Switching device for a DC network and operating procedure

Abstract

A DC switching device is described in which, assuming a linearly increasing fault current, the actual flowing current is extrapolated from a measured current equalized with an inverted transfer function, its slope and a stored time offset between the measured value and the actual current, and a shutdown is effected when the extrapolated current reaches a shutdown threshold.

Description

[0001] The invention relates to a switching device for a direct voltage network, which comprises a current sensor and at least one semiconductor switch, as well as an operating method for such a switching device.

[0002] Protective elements (DC switches) are used to protect systems and lines. These measure the current and use the current to detect short circuits or overloads. To effectively protect systems and lines, high demands must be placed on current measurement, particularly with regard to dead times and bandwidth. These demands have increased significantly with the introduction of DC networks, as the time constants of DC networks are significantly lower than those of AC networks. Short-circuit currents can therefore rise to several kiloamperes within a few microseconds and cause severe damage if they cannot be detected and switched off in time. At the same time, however, it is also important for operation in a DC voltage network that such protective elements only switch off when the flowing current reaches their assigned switch-off threshold; in other words, that the switch-off does not occur too early.

[0003] The object of the invention is to provide a switching device for a DC voltage network that is improved in terms of turn-off speed. A further object is to provide an operating method for the switching device.

[0004] This object is achieved by an operating method having the features specified in claim 1. A further solution consists in the switching device having the features of claim 6.

[0005] In the operating method according to the invention for a switching device for a DC voltage network, a measured value for the current flowing in the DC voltage network is determined. The current is expediently the current flowing through the switching device and is expediently recorded in the region of the switching device.

[0006] Furthermore, the temporal change of the measured value is determined. This temporal change can also be referred to as the derivative with respect to time. It should be understood that the mathematical construct of the derivative is not determined, but rather a real value, which can only approximate the derivative.

[0007] Furthermore, a stored time offset between the measured value and the current is provided. The time offset refers to the time that elapses until a real current value is available as a measured value for evaluation.

[0008] An extrapolated current value is calculated from the measured value, the change in the measured value over time, and the offset. If the magnitude of the extrapolated current value exceeds a shutdown threshold, the switching device triggers a current shutdown.

[0009] The switching device according to the invention for a DC voltage network comprises a current sensor for determining a measured value for the current flowing in the DC voltage network. It further comprises a controller configured to record the measured value, determine a temporal change in the measured value, and provide a stored temporal offset between the measured value and the current.

[0010] The controller is further configured to determine an extrapolated current value from the measured value, the temporal change of the measured value and the offset and to trigger a shutdown of the current by the switching device if the amount of the extrapolated current value exceeds a shutdown threshold.

[0011] This advantageously creates a control and operating procedure that enables very rapid shutdown of faults in a DC network.

[0012] It was recognized that the rate of current rise in the event of a fault, i.e. the change in current over time, is essentially constant in DC networks. If a sufficiently strong increase in current is detected, the actual current flow can be determined using a previously determined time offset between the measured current values ​​and the actual current flow and the change in time. This allows a comparison of the extrapolated current value with a shutdown threshold, whereby the shutdown threshold indicates the actual current flowing at which shutdown should occur. Shutdown can then be triggered when the actual current flowing reaches the shutdown threshold, even if a current signal of the corresponding magnitude is not yet present.

[0013] This advantageously ensures that a shutdown essentially occurs when the actual flowing current reaches the shutdown threshold and not only when the corresponding measurement signal reaches this threshold.

[0014] Advantageous embodiments of the operating method and switching device according to the invention are set forth in the dependent claims. The embodiment of the independent claims can be combined with the features of one of the subclaims or, preferably, with those of several subclaims. Accordingly, the following additional features can be provided: The extrapolated current value is conveniently determined as the sum of the measured value and the product of the time offset and the change in the measured value over time. This calculation can be performed quickly and easily by a dedicated component such as an FPGA or ASIC.

[0015] It is advisable to store a number of previous measured values ​​for the current and to determine the change in the measured value over time from the measured value and the previous measured values.

[0016] The shutdown is only triggered when the magnitude of the measured value change over time exceeds a derivation threshold. This prevents false shutdowns where no fault (short circuit) is present.

