Power In the Lab

Power quality and reliability are as critical to research and development (R&D) facilities as they are for production, manufacturing and data centers. A power event in the lab can result in lost data, specimen destruction, equipment failure and lost preparation time—unacceptable in today's costly R&D programs.

Before proper solutions can be developed, the nature of the electrical load and use of the building must be understood. Can a cryogenic freezer be off-line for 10 seconds without damaging specimens? Will a 70% voltage sag trip test equipment or produce erroneous results? Each concern has several solutions with a variety of costs.

Electric Power Research Institute studies have shown that roughly 83% of outages last less than 10 seconds, many occurring during summer thunderstorms, when lightning strikes on or near electrical equipment cause voltage fluctuations. In the case of a direct strike, a short interruption occurs while a breaker is automatically opened and closed through a recloser to clear the faulted line. Trees falling into overhead lines, or high winds slapping them together, may also cause the breaker/recloser operation. Consequently, the first line of defense for any R&D facility is underground feeders from the utility to prevent weather-related power events.

Another source of concern with the utility systems is that they can transmit power-quality problems from one facility to another. For example, an industrial plant with large motors could be the source of transients and disturbances that cause voltage sags, swells, electrical "noise" and other disturbances in a nearby R&D facility. Solutions would include installing surge protective devices, obtaining a dedicated utility feeder or modifying the industrial facility's equipment.

Many problems that affect test equipment can be mitigated by a carefully designed electrical distribution system within the R&D facility. Electrical noise can be caused by motor starters and adjustable-speed drives (ASDs) in a facility's mechanical systems. Separation of load types can be an effective means of mitigating the transmission of electrical noise to critical equipment.

A design that feeds sensitive loads from separate transformers or feeders will attenuate the impedance caused by mechanical equipment in cables and transformers. Maintaining this separation is not difficult, but it does take planning and diligence throughout the facility's life. In addition, panelboards serving critical loads should be located near the load, in an electrical room next to the lab or in the lab itself—in other words, far from loads that generate the electrical noise.

Harmonics can be a particular source of concern in R&D facilities. Harmonics are generated by the non-linear loads of testing equipment and computers—the type of equipment that R&D facilities are full of. Possible solutions include filters, K-rated transformers and panelboards with 200-percent neutrals.

To protect equipment that is susceptible to power disturbances from the utility, other facilities or the R&D facility itself, an uninterruptible power supply (UPS) may be necessary. There are basically two types of UPS: static (electronic) and rotary. Both types provide power-conditioning functions that boost the voltage to normal levels when there is a sag, and provide reduction if there is a swell.

Static UPS units have electronic components that take the incoming alternating current (AC) power, convert it to direct current (DC) and then invert it back to AC to serve the load. When the power is in a DC state, a bank of batteries provides the energy needed during the voltage sags to keep a steady 60-hertz voltage at the load. In case of a power outage, the batteries are also able to provide backup power for a limited duration, normally 5 to 15 minutes.

Rotary UPS units use flywheel inertia rather than batteries to provide energy. The rotary takes the "noisy" voltage and spins a generator that produces a clean voltage at 60 hertz. No upstream electrical noise passes to the load, because the motor and generator are mechanically connected. The generator's flywheel inertia also provides additional power to the load when voltage sags are present. Most systems can provide power to the load for 15 to 30 seconds, long enough for an engine generator to start and power the load. (See "High Power Computing.")

Load type—and the disturbances that can be tolerated—dictate the UPS application. It may be cost-effective, for example, to size the UPS for the test equipment and put only those critical pieces on the UPS system; other R&D equipment may be able to ride through a short outage without consequences. A research facility may use heat in its experiments, and the furnace or oven can go off-line for a short duration and come back on-line with temperature remaining in spec. In this case, the UPS should be applied only to the test equipment and computers.

Surge protective devices (SPDs), commonly referred to as transient voltage surge suppressors, are an integral part of an R&D facility's distribution system and must be incorporated from initial design. SPDs are electronic devices that limit transient overvoltages—caused by lightning, utility switching, internal load switching and power electronic devices such as ASDs and UPSs—and divert dangerous surge currents away from critical loads to ground.

