Shared resonant turn off circuit

The present inventions relate generally to a static transfer switch for transferring power from one power source to another power source to supply an electrical load.

Static transfer switches are used in the industry to control the electrical power supply to critical electrical components. In particular, static transfer switches are used for electrical loads like data centers where a constant, high-quality electrical supply is required.

An example of a static transfer switch 10 is shown in FIG. 1 . As shown, two different electrical power sources 12 A, 12 B are coupled to the static transfer switch 10 . The output of the static transfer switch 10 is coupled to an electrical load 14 . Typically, the output is directly connected to a Power Distribution Unit (PDU) 14 , which includes a transformer 16 . The final electrical load may be racks of computer servers 30 ( FIG. 2 ) in a data center. However, it is understood that static transfer switches 10 may also be used to supply power to other types of electrical loads.

The static transfer switch 10 may include a variety of sensors 18 A, 18 B, 20 A, 20 B to monitor electrical properties of the first and second power sources 12 A, 12 B and the power output. For example, it may be desirable to monitor the voltage 18 A, 18 B of each of the power sources 12 A, 12 B and to monitor current 20 A and voltage 20 B of the output. The static transfer switch 10 also includes one or more switches 22 A, 22 B associated with each of the power sources 12 A, 12 B. This allows the static transfer switch 10 to supply power to the output from either of the power sources 12 A, 12 B. For example, the first power source 12 A may be the preferred power source 12 A (e.g., the grid), and the second power source 12 B may be a backup power source 12 B (e.g., a generator). In normal use, power can be supplied from the first power source 12 A to the load 14 by closing the first switch 22 A and opening the second switch 22 B (to disconnect the second power source 12 B). In the event that the first power source 12 A suffers from degraded performance (e.g., drop in voltage) as determined from one or more of the sensors 18 A, 18 B, 20 A, 20 B, the power supply can be transferred to the second power source 12 B by opening the first switch 22 A and closing the second switch 22 B. Thus, the electrical load 14 is provided with a constant source of power despite the possibility of degraded performance events in one of the power sources 12 A, 12 B.

A static transfer switch is described for increasing the speed of switching from one power source to another power source. The system senses degraded performance of the power source supplying power to the load. In response to sensing degraded performance, the system turns off a gate signal to a first switch coupled between the power source and the load. The system also includes a resonant turn off circuit which forces the first switch to open during a switching event. The resonant turn off circuit is shared between the first switch associated with the first power source and a second switch associated with a second power source. The resonant turn off circuit can be selectively connected to either the first or second switch depending on which power source is supplying power to the load. The invention may also include any other aspect described below in the written description or in the attached drawings and any combinations thereof.

The invention may be more fully understood by reading the following description in conjunction with the drawings, in which:

FIG. 1 is a schematic of a static transfer switch;

FIG. 2 is a schematic of the static transfer switch showing three-phase power sources and a three-phase load;

FIG. 3 is a schematic of a switch circuit coupled between one phase of the power source and the load;

FIG. 4 is a series of graphs showing electrical properties of the switch circuit during opening of the switch between the power source and the load;

FIG. 5 is a schematic of a charge circuit for the energy storages;

FIG. 6 is another schematic of the static transfer switch;

FIG. 7 is a flow chart showing a transfer of power from one power source to another power source to supply a load;

FIG. 8 is a schematic of a digital signal processor used for controlling power through the static transfer switch;

FIG. 9 is another schematic of a digital signal processor used for controlling power through the static transfer switch;

FIG. 10 is another schematic of a static transfer switch;

FIG. 10A is a schematic of a double throw switch;

FIG. 11 is a schematic of a resonant turn off circuit shared between two main circuits;

FIG. 12 is another schematic of a static transfer switch;

FIG. 13 is another schematic of a static transfer switch;

FIGS. 14A-E are schematics of the operation of a resonant turn off circuit; and

FIG. 15 is a series of graphs showing electrical properties of the switch circuit of FIGS. 14A-E .

