Continuous battery monitoring improves UPS reliability, reduces costs-EEWORLD

Collect

Western societies have become increasingly accustomed to very high reliability of basic services. We expect infrastructure such as power grids, telecommunications and IT networks to be available when we need them, and when something goes wrong or the service provider doesn’t deliver the service we expect, we assume that the service level is seriously inadequate. We sometimes talk about “five nines” or “six nines” availability, but we don’t usually think deeply about what they really mean. In reality, 99.999% availability means that the power outage time in a year is about 5 minutes; while “six nines” only allow for an outage time of half a minute in a year.

Therefore, many critical infrastructure and business systems use uninterruptible power supplies (UPS) as backup power. When the power utility fails, the UPS accurately switches to the on-site power supply. The UPS market specializes in serving small-capacity systems, from 1 kVA or less to tens of kVA, which can support the work of 1 or 2 computers; and medium-sized devices of tens to hundreds of kVA are used to support the IT infrastructure of office buildings. In recent years, many large-scale devices with power of several million watts have emerged to support data centers and ensure the continuous availability of online services. In recent months and even years, large server farms and data centers have become the most important applications, and UPS continues to serve the long-established critical system market: such as operating rooms and air traffic control.

UPS are classified by their power output, not their capacity, which means how long they can supply power. UPS can be used to replace power for a long time, up to several hours; or it can be used to provide "bridging" power, that is, to start supplying power immediately when the power fails until a power source such as a diesel generator starts and reaches a stable output, and then stop supplying power, which lasts only a few seconds or minutes. There is another short-term application mode, which requires not only the UPS to support continuous operation, but also to provide sufficient power to ensure the scheduled shutdown of the customer system.

Battery packs power most UPS

Although UPS can use other technologies (such as flywheel energy storage), most UPS still use battery packs for energy storage. Battery packs not only provide considerable capacity, but also can provide power almost instantly. For UPS to operate reliably, the battery pack must be fully charged and in good condition.

Despite many advances in battery technology in other areas, the batteries used in UPSs are still the oldest chemistry, lead-acid batteries. There are many reasons why lead-acid batteries have never been replaced in backup power. In terms of the amount of electricity stored per unit, lead-acid batteries have a high energy density (although not outstanding by today's standards); they have a very high power density and can deliver large currents as needed without damage; although their relatively large weight makes them unsuitable for many portable applications, it is not a big problem for UPS equipment; especially, perhaps because of their relatively low cost. In addition, from today's environmental sensitivity, almost all of the components of lead-acid batteries are associated with "green", especially the lead, which is completely recyclable.

The charging and life cycle characteristics of lead-acid batteries are also very suitable for UPS use. VRLA (valve-regulated lead-acid) batteries can not only tolerate continuous "float" charging, but their service life can be maximized as long as they are always kept fully charged and rarely deeply discharged. However, the service life of a battery pack is limited, and if environmental conditions (especially temperature) are out of the optimal range, the battery life will be greatly shortened. The batteries on most equipment can be replaced regularly according to the shelf life, usually every 5 years. However, this approach has its drawbacks: batteries operating in unexpected environmental conditions may fail faster, while well-maintained batteries may have a longer service life.

Fault Tolerance

Modern UPS needs to provide high power output, so many battery cells are required to be connected in series on a large scale. The failure of a single battery cell can cause the failure of the entire series battery pack. Large and medium UPS implement redundancy to ensure that a single battery failure does not cause the entire UPS to fail. The UPS will continue to operate, but the peak current output will be reduced, and the time the system can run using the UPS will be shortened. In addition, the failed battery may also damage other cells in the battery pack, reducing their service life.

Battery monitoring and maintenance represents a significant cost associated with UPS operation. Engineers are typically required to visit the site on a regular basis (perhaps monthly) to measure the electrical characteristics of the batteries within the system. Batteries that are not functioning properly are usually identified and replaced by measuring the battery voltage. Output voltage is not always a good predictor of battery failure. It is easy to identify a battery as likely to fail when its terminal voltage is significantly below its nominal value, but lead-acid batteries can still exhibit satisfactory terminal voltages when their capacity is reduced or in the early stages of failure. Therefore, batteries may also fail between regular maintenance cycles, requiring additional inspections by engineers.

Continuous monitoring reduces costs

Continuous monitoring of batteries can reduce the time engineers spend on personally checking the status of each battery group, thereby improving their on-site inspection efficiency, and can also achieve preventive maintenance. By identifying potential battery failures, engineers can replace batteries during routine inspections, thereby ensuring higher reliability and avoiding emergency inspections by engineers.

