Managing risks, benefits with closed transition transfer switches
Here are best practices for closed transition transfer switches when two live sources are connected together.
- Understand advantages and risks associated with using closed transition transfer switches in a standby power system.
- Recognize all the factors that contribute to breakers tripping during closed transition transfer.
- Identify methods to mitigate the risk of tripping breakers during closed transition transfer.
Closed transition transfer switches are becoming more popular for transferring power for life safety and critical processing loads. The benefits are that the emergency power system can be tested without interrupting power to loads and power can be retransferred to the utility after a failure without interrupting power to loads. However, there are risks associated with closed transition transfer switches as two live sources are connected together.
Problems with closed transition transfer originate from a difference in voltage between the two sources at the instant when the two sources are connected. The difference in voltage can be caused by several factors:
- A difference in root mean square (RMS) voltage between the sources
- A phase angle difference between the two sources
- A transient condition on one of the sources caused by a load switching on or off, or instability of one of the sources.
The instantaneous voltage difference between the sources results in a current surge from the source with the higher voltage to the source with the lower voltage at the instant of the sources’ interconnection. This current is limited only by the impedance of the sources and the current-carrying capacity of the cable or bus connecting them. It is this current surge that can trip breakers or, in more extreme cases, damage equipment.
Recommendations for minimizing risks of out-of-phase closure include:
- Recognize that all sync check systems allow for sources to be a few degrees out-of-phase at closure, resulting in some level of surge current between the sources. Breakers, transfer switches, and cable must be sized accordingly.
- Consider active synchronizing with voltage matching to minimize the phase and voltage differences between sources.
- Minimize transient conditions at the moment of transfer by inhibiting multiple transfer switches from transferring at the same time and preventing other loads from cycling during the transition.
- Use a transfer switch “fail to disconnect” or maximum parallel timer relay to shunt-trip an upstream breaker to prevent extended paralleling in the event that a transfer switch fails.
Phase difference at closure: How much is too much?
Very rarely will two sources be exactly in sync at the instant that a switch or breaker closes the two sources together, so it is reasonable to ask how far out-of-phase two sources can be and still have a reliable closed transition. IEEE 1547 allows generator sets or systems of paralleled generator sets between 1.5 and 10 MVA to be up to 10 deg out of phase with the utility when closing to the grid, with higher phase difference limits for smaller systems.
Alternators typically can handle connecting to a source that is 10 deg out of phase with it. For other equipment the answer depends on how much surge current the system can handle without tripping breakers or damaging equipment. For high reliability, one must consider the magnitude of current that can flow between the sources at the instant of transfer.
The surge current will be proportional to the voltage difference between the phases divided by the total impedance in the system. Surge current can be modeled as:
Isurge = V diff / Zsystem
V diff = the instantaneous voltage difference between the sources
Zsystem = the total impedance of the system.
Total system impedance is the sum of the subtransient reactance of the alternator, the impedance of the utility transformer, and the impedance of the cable or bus connecting the sources. In many applications where a single standby generator set is backing up the utility, the impedance at the instant of closure will be dominated by the subtransient reactance of the alternator. However, a thorough analysis will include all the sources of impedance in the calculation.
This example uses only the subtransient reactance of a single alternator. Note that in applications with paralleled generators, one must account for the contribution to the current from all the generator sets. This can be done by calculating an equivalent subtransient reactance for the paralleled generators according to the following equation:
Xd"equivalent = 1/(1/ Xd"gen1 + 1/ Xd"gen2 + …)
Neglecting the reactance of the utility transformer results in a worst-case scenario for calculated surge current. With this assumption, surge current can be modeled as:
Isurge = V diff /Xd"
Xd" = the subtransient reactance of the alternator, or the equivalent subtransient reactance of paralleled alternators
Assuming that the root mean square (RMS) voltages of the two sources are identical and no other loads are being switched at that moment, then the instantaneous voltage difference between the sources will be a function of the phase difference between the sources at the moment in the cycle at which closure occurs. Figure 1 is a representation of two voltage sine waves that are 10 deg out-of-phase and the difference between the two waveforms at each point in the cycle. The dashed line represents the instantaneous voltage difference between the two sources. This line is also a sine wave at the same frequency as the two sources. The maximum voltage on this line is the worst-case scenario for the differential voltage at the instant the two sources are paralleled.
The equation for the worst-case differential voltage is:
V diff (per unit) = 2* sin(delta/2)
delta = the phase angle difference between the sources in degrees (10 deg in this case). The worst-case voltage in this case is 0.17 per unit (pu).
