SWITCHING DEVICE FOR A DC NETWORK AND OPERATING PROCEDURES FOR THE SWITCHING DEVICE

Description

[0001] Switching devices for DC networks are used in such networks to isolate individual loads, as well as subnetworks (network branches or load areas), from the rest of the DC network. This isolation can represent an intentional shutdown by a switching command from a higher-level control system, or it can be effected by the switching device itself as a protective function, for example in the event of a fault.

[0002] It has proven advantageous to use controllable semiconductor switches, such as IGBTs, in such switching devices. Typically, at least one of the semiconductor switches in a switching device is located in the load current path and carries the full load current when switched on. In normal operation, the semiconductor switch is therefore permanently switched on. During such normal operation, only defects that cause the switch to exhibit high resistance can be detected. Any defects that prevent the switch from generating a blocking voltage are only detected the next time it is switched off, i.e., when the switch fails. In this case, the system is no longer protected.

[0003] In bidirectional switching devices, two semiconductor switches are typically arranged anti-series in the load current path. Often, semiconductor switches are used that include a parallel freewheeling diode, either as a component or intrinsically, such as IGBTs. With these, even during the next switch-off, a conductive defect in one of the semiconductor switches remains undetected if the current direction is such that the freewheeling diode prevents the voltage build-up across that semiconductor switch, meaning the other – non-defective – semiconductor switch builds up the blocking voltage.

[0004] From US 2013 / 0009491 A1 it is known to test freewheeling diodes of a switched-off IGBT by applying a voltage in the forward direction to the IGBT and detecting the maintenance of the voltage.

[0005] From US 7,567,095 B2, a safe output circuit with a single-channel peripheral connection for the output of a bus participant is known.

[0006] From FR 2 991 461 A1 it is known to test a switch comprising two anti-series connected MOSFETs by measuring two voltages and comparing the measured voltages.

[0007] The object of the present invention is to provide a switching device that avoids the aforementioned disadvantage, in particular allowing functional testing independent of the current direction. A further object is to provide an improved operating method for such a switching device.

[0008] This problem is solved by a switching device with the features of claim 1. Another solution consists of the operating method with the features of claim 6.

[0009] The switching device according to the invention is defined in claim 1.

[0010] By switching off the first semiconductor switch and applying the current to the test circuit that includes the first semiconductor switch, a functional test for the switch is advantageously enabled, which does not depend on the first semiconductor switch actually generating a blocking voltage itself.

[0011] This is particularly advantageous because the switching device includes a second controllable semiconductor switch. The control system is designed to activate both semiconductor switches with the control pulse to briefly interrupt the load current. It is advantageous if the switching device includes a load relief network arranged in parallel to the semiconductor switches, which takes over the load current for the short interruption period without generating an excessively high voltage.

[0012] If each semiconductor switch includes a parallel freewheeling diode, the control pulse causes a voltage buildup in only one of the two semiconductor switches. Which switch is affected depends on the direction of the load current. In the following, it is assumed that it is the second semiconductor switch. The voltage buildup in the first semiconductor switch is then limited by the freewheeling diode. The control system can be configured to additionally determine a second value representing the voltage or voltage change across the second semiconductor switch as a result of the control pulse. A signal is then calculated from both values ​​and output, representing the functionality of both semiconductor switches.

[0013] Further advantageous embodiments of the switching device and the operating method according to the invention are described in the dependent claims. The embodiments of the independent claims can be combined with the features of one of the dependent claims or, preferably, with those of several dependent claims. Accordingly, the following additional features can be provided: The control can be configured to operate with a time interval of less than 10 µs, in particular less than 5 µs. These short switching times result in a significantly reduced impact on ongoing operation. The voltage build-up caused by the energy stored in the DC network via capacitors connected in parallel to the semiconductor switches, for example, in a load balancing network, remains low. This keeps the equalizing current during re-energization low, especially for capacitive loads.

[0014] The control system can be designed to reduce the conductivity of the semiconductor switch(es) with the control pulse, but without switching it off. While the corresponding control circuit, for example, a gate circuit, must be suitable for this, this advantageously allows for a gentler functional test with less impact on mains operation than a complete shutdown.