[0017] The current sensor can be a magnetoresistive current sensor. These sensors are characterized by a high bandwidth. They use a material that changes its resistance in the presence of a magnetic field. This change in resistance can be based, for example, on the anisotropic magnetoresistive effect (AMR) or the magnetic tunneling resistance (TMR).

[0018] The current sensor is a coreless, open-loop sensor. This means it is constructed without any magnetic material that focuses the field lines and also does not use a feedback loop that internally regulates the magnetic field emanating from the conductor to zero and outputs the controlled variable as a measured value. Instead, the current sensor prefers to directly measure the magnetic field emitted by the conductor at a single point.

[0019] The current sensor can comprise one or more sensor elements, each of which generates a signal. The sensor elements of the current sensor can be positioned on a specially shaped conductor, so that either only the magnetic field in one direction is measured or – if two sensor elements are positioned next to each other – the differential field is measured. A differential field measurement is significantly less sensitive to external fields that do not originate from the conductor being measured.

[0020] The current sensor can comprise a conductor section with two parallel conductor segments through which the current to be measured flows. The two conductor segments share the current to be measured. It is possible for the conductor segments to carry the current to be measured in equal proportions in the same direction. For this purpose, the two conductor segments are designed, for example, as cuboids and are physically and electrically parallel. The magnetoresistive sensor element or the multiple sensor elements are arranged centrally between the spaced-apart conductor segments.

[0021] Alternatively, the conductor sections can conduct current in opposite directions. For this purpose, the conductor sections are electrically connected in series and form part of a conductor loop (U-turn).

[0022] At frequencies around 100 Hz, the proximity and skin effects begin to occur in conductors constructed in this way. These effects lead to a current flow that is no longer uniform through the conductors, resulting in distortion of the magnetic field. As a result, the magnetic field to be measured at the sensor decreases or increases with increasing frequency, depending on the conductor geometry. For example, the magnetic field, relative to a reference strength at 1 Hz, for a split conductor with parallel current flow can drop from 0 dB to as low as -6 dB between 100 Hz and 10 kHz.

[0023] An improvement in frequency response can be achieved through compensation blocks. For example, copper compensation blocks can be placed next to the actual conductors. The compensation blocks do not participate in the current flow. However, at high frequencies, a current flow is induced in them, which counteracts the other effects. This can reduce the ripple to a range of -0.25 dB to 1 dB.

[0024] In an advantageous embodiment of the invention, a stored inverted transfer function can be applied to the measurement signal. This advantageously eliminates distortion of the current signal caused by a frequency-dependent sensor response. This advantageously results in more precise values ​​for the current and, above all, more precise values ​​for the temporal change of the current.

[0025] Preferably, the current sensor signal is digitized using a low-latency analog-to-digital converter. This makes a digital signal available for further processing in a comparatively short time. This is advantageous because the actual current flowing during shutdown only fails to rise significantly above the shutdown threshold if the time offset between the actual current and the sensor signal is not greater than the time required for a short-circuit current to reach the shutdown threshold. In other words, it is advantageous if a current increase can be detected at all before the actual current reaches the shutdown threshold.

[0026] Fast processing of the subsequent steps, i.e. the determination of the temporal change of the current and the formation of the extrapolated current value, preferably takes place in an FPGA or ASIC connected to the analog-to-digital converter.

[0027] Further control elements are preferably implemented in a microcontroller connected to the FPGA. Control variables, such as the shutdown threshold and the discharge threshold, are preferably provided from the microcontroller to the FPGA. The FPGA, in turn, can transmit current measured values ​​or a signal indicating the presence of an error to the microcontroller.

[0028] Shutdown in the event of a fault can be triggered by the microcontroller or directly by the FPGA.

[0029] The invention is described and explained in more detail below with reference to the exemplary embodiments illustrated in the figures. They show schematically: Fig. 1 a DC network with subnetworks, Fig. 2 a bidirectional DC switch with a relief network and a current sensor, Fig. 3 a split conductor configuration for the current sensor, Fig. 4 the arrangement of the current sensor in the split conductor, Fig. 5 a diagram showing the course of the current flowing through the switching device and the measuring signal of the current sensor.