Figure 1: Output voltage of a battery cell measured by a continuous monitoring system

Figure 1 shows the combined monitoring results of a battery measurement on a broadcast device equipped with an 800kVA UPS. The graph indicates the output voltage of several cells in a series battery pack. In this example, each series battery pack consists of 200 cells (in other articles, cells are also called battery packs: multiple battery cells enclosed in a housing) and can provide a voltage of about 440V. There are considerable fluctuations in the voltage because the battery pack is not configured correctly: this will be discussed below.

The graph clearly shows that one cell was delivering 2V instead of the nominal 2.2V. Although the cell was delivering a lower voltage than expected, the difference was relatively small (within the normal acceptable range) and was stable. This behavior is typical, and using output voltage as an indicator of impending battery failure is no longer reliable because the voltage value can remain within the threshold range and thus not trigger an alarm. This is what happened in this example; when this data was collected, the monitoring system was used to evaluate the effectiveness of a scheduled maintenance program rather than to warn of potential problems. Because no action was taken, the cell later failed catastrophically during the period covered by Figure 2; the horizontal axis indicates the date the test was performed. However, until the point of failure (the cell voltage dropped to 0.7V), the voltage of the failed battery group remained essentially constant, with no signs of impending failure. Finally, maintenance personnel replaced the cell in November, and the voltage subsequently returned to the average level for the entire battery array.

Figure 2: Battery cell completely fails – voltage barely drops before failure

This proves that output voltage is not an accurate predictor of a possible failure: another parameter, impedance, is a better indicator. As shown in Figure 3, the graph shows that impedance is on the rise in June, and by early July the value had increased by more than 20%. This trend easily shows that measuring impedance can detect problems 3 months before battery failure. With this data, the battery can be replaced during regular preventive maintenance, rather than waiting for it to deteriorate and fail.

Figure 3: Retrospective investigation of battery cell impedance trajectories proves that impedance is a better predictor of failure

Online impedance measurement

Voltage measurements are simple and can be made while the battery is running alone; using the latest monitoring technology, impedance can also be accurately measured using non-intrusive methods. These graphical overlays show the output of a monitoring system taking periodic measurements, with the result being a calibrated waveform with specific frequency characteristics, on top of the float voltage. The voltage and current measured at these frequencies reflect the basic performance of the battery pack. The waveforms tested reflect any potential changes in cell performance, but even so, considering only the test voltage gives limited warning until the point of failure.

Continuous monitoring of the batteries also provides other useful information for improving UPS reliability. As shown in Figure 1, it is clear that there are many charge/discharge cycles (indicated by the pulses on the voltage trace). Although all battery packs need conditioning, this one is discharged too frequently, 4-5 times per month. While some battery conditioning can extend life, too many discharge cycles can reduce life: two to three discharge cycles per year are a normal configuration. Typically, battery cells are warrantied for a lifespan of 20 to 50 cycles. In this case, we are considering that the battery may exceed this warranty after only a few months, and a battery replacement plan every 5 years may mean that the battery needs to go through several times more discharge cycles than the design warranty.

Frequent charge/discharge cycles in the field are caused by the installer setting the UPS in commissioning mode, which cycles the battery frequently for testing. This surprisingly common mistake can significantly shorten the battery life. The incorrect setting may not be obvious during continuous automatic monitoring by engineers, but the resulting problems are obvious.

Another cause of shortened battery life is high temperature. Even a small increase in temperature can increase the probability of unwanted chemical reactions in the battery, which can eventually lead to battery failure. Typically, battery manufacturers mention a battery operating temperature of 20 °C. Figure 4 shows the ambient temperature over time in the system, at one point reaching 22 °C. Failure of the air conditioning system to maintain the temperature within an acceptable range can result in shortened battery life. Also, elevated temperatures can void the battery manufacturer’s warranty.

Figure 4: The same system is used to collect temperature monitoring data and issue an alarm when the temperature exceeds the range.

We have found that long-term monitoring of batteries through monitoring systems has many advantages in addition to making on-site inspections more efficient and reducing costs. In this example, automatic monitoring of battery impedance can identify batteries that are about to fail three months before they fail.

Continuous monitoring also makes it easy to identify UPS configuration problems: in particular, incorrect charge/discharge frequencies that can significantly shorten battery life. Monitoring can measure environmental conditions to ensure that battery life is not shortened by the effects of high temperatures.

Monitoring maximizes the life of the battery, reduces the risk of battery failure and saves money by ensuring that the battery pack does not need to be replaced prematurely and that deteriorating batteries are detected early so that they can be replaced before the battery pack is exhausted. Although cost saving is not usually the first goal for critical systems such as UPS, it is important for users to convert their systems to long-term online monitoring because this not only cuts costs but also improves system reliability.

Reference address：Continuous battery monitoring improves UPS reliability, reduces costs

Previous article：How to choose digital camera batteries
Next article：Design of battery capacity estimation in UPS system using internal resistance method

Popular Resources
Popular amplifiers