Method for operating a power semiconductor switch, operating circuit for a power semiconductor switch, and electronic protection switch

Abstract

This invention relates to a method for controlling a power semiconductor switch (14A, 14B) of an AC circuit, the AC circuit being able to be turned on and / or off by the power semiconductor switch. The method comprises the following steps: a) determining an instantaneous current value and an instantaneous voltage value of the AC circuit; b) determining whether the instantaneous current value exceeds a pre-given maximum value; if so, c1) generating a control signal to turn off the circuit; c2) generating a control signal to turn on the circuit within a time interval from the generation of the control signal to turn off the circuit, wherein the time interval is less than or equal to the duration of a voltage cycle; and c3) determining whether the instantaneous current value after the circuit is turned on exceeds a pre-given maximum value, the maximum value corresponding to or less than a previous maximum value. If so, steps c1), c2), and c3) are repeated. In step d), if the number of repetitions of steps c1), c2), and c3) exceeds a value n, a fault condition is identified and a control signal to disconnect the circuit is persistently output, and the method is terminated. Otherwise, in step e), if it is determined in step b) or step c3) that the instantaneous current value does not exceed the corresponding maximum value, then the control signal for turning on the circuit continues to be generated, and the method continues with step a).

Description

Technical Field

[0001] The present invention relates to a method for controlling a power semiconductor switch, a control circuit for a power semiconductor switch, and an electronic protection switch. Background Technology

[0002] High inrush currents, also known as surge currents, often occur when connecting electronic loads, such as those with switching power supplies and / or rectifiers, to an AC voltage power supply. Even with relatively low power ratings, such as switching power supplies with a power rating of less than 100W, current spikes of hundreds of amperes can occur. These inrush currents can cause unintended triggering of conventional power protection switches (MCBs).

[0003] A typical example of a consumer with low continuous power and high turn-on current is an LED emitter, which is increasingly replacing other light sources in homes and building technology. LED emitters used for operation on a 230V mains grid also have a power supply that exhibits capacitive behavior at the moment of turn-on. In practice, many of these emitters are often connected in parallel and thus turned on simultaneously, resulting in a corresponding multiplication of the turn-on current and potentially overloading the circuit and / or undesirable triggering of the MCB (Mechanical Control Block).

[0004] Furthermore, for example, if a high current-carrying capacity occurs in combination with mechanical contact jitter, the high current-carrying capacity can also cause increased wear in conventional switching devices such as relays, contactors, or switches. The jitter associated with high current can then lead to brief arcing at the switch contacts, which in turn causes contact burn-out and corresponding wear of the switching elements, and in extreme cases, contact welding.

[0005] Another problem with using modern semiconductor circuit breakers (SCCBs, sometimes also called solid state circuit breakers, or SSCBs; hereinafter referred to as SCCBs) is that these SCCBs themselves can be damaged by excessive current.

[0006] The entire current circuit must be designed using conventional components to ensure that the maximum on-state current does not cause interference or overload. In practice, this means that all components of the circuit must be designed for multiples of the set continuous load. This correspondingly increases costs and results in the power protection switch not protecting against the actual application conditions defined by the set continuous load, but rather protecting against a trade-off between the continuous load and the on-state current. While this trade-off can be achieved relatively cost-effectively in conventional electromechanical MCBs by selecting the MCB's trigger characteristics, the following trade-off persists: the MCB is not optimally designed for the corresponding application conditions, especially using an MCB with slower trigger characteristics, which in turn results in higher fault currents in fault conditions. In many cases, the use of SCCBs is simply not economically feasible. Summary of the Invention

[0007] Therefore, the object of the present invention is to describe an improved method for controlling a power semiconductor switch, an improved control circuit for a power semiconductor switch, and an improved electronic protection switch, which are suitable for use in circuits with potentially very high turn-on current.

[0008] This task is accomplished by a method having the features of independent patent claim 1, by a control circuit having an apparatus for performing the method according to the invention, and by an electronic protection switch having such a control circuit.

[0009] Advantageous improvements of the invention are described in the dependent claims.

[0010] The advantages of this invention can be seen in that, upon detecting an excessively high turn-on current, the excessive turn-on current is suppressed by briefly interrupting the circuit, and then the circuit is turned on again, wherein the duration between the interruption and re-enabling corresponds at most to the cycle duration. In this case, in many consumables of the type discussed at the beginning, the capacitive elements of the consumable are pre-charged by the turn-on current surge limited in time by this method, resulting in a lower turn-on current during re-enabling. Preferably, re-enabling occurs at a (lower) voltage value, for example, a voltage value occurring in a time window near a zero crossing, such that the voltage value at re-enabling is less than approximately 35% of the peak AC voltage, the time window beginning with a voltage phase angle of -20° relative to the next zero crossing and ending with a voltage phase angle of +20° relative to the next zero crossing. In other designs, a smaller phase angle is chosen, and therefore a smaller voltage, for example, approximately 25% of the peak voltage at + / -15° near the zero crossing, or approximately 17% of the peak voltage at + / -10° near the zero crossing. In a particular design of the invention, the circuit is reconnected before the next zero crossing of the voltage, that is, before the voltage follows the zero crossing after being cut off. This additionally limits the on-current during reconnection.