For example, in a 480 V system with the two sources 10 deg out-of-phase, the worst-case instantaneous voltage between the sources would be 82 V (480*0.17 = 82 V). If these two sources were paralleled, the voltage difference between them at the instant of closing could be as high as 82 V.
Whether that voltage is too high depends on how much current it causes to flow and whether the equipment in the circuit can handle it.
Consider a 2.5 MW generator set. To calculate how much current would flow, use the alternator data sheet to determine the kVA rating and subtransient reactance of the alternator. In the data sheet in Figure 2, we see that the alternator has a subtransient reactance of 0.144 pu based on an alternator kVA rating of 3660 kVA.
Current resulting from the 0.17 pu difference in voltage is given by:
Isurge = V diff /Xd"
V diff = 0.17 and Xd" = 0.144 the surge current Isurge = 1.2 pu.
To convert the per unit current to amps, use the following:
Iamps = Ipu * alternator kVA rating/(√(3) * 480) = 5329 amps (RMS)
Note that in a thorough analysis, the reactance of the transformer would be added to the subtransient reactance of the generator or the equivalent subtransient reactance of paralleled generators. When adding per-unit quantities, the per-unit values for the alternators and transformer must be based on the same base kVA rating.
Whether this level of surge current can damage equipment or trip a breaker depends on the equipment through which the current is flowing. Circuit breakers typically have their instantaneous trip current set to 7 to 10 times the long time-trip setting. The surge will last only for one or two cycles, so as long as the level of surge current is not in the instantaneous trip range of the breaker, the breaker will not trip. Keep in mind that if the breaker is a current-limiting breaker designed to trip in the first half cycle of a fault, this will have to be considered.
A transfer switch that is listed to UL 1008 can withstand a repeated overload of 6 times rated current and maintain that current for 10 electrical cycles (167 msec) on each iteration, and continue to function at rated load after the test.
In the example above, if the load is transferred by a 2000-amp transfer switch protected by 2000-amp breakers, the maximum current surge is less than three times the long time-trip rating of the breaker and less than three times the full load current rating of the transfer switch. This will not cause a problem for the switch or the breakers. However, if this load were being transferred by a 400-amp transfer switch protected by 400-amp breakers located lower in the system, there is now a chance that the current surge will trip one of the breakers.
Surge current generated by an instantaneous phase difference between sources at transfer must be considered in the design of the system. Equipment must be sized to handle the surge current. Where this is not practical, open transition switches should be used. Loads that cannot tolerate a momentary interruption in service should be fed by an uninterruptible power supply (UPS).
Passive synchronizing systems
Closed transition transfer switches have successfully used passive synchronizing systems in many applications. Transfer switches use a sync check function for initiating closure to the oncoming source when the two sources are in phase. There are two basic algorithms used by sync check systems: a permissive window algorithm and a predictive algorithm.
A permissive window algorithm is commonly used in both active and passive synchronizing systems. The sync check system measures the voltage, frequency, and phase difference between the two sources. When the three parameters are within some predefined limits, the sources are said to be within a “permissive window.” When the sources have been in the permissive window for some preset period of time, the controller closes to the oncoming source. The required time in the permissive window is typically set to 0.1 to 0.2 seconds for passive synchronizing systems and 0.5 seconds for active synchronizing systems.
A predictive algorithm operates similar to a sync check system except that rather than waiting for the two sources to be in a permissive window for some period of time, it measures the rate of change of the phase angle difference between the two sources and calculates an “optimum phase angle” at which to initiate closure so that at the instant the switch closes, the two sources are as close to in-phase as possible.
Both types of algorithms have been used successfully. Generally speaking, the permissive window algorithm is more robust because the predictive algorithm is susceptible to transients on the voltage sources, which could skew the calculation of the optimum phase angle.
In many applications a slight frequency difference known as a “slip frequency” is imposed between the sources to make sure that they will come into sync with each other at a controlled rate. A slip frequency of 0.1 Hz has been used effectively.
Active synchronizing is the process of adjusting the generators’ engine governor to bring the waveform into phase with the utility waveform. Many synchronizing systems also include a voltage matching function, in which the generator sets will adjust the voltage regulator to drive the generator voltage level to match the utility voltage level. The voltage matching function is important in applications where the voltage on the utility transformer varies with load.
