Solid-state subcycle switching devices can eliminate the need for a UPS, if two sources of power are available.

For more than 70 years, we've had the motor-generator and motor-alternator helping us to provide reliable conditioned power. And, believe it or not, the uninterrupted power supply (UPS) has been with us for more than 40 years. Some 12 to 14 years ago, one manufacturer, as an offshoot of the UPS, developed the subcycle automatic transfer switch (ATS) for stand-alone applications at 600V and below. Now, thanks to further development of power level silicon technology, this capability of high-speed transfer between two power sources is available in 600V as well as the following MV distribution voltage ranges: 2.4kV, 4.8kV, 15 kV, 27kV, and 38kV.

Looking back

In our examination of solid-state UPS systems, we've seen the value of the static switch in complementing the complete UPS function. When the inverter output tries to handle a large inrush or a fault condition downstream of the UPS, the inverter overloads itself so quickly that it's in danger of current limit. This power electronics protective feature, which consists of the inverter shutting itself off, provides damage avoidance. The heart of this self-protection feature is that the inverter is a "soft" power source and not capable of handling high load inrushes in any way similar to the utility, which is a "stiff" power source. As a result, the quick switching to bypass line by the static switch is the means of avoiding the shutdown problem when inrush occurs.

When an internal problem inside the static switch arises, one that may affect the output (such as a phase fault or component failure), the switch transfers to the bypass line to avoid loss of output. This transfer function keeps the load on line. The diagnostics inform you of the internal fault, and lock the switch in the bypass position until the condition is corrected through inspection and/or maintenance.

The "stand-alone" concept

As switch techniques improved over the years of UPS operations, equipment designers made possible the switching between two energized power lines, rather than transferring from inverter to power line. So perfect is this technology that it's even possible to make a device capable of a 4-ms transfer, with a zero crossing at the midpoint of the transfer, thus avoiding any reverse current flow when the transfer takes place.

The above technique readily adapts to a "stand-alone" product, one not tied to an inverter output but capable of switching between two power lines. In fact, this application has found favor with those users having two independent power sources at 600V and below; they now could count on these energy sources to back up each other. (A typical "stand-alone" configuration is shown in [ILLUSTRATION FOR FIGURE 1 OMITTED].) The beauty of this concept is the potential elimination of a battery-support system, along with the associated high cost of rectifier/inverter assemblies. By using a "stand-alone" solid-state subcycle transfer switch, you can save from 75% to 80% of installed UPS cost. In fact, many dual-feed locations such as hospitals, computer service centers, banks, and insurance operations have made use of this technology, reducing their expense for continuous service to as little as a quarter or a fifth of the cost of an appropriately sized UPS.

Certain prerequisites, however, must be considered to ensure the above application is a valid one. First, you must verify two power sources are, in fact, available. Second, you have to make sure no high shock torque will be applied to downstream motors during the operation of these solid-state subcycle devices.

Although limited in ampere rating and to 600V and below applications, these transfer switches are now available in sizes up through 4000A and capable of handling up to 3000kVA of load. Equipment costs for this low-voltage technology are in the range of $500 to $900 per kVA. (Smaller sizes are usually at a higher cost per unit.)

Space considerations are just as important as equipment costs, however, especially when tenant occupancy costs in major buildings run from $30 to $75 per sq ft. Here again, the "stand-alone" solid-state, sub-cycle transfer switch provides a cost benefit: Along with a much lower cost of maintenance, it takes up considerably less room than a UPS and its peripheral support equipment. By eliminating the needed space for a large battery system, equipment cabinets, etc., real savings are achieved. In several static switch installations, the recovery of space cost (through additional rental revenues) has made construction and material financial arrangements even more attractive.

Case history: Multiple feeds

We were asked to come to a Northeastern bank's operations center to recommend specifications for a new static UPS that was to serve a critical computer room. The owner showed us the existing electrical service, and we obtained transformer size information so we could do the required load calculations for sizing the proposed UPS. Immediately noticed were four separate power lines entering the facility, along with a standby turbine generator for backup.

We asked the owner why such a service arrangement needed a UPS in addition to the multiple power sources; high-speed transfer switch technology could do the job just as reliably yet much less expensive. The owner had never heard of such technology.

One of the proposals for the site includes a pair of low-voltage static transfer switches, one for each pair of power lines. Both switches feed a double-ended substation. [ILLUSTRATION FOR FIGURE 2 OMITTED] In this configuration, the power sources can be switched fast enough between the multiple utility lines that any single-line disturbance will not cause a shutdown. Thus, the redundancy of the service design is preserved.

The next step: MV subcycle switching

The obvious question to equipment designers is: "Can a stand-alone solid-state subcycle transfer switch be designed for use at higher voltages, especially those associated with primary service from a utility?" The answer is "yes." In 1990, a medium-voltage (MV) version was put in service at a U.S. Navy facility. But, it would take several years for the cost and availability of new silicon component devices to be attractive for commercial use in a "civilian" version.

In 1994, with increasing interest from primary service customers, one utility began a process that led to the installation and operation of the first 15kV, 600A, 14MVA, 4-ms switch. Resultant testing and field verification have shown consistent operation and performance characteristics that are all within design parameters. This installation has led a number of utilities to order these MV devices for their systems and specific customer installations.

In Fig. 3, we see a typical transfer from one incoming line to a second one. There is very little difference between the low-voltage and MV versions. First, both operate as gated thyristors. The "gate" here is the switch's control device, which issues the signal to the silicon to "conduct" or to "turn off." These functions are done in microseconds, so that the speed of transfer will never interfere with sensitive loads. Second, both versions have no moving parts, so that repeated transfers will not cause equipment deterioration. Finally, the use of an old established technique (the gated thyristor) adds higher speed sensing and control advances.

One innovation recently applied to subcycle transfer switching is the control of the gating signals by fiberoptics rather than shielded copper wire. The result is a lighter and smaller equipment package and more secure signal transmission.

The control system for the MV version includes digital signal processing (DSP), which increases the speed and quantity of information used to perform the transfer functions. For example, when the MV controller senses an irregularity on Line 1, it collects information from 192 points per electrical cycle. It then uses 2.5 ms to store more than 30 points of information in making the decision to transfer to Line 2. One of the deciding factors in a transfer is the presence of a downstream fault; this condition needs to be cleared by circuit protective devices. This is evaluated to prevent initiation of transfer to the second line, which would only feed the faulted portion.

Another innovation is the use of the MV version in a metal-clad or metal-enclosed lineup of switchgear. A complete lineup, including electromechanical backup transfer switches, is really a replica of the UPS function. In Fig. 4, we see a conventional electromechanical MV transfer switch available (with motor operators designated by the letter "M") being used in a bypass arrangement. Here, the silicon devices are the primary transfer devices; however, the static and electromechanical control systems exchange information with each other to ensure automatic operation. If the silicon switch needs to be taken out of service, the bypass device (electromechanical transfer switch) is signaled to close in parallel before the silicon device is opened. As you can see, continuity is just like a UPS with its bypass circuit.