An example of a three-phase static transfer switch 10 is shown in FIG. 2 . Typically, static transfer switches 10 are designed to complete a switching event between the two power sources 12 A, 12 B within one electrical cycle of the power sources 12 A, 12 B. This is desirable so that a high-quality, constant power supply is provided with minimal effect on the electrical load 14 , 30 . In order to achieve switching events this quickly, it is typically necessary to use solid-state switches 22 A, 22 B (first switches 22 A and second switches 22 B) to perform the switching event since solid-state switches 22 A, 22 B can be switched on and off in less than one electrical cycle. Preferably, the switches 22 A, 22 B are silicon controlled rectifiers (SCR). Various types of thyristors may be used for the solid-state switches 22 A, 22 B, such as integrated gate-commutated thyristors (IGCT), reverse blocking integrated gate-commutated thyristors (RB-IGCT), or gate turn-off thyristors (GTO). In the case of a multiphase static transfer switch 10 , each of the main switches 22 A, 22 B will be made up of multiple individual switches 26 , with at least one switch 26 for each phase A, B, C. A pair of anti-parallel thyristors 26 is particularly well-suited for each switch 26 associated with a phase A, B, C. Because the power sources 12 A, 12 B are AC power sources, each switch 26 typically includes two switches 26 A, 26 B in an anti-parallel arrangement. However, each pair of anti-parallel thyristors 26 A, 26 B are often treated as a single switch 26 because they typically turn on and off together. The static transfer switch 10 may also include a series of manual switches 28 that are primarily used during maintenance to isolate sections of the circuit.

It is often preferred for the first and/or second power sources 12 A, 12 B to include an uninterruptible power supply (UPS) 32 to provide control over the electrical properties of the original source 12 A, 12 B and manage power drops or losses in the original source 12 A, 12 B. As noted, the output is typically coupled to the transformer 16 of a PDU 14 , and the final electrical load 30 is often racks of computer servers 30 in a data center.

An improvement of the invention herein is that it uses a resonant turn off topology adjusted to force commutate a three-phase power system and is operated by an embedded digital processor controlling the signals with software intelligence and algorithms in order to achieve autonomy and make possible a sub-millisecond transfer switch.

The turn off circuit 40 with RTO topology is shown in FIG. 3 . As shown, the circuit 40 includes a main circuit 42 and a resonant circuit 44 . The main thyristors 26 A, 26 B (first switches) are S m1 and S m2 . The resonant circuit includes four auxiliary thyristor switches 34 (third switches) S r1 , S r2 , S r3 , S r4 , resonant capacitor C (energy storage) 36 , and resonant inductor L 38 . The capacitor C 36 is pre-charged to provide resonant current to create a zero-current crossing for the main thyristors 26 . The inductor L 38 limits di/dt for main thyristors 26 during turn-off. During normal conduction, only the main thyristors S m1 (or S m2 ) 26 A, 26 B are conducting and all the auxiliary switches 34 are off. Thus, the pre-charged resonant capacitor 36 is isolated from the main thyristor switches 26 . In resonant turn-off operation, the auxiliary switches 34 S r1,2 (or S r3,4 depending on current direction) are triggered to open by sending a gate signal thereto. As a result, the energy stored in the resonant capacitor 36 is discharged. When the resonant current through the inductor 38 L exceeds the load current, the current through the main thyristor 26 is commuted to the resonant circuit. In the meantime, the capacitor 36 voltage provides a negative bias voltage to help the main thyristor 26 turn off (i.e., open and stop conducting). When the main thyristor 26 current reaches zero, it starts turning off with reverse bias voltage from the resonant capacitor 36 . In the RTO topology, there are three possible design choices to control performance: resonant capacitance value C, pre-charged initial capacitor voltage V c0 , and resonant inductance value L. These parameters can be used to determine how much and how fast the main thyristor current can be turned off, as well as the size and cost of the auxiliary resonant circuit.