[0015] Alternatively, the control system can be designed to completely switch off the semiconductor switch(es) with the control pulse, i.e., to put them into a non-conductive state, provided they are functional. This results in a clearer signal compared to a decrease in conductivity, and controlling the contact requires no additional effort, since switching off must be possible anyway.

[0016] According to the invention, the switching device has a test circuit for each of the semiconductor switches. Since the test circuit is necessary for the functional test of the semiconductor switch that remains largely de-energized when switched off, this enables the functional test of both semiconductor switches regardless of the current direction.

[0017] The resistors in the resistive voltage dividers can advantageously have resistance values ​​of less than 10 kΩ, particularly less than 5 kΩ. The current flow through the load contacts of the semiconductor switch under test is determined by the resistance values ​​and an applied test voltage. This current flow, in turn, determines the rate at which a voltage is built up across the switched-off semiconductor switch, assuming it is blocking. This, in turn, determines the measurement speed. The aforementioned values ​​advantageously enable a sufficiently fast measurement within the time window defined by the switch-off signal from the control pulse, which, depending on the design, is only 10 µs or less.

[0018] The operating method according to the invention is defined in claim 6.

[0019] A key advantage of the invention is that it enables a reliable assessment of the functionality of switching devices with anti-series connected semiconductor switches. This testing allows for the early detection of failures due to aging, spontaneous component failure, or overloading, for example, by excessive temperature or voltage. This prevents damage from consequential errors or production downtime resulting from the failure of the protective or switching function, increases the availability of electrical systems connected to a power grid, and can also extend the intervals between maintenance.

[0020] The test can be advantageously carried out during operation. Therefore, there is no need to disconnect any connected load(s) for testing the switching device. It is also advantageous that both current directions do not need to occur within a test interval. A particular advantage is that testing is even possible when there is no load current at all.

[0021] The invention will now be described and explained in more detail with reference to the exemplary embodiments shown in the figures.

[0022] They show schematically: Figure 1 shows a DC network with subnetworks, Figure 2 shows a bidirectional DC switch with a relief network and test circuits.

[0023] Figure 1Figure 1 shows a schematic DC network 10. The DC network 10 comprises a DC voltage source 11 that supplies the DC network 10 with a DC voltage. The DC voltage source 11 can, for example, be a rectifier connected to a supply network. However, it is also possible that the DC voltage source is a generator in conjunction with a rectifier.

[0024] The DC network 10 continues to include a number of network participants 16. In the Figure 1While these network participants 16 are all depicted as identical, in reality, they can have very different properties. For example, one or more of the network participants 16 may be a partially capacitive or inductive load instead of a purely resistive one. For instance, some of the network participants 16 may be electric motors. Similarly, network participants 16 may be prosumers, meaning they not only consume energy but can also feed electrical energy back into the DC network 10. For example, network participants 16 may contain a battery. Finally, some of the network participants 16 may also be or contain generators, such as a photovoltaic system.

[0025] The in Figure 1The exemplary DC network 10 shown comprises a first and a second subnetwork 12, 13. The first subnetwork 12 includes the DC voltage source 11 and some of the network participants 16. The second subnetwork 13 includes some further network participants 16. The two subnetworks 12, 13 are connected to each other in the DC network 10 via a DC switch 14, or simply DC switch 14. The subnetworks 12, 13 can therefore be disconnected from each other by the DC switch 14. Since the direction of current flow cannot be reliably determined at the DC switch 14, it is a bidirectional DC switch 14.

[0026] Near each of the other network participants 16, the DC network also includes a DC switch 14. For some applications, unidirectional DC switches could be used here. However, for this example, it is assumed that bidirectional DC switches 14 are also used near the network participants 16.

[0027] Figure 2Figure 1 shows such a bidirectional DC switch 14, as used in the exemplary DC network 10. The DC switch 14 comprises a first IGBT 151 with a parallel freewheeling diode 153. Furthermore, the DC switch 14 includes a controller 152, which is connected to the gate terminal of the IGBT 151 and controls it during operation by means of control signals in the form of different voltage levels. It also has a second IGBT 155, which is connected anti-series, i.e., in series but with the opposite orientation to the IGBT 151. The second IGBT 155 also has a parallel freewheeling diode 156, which is also arranged opposite to the orientation of the freewheeling diode 153. The second IGBT 155 is also controlled by the controller 152, i.e., supplied with a gate voltage. The DC switch 14 of the Figure 2 This makes it a bidirectional DC switch, meaning it can switch off a load current regardless of its direction.

