Today, a typical organisation’s performance is only as good as its ICT resource – which depends in turn on the quality and continuity of its power supply. Accordingly, data facilities of all sizes invariably protect their data processing and communications equipment with a UPS system. This provides protection from short-term blackouts as well as other mains-born electrical disturbances.
However, simply sourcing a UPS of suitable capacity isn’t necessarily enough to provide appropriate protection. What happens if an extended blackout of several hours or more occurs? Below, we look at the factors to consider when accurately matching a power backup system to your particular data facility’s real needs.
We can start by recognising that power supply issues divide into long-term blackouts of extended duration, or short-term problems such as transients, noise, brownouts and brief-duration blackouts.
The long-term blackouts need special consideration because a UPS alone is never sufficient to handle them; no matter how much the battery autonomy, a blackout of longer duration is always possible. Some facilities can tolerate this provided they can shut down safely, but most – especially if the system is processing on-line transactions – must continue running under all circumstances.
If the facility is ‘shut-down tolerant’, it can survive with a UPS alone. During short-term blackouts, the UPS simply switches its inverter to battery power on detection of utility failure. If power is restored well within the battery autonomy, the inverter is switched back to the incoming mains supply. The load has enjoyed uninterrupted operation, with complete protection from the temporary problem.
However if the blackout period starts to threaten the battery autonomy, the UPS must signal the load, allowing it to shut down gracefully in the remaining battery time. System hardware and data has been protected from damage, although the organisation loses its ICT resource.
If such loss is unacceptable, the only solution comprises a UPS in combination with a backup generator, which, with sufficient fuel, can outlast any blackout.
Such generator-UPS combinations need reliable communications between the generator, UPS and critical load, as well as a suitable control system on the generator. During short-duration blackouts, the UPS uses its battery power to avoid unnecessary generator start-ups. If a blackout becomes extended, though, the UPS signals the generator, through an Automatic Mains Failure (AMF) panel, to start up. The UPS battery autonomy allows the generator to reach full speed and synchronise its voltage output with the UPS, while interrupting the load’s power. Once the generator is on line, the UPS uses its power to feed the load and recharge its batteries.
When the UPS detects mains power restoration, it delays, typically for at least two minutes, before switching back to the mains. This ensures that the power is sustained rather than, say, a temporary utility company fault location procedure.
The generator itself must be adequately prepared to perform successfully as a power backup component. It should have an electronic governor to regulate engine speed and therefore the alternator output frequency, to ensure sufficiently fast response and accuracy. Mechanical governors are lower-cost, but lack the responsiveness, accuracy and stability needed by the UPS and critical load.
If the generator’s output frequency is outside the UPS’s critical limits, or is changing (slewing) too fast, the UPS will not synchronise. It will also generate a warning that if a fault occurs the load will not transfer from the UPS to the raw generator supply.
To avoid such problems, it’s important to inform the generator supplier that his equipment is intended for use with a UPS, and ensure that it’s designed and tested accordingly.