[0030] Fig. Figure 1 shows a schematic of a direct voltage network 10 or DC network 10. The DC network 10 includes a DC voltage source 11 that supplies the DC network 10 with a direct voltage. The DC voltage source 11 can, for example, be a rectifier connected to a supply network. However, it is also possible for the DC voltage source to be a generator in conjunction with a rectifier or, in other examples, a generator that generates a direct voltage from the outset, for example, a photovoltaic system.

[0031] The DC network 10 also includes a number of network participants 16. In the Fig. 1, these network participants 16 are all depicted as similar, but in fact, it is understood that the network participants 16 can have very different properties. For example, one or more of the network participants 16 can represent a partially capacitive or inductive load instead of a purely resistive load. For example, some of the network participants 16 can be electric motors. It is also possible for network participants 16 to be prosumers, meaning they not only act as energy consumers but can also feed electrical energy back into the DC network 10. For example, network participants 16 can contain an accumulator. Finally, some of the network participants 16 can also be or contain generators, for example, a photovoltaic system.

[0032] The Fig. The exemplary DC network 10 shown in Figure 1 comprises a first and a second subnetwork 12, 13. The first subnetwork 12 comprises the DC voltage source 11 and some of the network devices 16. The second subnetwork 13 comprises several additional network devices 16. The two subnetworks 12, 13 are connected to one another in the DC network 10 via a DC voltage switching device 14, or DC switch 14 for short. The subnetworks 12, 13 can therefore be separated from one another by the DC switch 14. Since the direction of current flow cannot be reliably determined at the DC switch 14, this is a bidirectional DC switch 14.

[0033] The DC network also includes a DC switch 14 near each of the other network nodes 16. For some applications, it would be possible to use unidirectional switching devices here. However, for the present example, it is assumed that bidirectional DC switches 14 are also used near the network nodes 16.

[0034] Fig. Figure 2 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 comprises a controller 152, which is connected to the gate terminal of the IGBT 151 and, during operation of the DC switch 14, controls it by means of control signals in the form of various voltage levels. In addition, it has a second IGBT 155, which is connected anti-serially, i.e., in series and in 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 Fig. 2 is therefore a bidirectional DC switch, meaning it can switch off a load current regardless of its direction.

[0035] The DC switch 14 further comprises a relief network 20. This has three branches, a first and a second of which are connected in parallel to each other and in parallel to the two IGBTs 151, 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.

[0036] The relief network 20 ensures that energy stored inductively in the DC network 10, for example in the line inductances, is dissipated when the DC switch 14 is switched off, without leading to such high overvoltages that there is a risk of damage to the IGBTs 151, 155. In addition to the shape of the relief network 20, which is shown in Fig. 2, there are other forms that can also perform the tasks described.

[0037] The DC switch 14 further comprises a current sensor 30, which is connected in series with the IGBTs 151, 155 and measures the current flowing in the line 101, i.e. through the DC switch 14, and provides a corresponding measured value for the controller 152.

[0038] Since the current in DC voltage networks 102 rises very quickly in the event of a short circuit, shutdown in the event of a fault must also occur very quickly. Current sensor 30 and controller 152 are designed to enable this rapid shutdown and are described below.

[0039] The Fig. 3 and Fig. 4 together show an exemplary structure for the current sensor 30 used. The current sensor 30 comprises a current conductor 31 through which the current to be measured flows. For this purpose, the current conductor 31 is expediently inserted into the line 101 to be measured, so that the current in the line 101 also flows through the current conductor 31. In this example, the current conductor 31 is divided into two electrically and physically parallel sections 32a, b of the same cross-section. The current flows in these sections in essentially equal parts and in the same direction.

[0040] The current sensor 30 further comprises a magnetoresistive component 33, which is arranged centrally on the current conductor 31 in the region of the sections 32a, b, as in Fig. 4 in cross-section. Component 33 here comprises two magnetoresistive sensor elements incorporated in an IC, but in other embodiments, it may also comprise a single sensor element.

[0041] The component 33 is connected to a low-latency analog-to-digital converter 34 (low-latency ADC), which converts the analog signal from the component 33 into a digital signal. The digital signal is provided to the controller 152. The controller 152 does not have to be a single component, but can consist of several discrete components. The processing of the digital signal from the analog-to-digital converter 34 can, for example, take place in a dedicated FPGA 35. The resulting signal can be provided as a digital value to a microcontroller 36, which is part of the controller 152, and used for overload scenarios to calculate the shutdown time, since small delays are not important here. A resulting signal for shutdown in the event of a fault can also be passed on to the microcontroller 36, or shutdown can be initiated directly from the FPGA 35.