[0011] If the existing capacitor has not been sufficiently precharged, an excessively high turn-on current is detected according to the invention, and the above-mentioned steps are repeated until either the excessively high turn-on current no longer occurs or, after n repetitions of the method, it is determined that the turn-on current continues to be excessive and thus a consumer with excessively large capacitance, a short circuit, or more generally, a fault is determined.

[0012] In other words, this invention reliably prevents excessive turn-on current, and unlike the prior art, it eliminates the need to design the entire circuit to handle high turn-on currents generated by permissible consumers. It also eliminates the need to intentionally reduce the speed of the protective switching device, for example, by using a protective switching device with C characteristics instead of a protective switching device with B characteristics, to prevent erroneous triggering of the protective switching device when a permissible and properly functioning consumer is connected. In this case, in the embodiment, turn-on current is advantageously limited centrally via an SCCB, eliminating the need for special components for turn-on current limiting, such as DALI circuits and phase cutoff controllers, thereby saving costs on the consumer side.

[0013] Additionally, it can be advantageously achieved, through the method according to the invention, that the SCCB can be used entirely without the risk of damaging the SCCB due to excessively high turn-on current. At least in any case, a given SCCB can be extended according to the method according to the invention and thus adapted to circuits comprising power consumers having typically high turn-on current and low continuous current.

[0014] In summary, this achieves improved protection for electrical consumers in circuits and the entire electrical system. In this case, the method according to the invention can be arbitrarily extended and used in virtually any voltage and current range.

[0015] A particular advantage is that by repeatedly cutting off and reconnecting, loads with very large turn-on currents, such as rectifiers with very large intermediate circuit capacitors, can also be started with a limited current. Only the number of repetitions, n, must be adapted accordingly.

[0016] Furthermore, it is advantageous that the necessary measuring devices for detecting instantaneous values ​​of current and voltage in modern SCCBs already exist or can be implemented with very little additional cost. In addition, a controller is usually present, making it often possible to implement the invention using existing hardware.

[0017] Furthermore, it is advantageous that the power consumers already connected to the circuit are not significantly affected by the cut-off and reconnection processes performed within a single full wave or, in embodiments, within a single half wave. Considering typical household circuits, it can be assumed, for example, that for reasons of interference immunity, any power consumer is generally unaffected relative to multiple short cut-off and reconnection processes.

[0018] Another important advantage of the present invention can be seen in that, unlike conventional soft starters which are separately allocated to the motor, a 1:1 relationship does not necessarily exist between the consumer and the circuit for limiting the on-current, which, for example, brings cost advantages. More specifically, in typical applications, the protective switch controlled according to the present invention is used to protect circuits with multiple consumers. Attached Figure Description

[0019] In the following sections, embodiments of the invention are explained in more detail with reference to the accompanying drawings.

[0020] In the attached diagram:

[0021] Figure 1 A schematic diagram of a semiconductor power protection switch according to an embodiment of the present invention is shown;

[0022] Figure 2 An exemplary time distribution of current and voltage is shown in relation to an application of an embodiment of the method according to the invention;

[0023] Figure 3 An exemplary process is shown, illustrating an embodiment of the method according to the present invention; and

[0024] Figure 4 An exemplary time distribution of current and voltage is shown in relation to an application of another embodiment of the method according to the invention. Detailed Implementation

[0025] Figure 1 A schematic diagram of a semiconductor power protection switch (hereinafter: SCCB) 10 according to an embodiment of the present invention is shown. The SCCB 10 has grid-side terminals 11A and 11B, which enable the SCCB 10 to be connected to a power grid (not shown). The first grid-side terminal 11A is used here to connect to the neutral conductor N, and the second grid-side terminal 11B is used to connect to the phase conductor L.

[0026] Voltage measuring device 12 is used for voltage measurement on the input side of SCCB 10. The voltage value determined by voltage measuring device 12 is transmitted to controller 13 (shown by dashed lines). In this case, voltage measuring device 12 can be designed such that a signal representing the voltage between input terminals 11A and 11B is continuously transmitted to controller 13 in analog form. In an alternative design, the voltage between input terminals 11A and 11B is sampled by voltage measuring device 12 at discrete time points and provided to controller 13 as a time-discrete digital signal, wherein the sampling frequency is selected relative to the mains frequency, so that controller 13 can determine the time distribution of the voltage applied to input terminals 11A and 11B from the signal sequence, especially the zero-crossing time points, and determine them by interpolation if necessary.