Figure 3 represents a generator waveform coming into phase with a utility waveform using an active synchronizer with voltage matching. Note that the utility waveform is constant and the synchronizer drives the generator set waveform in to sync with the utility. The voltage matching function forces the generator voltage to be at the same level as the utility voltage.
The synchronizer will hold the generator in sync with the utility until the synchronizer is turned off, unless a sudden load change causes a frequency change. Load changes on a system bus cause a sudden change in phase angle difference as frequency surges or sags in response to the load transient. This can cause the two sources to momentarily be out of sync until the synchronizer forces them back into synchronization. This is why for systems with multiple closed transition transfer switches, best practice is to allow only one switch to transfer at a time. With an active-phase lock-loop synchronizer, the time to synchronize is relatively short and reliable, so timing between switch operations need not be long.
Transfer and retransfer inhibit
It is common for changes in load on a generator set to cause sudden changes in the voltage and in the phase angle relationship between two sources that have been synchronized. For this reason the possibility of load transients at the moment of transfer should be minimized. For systems with multiple transfer switches, best practice is to allow only one switch to transfer at a time. This can best be achieved either by staggering transfer time delays or by using the transfer and retransfer inhibit functions.
The inhibit functions are used to prevent transfer to either the emergency source (transfer inhibit) or the normal source (retransfer inhibit). When transferring loads with closed transition transfer switches, only one transfer switch should be allowed to transfer at any given time.
The inhibit function can be controlled by a master control used in conjunction with a paralleling system. All switches initially are inhibited from transferring and the master releases the inhibit on one switch at a time.
In simple applications one switch can inhibit another. For example, consider the system in Figure 4 consisting of two closed transition transfer switches. The normally closed aux contact from the normal side of the automatic transfer switch (ATS) 1 is wired into the retransfer inhibit input ATS 2. This will inhibit ATS 2 from beginning its retransfer sequence (including all time delays) until after ATS 1 has transferred back to the normal source.
This configuration does create a potential failure mode. If the first switch fails to transfer or if the aux contact fails, the second switch will not transfer without manual intervention. For this reason it is preferable in some cases to use staggered time delays to prevent the switches from transferring at the same time. As these time delays are typically set on-site, it is important to clearly specify the time delays in commissioning documentation.
It should also be noted that the inhibit function is only required when transferring between two live sources. It is not a requirement in the event of a utility failure, so there is no need to be concerned about not getting the emergency source online quickly enough.
Many utilities require that closed transition transfer switches provide means to shunt-trip the breaker on the normal (utility) side of the transfer switch if there is a failure of the transfer switch that causes the two sources to remain paralleled in excess of 100 msec. Many transfer switches have a “fail to disconnect” output which can be used for this. It is up to the installer to connect these devices to the shunt trip of the breaker. Some utilities require maximum parallel timer and lockout relays that are separate from the transfer switch control to implement this function.
Regardless of whether this is required by the local utility, it is considered a best practice to use this function to make sure that two sources are not unintentionally paralleled for an extended period of time. Tripping either the normal side or emergency side breaker will provide the same level of equipment protection, although many utilities require tripping the normal side breaker.
Closed transition transfer switches allow for transferring loads without interruption during a test or when returning to the utility after an outage. This is a significant benefit in some applications; however, there are risks associated with closed transition transfer as two live sources are connected together. For loads that are not protected by a UPS, it is worth considering if the value of not having an interruption during a test or a retransfer to the utility justifies the risk of a closed transition transfer.
There are several methods for mitigating the risk of closed transition transfer:
- Make sure that breakers, transfer switches, and cable are sized to handle the surge current that may result from sources being a few degrees out-of-phase at closure as allowed by the sync check system.
- Consider using active synchronizing with voltage matching to minimize the phase and voltage differences between sources.
- Minimize the possibility of transient conditions at the moment of transfer by inhibiting multiple transfer switches from transferring at the same time and preventing other loads from cycling during the transition.
- Use a transfer switch “fail to disconnect” or maximum parallel timer relay to shunt-trip a feeder breaker to prevent extended paralleling in the event that a transfer switch fails.
Rich Scroggins is a technical specialist in the application engineering group at Cummins Power Generation. Scroggins has been with Cummins for 18 years in a variety of engineering and product management roles. He has led product development and application work with transfer switches, switchgear controls, and networking and remote monitoring products, and has developed and conducted seminars and sales and service training internationally on several products. Rich received his BSEE from the University of Minnesota and an MBA from the University of St. Thomas.