[0042] The controller 152 is configured to process the incoming signal with an inverted transfer function, wherein the inverted transfer function is stored in the controller 152. The inverted transfer function is typically stored in the form of a set of discrete factors.

[0043] The inverted transfer function is determined in advance. It is an inverse function of the transfer function exhibited by the structure consisting of current conductor 31 and component 33. This transfer function is a function of signal attenuation or amplification as a function of frequency. Due to the skin effect and proximity effect, the transfer function deviates significantly from the ideal value of 1 at frequencies above 100 Hz.

[0044] The present transfer function can be measured for the present combination of current conductor 31 and component 33, for example via a frequency generator that impresses a defined signal as a current, whereby the output signal of sensor 30 is measured. Alternatively, a FEM simulation can be performed that simulates the magnetic field at different frequencies based on the conductor geometry. In a second step, the frequency response must be simulated using a transfer function. Any transfer function of any order can be used for this purpose. If the transfer function is known, it can be inverted, thus forming the inverted transfer function. The application of this inverted transfer function to the measured current signal is implemented in controller 152 in order to compensate for the effect of the transfer function on the signal.

[0045] Fig. Figure 5 shows the current flow curve of a test signal 31. The ramp-shaped test signal 31 increases from 0 A to 100 A with a rise time of 20 µs. Fig. 5 the resulting measurement signal 32 of the component 13. It can be seen that the test signal 31, due to its high frequency components, is reproduced by the component 13 with a significantly lower initial amplitude and only gradually approaches the actual final value.

[0046] Fig. Finally, Figure 5 shows the output signal 33 of the current sensor 10. The output signal 33 is obtained by multiplying it by the inverted transfer function. The output signal 33 corresponds very well to the test signal 31, but exhibits a time offset (delay) of, for example, 10 µs. The delay arises during signal processing in the analog-to-digital converter 34 and in the controller 152. By using a low-latency analog-to-digital converter 34 and a fast FPGA 35, this delay can be reduced but not completely eliminated.

[0047] The controller 152 now determines the temporal change (derivative) of the output signal 33. This can be done, for example, by at least temporarily storing the output signals 33. The temporal change is then calculated using the current output signal 33 and previous output signals 33 together with their timestamps or a known time interval.

[0048] The magnitude of the change over time is compared with a derivation threshold. If the derivation threshold is exceeded, the current increase is so great that a fault is likely present. In principle, this information alone could trigger a shutdown of the IGBTs 151, 155, regardless of the actual absolute current strength. In practice, however, a hierarchy regarding the shutdown may exist, according to which certain DC switches 14 should switch off before other DC switches 14, which is expressed in different thresholds for the respective shutdown. Such DC switches 14 with higher thresholds must therefore adhere to these and thus only switch off when the current, not the current increase, reaches the respective threshold.

[0049] The controller therefore determines an extrapolated current value. This advantageously utilizes the knowledge that in the event of a fault in DC networks, a nearly linear current increase occurs – at least well beyond the typically used threshold values.

[0050] This means that, in the presence of an error, the true current continues to increase with the determined temporal change in the time period corresponding to the temporal offset between the true current and the output signal 33. Since the temporal change is constant, the true current is therefore higher than the instantaneous measurement signal for the current, i.e., the output signal 33, by the product of the temporal offset and the temporal change.

[0051] The extrapolated current value I ex is determined as follows: Iex=IM+mI⋅tV

[0052] I M stands for the instantaneous measured value for the current, mI for the temporal change (derivative, slope) of the current and t v for the temporal offset. The temporal change m I is determined by measurement, but corresponds to m I = U / L, where L is the inductance present in the DC voltage network.

[0053] In the example according to Fig. 4, the current signal increases from 0 A to 100 A within 20 µs. The resulting change in current over time is therefore: mI=100 A / 20 μs=5 A / μs=5 MA / s

[0054] The exemplary time offset t v is 10 µs. This time offset depends on the specific circuit design, i.e. mainly on the hardware components used, in particular on the analog-to-digital converter 34 used and the specific design of the controller 152, for example the FPGA 35 used here. The time offset t vis determined and saved in advance and can therefore be retrieved at any time.

