When planning power requirements and infrastructure for critical systems, facility executives face difficult decisions that can have costly business repercussions. Selecting a system too large for an organization’s needs can lead to inefficient use and an unnecessarily high outlay of capital. A system that does not provide enough carryover protection in a mission-critical situation can set off a cascade of expensive consequences, racking up both direct and indirect costs.
Asking the right questions, understanding the answers and then applying the appropriate systems, can help ensure a cost-effective and risk-appropriate solution.
Uninterruptible power supply (UPS) systems generally fall into four configurations. They are, from simplest to most complex:
The N is the “need” of the critical load, or the minimum backup protection for the critical load.
The simplest configuration, capacity, or N systems, are typically comprised of one double-conversion UPS with a maintenance bypass, sized only to accommodate the critical load. This configuration is the least expensive option, but requires UPS system downtime for maintenance. Also, the critical load is exposed to unconditioned power if the unit fails or is unprotected during maintenance periods. Scheduled preventive maintenance requires that the business tolerate about two to four hours of downtime per year. The lack of redundancy also opens many single points of failure, with system reliability hinging on the weakest link in the system.
There are two types of N+1 configurations: isolated redundant systems and parallel redundant systems. Both employ a single emergency generator and automatic transfer switch but, compared to the N configuration, have an additional level of redundancy at the UPS level. Both types of N+1 configurations have a second UPS system to back up the first.
Isolated-redundant systems utilize series-connected UPS units, each sized to the critical load. While the primary UPS is fully loaded, the secondary unit only assumes the load if the primary UPS fails or is taken out of service for maintenance. This is a relatively cost-effective scenario that enables facility executives to retain flexibility and control with the option to mix and match products from different manufacturers. This especially applies when a system is upgraded from an N configuration.
Parallel redundant systems use dual, parallel-connected UPS units to serve the load, each individually capable of accommodating the entire critical load. The load is divided among the different UPS units. Because additional units can be combined in the same installation, this configuration allows flexibility as the critical load grows. The UPS units typically must be identical in design, rating, technology and manufacturer, and a paralleling board is necessary to provide UPS coordination and synchronization.
For parallel-redundant systems, most manufacturers require both an external static bypass switch to allow the UPS units to share the load and an external maintenance bypass switch, each making the system a bit more complex.
Some additional considerations for both isolated redundant and parallel redundant systems include:
• Single points of failure exist at the transfer switch and its associated switchgear, and also downstream of the UPS output bus at the power distribution unit and its related switchgear.
• The UPS units do not use full capacity, making this configuration less operationally efficient than the capacity system.
• The system requires two to four hours of maintenance downtime annually, exposing the critical load to unprotected utility power. Individual UPS units, however, can be removed for preventive maintenance without taking the entire system down.
A distributed-redundant system achieves nearly complete redundancy without the cost of a system-plus-system configuration. It employs two utility power sources, an emergency generator, and multiple, parallel and independent automatic transfer switches, UPS systems, and UPS system input and output feeders and busses.
This type of system is capable of continuous availability because it can transfer the critical load to alternate UPS busses, always providing protected power to the critical load during maintenance or load reconfiguration. This design also minimizes single points of failure.
However, this solution requires more switchgear to build up the independent portions of the system and multiple utility services; consequently, more space is needed to house the additional hardware and switchgear. The system is also more complex because it is difficult to ascertain which UPS system branches are feeding which loads. Like the parallel-redundant configuration, the critical load is often served via UPS units that are not fully loaded, adding to operating costs.
The highest cost solution in terms of space and money is a system-plus-system redundant configuration. This design is typically installed in stand-alone and specialized buildings such as a central data center building for a corporation. Single points of failure are virtually eliminated, but the costs increase rapidly. This system employs multiple utility sources, emergency generators, automatic transfer switches, UPS systems and power distribution units.
Basically, identical and duplicate infrastructure components serve the critical systems. The multiple and separate power paths allow for any piece of equipment to shut down without transferring the load to utility power. As in some of the other redundant configurations, the systems tend to be less efficient because the UPS units are not fully loaded. The additional components increase maintenance costs.
Understanding configuration options is the easy part. Deciding what type of configuration to use is more difficult. Gaining a better understanding and clarification of system availability, business risk, life-cycle costs, space planning, testing and installation, and maintenance will help facility executive make an informed decision.
