Imagine you're watching a football game. The wide receiver is in perfect position to catch a pass and score the winning touchdown. You can tell his eyes are locked on the ball, but it hits him in the chest and bounces to the ground. What happened? Although he was obviously monitoring the ball, the player didn't take the right course of action at the right time. Something similar can happen when monitoring electrical equipment. If you have a disconnect between watching and doing, the monitoring becomes pointless.
Good battery monitoring, on the other hand, has three key components: seeing the right things, understanding what you see, and responding quickly and correctly. Someone must watch the monitoring system, correctly interpret the data, and initiate the correct response. Just as in football, this must happen in a quick and expert manner. If you don't have that expert on your team, outsourcing (Sidebar) is a far more cost-effective and reliability-centric option than stumbling around or dropping the ball. Whether you're approaching the process in-house or out, the following actions will help you stay in the game, monitoring your battery systems more effectively.
Watching. Monitors allow us to see what is going on inside and around the battery. IEEE 1491-2005, “Guide for Selection and Use of Battery Monitoring Equipment in Stationary Applications,” (Sidebar on page 30) defines 17 battery measurement parameters. Things to watch, archive, and trend include:
- Voltage segments
- Battery ohmic values (resistance, impedance, conductance)
- Charge and discharge current
- Planned and unplanned discharge data
One costly problem is most people don't know what's normal and what indicates a problem. It is the exception when someone other than a battery specialist gets this right.
Interpreting. When examining battery monitoring data, we can't look at isolated points. We must look at the data over time, in relation to each other and the installation particulars. That is where experience quickly becomes crucial.
Suppose you read what appears to be a high voltage. You might assume something is wrong with the battery. But that voltage might be normal for that battery. How do you know? You need to analyze the degree and rate of change between historical measurements. Turn this around and think of the implications of assuming a “normal” reading means you don't have a problem.
A battery monitoring expert will look at outlying data points, even when they fall within the normal range, and try to determine what factors account for variances. Sometimes, the issues are measurement-related. This means you aren't seeing what you think you see. The ability to catch and correct for measurement errors is a skill developed through experience.
Correctly interpret the data, and you can most economically perform the repairs or replacements needed to prevent failures.
Correcting. Corrective actions can vary from adding electrolytes to replacing major components. If you understand which actions to take, you can cost-effectively prevent loss of critical operations.
Correction differs from maintenance. Most maintenance procedures require performing specified tasks at specified times (e.g., quarterly). But batteries can rapidly decline to failure within weeks. You prevent failure by correcting problems between maintenance sessions and correcting those you don't detect during routine maintenance.
If you have a battery monitoring system in place, such failures can be investigated. But is this the best you can do? Why not use the battery monitoring system to prevent failures in the first place? Using it correctly, you can save money (Sidebar on page 30) and improve reliability. Here are a few things to consider.
Hydrogen lurks. UPS systems are safe and reliable if you perform the required maintenance. In Photo 1 on page 26, see what can happen when you don't. Vented batteries have cost and performance advantages over valve regulated (VRLA) batteries. But you lose those advantages if you fail to do things like accommodate the small amounts of hydrogen these batteries emit.
In a battery room designed for vented batteries, the hydrogen theoretically doesn't accumulate. But in the real world, filters clog and fans fail. If some minor equipment failure renders the ventilation system inadequate, how will you know? And if you don't know, how can you take preventive action? Proper battery monitoring is the solution.
Corrosion kills. Runaway corrosion can produce devastating results (Photo 2 on page 28). In this situation, it caused a battery explosion and subsequent loss of load. This was preventable.
A VRLA is sealed with venting valves, and hydrogen gas fills its airspace. As VRLA batteries age, they corrode internally. When they are called upon to deliver large amounts of current, internal pitting and corrosion can create a spark that ignites the hydrogen, causing an explosion. How can you see and correct for excessive corrosion to prevent an explosion? Once again, battery monitoring is the solution.
Correctly using your battery monitoring system can hugely effect the safety, reliability, and profitability of your facility. You need to honestly assess your internal resources — training, experience, and time — and determine if you can do the job right. If not, outsource the work. If you use your battery monitoring system as a predictive maintenance tool and quickly follow up with the correct response, you prevent failures. But if you don't, failure becomes a matter of when and how bad.
Sidebar: Outsourcing Battery Monitoring
A good battery outsourcing solution has the following features:
- Scalable. Stores data for years
- Flexible. Changes are easy.
- Fast. Analyzes hundreds of sites simultaneously
- Easy-to-use. Accessible from any Web browser
A quality summary report on your battery monitoring system should include drill downs to highlight:
- Failing, critical, marginal, and trending batteries (predictive technology forcing maintenance)
- Power outage occurrences (time and duration)
- Temperature measurements
- Alarm summaries and history
- Battery replacement logs
- Service history and archives
- Asset management items, such as aging, replacement percentages, and replacement recommendations
- Automatic e-mails to designated personnel
Sidebar: The Cost of Failure
The standard for data center power reliability is five nines. That means the facility has reliable power 99.999% of the time. Out of 8,760 hours in one year, there are 5 minutes left over for something to fail. Battery failure isn't built into these calculations.
Computer equipment can withstand a power interruption, if the power comes back in about 20 milliseconds. Let's call this 20-millisecond period a Computer Tolerance Period (CTP). Once a loss of power reaches the CTP threshold, you lose the load.
Costs vary by facility, but in a data center this one event can cost hundreds of thousands of dollars. There are about 15,000 CTPs in 5 minutes. Multiply just a fraction of those by the previous cost per outage per incident and the costs are staggering. Remember, those 5 minutes aren't necessarily going to happen in one lump. In a worst-case scenario, you would have 15,000 CTPs spread out evenly across the year. Fortunately, reality is more kind than that. But reality is expensive — and sometimes deadly — nonetheless.
Sidebar: The Standard
IEEE 1491-2005, “Guide for Selection and Use of Battery Monitoring Equipment in Stationary Applications,” is an IEEE and ANSI Standards publication. This document defines battery monitoring equipment for stationary batteries (vented lead acid, valve-regulated lead acid, and nickel cadmium) in applications such as UPS, telecommunications facilities, and utilities. It defines 17 battery measurement parameters and includes safety, communications, security, software, and electrical references. If you work with batteries, this reference is indispensable.