[0055] Such delays can also occur in the time offset t v Factors that do not directly result from signal evaluation must be taken into account. For example, a subsequent shutdown in the event of a fault still requires some time until the corresponding signal is implemented in the gate driver and the IGBTs 151, 155 actually begin to build up voltage.

[0056] In the example of Fig. 4, at a first measurement time t1 = 22 µs, which follows directly after the start of the strong current increase, an output signal 33 results, which represents a current of 10 A. The extrapolated current value is therefore: Iex=10 A+10 μs*5 A / μs=60 A

[0057] The discharge threshold was assumed to have been exceeded here. With a threshold of 100 A for triggering the shutdown, the DC switch does not yet switch off at this point.

[0058] Even at subsequent measurement times of 24 or 26 µs, an extrapolated current value that reaches 100 A is not obtained. At measurement time t2 = 30 µs, an output signal 33 is obtained, which represents a current of 50 A. The extrapolated current is then: Iex=50 A+10 μs*5 A / μs=100 A

[0059] Thus, the threshold is reached here, and shutdown is triggered at time t2 = 30 µs. The shutdown occurs at a time when the measured value (50 A) has not yet reached the shutdown threshold (100 A). The measured value would not reach the shutdown threshold until t3 = 40 µs. At this time, however, the actual current flowing is already 150 A.

[0060] The shutdown therefore occurs significantly earlier and at a time when the actual current flowing approximately reaches the shutdown threshold. Both the shutdown threshold and the discharge threshold are specified to the FPGA 35 in the controller 152 by the microcontroller 36 in this example. List of reference symbols 10 DC network 11 DC voltage source 12, 13 subnetworks 14 DC switches 16 network participants 151, 155 IGBT 152 Control 153, 156 freewheeling diode 20 Relief Network 201 Varistor 202, 204 capacitor 203, 205 Relief resistance 30 current sensor 31 conductors 32a, b sections 33 magnetoresistive component 34 low-latency analog-to-digital converters 35 FPGA 36 microcontrollers t v time offset t1, t2 measurement time

Claims

[1] Operating method for a switching device (14) for a DC voltage network (10) comprising the steps: - Determining a measured value for the current flowing in the DC voltage network (10), - Determination of the temporal change of the measured value, - Provision of a stored time offset (t v ) between the measured value and current, - Determination of an extrapolated current value from the measured value, the temporal change of the measured value and the offset (t V ), - triggering a switch-off of the current by the switching device (14) when the amount of the extrapolated current value exceeds a switch-off threshold. [2] Operating method according to claim 1, wherein the determination of the extrapolated current value is carried out as the sum of the measured value and a product of the time offset (t v ) and the temporal change of the measured value. [3] Operating method according to claim 1, wherein a stored inverted transfer function is applied to the measurement signal. [4] Operating method according to claim 1, in which a storage of a number of previous measured values ​​is carried out and the determination of the temporal change of the measured value is carried out from the measured value and the previous measured values. [5] Operating method according to claim 1, wherein the triggering of the shutdown only occurs when the temporal change of the measured value exceeds a derivative threshold. [6] Switching device (14) for a DC voltage network (10) with a current sensor (30) for determining a measured value for the current flowing in the DC voltage network (10) and a controller (152) which is designed - record the measured value, - to determine a temporal change in the measured value, - a stored time offset (t v) between the measured value and current, - an extrapolated current value from the measured value, the temporal change of the measured value and the time offset (t v ) to determine - to trigger a current cut-off by the switching device if the amount of the extrapolated current value exceeds a cut-off threshold. [7] Switching device (14) according to claim 6, wherein the current sensor (30) is a magnetoresistive current sensor (30). [8] Switching device (14) according to claim 6 or 7, wherein the current sensor (30) comprises a conductor section with parallel conductor pieces (32a, b) through which the current to be measured flows. [9] Switching device (14) according to claim 8, in which the conductor pieces (32a, b) carry the current to be measured in equal proportions in the same or opposite direction. [10] Switching device (14) according to one of claims 6 to 9 with a low latency analog-digital converter.

Images