Proper backup battery reconditioning methods lead to increased reliability, performance, and savings.
With an aging infrastructure, power outages are a concern across the U.S. It is estimated
that U.S. businesses lose on average $15,709 every 30 min during a power outage, totaling $104 billion annually.
New data center protects its power supply
Not surprisingly, electrical reliability was a top priority at one Detroit company when it constructed a new 5,040-sq-ft data center in mid-2012. Designed to support 48 workstations, the state-of-the-art data center relies on a flywheel system to provide emergency power in case of an outage. Energy storage, in the form of a flywheel, provides backup power until diesel generators can be started. As a part of the system that routes the flow of backup power, a bank of nickel cadmium (NiCd) batteries supports the switchgear, which helps to ensure stable delivery of power to the load until the end of the disturbance.
As such, the data center’s stored energy system is designed to provide:
- Instant availability of supply power to the critical load via the UPS in the event of sags, spikes, complete utility failure, or any other power disturbance that requires a switchover to a backup power source
- Proper sizing to supply the critical load that is normally supported by the utility via the UPS
- Sufficient operating time for backup power to come online (typically the time required for a generator to start and be able to accept load).
While there are several types of rechargeable batteries available for backup applications, the data center designer turned to EnerSys for NiCd batteries to power its switchgear because of their good cycle life, capacity, and performance at low temperatures. The selected NiCd batteries—with a nominal cell voltage of 1.2 Vdc—were specifically designed for mixed loads that include both high and low discharge rates.
Testing one, two, three
The batteries were delivered on Sept. 20, 2012. Almost two weeks later, they were installed in two strings of 92 with two 100-A, 125 Vdc chargers, but were never commissioned. While awaiting commissioning, the data center noticed a significant decrease in the voltage level of the battery bank discharge curve.
Data center personnel applied power to the batteries and ran three separate load tests. During each test, the voltage per cell fell below acceptable limits in the 1-hr testing timeframe. After attempts to troubleshoot the situation, the data center reached out to the battery manufacturer for assistance.
Diagnosis: Floating effect
After reviewing the test data, EnerSys
engineers realized that the NiCd batteries were experiencing a phenomenon known as “floating effect.”
This is a common occurrence. The conditions that can cause this effect include:
- Maintaining the NiCd cell at a fixed floating voltage for a period of time
- Storing a battery for long periods of time without proper activation
- Applying an improper commissioning charge.
When a NiCd cell is maintained at a fixed floating voltage over a period of time as required in switchgear applications, there is a decrease in the voltage level of the discharge curve. Known as floating effect, this process begins after one week at a fixed floating voltage and reaches its maximum in about 3 months. It can yield very poor capacity, with results as low as 10%. It also can significantly decrease available battery run time.
With NiCd batteries, the floating effect cannot be eliminated by a boost charge as it can with lead acid batteries. As such, they need to be fully drained and recharged. This is called a reconditioning cycle. In switchgear and other similar applications that require the batteries to remain on float for long periods of time, it is important to make the reconditioning cycle part of routine maintenance practices.
Following the battery manufacturer’s recommended directions, the data center completed a reconditioning cycle and load test by:
- Discharging the cells as a string to 1.1 V per cell at 211 A for 5 hr
- Recharging the cells as a string at 1.62 V per cell for 30 hr
- Allowing the cells to remain on open circuit for 4 hr before discharge
- Discharging the cells as a string at 456 A to 1.15 V per cell.
During the charge and discharge steps, the data center monitored the cell temperatures, cell voltages, and string voltage. In addition, it verified the voltage readings with an independent monitor, verified the functionality of the load bank before discharging the strings, and verified that the string voltages matched the totals of the individual cell voltages (see Figure 1).
Real-world trial success
The reconditioning charge worked, and the floating effect was completely reversed. The real proof came just a month later when the company lost a utility feed to its new data center and experienced a full power outage. The emergency backup power system worked perfectly.
There is no right or wrong battery for float applications. However, it is important to understand that different battery chemistries perform differently and, therefore, must be sized properly for the load and the application. For floating applications, such as the one described here, it is necessary to consider the derating factor when determining battery size. NiCd batteries offer a cost-efficient option for float applications when properly specified and commissioned—and demonstrate excellent resiliency even after repeated reconditioning.
Jennifer A. Eirich is a marketing manager, Utilities/Rail at EnerSys. She joined EnerSys in 2012 as part of the team responsible for launching the company’s first utility-scale energy optimization system. Eirich has more than a decade of experience in mixing and systems engineering. She is a member of the U.S. Technical Advisory Group to IEC TC120 for the Standardization of Electrical Energy Storage Systems. Eirich holds a BS in Chemical Engineering from Pennsylvania State University.