[0027] The neutral conductor connected to the first grid-side terminal 11A is directly connected to the first output-side or consumer-side terminal 18A of the SCCB 10. The phase L applied to the second grid-side terminal 11B is connected to the second output-side or consumer-side terminal 18B via power semiconductor circuits 14A and 14B. Figure 1 In one example, the power semiconductor circuit has two self-commutated power semiconductor switches 14A and 14B, which are controlled by a controller 13 to connect phase L to the second output side or consumer side terminal 18B or to disconnect the connection between the second grid side terminal 11B and the second consumer side terminal 18B. In embodiments of the invention, parallel-connected power semiconductors (not shown) may also be used.

[0028] exist Figure 1In the example, the energy absorber 16 is connected in parallel with the power semiconductor circuits 14A and 14B. The energy absorber is connected between the grid-side jumper terminal 15A and the load-side jumper terminal 15B and is used for voltage limiting and therefore for protecting the power semiconductor circuits in the event of a defined switching event.

[0029] A current measuring device 17 is arranged in the phase conductor between the power semiconductor circuits 14A, 14B and the second consumer-side terminal 18B and is used to measure the load current in the phase conductor. In this case, the current measuring device 17 can be designed such that a signal representing the current flowing in the phase conductor is continuously supplied to the controller 13 in analog form (indicated by the dashed line between the current measuring device 17 and the controller 13). In an alternative design, the current is detected by the current measuring device 17 at discrete time points and provided to the controller 13 as a time-discrete digital signal, wherein the sampling frequency is selected to be high so that, for example, a significant increase in current distribution caused by a short circuit for the corresponding application can be detected in a timely manner and translated into appropriate action.

[0030] Three power consumers 20, 30, and 40 are connected to the load-side terminals 18A and 18B of the SCCB 10. Consumers 30 and 40 should be arbitrary and are only shown to illustrate that if other consumers 30 and 40 in the corresponding circuit protected by the SCCB 10 are already active before consumer 20 is turned on (which corresponds to a typical practical application), the method for limiting the on-state current of consumer 20, as described below, can also be implemented.

[0031] exist Figure 1 In the example, consumer 20 is a consumer of the type described at the beginning, that is, a consumer with low continuous power and high turn-on current, such as an LED lighting system in a large room comprising multiple individual LED emitters. This is indicated by rectifier 21 and capacitive load 22. In embodiments of the invention, not only rectifier 21 but also capacitive load represents multiple rectifiers and capacitive loads connected in parallel, as is the case, for example, in an LED lighting system in a large hall. Consumer 20 can be turned on or off by switch 23. In this case, switch 23 can be a mechanical wall switch that can be operated by a user, or it can be an electronically controlled switch.

[0032] A first variation of the method according to the invention is incorporated herein by reference. Figure 2 Let me explain. Figure 2An exemplary time distribution 220 (right scale axis of the graph) of the current I detected by the current measuring device 17 and an exemplary time distribution 210 (left scale axis of the graph) of the voltage U detected by the voltage measuring device 12 are shown in a single graph. The case of a 230V low-voltage power grid is shown purely exemplarily, in which the peak voltage of a single phase L relative to the neutral conductor N at time t=0 is approximately 325V.

[0033] At time t=0, the power consumer 20 is connected, causing a very high turn-on current 221. This current is determined by the current measuring device 17 and processed by the controller 13. In this embodiment, the controller 13 is configured such that a current exceeding 80A is not allowed. Therefore, when the current value of 80A is reached, the controller 13 controls the power semiconductor circuit to turn off, that is, the current flow in the phase conductor L between the input terminal 11B and the output terminal 18B is interrupted, and then the current and the current measurement value drop to zero.

[0034] Then, shortly before the sinusoidal (grid) voltage U reaches zero at t=0.005s, the power semiconductor circuit is switched on by controller 13, and the measured current value again shows a large increase 222. However, this increase is flatter than at t=0 because the capacitor of consumer 20 has been partially charged and / or because the instantaneous voltage value is low and has a downward trend. Therefore, the increase and maximum value of the current are limited independently of consumer 20.

[0035] If the consumer 20 is as follows Figure 1 As shown in the diagram, a consumer with capacitive element 22 typically forms a zero-current phase 230 near the voltage zero crossing due to the partially charged capacitor. This fact and the term "zero-current phase" are further explained in detail below.

[0036] After the voltage zero-crossing (characterized here by entering the negative half-wave) and at the end of the zero-current phase 230, the current rises again to a value 224 significantly higher than the normal value for the corresponding circuit, but no longer reaches the trigger value of 80A, and returns to a low continuous current value of a few hundred milliamps for the consumer 20 during the negative half-wave of the mains voltage. Figure 2 The continuous current value in the current model cannot be distinguished from the current value I=0 due to the chosen scale.

[0037] Controller 13 determines that there is no short circuit based on the fact that the current does not reach the 80A limit shortly before the voltage crosses zero after being switched on again. SCCB 10 continues to operate by monitoring the current.