FIG. 4 shows the voltage and current waveforms of resonant turn-off operation for the main thyristors 26 and auxiliary switches 34 . As shown in the bottom graph, at time 0.1 seconds a gate signal is sent to turn off the main thyristor 26 , and a gate signal is sent to turn on the auxiliary thyristors 34 . This occurs as a result of degraded performance being identified by a sensor 18 , 20 . Because the second switches 22 B cannot be closed to switch power to the second power source 12 B until the first switches 22 A have been opened, it would be beneficial to be able to stop current conduction through the first switches 22 A faster than in a conventional static transfer switch. That is, once a turn on gate signal has been applied to an SCR, current will continue to be conducted through the SCR even after a turn off gate signal is applied. However, after a turn off gate signal is applied to the SCR, the SCR will stop conducting current after the current drops below a threshold. That normally occurs with an AC current when the current waveform crosses zero. However, that may require up to half of an AC cycle for the power source waveform to cross zero (i.e., 8 ms). It would be beneficial to be able to stop the first switches 26 A quicker so that the second switches can be closed more quickly to transfer power as quick as possible from the first power source 12 A to the second power source 12 B. As shown in the top graph in FIG. 4 , current to the load 14 , 16 through the main and auxiliary thyristors 26 , 34 stops in less than 0.18 ms after applying the gate signals to the main and auxiliary thyristors 26 , 34 , and preferably within 0.5 ms. Thus, the turn off speed is greatly increased.

Referring to FIG. 4 , after the auxiliary resonant thyristors 34 are triggered, the resonance starts and the resonant current through L 38 increases fast. Once the resonant current is larger than the load 14 , 16 current, the main thyristor 26 current decreases to zero and starts turning off. In the meanwhile, the load 14 , 16 current is commutated or bypassed to the auxiliary circuit 44 ( 34 , 36 , 38 ). The resonant capacitor 36 is re-charged by the load 14 , 16 current but in reversed voltage polarity. After the resonant capacitor 36 voltage is high enough and the load 14 , 16 current is interrupted, the auxiliary thyristors 34 are turned off. As a result, the RTO has disconnected the first power source 12 A and the load 14 , 16 is ready to be transferred to the alternative power source 12 B.

The resonant circuits 44 (one for each phase) are shown in FIG. 5 . As shown, a pre-charge circuit 46 is coupled to the resonant circuits 44 with a DC bus 48 . Preferably, the DC bus 48 is pre-charged slowly through the single bridge rectifier diodes 50 with resistors 52 to make sure capacitive inrush is prevented. Also, in case the DC bus 48 runs above the set limit, a bleeding resistor 54 can be used to adjust that DC level to the desired value. A pre-charge relay 56 and bleed relay 58 are used to turn on and off the pre-charging and bleeding. Preferably, a fourth switch 60 is provided between the pre-charge circuit 46 and the capacitors 36 to disconnect the pre-charge circuit 46 from the capacitors. The fourth switch 60 is preferably opened to stop charging of the capacitors in response to the switching event. If desired, fourth switches 60 may be provided for each resonant circuit 44 to disconnect the respective circuit 44 from the DC bus 48 .

A phase locked loop (PLL) 62 (see FIG. 9 ) is used to determine the frequency and phase angle of the AC waveform of the power source 12 A. The PLL 62 can be used in determining the firing sequence for the auxiliary switches 34 because the switches 34 need to be fired at the correct polarity. Since this is a three-phase system, the polarity of phase A, phase B and phase C will not be the same at the same time. The firing sequence is determined by the following algorithm:

IF 0<=PLL_RADIAN [PHASE_A]<=π

Turn ON OUT_RTO_PLUS_A Turn OFF OUT_RTO_MINUS_A

Turn ON OUT_RTO_MINUS_A Turn OFF OUT_RTO_PLUS_A

The same algorithm is applied for the other remaining phases to complete the firing sequence in case a transfer is needed. In other words, only two of the auxiliary switches 34 are turned on to release current from the capacitor 36 , with one switch 34 being coupled to the input of the main thyristors 26 and the other switch 34 being coupled to the output of the main thyristors 26 . The two auxiliary switches 34 that are turned on depends on current flow of the AC current through the main switches 26 . Thus, when the current is positive, one pair of switches 34 will be turned on (while the other pair remains off), and when the current is negative, the other pair of switches 34 will be turned on (again, with the other pair remaining off). Preferably, all of the main switches 22 A, 26 are turned off (i.e., opened) and the auxiliary switches 34 are turned on (i.e., closed) at the same time in the static transfer switch 10 . Because some phases A, B, C will have a positive current and some phases A, B, C will have a negative current at the switching instant, the pair of auxiliary switches 34 that is closed in each phase resonant circuit 44 will vary depending on the current of the particular phase A, B, C.

The main circuits 42 are shown in FIG. 6 . The table below lists the names of the signals used in the main circuits 42 and the resonant circuits 44 of FIGS. 5-6 . As explained below, a DSP 64 ( FIGS. 2 and 8-9 ) may be used to generate the control signals used by the main circuits 42 and resonant circuits 44 .