Service-entrance equipment should have an SPD to reduce externally generated surges. The branch-circuit panelboards that serve the research test equipment should have SPDs to minimize surges from within the R&D facility. Additional SPDs at the intermediate distribution panels may be needed, depending on distribution system exposure and how critical a given load may be.

Properly designed grounding is crucial for assuring power quality, but these systems are often overlooked during preliminary development of an R&D facility. Grounding systems are relatively inexpensive. The National Electrical Code (NEC) only outlines the grounding system with respect to life-safety—not from a power-quality perspective. Design professionals, however, must make a special effort to assure that:

• All metallic components of the building are bonded together to the service.
• The service neutral and ground are properly bonded together.
• Water pipes are bonded to the service.
• Separately derived systems such as UPSs and dry-type transformers are bonded to the building steel and service.

An effective earth-ground electrode system is required and usually consists of the building steel, a buried ground ring, a metal underground water pipe and other NEC-required electrodes. Every feeder and branch circuit should have an insulated grounding conductor—that is, a green wire—and not rely on the conduit for a ground path.

Special grounding areas—computer rooms, raised floor areas and high-frequency generator locations—must be bonded to building steel and the service. An isolated ground system is not needed if the facility has a properly designed and installed equipment-grounding system.

Power-quality and power-reliability solutions are often the same. Some reliability solutions, however, go beyond those for power quality. What is the impact if the electrical system fails and causes an interruption to the critical loads? What duration of an interruption can be tolerated? The answer to these questions drives the type of system and preventive steps that are needed to provide reliable electrical service to the critical loads.

• Utility feeders. Most R&D facilities are fed either underground or overhead from a single radial line from the utility. This exposes the facility to single-point failure—from cars hitting poles, lightning strikes and backhoe work. The outage in these cases could be many hours.

A more reliable approach would be to bring in a second utility feeder, preferably from a different utility substation, and install a transfer switch to transfer from the feeder that fails to one that remains operational. Manual operation of the switch could be performed for isolation of a feeder for maintenance. If the loads are sensitive to the loss of voltage during the transfer time—five to seven cycles—a static transfer switch can be installed which performs the transfer in a quarter-cycle.

• Owner system. Another method of terminating two utility feeders is to install two transformers and a double-ended substation with two 600-volt main breakers and a tie breaker. By sizing each transformer to carry the entire load of the facility, either transformer or associated utility feeder can be taken out of service and the facility can remain operational. The 600-volt main breakers and tie breaker can be set up to automatically transfer the load to the source with remaining power if one feeder is lost.

Transfer time for this solution is five to seven cycles. If the loads are sensitive to the loss of voltage during the transfer time, the static transfer switch option can be used.

Note that the double-ended substation allows for easy separation of mechanical loads—HVAC components, pumps, chillers—from sensitive test equipment and research loads. Beyond providing power quality, this helps with the isolation of loads for maintenance.

• UPS/generator. If the utility has a record of extended outages that the R&D critical loads cannot tolerate, a generator may be warranted to provide standby power. A generator requires five to 10 seconds to sense a voltage loss, start working, reach the needed voltage and frequency and then transfer to the critical loads. A UPS can be installed to ride through the interruption until the engine generator can transfer to the critical loads (see "Keeping Specimens on Ice" below).

When generators are used, an automatic-transfer switch (ATS) is used to transfer the load from the utility system to the engine generator. The ATS must be maintained for continued reliability and can be provided with a full maintenance bypass switch that allows testing and maintenance of the automatic portion of the switch without interrupting the load. This is an important issue for the double-ended substation/dual-feeder configuration of facilities that cannot be down for any scheduled maintenance. This type of system, while increasing reliability by having a second source of power, also enables safe maintenance procedures by completely de-energizing the portion of the system to be maintained. By moving the "dead" portion around, the entire system can then be maintained, further increasing the reliability.

By understanding a specific R&D program's needs and equipment operating constraints, designers can create an electrical distribution system that meets the budget and needs of the users. A 500-watt under-desk UPS may be all

the cost a start-up company can afford for power quality and reliability. Ongoing medical research, however, with controlled environments and strict temperature and humidity control may require a dual feeder system with UPS and engine generator. No single solution fits every case.

From Pure Power, Fall 2001.