[0038] In a preferred embodiment of the invention, the timing for reactivating the power semiconductor circuit, as described above using "shortly before reaching zero crossing," can preferably be selected by the controller 13 as a timing point where, depending on whether the circuit was interrupted by the controller during the positive or negative half-wave of the mains voltage, the phase angle of the sinusoidal mains voltage (i.e., at the input terminals 11A, 11B of the SCCB 10) is between 160° and 170° or between 340° and 350°.

[0039] In embodiments of the invention, the zero-current phase can be used to distinguish between short circuits and turn-on processes. This is because a characteristic current time distribution typically occurs during a short circuit, and a zero-current phase does not actually occur, as short circuits exhibit ohmic behavior, meaning the current time distribution has high current values ​​and is very similar to the voltage time distribution. The situation is different in the case of a turn-on process of a consumer 20 with capacitive elements that results in a very high turn-on current 221: here, if otherwise no consumer is turned on in the circuit, the desired current time distribution near the zero-crossing of the voltage is characterized by a low current value that can be zero – hence the simplified term "zero-current phase" used here.

[0040] Figure 3 The figure above illustrates an embodiment of the method according to the present invention. Figure 2 The described process. The method begins with an initialization step 310, which provides a maximum current value, for example, by reading from memory, and provides a maximum current value for monitoring step 320.

[0041] In monitoring step 320, the current current value is compared with the maximum current value. If the current current value does not exceed the maximum current value, then the power semiconductor circuits 14A and 14B remain on, step 330, and the method continues with step 320.

[0042] If the current current value corresponds to or exceeds the maximum current value, then the power semiconductor circuits 14A and 14B are cut off in step 340, and in step 350 it is checked whether a short circuit or fault condition can be identified, because, for example, the counter (not shown) incremented in step 340 exceeds a determined value n.

[0043] If a short circuit or fault condition is determined in step 350, the method ends at step 370 and the power semiconductor circuits 14A and 14B remain disconnected until, for example, the short circuit is eliminated and manually reconnected (not shown). In embodiments of the invention, if a fault condition or short circuit condition is determined, an isolating switch (not shown) arranged in series with the power semiconductor switch can always be additionally disconnected. This isolating switch is preferably arranged on the grid side of the power semiconductor switch in conductor L1, but it can also be arranged on the load side of the power semiconductor switch. In this case, a single-pole or double-pole isolating switch may be used.

[0044] If it is determined in step 350 that the conditions for determining a short circuit or fault condition have not been met, then in step 360, the power semiconductor circuits 14A and 14B are reconnected. Figure 2 The comments made apply to the time when the connection is reconnected.

[0045] When re-entering monitoring step 320 from step 360, it can be additionally specified that another, for example, lower, maximum current value is used instead of the previous maximum current value. Additionally or alternatively, when re-entering monitoring step 320 from step 360, it can be specified, as in combination with... Figure 2 As explained, the method monitors the presence of a zero-current phase and identifies short-circuit or fault conditions when no zero-current phase occurs, thus terminating the method (not shown).

[0046] Another variation of the method according to the invention is incorporated below. Figure 4 An explanation will be provided. Figure 4 (like Figure 2 Thus, an exemplary time distribution 220A of the current I detected by the current measuring device 17 (the right scale axis of the graph) and an exemplary time distribution 210 of the voltage U detected by the voltage measuring device 12 (the left scale axis of the graph) are shown in a single graph. The case of a 230 V / 230 V low-voltage grid is also shown purely exemplarily, in which the peak voltage of a single phase L relative to the neutral conductor N is approximately 325 V.

[0047] At time t=0.005s, the power consumer 20 is connected, causing a very high turn-on current 221. This current is determined by the current measuring device 17 and processed by the controller 13. In this embodiment, the controller 13 is configured such that a current exceeding 50A is not initially allowed. Therefore, when the current value of 50A is reached, the controller 13 controls the power semiconductor circuit to turn off, i.e., the current flow in the phase conductor is interrupted, and the measured current value immediately returns to zero.

[0048] The control signal output to the power semiconductor circuit is in Figure 4 The lower edge of the graph is shown as a sequence of digital signals 240. From t=0 to t=0.005s, an on (EIN) signal is shown; from t=0.005s to shortly before t=0.01s, an off (AUS) signal is shown; then a very short on signal, followed by a very short off signal, then an on signal at approximately t=0.01s, and so on.

[0049] As is clear from signal sequence 240, shortly before the sinusoidal voltage U reaches zero at t = 0.01s, the power semiconductor circuit is switched on by controller 13 (first re-on), and immediately thereafter the current and its measured value have a large increase 222 and rise to over 50A. Controller 13 then re-controls the power semiconductor circuit, causing it to turn off, i.e., the current flow in the phase conductor is interrupted, and immediately the current and its measured value return to zero.