[0028] Furthermore, the DC switch 14 includes a load-suppression network 20. This network has three branches, the first and second of which are connected in parallel to each other and in parallel to the two IGBTs 151 and 155. The first branch comprises a series connection of a varistor 201 with a capacitor 202. The second branch comprises a series connection of a resistor 203 with a second capacitor 204. The third branch comprises a second resistor connected in parallel to the second capacitor 204. The center terminals between the elements in the first and second branches are also connected.

[0029] The load relief network 20 ensures that the energy inductively stored in the DC network 10, for example in the line inductors 191, 192, is dissipated when the DC switch 14 is turned off, without causing overvoltages that could damage the IGBTs 151, 155. In addition to the form for the load relief network 20, which is described in Figure 2 As depicted, there are other forms that can also perform the described tasks.

[0030] The DC switch 14 is connected in series to a DC line carrying current in direction 18. A defect in the DC switch 14, leading to a permanent and unintended high-resistance state (i.e., not initiated by the controller 152), is therefore directly detectable during operation. However, a state in which the switch remains permanently low-resistance cannot be detected directly, but only during the next attempted shutdown if it fails. The controller 152 of the DC switch 14 is therefore designed to briefly shut down the IGBTs 151 and 155 using a suitable control pulse for functional testing of the DC switch 14.

[0031] Provided that the IGBTs 151 and 155 are functioning, the load relief network 20 then takes over the load current. Because the load relief network 20 has capacitors in its current path, a brief shutdown of the IGBTs 151 and 155 results in an approximately linear increase in the voltage across the DC switch 14, while the load current flows through the load relief network 20 and the capacitors 202 and 204 are charged.

[0032] This only applies, however, if the IGBTs 151 and 155 actually become non-conducting when they receive a control pulse from controller 152, which normally causes them to switch off (for example, a reduction in gate voltage). Thus, by detecting the voltage across the IGBTs 151 and 155, it can be determined whether they are responding to the control pulse as intended. If they are, a voltage build-up occurs in response to the control pulse. If no voltage build-up occurs, or if the voltage build-up is insufficient, a malfunction has occurred. A threshold voltage can be defined, for example, to determine the fault condition. If this threshold is not reached due to the control pulse, a malfunction has occurred.

[0033] It is advantageous if the switch-off time is dimensioned such that the voltage difference between the input and output of the DC switch 14 does not become too large during the duration of the control pulse. A control pulse duration of less than 1 ms is suitable for this purpose. In particular, durations of less than 10 µs are sufficient for a functional test of the DC switch 14. Short durations in the µs range generally minimize the influence of the functional test on ongoing operation and, especially with capacitive loads, keep the equalizing current low when switching on again, i.e., at the end of the control pulse.

[0034] In an alternative embodiment of the DC switch 14, the control pulse is dimensioned such that it does not switch off the IGBTs 151 and 155, but only reduces their conductivity. For example, the reduction in gate voltage can be smaller than what would be required for a complete shutdown of the IGBTs 151 and 155. The IGBTs 151 and 155 then continue to operate in the active region and conduct the load current. Nevertheless, the voltage across the IGBTs 151 and 155 increases due to the control pulse.

[0035] In this way, a functional test for the IGBTs 151 and 155 is possible, whereby the resulting voltage across the IGBTs 151 and 155 is lower when completely switched off, thus reducing problems with overvoltage or equalizing currents for the same duration of the control pulse. Conversely, the controller 152 must be designed to generate a corresponding control pulse, for example, a reduction in the gate voltage.

[0036] Since IGBTs 151 and 155 each have a parallel freewheeling diode 153 and 156, the voltage buildup resulting from the control pulse actually affects only one of the IGBTs 151 and 155 to a large extent. The freewheeling diode 153 and 156 prevents the other IGBT 151 and 155 from developing a voltage. Which of the two IGBTs 151 and 155 remains almost voltage-free depends on the direction of the load current. If the load current flows in the Figure 2 From the side of IGBT 151 to the side of IGBT 155, the voltage build-up occurs on IGBT 151. In the other case, a voltage build-up only occurs across IGBT 155.

