Feature
B
attery Testing — Then and Now
M
uch has been written about battery chemistry and battery maintenance over the years. During the last ten years much has changed
in our understanding of batteries and how they can be tested more
effectively and efficiently at less cost and higher reliability.
A stationary battery is needed for two basic reasons: to support critical
equipment and to protect revenue streams.
Some of the typical battery systems are substation batteries, electric
lighting units, fire and security systems, telecommunications systems,
hospitals, airports, UPS, and financial data.
A battery system includes the supported equipment, the rectifier/
charger, and the battery. The battery itself has two main components:
cells and intercell connectors. Cells in a bank behave similarly to resistors in a series-connected circuit. Parallel banks also behave similarly to
parallel resistor circuits. Ohm’s law rules. As with any series circuit, if one
component or resistor opens the entire circuit opens. The same is true of
a bank of batteries; that is, if any one cell fails then the whole battery fails
to support the load. The battery charger has a dual role in that it keeps
the battery charged and also supplies the load while the ac is on.
Existing Methods
There are two battery terms that need to be defined: state-of-charge
(SOC) and state-of-health (SOH). State-of-charge indicates, as its name
implies, whether the cells and the bank are fully charged or at some state
of discharge as measured by voltage. State-of-health is an indicator of
whether the bank can support the load during an outage as measured by
capacity. All testing methods indicate either SOC or SOH but not both. In
order to obtain reliable measurements, the bank must be fully charged.
Many battery users have relied on voltage and specific gravity measurements only to find that the battery frequently fails in spite of having
normal cell voltages and specific gravities. Basically, voltage and specific
gravity indicate the location of the sulfate in the battery. If the sulfate is on
the plates, then the cell will have a low voltage and a low specific gravity.
If the sulfate is in the acid, then the battery will have a normal voltage
and normal specific gravity. The basic chemical reaction is:
Pb + PbO2 + 2H2SO4 = 2PbSO4 + 2H2O
(charged)
(discharged)
Winter 2003-2004
by Rick Lawrence
Megger
That is not to say if the cell is
fully charged, an abnormal voltage
provides no SOH information. On
the contrary, an abnormal voltage
can provide much information
about SOH. It is just that normal
voltages indicate nothing about
SOH, only SOC.
Abnormal voltages can indicate
several factors about the condition
of a battery. If the voltage is high
then that cell is probably compensating for a weaker low-voltage
cell. Abnormally high voltage
means that the cell is being overcharged, and this overcharged
condition, over time, will shorten
the cell’s life. If a cell has a low voltage (while the others are normal)
this can mean any one of several
factors, the first being that it is
undercharged and is sulphated,
a temporary defect. An equalization charge will usually reverse
this problem. The second possible
low-voltage defect is a short circuit
caused by a paste lump or material
lodged between the plates or sedi-
1
ment build-up creating a shorted electrical pathway.
A short circuit becomes obvious at approximately 2.05
volts dc. An equalization charge may reverse this defect as well but more than likely will not. A short circuit
is typically a permanent defect requiring the cell to
be replaced. The amount of sediment is a function of
the number of charge-discharge cycles a battery has
endured, whether from testing or from actual usage.
Sediment is a sloughing off of active material due to
cycling, and as it builds up in the bottom of a cell it
can accumulate to the point of shorting across the bottoms of two opposite polarity plates. In rare cases, a
soft bumping of the cell may compact the sediment
for the short term but that will not fix it for long. This
defect will usually be seen in a number of cells, not just
one or two, since the entire bank will be experiencing
the cycling.
Temperature is frequently overlooked since it usually means some form of HVAC, which is expensive,
should be installed. As temperatures increase, lifetimes
decrease. Arrhenius, a Swedish chemist, discovered
that for every 18°F (10°C) increase in temperature,
the battery life is halved. Unfortunately, reducing the
temperature by 18°F does not reclaim the battery’s life.
If HVAC is too expensive based on the cost-benefit
analysis, then a thermostatically-controlled fan would
be a minimum to control high temperatures.