[0050] In one design of this method, the power semiconductor circuit can be switched on again directly by the controller 13, either shortly before or during the zero-crossing of the voltage U (second re-on), and then, due to the diodes in the rectifier 21 and / or the partially pre-charged capacitor 22, the current increase 224 is only performed when the voltage value exceeds the threshold voltage of the diodes in the rectifier and / or the voltage of the partially pre-charged capacitor.

[0051] In point 224, the current also exceeds the maximum value after the second reconnection, wherein in the illustrated embodiment, the maximum value in the second reconnection case is selected to be less than the maximum value in the first reconnection case and is, for example, 40A.

[0052] Since the maximum value is exceeded, the controller 13 re-controls the power semiconductor circuit, causing the power semiconductor switch to turn off, i.e., the current flow in the phase conductor is interrupted, and then the current and its measured value return to zero.

[0053] At the end of the (negative) half-wave of voltage U, the power semiconductor circuit is switched on by controller 13, and the circuit is reconnected (the third reconnection). In this case, the same time interval as in the first reconnection case can be selected before the voltage crosses zero, or a slightly later time point closer to the voltage crosses zero can be selected for the third reconnection.

[0054] The current increases significantly again at point 225, but does not exceed the maximum value. For this reason, controller 13 switches the power semiconductor circuit to conduct. However, after crossing zero, at the end of the zero current phase 231, the current at point 226 exceeds the maximum value of 40A again. Immediately afterwards, controller 13 re-operates the power semiconductor circuit, causing it to turn off, that is, the current flow in the phase conductor is interrupted, and then the current and its measured value return to zero.

[0055] At the end of the (positive) half-wave of voltage U, the power semiconductor circuit is switched on by controller 13, and the circuit is reconnected (the fourth reconnection). In this case, the same time interval as in the third reconnection case can be selected before the voltage zero crossing, or the fourth reconnection can be selected at a slightly later time point closer to the voltage zero crossing, or the fourth reconnection can be selected at a slightly earlier time point further away from the voltage zero crossing.

[0056] exist Figure 4 In the example, no current increase occurs shortly before t=0.03s after the fourth reconnection, because the connection time falls within the zero-current phase 232 caused by the partial charging of the capacitor 22 of the consumer 20.

[0057] After crossing zero and at the end of the zero current phase 232, the current at point 227 exceeds the maximum value of 40A again. Then, the controller 13 re-operates the power semiconductor circuit, causing the power semiconductor circuit to shut down, that is, the current flow in the phase conductor is interrupted, and then the current and its measured value return to zero.

[0058] At the end of the (negative) half-wave of voltage U, the power semiconductor circuit is switched to conduction by controller 13 and then turned on again (the fifth time it is turned on again).

[0059] exist Figure 4 In the example, no current increase occurs shortly before t=0.04s after the fifth reconnection, because the reconnection time falls back into the zero-current phase 233 caused by the partial charging of capacitor 22 of consumer 20.

[0060] After the zero-crossing and at the end of the zero-current phase 233, the current at point 228 increases significantly again, but does not reach the maximum value of 40A. Therefore, controller 13 keeps the power semiconductor circuit in the ON state, and the method continues to monitor the current, using the original maximum value of 50A either directly or after a defined time has elapsed, instead of the modified maximum value of 40A.

[0061] More generally, the method consisting of the steps of cutting off and re-connecting the power semiconductor can be repeated up to n times. If the current no longer exceeds the maximum value before the nth repetition, the controller determines that the turn-on process is normal. If this does not occur, i.e., the maximum current is exceeded even on the nth repetition of the method, the method is terminated, a fault condition is identified, and the operator is notified by signal if necessary.

[0062] As illustrated above, the maximum value can be constant for each reconnection, or it can be redefined for each reconnection, for example, reduced, in order to limit the total I of the system. 2 t load. Furthermore, the timing for re-energizing relative to zero can be altered, for example, by setting the re-energizing timing closer to zero with each repetition to achieve a lower energizing current based solely on the subsequent lower absolute value of the voltage, while still charging the capacitors on the consumer side, particularly the intermediate circuit capacitors of the rectifier, or to move the re-energizing to a desired zero-current phase in the case of rectifier consumers, and to identify a short-circuit condition when an undesirably high current occurs in that phase.

[0063] The number of repetitions, n, also depends on which other consumers 30, 40 are connected to the circuit protected by the SCCB 10, and especially on how many half-waves can fail without causing faults or damage at the other consumers. This can be derived from the standards applicable to the relevant circuit, or determined on a case-by-case basis according to the typically connected consumers.

[0064] exist Figure 4 In the example, n=5 was chosen. In other embodiments, n≤5 is applied, and in yet another embodiment, n≤10 is applied. In this case, the quantity n can be related not only to the type of consumer, as explained above, but also to the total load already present in the circuit in the specific case, which is caused by other active consumers 30, 40. Similarly, the maximum current value can be related to the existing load, especially when a load already exists in the circuit, particularly when the existing load exceeds, for example, 50% or 75% or 90% of the circuit's rated load or permissible continuous current, the initial maximum current value and all other maximum current values ​​can be chosen to be less than the standard maximum current value.

