Published on Thursday, Apr 29 2010 by
Most organisations’ sites today play host to a significant level of sensitive ICT equipment. And, with 24/7 online processing of business critical transactions now the norm, protecting this equipment from mains problems becomes vital for business survival. Uninterruptible Power Supplies (UPSs) are typically used in this role, as they provide protection from electrical disturbances such as noise or transients as well as mains brownouts or failures.
But it’s not enough to simply provide protection for conditions prevailing today. Site operators investing in a UPS system will expect their outlay to buy them security for the working life of the installation. Some further expenditure on ‘top-up’ equipment would be acceptable; complete replacement of the UPS installation at regular intervals certainly would not be. Yet complying with this apparently simple expectation is fraught with difficulties. Providing extra capacity for future demand, or oversizing, consumes extra money and floorspace, and may degrade energy efficiency as well. Yet correctly sizing the UPS to the initial load will also quickly lead to problems if the load grows beyond the installed UPS capacity. So what is the solution to a scenario in which the UPS seems destined inevitably to be either oversized or undersized?
The answer comes from an advance in UPS topology known as modular technology, better known as true modular technology to distinguish it from other approaches that do not offer the same UPS reliability or availability. A true modular technology UPS comprises a 19” rack containing a set of entirely independent UPS modules that, operating in parallel, together provide the capacity demanded by the critical load. As time passes, more modules can simply be added to meet any increase in load demand. The incremental increase in capacity added by each module means that, unlike a standalone UPS, total UPS capacity remains closely matched to the critical load size.
We can demonstrate this by example. Suppose we have a load that is currently sized at 120 kVA and that the load’s critical nature demands a redundant power solution. This can be satisfied by any UPS configuration comprising N+1 modules, of which N add up to sufficient capacity for the critical load. Therefore any one of the N+1 modules can fail without compromising the load supply. Using a traditional system, a 2 x 120 kVA configuration would be required to meet the 120 kVA load with N+1 redundancy. By contrast, a modular system could provide 120 kVA N+1 redundancy using 4 x 40 kVA modules mounted within a single UPS rack. Only 40 kVA of extra capacity has been provided, compared with 120 kVA for the standalone system. And this is capacity that may never be called upon at all.
In reality, the surplus capacity gap will probably be even greater. The standalone UPS installers may well have supplied a pair of 150 kVA units to allow for possible future growth. So, while the load is 120 kVA, 180 kVA of extra capacity is actually being provided. On the rackmounting modular system, there is no need to add extra capacity until the 150 kVA is seen as a certain requirement. And even when it is, it can be met in minutes by slotting another 40 kVA module into the rack. The extra capacity is now just 50 kVA.
This flexibility in adding or removing modules from the UPS rack is known as the UPS’s vertical scalability. The UPS can be quickly and easily scaled to maintain its rightsizing to the critical load, even while the load size is either growing or shrinking. Such a UPS rack typically has five module slots available. Since each module can be rated from 10 to 50 kVA, the UPS rack can be configured for any capacity from 10 to 250 kVA, or 200 kVA with N+1 redundancy. If further capacity is needed, this can be achieved by adding further racks as required. The UPS’s support for multiple racks is known as its horizontal scalability. Up to four racks can be paralleled, providing up to 1 MVA capacity.
This example shows how modular topology allows changing critical loads to be tracked closely and easily. In terms of flexible capacity matching, future-proofing is assured. But simply future-proofing in terms of size is not enough; the UPS cannot compromise its reliability or availability as it tracks its changing load. A look at modular technology design shows how it achieves this.
Early UPS designs used transformers in their output stage to boost their output AC voltage to the same level as the mains input supply. Then advances in power semiconductor technology allowed elimination of this output transformer. This brought great improvements in energy efficiency, input power factor and input current harmonic distortion. Of equal importance was the very significant reduction in physical size and weight – a reduction that led to the concept and development of a complete UPS within a portable rack mounting module rather than a large floorstanding installation. And a truly modular system has decentralised architecture, meaning that each module is a complete unit able to operate independently from any other component. Some modular systems yield to the temptation for modules to share some expensive components such as the static switch to cut costs. However such compromises reduce system reliability, since any failure of a shared component will disable not one but all of the modules relying on it.
Reliability is of course vital, yet it is only one factor within what ultimately matters to a critical load: system availability. Availability is defined as a ratio between system uptime and system downtime, or, to put it another way, between Mean Time Between Failures (MTBF) and Mean Time To Repair (MTTR).
Availability improves in direct proportion to reducing MTTR. And, whereas a traditional system takes typically six hours to repair by bypassing the UPS before identifying and rectifying the fault on site, a faulty modular UPS system can usually be restored to full operating capacity in less than half an hour simply by swapping out the faulty module. In fact this is known as hot swapping because it can be done with no need to isolate the UPS from its load. After removal, the faulty module can be repaired off site at any time.
The truly modular approach is not the cheapest future-proofing solution available. However in most applications a consideration of the consequences of failure will show it as the best if not the only solution to use. By slotting in modules to increment capacity, loads can be efficiently, easily and closely tracked as they vary from 10 kVA to 1 MVA. And the modules’ decentralised architecture and hot swappability ensure that module reliability and system availability remain optimised irrespective of the UPS’s current capacity.