case, then it should be easily identified
with a visual inspection and replaced
as soon as possible. Sometimes,
however, the damage and leak is not
visible if it is at the rear of the installation, against a wall or inside a closed
cabinet and not easily seen. However,
in many cases they can be detected
by changes in internal resistance,
open circuit voltage or float current.
EN310TRA-BatOpt / 0113
deposits built up until a conductive path
was formed between the battery post
and the shelf (which was grounded).
Figure 33. Leak Caused by Cracked Case
Another danger of this type of fault is
ground fault currents. A leak or other
external deposits of electrolyte, sulfates
or just a general accumulation of grime
can create conductive paths where
parasitic currents can flow between the
battery and ground through the rack or
shelves. Ground fault currents can cause
excessive current to flow through all
series connected cells, leading to overcharging, overheating and accelerated
aging and, occasionally, thermal runaway. Replacing single blocs can prevent
premature failure of the entire battery.
Figure 34. Sulfate Deposits from
Electrolyte Contamination
This created a ground fault current
which caused burning around the post
and damage to the metal shelf [Figure
35]. The increased current also passed
through all the series connected cells
leading to further overheating, leakage and corrosion. In this case, the
only option is to completely strip,
clean and rebuild the battery.
7.4Cleanliness
Cleaning batteries is an essential step
in any maintenance regime. Apart
from the potential for oxidized links
and poor connections, there is a very
real possibility of ground faults occurring if regular maintenance does not
include a wipe down with a clean cloth
and water (with a solution of sodium
carbonate to neutralize any acid leaks).
Figure 34 shows a deposit of lead sulfate (white) and copper sulfate (blue)
caused by electrolyte contamination.
The electrolyte came from leaking pillar seals, possibly due to manufacturing
defect or possibly due to overheating.
The maintenance on this battery did not
include cleaning; consequently, these
Figure 35. Burning Around Post
Caused by Fault
32
Battery Optimization Services
8.0Conclusions
VRLA batteries need regular maintenance and inspection to ensure that they
are always able to provide sufficient
capacity to support the load under all
conditions, consistent with the specified requirements. The various failure
modes of VRLA batteries are more difficult to detect than for VLA (flooded)
batteries, where visual inspection of the
cells is possible and new techniques,
particularly internal resistance measurements, have been developed over a
number of years to determine battery
condition or state-of-health. The Ohmic
measurements; either conductance,
impedance or DC internal resistance
measurement, have been shown to correlate with battery capacity and can
be used to make a good assessment of
battery condition. These Ohmic tests
may be carried out on-site or remotely.
Capacity checks are still needed for a
more accurate measure of battery condition. These tests may also be made
on-site or remotely using the system
load. Capacity checks give a measure of
the battery condition by comparing the
discharge behavior with manufacturer’s
data. The preferred delivery method is
remote testing for economic reasons. For
a more accurate determination of battery capacity, an on-site discharge test
is needed where the battery is discon-
33
9.0References
nected from the plant and connected to
an external load bank. This will provide
a precise figure for the battery capacity at the time of the test. In addition
to these measurements, visual inspection of the battery and a check on the
mechanical integrity of the connections
needs to be made from time to time.
The various testing methods may be
packaged in different combinations,
leveraging the optimal technique(s) for
the site with regard to cost and accuracy. An optimal battery management
program will be reached when the tests
are applied in a dynamic way over the
battery life based on earlier test results,
age of the batteries and the criticality
of the site. Larger sites with batteries
protecting power supplies for larger
switching centers, data centers and
important nodes in the telecommunications network need to be assured
of the highest levels of reliability. A
comprehensive battery optimization
solution must include IR and capacity
trend reports as well as information on
all other observed parameters. This gives
the user a clear indication of the battery
condition, a forecast of remaining life
and recommendations for corrective
action to ensure battery reliability.
[1]G J May; Stationary Batteries, in
Jurgen Garche, Chris Dyer, Patrick
Moseley, Zempachi Ogumi, David
Rand and Bruno Scrosati, editors;
Encyclopaedia of Electrochemical
Power Sources, 4; Elsevier,
Amsterdam, 2009, 859-864.