[0065] In embodiments of the invention, it may be specified that, in the event of a fault being identified, the circuit is automatically reconnected after a configurable waiting time (which may be from milliseconds to seconds or from seconds to minutes) has elapsed, and the method described above is then repeated to determine whether the fault condition persists or has subsided, for example, in the case of a temporary fault caused by heat from the consumable.

[0066] This can be implemented, for example, as follows: During the course of the method described above, for each reconnection, it is determined, for example, according to the criterion of whether a zero-current phase can be determined near the zero crossing, whether the subsequent current peaks 222, 224, 226, 227, 228 are turn-on currents or short-circuit currents. If a turn-on current event is recognized, the counter inr_cnt is incremented. If a short-circuit current event is recognized, the counter sc_cnt is incremented.

[0067] These counters are continuously compared with limit values. If inr_cnt > n, reconnection is no longer possible, as described above, where the counter inr_cnt is gradually decremented again after a pre-given time until the counter has reached its initial value, for example zero. The pre-given time can be, for example, one power grid cycle, i.e., 20 ms in the case of the European domestic power grid, or several power grid cycles up to several seconds.

[0068] Additionally, the counter sc_cnt can be compared with another limit value k, and only when sc_cnt > k is reconnection after a short-circuit current event or a fault event prohibited, such that even an event classified as a short-circuit current event does not directly lead to the determination of a fault, but rather a fault is only determined and output when more than k short-circuit current events are detected.

[0069] In this case, k = 0 can be selected, that is, the first short-circuit current event already leads to the determination of a fault situation, or k > 0 can be selected, preferably 0 ≤ k < n. For the counter sc_cnt, it can also be stipulated that the counter is gradually decremented again after a pre-given time until the counter has reached its initial value, for example zero. The pre-given time can likewise be, for example, one power grid cycle, or it can be selected to be longer than the pre-given time after which the counter inr_cnt is decremented. Preferably, the pre-given time waited before the counter sc_cnt is gradually decremented again is several seconds and in certain embodiments even several minutes.

[0070] The already explained decrease in the maximum value of the current (in which case the power semiconductor circuit is switched off) can be coupled to the value of the counter inr_cnt and / or sc_cnt in this case, for example, in such a way that the maximum value is reduced with each increment of one of these values, or in such a way that the maximum value is reduced in a few stages, for example, first reduced when inr_cnt = 2 is reached and reduced again when inr_cnt = 4 is reached.

[0071] It should be noted that, in the above description of the embodiments, the current consumption of consumers 30, 40 that were connected before other consumers 20 were connected is simply assumed to be low, and this current consumption is therefore... Figure 2 and Figure 4 This cannot be distinguished from I=0. Similarly, the period during which the rectifier-equipped consumer 20 does not consume current because the forward voltage of the rectifier diode has not been reached is simply referred to as the "zero current stage". The term "zero current stage" generally refers to the period during which current is consumed. Figure 1 The circuit shown is unaffected by the connection of the newly connected consumer 20, typically during periods when consumer 20 does not cause any or any significant additional current flow, for example because capacitor 22 has already been partially charged.

[0072] The final example should be understood as follows: if a consumer 21 with a rectifier and parasitic or effective capacitance 22 is connected in a DC circuit via a switch 23, these capacitors are charged at least partially by current pulses as already explained until they are switched off by the controller 13 and the voltage is maintained fully or partially for a short period of time until the power semiconductor circuit is switched back on at the latest at the end of the current half-wave of the voltage.

[0073] Due to this partial charging of the capacitor, after reconnection, current only begins to flow into the consumer 20 when the instantaneous value of the voltage on the load side of SCCB 10 exceeds the value of the voltage applied to rectifier 21 on the DC voltage side plus the threshold voltage of the diode located in the current path. Conversely, the zero-current phase thus derives from the fact that once the voltage defined by the previous current pulse has reached capacitor 22, no charge flow occurs as long as the instantaneous value of the voltage fed into consumer 20 is below that defined voltage. Of course, this can also be easily expressed as a phase angle or time interval via a sinusoidal relationship of the voltage distribution, as detailed below.

[0074] In certain embodiments of the invention, the defined voltage value can be at least approximately determined by measuring the voltage between load-side terminals 18A, 18B, or based on such measurement as a comparison standard. In other embodiments, the defined voltage value can be determined based on an estimate, particularly based on an estimate of the number of reconnection cycles. In this case, the typical switching behavior (e.g., current and voltage distribution and the required number of reconnection cycles) can more typically be based on the consumer 20, which can be connected to a specific AC circuit.

