Scheming for Power
The challenges posed by electrical load growth and utility deregulation have brought power-quality issues to the forefront. Load growth has always influenced power quality. The need for increased infrastructure-distribution, transmission and generation-and the economics of its development are a balancing act in which power needs are carefully evaluated and weighed against regulatory requirements and environmental impact.
However, power quality and reliability problems are increasingly associated with utility deregulation, and until these issues are resolved, one can expect significant regional brownouts-or even blackouts. No matter what the cause, design professionals must work closely with clients to evaluate and identify their on-site needs for guaranteeing high-quality power.
Power quality is relative. What can cause devastating problems for electronic devices can be completely innocuous to motor and lighting electrical loads. Mission-critical elements depend on microprocessors-which are highly susceptible to power events-and discussions of power quality often focus on events that jeopardize manufacturing, commercial and institutional applications of microprocessors.
The most common sources of power-quality problems are voltage sags and swells, harmonics, transient voltage surges, undervoltages, overvoltages and outages (see “Power Events, Defined,” page 52). Underfrequency and overfrequency are also power-quality issues. However, they normally result from one of the aforementioned events, or are directly related to generation problems.
Planning a strategy
Strategies for solving power-quality problems are available-with an economic impact. A cost/benefit analysis is required to evaluate the cost of interruptions versus the on-site equipment and installation costs necessary to insure power quality. For the power consumer, quality involves protection against two types of power events: short-duration and long-term.
In 1993, the Electric Power Research Institute developed the Computer and Business Equipment Manufacturers Association (CBEMA) curve from empirical data of actual voltage excursions versus duration, as shown in Figure 1 (on opposite page). CBEMA has since been superseded by the Information Technology Industry Council, but the curve still goes by that name. The vast majority of excursions-approximately 98 percent-fell between 25 and 125 percent of nominal for durations of 16 milliseconds to 15 seconds. Based on the CBEMA curve, one would only need to protect against momentary events. However, with an eroding generation base, it is important to be cognizant of the need for protection against extended power outages in excess of 15 seconds.
External solutions for improved reliability involve utility infrastructure and are not readily influenced by the design professional or the client. Internal solutions, however, are driven by the power consumer and can be determined by identifying and analyzing the problem source. It is possible to provide protection for all power-quality scenarios. The more critical the application, the greater the capital investment.
Technologies old and new
One of the most cost-effective protections against catastrophic electronic equipment damage caused by voltage transients is the traditional transient-voltage surge suppressor (TVSS). These have been in use for decades and originally consisted of gas-discharge tubes and metal-oxide varistors. TVSS devices have evolved to the use of avalanche-diode technology coupled with event logs.
In geographic areas of high lightning occurrence, the TVSS is a mandatory element of good design. In all other areas it is inexpensive insurance and should definitely be considered. Engineers generally protect only the main distribution switchboard at the service entrance. However, some designers provide additional TVSS protection at selected panelboards throughout the facility.
The strategy that addresses all power-quality elements-with the exception of internally generated harmonics-incorporates an uninterruptible-power supply (UPS) in conjunction with on-site generation. The UPS provides near-perfect waveform output coupled with excellent voltage regulation by rectifying the input alternating-current (AC) source and inverting the resulting direct-current (DC) power to a sinusoidal output. The UPS internal DC bus is supplemented with a battery back-up source, which is paralleled with the rectifier output. If there is a power disruption, the UPS automatically switches to the battery power source.
Both for the rectifier section and the inverter section, most UPS manufacturers use high-speed switching insulated-gate bipolar-transistor (IGBT) technology, which not only enhances the waveform output, but also decreases the total harmonic distortion reflected back into a facility’s electrical distribution system. IGBT technology also improves the availability of power to the load, because there is a higher degree of reliability over the older silicon-controlled rectifier technology. Single UPS systems approach 89,000 hours of continuous availability.
There are other UPS scenarios that extend the hours of availability. These include parallel-redundant and isolated-redundant schemes, both of which require additional UPS modules, and consequently, added cost.
The typical battery discharge time is between 5 and 15 minutes at 100-percent continuous load. This allows time for the generator to come on line and reenergize the AC bus. The batteries of today no longer require separate battery rooms with continuous ventilation. In the last two decades, battery designs have incorporated a recombinant-gas feature that essentially results in a nongassing, maintenance-free apparatus. The advantage of this is space utilization: The recombinant batteries can be located adjacent to the UPS, shortening the distance between the DC source and the UPS and eliminating the need for the traditional battery room.
One generally specifies these types of installations for critical loads of 250 to 500 kilowatts (kW). Parallel- and isolated-redundant schemes are more expensive. There is a significant increase in the availability time with these two design approaches, but the cost must be weighed against the criticality of the mission.
Intriguing and new
For loads of greater than 500 kW, there are two technologies that are very intriguing. The first utilizes a rotary motor/generator on a single axis with an integral storage mass coupled with battery and generator backup.
This scheme has been used successfully in Europe for a number of years and was introduced into North America during the last two years. By utilizing it, electrical loads in excess 2,000 kW can be accommodated. The operational characteristics allow the rotating mass with the electronic control to deliver constant power throughout the entire range of power-quality events. The rotating storage mass allows time for the generator to come on-line in the event of an extended power outage.
The second new technology recognizes that in manufacturing facilities with robotic devices, there is a genuine need for distributed and conditioned power. The cost of providing individual UPS sources for all locations would be prohibitive. An ingenious solution has been to develop a medium-voltage UPS system that is rated up to 34.5 kilovolts and 20 millivolt-amperes. The standard integral battery provides sufficient power to allow the medium-voltage UPS to control 98 percent of all power-quality issues. If the system is specified with adequate battery capacity and is used in conjunction with a generator, it can meet all the requirements for assured power reliability. Surprisingly, the installed cost for both of these new systems is comparable to standard UPS systems.
Disruption from within
The power-quality problems and solutions discussed so far are from external sources-primarily utility-based. The power-quality issue linked to internal sources within a facility is harmonics. The proliferation of nonlinear load that is connected to the building’s electrical-distribution system generates harmonics. Harmonics-generating equipment such as personal computers and variable-frequency drives can cause waveform distortion, which in turn can produce neutral heating. If left unchecked, neutral heating can result in significant damage to panelboards and switchboards.
One technique used with a great degree of success is the provision of a zigzag-wound inductor located in close proximity to the harmonics-producing source. This solution mitigates harmonics so that their effect is not transferred to downstream conductors and equipment, thereby precluding the need to increase the neutral size for all downstream devices including conductors and panelboards. Figures 2 and 3 on page 50 are from a successful implementation of this harmonic-mitigating technology. Care must be taken, however, in the application of this equipment. Some harmonic filters contain capacitance in their tuned circuitry and can actually form a sink for all the harmonic current within a distribution system. This could lead to their eventual failure due to excess current heating.
There is hope that with the emergent technologies of distributed resources-including traditional generators, microturbines and fuel cells-power quality will improve and system reliability will be enhanced.