FIG. 7 shows a flowchart of the method of transferring the power supply for the load 14 from one power source 12 A to another power source 12 B in a static transfer switch 10 . As understood, the method may be implemented by a controller 64 , which may be in the form of a DSP 64 . In steps 66 - 68 , the energy storage (capacitors 36 ) are precharged by the charging circuit 46 . The voltage of the DC bus 48 between the charging circuit 46 and the resonant circuits 44 is monitored and controlled to maintain a desired charge on the capacitors 36 . That is, the relays 56 , 58 are opened and closed as needed to supply voltage and bleed voltage to charge the capacitors 36 . In step 70 , the quality of the power being supplied by the power source 12 A connected to the load 14 is monitored for degraded performance events. When such a situation is identified, the main switches 26 for the first power source 12 A are opened to disconnect the first power source 12 A from the load 14 in steps 72 - 78 . That is, in step 72 turn off gate signals are sent to the main switches 26 . The auxiliary switches 34 that must be turned on in each resonant circuit 44 are then determined in step 74 by phase locked loops 62 (see OUT_RTO_MINUS, OUT_RTO_PLUS algorithm above). The auxiliary switches 34 that have been determined in step 74 are then turned on with gate signals in step 76 .

In step 78 , the DSP 64 verifies that the preferred power source 12 A is completely disconnected by confirming that the net value of the current and voltage passing through the main switches 22 A, 26 for each phase of the preferred power source 12 A is zero or negligible enough to confirm that the resonant circuits 44 did in fact reverse the bias for each of the main circuits 42 . Finally, the DSP 64 can initiate a turn on command to the alternate power source 12 B in step 80 . Due to power quality considerations like inrush and soft start and preferred turn on conditions, it is possible to vary the method used to turn on the second power source 12 B while still making use of the improved method of turning off the main switches 26 of the first power source 12 A. After the main switches 22 B, 26 for the alternate power source 12 have been turned on, the power transfer has been completed in step 82 .

FIGS. 8-9 show a control system that may be used to control the resonant circuits 44 , and associated charging circuit 46 . It is understood that the control system may also be used to control the main circuits 42 as well. The Digital Signal Processor (DSP) 64 may be used to control all the relays, sense the voltages for the PLL synchronization and control the main switches 26 , 42 and RTO thyristors 34 , 44 . As shown, a scaling processor 84 may be provided to adjust the inputs for use by the DSP 64 . As also shown in FIG. 9 , the DSP 64 may include a PLL element 62 and a phase decoder element 86 to evaluate the current direction of each phase and generate the gate signals for the auxiliary switches 34 .

Static transfer switches 10 are designed to transfer power to a load 14 between a preferred electrical power source 12 A and alternate electrical power source 12 B in response to power source quality disturbances. Traditionally, thyristors 26 are commonly used in conventional static transfer switches 10 due to their high current handling capability. As a semi-controlled device, thyristors 26 cannot be turned off by simply removing the gate signal to the thyristor 26 , because a thyristor 26 is turned off when the conduction current through the thyristor 26 commutates below its minimum holding current (usually less than 1 A) or the voltage applied to the thyristor 26 is reversed, which can be as long as half a cycle of the power source 12 fundamental frequency (e.g., 8.3 ms in 60 Hz AC systems). To speed up commutation of the thyristors 26 , a resonant turn off (RTO) circuit 44 may be used as described above and below, where the resonant turn off circuit 44 is added in parallel to the thyristors 26 . During thyristor 26 turn-off, a reverse current is injected by the resonant turn off circuit 44 to force the thyristor 26 current to commutate quickly to zero. In this way, the resonant turn off circuit 44 can interrupt the current within 1 ms, or preferably less than a half cycle of the power sources 12 , which greatly shortens the disconnect time of the off going electrical power source 12 . Typically, each phase A, B, C of each power source 12 has an individual resonant turn off circuit 44 in order to quickly turn off each phase during a transfer and to allow quick transfers back-and-forth in both directions between the power sources 12 . Thus, six resonant turn off circuits 44 are required in for a two source static transfer switch 10 . Likewise, nine resonant turn off circuits 44 would be needed in a three source static transfer switch 10 .

