Event Recap
RECAP | ROUNDTABLE | Battery Service and Monitoring
Introduction:
Battery systems continue to be the life blood of most data centers. Poor preventive and predictive battery maintenance can lead to failures and outages. Juan Orozco, Uptime Institute consultant, joins a roundtable discussion that will allow attendees to engage on battery maintenance and service practices, and to also discuss battery monitoring systems and vendors. What are the best-in-class maintenance practices for battery service and load testing? Should battery maintenance be performed by onsite staff or vendors? What are the member preferred battery monitoring systems?
In this roundtable, members discussed battery system alarm thresholds, service and monitoring practices, as well as long term discharge testing, to main a few topics.
Alarm threshold: A member stated they are seeing 30% internal resistance alarms for VRLA batteries in 3-5 years. Batteries have been checked and are found not to be bad. Should threshold be simply increased? Juan Orozco mentioned more investigation should be conducted to find the root cause of the alarm. A member responded that the battery monitoring system is designed to detect “bad” batteries, so raising slightly should not be an issue. Another member indicated manual monitoring should still be conducted at a set frequency even with the battery monitoring system in place. Conduct quarterly checks initially, and as years increase and monitoring indicates issues, increase maintenance check frequency. 30% first threshold is really the annoyance level. When reaches the 2nd level of 45%, battery should be replaced. You can expect issues to occur at about years 6 to 7 and replace in full around year 8.
From a best practice perspective, the question was asked is having only a battery monitoring system enough? The consensus was that best practice is for a battery monitoring system to be coupled with onsite battery maintenance. Vendor maintenance is necessary to support battery monitoring. Many times, vendor maintenance will find an issue that the battery monitoring system won’t detect. This is because vendor maintenance and testing look at more things than what the battery monitoring system checks. Use the battery monitoring system for more of a predictive maintenance type of approach. The owner should also be heavily involved to gain insight.
Additionally, from a best practice perspective, battery load testing based on IEEE standards should be conducted. There were varying views on what is best from a load testing perspective. One member commented they are conducting full load testing the 1st year of installation, then every 4 years. Another member commented he believes IEEE specifies initial full load test within 2 years, then every 2-3 years, and then once you reach 90% life test every year. A member pointed out that pull-the-plug type of testing, or quarterly engine-generator system testing, where you test the system to validate it will automatically pick up the building load does a good job of giving you a read out of battery validity since batteries can sit floating without load and give you a false sense of security. However, this does not substitute for full discharge testing as required by IEEE.
Lastly, it was discussed how published battery maintenance recommendations are hard to find. Juan Orozco stated the IEEE standards for stationary batteries as being IEEE 1491, IEEE 1188 (VRLA batteries), and IEEE 450 (VLA batteries). Juan also searched for additional information, which is included below as attachments for your reference. Both attachments consist of information from Vertiv and are public documents.
Battery systems continue to be the life blood of most data centers. Poor preventive and predictive battery maintenance can lead to failures and outages. Juan Orozco, Uptime Institute consultant, joins a roundtable discussion that will allow attendees to engage on battery maintenance and service practices, and to also discuss battery monitoring systems and vendors. What are the best-in-class maintenance practices for battery service and load testing? Should battery maintenance be performed by onsite staff or vendors? What are the member preferred battery monitoring systems?
In this roundtable, members discussed battery system alarm thresholds, service and monitoring practices, as well as long term discharge testing, to main a few topics.
Alarm threshold: A member stated they are seeing 30% internal resistance alarms for VRLA batteries in 3-5 years. Batteries have been checked and are found not to be bad. Should threshold be simply increased? Juan Orozco mentioned more investigation should be conducted to find the root cause of the alarm. A member responded that the battery monitoring system is designed to detect “bad” batteries, so raising slightly should not be an issue. Another member indicated manual monitoring should still be conducted at a set frequency even with the battery monitoring system in place. Conduct quarterly checks initially, and as years increase and monitoring indicates issues, increase maintenance check frequency. 30% first threshold is really the annoyance level. When reaches the 2nd level of 45%, battery should be replaced. You can expect issues to occur at about years 6 to 7 and replace in full around year 8.
From a best practice perspective, the question was asked is having only a battery monitoring system enough? The consensus was that best practice is for a battery monitoring system to be coupled with onsite battery maintenance. Vendor maintenance is necessary to support battery monitoring. Many times, vendor maintenance will find an issue that the battery monitoring system won’t detect. This is because vendor maintenance and testing look at more things than what the battery monitoring system checks. Use the battery monitoring system for more of a predictive maintenance type of approach. The owner should also be heavily involved to gain insight.
Additionally, from a best practice perspective, battery load testing based on IEEE standards should be conducted. There were varying views on what is best from a load testing perspective. One member commented they are conducting full load testing the 1st year of installation, then every 4 years. Another member commented he believes IEEE specifies initial full load test within 2 years, then every 2-3 years, and then once you reach 90% life test every year. A member pointed out that pull-the-plug type of testing, or quarterly engine-generator system testing, where you test the system to validate it will automatically pick up the building load does a good job of giving you a read out of battery validity since batteries can sit floating without load and give you a false sense of security. However, this does not substitute for full discharge testing as required by IEEE.
Lastly, it was discussed how published battery maintenance recommendations are hard to find. Juan Orozco stated the IEEE standards for stationary batteries as being IEEE 1491, IEEE 1188 (VRLA batteries), and IEEE 450 (VLA batteries). Juan also searched for additional information, which is included below as attachments for your reference. Both attachments consist of information from Vertiv and are public documents.
Request an evaluation to view this report
Apply for a four-week evaluation of Uptime Intelligence; the leading source of research, insight and data-driven analysis focused on digital infrastructure.
Already have access? Log in here
Posting comments is not available for Network Guests