[0075] In embodiments of the invention, to identify short-circuit or fault conditions, the current time distribution can therefore be additionally evaluated during the desired or as defined above-defined zero-current phase, that is, immediately before and immediately after the zero-crossing of the instantaneous voltage value. In this case, the current time distribution is compared with the current time distribution. The desired current time distribution may, for example, be the current time distribution of a short circuit or an unacceptably high load, and a short-circuit or fault condition is identified when the actual current time distribution at least approximately corresponds to such a short-circuit distribution or unacceptable overload distribution.

[0076] In other embodiments, the current time distribution is compared with the expected current time distribution in the absence of short circuits or unacceptable overload conditions. This rated current time distribution can be pre-defined by configuration for a specific AC circuit and, for example, can correspond to the current time distribution of a circuit with its rated load.

[0077] Alternatively or additionally, the rated current time distribution can be selected as a current time distribution that normally exists before the first determination of a value exceeding a pre-given maximum, i.e., before the first turn-on current event or short-circuit current event 221. Here, "normal" means "normal condition" that can be determined in different ways. For example, the current time distribution between a selectable phase angle before the voltage crosses zero, e.g., -10°, and a selectable phase angle after the voltage crosses zero, e.g., +10°, can be stored as a reference for each zero crossing.

[0078] If a current-on event or a short-circuit current event is detected, the reference is used to compare the zero-current phase following the event with the reference and to determine the presence of a short-circuit current event if the deviation exceeds a defined tolerance, and otherwise, to determine the presence of a current-on event if the deviation between the current distribution and the reference is less than a defined tolerance.

[0079] The aforementioned phase angle is preferably selected such that the absolute value of the voltage is suitable for charging the (potentially pre-charged) capacitors in the newly connected consumer, that is, a certain value higher than the voltages already reached by these capacitors, which can be derived, for example, from the previous distribution of the switching process. In embodiments, particularly those with a mains frequency of 50 Hz, the zero-current phase is defined as a period of 0.5 milliseconds near the zero-crossing of the voltage, wherein this period is preferably symmetrically positioned near the zero-crossing, i.e., starting 0.25 milliseconds before the zero-crossing and ending 0.25 milliseconds after the zero-crossing. In other embodiments, or in the case of repeated re-connection, for example from the second or third repetition, this period can be chosen to be longer in order to provide a higher voltage for further charging of capacitors that have already been partially pre-charged. For example, a 1-millisecond zero-current phase can be selected, which is preferably again symmetrically positioned near the zero-crossing, i.e., starting 0.5 milliseconds before the zero-crossing based on a voltage value of approximately 50V and ending 0.5 milliseconds after the zero-crossing.

[0080] The zero-current phase can also be described by the equation: abs(I(t)) ≤ I_Lim(t). Here, I(t) represents the value of the current in the SCCB 10 determined by the current measuring device 17, I_Lim(t) represents the expected current during the zero-current phase, and abs() represents the numerical value. I_Lim(t) can be constant, i.e., independent of t. In general, I_Lim(t) describes the expected time distribution of the current in the considered time window near the zero crossing. In this case, a tolerance of, for example, 10%, 20%, 25%, or 50% of the actual expected value can be considered together.

[0081] In other embodiments, a constant limit value can be used, to which a time-dependent value proportional to the grid voltage is added, i.e., for example, I_Lim(t) = I_const + I_dyn(t), where I_const is a constant limit value and I_dyn(t) is a time-dependent value. In this case, I_dyn(t) = U(t) / R_cur can be chosen, where U(t) is the instantaneous value of the voltage and R_cur is the ohmic load currently connected, such as the ohmic resistance of ohmic consumables 30 and 40. The value determined by the voltage measuring device 12 can be readily used for U(t), or an additional load-side voltage measuring device (not shown) can be used.

[0082] In addition to or alternative to the methods described above for distinguishing between current-on events and short-circuit current events, the increase in the current time distribution can also be compared with a maximum or reference value to distinguish between current-on events and short-circuit current events. This is particularly suitable in implementations that allow multiple (e.g., n=5) repetitions of the switching process during subsequent re-switching. Capacitive loads are charged a little with each re-switching process, such that as the number of re-switching processes increases, a smaller increase in the current time distribution can be assumed, whereas the same increase can always be expected in the case of a short circuit.

[0083] In one improved approach, this consideration is extended to the shape of the current-time distribution curve between the power semiconductor circuit's re-on and the subsequent cutoff due to exceeding the maximum value. In this case, the shape of the current-time distribution curve is compared to a pre-defined fault curve shape. If the current-time distribution curve shape extends at least primarily above the fault curve shape, a fault / short-circuit event is identified. Otherwise, a turn-on current event is identified.

[0084] It should be noted that the embodiments described above can be combined with each other arbitrarily. Furthermore, it should be noted that the above specifically considers the case of a single-phase low-voltage circuit in a three-phase 400V / 50Hz power supply network. Based on the above disclosure, those skilled in the art can find variations and applications of the invention described above without inventive effort, such as applying the invention to other voltages and / or frequencies and / or to any circuit in a three-phase system, for example, for a circuit connected between two phases L1 and L2.