As shown in FIGS. 10-11 , the number of resonant turn off circuits 44 may be reduced by sharing the resonant turn off circuit 44 between the phases A, B, C of different power sources 12 A, 12 B. This may be useful in reducing the cost and complexity of the static transfer switch 10 . As shown, an AC contactor or relay 90 may be used to select which power source 12 A, 12 B the resonant turn off circuit 44 is connected to. Various types of switches 90 (i.e., a fourth and fifth switch) may be used to connect the resonant turn off circuit 44 to a selected power source 12 A, 12 B, although an electro-mechanical switch 90 may be preferred due to the low cost, high capacity and reliability of such switches. Since the switching times of the contactors 90 is not critical because the contactors 90 are not directly involved in the actual switching event of turning off the thyristors 26 A of one source 12 A and turning on the thyristors 26 B of the other source 12 B, contactors 20 with slower switching times than the main thyristors 26 A and/or the auxiliary thyristors 34 may be used. Although individual contactors 90 A, 90 B may be used, with each contactor 90 A, 90 B corresponding to one of the power sources 12 A, 12 B and one of the phases A, B, C, a single double throw switch 90 as shown in FIG. 10A may incorporate two switches 90 A, 90 B (i.e., the fourth and fifth switch) into one physical switch 90 . In use, the contactors 90 A, 90 B connect the shared resonant turn off circuit 44 to the main circuit 42 of the power source 12 A, 12 B that is currently supplying power to the electrical load 14 . As a result, the resonant turn off circuit 44 is connected to and prepared to quickly turn off current through the operating main circuit 42 when a power transfer switching event is initiated. For example, when the thyristors 26 of the first power source 12 A are closed such that power is being supplied from the first power source 12 A to the load 14 , the first contactors 90 A are closed to connect the shared resonant turn off circuits 44 to the first thyristors 26 and the second contactors 90 B are open to disconnect the shared resonant turn off circuits 44 from the second thyristors 26 . Conversely, when the thyristors 26 of the second power source 12 B are closed such that power is being supplied from the second power source 12 B to the load 14 , the second contactors 90 B are closed to connect the shared resonant turn off circuits 44 to the second thyristors 26 and the first contactors 90 A are open to disconnect the shared resonant turn off circuits 44 from the second thyristors 26 . As a result of the shared arrangement of the resonant turn off circuits 44 , the number of resonant turn off circuits 44 can be reduced to three instead of six as required in a conventional two source static transfer switch 10 . Similarly, in a three source system, the number of resonant turn off circuits 44 could be reduced from nine to three. The contactors 90 may also be useful in protecting the resonant turn off circuits 44 and/or isolating the circuits 44 during fault conditions (e.g., thyristor 26 , 34 or IGBT 96 failures).

As shown in FIG. 12 , the number of resonant turn off circuits 44 may also be reduced by eliminating the resonant turn off circuits 44 from one phase C of a three phase system. Thus, as shown resonant turn off circuits 44 may be directly connected to the main circuits 42 of two of the phases A, B. In this arrangement, when the two main circuits 42 are turned off with the respective resonant turn off circuits 44 , current flow from the third phase C is also forced to zero automatically through the delta input side 92 of the transformer 16 between the main circuits 42 and the electrical load 14 . For example, in a power transfer switching event where power is initially being supplied by the first power source 12 A, the resonant turn off circuits 44 directly connected to the main circuits 42 for phases 12 A-A and 12 A-B (i.e., phases A & B of source 12 A) are activated. This causes the respective main circuits 42 to open and phases 12 A-A and 12 A-B to be disconnected from the load 14 . One or both of the two resonant turn off circuits 44 then force the remaining main circuit 42 for phase 12 A-C to open through the delta side 92 of the transformer 14 which disconnect the remaining main circuit 42 from the load 14 . Even though the third phase C may not have its own resonant turn off circuit 44 , the current released by the two resonant turn off circuits 44 and the current conducted by all three main circuits 42 may cease flowing to the transformer 16 within 0.5 ms of the gate signals sent to the main circuits 42 or within less than half a cycle of the power source 12 . The main circuits 42 for phases 12 B-A, 12 B-B, 12 B-C are then closed to connect the second power source 12 B to the load 14 . Preferably, the two resonant turn off circuits 44 are activated at the same time by closing the corresponding auxiliary switches 34 at the same time such that current from both resonant turn off circuits 44 forces the remaining main circuit 42 to open through the delta side 92 of the transformer 16 . Preferably, the output side of the transformer 16 is a wye side output 94 . Thus, whereas a conventional static transfer switch 10 would have a separate resonant turn off circuit 44 for each phase A, B, C of a three-phase electrical power source 12 , the present embodiment may be used to eliminate the resonant turn off circuit 44 from one of the phases C such that no resonant turn off circuit 44 is directly connected to the main circuit 42 of one phase C. As a result of the delta connected resonant turn off circuits 44 , the number of resonant turn off circuits 44 can be reduced to four instead of six as required in a conventional two source static transfer switch 10 . Similarly, in a three source system, the number of resonant turn off circuits 44 could be reduced from nine to six.