[0037] Using the described method for determining whether IGBTs 151 and 155 are conductively defective, only one of the two IGBTs 151 and 155 can actually be tested unless both load current directions are available for measurement. If the load current drops out completely for a period of time, neither IGBT 151 or 155 can be tested.

[0038] To address this problem, the DC switch 14 additionally includes a first and a second test circuit 41, 42, wherein the first test circuit 41 serves to functionally test the first IGBT 151 and the second test circuit 42 serves to functionally test the second IGBT 155. The first and second test circuits 41, 42 are mirror images of each other.

[0039] Each circuit comprises a first voltage divider formed by a first and second test resistor 411, 412, 421, 422. It also comprises a second voltage divider comprising a third and fourth test resistor 413, 414, 423, 424. The first and second voltage dividers are connected to each other at one of their external terminals and to node 48 between the two IGBTs 151, 155, i.e., to their emitters. The other external terminals are connected to an auxiliary potential 416, 426, which is, for example, +15V relative to node 48.

[0040] The midpoint of the second voltage divider is connected to the collector of the respective IGBT 151, 155 via one or more decoupling diodes 417, 427, which must be able to block the voltage applied during operation. Furthermore, the midpoint of the second voltage divider is connected to the inverting input of an operational amplifier 415, 425 of the respective test circuit 41, 42. The midpoint of the first voltage divider is connected to the non-inverting input of the respective operational amplifier 415, 425. The outputs of the operational amplifiers 415, 425 are connected to the controller 152 and transmit the result of the functional test to the controller 152.

[0041] The first voltage dividers establish a reference voltage. For example, resistance values ​​of 1 kΩ can be used for the first test resistor 411, 421 and 2 kΩ for the second test resistor 412, 422, which in this example results in a reference voltage of 10 V at the non-inverting input of the operational amplifier 415, 425.

[0042] For the second voltage divider, resistance values ​​of 1 kΩ can be used for the third test resistors 413 and 423, and 3 kΩ for the fourth test resistors 414 and 424, respectively. One of the two IGBTs 151 and 155 is connected in parallel to the fourth test resistor 414 or 424, respectively. The IGBTs 151 and 155 are therefore also part of the respective test circuit 41 and 42.

[0043] If each IGBT 151, 155 is functioning correctly (i.e., actually high-impedance) during the brief period in which both IGBTs 151, 155 are switched off by the control pulse, this results in a voltage corresponding to the second voltage divider, i.e., 11.25 V, at the inverting input of the respective inverting input of the operational amplifier 415, 425. Since this voltage is higher than the reference voltage, the respective operational amplifier 415, 425 outputs a "low" signal, which can be processed as a signal indicating a functioning IGBT 151, 155.

[0044] If, on the other hand, the tested IGBT 151, 155 conducts despite the control pulse, i.e., is defective, the voltage at the inverting input drops to a value corresponding to the sum of the forward voltages of the decoupling diodes 417, 427 and IGBT 151, 155, i.e., approximately 3 V relative to node 48. The respective operational amplifiers 415, 425 subsequently output a "high" signal, which can be interpreted as a signal indicating a conducting defect in IGBT 151, 155.

[0045] Since the fastest possible measurement is of interest for this measurement function, higher currents are used in contrast to known saturation measurements for IGBTs, and therefore lower third and fourth test resistors 413, 414, 423, 424 are used.

[0046] In this way, a DC switch 14 is created that is capable of testing its controllable semiconductor switches for conductive defects during operation, i.e., under load current, and detecting such defects before the next switch-off operation actually occurs. This also allows testing of the one of two anti-series IGBTs 151, 155 that, due to the current direction, cannot be energized even during a switch-off. This is achieved by exploiting the fact that the load-dissipating network takes over the load current during switch-off, allowing a current to be applied to the affected, de-energized IGBT 151, 155. The voltage drop, which only occurs with a functioning IGBT 151, 155, is evaluated as a signal for the functional test.

[0047] Furthermore, the design of the DC switch according to the Figure 2even a functional test in the load-free state. If there is no load current at all in the respective subnetwork 12, 13 in which the DC switch 14 is located, then the blocking and conductivity can even be tested by applying the current through the test circuits 41, 42, whereby the controller 152 must switch the IGBTs 151, 155 on and off appropriately for short time periods.