One important criterion is the need for system availability. Some businesses can function with very basic protection, using just transient voltage surge suppression (TVSS) or power conditioning for sensitive equipment.
For many companies, however, downtime can result in thousands of dollars of lost revenue. The payback on a system that prevents downtime becomes a primary driver for selecting a system, but the system configuration should be appropriate for the situation.
Compared to very basic protection, the capacity system provides better operational support. This system’s single UPS system provides temporary ride-through protection for short-term interruptions. With adequate battery capacity, it also allows for orderly shutdown of critical systems to prevent data loss if a sustained utility power outage occurs. Depending on the reliability of the utility source, this system results in power availability of 99.9 percent, or less than one hour of unplanned downtime per year.
High availability systems with various levels of redundancy allow a business to continue operations for extended periods, decreasing downtime to between five minutes and three seconds annually (99.99 to 99.999 percent availability). Continuous availability provides conditioned power to the critical load very nearly 100 percent of the time on an annual basis (99.9999 percent availability). High or continuous available systems make good choices where downtime cannot be tolerated or is cost-prohibitive.
Business risk also plays into the availability discussion. In addition to direct monetary costs — for example, revenue lost by a health care provider because downtime caused procedures to be canceled — downtime brings indirect costs, such as the loss of key business-sustaining customers. The configuration should align with an appropriate level of business risk tolerance.
Initial costs always seem to be in the forefront of the decision-making process. Modular systems that allow the system to grow incrementally seem like an attractive option because they reduce initial costs and permit the user to pay for capacity as it is needed. However, when systems are expected to grow significantly, installing the larger traditional UPS system upfront may be more cost-effective than the installation and expansion of a modular system.
In evaluating the life-cycle costs of system options, it’s important to remember that the emergency generator, electrical distribution systems and HVAC systems must grow with the UPS system — it’s not as simple as just adding more modules. Facility executives should be aware of the costs and headaches of maintaining and testing two generators and the batteries for multiple UPS systems, not to mention the related environmental cooling systems.
Maintenance costs should always factor into the decision-making process. Newer flywheel UPS systems can be modularized to achieve redundancy, eliminating battery maintenance, but these systems are not without maintenance needs. Flywheel bearings need replacement on a regular basis, requiring the services of qualified technicians, and vacuum system filters require inspection and periodic changing.
Another important factor to consider is the amount of space required to house the systems. Consider a system-plus-system configuration. Space is required for multiple transfer switches and switchgear (and the clearances required by the National Electrical Code for these devices), redundant UPS equipment, redundant power distribution units, and other equipment. This type of system takes up much more space than a capacity system with its single UPS. Opinions vary as to whether flywheel UPS systems occupy more space than their battery counterparts; project-specific comparisons are required to make an accurate determination. When used, floor-mounted cooling equipment will also occupy valuable floor space.
High availability system components work together to provide seamless operation, making testing and commissioning a more complicated matter for these systems than for systems with discrete and independent components. All components should be tested before they are inserted into a high or continuously available configuration to ensure system interoperability for varying operating conditions.
Working with an experienced manufacturer representative can help to cover the bases of infrastructure issues, such as proper electrical system grounding and electrical device coordination, cooling requirements and integration with building management systems. If an installation is planned on a floor above slab on grade, structural integrity must be verified to allow for proper floor support. A structural retrofit can be disruptive and expensive.
Components outside the telecommunications room should not be forgotten. The two generators required for system-plus-system operation require not only footprint area, but also clearances for controls access and radiator cooling if not ducted. It’s also imperative to keep generators located outdoors at least 25 feet away from HVAC fresh air intakes to help maintain good indoor air quality.
No matter the system, preventive maintenance is required. The number of parts in the system affects its maintainability, as coordination of maintenance becomes more complex. Fewer and larger UPS systems are easier to maintain than a larger quantity of smaller systems.
Manufacturers have readily available resources and easily accessible information to aid the decision making process. Much of the information is available online. Facility executives should understand critical system configurations and also what factors to consider when choosing a particular design. A detailed analysis is the key to providing the simplest, most cost-effective and most appropriate solution to support mission-critical systems.
Mark Hoskins is the electrical engineering manager at Marshall Erdman & Associates, a fully integrated health care facility design-build firm headquartered in Madison, Wis.