There are two alternatives to full-duration load
tests, and they are two forms of a “partial” test.
(Duration refers to the time of a test.) The example
used here is a 100 Ah battery designed to provide
12.5 amperes for eight hours and 55 amperes for one
hour. The reason for such a discrepancy in current
at the two durations is the design of the battery. A
long-duration battery is not very efficient at shorter
rates. This has more to do with plate thickness than
any other parameter. The first partial test is a load
test performed for a short duration (one hour) but at
the long-duration rate (12.5 amperes). The other is a
short-duration test (one hour) at the commensurate,
short-duration rate (55 amperes). Using the battery
example -- in the first version of partial test, an eighthour rate is used but the test is performed for only one
hour. In the second version, a one-hour rate is used
for the one-hour duration.
The first version of partial-load test will supposedly
indicate whether the battery will last by using a prediction of run time based on the voltage curve. Figure 1
depicts the differences in the possible scenarios of run
time based on voltage measurements.
Discharge Testing
Another test upon which battery users have relied
is discharge testing. Although in recent years, with
cutbacks in personnel and capital expenses, discharge
testing has declined. Discharge testing does, indeed,
provide accurate information about the SOH of a battery, especially in conjunction with IR thermography
during the discharge test. IR thermography shows
high resistance connections. If correctly performed, a
discharge test is expensive and time consuming and
generally provides very little predictive information.
For example, in a substation battery that needs to operate for eight hours and is the only battery installed,
a second battery should be connected to the system
in the event an ac outage occurs during the discharge
test. All of the voltage leads must be connected to the
main battery to monitor voltage during the discharge.
It takes one full day to run the test, then two-to-three
days to fully recharge the main battery. At that point,
the second battery can be disconnected and removed.
All-in-all, a discharge test can take three-to-five days
with some overtime. Now consider that in a utility
transmission and distribution network there can be
several hundred battery strings or more. Furthermore,
it is possible that if the battery is in bad condition, it
will fail during the test. At this point, an emergency
battery should be purchased and installed immediately, dramatically increasing battery maintenance
costs.
2
Figure 1
The blue curve (with circular points) is the fullduration load test at the time-appropriate rate. If the
test is stopped before the full duration, then the red
(triangular points) and green (square points) curves
depict two possible run times which are not even
close to the actual blue curve run time. Since there
is so much variability in the possible scenarios, it is
not worth the effort to perform this version of a load
test except that the very worst of cells will show a
considerably lower voltage and would have probably
failed in an actual outage. In that respect, it is a valid
test — it found a weak cell.
The second version of a partial test is much more
effective at determining actual battery capacity since
it is using the actual rate tables from the battery
manufacturer. To save time, do a one-hour test at the
one-hour rate. This can tell much about the battery’s
capacity and is an acceptable test. The caveat, however,
NETA WORLD
is whether the bus work and intercell connectors, etc.,
can tolerate 55 amperes of current for one hour. The
other drawback is that all of the preparation work still
needs to be done, i.e., all of the voltage leads, a second
battery, a load bank, etc. So why not do the appropriate
test, other things being equal?
As the time between discharge tests is extended, the
cost of testing decreases but the risks can then increase.
The cost-benefit analysis enters the financial picture.
So if voltage is mostly inadequate for determining
state-of-health and discharge testing is expensive with
little predictive value, how is one expected to have
confidence that the battery will support the load during an outage?
New Test Methods
There is a relatively new test that was developed
and patented in the late 1980s by Commonwealth
Edison and was commercialized in the early ‘90s. It
is called impedance, and it basically views the cells in
a bank of batteries as resistors in a series connection.
The higher the impedance of a cell, the less current
it can generate (when a voltage is applied). Impedance is just ac resistance. Impedance includes terms
for inductance, capacitance, and frequency but at a
constant frequency, this term becomes a constant.
Changes in measured impedance are then due to
actual differences in the internal impedance of a cell
and not due to changes in frequency. By applying an
ac current signal into a cell or across the entire bank,
and measuring the ac voltage drop, impedance can be
calculated using Ohm’s law, Z = V/i.