As shown in FIG. 13 , the designs of FIGS. 10 and 12 may also be combined to reduce the number of resonant turn off circuits 44 even further. Thus, the static transfer switch 10 may have shared resonant transfer circuits 44 between two corresponding phases A, B of the power sources 12 A, B and may have one phase C connected through the delta side 92 of the transformer 16 that does not have its own resonant transfer circuit 44 . As a result, the number of resonant turn off circuits 44 may be reduced from six to two in a two source system and from nine to two in a three source system.

The operation of the resonant turn off circuit 44 of FIG. 11 is illustrated in FIGS. 14A-E (refer to FIG. 11 for detailed reference numbers), and the voltage and current is shown in FIG. 15 . It is understood, however, that any type of resonant turn off circuit 44 may be used including the resonant turn off circuit 44 of FIGS. 3 and 5 . The capacitor 36 may be pre-charged to a relatively low voltage (40-50 V) and provides a resonant current to create a zero-current crossing for main circuit 42 . A pre-charge circuit 46 may be provided for pre-charging the capacitor 36 . The resonant inductor 38 limits the change in current to provide a softer commutation. As shown in FIG. 14A , before t 1 , the main thyristors 26 conduct current to the load 14 . In FIG. 14B , at t 1 , when a switching event is initiated, the gate signal to at least one of the main thyristors 42 is turned off, depending on the load current direction, and the corresponding auxiliary thyristors 34 and the IGBT 96 are closed with a gate signal. For example, in FIG. 4( a ) , S m1 conducts the load current. From t 1 to t 2 , in FIG. 14B , load current is forced to commutate from the main thyristor 26 to the resonant turn off circuit 44 . In FIG. 14C , at t 2 , reversed bias voltage from the capacitor 36 is applied to the anode of the main thyristor 26 to speed up its turn-off. As shown in FIG. 15 , the capacitor 36 discharges and voltage of the capacitor 36 drops during this period. The energy stored in the capacitor 36 is preferably large enough to maintain the voltage polarity and reverse the bias to the main thyristor 26 . The period between t 2 and t 3 allows sufficient time for the main thyristor 26 to commutate. In FIG. 14D , at t 3 , the main thyristor 26 is completely blocks forward voltage, and the auxiliary thyristors 34 and IGBT 96 are commanded to turn-off. From t 3 to t 4 , in FIG. 14D , the voltage of the resonant capacitor 98 increases until reaching the clamping voltage. In FIG. 14E , loop energy has been absorbed by the snubber circuit 100 , and the load current is zero and has been completely interrupted at t 4 . It is understood that a variety of resonant turn off circuits 44 may be used herein, with additional examples being provided in U.S. patent application Ser. No. 16/795,988, which is incorporated by reference herein.

While preferred embodiments of the inventions have been described, it should be understood that the inventions are not so limited, and modifications may be made without departing from the inventions herein. While each embodiment described herein may refer only to certain features and may not specifically refer to every feature described with respect to other embodiments, it should be recognized that the features described herein are interchangeable unless described otherwise, even where no reference is made to a specific feature. It should also be understood that the advantages described above are not necessarily the only advantages of the inventions, and it is not necessarily expected that all of the described advantages will be achieved with every embodiment of the inventions. The scope of the inventions is defined by the appended claims, and all devices and methods that come within the meaning of the claims, either literally or by equivalence, are intended to be embraced therein.