Claims

1. A method for controlling a power semiconductor switch (14A, 14B) of an AC circuit, the AC circuit being switchable and / or disconnectable by the power semiconductor switch, the method comprising the following steps: a) Determine the instantaneous current and instantaneous voltage values ​​of the AC circuit; b) Determine whether the instantaneous current value exceeds a pre-defined maximum value, and if so: c1) Generate a control signal for disconnecting the circuit; c2) Generate a control signal for turning on the circuit within a time interval from the generation of a control signal for turning off the circuit, wherein the time interval is less than or equal to the duration of the voltage cycle. c3) Determine whether the instantaneous current value after the circuit is turned on exceeds a pre-defined maximum value, which corresponds to or is less than a previous maximum value, and either c4a) Determine whether the current time distribution immediately before and immediately after the next zero crossing of the instantaneous voltage value at least approximately corresponds to the desired current time distribution of a short circuit, or c4b) Determine whether the current time distribution immediately before and immediately after the next zero crossing of the instantaneous voltage value does not at least approximately correspond to the desired current time distribution, and c5) If condition c3) is satisfied and condition c4a) or condition c4b) is not satisfied, then repeat steps c1), c2), c3), c4a) or c4b) and c5). d) If the number of repetitions of steps c1), c2), c3), c4a), or c4b), c5) exceeds the value n and / or if condition c4a or c4b is met, then a fault condition is identified and a control signal for disconnecting the circuit is continuously output and the method is terminated. e) If it is determined in step b) that the instantaneous current value does not exceed a pre-given maximum value or if it is determined in step c5) that condition c3 is not met, then continue to generate a control signal for turning on the circuit and continue the method with step a).

2. The method of claim 1, wherein in step c3), a control signal for turning on the circuit is generated before and / or within a time window, the time window beginning with a phase angle of -20° relative to the next zero crossing of the voltage and ending with a phase angle of +20° relative to the next zero crossing of the voltage.

3. The method according to claim 1 or 2, wherein in step c3), in addition to the instantaneous current value, an increase in the current time distribution is determined, wherein the increase is compared with a pre-given maximum value, and wherein if the increase is above the maximum value, the method is terminated, a fault condition is identified, and a control signal for disconnecting the circuit is persistently output.

4. The method of claim 3, wherein instead of the increase, the shape of the current time distribution curve is determined, the curve shape is compared with a pre-given fault curve shape, wherein if the shape of the current time distribution curve extends at least primarily above the fault curve shape, the method is terminated, the fault condition is identified, and a control signal for disconnecting the circuit is persistently output.

5. The method of claim 1 or 2, wherein the current time distribution that normally exists before the first determination of a maximum value that can be given in advance in step b) is used as the desired current time distribution.

6. The method according to claim 1 or 2, wherein the current time distribution is evaluated in a time window near the zero crossing of the instantaneous voltage value, the time window being defined such that the value of the instantaneous voltage value is below a predefined value within the time window.

7. The method according to claim 6, wherein the predefined value at least approximately corresponds to a voltage present on the DC voltage side of the rectifier (21) of the consumer (20) to be connected to the AC circuit.

8. The method of claim 6, wherein the predefined value is determined based on a measurement of the voltage at the connection terminals (18A, 18B) on the load side of the power semiconductor switch or based on an estimate for a consumer (20) typically connectable to the AC circuit.

9. The method of claim 8, wherein the estimation is an estimate based on the number of repetitions performed in steps c1), c2), and c3).

10. The method according to claim 1 or 2, wherein in step c2), an operating signal for turning on the circuit is generated such that the circuit is turned on at a predetermined or pre-given phase angle of the voltage.

11. The method of claim 10, wherein in step c2), a control signal for turning on the circuit is generated such that the circuit is turned on at a voltage phase angle of 10° to 20° prior to the next zero crossing.

12. A control circuit (13) for a power semiconductor switch (14A, 14B) for an AC circuit, the control circuit having means for performing the method according to any one of claims 1 to 11.

13. An electronic protection switch (10) for an AC circuit, the electronic protection switch having the following components: - Power semiconductor switches (14A, 14B). - Current measuring device (17) for determining the instantaneous current value of the current flowing in the AC circuit; - Voltage measuring device (12) for determining the instantaneous voltage value of the AC circuit relative to a reference potential; - The control circuit (13) according to claim 12 controls the power semiconductor switch.

14. The electronic protection switch according to claim 13, wherein the power semiconductor switch (14A, 14B) is a self-commutated power semiconductor switch for connecting and / or disconnecting the AC circuit.

15. The electronic protection switch according to claim 13 or 14, wherein the electronic protection switch additionally has an electromechanical isolation contact arranged in series with the power semiconductor switch, and means for disconnecting the isolation contact if a fault condition is detected.

Images