There is a strong correlation between battery capacity and impedance, an internal ohmic test. As capacity
decreases, impedance increases. Internal impedance
can find weak cells in a battery bank with extremely
high reliability. The Electric Power Research Institute
conducted a study over about four years on approximately 30,000 cells comparing the value of internal
ohmic testing, including impedance to battery capacity. The study determined that impedance is an
excellent tool for filling in the gaps in load tests while
simultaneously decreasing battery back-up risks. In
other words, it reduces the cost of battery maintenance
and reduces risk.
This is not to say that discharge tests are passé.
Impedance, with its excellent ability to find weak
cells, can greatly reduce the frequency of performing
discharge tests. In addition, by decreasing the frequency of discharge tests, battery maintenance costs
are also reduced. In order to get a better correlation
between capacity and impedance, it is recommended
an impedance test be performed directly preceding a
discharge test.
Winter 2003-2004
There are a couple of other tests that are quick
and provide useful information about the battery:
float current and ripple current. Float current is the
current used to keep the battery fully charged while
ripple current is an artifact of the charger. In the case
of VRLA cells, impending thermal runaway can be
seen by an increase in float current. Depending upon
the battery model, the increase in float current before
thermal runaway occurs can be from a 50 percent to
400 percent. Furthermore, the time between an increase in float current is found and thermal runaway
occurs can two weeks to two months. Measuring float
current monthly, if practical, is recommended; otherwise, every three months is the least often it should
be measured. It is noteworthy that flooded batteries
cannot thermally run away due to the large volume
of acid. It just boils, releasing the heat as steam rather
than heating the lead.
Another test that is recommended is measuring
the ripple current. Ripple current is a function of
the charger which is designed to convert ac into dc
and keep the battery charged while simultaneously
driving the load. The conversion process is not 100
percent efficient, and some ac will carry over into the
dc bus as ripple current. Generally speaking, the battery manufacturers have stated that the upper limit
of ripple current is about 5 amperes rms for every 100
ampere-hertz of battery capacity, so about five percent.
Above five percent heating of the battery will occur,
thus reducing battery life in accordance with the Arrhenius equation mentioned previously. As chargers
age, ripple current will slowly increase, but if a diode
fails then it can double or even triple. Without measuring ripple current, no one would ever really know
that the charger has a problem that may be shortening
the life of the battery.
There is one other test that is not given a lot of
importance but is quite critical. It is a simple visual
inspection. A visual inspection is important to find the
physical defects of a battery regardless of whether the
battery is flooded or sealed. Visually inspect flooded
batteries for post-seal leaks, corrosion, and the general
appearance and condition of the battery. Although it
is not possible to see inside a sealed battery, a visual
inspection is just as important for detecting bulging
cells, post-seal leaks, and corrosion.
Once all of these data have been collected, how are
they to be analyzed? A snapshot in time can indicate
a great deal about the condition of a battery. However,
the best method is to accumulate a database and trend
battery data. A reporting mechanism is extremely useful. Management can be easily informed of battery
condition; it also allows for budgetary planning of
battery replacements (rather than emergency replacements). When all is said and done, the most prudent
battery-testing regime is one which keeps maintenance
costs low while simultaneously reducing the risk of
3
a battery failure. The basic measurements are cell
and string voltages, cell impedance, cell and ambient
temperature, float current, ripple current, and the occasional discharge test performed on a periodic basis
in addition to analyzing the data to prevent battery
failures.
Rick Lawrence is the Marketing Manager for the Battery
Group of Megger, a manufacturer of electrical test and measurement equipment. He has a BS in chemistry from Randolph-Macon
College and an MBA from St. Joseph’s University. He has extensive
lab experience, and for the last several years has been involved
with lead-acid batteries with a major battery manufacturer and
Megger. Previous papers cover topics such as tubular batteries,
x-ray spectrometric analyses of metals, and battery capacity via
battery impedance data.
4
NETA WORLD