[2] D
Linden and T B Reddy, Handbook
of Batteries, 3rd Edition, 2001,
McGraw-Hill, New York, 2.26-2.29.
[3]D O Feder, T G Croda, K S Champlin
and H J Hlavac, Intelec 92, IEEE,
Washington, 1992, 218.
[4]D O Feder and M J Hlavac, Intelec
94, IEEE, Vancouver, 1994, 282.
[5]K Kozuka, K Takano, Y Konya
and Y Kawagoe, Intelec 97,
IEEE, Melbourne, 1997, 397.
[6]M Kniveton and A I Harrison, Intelec
98, IEEE, San Francisco, 1998, 298.
[7]K Takahashi and Y Watakabe, Intelec
03, IEEE, Yokohama, 2003, 664.
[8]Enersys, Ohmic Measurements
as a Maintenance Tool for
Lead Acid Stationary Cells,
White Paper, August 2005
[9]C&D Technologies, Impedance
and Conductance Testing,
White Paper 41-7271
EN310TRA-BatOpt / 0113
Battery Optimization Services
10.0Annex
10.1 Design Life
STANDARD COMMERCIAL (3-5 years):
This group of batteries is at the consumer end of standby applications,
popular in small emergency equipment.
GENERAL PURPOSE (6-9 years): This
group of batteries is usually used when
an improved life is required in comparison to the Standard Commercial
product; or in cases where operational conditions are more severe.
HIGH PERFORMANCE (10-12 years):
This group of batteries is used
where high power, long life and high
safety standards are required.
LONG LIFE (12 years and longer): This
group of batteries is used in applications where longest life and highest
safety standards are required.
10.2 Cycle Life
Most battery manufacturers will provide
a specification for cycle life, which gives
an indication of how many duty cycles
a battery can deliver to a given DoD.
For an OPzV battery, this specification
12000
11000
10000
Number of Cycles
Battery manufacturers often specify batteries based on definitions of design life
provided by the European battery trade
organization, Eurobat. The design life
is the estimated life determined under
laboratory conditions, and is quoted at
20°C ambient temperature using the
manufacturer’s recommended float
voltage conditions. Design life is categorized in four main groups as follows:
9000
8000
7000
6000
5000
4000
3000
2000
1000
0
10
20
30
40
50
60
70
80
Depth of Discharge (%)
Figure 36. Cycles Versus Depth of Discharge
gives 11,000 cycles to a DoD of 10% and
~1000 cycles to a DoD of 80%. In standby
applications the number of duty cycles
is not generally a life-limiting factor.
actual life. A battery with a 10-12 year
In a cyclic application [Figure 36] the
number of available duty cycles at a
given DoD and also at a given rate of
discharge and temperature will be a
factor in determining service life.
that the battery condition is satisfac-
design life may need to be replaced after
5-6 years of service or less in extreme
conditions. The only way to be sure
tory is to have an ongoing maintenance
program of battery monitoring, testing
and surveillance. This will ensure that
the battery can always support the load
10.3 Service Life/End-of-Life
Service life to end-of-life is defined as the
point at which actual capacity reaches
80% of nominal capacity. Service life is
usually significantly less than the design
life. Environmental factors, installation and operational conditions and
maintenance have a strong influence on
within system design limits and that the
reliability of the network is assured.
High temperatures are very damaging
for batteries. Design life is specified at
20° to 25°C ambient temperature; every
10°C above this reference effectively
halves the service life of a VRLA battery.
EmersonNetworkPower.com/EnergySystems (North America)
EmersonNetworkPower.eu/EnergySystems (EMEA)
© Emerson Network Power Energy Systems North America 2013.
Business-Critical Continuity™, Emerson Network Power™, the Emerson Network Power logo, Emerson® and Consider it Solved are service marks and trademarks of Emerson
Electric Co. EnergyMaster ™, eSure™, NetPerform™, NetReach™, NetSpan™, NetSure® and NetXtend™ are trademarks of Emerson Network Power Energy Systems North America.
34
EN310TRA-BatOpt / 0113