[0048] It is understood that the resulting voltages from switching off and applying a current can be evaluated in various ways, for example, analogously or digitally, using a comparator, or by direct connection to a microcontroller. The resulting signal, representing the functionality of the IGBTs 151 and 155, can also take various forms. For instance, the signal can be stored in a protocol, output as an optical or acoustic signal, or fed as an electrical signal to a higher-level control or evaluation unit, or a combination of these methods.

Claims

1. Switching device (14) for a DC voltage grid (10), comprising a first and a second controllable semiconductor switch (151, 155) that are connected in anti-series and each have a control contact, and a controller (152) for the semiconductor switches (151, 155), configured to control the semiconductor switches (151, 155) by means of an electrical control signal at the control contact, wherein the controller (152) is configured to carry out the following steps: - actuating the semiconductor switches (151, 155) using a control pulse that, in the case of a functional semiconductor switch (151, 155), causes the electrical conductivity of the semiconductor switch (151, 155) to reduce for a time period of less than 1 ms, - applying a current to a test circuit (41, 42), - ascertaining a first value representing the voltage or the change in voltage across one of the semiconductor switches (151, 155) as a result of the control pulse and the application, - ascertaining from the first value, and outputting, a signal that represents the functionality of the semiconductor switch (151, 155), characterized by a first and a second test circuit that each comprise a first and a second resistive voltage divider, wherein - the first and the second voltage divider are connected to one another at one of the outer connections thereof, and connected to a node (48) between the two semiconductor switches (151, 155), - the voltage dividers are connected to an auxiliary potential (416, 426) at a second one of the outer connections thereof, - in the case of the first voltage dividers, the respective central connection thereof is connected to a load connection of a respective semiconductor switch (151, 155) via one or more respective decoupling diodes (417, 427), and - the central connections of both voltage dividers are connected to inputs of a respective amplifier (415, 425).

2. Switching device (14) according to Claim 1, in which the controller (152) is configured to use less than 10 µs as the time period.

3. Switching device (14) according to Claim 1 or 2, in which the controller (152) is configured to deactivate the semiconductor switch (151, 155) using the control pulse.

4. Switching device (14) according to one of the preceding claims, in which the semiconductor switches (151, 155) each comprise a parallel freewheeling diode (153, 156), wherein the control pulse causes a build-up of voltage at the second semiconductor switch (151, 155), whereas the build-up of voltage at the semiconductor switch (151, 155) is limited by the freewheeling diode (153, 156), wherein the controller (152) is configured to additionally carry out the following steps: - ascertaining a second value representing the voltage or the change in voltage across the second semiconductor switch (151, 155) as a result of the control pulse, - ascertaining from the values, and outputting, a signal that represents the functionality of both semiconductor switches (151, 155).

5. Switching device (14) according to one of the preceding claims, in which the resistors (413, 414, 423, 424) of the resistive voltage divider have resistance values of less than 10 kohm, in particular less than 5 kohm.

6. Operating method for a switching device (14) for a DC voltage grid (10) according to one of the preceding claims, in which the semiconductor switches (151, 155) are controlled by means of an electrical control signal at the control contact, in which: - the semiconductor switches (151, 155) are actuated using a control pulse that, in the case of a functional semiconductor switch (151, 155), causes the electrical conductivity of the semiconductor switch (151, 155) to reduce for a time period of less than 1 ms, - a current is applied to a test circuit (41, 42), - a first value representing the voltage or the change in voltage across the semiconductor switch (151, 155) is ascertained as a result of the control pulse and the application, - a signal that represents the functionality of the semiconductor switch (151, 155) is ascertained from the first value, and output.

7. Operating method, according to Claim 6, for a switching device (14) for a DC voltage grid (10), wherein the switching device (14) has a first and a second controllable semiconductor switch (151, 155), wherein the semiconductor switches (151, 155) each comprise a parallel freewheeling diode (153, 156) and are connected in anti-series, in which: - both semiconductor switches (151, 155) are actuated using the control pulse, as a result of which a build-up of voltage is caused at the second semiconductor switch (151, 155), whereas the build-up of voltage at the first semiconductor switch (151, 155) is limited by the freewheeling diode (153, 156), - a second value representing the voltage or the change in voltage across the second semiconductor switch (151, 155) is ascertained as a result of the control pulse, - a signal that represents the functionality of the semiconductor switches (151, 155) is ascertained from the first value and the second value, and output.